Trace vapour detection at room temperature using Raman spectroscopy.

Analyst View Article Online

Published on 24 January 2014. Downloaded by Northeastern University on 04/10/2014 10:22:03.

PAPER

Cite this: Analyst, 2014, 139, 1960

View Journal | View Issue

Trace vapour detection at room temperature using Raman spectroscopy Alison Chou,*a Babak Radi,b Esa Jaatinen,b Saulius Juodkazisc and Peter M. Fredericksb A miniaturized flow-through system consisting of a gold coated silicon substrate based on enhanced Raman spectroscopy has been used to study the detection of vapour from model explosive compounds. The measurements show that the detectability of the vapour molecules at room temperature depends sensitively on the interaction between the molecule and the substrate. The results highlight the capability

Received 12th August 2013 Accepted 23rd January 2014

of a flow system combined with Raman spectroscopy for detecting low vapour pressure compounds with a limit of detection of 0.2 ppb as demonstrated by the detection of bis(2-ethylhexyl)phthalate, a

DOI: 10.1039/c3an01522j

common polymer additive emitted from a commercial polyvinyl chloride (PVC) tubing at room

www.rsc.org/analyst

temperature.

Introduction Two key considerations in developing vapour sensing systems have been lowering the concentration of vapour required for a positive detection, as well as making the overall system as compact as possible to aid portability. Detection of vapours emitted from a hazardous solid or liquid is the preferable mode of operation as it does not require contacting the surface of the sample. However, the concentration of vapour depends on the vapour pressure of the compound and in still air at room temperature, most explosives have vapour concentrations in the range of parts per billion (ppb) to parts per trillion (ppt) and even lower for mixtures of explosive materials.1 Capturing enough explosive material vapour in the air for detection and analysis is thus difficult. Until the discovery of Raman scattering enhancement of pyridine adsorbed on roughened silver electrodes2,3 and silver nanoparticles4 in the 1970s, the application of Raman spectroscopy in chemical sensors had been constrained by the limits of detection due to small fractions of the inelastic Raman scattering process, which typically has cross-sections of the order of 10 30 to 10 25 cm2 per molecule.5 Since then, the amplication of Raman scattering signals by metal nanostructures, which is now referred to as surface enhanced Raman spectroscopy (SERS), has been pursued with great interest because of its potential to a

Central Analytical Research Facility (CARF), Institute for Future Environments Queensland University of Technology, Australia. E-mail: [email protected]; Tel: +61 7 3138 9403

b

School of Chemistry, Physics and Mechanical Engineering Faculty of Science and Engineering, Queensland University of Technology, Australia. Tel: +61 7 31382275

c Centre for Micro-Photonics, Faculty of Engineering and Industrial Sciences, Swinburne University of Technology, Australia. E-mail: [email protected]; Fax: +61 3 9214 5435; Tel: +61 3 9214 8718

1960 | Analyst, 2014, 139, 1960–1966

substantially lower detection limits. The signal enhancement relies on the optical properties of the metal nanostructures with the mechanism depending on the surface roughness and the geometry of the experiment setup. Generally, the mechanism has been attributed to surface plasmon,6 charge transfer,7 and resonant light scattering.8 Available data in the literature on SERS show many of the most sensitive methods use silver9,10 or gold nanoparticles11–14 in their vapour or gas sensing scheme for signal enhancement with detection limits at the picomolar level possible.15 More recently developed SERS ow-through vapour detection methods report detection limits of 0.5 ppb and 1 ppb as demonstrated by Oo et al.16 and Piorek et al.17 respectively. Oo et al. used a ow through detection system where vapours of pyridine and 4-nitrophenol are passed through capillaries of various inner diameters ranging from 4.7 to 700 mm lined with gold nanoparticles on the inner walls of the capillaries. Piorek et al. demonstrated the detection of 2,4-dinitrotoluene (2,4-DNT) vapour using a microuidic ow system lled with silver colloid in aqueous solution. Analyte vapours ow over the microuidic channel and diffuse into the aqueous phase therein interacting with the silver colloid followed by detection with a Raman spectrophotometer. A disadvantage of both methods is that they require synthesis of nanoparticle systems with appropriate dimensions and specic optical properties in order to achieve their detection objectives. This work describes a detection method that requires only a silicon substrate coated with a fresh sputtered gold lm, without the complexity of incorporating nanoparticles into a polymer layer or a separate step of nanoparticle synthesis. Our ow system is constructed by adapting a modied approach based on our recently published static test of nitro compound vapour detection using Raman spectroscopy.18 The modied detection system is an automated recirculating ow system that

This journal is © The Royal Society of Chemistry 2014

View Article Online


Paper

Analyst

delivers vapours to the sensor surface allowing real-time detection. Vapours from sublimation of solid analytes at room temperature ow over a 5 mm2 square silicon substrate coated with sputtered gold used for capturing vapour molecules and detection using a commercially available Raman spectrophotometer. The vibrations of the molecules on the substrate surface produce a spectrum that allows the method to detect and distinguish compounds at concentrations of a few parts per million at room temperature. The detected concentration of the analyte was estimated from the vapour pressure of the analyte at room temperature. The main objectives of this paper were to verify the feasibility of our proposed detection method and to assess the differences of dynamic versus static measurements. Surface fouling of gold substrates which could pose a limitation for the SERS based sensor using gold is also discussed. Experiments were carried out using six explosive model compounds with vapour pressures ranging between 2 10 4 and 0.1 mmHg at 25 C. To demonstrate the trace detection capability of the ow system, a commercial sample of polyvinyl chloride (PVC) tubing containing bis(2-ethylhexyl)phthalate (DEHP, or DOP) was used as an analyte. DEHP is a commonly used plasticiser for PVC consumables. When incorporated in PVC, it can escape by evapouration or extraction by contact with liquids because it is not chemically bonded to the PVC.19 This is due to the polar interactions

Chemical structure and vapour pressure (at 25 C) of tested compounds Table 1

Compound (abbreviation)

Vapour pressure at 25 C mm 1 Hg

Bis(2-ethylhexyl)phthalate (DEHP)

1.43 10

3-Dinitrobenzene (1,3-DNB)

2 10

1-Chloro-2,4-dinitrobenzene (CNDB)

0.001b

Chemical structure

7a

between the ester groups and the aromatic ring of the phthalate molecule and the positively charged areas of the vinyl chain.20 Pure DEHP has a vapour pressure of 1.43 10 7 mmHg at 25 C.21 It is expected that the vapour pressure of the DEHP emitted from DEHP incorporated PVC would be lower than that of pure DEHP.22 The measurement of DEHP is also of concern because adverse effects on human health have been expressed even for low level exposure.23,24 The chemical structures and vapour pressures for the tested compounds are given in Table 1. The test compounds were selected to represent a range of different vapour pressures and functional groups in order to test the limit of detection (LOD) and the selectivity of the sensor.

Experimental section Materials 1,3-Dinitrotoluene, 1-chloro-2,4-dinitrobenzene (97% purity), 2,4dinitrotoluene (97% purity), 2-naphthalenethiol (99% purity), 4methyl-2-nitrophenol (99% purity), and 4-nitrotoluene were purchased from Sigma-Aldrich, Australia. Tygon® formula R-3603 laboratory tubing (containing DEHP phthalate as per product description) was purchased from Sigma-Aldrich, Australia. Sputtered-gold lm preparation Gold (99.99%) was sputtered onto clean silicon substrates using a Leica EM SCD005 sputter coater under argon. The vacuum chamber was evacuated to 0.05 mbar and the sputter coating attachment was charged at 30 mA with an argon carrier gas. The silicon substrates were placed 50 mm from the gold source. The sputtering time was 150 s. This results in a gold coating of approximately 30 nm with a root-mean-square surface roughness of 5.5 nm measured by atomic force microscopy (AFM). The gold sputter target was purchased from ProSciTech. Atomic force microscopy measurements

4b

Sputtered gold lms were imaged at room temperature under an ambient atmosphere with an atomic force microscope (NTMDT) in contact mode. The cantilever used was a silicon cantilever with platinum coating (NT-MDT, CSG11/Pt/20). Raman spectroscopy

b

Raman measurements were carried out at room temperature using a Renishaw inVia Raman Microscope. All spectra were acquired with a 50 magnication objective, 10 s exposure of 20 mW of 785 nm excitation and an acquisition time of 30 s.

2,4-Dinitrotoluene (2,4-DNT)

0.002

2-Naphthalenethiol

0.005b

4-Methyl-2-nitrophenol

0.04b

XPS spectra were acquired using a Kratos Axis Ultra photoelectron spectrometer which uses Al Ka (1253.6 eV) X-rays with a pass energy of 20 eV and a spot size of 700 300 mm.

4-Nitrotoluene

0.1b

Vapour detection test

X-ray photospectroscopy analysis

a

Ref. 19. b Royal Chemical Society chemical structure database.


Fig. 1 shows the schematics of the ow-through system. 1 mg of the compound to be tested is transferred into a 50 ml aluminium sample pan and placed inside the analyte chamber. The analyte

Analyst, 2014, 139, 1960–1966 | 1961

View Article Online


Analyst

Schematic diagram of the flow setup. Flow-through system used in conjunction with the inVia Renishaw Raman microscope for vapour detection on a gold coated silicon substrate. Vapours from analyte sublimation inside the sample chamber are delivered to the substrate surface using a micropump and then excited with a 785 nm laser. The sensing information is obtained from the measured spectrum of the inelastically scattered photons due to molecular vibrations. The detection axis is the same as the laser beam axis, y.

Paper

Fig. 1

chamber follows the basic cyclone design using the benet of swirl ow.25 The swirling ow eld inside the chamber is created by the addition of baffles on the inside of the chamber walls creating a turbulent ow which makes the vapour concentration constant and uniform inside the system. The system was ushed with dry nitrogen prior to use. The carrier gas (nitrogen) stream enters from the bottom of the chamber, spiraling upwards and outows from the top of the chamber. As the nitrogen carrier gas ow is adjacent to the solid analyte surface, sublimation occurs and the mass is transferred to the carrier gas stream. Thus the cyclone design ensures that the saturation concentration is reached faster. The substrate onto which the vapour molecules are captured for detection consists of a 30 nm thick freshly sputtered gold lm on a silicon wafer (5 5 mm). A micropump (RS Components) with a ow rate of 200 ml min 1 is used for delivering the analyte to the substrate surface. The substrate chamber consists of a cell incorporating a glass window and is placed directly under an objective lens of a Renishaw inVia Raman microscope through which the laser beam is focused on the substrate and scattered light is collected for detection. System evaluation was carried out by measuring the response of the gold substrate inside the analyte chamber aer exposure to the analyte vapour at room temperature.

Results and discussion Fig. 2 shows the AFM images of the gold coated silicon substrate. The morphology of the sputtered lm on silicon appears very similar to that of selective chemical etching of Ag from Ag/Au alloy lms on planar substrates reported by Dixon et al.26 The lm does not appear homogeneous with clusters of round features of approximately 50 nm in diameter. This is typical of sputtered gold lms as gold atoms tend to cluster into islands minimizing the overall surface energy.27 Fig. 3a shows the Raman spectra for the gold coated silicon substrate with 785 nm excitation aer exposure to 2,4-DNT at room temperature inside the ow system. The ow rate of the

1962 | Analyst, 2014, 139, 1960–1966

Fig. 2 Contact mode AFM image of the sputter-deposited gold film on silicon used for the sensor in two different scan sizes: (a) and (b). (c) and (d) are line scans corresponding to the blue line in the AFM images above, showing the height of the structures from the sputtering process.

nitrogen carrier gas inside the system was 200 ml min 1 and the intensity of the signal was found to be independent of the ow rate range tested (100–380 mL min 1), but could be improved if the measurement was taken 1 min aer turning the pump off. This result is in agreement with the static measurement performed separately shown in Fig. 3b. The fact that the signal from the dynamic measurement is weaker and noisier is attributed to the interaction between the molecule and the gold surface. When the pump is switched on, the molecules are being swept over the surface of the substrate due to its motion in the ow, thus limiting the probability of the vapour molecules colliding with the substrate for adsorption and detection. In the static measurement, the mobility of the vapour molecules is reduced and there is sufficient time for adsorption to occur hence producing a detectable signal. The responses of the other nitro analytes tested in the ow system are shown in Fig. 3b–e. Results were recorded with the pump turned off aer recirculating the vapour for 5 mins at room temperature. The intense peak at 1380 cm 1 seen in all four spectra is attributed to the symmetric stretching of the nitro group (–NO2) and it reects the interaction between the nitro group and the gold surface. It has been shown by electron energy loss spectroscopy and temperature programmed desorption that NO2 adsorbs on the surface of gold through its oxygen atoms to form a bidentate surface chelate.28 Upon adsorption, the valence electrons of the oxygen in the NO2 group change the metal surface electron distribution close to the highest occupied molecular orbital in the valence band. The degeneracy at the surface is lied by interaction of these lobes causing the band to be more intense than other bands in the spectrum. The broadness of the peak suggests intermolecular interactions between the DNT molecules on the gold surface. For 2-naphthalenethiol (Fig. 3f), there is no difference between the static and the dynamic measurement in terms of


View Article Online


Paper

Fig. 3 Raman spectra of six different compounds detected in the flow system on a sputtered gold coated silicon substrate at room temperature. Dynamic response (blue) was recorded with the pump turned on at a flow rate of 200 ml min 1. Static response was recorded with the pump turned off after recirculating the vapour for 5 mins at room temperature. Spectra of the solid analyte (black) are also shown for comparison. (a) 2,4-Dinitrotoluene, (b) 1,3-dinitrobenzene, (c) 1-chloro-2,4-dinitrobenzene, (d) 4-methyl-2-nitrophenol, (e) 4-nitrotoluene, and (f) 2-naphthalenethiol.

signal strength. All bands in the dynamic and static spectra appear very narrow and sharp. The position of the bands for the adsorbed 2-napthalenethiol closely matches that for the pure solid. This means that the vibrational frequencies that are most visible in the spectra are not affected much by surface binding. Noticeable differences between the sensor response spectrum and the reference spectrum are the change in intensities for the ring deformation modes at 767, 514 and 354 cm 1.29 Probable adsorbed conformation of 2-naphthalenethiol on the gold surface is with the aromatic rings perpendicular to the metal surface29 as the attachment of 2napthalenethiol occurs via the Au–S bond. This covalent bond formation causes localized charge uctuations at the metal– molecular interface inducing a series of dipoles,30 probably the reason why the bands are so intense in comparison with 2,4dinitrotoluene. Similar observations have also been made by Hill and Wehling.31 They compared different molecules adsorbed on the silver and gold surface and reported that pmercaptoaniline gave signicantly larger intensities than other non-thiol molecules tested. The sharpness of the bands probably points at relatively low inhomogeneous broadening, since the thiol group binds strongly to gold with the sulfur–gold


Analyst

interaction being the predominant mode. The sharpness of the Raman bands of thiol molecules has also been explained by the effective resonant model proposed by Ueba8 that the electronic resonant level of the adsorbed molecule shis towards lower energies and the effective resonant condition is realised even when the molecule is excited by the off-resonant energy of the isolated molecule. The detection of DEHP emitted from a sample of Tygon® tubing is shown in Fig. 4a. Characteristic bands for phthalate esters at 1040, 1581, and 1600 cm 1 are associated with the various vibrations of the ortho-phenyl group32 and the band at 1726 cm 1 is assigned to the carbonyl C]O stretching vibration of the ester group. Measurement of DEHP from commercial PVC products by Uhde et al.33 reported that, in general, DEHP emission from wallpaper measured in a CIS/ GC-MS system results in a maximum concentration of approximately 1 mg m 3 under ambient conditions. Thus this example shows that the LOD of our technique is at the level of sub-ppb, comparable with the detection level reported by Oo et al. and Piorek et al., demonstrated using pyridine and 4nitrophenol. Further investigation to conrm DEHP on the gold surface was performed in a 0.75 ml stainless-steel sealed cell containing a sample of the PVC tubing and a gold coated silicon substrate. The cell was kept at 50 C for 17 h to increase the DEHP concentration. As can be seen from Fig. 4a, the bands become more pronounced and the entire spectrum is a closer match to the spectrum of the PVC tubing than that obtained inside the ow system due to higher concentration of DEHP at higher temperature. The uncoated silicon (Fig. 4b) substrate shows no Raman bands that correspond to DEHP except for the main silicon band at 521 cm 1, a weak peak at 308 cm 1 corresponding to Si crystals,34 and another weak broad band at 900–1100 cm 1 is the two phonon overtone of silicon.35 All compounds tested gave identiable Raman signal with some showing more spectral difference in both relative intensity and band positions compared to the solid analyte. Most noticeably, 2-naphthalenethiol has a lower vapour pressure than 4-methyl-2-nitrophenol and 4-nitrotoluene, but out of all compounds tested, 2-naphthalenethiol gave the strongest signal on gold with its Raman spectrum closely resembling the

Fig. 4 Raman spectra of (a) the gold coated silicon substrate after exposure to a sample of Tygon® tubing inside the flow system at room temperature and a sealed cell at 50 C. A spectrum of the Tygon® tubing and pure PVC powder are also shown for comparison. * denotes peaks due to the silicon underneath the gold. (b) An uncoated silicon substrate after exposure to DEHP at 50 C for 17 h.

Analyst, 2014, 139, 1960–1966 | 1963

View Article Online


Analyst

Paper

spectrum recorded of the material alone. Furthermore, the static and the dynamic spectra for 2-napthalenethiol are almost the same; suggesting that a ow rate of 200 ml min 1 has no effect on the adsorption capacity of 2-naphthalenethiol. In contrast, the inuence of the ow is more signicant in the case of nitro analytes. Thus the LOD of the sensor could also be related to the molecule–gold interaction energies as well as the vapour pressure of the solid sample. Table 2 presents the bond enthalpy between the different functional groups and gold substrate for the three types of compounds studied in this work. From the table, 2-naphthalenethiol forms the strongest bond with gold compared to other compounds tested. In general, sulfur containing compounds have strong specic interaction with gold and tend to chemisorb onto its surface resulting in a metal–S bond formation.36 Estimated using alkanethiol adsorption, this bond energy is approximately 167 kJ mol 1.37 For the nitro analytes, the bonding between the nitro group and the gold surface is dative through the interaction between the oxygens on the nitro groups and the gold surface. This interaction has been estimated to be 59 kJ mol 1.38 As for DEPH, the adsorption is a physisorption through the aromatic p electrons and the gold surface,39 this interaction energy is only about 25 kJ mol 1.40 van der Waals interaction between the ester group of the DEHP and the gold can also occur. Of all compounds tested, DEHP has the lowest vapour pressure at room temperature and the weakest interaction with gold. The detected DEHP Raman intensity was also the lowest amongst all compounds tested. This result suggests that the interaction between the molecule and the gold substrate could be a critical parameter in detectability as well as the vapour pressure of the compound. In any case, the reference spectra were used as a guide only and exact match of the band positions were not expected. This is because vibrations of a molecule depend on its physical state and Raman scattering is proportional to the polarizability of the molecule.8 While the detected signals of molecules are due to the adsorbed state, the references are of the pure material in the solid state. When a molecule is adsorbed on a surface, a new species is formed (molecule–metal complex). Hence a change in the electronic and the vibrational states is expected due to the rearrangement of the electron conguration of the molecule.41

The resulting physical interaction between the molecule and the gold substrate, the electromagnetic eld of the incident light is also modied by the surface in magnitude, direction, and spatial dependence.42 Substrate contamination and storage It has been observed that within minutes of exposure to the laboratory atmosphere, a clean hydrophilic gold surface can be rendered hydrophobic by adsorption of nonpolar contaminants.43–47 The storage condition of the substrates is thus crucial as the reliability of the sensor could be compromised by surface fouling. Preserving the activity of the substrate for practical use was considered by studying different storage environments. In a static test, a signicant decrease in signal intensity was observed for the air-exposed substrate, which showed a 93% reduction in signal intensity. Identiable signal was detected with the substrates stored in a nitrogen atmosphere at room temperature and 20 C (Fig. 5). The decrease in the Raman intensity for the air exposed substrate indicates that the detection mechanism is sensitive to the surface chemistry of gold. While bulk gold is generally considered inert, nano-sized gold can form surface intermediates of moderate stability48 thus causing unnecessary background which may interfere with signal readouts. A sputtered gold lm on silicon was investigated by X-ray photoelectron spectroscopy within 2 hours of lm preparation. The sample had been stored inside a 15 ml glass vial immediately aer retrieving the sample from the sputter coater. Detectable oxygen and carbon species suggest that organic contamination from the air inside the vial is evident in Fig. 6. The O 1s signal at 532 eV is assigned to oxygen adsorbed on the gold surface.49 The value at 533 eV can be attributed to Au(OH)3 or Au2O350,51 as gold can be readily oxidized under ambient conditions.52 However, for the majority of all known oxides or hydroxides, the binding energy of the oxygen 1s electron varies between 525 and 531 eV,53 thus the oxygen peak

Energetics of adsorption of different functional groups on the gold surface

Table 2

Molecule

Nature of bonding

Interaction

Bond enthalpy DH, kJ mol 1

2-Naphthalenethiol Chemisorption Sulfur–gold 167a Di(2-ethylhexyl)Physisorption p electrons–gold 25b phthalate van der Waals ester–gold (through charge transfer coordination) 2,4-Dinitrotoluene Dative Nitro-gold 59c and other bonding nitroaromatics tested a

Ref. 37. b Ref. 40. c Ref. 38.

1964 | Analyst, 2014, 139, 1960–1966

The effect of the storage condition on the Raman spectrum of 2,4-DNT on Au/Si substrates. In each case, the gold film was freshly prepared and stored under the specified conditions for 24 h before exposing the substrate to 2,4-DNT vapours at room temperature.

Fig. 5


View Article Online

Paper

Analyst

discussions regarding Raman spectroscopy, Dr Barry Wood for XPS analysis, QUT Central Analytical Research Facility, QUT electrical workshop. This work was supported by the Australian Research Council, Linkage Project LP0882614: a new nanosensor technology for the detection and identication of residual vapour explosives, drugs and chemicals in the air.


References Fig. 6 High-resolution XPS spectra of (a) the O 1s and (b) the C 1s photoelectrons of a gold coated silicon substrate. The spectra were acquired with a pass energy of 20 eV and a spot size of 700 300 mm.

at 530 eV provides indication but not conclusive evidence of the presence of gold oxide. To be sure, there should also be a peak at 85.7 eV in the gold 4f for the gold oxide. This was not observed, probably there was not enough gold oxide in the samples to be detected. Further evidence suggesting contamination is the Raman spectrum of the gold substrate without exposure to analytes shown in Fig. 3a. Bands at 1037, 982, 621, 388 and 289 cm 1 are typical of carbonate54,55 and sulfate56 species. Conveniently, these background signals do not interfere signicantly with the detection of the selected analyte and can be distinguished from the signal response spectrum. Nevertheless, we recognize the importance of the cleanliness of the gold surface for obtaining unambiguous signals.

Conclusions This paper presents a vapour detection method using Raman spectroscopy. Differences in spectra were found between the static and the dynamic test measurements with the static measurement giving more pronounced bands. We examined the effect on the Raman spectrum of three different bonding types with gold: van der Waals, dative and covalent bonding by comparing the quality of the spectrum obtained from three different types of compounds selected. As well as the vapour pressure of the compound, the detection of the vapour molecules on the surface of the gold coated silicon substrate appears to be dependent on the net physicochemical force which consists of: (1) the interaction strength between the molecule and the gold surface and (2) the dynamics in the ow system. The technique is suitable for detection of molecules that bear functional groups that will interact favourably with gold with a LOD of parts per million or greater for the nitro analytes tested and parts per billion for the DEHP emitted from PVC tubing. The current experiments were conducted within 1– 5 mins because the gold lm quickly ages when exposed to atmospheric conditions. This could be overcome by storing the substrates under an inert atmosphere at low temperatures or vacuum until use as demonstrated in this work.

Acknowledgements The authors would like to thank Mr Armin Leibhardt for fabricating the ow-through cells, Dr Llew Rintoul for helpful This journal is © The Royal Society of Chemistry 2014

1 B. C. Dionne, D. P. Rounbehler, E. K. Achter, J. R. Hobbs and D. H. Fine, J. Energ. Mater., 1986, 4, 447–472. 2 M. Fleischmann, P. J. Hendra and A. McQuillan, Chem. Phys. Lett., 1974, 26, 163. 3 D. L. Jeanmaire and R. P. Van Duyne, J. Electroanal. Chem. Interfacial Electrochem., 1977, 1–20. 4 M. G. Albrecht and J. A. Creighton, J. Am. Chem. Soc., 1977, 99, 5215–5217. 5 D. A. Long, The Raman Effect: A Unied Treatment of the Theory of Raman Scattering by Molecules, John Wiley & Sons, Ltd, West Sussex, England, 2001. 6 M. Moskovits, Notes Record Roy. Soc. Lond., 2012, 66, 195– 203. 7 J. R. Lombardi, R. L. Birke, T. H. Lu and J. Xu, J. Chem. Phys., 1986, 84, 4174–4180. 8 H. Ueba, J. Chem. Phys., 1980, 73, 725–732. 9 T. Vo-Dinh and D. L. Stokes, Field Anal. Chem. Technol., 1999, 3, 346–356. 10 D. A. Stuart, K. B. Biggs and R. P. Van Duyne, Analyst, 2006, 131, 568–572. 11 J. Bowen, L. J. Noe, B. P. Sullivan, K. Morris, V. Martin and G. Donnelly, Appl. Spectrosc., 2003, 57, 906–914. 12 M. K. K. Oo, C. F. Chang, Y. Z. Sun and X. D. Fan, Analyst, 2011, 136, 2811–2817. 13 J. M. Sylvia, J. A. Janni, J. D. Klein and K. M. Spencer, Anal. Chem., 2000, 72, 5834–5840. 14 J. Wang, L. L. Yang, S. Boriskina, B. Yan and B. M. Reinhard, Anal. Chem., 2011, 83, 2243–2249. 15 K. Kneipp, Y. Wang, R. R. Dasari, M. S. Feld, B. D. Gilbert, J. Janni and J. I. Steinfeld, Spectrochim. Acta, Part A, 1995, 51, 2171–2175. 16 M. K. K. Oo, Y. B. Guo, K. Reddy, J. Liu and X. D. Fan, Anal. Chem., 2012, 84, 3376–3381. 17 B. D. Piorek, S. J. Lee, M. Moskovits and C. D. Meinhart, Anal. Chem., 2012, 84, 9700–9705. 18 A. Chou, E. Jaatinen, R. Buividas, G. Seniutinas, S. Juodkazis, E. L. Izake and P. M. Fredericks, Nanoscale, 2012, 4, 7419–7424. 19 S. Vainiotalo and P. Pfaffli, Ann. Occup. Hyg., 1990, 34, 585– 590. 20 A. S. Wilson, Plasticisers: principles and practice, The Institute of Materials, London, 1995. 21 D. A. Hinckley, T. F. Bidleman, W. T. Foreman and J. R. Tuschall, J. Chem. Eng. Data, 1990, 35, 232–237. 22 A. Afshari, L. Gunnarsen, P. A. Clausen and V. Hansen, Indoor Air, 2004, 14, 120–128. 23 S. Sathyanarayana, C. J. Karr, P. Lozano, E. Brown, A. M. Calafat, F. Liu and S. H. Swan, Pediatrics, 2008, 121, E260–E268.

Analyst, 2014, 139, 1960–1966 | 1965

View Article Online


Analyst

24 N. Hallmark, M. Walker, C. McKinnell, I. K. Mahood, H. Scott, R. Bayne, S. Coutts, R. A. Anderson, I. Greig, K. Morris and R. M. Sharpe, Environ. Health Perspect., 2007, 115, 390–396. 25 A. K. Gupta, D. C. Lilley and N. Syred, Swirl Flow, Abaacs Press, Tunbridge Wells, England, 1984. 26 M. C. Dixon, T. A. Daniel, M. Hieda, D. M. Smilgies, M. H. W. Chan and D. L. Allara, Langmuir, 2007, 23, 2414–2422. 27 C. Schug, S. Schempp, P. Lamparter and S. Steeb, Surf. Interface Anal., 1999, 27, 670–677. 28 M. E. Bartram and B. E. Koel, Surf. Sci., 1989, 213, 137–156. 29 R. A. Alvarez-Puebla, D. S. Dos Santos and R. F. Aroca, Analyst, 2004, 129, 1251–1256. 30 G. Heimel, L. Romaner, J. L. Bredas and E. Zojer, Phys. Rev. Lett., 2006, 96, 196806. 31 W. Hill and B. Wehling, J. Phys. Chem., 1993, 97, 9451–9455. 32 R. A. Nyquist, Appl. Spectrosc., 1971, 26, 81–85. 33 E. Uhde, M. Bednarek, F. Fuhrmann and T. Salthammer, Indoor Air-International Journal of Indoor Air Quality and Climate, 2001, 11, 150–155. 34 Y. Kanzawa, S. Hayashi and K. Yamamoto, J. Phys.: Condens. Mattert, 1996, 8, 4823–4835. 35 P. A. Temple and C. E. Hathway, Phys. Rev. B: Solid State, 1973, 7, 3685. 36 R. G. Nuzzo, F. A. Fusco and D. L. Allara, J. Am. Chem. Soc., 1987, 109, 2358–2368. 37 R. G. Nuzzo, L. H. Dubois and D. L. Allara, J. Am. Chem. Soc., 1990, 112, 558–569. 38 J. Wang and B. E. Koel, J. Phys. Chem. A, 1998, 102, 8573– 8579.

1966 | Analyst, 2014, 139, 1960–1966

Paper

39 C. Woll, J. Synchrotron Radiat., 2001, 8, 129–135. 40 H. Dahms and M. Green, J. Electrochem. Soc., 1963, 110, 1075–1080. 41 A. Campion and P. Kambhampati, Chem. Soc. Rev., 1998, 27, 241–250. 42 S. Efrima, J. Chem. Phys., 1985, 83, 1356–1362. 43 E. H. Loeser, W. D. Harkins and S. B. Twiss, J. Phys. Chem., 1953, 57, 251–254. 44 G. L. Gaines, J. Colloid Interface Sci., 1981, 79, 295. 45 M. Schneegans and E. Menzel, J. Colloid Interface Sci., 1982, 88, 97–99. 46 T. Smith, J. Colloid Interface Sci., 1980, 75, 51–55. 47 C. D. Bain, E. B. Troughton, Y. T. Tao, J. Evall, G. M. Whitesides and R. G. Nuzzo, J. Am. Chem. Soc., 1989, 111, 321–335. 48 A. Haruta, Chem. Rec., 2003, 3, 75–87. 49 M. Peuckert, J. Phys. Chem., 1985, 89, 2481–2486. 50 M. Peuckert, F. P. Coenen and H. P. Bonzel, Surf. Sci., 1984, 141, 515–532. 51 T. Dickinson, A. F. Povey and P. M. A. Sherwoond, J. Chem. Soc., Faraday Trans. 1, 1975, 71, 298–311. 52 M. White, J. Phys. Chem., 1964, 68, 3083–3085. 53 C. D. Wagner, D. A. Zatko and R. H. Raymond, Anal. Chem., 1980, 52, 1445–1451. 54 I. Pockrand and A. Otto, Appl. Surf. Sci., 1980, 6, 362– 371. 55 R. P. Cooney, M. R. Mahoney and M. W. Howard, Chem. Phys. Lett., 1980, 76, 448–452. 56 K. H. Fung and I. N. Tang, J. Colloid Interface Sci., 1989, 130, 219–224.


View Article Online

Analyst


PAPER


Methanol-induced conformation transition of gland fibroin monitored by FTIR spectroscopy and terahertz spectroscopy Chao Yan,a Bin Yang*a and Zhicheng Yuab Silk fibroin extracted from the gland of Bombyx mori silkworm is employed as an ideal system to investigate its conformation transition in methanol/D2O solution. The transition process was monitored by Fourier transform infrared (FTIR) spectroscopy coupled with terahertz time domain spectroscopy (THz-TDS). Analysis of FTIR spectra suggests that, with increasing time of treatment, an increasing band at 1634 cm1 is observed indicating the formation of b-pleated sheets coincident with the loss of intensity of a band at 1673 cm1 indicating decrease of the random coil structure. In addition, there is a burst phase of 33% occurring during the first 2 minutes when the gland fibroin membranes are immersed into methanol/D2O solution. THz spectra present distinct features for conformations of silk fibroin, in combination with the results obtained from FTIR; the peaks observed at 1.54 THz (51 cm1), 1.67 THz (55 cm1), and 1.84 THz (61 cm1) can be attributed to a b-pleated sheet, a-helix, and random coil, respectively. Intensity change of bands centered at 1.54 THz and 1.84 THz confirms the formation of the

Received 14th August 2013 Accepted 27th January 2014

b-pleated sheet and the disappearance of the random coil. Kinetic curves obtained from THz spectra indicate that the methanol-induced conformation transition from the random coil to the b-pleated sheet is fitted with an exponential function. The results suggest that THz-TDS presents great potential as a

DOI: 10.1039/c3an01547e

complementary approach in studying the secondary structure of a protein, providing significant insight

www.rsc.org/analyst

into the silk-spinning process in vivo.

1. Introduction Silk has recently emerged as a highly promising biopolymer owing to its excellent mechanical1 and optical properties,2 as well as biocompatibility.3–7 It is remarkable that silk laments can be formed by Bombyx mori silkworms at room temperature, with water as a solvent and without high hydrostatic pressure. Thus, it is important to know the structure of the silk broin before and aer spinning, to elucidate the mechanism of ber formation and design new silk-like materials that can be produced under similar mild conditions. Bombyx mori heavy chain broin, a major structural protein of silk, is primarily composed of repeats of the motif (Gly-Ala-GlyAla-Gly-Ser)n. Within the lumen of the posterior and middle divisions of the gland, the heavy chain broin is thought to be present as a silk I-like conformation lacking an ordered secondary structure and largely containing extended chains repeatedly folded back on themselves. During natural spinning, this conformation is thought to be converted into silk II, which is largely

a

Key Laboratory of Advanced Textile Materials and Manufacturing Technology, Zhejiang Sci-Tech University, Hangzhou, Zhejiang, China. E-mail: yangbin5959@ yahoo.com.cn

b

Ecological Engineering Research Center of Dyeing and Finishing Technology, Hangzhou, Zhejiang, China


formed from crystalline b-sheets. These b-sheets are well orientated along the axis of the ber except for the short disordered regions, giving it a strength and stiffness comparable or even superior to that of high-performance synthetic materials. Thus, establishing effective methods to determine broin conformation is a fundamental problem for understanding the mechanism of the change in conformation as water-soluble Bombyx mori broin in silk glands is spun into a solid cocoon ber. Various other techniques have been demonstrated as available approaches to characterize protein conformation, including X-ray crystallography or multidimensional nuclear magnetic resonance (NMR) spectroscopy.8–12 However, highresolution studies of proteins are not always feasible. For example, crystallographic studies require high-quality single crystals which for many proteins are not available. Furthermore, the question arises as to whether the relatively “static” structure in single crystals adequately represents the protein conformation in a complex and dynamic environment. NMR offers a somewhat better exibility in studying protein structures in “biologically relevant” environments. However, the interpretation of NMR spectra of larger proteins is very complex, and the assignment of interproton distances generated by the NMR experiment is not always feasible. To date, this technique has been restricted to small proteins of less than approximately 15–20 kDa.19

Analyst, 2014, 139, 1967–1972 | 1967

View Article Online


Analyst

The practical limitations encountered in high-resolution structural studies of proteins motivate continual progress in the development and improvement of “low-resolution” spectroscopic methods (e.g., circular dichroism,13–15 various forms of vibrational spectroscopy) which provide global insight into the overall secondary structure of proteins. For example, FTIR spectroscopy has been established as a powerful tool in the study of the secondary structure of proteins due to its apparent easiness and rapidity of acquiring well-resolved spectra.16–22 In the infrared spectra of proteins, the secondary structure is most clearly reected by the amide I and amide II bands, particularly the former, which absorbs in the region between approximately 1600 cm1 and 1700 cm1 and is primarily associated with the stretching vibrations of peptide carbonyl groups. Furthermore, infrared spectroscopy allows studies of proteins in a variety of environments. This makes the technique uniquely well suited to probe the structure of membrane proteins as well as of protein assemblies which are difficult to study by other spectroscopic methods. Despite a well-recognized conformational sensitivity of protein infrared bands, the quantitative analysis of the spectra in terms of a different secondary structure is not straightforward but relies on subsequent calculation of the obtained spectra with a multivariate statistical technique. So, it is desirable to develop an accurate and convenient method which can directly reect different conformations of proteins. In the past decade, the signicant development of THz spectroscopy has presented great potential as a complementary tool to other methods in structural analysis. The THz region of the electromagnetic spectrum can be described as being from 0.1 to 10 THz (3.3 cm1 to 333.3 cm1), bridging the microwave region and the midinfrared. As a remarkable progress in the low scale semiconductor and ultrafast optoelectronic technique, THz-TDS has been demonstrated to be an effective probe in the structural analysis based on the unique resonant properties that materials exhibit in the THz frequency range.23–34 THz spectra offer structural information resulting from intramolecular interaction, intermolecular interaction and their coupling that is not accessible with other wavelengths. Therefore, THz-TDS is advantageous over other methods in probing low frequency collective modes associated with non-local vibrations and conformation, giving the opportunity to study the secondary structure of proteins where a number of atoms move in a concerted fashion. Previous investigations have conrmed that some amino acids with denite structures exhibit “ngerprint” characteristics in the THz regime corresponding to functionally relevant, collective modes with periods on the picosecond timescale.35–37 In our previous work, we have investigated the response characteristics of regenerated silk broin in the THz region, which indicates that THz spectroscopy presents great potential as a complementary method to FTIR in determining the secondary structure of silk broin.38 In order to better understand the conformation of broin in vivo, the broin membranes formed from the middle division of the gland were employed as an ideal model in this paper. Methanol-induced conformation transition of gland broin was determined by THz-TDS coupled with FTIR spectroscopy.

1968 | Analyst, 2014, 139, 1967–1972

Paper

2.

Experimental

2.1. Chemicals and materials All chemicals used in this study were purchased from Aladdin, and were of analytical grade and used without any further purication. Bombyx mori silkworms were raised in our silkworm-breeding base. 2.2. Sample preparation The matured silkworm of 5th instar larvae was anesthetized in CO2 gas and stabilized on a dissecting tray, and subsequently dissected using a scalpel along the back. Then, two symmetric glands were carefully pulled out of the body of the silkworm, and immediately immersed in phosphate buffer solution (pH ¼ 6.6) for about ten minutes. The middle divisions were separated from the whole gland, and the epidermis was carefully removed by tweezers; then the remnant gel-like substance was rinsed with distilled water to remove sericin coated on the surface. Finally, the resulting silk broin was dissolved in distilled water for 24 hours and cast onto a at piece of polyethylene substrate, le to dry overnight under ambient conditions to yield uniform membranes with approximately 0.9 mm in thickness. Subsequently, the membranes were carefully separated from the substrate and used for the treatment or measurements. 2.3. FTIR and THz measurements The prepared samples were membranes with a certain thickness, which could meet the requirement of FTIR and THz, so measurements were easily performed on them without further treatment. For FTIR measurements, the absorption features were recorded by the ATR method using a Nicolet 5700 FTIR spectrometer over the range of 1800 cm1 to 1500 cm1, with a resolution of 4 cm1, and averaged over 32 scans. The prepared silk broin membranes were cut into 15 pieces of discs with a diameter of 5 mm. As soon as the discs were immersed into methanol/D2O (60%/40%) solution, the measurements were started. Each disc was measured for an individual spectrum, the interval between two successive spectra was 2 minutes due to the parameters chosen for measurement, and the total data collection time was 30 min in this measurement. For THz measurements, the THz-TDS system used in this paper was described in detail elsewhere.39 Briey, a Ti:sapphire laser (Femtosource) provides ultrafast pulses of 15 fs duration at a center wavelength of 800 nm with a repetition rate of 76 MHz. The pulse train was rst split into two optical beams. One optical beam controlled by an optical delay apparatus was employed to trigger a biased photoconductive semiconductor antenna, radiating the THz pulses. The obtained THz pulses were collected and collimated by a pair of gold-coated off-axis parabolic mirrors and transmitted through the investigated sample, ultimately refocused onto an electro-optic (EO) detector at normal incidence. Pulses from the other optical beam were transmitted and focused onto the same position where the THz pulses were focused. The output signals from the balanced photodiodes were monitored by a lock-in amplier. THz spectra


View Article Online


Paper

Analyst

were measured under a nitrogen atmosphere with a spectral resolution of 40 GHz and bandwidth between 0.5 and 2.6 THz. All the measurements were conducted under room temperature, and the whole testing system was located in an airtight enclosure to mitigate the effects of water vapor on the THz beams. As it takes longer time to obtain an individual THz spectrum than an FTIR spectrum, the interval between two successive THz spectra was set for 5 minutes. It is worth noting that prior to both FTIR and THz measurement each sample was removed from the solvent, subsequently the surface drop was scrubbed and used for measurements immediately. Additionally, the thicknesses of samples changed slightly aer treatment with solution, which were determined and taken into account during the extraction of optical parameters for THz spectra. Variation of FTIR spectral features with time. (The “Tn” spectrum corresponds to the gland fibroin membrane treated with methanol/ D2O solution for “n” minutes.)

Fig. 1

2.4. Optical parameters extraction process for THz measurements In order to extract optical information from the time-domain data, the theory put forward by T. D. Dorney and L. Duvillaret40,41 is employed to calculate absorption spectra. During the analysis of the transmission spectrum, sample information can be calculated using a complex transmission function “T(u)” which is obtained from the comparison between the signal spectrum and reference spectrum. “T(u)” is a function of the complex refractive index N ¼ n + ik, where “n” is the factual refractive index and “k” is the extinction coefficient, which stand for chromatic dispersion and absorption characteristics, respectively. The complex transmission function presents as the following equation: TðuÞ ¼

iðN1Þdu Esample ðuÞ 4N ¼ e c ¼ Aei4 2 Eref ðuÞ ð1 þ NÞ

(1)

where “A” is the amplitude ratio of the two waves, “u” is the angular frequency, “d” is the thickness of the sample, and “f” is the phase difference between the sample and reference signals. Then the refractive index “n(u)” and absorption coefficient “a(u)” can be calculated from the following equations: nðuÞ ¼

4c þ1 ud

" # 2 4nðuÞ aðuÞ ¼ ln d A½nðuÞ þ 12

3.

without any treatment (T0) is a broad, slightly asymmetric envelope with a center at around 1670 cm1, which is the characteristic of the random coil or silk I conformation.17,19 With increasing time of treatment in methanol/D2O solution, continuous characteristic changes in the absorbance spectra are observed when the amide I band gradually changes from 1673 cm1 to 1634 cm1. These time-resolved changes are emphasized in the difference spectra (Fig. 2) which are calculated by the subtraction of the rst absorbance spectrum (T0) from the absorbance spectrum recorded at time t aer addition of methanol/D2O solution. The increasing positive band at 1634 cm1 reects the increasing b-pleated sheet structure that develops with the time of treatment. Correspondingly, the one negative band around 1673 cm1 indicates the loss of the random coil or silk I structure during the process. The changes of spectral features in the amide I band indicate that the disappearance of the random coil or silk I structure coincides with the formation of the b-pleated sheet structure. It is worth

(2)

(3)

Results and discussion

3.1. Analysis of conformation transition from FTIR spectra Fig. 1 shows the infrared absorbance spectra of the gland broin membranes during the conformation transition aer the addition of methanol/D2O solution, from 0 to 30 min. To characterize the structure of the polypeptide, we focused on the FTIR spectra in the amide I region (1700–1600 cm1). It can be seen that the amide I band of the gland broin membrane This journal is © The Royal Society of Chemistry 2014

Variation of FTIR difference spectra with time. (The notations “Tn” can be identified in Fig. 1.)

Fig. 2

Analyst, 2014, 139, 1967–1972 | 1969

View Article Online

Analyst

Paper

noting that an isosbestic point is observed around 1650 cm1, which means that the amount of a-helix conformation does not change a lot during the treatment.


3.2. Burst phase in methanol/D2O solution It is of particular interest that the spectral features recorded at 2 min aer the immersion in methanol/D2O solution (T2) are vastly different from those of the untreated membrane (T0), which is highlighted in the inset in Fig. 1. It indicates that a signicant burst phase occurs during the rst 2 minutes when the gland broin membrane is immersed in the methanol solution. The conformational change of silk broin was found to be closely related to the microstructure of the solvent system, and to be determined by the interaction between the peptide unit of silk broin and the cluster of the mixed solvent. Once the silk broin is immersed in methanol, the structural transition of the solvent from a tetrahedral-like water structure to a chain-like methanol structure will induce the formation of new hydrogen bonds, which leads to the burst phase transition.42 In order to determine the specic amount of burst phase in the whole measurement, a quantitative method based on deconvolution was employed as an effective way to estimate the amount of each conformation (Fig. 3). The deconvolving process was performed using Origin 8.0 according to the Gaussian tting function. In the deconvolving process, the

Fig. 4 Conformation proportion for gland fibroin membranes measured at different times.

bands centered at 1673 cm1, 1658 cm1, and 1634 cm1 are considered as corresponding to the random coil, a-helix, and b-pleated sheet, respectively.16,19 The calculation results for all the samples are shown in Fig. 4, which suggest that the increase of the b-pleated sheet is accompanied by an approximately identical loss of the random coil in terms of the amount, while the amount of a-helix does not vary a lot in this process. Ideally, in view of the change of the b-pleated sheet, the amount of burst phase can be easily obtained by calculating (T2b T0b)/ (T30b T0b), where T0b, T2b, T30b represent the amount of b-pleated sheet of gland broin membranes measured at T0, T2, T30, respectively. This analysis indicates a burst phase of 33% in the b-pleated sheet structure during the rst 2 minutes of treatment in the methanol solution.

3.3. Response of conformation transition in the THz region

Fig. 3 Deconvolved spectra of gland fibroin membranes measured at times 0 (T0), 2 (T2), 30 (T30) min. (a) Original FTIR spectrum, (b) total fitting curve, b1, b2, b3: deconvolved curves for random coil (1673 cm1), a-helix (1658 cm1), and b-pleated sheet (1634 cm1), respectively.

1970 | Analyst, 2014, 139, 1967–1972

The absorption features of gland broin membranes in the range of 0.5–2.6 THz were obtained by THz-TDS, with the THz signal transmitting through the nitrogen as the reference waveform for the samples. It can be clearly seen from Fig. 5 that all the samples present signicantly different absorption features in the range of 1.3–2.0 THz, particularly at 1.54 THz, 1.67 THz and 1.84 THz, while sharing the identical characteristics beyond that region. Specically, with increasing time of treatment in methanol/D2O solution, the intensity of the 1.84 THz band gradually becomes weaker, that of the 1.54 THz band presents a contrary trend, and the features at 1.67 THz change slightly in this process. The intensity change of these bands centered at 1.84 THz, 1.67 THz, and 1.54 THz is in excellent agreement with that of FTIR peaks observed at 1673 cm1, 1658 cm1, and 1634 cm1, respectively. These interesting changes are emphasized in the THz difference spectra (Fig. 6) which are obtained by the previous method used to calculate FTIR difference spectra. In combination with the results obtained from FTIR, it can be inferred that the band centered at 1.84 THz can be regarded as a probe for the random coil structure, the peak observed at 1.54 THz corresponds to the b-pleated sheet, and the feature


View Article Online


Paper

Variation of THz spectral features with time. (The “Tn” spectrum corresponds to the gland fibroin membrane treated with methanol/ D2O solution for “n” minutes.) Right column: intensity–time plots for absorption features observed at 1.54 THz, 1.67 THz and 1.84 THz.

Fig. 5

Analyst

Fig. 7 Variation of THz spectral features with time. (The “Tn” spectrum corresponds to the regenerated fibroin membrane treated with methanol/D2O solution for “n” minutes.) Right column: intensity–time plots for absorption features observed at 1.54 THz, 1.67 THz and 1.84 THz.

45% in the b-pleated sheet structure during the rst 5 minutes of treatment in the methanol solution, while the counterpart inferred from FTIR (Fig. 4) is about 48%; this conrms that the information obtained from THz spectra is in good agreement with that of FTIR spectra. It is worth noting that the gland broin studied in this paper shares identical conformational response located at the region between 1.5 THz and 2.0 THz with the regenerated broin investigated in our previous work.38 However, it presents slightly different absorption intensity and spectral features above 2.0 THz, which can be attributed to the native difference between tested samples (Fig. 7).

4. Conclusion Variation of THz difference spectra with time. (The notations “Tn” can be identified in Fig. 5.) Fig. 6

presenting at 1.67 THz may be attributed to the a-helix structure. As each conformation presents discrete and well-resolved features in THz spectra, intensity–time plots (right column within Fig. 5) can be easily generated without the need to resort to further processing, which provides rapid insight into the kinetics of the conformational transition. It can be seen in the right column within Fig. 5 that just like the transition trend obtained from FTIR (see Fig. 4), the conformation transitions of both the random coil (1.84 THz) and b-pleated sheet (1.54 THz) are tted with an exponential function; the relative amplitudes for the increasing structure of the b-pleated sheet at 1.54 THz are coincident with those observed for the disappearance of the random coil structure at 1.84 THz. Additionally, the burst phase in the b-pleated sheet can be obtained by calculating (T5b T0b)/(T30b T0b), where T0b, T5b, T30b represent the absorption coefficients of gland broin membranes measured at T0, T5, T30, respectively. This analysis indicates a burst phase of


The conformation transition of gland broin in methanol/D2O solution was investigated by conventional FTIR coupled with the method of THz-TDS. FTIR analysis shows that, with increasing time of immersion in methanol solution, an increasing band at 1634 cm1 is observed indicating the formation of a b-pleated sheet coincident with the loss of intensity of a band at 1673 cm1 indicating decrease of the random coil structure. In addition, there is a burst phase of 33% occurring during the rst 2 minutes when the gland broin membranes are immersed into methanol/D2O solution. THz spectra present distinct features for conformations of silk broin, in combination with the results obtained from FTIR spectra; the peaks observed at 1.54 THz, 1.67 THz and 1.84 THz can be attributed to the b-pleated sheet, a-helix, and random coil, respectively. Intensity changes of bands centered at 1.54 THz and 1.84 THz conrm the formation of the b-pleated sheet and the disappearance of the random coil. Kinetic curves obtained from THz spectra indicate that the methanol-induced conformation transition from the random coil to the b-pleated sheet is tted with an exponential function. Although THz-TDS presents great potential as a complementary approach in studying the secondary structure of a protein, the resonant

Analyst, 2014, 139, 1967–1972 | 1971

View Article Online

Analyst

mechanism of spectral features for conformations has not yet been completely interpreted, this will be the focus of our future work.


Acknowledgements The authors would like to acknowledge the nancial support provided by the scientic research innovation fund in Zhejiang province (YK2011055), as well as an open project of Key Laboratory of Terahertz Optoelectronics at Capital Normal University.

Notes and references 1 Z. Z. Shao and F. Vollrath, Nature, 2002, 418, 741. 2 F. G. Omenetto and D. L. Kaplan, Nat. Photonics, 2008, 2, 641–643. 3 B. Panilaitis, G. H. Altman, J. Chen, H. J. Jin, V. Karageorgiou and D. L. Kaplan, Biomaterials, 2003, 24, 3079–3085. 4 G. H. Altman, F. Diaz, C. Jakuba, T. Calabro, R. L. Horan, J. Chen, H. Lu, J. Richmond and D. L. Kaplan, Biomaterials, 2003, 24, 401–416. 5 H. J. Jin, J. S. Chen, V. Karageorgiou, G. H. Altman and D. L. Kaplan, Biomaterials, 2004, 24, 1039–1047. 6 J. H. Qian, Y. C. Liu, H. Y. Liu, T. Y. Yu and J. Q. Deng, Anal. Biochem., 1996, 236, 208–214. 7 R. E. Unger, M. Wolf, K. Peters, A. Motta, C. Migliaresi and K. C. James, Biomaterials, 2004, 25, 1069–1075. 8 H. Saito, R. Tabeta, T. Asakura, Y. Iwanaga, A. Shoji, T. Ozaki and I. Ando, Macromolecules, 1984, 17, 1405–1412. 9 T. Asakura, K. Ohgo and K. Komatsu, Macromolecules, 2005, 38, 7397–7403. 10 T. Asakura, H. Suzuki and Y. Watanabe, Macromolecules, 1983, 16, 1024–1026. 11 T. Asakura, T. Watanabe, A. Uchida and H. Minagawa, Macromolecules, 1984, 17, 1075–1081. 12 P. Zhou, X. Xie, D. P. Knight, X. H. Zong, F. Deng and W. H. Yao, Biochemistry, 2004, 43, 11302–11311. 13 N. J. Greeneld, Nat. Protoc., 2006, 1, 2876–2890. 14 W. C. Johnson, Annu. Rev. Biophys. Biophys. Chem., 1988, 17, 145–166. 15 L. Whitmore and B. A. Wallace, Biopolymers, 2007, 89, 392– 400. 16 K. Eckert, R. Grosse, J. Malur and K. R. H. Repke, Biopolymers, 1977, 16, 2549–2563. 17 D. M. Byler and H. Susi, Biopolymers, 1986, 25, 469–487. 18 A. Dong, L. S. Jones, B. A. Kerwin, S. Krishnan and J. F. Carpenter, Anal. Biochem., 2006, 351, 282–289. 19 W. K. Surewicz, H. H. Mantsch and D. Chapman, Biochemistry, 1993, 32, 389–394.

1972 | Analyst, 2014, 139, 1967–1972

Paper

20 M. E. Goldberg and A. F. Chaffotte, Protein Sci., 2005, 14, 2781–2792. 21 J. L. Kong and S. N. Yu, Acta Biochim. Biophys. Sin., 2007, 39, 549–559. 22 X. X. Feng, Y. H. Guo and J. Y. Chen, J. Biomater. Sci., Polym. Ed., 2007, 18, 1443–1456. 23 M. X. He, A. K. Azad, S. H. Ye and W. L. Zhang, Opt. Commun., 2006, 259, 389–392. 24 A. D. Burnett, W. Fan, P. C. Upadhya, J. E. Cunningham, M. D. Hargreaves, T. Munshi, H. M. Edwards, E. H. Lineld and A. G. Davies, Analyst, 2009, 134, 1658– 1668. 25 R. S. Albert, S. Gerard, G. Regina, R. Eva, G. J. Antonio, C. Massimo and T. Javier, Analyst, 2011, 136, 1733–1738. 26 M. Walther, P. Plochocka and B. M. Fischer, Biopolymers, 2002, 67, 310–313. 27 B. M. Fischer, M. Walther and P. U. Jepsen, Phys. Med. Biol., 2002, 47, 3807–3814. 28 Y. C. Shen, P. C. Upadhya and E. H. Lineld, Vib. Spectrosc., 2004, 35, 111–114. 29 H. B. Liu, Y. Q. Chen and X. C. Zhang, J. Pharm. Sci., 2007, 96, 927–934. 30 M. R. Leahy-Hoppa, M. J. Fitch and X. Zheng, Chem. Phys. Lett., 2007, 436, 227–230. 31 N. Li, J. L. Shen and J. H. Sun, Opt. Express, 2005, 13, 6750– 6755. 32 A. R. Sanchez, G. Salvatella, R. Galceran, E. Roldos, J. G. Reguero, M. Castellari and J. Tejada, Analyst, 2011, 136, 1733–1738. 33 W. Zhang, E. R. Brown, M. Rahman and M. L. Norton, Appl. Phys. Lett., 2013, 102, 023701. 34 A. Arora, T. Q. Luong, M. Kruger, Y. J. Kim, C. H. Nam, A. Manz and M. Havenith, Analyst, 2012, 137, 575–579. 35 M. Yamaguchi, F. Miyamaru, K. Yamamoto, M. Tani and M. Hangyo, Appl. Phys. Lett., 2005, 86, 053903. 36 K. Yamamoto, K. Tominaga, H. Sasakawa, A. Tamura, H. Murakami, H. Ohtake and N. Sarukura, Biophys. J., 2005, 89, 22. 37 P. F. Taday, I. V. Bradley and D. D. Arnone, J. Biol. Phys., 2003, 29, 109–115. 38 C. Yan, B. Yang and Z. C. Yu, Anal. Methods, 2014, 6, 248–252. 39 C. Yan, B. Yang and Z. C. Yu, Anal. Lett., 2013, 46, 946– 958. 40 T. D. Dorney, R. G. Baraniuk and D. M. Mittleman, J. Opt. Soc. Am. A, 2001, 18, 1562–1571. 41 L. Duvillaret, F. Garet and J. L. Coutaz, Appl. Opt., 1999, 38, 409–415. 42 L. Ma, W. R. He, A. M. Huang, L. S. Li, Z. F. Tong, Q. N. Wei and Z. L. Huang, Spectroscopy and Spectral Analysis, 2010, 30, 3047–3051.


View Article Online

Analyst


PAPER


Conformational and mechanical changes of DNA upon transcription factor binding detected by a QCM and transmission line model† Jorge de-Carvalho,‡a Roge´rio M. M. Rodrigues,‡a Brigitte Tome´,a S´ılvia F. Henriques,b Nuno P. Mira,b Isabel Sa´-Correiab and Guilherme N. M. Ferreira*a A novel quartz crystal microbalance (QCM) analytical method is developed based on the transmission line model (TLM) algorithm to analyze the binding of transcription factors (TFs) to immobilized DNA oligoduplexes. The method is used to characterize the mechanical properties of 00 0 and Gfilm, and the biological films through the estimation of the film dynamic shear moduli, Gfilm film thickness. Using the Saccharomyces cerevisiae transcription factor Haa1 (Haa1DBD) as a biological model two sensors were prepared by immobilizing DNA oligoduplexes, one containing the Haa1 recognition element (HREwt) and another with a random sequence (HREneg) used as a negative control. The immobilization of DNA oligoduplexes was followed in real time and we show that DNA strands initially adsorb with low or non-tilting, laying flat close to the surface, which then lift-off the surface leading to final film tilting angles of 62.9 and 46.7 for HREwt and HREneg,

Received 5th September 2013 Accepted 30th October 2013

respectively. Furthermore we show that the binding of Haa1DBD to HREwt leads to a more ordered and compact film, and forces a 31.7 bending of the immobilized HREwt oligoduplex. This work demonstrates the suitability of the QCM to monitor the specific binding of TFs to

DOI: 10.1039/c3an01682j

immobilized DNA sequences and provides an analytical methodology to study protein–DNA biophysics

www.rsc.org/analyst

and kinetics.

Introduction The understanding of the role of transcription factor (TF) proteins in gene expression is very important in bioengineering and biomedical applications. TFs act as transcription triggers having inuence on efficient biological response to different stimuli.1,2 Thus, research on the regulation of gene transcription by TFs, leading to system manipulation with engineered transcription pathways, is being carried out to understand cellular mechanisms with relevance, for instance, in cancer research3 and to enhance the microbial stress tolerance in industrial processes.4,5 The mechanism of TF recognition of specic DNA sequences is believed to involve two distinct steps. In the rst step the protein binds the DNA strand, through non-specic electrostatic interactions. The TF is assumed to probe the DNA a

IBB-Institute for Biotechnology and Bioengineering, Centro de Biomedicina Molecular e Estrutural, Universidade do Algarve, 8005-139 Faro, Portugal. E-mail: gferrei@ualg. pt; Fax: +351-289818419

b

IBB-Institute for Biotechnology and Bioengineering, Center for Biological and Chemical Engineering, Department of Bioengineering, Instituto Superior Tècnico, Universidade de Lisboa, Avenida Rovisco Pais, 1049-001 Lisbon, Portugal † Electronic supplementary 10.1039/c3an01682j

information

(ESI)

available.

‡ These authors contributed equally as rst authors to this work.


See

DOI:

nucleotide sequence by a sliding process.6 When the specic recognition element is found, the TF specically binds to it through hydrogen bonds, resulting in a molecular dehydration7 and conformational modication, usually known as DNA bending.8 Novel label free technologies, such as bulk acoustic wave sensors, allowing a quick assessment of binding kinetics and mechanics of the TF interaction with different DNA motifs are highly demanded.9–11 The use of bulk acoustic wave sensors has gained signicant interest for the study of molecular interactions.12–14 These sensors, commonly known as quartz crystal microbalances (QCM), respond linearly with the increase of the mass deposited at the sensor surface.15 This property is very useful for the determination of kinetics constants by simply monitoring in real-time the variation of the sensor resonance frequency.16–19 This ideal response, known as the Sauerbrey equation,15 is however affected by the loss of energy of the propagating acoustic wave. Such loss is associated with the viscoelasticity of the deposited lms, and with the viscosity, density and charges of the loaded liquid solutions.12,20–22 It was shown that when detecting analytes in liquid environments, or when exible and viscoelastic biomaterials are deposited at the sensor surface, both the mass and the liquid contribute to the measured resonance frequency variations.20,23 These contributions, which are additive, cannot be distinguished by simple

Analyst, 2014, 139, 1847–1855 | 1847

View Article Online


Analyst

frequency counting.20,24 As such, complex impedance analysis has been used to gather more detailed information related to the event differentiation occurring at the sensor surface.20,21,24 Recent efforts on the application of physical models and mathematical algorithms to interpret the diverse contributions affecting the acoustic wave propagation fostered a new generation of QCM-linked technologies.12,14,25,26 The viscoelastic properties of deposited lms can be assessed by Voigt model description of the physical properties of materials. This model is commonly used in QCM-D data analysis. QCM-D is based on the measurement of the sensor electrical response at several resonance harmonics. The variations of the resonance frequency (Df) and acoustic wave dissipation (DD) are determined by periodically switching on and off the driving alternating current (AC) voltage and by measuring the decay signal of the crystal oscillation. The data are later used in a complex tting process for the determination of materials physical properties, as described elsewhere.27 Another proposed manner to derive quantitative information related to the conformation of the deposited lm layers is the Transmission Line Model (TLM).28,29 The TLM is an electromechanical model derived from a one-dimensional solution of the wave equation.28 It enables the determination of the dynamic shear modulus (G*) from the sensor admittance (Y) around the fundamental resonance. Both Voigt and TLM models rely on the characterization of the biological layers through the determination of the lm thickness (hlm) and dynamical shear modulus (G*lm). Both models were successfully applied to measure the viscoelastic properties of distinct targets, such as polymers,27,28,30 lipid vesicles,31 proteins,23,32 DNA,33 and cells.34 The major disadvantage of the QCM-D technique is the extrapolation of G*lm to different harmonics. As G*lm is known to have a dependency on the frequency,29 QCM-D relies on the rough assumption of linearity for such dependency.35 The application of the TLM, on the other hand, is constricted to the sensor resonance frequency. Therefore, the determination error associated with the dependency of G*lm on the resonance frequency is minimized.29 In this work we demonstrate the applicability of the TLM to derive the mechanical properties of the biological molecules deposited at the QCM surface. The use of the TLM clearly shows that specic binding of the transcription factors to immobilized DNA sequences increases the lm viscosity while decreasing its thickness. These variations are associated with the DNA strand bending upon the TF specic binding. The biological model used is based on the interaction between the Saccharomyces cerevisiae transcription factor Haa1 and immobilized DNA strands (HREwt) that contain this TF's target DNA, the Haa1-responsive element (HRE).36 The immobilized HREwt DNA strands used are a part of the promoter region of the Haa1-regulated TPO3 gene that is recognized in vitro and in vivo by Haa1.36 Haa1 plays an important role in yeast adaptation and tolerance to stress induced by acetic and propionic acids,37–39 with potential implications in biotechnology and the food industry.4,5

1848 | Analyst, 2014, 139, 1847–1855

Paper

Theory – the transmission line model The TLM assumes that the mechanical properties of a viscoelastic adlayer can be deduced from a one-dimensional solution of the wave equation.28 The mechanical properties of the viscoelastic material are determined by the dynamic shear modulus, G*lm, which describes the material's response to shearing strain caused by the acoustic wave. Under that sinusoidal deformation, this is dened as a complex quantity: 00

0 + iGfilm G*film ¼ Gfilm

(1)

0 , represents the component stress in-phase The real part, Glm with the strain, leading to energy storage and is called the 00 storage (or elastic) shear modulus. The imaginary part, Glm , which represents the component of stress out-of-phase with the strain, is a measure of energy dissipation and is called the loss (or viscous) shear modulus. The TLM is a one-dimensional model since it assumes isotropic, homogeneous, and uniform layers in which lateral dimensions have no effect on the wave propagation.29 Usually the TLM is applied to describe the analytical structure based on a three layer system: (i) the quartz crystal electrode, (ii) the viscoelastic layer and (iii) the semi-innite Newtonian liquid medium (Scheme 1).21,29

Scheme 1 Representation of the three layer system for TLM application. In QCM sensing applications the TLM is usually applied for the electromechanical characterization of systems composed of a quartz crystal electrode layer (i), a viscoelastic layer (ii) and a semi-infinite Newtonian liquid medium (iii). Since the immobilized biotin–streptavidin (SAM:SA) is very rigid and adopts an ordered orientation, it could be included within the quartz crystal electrode layer. The determination of the eight unknowns: the quartz effective thickness, viscosity and static capacitance (hquartz, hquartz, and C0), the liquid density and viscosity (rliquid and hliquid), and the biological layer thickness, density and dynamic shear modulus (hfilm, rfilm, and G*film) allow us to describe the system electromechanically. Using the theoretical length of DNA it was possible to estimate the tilting angle (s) after HREwt and HREneg adsorption and the DNA bending angle (b) upon Haa1DBD binding.


View Article Online

Paper

Analyst

The global admittance of the TLM circuit can be represented by a static capacitance in parallel with the motional impedance28 and it is described by the equation


Y ¼ iuC0* þ

1 Zm

(2)

where C*0 is the sum of the static quartz capacitance C0 and an added external capacitance Cex that accounts for packaging, connections, crystal holder, etc. The motional impedance Zm is given by 0 1 ZL 1i cot a C 1 B Zquartz B C 1C Zm ¼ (3) B 2 A iuC0 @K a ZL 2 tan i a Zquartz 2 where K2 is the electromechanical coupling of the quartz, a is the acoustic phase shi inside the quartz crystal, Zquartz is the characteristic acoustic impedance of the quartz crystal, and ZL is the acoustic load impedance acting at the surface of the quartz crystal. With the exception of the acoustic load impedance (ZL), all these parameters depend on the resonance frequency of the quartz crystal and on the intrinsic properties of the quartz28 – Table 1. When the quartz crystal sensor is loaded only with a semiinnite Newtonian liquid, the storage or elastic shear modulus 0 is null (Gliquid ¼ 0 Pa), given the purely viscous liquid nature, and ZL in eqn (3) is given by Zliquid as dened in eqn (10). qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Zliquid ¼ iurliquid hliquid (10) where rliquid and hliquid are the liquid density and viscosity, respectively. When the quartz crystal sensor is simultaneously loaded with viscoelastic mass and liquid, the acoustic load impedance is represented by the contribution of the physical and geometrical properties of both layers: ZL ¼ Zfilm

Table 1

Zliquid þ Zfilm tanhðgfilm hfilm Þ Zfilm þ Zliquid tanhðgfilm hfilm Þ

(11)

where Zlm is the characteristic impedance of the coating and depends on the complex dynamic shear modulus G*lm and its density rlm (Zlm ¼ (rlmG*lm)1/2). glm is the complex wave propagation factor in the coating (glm ¼ iurlm/Zlm) and hlm is the coating thickness. Eqn (1)–(11) allow the calculation of the electrical response of the loaded QCM from the physical properties of the loading layers (G*lm, hlm, rlm, hliquid and rliquid), the quartz crystal parameters (Table 1) and parallel static capacitance (C*0).

Experimental Preparation of the acoustic QCM sensor AT-cut quartz crystals covered on both sides by gold electrodes of 100 nm with a fundamental resonance frequency of 10 MHz were purchased from International Crystal Manufacturing Company (USA). The crystals were cleaned by incubation in a UV/ozone oven (NovaScan, USA), exposure to piranha solution (H2SO4–H2O2 at 3 : 1 (v/v)), rinsing with ultra-pure water and a nal N2 stream drying. The crystals were then immersed in a mixture of 10% biotin–PEG disulde (LCC Engineering & Trading, Switzerland) and 90% 11-hydroxy-1-undecanethiol (Dojindo, Japan) at a net concentration of 1 mM in absolute ethanol and further incubated for at least 24 hours at 4 C. Functionalized QCM crystal sensors were mounted on a 25 mL acrylic cylindrical ow cell which was custom designed and optimized by computational uid dynamics tools. A syringe pump system (Cetoni GmbH, Germany) and a selector valve (Valco Instruments, USA) were used to control the ow onto the acoustic QCM sensor cell. Immobilization of DNA sequences and preparation of the Haa1DBD peptide used for protein–DNA interaction assays To prepare streptavidin-modied surfaces for DNA oligoduplex immobilization, the biotin-functionalized QCM electrodes were exposed to 3 mg mL1 streptavidin (Roche). Two complementary oligonucleotides (50 -[biotin]-TTC TCT GTG CTT GGC GAG GGG TTT ACT GGA GCC CAA TC and 50 -GA TTG GGC TCC AGT AAA CCC CTC GCC AAG CAC AGA GAA-30 ), including a HRE motif

Quartz crystal parameters used in the TLM for electromechanical characterization of the QCM electrodea layer28,29

Parameter

Variable/equation

Value

Viscosity (Pa s) Thickness (m) Density (g dm3) Mechanical stiffness (N m2) Piezoelectric constant (A S m2) Permittivity (A2 s4 kg1 m3) Piezoelectrically stiffened shear modulus

hquartz hquartz rquartz c66 equartz h e26 3quartz h 322

(4) (5) (6)

3.500 104 1.670 104 2.651 103 2.947 1010 9.530 102 3.982 1011 FD

Electromechanical coupling

K2 ¼

(7)

FD

(8)

FD

(9)

FD

Internal acoustic phase shi Characteristic acoustic quartz impedance a

cquartz hc66 þ

e26 2 þ iuhquartz 322

equartz 3quartz cquartz rffiffiffiffiffiffiffiffiffiffiffiffiffi rquartz a ¼ uhquartz cquartz pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Zquartz ¼ rquartz cquartz

FD ¼ frequency-dependent.


Analyst, 2014, 139, 1847–1855 | 1849

View Article Online


Analyst

Paper

(indicated in bold), were annealed and the oligoduplex obtained was immobilized on the streptavidin-coated surface of the QCM electrode. 0.1 mM of the HRE oligoduplex (HREwt) was used for immobilization on the surface of the QCM sensor via biotin– streptavidin affinity coupling. The same experimental procedure was carried out to immobilize the oligoduplex used as a negative control (HREneg; 50 -[biotin]-GTA TAC TTA TAT GTA TAT TAT ATT CAA ATA CAT TAT GT-30 and 50 -AC ATA ATG TAT TTG AAT ATA ATA TAC ATA TAA GTA TAC). The Haa1DBD peptide used for protein–DNA interaction assays was over-expressed in Escherichia coli and puried by affinity chromatography, as described before.36 For the protein–DNA interaction assays 50 nM of the Haa1DBD peptide were used for each type of DNA (HREwt and HREneg). Tris buffer pH 8.0 (10 mM Tris, 100 mM KCl, 0.005% Tween20) was used as running buffer during all experiments with a constant ow rate of 100 mL min1. Impedance spectroscopy data acquisition The complex reection coefficient (S11) was acquired with an Agilent 4395A impedance analyzer. S11 was converted in complex admittance (Y ¼ G + iB), where G is the conductance and B is the susceptance. The system was calibrated with standards (open, short and 50 U load) prior to any measurement. Aer calibration 801 points over 50 kHz span centered on the maximum of the conductance (Gmax) were continually measured during biomolecule adsorption. The frequency at the maximum of the conductance, fGmax, is very close to the series resonance frequency of the motional branch (fs). If the S11 measurement is performed away of any of the frequencies where the QCM vibrates (minimized motional behavior), the C*0 can be calculated by: C*0 z B/w

(12)

These experimental data are the input basis for the application of the TLM algorithm. Estimation of biological lm density The impedance spectroscopy data were treated with an adapted TLM-based algorithm. The biological lm density is another important input value. This can be either determined by an alternative technique or estimated using values from the literature. Using published density values of 1000, 1350, and 1700 g dm3 for liquid buffer, protein, and DNA, respectively, we have estimated the lm density at the adsorption equilibrium.7 In the case of a DNA lm, the streptavidin mass, which can be obtained directly from the Sauerbrey equation due to its lm rigidity, was used for stoichiometric determination of the number of DNA molecules. The molecular model of DNA allowed us to estimate the weighted density of the binary mixture (DNA + liquid buffer). In the case of DNA–protein complex lms, the density was assumed estimating the effect on lm density provoked by replacing an amount of liquid buffer by adsorbed protein into the DNA lm. In both cases the lm densities during the adsorption transient were normalized using the variation of fs.

1850 | Analyst, 2014, 139, 1847–1855

Results and discussion Uniqueness solution issue on TLM application The use of the TLM in sensing applications aims at the extraction of the mechanical properties of the adlayers from the experimentally measured complex admittance spectra near the sensor resonance frequency. We are interested in estimating the thickness (hlm), the density (rlm) and the dynamic complex shear modulus (G*lm) of biological layers. Aer discounting C*0 from experimental complex admittance spectra (Gm,exp and Bm,exp), the estimation of these parameters requires nding the solution of eqn (3) and (11) for the three layer system depicted in Scheme 1. In this process however, there are eight unknowns: the quartz effective thickness, viscosity and static capacitance (hquartz, hquartz, and C0), the liquid density and viscosity (rliquid and hliquid), and the biological layer thickness, density and dynamic shear modulus (hlm, rlm, and G*lm). As the TLM does not give an explicit mathematical expression for G*lm, and given the interdependency of the lm layer variables (hlm, rlm, and G*lm), the use of Gm,exp and Bm,exp spectra results in an undetermined problem. In fact, several combinations of hlm, rlm, and G*lm can generate the same admittance spectra resulting in a uniqueness solution issue.29,40 As such, an iterative approach was proposed to numerically solve this issue and extract hlm, rlm, and G*lm.29 A scheme of the adapted algorithm from the work by Jim´ enez et al.29 is shown in Fig. 1. The strategy is to parameterize the unknown variables, using only solutions with physical meaning to generate admittance spectra based on TLM equations (Gm,TLM and Bm,TLM). The best solution is further selected based on error minimization between the experimental and TLM-derived admittance spectra. This algorithm is applied in two steps: (A) electrode layer calibration and (B) extraction of the mechanical properties of the biological adlayer. Electrode calibration. The rst step is the calibration of the QCM sensor immersed in the liquid. In this step, the liquid load parameters (rliquid and hliquid) are xed (Table 2) and the quartz crystal parameters hquartz and hquartz are parameterized. For each generated pair of hquartz and hquartz, eqn (3) is solved to calculate the complex C0 using the acoustic load impedance (ZL) equal to Zliquid at resonance frequency (fs,i). The motional impedance is obtained from the experimental motional admittance spectra values at resonance frequency (Zm,fs,i). This generates triads (hquartz, hquartz, and C0) of possible solutions, which are narrowed down applying the physical criteria stating that the imaginary part of C0 has to be null (C0,imag ¼ 0).29 The variation of C0 with hquartz and hquartz is shown in Fig. 2A. As expected the variation of the crystal effective viscosity (hquartz) has no effect on the imaginary part of C0 (Fig. 2A). The zero of C0,imag was calculated using a regula falsi-based numerical method with a tolerance equal to 1015. This constriction was achieved for a crystal thickness (hquartz) of 167.2 mm (Table 2). Nevertheless, it is clear in Fig. 2A that any crystal viscosity can be considered a solution based only on the physical criteria. The nal selection is made by comparing the motional admittance spectra generated using eqn (3) (Gm,TLM,i and Bm,TLM,i) for these


View Article Online

Analyst


Paper

Flowchart of the mathematical algorithm adapted from ref. 29 to numerically solve TLM equations. This algorithm is applied in two steps: (A) electrode calibration and (B) extraction of the mechanical properties of the adsorbed biological layer. In both steps the strategy is to parameterize the unknown variables, using only solutions with physical meaning to generate admittance spectra based on TLM equations (Gm,TLM and Bm,TLM). The best solution is further selected based on an error minimization between the experimental (Gm,exp and Bm,exp) and TLMderived admittance spectra. Fig. 1

The selected triad is summarized in Table 2. Since the immobilized biotin–streptavidin layer is very rigid and adopts an ordered orientation,41 no signicant deviations to the

triads with the experimental spectra (Gm,exp,i and Bm,exp,i) (Fig. 2B). The solution is the triad that minimizes the error between spectra (eqn (13)).

n ofP points

Error ¼

k¼1

Gm;exp;i ðuk Þ Gm;TLM;i ðuk Þ n ofP points k¼1


Gm;exp;i ðuk Þ

2

2

þ þ

n ofpoints P

Bm;exp;i ðuk Þ Bm;TLM;i ðuk Þ

k¼1 n ofP points k¼1

Bm;exp;i ðuk Þ

2

2 (13)

Analyst, 2014, 139, 1847–1855 | 1851

View Article Online

Analyst Table 2

Paper Summary of parameter extraction using the TLM-based algorithm for extraction of the different studied biomolecular filmsa

Layer

Film

h (mPa s)

r (g dm3)

C0 (pF)

G0 (kPa)

G00 (kPa)

h (nm)

Electrode

HREwt HREwt–Haa1DBD HREneg HREneg–Haa1DBD HREwt HREneg HREwt–Haa1DBD HREneg–Haa1DBD HREwt HREneg HREwt–Haa1DBD HREneg–Haa1DBD

13.14

2651

5.13

NA

NA

167.2 103

24.74

2651

5.14

NA

NA

167.2 103

1180 1130 1210 1170 1000

NA NA NA NA NA

18.9 20.9 19.7 16.8 0

89.2 1.3 91.3 1.4 135.3 1.7 83 3.7 60

11.5 0.3 9.4 0.3 9.1 0.1 10.3 0.3 NA


Biological lm

Liquid

a

1.42 1.45 2.16 1.33 0.001

0.02 0.02 0.03 0.06

1.3 1.5 0.4 1.9

NA – not applicable.

calibration parameters of the system are expected.33 Therefore, our system can still be represented as a three layer system and this calibration process was applied to the QCM modied with biotin–streptavidin (Scheme 1). Furthermore, given the strong binding of streptavidin–biotin, this construction is no-longer disrupted and therefore was explored to generate biosensors. Aer immobilization of the biotin-modied oligoduplexes and binding of the TFs, the biosensor regeneration is possible simply by removing the bound TF. Clearly, the use of a high salt concentration removes the bound TF from the immobilized oligoduplexes as shown by the restoration of the sensor motional series resonance frequency (see Fig. S1, ESI†). Extraction of mechanical properties of the adsorbed biological layer. In the second step of the algorithm application, the parameter sets in the calibration are used to calculate the mechanical properties of the immobilized biomolecules. A similar strategy to numerically solve the TLM equation is now used to calculate hlm, rlm, and G*lm. However, at this stage, it is more relevant to estimate the surface mass of the lm (msurf,lm) rather than the lm density. The lm storage (or

0 elastic) shear modulus (Glm ) and loss (or viscous) shear 00 modulus (Glm) are parameterized to calculate the characteristic impedance of the biological layer (Zlm). The acoustic load impedance (ZL,fs,n) is calculated from eqn (3), where Zm,fs,n is obtained from the experimental motional admittance spectra at resonance frequency as done before. Zlm, Zliquid and ZL,fs,n are then used to calculate the complex surface mass using eqn (14):

ms ¼

Zfilm ½lnðrÞ þ iðq þ 2kpÞ 2iu

(14)

where k is related to the mechanical resonance of the coating and demonstrated to be null below its rst mechanical resonance.29 r and q are, respectively, given by Z Zfilm þ Zliquid 1 L Z film r ¼ (15) Z film Z 1 þ Z film liquid Zliquid q ¼ phase(r)

(16)

Fig. 2 Plot outputs for the electrode calibration step of the adapted TLM-based algorithm. (A) The quartz effective viscosity (hquartz) was parameterized between 0.01 and 1 Pa s and the quartz effective thickness (hquartz) was parameterized between 120 and 220 mm. The real (squares) and the imaginary (triangles) part of the static quartz capacitance (C0 ¼ C0,real + C0,imag) were calculated using eqn (3) for each generated pair of hquartz and hquartz. A physical criterion stating that the imaginary part of C0 has to be null (tolerance ¼ 1015) was applied to narrow the triad (hquartz, hquartz, and C0,real) solutions (open circles). (B) The solutions with physical meaning (circles) were used to calculate motional admittance spectra (GTLM and BTLM). The minimum error between experimental and TLM-based spectra (of each triad) was found using eqn (13). The relative error to the selected solution (open circle) error is shown as a function of hquartz and hquartz narrowed variations.

1852 | Analyst, 2014, 139, 1847–1855


View Article Online


Paper

Analyst

It is clear from eqn (14)–(16) that only a restricted set of solutions have physical meaning. These correspond to the solutions leading to the non-imaginary part of the mass. This was therefore used as the physical criteria to narrow down the 00 0 set of possible solutions. The triads Glm , Glm and msurf,lm are then used to generate motional complex admittance spectra (Gm,TLM,n and Bm,TLM,n) which are compared with the experimental spectra (Gm,exp,n and Bm,exp,n). Similar to electrode calibration, the best solution is the triad minimizing an analogous error function (eqn (13)). The lm thickness (hlm) is another important parameter extracted from the calculated surface mass: hfilm ¼

Aelectrode rfilm mfilm

(17)

Immobilization of biotinilated DNA oligoduplexes over the QCM surface previously functionalized with streptavidin. Two different nucleotide sequences were tested: a DNA oligoduplex with the specific recognition element (HREwt, solid squares) and non-specific nucleotide sequence (HREneg, open squares). A mathematical algorithm based on the TLM was applied to extract the mechanical properties during the entire DNA film formation: (A) storage shear modulus 00 0 (Gfilm ); (B) loss shear modulus (Gfilm); and, (C) thickness. Time-average over 30 s is presented corresponding to three measured data points. Fig. 3


where Aelectrode is the overlaid area between QCM gold electrodes. Monitoring the immobilization of DNA with the TLM The HRE oligoduplex36 and the negative control were immobilized on the sensor surface via biotin–SA affinity coupling. The electrical admittance was recorded in real time during the entire DNA lm formation. The described TLM algorithm was then applied for the extraction and characterization of the lm dynamic shear moduli and thickness (Fig. 3). As expected both HREwt and HREneg sequences show similar mechanical properties since they have the same length. It is clear that a very 00 0 rapid variation of Glm and Glm occurs at the beginning of the deposition of the DNA sequences (Fig. 3A and B). This shows that DNA is deposited very quickly and promotes signicant alterations of the dynamic shear moduli of the system. This response is in accordance with the formation of a more rigid 0 layer, revealed by the increasing Glm , which is also more 00 viscous, as revealed by the out of phase strain component Glm. 00 0 Aer approximately 100 s and 200 s, Glm and Glm tend to stabilize, respectively. It is noticeable that during this period, the lm thickness also increased (Fig. 3C). This suggests that DNA strands initially adsorb with low or no-tilting, laying at close to the surface, and then start liing off the surface as more molecules adhere. This mechanism of surface coverage by DNA oligoduplex molecules has also been demonstrated by other techniques such as AFM42 and neutron reectivity.43 At the end of the deposition process the thickness of the DNA layer was 11.5 0.3 nm for the HREwt and 9.4 0.3 nm for the HREneg, which are in the same range as previously estimated thicknesses of DNA oligoduplexes of similar length.7,43,44 The molecular model of DNA estimates a contour length of 12.92 nm for the 38 base pair HREwt and HREneg oligonucleotide sequences. Thus, we predict from the TLM that at the end of the deposition process, HREwt and HREneg oligonucleotide molecules are tilted by 62.9 2.9 and 46.7 1.9 , respectively. These tilting angles (s in Scheme 1) are also in accordance with the physical models of immobilized DNA molecules shown by high resolution in situ visualization techniques.42,43 The explanation for the discrepancy of the tilting angles of both oligonucleotides relies on the estimation of the immobilized lm thickness. The lm thickness is obtained from the TLM surface mass assuming the density of buoyant DNA (rDNA ¼ 1700 g dm3)7 and normalizing to the total amount of oligonucleotides deposited at the QCM surface (Table 2). Since the total surface mass of HREneg is lower than the surface mass of HREwt, the thickness difference and also the tilting angle discrepancy are due to a calculation error inherent to this assumption. Despite this apparent drawback, the TLM was clearly effective in determining the intrinsic mechanical properties of the oligonucleotide lms. The TLM shows bending of DNA upon specic binding of TFs The TLM was further used to monitor the binding of Haa1 to the immobilized oligonucleotide sequences. Because the heterologous expression of full-length Haa1 in E. coli is difficult and leads to a very low yield of puried protein, the interaction

Analyst, 2014, 139, 1847–1855 | 1853

View Article Online


Analyst

Fig. 4 Monitoring of Haa1DBD at 50 nM interaction with HREwt (full squares) and HREneg (open squares). The variation of the (A) storage 00 0 modulus (DGfilm ), of the (B) loss modulus (DGfilm) and of (C) the film thickness (Dhfilm) shows that binding mechanisms of the Haa1DBD to HREwt and of the Haa1DBD to HREneg can be distinguished by applying the TLM-based algorithm. Time-average over 30 s is presented corresponding to three measured data points.

assays with DNA were carried out with a peptide that comprises solely the DNA-binding domain of Haa1 (mapped to the 123 Nterminal residues). The interaction of this peptide with the HRE motif was proved to mimic the interaction observed with the full-length TF.36 The modied QCM sensors with HREwt and HREneg were challenged with 50 nM solution of the Haa1DBD peptide with a constant ow-rate of 100 mL min1. The impedance of the system was measured and the TLM was used to calculate the 00 0 variations of Glm , Glm and hlm. In Fig. 4 we show the calcu0 lated variation of the storage modulus (DGlm ), of the loss 00 modulus (DGlm) and of the lm thickness (Dhlm). It is observed that binding of the TFs to specic oligonucleotide sequences has less effect on the immobilized lm stiffness as the storage modulus is constant throughout the binding

1854 | Analyst, 2014, 139, 1847–1855

Paper

process, while for the unspecic oligonucleotide sequence it has slightly decreased (Fig. 4A). The binding of Haa1DBD to the specic DNA sequences however is clearly different from the unspecic binding when looking at the variation of the loss modulus (Fig. 4B) and lm thickness (Fig. 4C). In fact, binding of Haa1DBD to its specic HREwt sequence resulted in a lm with higher viscosity, suggesting the formation of a more ordered closed pack structure as compared with binding to HREneg. This is further demonstrated by the lower end-value thickness achieved for the binding to the specic sequence (Table 2). It is important to note that the results achieved by the TLM are in accordance with the general accepted biological model for TF binding to DNA. The formation of a protein–DNA complex comprises two distinct steps. Initially, the protein binds to the DNA non-specically, mostly by electrostatic interactions. This initial step is thus expected to be similar for different DNA sequences and thus not possible to be discriminated by several measuring techniques. Fig. 4C clearly shows this as up to 100 s the TF binding to HREwt and HREneg revealed no alteration of the lm thickness. Aer this initial binding the TF is assumed to slide through the DNA strand, probing the nucleotide sequence.6 When the specic recognition element is found, the TF specically binds, forming hydrogen bonds and releasing water molecules from the hydration spheres,7 resulting in a modication of the conformation of the DNA molecule.8 In an acoustic wave biosensor this is signaled by an increase of the lm viscosity and a decrease of the lm layer thickness, as indeed shown by the output of the TLM (Fig. 4). The lm thickness decrease for the specic binding of Haa1DBD to the HREwt sequences was approximately 2.4 0.1 nm (Fig. 4C), resulting in a nal thickness of 9.1 0.1 nm (Table 2). It is well known that the binding of proteins to DNA promotes the bending of the DNA at the specic recognition element.8 The decrease of the thickness of the immobilized HREwt is thus expected and is in accordance with the TF–DNA binding mechanism. In fact, bending of the DNA strands is considered to be of major importance for the regulation and control of gene expression.45,46 As the specic recognition element of the immobilized HREwt is located at the middle of the oligonucleotide sequence, using simple trigonometric relations (b in Scheme 1, see ESI†) we estimate a bending of about 31.7 1.0 in the immobilized HREwt aer binding of Haa1DBD, which is within the range of bending as estimated from molecular modeling.6,47 Herein, these results show that TLM application for impedance spectroscopy data analysis of the QCM allows us to understand the DNA bending mechanism occurring upon specic interaction with TFs.

Conclusion Impedance spectroscopy was used to monitor the changes in admittance spectra of a QCM around its resonance frequency, which occurred during the formation of DNA layers and sequential interaction of these with a TF. A mathematical algorithm based on the TLM was successfully applied for data analysis and extraction of the mechanical properties of different biological layers. The values obtained are in accordance with


View Article Online

Paper

the biological models of DNA lm formation and the DNA–TF recognition process indicating the potential of this approach in biophysical studies.


Acknowledgements This work was supported by national Portuguese funding through FCT – Fundaç˜ ao para a Ciˆ encia e a Tecnologia, project ref. PEst-OE/EQB/LA0023/2013. The authors further acknowledge the FCT for the nancial support through the research projects PTDC/EBB-EBI/108517/2008 and PTDC/SAU-BEB/ 105189/2008 and the grants SFRH/BD/33720/2009 (J.C.), SFRH/ BD/38136/2007 (R.M.M.R.), SFRH/BPD/44895/2008 (B.T.) and SFRH/BD/78058/2011 (S.F.H.).

References 1 G. D. Amoutzias, D. L. Robertson, Y. Van de Peer and S. G. Oliver, Trends Biochem. Sci., 2008, 33, 220–229. 2 N. Mira, M. Teixeira and I. S´ a-Correia, in Transcriptional Regulation SE – 2, ed. A. Vancura, Springer, New York, 2012, vol. 809, pp. 27–48. 3 N. H. Colburn and T. W. Kensler, Cancer Prev. Res., 2008, 1, 153–155. 4 M. C. Teixeira, N. P. Mira and I. S´ a-Correia, Curr. Opin. Biotechnol., 2011, 22, 150–156. 5 Z. Lin, Y. Zhang and J. Wang, Biotechnol. Adv., 2013, 31, 986–991. 6 B. Bouvier, K. Zakrzewska and R. Lavery, Angew. Chem., Int. Ed. Engl., 2011, 50, 6516–6518. 7 W. Y. X. Peh, E. Reimhult, H. F. Teh, J. S. Thomsen and X. Su, Biophys. J., 2007, 92, 4415–4423. 8 C. G. Kalodimos, N. Biris, A. M. J. J. Bonvin, M. M. Levandoski, M. Guennuegues, R. Boelens and R. Kaptein, Science, 2004, 305, 386–389. 9 S. Tan and T. J. Richmond, Curr. Opin. Struct. Biol., 1998, 8, 41–48. 10 C. W. M¨ uller, Curr. Opin. Struct. Biol., 2001, 11, 26–32. 11 A. M. Florescu and M. Joyeux, J. Chem. Phys., 2009, 130, 015103. 12 G. N. M. Ferreira, A.-C. Da-Silva and B. Tom´ e, Trends Biotechnol., 2009, 27, 689–697. 13 M. A. Cooper and V. T. Singleton, J. Mol. Recognit., 2007, 20, 154–184. 14 B. Becker and M. A. Cooper, J. Mol. Recognit., 2011, 24, 754– 787. 15 G. Sauerbrey, Zeitschri fur Physik, 1959, 155, 206–222. 16 G. N. M. Ferreira, J. M. Encarnaç˜ ao, L. Rosa, R. Rodrigues, R. Breyner, S. Barrento, L. Pedro, F. Aires da Silva and J. Gonçalves, Biosens. Bioelectron., 2007, 23, 384–392. 17 J. M. Encarnaç˜ ao, L. Rosa, R. Rodrigues, L. Pedro, F. A. da Silva, J. Gonçalves and G. N. M. Ferreira, J. Biotechnol., 2007, 132, 142–148. 18 J. M. Encarnaç˜ ao, R. Baltazar, P. Stallinga and G. N. M. Ferreira, J. Mol. Recognit., 2009, 22, 129–137. 19 A. Y. Ryazanova, E. A. Kubareva, I. Grman, N. V. Lavrova, E. M. Ryazanova, T. S. Oretskaya and T. Hianik, Analyst, 2011, 136, 1227–1233.


Analyst

20 J. M. Encarnaç˜ ao, P. Stallinga and G. N. M. Ferreira, Biosens. Bioelectron., 2007, 22, 1351–1358. 21 R. Lucklum and P. Hauptmann, Sens. Actuators, B, 2000, 70, 30–36. 22 R. Etchenique and A. D. Weisz, J. Appl. Phys., 1999, 86, 1994– 2000. 23 F. H¨ o¨ ok, B. Kasemo, T. Nylander, C. Fant, K. Sott and H. Elwing, Anal. Chem., 2001, 73, 5796–5804. 24 S. J. Martin, V. E. Granstaff and G. C. Frye, Anal. Chem., 1991, 63, 2272–2281. 25 I. Reviakine, D. Johannsmann and R. P. Richter, Anal. Chem., 2011, 83, 8838–8848. 26 R. W. Cernosek, S. J. Martin, A. R. Hillman and H. L. Bandey, IEEE Trans. Ultrason. Eng., 1998, 45, 1399–1407. 27 M. V. Voinova, M. Rodahl, M. Jonson and B. Kasemo, Phys. Scr., 1999, 59, 391–396. 28 R. Lucklum and P. Hauptmann, Faraday Discuss., 1997, 107, 123–140. 29 Y. Jim´ enez, R. Fern´ andez, R. Torres and A. Arnau, IEEE Trans. Ultrason. Eng., 2006, 53, 1057–1072. 30 R. C. Holt, G. J. Gouws and J. Z. Zhen, Curr. Appl. Phys., 2006, 6, 334–339. 31 N. Cho, K. K. Kanazawa, J. S. Glenn and C. W. Frank, Anal. Chem., 2007, 79, 7027–7035. 32 S. X. Liu and J.-T. Kim, JALA (1998-2010), 2009, 14, 213–220. 33 G. Stengel, F. H¨ o¨ ok and W. Knoll, Anal. Chem., 2005, 77, 3709–3714. 34 F. Li, J. H.-C. Wang and Q.-M. Wang, Sens. Actuators, B, 2008, 128, 399–406. 35 C. Larsson, M. Rodahl and F. H¨ o¨ ok, Anal. Chem., 2003, 75, 5080–5087. 36 N. P. Mira, S. F. Henriques, G. Keller, M. C. Teixeira, R. G. Matos, C. M. Arraiano, D. R. Winge and I. S´ a-Correia, Nucleic Acids Res., 2011, 39, 6896–6907. 37 A. R. Fernandes, N. P. Mira, R. C. Vargas, I. Canelhas and I. S´ a-Correia, Biochem. Biophys. Res. Commun., 2005, 337, 95–103. 38 K. Tanaka, Y. Ishii, J. Ogawa and J. Shima, Appl. Environ. Microbiol., 2012, 78, 8161–8163. 39 N. P. Mira, J. D. Becker and I. S´ a-Correia, OMICS: J. Integr. Biol., 2010, 14, 587–601. 40 A. R. Hillman, A. Jackson and S. J. Martin, Anal. Chem., 2001, 73, 540–549. 41 K. M. M. Aung, X. Ho and X. Su, Sens. Actuators, B, 2008, 131, 371–378. 42 J. van Noort, F. Orsini, A. Eker, C. Wyman, B. de Grooth and J. Greve, Nucleic Acids Res., 1999, 27, 3875–3880. 43 R. Levicky, T. M. Herne, M. J. Tarlov and S. K. Satija, J. Am. Chem. Soc., 1998, 120, 9787–9792. 44 H. Ma, J. He, Z. Zhu, B. Lv, D. Li, C. Fan and J. Fang, Chem. Commun., 2010, 46, 949–951. 45 G. Patikoglou and S. K. Burley, Annu. Rev. Biophys. Biomol. Struct., 1997, 26, 289–325. 46 D. S. Lautchman, Eukariyotic Transcription Factors, Elsevier, 4th edn, 2004. 47 Y. Zhang, Z. Xi, R. S. Hegde, Z. Shakked and D. M. Crothers, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 8337–8341.

Analyst, 2014, 139, 1847–1855 | 1855

View Article Online

Analyst


PAPER


Characterization of micro- and nanocapsules for self-healing anti-corrosion coatings by highresolution SEM with coupled transmission mode and EDX V.-D. Hodoroaba,*a D. Akcakayiran,b D. O. Grigorievb and D. G. Shchukinc The observation of morphological details down to the nanometer range of the outer surface of micro-, submicro- and nanoparticles in a high-resolution scanning electron microscope (SEM) was extended with in-depth observation by enabling the transmission mode in the SEM, i.e. TSEM. The micro- and nanocapsules characterized in this study were fabricated as depots for protective agents to be embedded in innovative self-healing coatings. By combining the two imaging modes (upper and indepth observation) complementing each other a better characterisation by a more comprehensive interpretation of the “consistency” of the challenging specimens, e.g. including details “hidden” beyond the surface or the real specimen shape at all, has been attained. Furthermore, the preparation of the

Received 9th September 2013 Accepted 23rd January 2014

quasi electron transparent samples onto thin supporting foils enables also elemental imaging by energy dispersive X-ray spectroscopy (EDX) with high spatial resolution. Valuable information on the elemental

DOI: 10.1039/c3an01717f

distribution in individual micro-, submicro- and even nanocapsules completes the “3D” high resolution

www.rsc.org/analyst

morphological characterization at the same multimodal SEM/TSEM/EDX system.

Introduction More and more submicro- and nanomaterials have to be “quickly”, but accurately characterized with respect to their surface morphology, shape, size and size distribution, and chemical composition as well. A eld emission scanning electron microscope (FE-SEM) attached to an energy dispersive X-ray spectrometer (EDS) constitutes one established methodical approach which may be taken into consideration for such characterization.1 While a modern high-resolution SEM enables qualitative and quantitative (lateral) dimensional evaluation of specimen surface morphology at a scale down to the nanometer range, the high-resolution elemental analysis by X-ray spectroscopy is hindered by the large X-ray generation volumes in the specimen typically in the micrometer range. In many cases the inner “consistency” of particles with complex structure shall be complementary investigated, so that in-depth information by microscopy/imaging and spectroscopy/ chemistry can be provided. One relevant example of specimens with such challenging analytical requirements is innovative a

BAM Federal Institute for Materials Research and Testing, Division 6.8 Surface Analysis and Interfacial Chemistry, Berlin, Germany. E-mail: Dan.Hodoroaba@bam. de; Fax: +49 30 8104 1827; Tel: +49 30 8104 3144

b

Max-Planck Institute of Colloids and Interfaces, Department Interfaces, 14424 Potsdam, Germany

c Stephenson Institute for Renewable Energy, University of Liverpool, L69 7ZD Liverpool, UK

2004 | Analyst, 2014, 139, 2004–2010

micro- and nanocapsules lled with protective agents and incorporated in novel anti-corrosion coatings.2–4 Preparation of the electron transparent submicro- and nanocapsules on thin supporting foils as typically observed for the transmission electron microscope (TEM), i.e. on TEM grids, as well the “activation” of the transmission mode optionally available at a modern SEM, i.e. transmission in SEM (TSEM), make possible a multi-modal analytical approach consisting of high-resolution characterization of the surface morphology, indepth inspection of the “inner” volume of the particles and elemental imaging of the thin specimens by X-ray spectroscopy with high spatial resolution. Especially for particles of sizes and complex morphology and structures as those in this study the combination of the methods above in only one instrument enables a thorough and relatively quick “3D” characterization of morphology, structure and elemental chemistry of individual particles.

Experimental and materials Micro- and nanocapsules of various sizes, structures and chemistry have been fabricated by various approaches as part of extensive projects for development of “smart” depots for healing agents to be embedded in self-healing anti-corrosion coatings.2–4 The “smart” behaviour of the capsules containing the active substance (e.g. corrosion inhibitors and biocides) consists of the ability to regulate their permeability in response


View Article Online


Paper

to the corrosion trigger. In practice, in the case when the new generation anti-corrosion coating is mechanically or chemically damaged, the nearby capsules are activated and release their load into the damaged area, so that the corrosion attack is terminated at an early stage, quickly and locally. In consequence, the durability of the innovative coatings is signicantly prolonged, the environmental sustainability is improved, and the overall performance of the protective coating is also enhanced.2–4 A large variety of capsules with sizes ranging from several tens of mm down to below 100 nm have been fabricated and systematically investigated in the present study. Representative examples were selected and are described in the Results section. In this study we present two types of capsules: polymer based polyepoxide (PE) capsules and SiO2 based inorganic capsules lled with a corrosion inhibitor. These core/shell particles were synthesized from oil in water emulsions as products of polymerisation (PE) or polycondensation (SiO2) reactions wherein the corrosion inhibitor is the oil phase. The PE capsules having 2-methylbenzothiazole as an inhibitor in their core were synthesized as described in ref. 5. The SiO2 based capsules

Analyst

having an organic inhibitor in their core part are synthesized according to the method proposed in ref. 6. As far as the methods are concerned, as already pointed out in the Introduction, the conventional high resolution SEM has been employed mainly for resolving morphological details on the top surface of the micro- and nanocapsules. An FE-SEM of type Supra 40 from Zeiss (Oberkochen, Germany) equipped with an In-Lens detector was used, see Fig. 1a. This type of electron detector is indispensable for high-resolution imaging of the top surface of a specimen in an SEM, due to its ability to collect selectively only the high-resolution, so-called SE1 electrons emitted locally, only at the point of impact of the primary electron beam onto the specimen surface. Further instrumental details can be found in ref. 7–9. In order to gain also valuable in-depth information on the specimen structure beyond the specimen surface, and complementing hence the surface sensitive In-Lens imaging mode, the transmission mode in SEM must be enabled. In spite of the fact that this special operation mode has demonstrated its proof-of-principle in the seventies,10 relevant applications have been reported just recently.1,11–15 In particular, the capability of

Fig. 1 (a) Scheme of an SEM with the common Everhart–Thornley detector collecting the secondary and the backscatter electrons and optionally equipped with a high-resolution In-Lens detector able to collect only the “SE1” electrons. The X-rays emitted by the massive specimen can be collected using an EDS detector; (b) SEM configuration with E–T and In-Lens detectors, with a thin specimen prepared on electrontransparent supporting foils and equipped with a STEM detector (able to operate in bright and dark field modes); the availability of high-sensitive, large-area EDS detectors for submicro- and nanoparticles is advantageous; (c) SEM with the transmission mode in the version of the single-unit setup and (d) a photo of the single-unit transmission setup as used in this study.


Analyst, 2014, 139, 2004–2010 | 2005

View Article Online


Analyst

metrological TSEM measurements of lateral dimensions in the nanometer range, e.g. for purposes of accurate determination of nanoparticles' size and size distribution, has been already successfully exploited. Basically, the so-called STEM detector to be placed immediately under the thin specimen can be used for acquiring images in transmission mode in a SEM, see Fig. 1b.16 In the present study a rather seldom, but commercially available transmission setup has been used, see Fig. 1c and d.17,18 This single-unit transmission setup acts as a specimen holder that enables to guide the transmitted electrons onto an electron multiplier (gold plate) and further to the conventional Everhart–Thornley detector. Further details on the operation and results of this type of producing transmission images at a SEM and comparison to the STEM detector are described elsewhere.13–15 Briey summarizing, the quality of the images obtained by the two modes is practically equally good. It should be also noted that even if working with high-end, high-resolution SEMs in the transmission mode, the spatial resolution which is typical for TEMs cannot be practically attained. The much higher accelerating voltages as well as the aberration correctors available in a TEM make the spatial resolution attainable in a TEM signicantly superior to that of any T-SEM system. Furthermore, by preparing the quasi electron transparent specimens onto thin TEM supporting foils the excitation volume responsible for the generation of X-rays is signicantly reduced, so that the acquired EDX spectra with a conventional EDX detector at a SEM operating in the transmission mode address a high spatial resolution. Of course, especially for nanoobjects, the EDX signals will be very weak. However, the employment of the newly developed large-area SDD (silicon dri detector) EDS can improve signicantly the signal-to-noise ratio and ensures valuable elemental information over the X-ray spectra also for nanoobjects. In this study two EDX spectrometers were employed, an SDD one of type XFlash® 5010 from Bruker (Berlin, Germany) and a Si(Li) one from Thermo Fisher Scientic (Waltham, MA, USA) both with a detector active area of 10 mm2.7–9,19

Paper

The micro- and nanocapsules have been prepared from the suspension or powder form aer corresponding dilution (in distilled water or in isopropanol) with, in some cases, 2 min treatment in an ultrasonic bath, pipetting 1 ml onto various TEM grids (the lacey carbon lms giving the best results) and nally drying. To what extent the specic sample preparation procedure affects the morphology and structure of the specimens investigated could not be established in the present study and remains a challenging task for further investigation. Changes at least in the silica particle size during the analysis in the electron microscope (under electron bombardment and vacuum) cannot be excluded.20 The remarkable powerful analytical characterization of the challenging specimens of complex morphology and structure chosen in this study can be demonstrated in a convincing way by employing the multi-method approach conceived above, namely the high-resolution imaging of the top surface morphology combined with in-depth transmission imaging and X-ray imaging for elemental distribution with high spatial resolution – all in one instrument. To our knowledge there are no other reports published until now about exploiting such a methodical approach. Furthermore, due to the on-going signicant increase of instrumental performance of SEM/EDX systems, such as imaging resolution of SEMs or EDS detectors with a very large collecting solid angle, the authors expect a broad application of high-resolution SEM/TSEM/EDX especially for quick (in comparison to TEM) and accurate characterization of submicro- and nanoparticles.

Results and interpretation Having described above the materials to be characterized as well as the methods and the instrumentation to be applied, representative examples of characterizing analysis were selected and are discussed in the following. Fig. 2 shows in “double pack” the two complementary imaging modes, upper observation with the In-Lens detector and the corresponding TSEM image of the same lateral scanned

Fig. 2 Example of complementary “upper” observation of a single SiO2 submicrocapsule (about 200 nm size) onto a string of the supporting lacey carbon foil using the In-Lens detector (a) and in the transmission mode (b). Note the “shell-like” structure of the particle that can be visualized only in the transmission mode.

2006 | Analyst, 2014, 139, 2004–2010


View Article Online


Paper

outer size. The particle is stuck to a string of the lacey carbon supporting foil onto the TEM grid and has practically no substrate beyond it, i.e. offering hence excellent analysis conditions due to the absence of mostly disturbing interactions with the substrate. While the In-Lens imaging mode supplies the expected high-resolution morphology of the outer surface of the particle, the transmission imaging mode reveals the shell-like structure of the capsule. Neither a conventional SEM alone nor a TEM alone is able to characterize both the nanomorphology of the outer specimen surface and the inner structure of the submicrometer particle. However, a clear statement on the composition of the capsule, whether the core is lled by the inhibitor (as expected by manufacturer) or the capsule is empty, cannot be given. Due to the closeness of the particle to the carbon support strand at which the particle sticks at, an EDX analysis cannot identify reliably the possible presence of carbon in the core of the particle. A somewhat similar tricky arrangement of several submicroparticles sizing from the micrometer down to the nanometer range is shown – again in a double pack as upper and in-depth observations – in Fig. 3. The imaging mode with the In-Lens is able to detect sensitively morphological details on the surface of the bigger particles and also detect small nanoparticles adhered onto the surface of the bigger particles. In turn, the TSEM image reveals such in-depth complementary information that demonstrates the same calotte shape of the two big particles in Fig. 3 spatially oriented in opposite directions. If in the TSEM mode the two particles look very similar, in the In-Lens surface sensitive mode one can observe the inner surface of one particle as well as the outer surface of the other (big) particle. Here again the use of either only a SEM or only TEM could have led to the misinterpretation of the real particle shape. The example in Fig. 4 reveals the real shape of another particle package if both imaging modes are taken into account. Only the conventional SEM imaging (Fig. 4a) suggests massive, nearly spherical particles with ne details on their surface. However, the real shape of the particles becomes evident when the complementary image in transmission mode is evaluated:

Analyst

the bottom particle is not massive, but rather shell-like, i.e. halfcapsules – as reported also elsewhere21 and “hides” also another particle beyond its surface. The result of imaging two neighbour silica particles of completely different structures (consistency) using the dual mode In-Lens/TSEM is presented in Fig. 5a and b, respectively. If the two particles look quite similar in the “upper” (In-Lens) observation mode (Fig. 5a), the difference between them becomes evident in the transmission imaging mode (Fig. 5b). The TSEM image suggests that the smaller particle is either constituted by a material heavier than the big particle or a thick (electron non-transparent) shell or lled with a substance – opposite to the big particle which seems to be “empty”, constituted only by a thin shell. Activating EDX onto the dened areas equally large on the two particles, it is quickly conrmed that the smaller particle is lled with a carbon rich substance and the big particle is mainly constituted by a SiO2 thin shell. A last example continuing the characterisation treatment applied to the same type of particle as those illustrated in Fig. 5 but at smaller magnications, and thus considering a large number of particles, is shown in Fig. 6. Similar features can be recognized in Fig. 6a and b. EDX qualitative elemental maps of selected elements of interest are shown in Fig. 6c (carbon) and 6d (sulphur). If the distribution of carbon is quite clear and can be explained by the presence of carbon in smaller particles as evidence for the successful lling of the capsules with the corrosion inhibitor, the presence of sulphur was not expected as no sulphur contained components were used in the manufacturing process. Therefore, an EDX spectrum has been extracted from a sulphur-“rich” area marked in Fig. 6d. The EDX spectrum shows a weak signal of the S K line, which is, however, at the limit of detection, so that the presence of sulphur cannot be undoubtedly conrmed. This example was specially selected with the purpose to show deliberately also the limits of the methodical approach applied in the present study. Due to the very small amount of substance excited by the electrons which in turn produce X-ray spectra of poor signal-to-noise ratios, the limits of detection of the EDX are rather poor. Nevertheless, the

Fig. 3 Example of complementary “upper” observation of polyepoxide-shell and oil-core particles submicrocapsules using the In-Lens detector (a) and in the transmission mode (b). Note the details on the outer particle surface as well as the presence of many nanoparticles in the left micrograph and the quite different occurrence of the big, bottom particle in the two imaging modes.


Analyst, 2014, 139, 2004–2010 | 2007

View Article Online


Analyst

Paper

Fig. 4 Example of complementary “upper” observation of polyepoxide-shell and oil-core particles using the In-Lens detector (a) and in the transmission mode (b). Arrows mark details visible on the outer particle surface morphology as well as the presence of adjacent nanoparticles in the left micrograph. Note the different occurrence of the big, bottom particle in the transmission mode revealing the real, shell-like structure of the particle, and hiding a smaller particle (indicated by arrow) underneath.

Fig. 5 Upper (a) and in-transmission (b) SEM observation of SiO2 submicrocapsules prepared on lacey carbon foil on TEM grids. Note the clear elemental compositional difference between the two particles in the T-SEM mode (b). The presence of carbon as a main element “filling” the left particle is confirmed by EDX spectra (c). The right particle is constituted mainly by a SiO2 shell. Cu and Zn signals come from the TEM grid. The EDX spectra are normalized to the Si peak so that the carbon peaks can be compared directly.

value of the elemental (distribution) information in such small particles is impressive. There is only reason to come towards the very recent markedly technological developments in the eld of

2008 | Analyst, 2014, 139, 2004–2010

large area SDD EDS detectors. Especially the signicantly larger collection solid angle as well as the quick processing electronics result in considerably higher signal-to-noise ratios in the EDX


View Article Online

Analyst


Paper

Fig. 6 Upper (a) and in-transmission (b) SEM observation of SiO2 submicrocapsules prepared on lacey carbon foils. EDX elemental maps of C K (c) and S K in the range of the detection limit (d) with the corresponding X-ray spectrum (e) from the marked area.

X-ray spectra, so that the high spatial resolution EDX can excellently complete the high-resolution SEM imaging both laterally and in transmission mode in a single instrument as a very powerful analytical tool for quick and accurate characterization of smaller and smaller particles. One basic question that could not be fully answered in the present study is to what extent the morphology, shape, structure and chemical composition of the investigated capsules are inuenced by the sample preparation, high-vacuum or the electron bombardment. The fact that the most larger capsules in Fig. 6 are both somewhat squashed and missing the expected oil


lling can be attributed to a faulty synthesis, but one of the artefacts above cannot be excluded. To our knowledge the literature helping one to elucidate this challenging issue is rather scarce.20,22

Conclusions A multimodal methodical approach has been successfully applied for the advanced characterization of micro-, submicroand nanocapsules lled with protective agents and incorporated them into new self-healing coatings. It is demonstrated by examples on how the complexity of both the morphology and

Analyst, 2014, 139, 2004–2010 | 2009

View Article Online


Analyst

composition of the challenging specimens emphasizes the analytical strengths of each particular method. By a combination of them, a thorough characterizing picture by means of only one measurement system nally results: high-resolution SEM imaging of the top surface of the specimens, in-depth observation of individual particles by the transmission mode in SEM (TSEM), and elemental imaging by EDX with high spatial resolution. The prerequisite for this detailed characterization is the preparation of the quasi electron transparent particles on thin foils as typically observed for TEM and the possibility to take images in transmission mode with a special transmission setup. Hence, the real shape of the capsules could be interpreted better and hidden details beyond the outer surface of the specimens could be visualised. Nevertheless, care must be taken with respect to possible inuences of the sample preparation, vacuum and electron bombardment on the real shape, structure and even chemical composition of the particles under investigation relative to their state immediately aer the particle synthesis.

Acknowledgements The authors would like to thank Mrs Sigrid Benemann (BAM6.8) for taking careful measurements. D. Grigoriev and D. Akcakayiran acknowledge nancial support from the European G8 project “Multiscale smart coatings with sustained anticorrosive action – smart coat”.

Notes and references 1 V.-D. Hodoroaba, S. Rades, N. Kishore, G. Orts-Gil and W. E. S. Unger, Imaging & Microscopy, 2013, 15, 54. 2 D. G. Shchukin, D. O. Grigoriev and H. M¨ ohwald, So Matter, 2010, 6, 720. 3 A. Latnikova, D. O. Grigoriev, H. M¨ ohwald and D. G. Shchukin, J. Phys. Chem. C, 2012, 116, 8181. 4 D. G. Shchukin and H. M¨ ohwald, Small, 2007, 3, 926. 5 D. O. Grigoriev, M. F. Haase, N. Fandrich, A. Latnikova and D. G. Shchukin, Bioinspired, Biomimetic Nanobiomater., 2012, 1, 101. 6 N. Lapidot, O. Gans, F. Biagini, L. Sosonkin and C. Rottman, J. Sol-Gel Sci. Technol., 2003, 26, 67.

2010 | Analyst, 2014, 139, 2004–2010

Paper

7 M. Procop and V.-D. Hodoroaba, Microchim. Acta, 2008, 161, 413. 8 V.-D. Hodoroaba, M. Radtke, L. Vincze, V. Rackwitz and D. Reuter, Nucl. Instrum. Methods Phys. Res., Sect. B, 2010, 268, 3568. 9 V.-D. Hodoroaba and M. Procop, X-Ray Spectrom., 2009, 38, 216. 10 L. Reimer, B. Volbert and P. Bracker, Scanning, 1979, 2, 96. 11 E. Buhr, N. Senleben, T. Klein, D. Bergmann, D. Gnieser, C. G. Frase and H. Bosse, Meas. Sci. Technol., 2009, 20, 084025. 12 T. Klein, E. Buhr, K.-P. Johnsen and C. G. Frase, Meas. Sci. Technol., 2011, 22, 094002. 13 V.-D. Hodoroaba, S. Benemann, C. Motzkus, T. Mac´ e, P. Palmas and S. Vaslin-Reimann, Microsc. Microanal., 2012, 18(2), 1750. 14 V.-D. Hodoroaba, C. Motzkus, T. Mac´ e and S. VaslinReimann, Microsc. Microanal., 2014, DOI: 10.1017/ S1431927614000014. 15 C. Motzkus, T. Mac´ e, F. Gaie-Levrel, S. Ducourtieux, A. Delvallee, K. Dirscherl, V.-D. Hodoroaba, I. Popov, O. Popov, I. Kuselman, K. Takahat, K. Ehara, P. Ausset, M. Maill´ e, N. Michielsen, S. Bondiguel, F. Gensdarmes, L. Morawska, G. Johnson, E. M. Faghihi, C. S. Kim, Y. H. Kim, M. C. Chu, J. A. Guardado, A. Salas, ˚ K. J¨ G. Capannelli, C. Costa, T. Bostrom, A. amting, M. A. Lawn, L. Adlem and S. Vaslin-Reimann, J. Nano Res., 2013, 15, 1919, DOI: 10.1007/s11051-013-1919-4. 16 J. P. Vermeulen and H. Jaksch, Imaging & Microscopy, 2005, 1, 22. 17 U. Golla, B. Schindler and L. Reimer, J. Microsc., 1994, 173, 219. 18 U. Golla-Schindler and B. Schindler, US Pat. 6,815,678 B2, 2004. 19 V.-D. Hodoroaba, K. J. Kim and W. E. S. Unger, Surf. Interface Anal., 2012, 44, 1459. 20 K. Y. Jung, B. C. Park, W. Y. Song, B.-H. O and T. B. Eom, Powder Technol., 2002, 126, 255. 21 V.-D. Hodoroaba, S. Rades and W. E. S. Unger, Surf. Interface Anal., 2014, 46, DOI: 10.1002/sia.5426. 22 F. Tiarks, K. Landfester and M. Antonietti, Langmuir, 2001, 17, 908.


View Article Online

Analyst


PAPER


Non-competitive aptamer-based quenching resonance energy transfer assay for homogeneous growth factor quantification† Kari Kopra,* Markku Syrja ¨npa ä ¨, Pekka Ha ¨nninen and Harri Ha ¨rma ¨ A non-competitive homogeneous, single-label quenching resonance energy transfer (QRET) assay for protein quantification is now presented using lanthanide-chelate labeled nucleic acid aptamers. A labeled ssDNA aptamer binding to a growth factor has been successfully used to provide luminescence signal protection of the lanthanide label. The QRET technology has previously been applied to competitive assay formats, but now for the first time a direct non-competitive assay is presented. The QRET system is based on the protection of the Eu(III)-chelate from a soluble quencher molecule when the aptamer interacts with a specific target protein. The direct QRET assay is possible as the aptamer structure itself cannot protect the Eu(III)-label from quenching. The dynamic range for the optimized vascular endothelial growth factor (VEGF) assay is 0.25–10 nM. A successful quantification of the basic

Received 23rd September 2013 Accepted 23rd January 2014

fibroblast growth factor (bFGF) is also demonstrated using the same QRET assay format with a dynamic range of 0.75–50 nM. These assays evidently show the suitability of the direct QRET technique to simple

DOI: 10.1039/c3an01814h www.rsc.org/analyst

and efficient detection of large biomolecules. The QRET assay can potentially be applied as a detection platform for any other protein targets with a known aptamer sequence.

Introduction Aptamers are traditionally perceived as short (15–100 nt) nucleic acid molecules obtained from large oligonucleotide libraries through a selection method called SELEX (Systematic Evolution of Ligands by EXponential enrichment).1–3 The affinity of aptamers towards their target is normally in the same range as that of antibodies (pM–mM), but they exhibit prominent advantages such as site-specic label conjugation, no batch-tobatch variation, good stability, structure-controlled design, and lower overall costs.4,5 Aptamers have become viable competitors for traditional antibodies, providing the basis for the development of a more comprehensive number of different kinds of assays. An aptamer oen undergoes a conformational change when bound to its target, which enables homogeneous assay development.6–8 Aptamers have the ability to interact specically with their targets, which can be anything from chelated ions and small molecules to proteins and whole cells. Aptamers have also been selected against many growth factor proteins.9–12 Growth factors

Laboratory of Biophysics, Department of Cell Biology and Anatomy, Institute of Biomedicine, University of Turku, Tykistökatu 6 A 5th oor, FI-20520 Turku, Finland. E-mail: khkopr@utu.; Tel: +358 2 333 7063 † Electronic supplementary information (ESI) available: Experimental procedures, spectral characterization, the 30 -VBA–Eu dissociation constant, 50 -FBA–Eu binding kinetics, VEGF titration in serum, and bFGF titration in serum. See DOI: 10.1039/c3an01814h

2016 | Analyst, 2014, 139, 2016–2023

have vital roles in stimulating cellular growth, as well as in cell proliferation and cellular differentiation, when bind to their target receptors. Two important classes of growth factors are vascular endothelial growth factors (VEGFs) and broblast growth factors (FGFs). VEGF has a role in vasculogenesis and angiogenesis. Overexpression of VEGF is also known to enable cancer growth and metastasis. Like VEGF, FGF is an angiogenic factor, which makes both of these growth factors potential anticancer drug targets.13–15 Both proteins, VEGF and FGF, also have a role in the nervous system making them potential players in numerous neuronal conditions like in Alzheimer's disease and Huntington disease.16–18 Growth factors are interesting therapeutic targets as well as important biomarkers showing low or high expression levels in different disease states. Because of a close relationship between growth factors and their acting domains, recognition elements with high affinity and specicity are needed, which makes aptamers attractive alternative binders to meet these demands. Recently, we have developed a signaling method, the Quenching Resonance Energy Transfer (QRET), which enables the use of only one luminescent label in the homogeneous assay format.19 High simplicity is achieved with the single-label QRET system in comparison with many commercial methods. Previously, we have proved the functionality and high throughput capability of the QRET technology in the competitive assay format.19–22 The QRET assay is a viable option for other singlelabel methods like uorescence polarization (FP), which is widely used especially in aptamer based small molecule


View Article Online


Paper

detection, but also in protein detection.20,23–26 However, more reliable data could be obtained with the QRET method than with FP.20 This is due to the high signal-to-background (S/B) ratio in the QRET assay, which ensures the assay reliability also under non-optimal conditions. Fluorescence protection with normal uorophores has been previously reported in a homogeneous uorescence protection assay utilizing antibody induced quenching based on steric hindrance.27 The specic interaction between antibodies and uorophores sets strict requirements for the assay optimization, and especially for the optimization of the label conjugation. Carbon nanostructures (CNs) have also been applied to single-label quenching assays using aptamers.28,29 In these assays, the quenching is based on interactions between ssDNA and CN or between uorophores and CNs. With CNs, there is no need for labeling point optimization, but these assays oen suffer from poor CN solubility and safety concerns. Also, an evident conformational change of the aptamer is needed. A free aptamer is usually in the ssDNA conformation and when it binds to the target the conformation is changed to a partly dsDNA form. This conformation change removes the aptamer from the carbon surface, thus enabling luminescence signal formation.28–31 The conformational change is also needed in label free assays, which are normally less sensitive than assays where a label is conjugated to the aptamer.7,8 With the novel QRET technique, there is no need for either conformational changes of an aptamer or the optimization of the label conjugation. This makes the method more straight-forward than the uorescence protection or the CN protection assays. The QRET assay can also be performed with either a DNA or RNA aptamer. Here, we introduce a direct non-competitive homogeneous QRET protein detection assay using two growth factors (GF) as model analytes. The growth-factor-specic aptamer can interact with its target bringing the lanthanide label to the close vicinity of the protein surface. This interaction forms a steric hindrance, thus protecting the Eu(III)-chelate from quenching by a soluble quencher molecule. The time-resolved luminescence signal is protected only when the growth factor is present. Previously, we have presented similar Eu(III)-chelate protection phenomena in a competitive QRET assay format.19–22 In the current method, no competition is needed which simplies the protocol and increases the assay suitability to detect a wide variety of target molecules. To highlight the suitability of the non-competitive QRET method, two growth factors (VEGF and bFGF) were investigated. The limit-of-detection (LOD) using the single-label QRET method was at a sub-nanomolar level with only a few minutes of total assay time. As a proof-of-principle, this work demonstrates the high sensitivity and selectivity of the QRET technique in the direct non-competitive protein detection assay. The assay also highlights the simplicity of the homogeneous QRET technique for protein detection in buffer and diluted serum.

Experimental section

Analyst

in E. coli, were purchased from Invitrogen (Carlsbad, CA, USA). Nonspecic proteins, human recombinant endothelial growth factors (EGFs), truncate human recombinant vascular endothelial growth factor receptor 1 (VEGFR-1, Met 1–His 687), streptavidin (SA), C-reactive protein binding antibody (Anti-h CRP 6404 SP-2), g-globulins, and bovine serum albumin (BSA) were purchased from R&D systems (Minneapolis, MN, USA), Merck Millipore (Darmstadt, Germany), Kaivogen (Turku, Finland), Medix Biochemica (Kauniainen, Finland), Sigma-Aldrich (St. Louis, MO, USA). The 45-nt nonspecic control oligonucleotide 50 -CTC CAC GCA TAA TCT TAA ATG CTC TAT ACA CTT GCT CAA TTA ACA*-30 , the 26-nt VEGF DNA aptamer (VBA),9 50 *CAA TTG GGC CCG TCC GTA TGG TGG GT*-30 , and the 30-nt bFGF DNA aptamer (FBA),10 50 -*GCG AAG GCA CAC CGA GTT CAT AGT ATC CCA*-30 , were all synthesized by Biomers.net GmbH (Ulm, Germany). All oligonucleotides were either 30 - or 50 -aminolink-C6 modied (*). Pooled normal human serum was purchased from Hytest Ltd (Turku, Finland). The soluble cyanine dye type quencher molecule, Quench V (Q5), and all other quenchers were obtained from QRET Technologies (Turku, Finland). The black assay plate, OptiPlate 384 F, was purchased from PerkinElmer (Groningen, Netherlands). 4-[2-(4Isothiocyanatophenyl)ethynyl]-2,6-bis{[N,N-bis(carboxymethyl)amino]methyl}pyridine europium(III) bearing one additional iminodiacetate coordinating arm was synthesized in our laboratories as described elsewhere.32 All other reagents, including analytical-grade solvents, were purchased from Sigma-Aldrich. Instrumentation The aptamer purication was performed using reversed-phase adsorption chromatography, a Dionex ultimate 3000 LC system from Thermo Fischer Scientic, Dionex (Sunnyvale, CA, USA), and an Ascentis RP-amide C18 column from Sigma-Aldrich, Supelco Analytical (St. Louis, MO, USA). In the QRET assays, time-resolved luminescence (TRL) emission signals were measured at 615 nm, using 340 nm excitation wavelength, and 400 ms delay and decay times. All measurements were made with a Victor2 1420 multilabel counter PerkinElmer Life and Analytical Sciences, Wallac (Turku, Finland). The time-resolved excitation and emission spectra and luminescence lifetimes were measured using a Varian Cary Eclipse uorescence spectrophotometer, Varian Scientic Instruments (Mulgrave, Australia). Excitation spectra were measured from 200 nm to 500 nm using a 20 nm slit, 0.4 ms gate time, 0.1 ms delay time, and 615 nm emission wavelength with a 5 nm slit. Emission spectra were measured from 550 nm to 750 nm using a 5 nm slit, 0.4 ms gate time, 0.1 ms delay, and 340 nm excitation wavelength with a 20 nm slit. Emission lifetimes were measured using 340 nm excitation wavelength with a 20 nm slit, and an emission wavelength of 615 nm with a 20 nm slit and a gate time of 0.005 ms. The oligonucleotide concentration was measured using a NanoDrop ND 1000-spectrophotometer (Thermo Scientic).

Materials

Eu(III)-chelate conjugation to aptamers

The recombinant human VEGF, expressed in human embryonic kidney 293 cells, and the recombinant human bFGF, expressed

Both aptamers, VBA and FBA, have aminolink-C6 modication either at the 50 - or 30 -end. The Eu(III)-chelate was conjugated to


Analyst, 2014, 139, 2016–2023 | 2017

View Article Online


Analyst

either end of the aptamer in 50 mM carbonate buffer, pH 9.8.32 10 nmol of aptamer was used in conjugation with 20-fold molar excess of Eu(III)-chelate in a total reaction volume of 50 ml. Reactions were incubated for 18 h at 37 C in slow shaking. Labeled aptamers were puried and analyzed using reversedphase adsorption chromatography under the following conditions: an eluent system for Eu(III)-chelate labeled aptamers containing 20 mM TEAAc pH 7.0/ACN 100%, linear gradient (1.0 ml min1) from 86 : 14 to 70 : 30 in 25 min. Also in the used control oligonucleotide has aminolink-C6 modication, and it was labeled at the 30 -end, and puried similarly as aptamers above. HPLC fractions were characterized by measuring the Eu(III) concentration, using the DELFIA technique and commercial Eu(III) standard from PerkinElmer Life and Analytical Sciences, Wallac. The oligonucleotide concentration was dened by measuring the absorbance at 260 nm. Growth factor assay The growth factor assays were performed in a 384-well plate using assay buffer containing 0.11 mM KH2PO4, 0.56 mM Na2HPO4, 15.4 mM NaCl, 0.01% Triton-X, and 1 mM MgCl2, pH 7.4. All assays were performed in four replicates except for twelve replicates made for background (no GF) reactions. Proteins (VEGF or bFGF) and aptamers (VBA–Eu or FBA–Eu) were added in 20 ml, and Q5 was added in 10 ml. Final concentrations in 50 ml reaction volume were 0.5–10 nM, 0–750 nM, and 1.4–2.3 mM for the aptamer, growth factor, and quencher, respectively. The target protein and the aptamer were preincubated 10 min before quencher addition and TRL-signal measurement. The TRL-signal was monitored before quencher addition and at various time points aer quencher addition (0– 24 h). The assay was performed similarly in water diluted serum (1–10%), using 3.8–6.0 mM Q5 concentration. The 30 -VBA–Eu specicity was tested with a panel of nonspecic biomolecules. The assay was done using the same protocol as that used for previous assays, 5 nM 30 -VBA–Eu, 1.85 mM Q5, and 100 nM target molecule (EGF, SA, BSA, g-globulins, CRP Ab, VEGFR-1, DNA), except for bFGF which was added in 10 nM concentration in nal 50 ml reaction volume. The control reaction using nonspecic DNA–Eu (5 nM), 1.85 mM Q5, and 100 nM VEGF was also carried out. Assays were performed with four replicates using 10 min incubation time. In all assays the S/B was calculated as mmax/mmin, coefficient of variation (CV%) (s/m) 100, and the LOD was dened as mmin + 3s. The Kd value was calculated using the linearized Cheng– Prusoff equation, IC50 ¼ (([Ki]/Kd) [C]) + Ki.33 The Ki value in the heparin assay was calculated using the Cheng–Prusoff equation, IC50/(1 + [C]/Kd). In all formulae IC50 is the half maximal inhibitory concentration, m is the mean value, s is the standard deviation, C is the aptamer concentration, Ki is the inhibitory constant, and Kd is the dissociation constant.

Results and discussion The direct non-competitive single-label QRET technique is depicted in Fig. 1. The QRET assay is based on different

2018 | Analyst, 2014, 139, 2016–2023

Paper

The principle of direct non-competitive QRET growth factor detection assay. The direct non-competitive QRET assay gives an increasing luminescence signal with increasing target concentrations. The TRL-signal occurs only when the aptamer is bound to its target, and thus the Eu(III)-label near the protein surface is protected from quenching. Fig. 1

luminescence properties of an Eu(III)-chelate label when bound to the target protein.19–22 In the absence of the target protein, the TRL-signal is quenched since soluble quenchers can weakly interact with the Eu(III)-chelate. The energy transfer between the Eu(III)-chelate and Q5 dramatically decreases the Eu(III)chelates' lifetime. The TRL-signal is monitored only when the Eu–aptamer is bound and protected due to the target (Fig. S-1†). Here, we have optimized the non-competitive aptamer-based detection for VEGF and further applied the method for the quantication of another growth factor, bFGF, without further optimization. Eu(III)-chelate conjugation to aptamers Aptamers are attractive binder molecules as they inherently contain multiple sites for specic labeling. Label conjugation at one end of the aptamer is the easiest and most cost effective way to introduce the label to the aptamer structure.27 End-labeling also reduces potential interference with the aptamer structure thus reducing the optimization. These facts led us to select only the end-labeling strategy for the QRET assay platform. For the GF assays, the VEGF binding aptamer (VBA) and bFGF binding aptamer (FBA) were labeled at 30 - or 50 -ends with the Eu(III)chelate.9,10,33 The Eu(III)-conjugated aptamers were readily separated from the unbound label and free aptamer using reverse phase HPLC purication. Puried fractions were characterized measuring oligonucleotide and Eu(III)-chelate concentrations. Based on these characterizations the correct fraction was selected. Retention times 9.1 min and 9.8 min with unlabeled and labeled aptamers were monitored, respectively. Based on the oligonucleotide and the Eu(III) concentrations the labeling efficiency for 30 -VBA–Eu and 50 -FBA–Eu was 38% and 56% and the purity was 89% and 88%, respectively. VEGF assay The QRET platform is based on the low affinity interaction between the Eu(III)-chelate and the quencher, as well as specic high affinity interactions between the aptamer and the target molecule. This leads to a low TRL-signal when only VBA–Eu and Q5 are present, while the addition of VEGF exhibits a signicantly increased TRL-signal. This is due to a changed luminescence lifetime of the Eu(III)-chelate (Fig. S-1†). The assay


View Article Online


Paper

principle was tested using 10 nM of VEGF, and 5 nM of either 30 or 50 -conjugated VBA–Eu. With the 30 -VBA–Eu the signalprotection was signicant, and the recorded S/B ratio was 7.0 (Fig. 2). This indicates that 30 -VBA–Eu has high affinity towards the VEGF, and the 30 -Eu(III)-chelate was protected when bound to the target. The signal protection was also monitored with 50 -VBA–Eu but the S/B ratio monitored was only 2.4. These assays indicate that the 30 -VBA–Eu brings the Eu(III)-label near the VEGF surface without interfering the binding itself. Signal stability was also tested with both aptamers. The TRL-signal stays rather stable over time, and the S/B ratio aer 24 h was still 78% and 63% of maximum with 30 -VBA–Eu and 50 -VBA–Eu, respectively (Fig. 2). Good signal stability indicates that only fairly mild protein denaturation and aptamer degradation occurs over time under current assay conditions. The reason for 50 -VBA–Eu weaker signal protection could be the compromised

Analyst

binding affinity, disordered aptamer structure, or label location in relation to the VEGF surface. However, the exact VBA binding structure to VEGF is not known.9 Homogeneous assays provide ultimately rapid assay performance compared to heterogeneous assay formats. The QRET technique gives a stable TRL-signal which is formed in minutes.19–22 The kinetic performance of the homogeneous QRET assay with both 30 - and 50 -VBA–Eu was found to be the same, and the maximal signals were monitored aer 10 min of incubation (Fig. 2). This indicates that optimal detection time was only dependent on equilibrium formation with the quencher. The aptamer binding to the specic target seems to be extremely fast. The property of fast and high environment specicity of aptamer binding has also been observed before.34,35 All assays with current VBA have been carried out in phosphate buffer; the buffer composition was only slightly re-optimized for this assay.9,34–36 This increases the importance of quencher molecule selection for the QRET assay. In this study, the buffer was selected based on the aptamer and thus the quencher molecule needs to be compatible under these conditions to produce efficient quenching. The panel of different quenchers was screened with 30 -VBA–Eu, and positively charged quencher molecules, especially the ones belonging to the cyanine dye family, showed the most efficient quenching properties. Based on these results the Quench V (Q5), and 10 min incubation time were selected for further QRET assays (Fig. S-1†).

VEGF assay performance

Typical signal-to-background ratios for VBA–Eu in the QRET assay at different time points during 24 h incubation. The results are represented as means SD of four replicates using 10 nM VEGF, and 5 nM 30 -VBA–Eu (black) or 50 -VBA–Eu (red).

Fig. 2

The linear range and LOD were determined from calibration curves to clarify assay performance with the selected 30 -VBA–Eu. The TRL-signal increase was monitored with increasing VEGF concentrations ranging from 0 nM to 50 nM (Fig. 3). The linear relationship with 2 nM 30 -VBA–Eu was found in the concentration range from 0.5 nM to 10 nM, and the LOD calculated was 238 8 pM. The LOD was somewhat lower than the reported Kd value of the aptamer (0.50 0.4 nM).9 To study the

The QRET assay calibration curves for VEGF. The outsets represent the S/B ratios in the QRET assay at increasing VEGF concentrations. The results are shown as means SD of four replicates, with 2 nM (A) or 10 nM (B) 30 -VBA–Eu. The insets represent the linear range in the VEGF assay, respectively. Fig. 3


Analyst, 2014, 139, 2016–2023 | 2019

View Article Online


Analyst

Eu(III)-chelate related effect on the VBA binding affinity, the Kd value for the 30 -VBA–Eu was determined using the linearized Cheng–Prusoff method.33 The Kd value calculated for 30 -VBA–Eu was 0.44 0.4 nM, which was in line with the corresponding literature value (Fig. S-2†).9 Based on these results we can state that the Eu(III)-chelate conjugation has no effect on aptamer binding properties. The performance of the direct single-label QRET assay was also highly comparable to more complicated aptamer assays, and it serves as one of the most sensitive VEGF assays reported using aptamers.37–41 Due to the character of the QRET technique, the unbound Eu(III)-label is quenched, the linear range can be extended by increasing the VBA–Eu concentration without signicantly sacricing the detection limit (Fig. 3). In the QRET assay with 10 nM 30 -VBA–Eu, the linear range from 0.5 nM to 25 nM was monitored, with the LOD of 450 43 pM. Thus the detection range was easily adjustable, and can be subjected to the needs of an assay. The maximal S/B ratios using 2 nM and 10 nM aptamers were 6.7 and 13.1, respectively. One of the important characters of the direct non-competitive QRET assay in the practical use was that no signal decline was observed at very high protein concentrations. This was because of the single binder molecule used. The maximal TRL-signal was monitored when all aptamers were bound to the target, and excess of GF did not change the TRL-signal monitored. This simplies the testing of unknown samples, thus only one measuring point in the calibration curve was needed. Also the “signal-on” protein detection sensor provides a possibility to create more sensitive and reliable assays for different targets.

bFGF assay The proposed QRET concept could be used as a universal detection tool for proteins and other biomolecules that have a known aptamer. To verify this we developed a second assay based on a direct QRET platform using bFGF as a second model analyte. As for the VEGF assay, a bFGF binding aptamer (FBA) was labeled at the 30 - and 50 -ends.10 Aer FBA purication and characterization we performed similar sensitivity and specicity

Paper

tests to those performed with the VBA–Eu earlier. To further demonstrate the simplicity to convert the QRET assay for different proteins, no buffer or quencher optimizations were made. This was possible because nearly similar selection buffers were originally used for VBA and FBA.9,10 Probably, this is not possible with all aptamers without major compromises in aptamer affinity. In this case the bFGF assay was carried out using exactly the same QRET assay principle and conditions as for the VEGF assay before. The bFGF assay showed a high TRLsignal protection capability aer the FBA–Eu binding to the bFGF protein. Unlike in the VEGF assay, both 50 - and 30 -FBA–Eu bound strongly to the bFGF, and showed high TRL-signal protection capability (data not shown). No clear difference could be detected in the performance between 50 -FBA–Eu and 30 -FBA–Eu. Both VBA and FBA bind to a heparin binding pocket of their target growth factor, but compared to the VEGF assay, TRLsignal protection in the bFGF assay was much higher. The maximal S/B ratio monitored in the bFGF assay with 2 nM and 10 nM 50 -FBA–Eu were 62 and 278, respectively (Fig. 4). The high TRL-signal protection capability seems to be a feature of FBA– Eu, and not because of the bFGF target. The results from VEGF and bFGF assays indicate that the central part of the FBA and the 50 -end of the VBA are the most important parts for efficient aptamer binding. When the central part of the FBA bound to bFGF the Eu(III)-label at either 30 - or 50 -end did not interfere with binding but the label was still protected. The 50 -FBA–Eu results in more stable TRL-signal protection over time compared to its 3-labeled counterpart and thus it was investigated in more detail. The maximal TRL-signal protection in the bFGF assay was high, fast, and stable, but the assay was less sensitive compared to the VEGF assay. This was surprising because the reported Kd for FBA against bFGF is in a low pM range.10 Compromised affinity could be due to the fact that the same assay conditions were used as those with VBA–Eu and more optimal conditions could be found for the bFGF assay. However, no further assay condition optimization was performed. Another important reason could be the photo-SELEX background of the FBA.10,42 The linear range with 2 nM and 10 nM

Fig. 4 The QRET assay calibration curves for bFGF. The outsets represent the S/B ratios in the QRET assay at increasing bFGF concentrations.

The results are shown as means SD of four replicates, with 2 nM (A) or 10 nM (B) 50 -FBA–Eu. The insets represent the linear range in the bFGF assay, respectively.

2020 | Analyst, 2014, 139, 2016–2023


View Article Online

Paper


50 -FBA–Eu was 1–50 nM and 2.5–100 nM, and the LODs were 725 40 pM and 998 50 pM, respectively (Fig. 4). To further prove the capability of fast protein detection using the QRET method, the aptamer binding kinetics was measured using a slightly modied assay protocol (Fig. S-3†). The Q5 cannot affect the aptamer–protein interaction, and thus the total assay time can be decreased in 3 min without compromising the sensitivity or specicity. This proves that the assay can be performed also in a single step and real-time data could be obtained. Growth factor specicity The lanthanide luminescence protection was shown to be dependent on the interaction between 30 -VBA–Eu and VEGF as tested with a panel of nonspecic biomolecules: bFGF, VEGFR1, EGF, SA, BSA, CRP Ab, g-globulin, and DNA. Specicity was tested with 5 nM 30 -VBA–Eu and 100 nM of proteins and DNA, except for 10 nM of bFGF. No detectable TRL-signal was monitored indicating the high specicity of the 30 -VBA–Eu (Fig. 5). The bFGF makes an exception of this pattern. The bFGF shares an analogous heparin binding pocket with VEGF.43 The S/B ratios with 5 nM 30 -VBA–Eu and 10 nM growth factor proteins were 7.6 and 3.1 with VEGF and bFGF, respectively. Even though the VEGF heparin binding site differs from motifs in other proteins, it was apparent that VBA was not entirely VEGF specic.43 Even though bFGF gives some TRL-signal protection, the VBA–VEGF interaction was still the main source of the signal. Proteins without the heparin binding site on the other hand cannot interact with VBA–Eu even in high concentrations, which is in accordance with previous ndings.42 Also no TRL-protection was monitored with 5 nM of nonspecic Eu(III)-chelate labeled oligonucleotide even with 100 nM VEGF concentration (data not shown).

Analyst

Because some cross-binding between bFGF and VEGF was monitored with VBA–Eu, the GF specicity was investigated more thoroughly. We analyzed 30 -VBA–Eu affinity against bFGF and similarly 50 -FBA–Eu binding affinity towards VEGF. The LOD for bFGF monitored with 2 nM 30 -VBA–Eu was 7.3 0.7 nM, which was 30-fold higher than with specic VEGF protein (data not shown). Similarly the LOD for 50 -FBA–Eu against VEGF was 17.5 2.1 nM, which was 25-fold higher than for bFGF (data not shown). These assays indicate that a low concentration of the specic target could not be detected sufficiently when a high concentration of other GF was present. Thus more selective aptamers are needed to efficiently separate two GFs from each other. One possibility could be multivalent aptamers, which have been demonstrated to increase the affinity and function compared to the monovalent aptamers.44 In this case, both GF binding aptamers interact with the heparin binding domain (HBD).9,10 To our knowledge there are no previous data for the specicity of either aptamer used, thus we can only speculate whether the lanthanide label decreases aptamer specicity. The VBA used here is a truncated form for a more widely used aptamer (VEa5) and the truncation can affect aptamer specicity although it improves affinity.9,34 Previous publications, with thrombin as a nonspecic target, have indicated that the VEGF–aptamer could have affinity towards some conserved part of HBD, which could be the reason why the bFGF was also recognized.42 Nevertheless, there are rather large differences in HBD in different proteins.43 The low selectivity of FBA can also be partly explained under non-optimal assay conditions, and the Photo-SELEX background of the FBA.10 In the Photo-SELEX, a 5-bromo-20 -deoxyuridine (BrdU) is used as a modied nucleotide, and in the current assay thymidine was used instead of BrdU. This could also affect FBA affinity as well as specicity which are known to be improved with photo-crosslinking.42

Suitability for a universal molecule detection platform

Fig. 5 Specificity of the direct non-competitive QRET assay. The bar plot represents signal-to-background ratios of 5 nM 30 -VBA–Eu with the panel of biomolecules. The specificity was tested using 10 nM VEGF as a control, and 100 nM of nonspecific DNA or nonspecific proteins: bFGF (10 nM), EGF, SA, BSA, g-globulins, CRP antibody and VEGFR-1. Results are shown as means SD of four replicates after 10 min of incubation.


The GF assays prove that the direct non-competitive QRET principle is highly suitable for detecting proteins in buffer. However, the assay should work also in a more complex biological solution, thus diluted human serum was tested. Human serum is one of the most challenging media due to its high protein and ion concentrations. Previously, high dilutions of serum or plasma were needed in homogeneous assays performed with unmodied aptamers.37–41,45–48 This is due to the environmental specicity of aptamers. Unfavorable solution with high ion concentration can remarkably affect aptamer binding properties.36 Also high serum concentrations can affect the excitation and emission of the used label molecule, especially when modied nucleotides are used as signaling molecules.43,45–48 Thus modied aptamers, which are selected directly in serum, should be preferred when possible. Normally in homogeneous aptamer based assays around 1% serum concentration has been used.45–48 With a solid support and a heterogeneous assay protocol, higher serum concentrations are more manageable. Unfortunately, there were no comprehensive data about currently used aptamers in a complex medium.9,10,42

Analyst, 2014, 139, 2016–2023 | 2021

View Article Online


Analyst

Like in most previously used homogeneous assays, both GFs can be detected in 1 : 100 and 1 : 50 diluted serum (1–2% serum) without signicant compromises in the dynamic range or specicity. Nevertheless, the maximal S/B ratios monitored were decreased in higher serum concentrations. The maximal S/ B ratios in the VEGF assay performed with 10 nM 30 -VBA–Eu were 4.3 and 2.6 in 1 : 100 and 1 : 50 diluted serum, respectively (Fig. S-4†). Similarly in buffer the S/B ratio was 13.1. The lower S/ B ratio was due to the higher quencher concentration needed in serum, which was due to the high protein content in the matrix. The selected quencher molecule can bind to membranes and lipophilic biomolecules, which increases the needed Q5 concentration even more. It is also probable that the aptamer affinity was somewhat compromised. With 10 nM 30 -VBA–Eu the quencher concentration used in buffer was 2.3 mM, but 3.8 mM and 5.8 mM Q5 concentrations were needed in 1 : 100 and 1 : 50 diluted serum, respectively. Similar results were observed in the bFGF assay. However, because of the better TRLsignal protection capability with 50 -FBA–Eu, up to 10% serum can be used still providing a measurable signal difference. The maximal S/B ratios in 0%, 1%, 2%, and 10% serum were 278, 43, 27, and 6, respectively (Fig. S-5†). These assays prove the natural robustness of the QRET assay.19–22 For the rst time the QRET technique was proved to work efficiently also in high serum concentrations. However, no undiluted serum can be used due to the negative effect on quenching. Lanthanides as label molecules on the other hand should enable detection also in complex media.49 The GF detection using the non-competitive QRET assay indicates that the assay sensitivity is not sufficient for clinical protein detection. This is a common feature for homogeneous assays, and especially when aptamers are used in homogeneous assays.37–41 For analytical purposes, like detection of circulating GFs, the sensitivity should be at least three orders of magnitude lower. This is impossible to achieve with aptamers used and with the direct QRET GF detection method.19–22,50,51 Still there are numerous other research areas where the proposed direct non-competitive QRET assay could be used with current sensitivity. For example direct measurement of the cell surface receptor amount could be performed using the correct receptor binding aptamer. We have shown before that the QRET method can be used with intact cells and also combined with other methods, which increases the number of possibilities.19–22 It is also possible to use the direct QRET assay in screening, which we have now demonstrated with heparin. Heparin shares the same binding site in GFs as those of VBA and FBA, thus it can compete with aptamers for the GF binding. Based on the competition between heparin and the aptamer, the IC50 value for heparin with 30 -VBA–Eu and 50 -FBA–Eu were 4.0 0.13 nM and 0.6 0.02 nM, respectively (Fig. 6). The Ki value for heparin binding was also calculated with VBA-EU. The Ki value for heparin was 0.17 0.14 nM. This was calculated using the Kd value for 30 -VBA–Eu which was determined using the linearized Cheng–Prusoff method (Fig. S-2†).33 Altogether the QRET technique offers one of the simplest assay methods for GF detection, working also in diluted serum. The direct QRET assay is not limited for one target, but is easily convertible for other targets.

2022 | Analyst, 2014, 139, 2016–2023

Paper

Fig. 6 The competitive heparin binding titration using the aptamer based QRET assay. The competitive binding of the aptamer (30 -VBA– Eu or 50 -FBA–Eu) and heparin in VEGF (red) and bFGF (black) heparin binding domains. The results are shown as means SD of four replicates.

This can be performed without any special requirement for the aptamer structure. The QRET platform gives fast and reliable result in minutes, and the QRET could be possible to convert to monitor binding reactions in real-time in the future (Fig. S-3†).

Conclusions Previously, we have shown the applicability of the QRET technique in a competitive format.19–22 Now for the rst time we have demonstrated a direct non-competitive QRET assay for GF protein detection. The QRET method is an easy, rapid, and inexpensive single-label approach, based on target induced TRL-signal protection. In the QRET assay, time-consuming assay optimization is reduced because only one end-labeled aptamer is needed. Like in most assay formats in the literature, the direct QRET assay does not depend on aptamers' conformational changes but only binding to a specic target. However, aptamers, specically selected for the QRET, could improve the TRL-signal protection. The introduced QRET method is performed using the “bind and detect” principle, and there is no need for complementary oligonucleotides, complex signal measurement or amplication strategies. Assay can also be performed in a single step reducing the total assay time down to 3 min. These abilities make the QRET technique suitable for a high number of aptamer–target pairs, with straightforward optimization, and without multiple rm-error test panels. In this study, we have proved that the QRET technique can be used in a direct non-competitive fashion for GF detection using VEGF and bFGF as model analytes. The direct QRET assay is one of the simplest assay formats for protein detection described so far. The direct QRET assay method could easily be applied to the detection of a variety of protein molecules, like cell surface receptors. Also small molecule detection could be possible in the future, which makes the non-competitive QRET assay useful as a universal detection platform.


View Article Online

Paper

Acknowledgements This work was supported by the Academy of Finland (138584), FP7 Collaborative Project, NANOGNOSTICS (HEALT-F5-2009242264), and the National Doctoral program in Informational and Structural Biology.


Notes and references 1 A. D. Ellington and J. W. Szostak, Nature, 1990, 346, 818–822. 2 C. Tuerk and L. Gold, Science, 1990, 249, 505–510. 3 L. C. Bock, L. C. Griffin, J. A. Latham, E. H. Vermaas and J. J. Toole, Nature, 1992, 355, 564–566. 4 S. D. Jayasena, Clin. Chem., 1999, 45, 1628–1650. 5 S. M. Nimjee, C. P. Rusconi and B. A. Sullenger, Annu. Rev. Med., 2005, 56, 555–583. 6 S. Rowsell, N. J. Stonehouse, M. A. Convery, C. J. Adams, A. D. Ellington, I. Hirao, D. S. Peabody, P. G. Stockley and S. E. Phillips, Nat. Struct. Biol., 1998, 5, 970–975. 7 D. Li, S. Song and C. Fan, Acc. Chem. Res., 2010, 43, 631–641. 8 D. L. Ma, H. Z. He, K. H. Leung, H. J. Zhong, D. S. H. Chan and C. H. Leung, Chem. Soc. Rev., 2013, 42, 3427–3440. 9 H. Kaur and L. Y. Yung, PLoS One, 2012, 7, e31196. 10 M. C. Golden, B. D. Collins, M. C. Willis and T. H. Koch, J. Biotechnol., 2000, 81, 167–178. ¨ 11 L. S. Green, D. Jellinek, R. Jenison, A. Ostman, C. H. Heldin and N. Janjic, Biochemistry, 1996, 35, 14413–14424. 12 T. Saito and M. Tomida, DNA Cell Biol., 2005, 24, 624–633. 13 H. Takahashi and M. Shibuya, Clin. Sci., 2005, 109, 227–241. 14 A. K. Olsson, A. Dimberg, J. Kreuger and L. Claesson-Welsh, Nat. Rev. Mol. Cell Biol., 2006, 7, 359–371. 15 N. Turner and R. Grose, Nat. Rev. Cancer, 2010, 10, 116–129. 16 T. Kiyota, K. L. Ingraham, M. T. Jacobsen, H. Xiong and T. Ikezu, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, E1339– E1348. 17 M. Chiappelli, B. Borroni, S. Archetti, E. Calabrese and M. M. Corsi, Rejuvenation Res., 2006, 9, 485–493. 18 E. Storkebaum and P. Carmeliet, J. Clin. Invest., 2004, 113, 14–18. 19 H. H¨ arm¨ a, A. Rozwandowicz-Jansen, E. Martikkala, H. Frang, I. Hemmil¨ a and P. H¨ anninen, J. Biomol. Screening, 2009, 14, 936–943. 20 H. H¨ arm¨ a, G. Sarrail, J. Kirjavainen, E. Martikkala, I. Hemmil¨ a and P. H¨ anninen, Anal. Chem., 2010, 82, 892–897. 21 K. Kopra, M. Kainulainen, P. Mikkonen, A. RozwandowiczJansen, P. H¨ anninen and H. H¨ arm¨ a, Anal. Chem., 2013, 85, 2276–2281. 22 K. Kopra, Shweta, E. Martikkala, P. H¨ anninen, U. Pet¨ aj¨ aRepo and H. H¨ arm¨ a, Analyst, 2013, 138, 4907–4914. 23 J. A. Cruz-Aguado and G. Penner, Anal. Chem., 2008, 80, 8853–8855. 24 S. Perrier, C. Ravelet, V. Guieu, J. Fize, B. Roy, C. Perigaud and E. Peyrin, Biosens. Bioelectron., 2010, 25, 1652–1657. 25 T. G. McCauley, N. Hamaguchi and M. Stanton, Anal. Biochem., 2003, 319, 244–250.


Analyst

26 M. Song, Y. Zhang, T. Li, Z. Wang, J. Yin and H. Wang, J. Chromatogr. A, 2009, 1216, 873–878. 27 H. Q. Wang, Z. Wu, L. J. Tang, R. Q. Yu and J. H. Jiang, Nucleic Acids Res., 2011, 39, e122. 28 R. Yang, Z. Tang, J. Yan, H. Kang, Y. Kim, Z. Zhu and W. Tan, Anal. Chem., 2008, 80, 7408–7413. 29 Y. He, Z. G. Wang, H. W. Tang and D. W. Pang, Biosens. Bioelectron., 2011, 29, 76–81. 30 S. M. Bachilo, M. S. Strano, C. Kittrell, R. H. Hauge, R. E. Smalley and R. B. Weisman, Science, 2002, 298, 2361– 2366. 31 R. R. Johnson, A. T. Johnson and M. L. Klein, Small, 2010, 6, 31–34. 32 Q. Wang, K. Nchimi Nono, M. Syrj¨ anp¨ a¨ a, L. Charbonniere, J. Hovinen and H. H¨ arm¨ a, Inorg. Chem., 2013, 52, 8461–8466. 33 P. Newton, P. Harrison and S. Clulow, J. Biomol. Screening, 2008, 13, 674–682. 34 H. Hasegawa, K. Sode and K. Ikeburo, Biotechnol. Lett., 2008, 30, 829–834. 35 B. Strehlitz, N. Nikolaus and R. Stoltenburg, Sensors, 2008, 8, 4296–4307. 36 E. J. Cho, J. R. Collett, A. E. Szafranska and A. D. Ellington, Anal. Chim. Acta, 2006, 564, 82–90. 37 H. S. Lee, K. S. Kim, C. J. Kim, S. K. Hahn and M. H. Jo, Biosens. Bioelectron., 2009, 24, 1801–1805. 38 M. J. Gong, K. R. Wehmeyer, H. B. Halsall and W. R. Heineman, Curr. Pharm. Anal., 2009, 5, 156–168. 39 R. Freeman, J. Girsh, A. F. Jou, J. A. Ho, T. Hug, J. Dernedde and I. Willner, Anal. Chem., 2012, 84, 6192–6198. 40 H. Cho, E. C. Yeh, R. Sinha, T. A. Laurence, J. P. Bearinger and L. P. Lee, ACS Nano, 2012, 6, 7607–7614. 41 Y. Li, H. J. Lee and R. M. Corn, Anal. Chem., 2007, 79, 1082– 1088. 42 D. Smith, B. D. Collins, J. Heil and T. H. Koch, Mol. Cell. Proteomics, 2003, 2, 11–18. 43 A. El-Sheikh, C. Liu, H. Huang and T. S. Edgington, Cancer Res., 2002, 62, 7118–7123. 44 P. R. Mallikaratchy, A. Ruggiero, J. R. Gardner, V. Kuryavyi, W. F. Maguire, M. L. Heaney, M. R. McDevitt, D. J. Patel and D. A. Scheinberg, Nucleic Acids Res., 2011, 39, 2458– 2469. 45 Y. Jiang, X. Fang and C. Bai, Anal. Chem., 2004, 76, 5230– 5235. 46 C. W. Chi, Y. H. Lao, Y. S. Li and L. C. Chen, Biosens. Bioelectron., 2011, 26, 3346–3352. 47 X. Wang, J. Zhou, W. Yun, S. Xiao, Z. Chang, P. He and Y. Fang, Anal. Chim. Acta, 2007, 598, 242–248. 48 E. Katilius, Z. Katiliene and N. W. Woodbury, Anal. Chem., 2006, 78, 6484–6489. 49 M. Zhang, H. N. Le, X. Q. Jiang, B. C. Yin and B. C. Ye, Anal. Chem., 2013, 85, 11665–11674. 50 L. Smolej, C. Andrys, S. Perkov´ a, J. Schwarz, D. Belada and P. Z´ ak, Haematologica, 2006, 91, 1432–1433. 51 R. T. Poon, S. T. Fan and J. J. Wong, Clin. Oncol., 2001, 19, 1207–1225.

Analyst, 2014, 139, 2016–2023 | 2023

View Article Online

Analyst


PAPER


A hydrodynamically optimized nano-electrospray ionization source and vacuum interface† M. Pauly,‡*a M. Sroka,b J. Reiss,b G. Rinke,a A. Albarghash,a R. Vogelgesang,a H. Hahne,c B. Kuster,c J. Sesterhenn,b K. Kernad and S. Rauschenbach*a The coupling of atmospheric pressure ionization (API) sources like electrospray ionization (ESI) to vacuum based applications like mass spectrometry (MS) or ion beam deposition (IBD) is done by differential pumping, starting with a capillary or pinhole inlet. Because of its low ion transfer efficiency the inlet represents a major bottleneck for these applications. Here we present a nano-ESI vacuum interface optimized to exploit the hydrodynamic drag of the background gas for collimation and the reduction of space charge repulsion. Up to a space charge limit of 40 nA we observe 100% current transmission through a capillary with an inlet and show by MS and IBD experiments that the transmitted ion beams are well defined and free of additional contamination compared to a conventional interface. Based on computational fluid dynamics modelling and ion transport simulations, we show how the specific shape

Received 26th September 2013 Accepted 11th December 2013

enhances the collimation of the ion cloud. Mass selected ion currents in the nanoampere range available DOI: 10.1039/c3an01836a

further downstream in high vacuum open many perspectives for the efficient use of electrospray ion

www.rsc.org/analyst

beam deposition (ES-IBD) as a surface coating method.

Introduction Electrospray ionization (ESI)1,2 generates intact molecular ions, which are used for mass spectrometry,3,4 ion mobility spectrometry (IMS)5 or for ion beam deposition methods like so landing.6,7 The coupling of an API source to the mass spectrometer's rst vacuum chamber is usually made via a pinhole, or more frequently through a transfer capillary. Its opening diameter denes the gas load on the rst stage of the differentially pumped vacuum system and is thus strongly limited by the available pumping speed. Typically capillaries with an inner diameter of 0.5–0.7 mm and a length of 5–20 cm or pinholes of less than 0.5 mm diameter are used for the initial transfer of ions into vacuum. With moderately sized pumps a pressure gradient of 2–3 orders of magnitude is generated. Upstream of the transfer capillary, the ESI source can be operated with high ionization efficiency. At low ow rates and low concentrations all molecules in a solution can be ionized.8,9

a

Max-Planck-Institute for Solid State Research, Heisenbergstr. 1, D-70567 Stuttgart, Germany. E-mail: [email protected]; s.rauschenbach@f.mpg.de

b

Technical University Berlin, Ernst Reuter Platz 555, D-01300 Berlin, Germany

c

Chair of Proteomics and Bioanalytics, Technische Universit¨ at M¨ unchen, EmilErlenmeyer-Forum 5, D-85354 Freising-Weihenstephan, Germany

d Institut de Physique de la Matière Condensée, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland

† Electronic supplementary 10.1039/c3an01836a

information

(ESI)

available.

See

DOI:

‡ Present address: Institut Charles Sadron, CNRS-University of Strasbourg, 23 rue du Loess, F-67034 Strasbourg, France.

1856 | Analyst, 2014, 139, 1856–1867

Downstream, sophisticated in vacuo ion optics can reach a transmittance of 50–100%.10 During the transfer from atmospheric pressure to vacuum however a major fraction of the ion current is lost. In principle these losses are higher for high ow rates and high concentration of the molecules in the solution. For the ion transfer alone, the present literature suggests a transmittance of up to 20% for low ow interfaces (micro- or nanospray, ca. 500 nL min1),8,11–15 an individual example shows close to 50%.15 Transmission ratios reported for conventional interfaces operated at high ow rates are usually much lower than for nanospray.8,11–13 More oen, the transmittance is known to be low, but is not quantied explicitly. Several loss mechanisms such as space charge expansion, diffusion, transport in the electric eld, or turbulence are suggested, however, only in a few cases were they identied and quantied. For short capillary vacuum interfaces, for instance, it has been shown that a large fraction of the ion current is lost at the rim of the entrance hole and a smaller fraction within the capillary itself, while the loss due to space charge and diffusion becomes more important for long capillaries.13,15 Improving the ion transmission by changing the geometry16 has been approached for instance by using multiple sprayers,17,18 multiple capillaries,19,20 specially shaped electrodes,21 and ared capillary inlets.22–24 Contouring of the eld at the capillary inlet and throughout the entire conductance pathways into vacuum has improved transmission efficiency and overcame some of the space charge limitations.25–29 More recent developments aim at directly inuencing the gas ow by adding gas streams29–33 or by using hydrodynamic devices.34–37 Even the complete avoidance of


View Article Online


Paper

the vacuum transfer by placing the ESI-emitter within vacuum has been demonstrated.9 All these methods do increase the ion transmission into vacuum, some signicantly. Unfortunately, most studies present relative ion intensities for certain compounds rather than electrical currents, which complicates the comparison between different sources and does not allow for absolute quantication of the transmission.34,35,38 Moreover, the calculation of the motion of the background gas in a complex geometry and further its effect on the ion motion are computationally demanding and thus complicate the modelling and predicting of the performance of an ion source. Therefore, the knowledge about the ion transport in API sources remains mostly empirical and there is no fundamental understanding of the principles that could be employed to systematically optimize this key component. Nevertheless, a few studies report absolute current measurements,12–15,39,40 from which transmissions in the range of 1–20% for capillary interfaces are calculated. Most importantly, Lin et al. showed that the losses at the inlet account for 50–99% for various capillaries.39 In contrast, Page et al. reported that 90% of the losses occur within the capillary itself for a short capillary to the emitter distance, whereas 50% of the losses occur at the inlet when the distance between the emitter and the capillary is increased.15 To understand the origin of the ion losses and rationally improve the performance, we initially consider the motion of an ion in an ESI interface close to atmospheric pressure, which is governed by static electric elds, space charge elds, hydrodynamic drag and diffusion, depicted in Fig. 1a. Electric forces originate from the high voltages applied between the emitter and the capillary and depend mostly on the magnitude of the voltage and the geometry of the electrodes in the interface. Additionally, the ion cloud generated by the electrospray plume expands due to the Coulomb repulsion of its own space charge. Space charge forces scale with the charge density and are further inuenced by the specic geometry of the surrounding electrodes. Hydrodynamic forces are caused by the interaction of the particle with the surrounding uid, which is usually air or N2 containing some vaporized solvent. The hydrodynamic force eld strongly depends on the interface geometry dening the uid ow. Furthermore, the acceleration of the ions by electric forces causes ion–gas-collisions that can be described macroscopically by a drag force compensating the electrical force. The resulting dri motion depends on the ion mobility and on the local properties of the gas. Finally, diffusion is dependent on the properties of the gas and the ion itself as well as on the concentration gradients present. Among all these acting forces, only the hydrodynamic drag force can be collimating i.e. it can have a vector component that is pointing towards the axis of the capillary. Electrostatic and space charge forces typically point outwards. This initial consideration shows that an optimization of an atmospheric pressure interface has to be pursued either by reducing the dispersing effects of the electric forces and diffusion or by acting on the uid ow to use it for actively collimating the ion cloud. It becomes further apparent that the reasons for ion losses are the high electric elds required for electrospray ionization and the electric eld of the space charge, while the This journal is © The Royal Society of Chemistry 2014

Analyst

Fig. 1 (a) Scheme of the inlet region of an electrospray interface with the main influencing factors: electric field, flow field, space charge field and diffusion. The forces acting on an ion located in the ion plume (red cloud) are indicated: drag force due to the gas flow (blue), electric forces due to applied voltages (black), Coulomb repulsion of the space charge (red), and diffusion (green). (b) Scheme of the nano-electrospray interface used in the experiments, with the funnel-shaped and flat transfer capillary.

radial velocity component of the uid ow must become zero at the wall and thus cannot be responsible for ion losses. To improve the performance of nano-ESI capillary vacuum interfaces we conducted a study, systematically comparing a specically designed, funnel shaped capillary with a conventional, at capillary geometry of an otherwise identical conguration (Fig. 1b). We reason that the hydrodynamic ow in the funnel will collimate the ion cloud while at the same time reduce the space charge expansion due to an axial acceleration of the ions at the place of their generation and thus signicantly improve the performance. In this study we quantify the ion losses in both types of nano-electrospray (ES) interfaces experimentally by absolute current measurements and interpret the ndings with the help of ion transport simulations based on computational uid dynamic calculations.

Experimental and computational methods Atmospheric pressure interface A stainless steel tube of 6 cm length with an inner diameter of 1.0 mm is the transfer capillary between the ambient

Analyst, 2014, 139, 1856–1867 | 1857

View Article Online


Analyst

environment and the vacuum chamber (see ESI† for further details). It is soldered into a stainless steel cylinder of 8 mm diameter and 1 cm length. Two different capillary types are studied: a conventional shape, in which the stainless steel cylinder at the capillary entrance is at, and a funnel shape, which is craed into the cylinder by electro-erosion (see Fig. 1b). The funnel has an entrance opening of 8 mm in diameter, which smoothly reduces to the capillary diameter along a 10 mm length, while the curvature angle is approaching 0 . The capillary is mounted, electrically connected to a copper cylinder, which contains a 50 W cartridge heater, a Pt100 thermometer and a high voltage (SHV) wire connection (Fig. S1 in the ESI†). This assembly is held by a PTFE socket, electrically insulated from the vacuum chamber. Voltages of up to 5 kV are applied to the whole assembly which is heated to 200 C during measurements. Electrosprays are generated using glass emitters made from pulling capillaries of 75 mm inner diameter and 365 mm outer diameter to create a dened apex of reduced diameter. For the measurement of the spatial acceptance, commercial emitters (uncoated silica PicoTips, 150 mm outer diameter and 10 mm inner diameter, New Objective) are used since they are stable for a very long time. The emitters are cut to 5 cm length and then connected to the spray solution feed through a nger-tight peek union containing a liquid junction electrode to apply the electrospray voltage. This assembly is mounted on a movable stage allowing the precise positioning of the emitter in three dimensions. We dene the zero positions as the capillary axis for the radial direction r and the capillary entrance for the azimuthal direction z (see Fig. 1b).

Transmission measurements For the transmission measurements the described API interface assembly is mounted to a vacuum chamber pumped by a roots pump (Leybold Ruvac WSU 151, 153 m3 h1) backed with a rotary pump (Leybold Trivac D16B, 16 m3 h1). During operation, the pressure, measured at a distant wall in the chamber, is 4 mbar. Solutions of 102 to 105 M Rhodamine B (RhoB) in a 1 : 1 H2O/MeOH solvent mixture are used for the transmission quantication (further experimental details are given in SII†). A syringe pump (Harvard Apparatus Pump 11) is used to set the ow of 0.5– 1 mL min1. High voltages (Vemitter ¼ 1–5 kV) are applied to the liquid junction electrode to generate the electrospray. The ion currents are measured with Keithley 616 electrometers or with equivalently sensitive, self-build current–voltage converters that can be oated on high voltages. Connected to the capillary and to a metal plate placed 1 cm downstream of the capillary exit in vacuum, these instruments are set to a sensitivity of 10 pA. Both the capillary and the detector plate are grounded. The current measured in vacuum is referred to as the transmitted current Itransmitted in the following. The current measured on the capillary is referred to as the loss current or capillary current Icapillary. These electrical currents are caused by analyte ions, but can also be due to charged droplets, clusters, solvent ions or other contaminants. The total emitted current is

1858 | Analyst, 2014, 139, 1856–1867

Paper

calculated as the sum of the transmitted and capillary currents i.e. Iemitted ¼ Icapillary + Itransmitted. In addition, undetected leakage currents Ileak are possible on the ambient pressure and on the low pressure sides of the capillary. However, on the vacuum side, the detector is placed close to the capillary exit, which excludes ions leaving for the pump or chamber wall. The observation of unity transmittance in some cases conrms this. The absence of any leakage current on the ambient pressure side was veried in a separate experiment: using only an emitter and a collector plate of otherwise the same geometry as the interface, the same current was measured at both terminals (see Fig. S4†). Moreover, by placing the emitter at a position where the transmitted current becomes zero, the relationship Iemitted ¼ Icapillary is veried several times during each measurement, which excludes any leakage current on the atmospheric pressure side. As a benchmark for the performance of the capillary interface alone, we dene the transmittance as the ratio of the transmitted current and the emitted current. The overall performance of an ion source moreover depends on the ionization yield as well and is expressed in the sampling efficiency, which is dened as the ratio of the ux of molecular ions entering the vacuum through an aperture or capillary and the ux of molecules entering the source dissolved in a liquid.41 Mass spectrometry and electrospray ion beam deposition By reducing the capillary diameter to 0.75 mm while keeping the same funnel shape the interface is matched to the pressure requirement of 1 mbar in the rst pumping stage of our ES-IBD apparatus,42–44 which contains a linear time-of-ight (TOF) mass spectrometer. This mass spectrometer is used for the verication of the purity of the ion beam. It is mounted in the 4th differentially pumped chamber at a pressure of 107 mbar and has a resolution of m/Dm ¼ 1000 and a dynamic range of three decades. Mass spectra are recorded using a 104 M RhoB solution in 1 : 1 H2O : MeOH solvent, and a 0.01 mg mL1 (z8 107 mol L1) Cytochrome C (CytC) solution in a 3 : 1 H2O : MeOH to which 2% formic acid has been added. To probe the ion beam for contamination that cannot be detected by mass spectrometry, like neutral molecules and clusters or very large charged clusters, we perform scanning tunnel microscopy (STM) imaging of atomically clean Cu(111) samples aer deposition of the CytC ion beam by ES-IBD from the funnel interface. The depositions take place in the sixth vacuum chamber at 1010 mbar and the samples were then transferred in situ to the STM for characterization.42–44 The current is continuously measured on the sample, which gives access to the total charge deposited expressed in pAh (1 pAh corresponds to the charge deposited from a 1 pA ion beam for 1 h, i.e. 3.6 109 C). Computational uid dynamics and ion trajectory simulations The motion of the ions in the interface is simulated using SIMION 8 (ref. 45) with the SDS package46 to include the inuence of a neutral background gas in motion. The simulation calculates the trajectory of an ion fully dened by mass, charge,


View Article Online


Paper

position and velocity in the electric eld generated by metallic electrodes dened by their geometry and applied voltage. The gas ow is described by the two dimensional compressible Navier–Stokes equations of the ideal gas (Section S-III in the ESI†). Standard conditions for air are used. The friction is described by the usual Sutherland law. The simulation builds on characteristic rewriting of the Navier–Stokes equations,47 which works in pressure, velocity and entropy, from which all thermodynamically relevant quantities can be derived. The simulation is run in time until a steady state of the ow is reached. Due to the symmetry of the geometry, the ow simulations are done in two dimensions. In reducing the dimensionality, we do not expect qualitative changes in the characteristics of the gas ow, albeit locally quantitative deviations will occur. Assuming identical conditions in the capillary, the velocities in the inlet section of the two dimensional geometry will be higher, because for a given diameter the same mass ux is distributed over a smaller cross-section. Following the same argument, the gradients for pressure and velocity will be steeper in a 2d simulation as compared to a full 3d model. The most signicant difference caused by the 2d simulation is a suppression of turbulences, which however would only occur within the transfer tube in a gas owing at a constant velocity and not in the interface where the gas is accelerated. The ions for the example trajectory calculations are dened as RhoB singly positive charged ions (m/z ¼ 443) and trajectories are calculated by applying a static voltage of 2.5 kV to the emitter electrode and 0 V to the capillary (Section S-IV in the ESI†). The space charge of the ion beam is not considered in the simulations and only later calculated along specic ion paths to be compared to the other inuences on the ion beam. Therefore the effect of the external electric eld from the voltage applied on the emitter and of the electric eld of the space charge is expressed in velocities. For their calculation the mobility is expressed as K ¼ K0(P0/P)(T/T0) with P the local pressure from the computational uid dynamic calculations and the local temperature T ¼ 300 K considered as constant (P0 ¼ 1000 mbar and T0 ¼ 273 K). The value of the reduced mobility (K0 ¼ 9.2 105 m2 V1 s1) is the value estimated using the SDS package of SIMION from the molecular mass of RhoB.

Results and discussion Ion transmission characteristics Spray voltage, solution concentration and ow rate are the main parameters that determine the ion current generated by an ES emitter. We use the strong dependencies on spray voltage and concentration to vary the emitted current over many orders in magnitude,13,48,49 while the ow rate is held constant at 0.5 mL min1 to avoid unstable sprays caused by too low ow rates and the generation of large droplets caused by too high ow rates. Currents ranging from several picoampere (pA) up to a few mA are generated from RhoB solutions with concentrations between 105 M and 102 M and spray voltages of 1–5 kV (Fig. 2a and b).


Analyst

Emission and transmission characteristics for the interfaces with flat (a, c and e) and funnel-shaped (b, d and f) capillaries measured with 102 M to 105 M RhoB solutions as a function of the emitter voltage. (a and b) The emitted current is the sum of the transmitted and capillary currents (not shown) i.e. Iemitted ¼ Icapillary + Itransmitted (log scale). (c and d) Current transmitted to vacuum (solid curves) and total current emitted (dashed curves). (e and f) Transmittance, calculated as the ratio of the transmitted current and the total current. Fig. 2

The electric eld at the emitter apex is a key parameter in determining the electrospray current. To comparatively investigate the transmission characteristics of conventional and funnel interfaces, emitter positions are selected such that the distance between the emitter apex and the capillary wall is approximately the same in both cases, which results in a similar electric eld distribution and magnitude in the vicinity of the apex. Because the funnel extends into the half space of z < 0 mm, opening up from 1 mm to 8 mm diameter (see Fig. 1) the emitter at position z ¼ 6.5 mm is located within the funnel at a distance of 2 mm to the capillary wall. Thus for the at capillary the equivalent position is z ¼ 2 mm. For these geometric parameters we nd emission current characteristics Iemitted(Vemitter) that are identical for both interface types over the studied concentration range (Fig. 2a and b). If, as in Fig. 2, different emitters are used for the characterization of the two different capillaries, the characteristics differ in current by a factor of less than one order of magnitude, but show otherwise identical behaviour. Our emitter fabrication method has only limited reproducibility. Thus each emitter has

Analyst, 2014, 139, 1856–1867 | 1859

View Article Online


Analyst

a unique geometry, with a slightly different tip diameter, which can cause a shi in the emission current for the same ow rate and concentration. If the same emitter is used with both interfaces, we nd the identical characteristics for both interfaces (see for instance Fig. S3B(a) and (b)†). The transmitted current measured on the metal plate detector in vacuum increases with the emitted current until a maximum value of approximately 20 nA is reached and then stays constant on further increasing the spray voltage (Fig. 2c and d). This current limit exists for both at and funnel-shaped interface types. It further depends on the individual emitter, the transfer capillary diameter, the emitter position, and the solution concentration (see ESI section S-IX†). Although the net current transmitted into vacuum reaches the same top values for both capillary types, the transmittance, dened as the ratio between the current transferred through the capillary and the total emission current, is signicantly different. The at capillary's transmittance never exceeds 20% (Fig. 2e), whereas up to 100% of the emitted current can be measured in vacuum if the capillary with a funnel shaped inlet is used (Fig. 2f). In the measurement shown in Fig. 2, the unity transmittance is observed for the 105 M and 104 M solutions over the entire range of 1–5 kV applied to the emitter, i.e. for emission currents up to 20 nA. In general, with the 1 mm capillary, up to 40 nA can be observed at 100% transmission using an optimal emitter (Fig. S2 in the ESI†). Above this threshold, only the loss current measured on the capillary increases further with the electrospray voltage or the solution concentration, which translates to a decreasing transmittance. It should also be noted that the transmittance of the at capillary is higher than values reported in the literature as well,11–14 because the capillary used here has a diameter of 1 mm, larger than those used for other studies (typically 0.4–0.6 mm). To compare both capillary types and better understand the inuence of the geometry, a transmission and loss current map were measured and a transmittance map was calculated (Fig. 3). This was done at a working point of Vemitter ¼ 2.5 kV. The maximal values for the transmission current for both capillaries are found when the tip is aligned with the capillary axis (r ¼ 0 mm) and is brought as close as possible to the 1 mm diameter capillary entrance at z ¼ 0 mm (Fig. 3a and b). Away from this position in the radial and axial direction the transmitted current decreases. The funnel interface reaches a peak transmission of 30 nA, while for the at capillary only 6.5 nA is detected. The corresponding relative transmittance reaches 100% for the funnel capillary and up to 23% for the at interface. The map of the loss current (Fig. 3c and d) approximately resembles an inverse image of the transmission current map, because the emission current is the sum of transmitted and loss currents. As the emission depends on the electric eld at the emitter apex, the emitted current increases when the distance between the emitter and the capillary is reduced. This occurs in the axial direction for both interfaces and in the radial direction for the funnel interface (see ESI Fig. S3†). For both interface types we see that electrosprays generated at such a small distance to the capillary wall provide higher emitted current but

1860 | Analyst, 2014, 139, 1856–1867

Paper

Fig. 3 Measured transmitted current for the (a) flat and (b) funnel interface and loss current for the (c) flat and (d) funnel interface mapped as a function of the radial position r and the azimuthal position z of the emitter (104 M RhoB, Vemitter ¼ 2.5 kV, Iemitted ¼ 20–30 nA). (e and f) Corresponding normalized transmittance, with the line marking out the volume of at least half maximum transmittance.

also cause signicant losses. Both the absolute current transmitted as well as the relative transmittance are lower in regions close to the capillary wall. The transmittance map (Fig. 3e and f) also offers a way to dene and measure the volume for which the ions are effectively sampled into the transfer capillary. For instance the volumes of at least half of the maximal transmittance are approximately of the same size for both geometries. Hence, positioning of the emitter requires the same precision for both interfaces. However, the absolute intensities transmitted within these volumes differ greatly (Fig. 3a and b). When interpreting the measured current- and relativetransmittance maps, one must keep in mind that the emitter position and the position where the ions are generated are not identical, but rather a plume of macroscopic size is generated downstream the emitter apex, in which gas phase ions are created out of charged droplets.50 The shape of the ES-plume, as well as the distribution of droplets and ions in it, is determined by space charge and external electric elds as well as the hydrodynamic drag. We see for both interface geometries that positioning the emitter close to the wall is one major cause of ion losses, which is certainly due to electric elds, whereas it cannot be decided whether space charge or external elds are the main cause for that. When the emitter is positioned away from the wall, the conned geometry of the funnel leads to a signicant


View Article Online


Paper

improvement of the transmission for emitter positions within the funnel, while the transmission of a current generated in the free space in front of the interface is the same for both interfaces. Summarized, the current measurements show that under conditions of a xed set of parameters, e.g. spray voltage, solution concentration, ow rate and emitter performance, the funnel interface transmits higher currents than the at capillary interface when the emitter is positioned at an optimal position within the funnel. Up to 100% of the ions generated by an electrospray can be transferred to the vacuum with the funnel capillary, whereas the transmittance for the at capillary is limited to 25%. Independent of the geometry of the inlet section, however, beyond a transmission current limit, increasing the electrospray emission current does not increase the transmitted current further and thus the relative transmission decreases. While for the vacuum transmittance of capillary interface values around 20% are reported,11–14 for modied aperture interfaces of larger than usual diameter very high sampling efficiencies of up to 80% can be achieved, which implies an inlet transmittance of at least this value.41 The direct comparison of our experiment with the work of Schneider et al.41 is very insightful as both experiments require larger than the typical pumping speed for the rst vacuum chamber to compensate for the increased gas load from the large diameter capillary or large aperture, respectively. Both experiments use approximately the same solution ow rate of 500 nL min1 and molecules in the same mass range. The solution concentration, however, differs by orders of magnitude. Schneider et al. worked with a solution of 1.6 109 M, which would generate a current of 1.3 pA of the singly charged analyte ions at a given ow rate and 100% ionization efficiency. They estimated that the detected 6.8 105 counts per second relate to a sampling efficiency of 80% and hence a vacuum transmission above that. For the funnel capillary source we achieved 100% vacuum transmission with solutions of 104 M and below, with 103 M we still reach 90%. Assuming that the current relates to analyte ions only, which is shown by mass spectrometry measurements (see section ‘Chemical characterization’), we can relate the maximal current of 40 nA detected in vacuum to a sampling efficiency of 50%, as we estimate 80 nA of analyte ion current from the 104 M solution. For 105 M solutions the maximum transmitted current is 10 nA, which suggests a sampling efficiency of around 100% (see ESI section S-VIII†). These numbers represent an ion source of high intensity and efficiency, provided that the transmitted current consists exclusively of analyte ions. This indeed crucial point will be addressed in detail in the following section. Nevertheless the measurements show that unity transmission of ion currents up to 40 nA to vacuum is possible.

Analyst

nature of the charged species that constitute this current. The enhanced transmittance does represent an improvement only if the performance of the funnel source with respect to the chemical composition and contamination of the ion beam is comparable to conventional sources. In particular, by measuring electric current one cannot distinguish between ionized, desolvated analyte molecules and charged contaminants like ionized solvent molecules, ionized clusters or even charged solvent droplets containing analyte molecules. Their presence, however, can have a negative effect on the performance of a mass spectrometer or an ES-IBD experiment. These concerns were addressed by characterizing the ion beam using mass spectrometry, current transmission of a rf-only quadrupole and ion beam deposition experiments with RhoB and CytC as reference compounds. The combination of these techniques allows us to exclude all relevant types of contaminants, including neutral and charged clusters, droplets, fragments, neutrals and ions other than the analyte. Fig. 4a shows mass spectra of ion beams generated by a nano-ESI source with a funnel capillary interface, acquired with the linear TOF mass spectrometer of our ES-IBD experiment.42–44,51–53 The integration of the funnel capillary did not alter the mass spectra qualitatively as compared to the commercial pneumatically assisted source (Agilent Techn.) used before. However the detected intensity at the TOF detector (ETP 14882) increased to an extent that the dynode voltage had

Chemical characterization We have shown in the previous section that the modied capillary allows transmitting up to 100% of the ions produced at atmospheric pressure up to a threshold current. However, the measurement of a net current does not yield information on the


Fig. 4 (a) Positive ion mass spectra of RhoB (top) and CytC (bottom) recorded with the funnel capillary interface. (b) Ion current in high vacuum as a function of the quadrupole rf-voltage (rf only mode) and as a function of the corresponding lower m/z cutoff.

Analyst, 2014, 139, 1856–1867 | 1861

View Article Online


Analyst

to be signicantly reduced from 2.9 kV to 2.4 kV in order to detect an unsaturated signal. According to the detector specications, this decrease compensates for an increase in intensity of 2–3 orders of magnitude, which ts with the increase in the absolute current detected in the TOF chamber. Solutions of 104 M RhoB in water/methanol, as used for the characterization of the transmission, as well as a solution of 0.01 mg mL1 CytC in water/methanol with 2% formic acid are electrosprayed in positive mode. The RhoB mass spectrum shows one single peak, which can be assigned to the singly charged molecular cation (m/z ¼ 443 Th), whereas the CytC mass spectrum shows the characteristic pattern of several charge states due to multiple protonation (m ¼ 12 384 Da, +10 < z < +18).52,54 In both experiments no other peaks can be detected, which demonstrates the absence of ionized contaminants. However, high mass clusters and other chemical noise that would not yield individual peaks in a mass spectrum can be a signicant fraction of the net current. Thus, in addition to the TOF-mass spectrometry, the ion current has been measured on an electrode plate in the fourth pumping stage (107 mbar) while sweeping the rf-voltage of a quadrupole ion guide operated in rf-only mode (Fig. 4b). In this mode the ion guides have a sharp lower m/z-cutoff, while high mass ions are indiscriminately transmitted. The obtained data are equivalent to a mass spectrum integrated over m/z, in which heavily charged clusters and other high m/z-chemical noise would appear as a current at high rf-voltages, whereas all peaks have vanished from the mass spectra. Moreover, this measurement detects the absolute ion current in the high vacuum stage of the instrument through an electrometer and is thus, in particular, not affected by saturation effects, which allows the determination of the fraction of the current that is transmitted into high vacuum. Above a certain rf-voltage required to establish ion transmission through the quadrupole (approximately 20 V), an ion current of up to 5 nA is detected for both compounds. The current is constant for low rf-voltages. For the RhoB beam, sweeping the quadrupole rf-voltage further engenders a sudden drop of the current to zero at 260 V, corresponding to the suppressed transmission of the RhoB ion, which is conrmed by TOF-MS (upper panel in Fig. 4b). For the CytC beam the decrease is not abrupt due to the distribution of the current into several charge states of different mass-to-charge ratios (lower panel in Fig. 4b). Thus one would expect steps that correspond to the transmission of each charge state, those, however, are smeared out due to the limited resolving power of the quadrupole in rf-only mode. Above Vrf ¼ 450 V the current recorded for CytC is equal to zero as well, coinciding with the disappearance of the lowest charge state peaks from the mass spectrum. In both cases the tolerance is a few picoamps. Thus, based on these and the mass spectrometry measurements, any signicant amount of charged contaminants, in particular charged clusters and droplets, can be excluded. The detected current of 5 nA in high vacuum for both ion beams further shows that a large fraction of the ions, passing through the capillary into the rst vacuum chamber, is further transmitted into the high vacuum of the fourth pumping stage. For the 0.75 mm capillary used in these experiments, a current

1862 | Analyst, 2014, 139, 1856–1867

Paper

limit cut-off around 20 nA was measured (see ESI Fig. S9†), from which we estimate a transmittance for the whole apparatus from leaving the capillary to high vacuum of 25%. This is in agreement with previous measurements on our ES-IBD apparatus using other ion sources. It further conrms that the funnel source emits a well-dened, pure ion beam, while for instance leakage currents, gas discharges or other artefacts misinterpreted as molecular ion current can be dismissed. Thus the data from TOF-MS and quadrupole sweeps measured in high vacuum are representative for the ion beam entering the vacuum system through the capillary. Finally, the CytC ion beam has been deposited on atomically at and clean surfaces of a Cu(111) single crystal in the ultrahigh vacuum (p ¼ 2 1010 mbar) of the 6th pumping stage at low and high coverage and imaged in situ by STM (Fig. 5, see ESI† S-V for experimental details). This measurement is the most sensitive to neutral contamination, which would be detected as adsorbates on the surface when present in small quantity. If a large amount of charged or neutral contaminants would be deposited, the surface quickly became contaminated to a point where STM imaging is entirely impossible as a strongly contaminated surface would not allow for a stable tunnel junction.51,53 Due to the possibility of chemical interaction with the surface,55,56 it is not possible to infer the chemical state of a molecule aer ion beam deposition directly from an STM image alone. It is apparent, however, that only string-shaped structures, characteristic of so-landed unfolded proteins, are observed on an otherwise clean surface.52 The observed morphology, molecular- as well as surface-structures are identical to previous experiments with the same molecules deposited under the same conditions using either a commercial, pneumatically assisted, orthogonal electrospray source52 or a conventional cylindrical ion transfer capillary with an in-line nano-ES. In addition, the observed molecular coverage ts to the amount of deposited charge, which indicates that nothing but CytC molecules were transporting the charge to the surface. The ion beam deposition experiment allows for a determination of the mass-to-charge ratio fully independent of the mass spectrometer. The deposited charge is found by integrating the ion current over the deposition time (15 pAh and 50 pAh for the lowand high-coverage respectively) and the number of molecules is measured by extrapolating the molecular coverage observed with STM to the whole sample (see Fig. 5a and b).40,53 For the presented experiments we nd values between 12 and 21 charges per molecule, which reasonably ts with the corresponding mass spectrum. Since this method includes all charge carriers that arrive at the surface this result allows us to exclude the presence of ions other than those seen in the mass spectra. Moreover, a large number of samples and STM images of similar quality using many different molecules and surfaces were produced since the incorporation of the funnel ion source into the ES-IBD apparatus. The data obtained from current measurements, mass spectrometry, quadrupole sweeps, and nally STM aer IBD in ultrahigh vacuum (UHV) characterize the ion source in many, distinctly different aspects. Their combination provides enough


View Article Online


Paper

Analyst

Fig. 5 STM micrographs of unfolded CytC soft landed in UHV after mass selection on an atomically flat and clean Cu(111) surface. (a) Low molecular coverage (deposited charge 15 pAh). (b) High molecular coverage (50 pAh). (c) Magnified single molecule.

evidence to the matter of whether the transmitted ion current consists of charged droplets or desolvated ions. This is crucial, since capillaries of larger diameter, here 1 mm (or 0.75 mm for the ES-IBD instrument), could be problematic regarding droplet desolvation. The indicators for a well-dened molecular ion beam without droplets, clusters, or contaminants are: ion currents stable over a long time, clean TOF mass spectra, a sharp step in the ion current to zero for the Q2 sweeps, and atomically clean surfaces aer ion beam deposition. In addition we see a clear indication of desolvation in the temperature characteristic of the transmitted current at 125 C. Around this temperature the transmitted current increases from 7 nA to 30 nA, marking the transition of a droplet containing beam to one of gas phase ions (ESI S-VII, Fig. S8†). One might even speculate that the funnel geometry provides a more efficient heating due to the close proximity of the heated capillary wall to the emitter, and hence desolvation works well despite the fact that larger diameter capillaries are used. Finally, we did not detect any sign of an ill-dened ion beam, or droplet-containing ion beams, such as increased pressure in the UHV stages, high currents that do not react to voltages applied to electrostatic lenses, or strongly uctuating, unstable ion currents.

Modelling of the vacuum transfer To understand the origin of both the enhanced transmittance through the capillary associated with the modied inlet shape and the observed current limit, in the following we consider the motion of individual ions through both capillary geometries. Several representative trajectories of individual ions are simulated (Fig. S5 and S6 in the ESI†). The inuence of the space charge is not included in the simulations, but later estimated


based on these trajectories. Rather than modelling the whole system including electrospray, droplet evolution, ion motion and space charge for a large number of charged particles, with our approach we are able to extract data for each of the inuencing factors and compare their effect on the ion transport. Like in the current measurement experiments, the capillaries considered in the simulations have a diameter of 1 mm. The emitter is placed at an axial position of z ¼ 10 mm and z ¼ 5 mm for the funnel shaped and at capillary respectively. A potential of 2.5 kV is dened for the emitter and 0 V for the capillary. The ions are modelled as RhoB singly charged ions with a mass of 443 u. Their starting position is located in the plane 2.5 mm downstream of the emitter, a position where gas phase ions are likely to be generated in an electrospray plume.50,57,58 The rst step in modelling is the uid dynamics calculations for both geometries (ESI S-III†). Fig. 6a and c display the results in the form of pressure, radial and axial velocity. In both cases the gas ows within the 1 mm tube with an approximately constant velocity of 200 m s1 at a pressure of 800 mbar, while the pressure reduction and the velocity increase occur in the funnel- or free-space region, respectively. Once the gas enters the cylindrical part of the capillary the ow becomes choked and turbulences can occur. This is not observed in the simulation results because the two dimensional geometry will not allow for the reproduction of the turbulent ow. In the interface region we do not expect turbulences to occur at all because the gas ow is accelerated, which suppresses turbulences even at Reynolds numbers larger than 4000.59 The ow in the inlet section is characterized by a smooth pressure gradient, evenly distributed across the whole funnel, while the pressure drop is stronger but very much localized at

Analyst, 2014, 139, 1856–1867 | 1863

View Article Online

Paper


Analyst

Fig. 6 (a) Pressure, (b) radial and (c) axial gas velocities as calculated by CFD. Overlaid in (a) the trajectory of individual ions as simulated with 2.5 kV applied on the emitter. Note that in (b) the radial velocity is negative, i.e. the radial gas velocity is focusing the ion beam. (d) Radial velocity of the gas (green), radial velocity due to the external electric potential applied to the emitter (black) and radial velocity due to space charge for currents of 1, 10 and 100 nA (blue, violet and red). (e) Axial velocity of the gas (red) and the ion (bold line in (a–c)).

the orice for the at interface. The large region of pressure gradient corresponds to a large region of moving gas in the funnel inlet with radial velocities of up to 12 m s1 pointing towards the axis. For the at inlet, high velocities are only found close to the orice, yet the velocities are even higher, reaching up to 70 m s1 radial velocity. The distribution of the gas motion intuitively shows how the funnel shape supports the sampling of the generated ions. Ions generated within the large region inside the funnel or close to the capillary entry, respectively, are collimated by the force exerted by the radial gas velocity component, while a high axial velocity drags the ions towards the capillary inlet. In the following step, trajectory simulations are performed to compare the effect of the hydrodynamic drag to that of the electric elds from electrodes and space charge. The ions move

1864 | Analyst, 2014, 139, 1856–1867

in the electric eld generated by the voltage applied to the emitter and are inuenced by the moving background gas. Some trajectories are superimposed in Fig. 6a. It appears that they are closely following the stream lines of the gas ow, hence that the hydrodynamic drag is dominating the ion motion. The potential difference between the emitter and capillary is necessary to generate an electrospray, however the ion trajectory appears not to be strongly affected by the presence of this electrostatic force. As a consequence, without taking space charge forces into account all ions from a starting disc of more than 5 mm diameter are effectively transmitted for both geometries. This would translate in the experimental result as both capillaries displaying 100% transmittance, which is clearly not the case for the at capillary. Losses are thus due to the space charge expansion of the electrospray plume or of the ion beam.


View Article Online


Paper

Analyst

The motion of ions is fully described by the superposition of the electric force, the drag force and the space charge forces. Ion losses due to diffusion can be ignored, because the diffusion length for the residence time of ions in the capillary is in the order of micrometres (ESI S-IVB†), which is negligible compared to the capillary dimensions and the radial displacement due to space charge forces. In the regime of high pressure background gas, the mean free path length of the ion motion becomes extremely low in the order of 30 nm (Section S-IVA in the ESI†). Thus on the macroscopic scale an external force like the electric force on the ion F ¼ Eq does not lead to an accelerated motion, but rather to a dri of constant velocity vd ¼ KE characterized by a mobility constant K, which depends on the properties of the gas, but is independent of the motion of the ion. This motion can be added to the motion of the gas with the velocity vgas leading to v ¼ vgas + K(Eext + Esc) ¼ vgas + vext + vsc

(1)

with Eext being the external electric eld due to the applied voltages and Esc the electric eld of the space charge. Based on the relationship in (1), we can now analyse and compare the inuence of each of the three interactions on the ion's trajectory independently, as all three velocities are directly accessible. The velocity of the gas vgas is directly acquired from the uid dynamics calculations. The velocity due to the external electric eld is calculated by solving the Laplace equation.45 Finally the space charge eld for an ion at radius r is approximated by considering an innitely long uniform beam of radius Rbeam and of current I moving with velocity v.60 This electric eld has only a component in the radial direction, which is proportional to the current and inversely proportional to the ion velocity and the beam size: Esc ¼

Ir 2p3nRbeam 2

(2)

We choose to consider the data for the outermost trajectories, plotted as a bold line in Fig. 6a, as they represent the borderline case in which all interactions are most pronounced and thus losses would occur rst. In Fig. 6d the radial components of the gas velocity (green), the dri velocity due to the external electric eld (black) and the dri velocity induced by space charge (for 1 nA, 10 nA and 100 nA; blue, violet and red) are displayed as a function of the axial coordinate for the ion path plotted as a bold line in Fig. 6a. The axial velocity of the gas and of the ion is shown in Fig. 6e, which is the same for all three currents, since no axial force component of the space charge was considered. For both geometry types only the radial gas velocity has a negative sign, meaning that it has a collimating effect on the ion cloud and it is the only collimating force in the inlet. Also this velocity component is higher than all the other velocities. For the at capillary its magnitude is much higher compared to that of the funnel shape, but the region of this high gas velocity is limited to the close vicinity of the capillary entrance, whereas it extends over a large region for the funnel capillary. The radial dri velocity caused by the external electric eld is always


positive, thus decollimating. Yet, it is only signicantly closer to the emitter for both capillary types and in addition it causes a strong decollimating effect at the very entrance of the at capillary, which is avoided in the funnel inlet due to the fact that the emitter is positioned within the funnel. The dri velocity induced by the space charge is plotted for different currents from 1 nA to 100 nA in Fig. 6d. It is approximately constant along the ion path, and found to scale mainly with the total ion current. Even though beam radius Rbeam, radial particle position r and particle velocity v change over orders of magnitude along the trajectory, they cancel each other's effects on the space charge force (see eqn (2)). Moreover we nd that the radial velocity component due to space charge is almost negligible for currents of 1–10 nA and only reaches values comparable to the gas velocity above 10 nA (for instance, vsc ¼ 1 m s1 for I ¼ 50 nA). In the axial direction the ions in the funnel are transported into the capillary exclusively by the hydrodynamic drag, which is represented by the matching curves for the axial velocities in Fig. 6e. The gap between those two curves for the at interface shows that only the electric eld moves the ions towards to the orice because the axial gas velocity here is almost negligible. The quantitative comparison shows that the hydrodynamic force in the funnel region is strong enough to collimate the ion cloud even for currents as high as 100 nA, while the decollimating external electric force is very weak in this region. For the at capillary, however, high gas velocities are only present within a small region that cannot cover the entire electrospray plume and at the same time the strong electric elds at the rim of the capillary add to decollimation. This is reected in the fact that 100% of the ion current can be transmitted with the funnel shape, whereas only 25% of the ion current is transmitted through the at shape even for currents below the observed space charge limit. In the sampling region of the funnel or in front of the at inlet, the gas velocities exceed the velocities of the space charge expansion for 100 nA by an order of magnitude or more. Thus the inlet cannot be responsible for the observed current limit. Within the 1 mm tube however, the space charge gives rise to the only signicant radial velocity component. An ion at 200 m s1 passes the 60 mm long capillary in 0.3 ms. For an ion at the edge of a beam of 100 nA this leads to a velocity of 2 m s1 and the trajectory deviates by 0.6 mm in the radial direction, which is larger than the capillary radius. Thus a signicant number of ions may be lost on the interior walls of the capillary at such a current value. As this effect is independent of the geometry of the inlet section, it can explain the observed current limit for both interface types. Moreover it is in agreement with our observations for smaller diameter capillaries, which show a reduced current limit for otherwise identical behavior.

Summary and conclusions The high molecular coverage shown in Fig. 5b is obtained aer a deposition time of a few minutes with a current of 1–2 nA under ultrahigh vacuum, whereas the same experiment took several hours with our previous ion source, which usually

Analyst, 2014, 139, 1856–1867 | 1865

View Article Online


Analyst

allowed for 10–100 pA of deposition current in UHV. This documents that the funnel interface is a substantial advancement over conventional and nano-ESI APIs, which is achieved by two major improvements: rst, a higher transmission ratio of up to 100%, and second, an increased net current for otherwise identical conditions like spray voltage and concentration. Currents as high as 40, 20 and 3 nA have been obtained for the 1, 0.75 and 0.50 mm diameter capillaries respectively. The improvement in terms of absolute ionic current in vacuum is comparable to what has been previously obtained with multicapillary inlets.19,20 The latter requires a more extensive modication of the atmospheric interface as compared to the ease of implementation of the funnel-shaped capillary presented here, it however also performs better with high ow rates, due to a larger active volume and better desolvation in the small diameter capillaries. Based on computational uid dynamics calculations and ion trajectory simulations, the evaluation of the different forces acting on the ions shows that the hydrodynamic ow dominates the ion motion at the capillary inlet, while electrostatic forces are of lesser importance. The funnel design generates a large volume of high radial velocity, which allows for efficient sampling and thus for the 100% transmittance observed. For high currents (>20 nA), the space charge force becomes relevant rst in the transfer tube, which is expressed in the absolute current limit. The tests with our electrospray mass spectrometer and ion beam deposition experiment42–44 illustrate that the improved performance of the funnel interface has the potential to enhance sensitivity or selectivity as well as sample utilization and throughput61,62 in mass spectrometry or ion beam deposition applications. Initial tests with a commercial mass spectrometer (i.e. an LTQ Orbitrap) and with small diameter capillaries however showed only small improvements. It seems that the high gas load passing through the large diameter capillaries of 1 mm and 0.75 mm is crucial to generate a collimating ow eld of sufficient gas velocity (Fig. S7 and S9 in the ESI†). While the modications made seem small and fairly simple as compared to conventional nanospray interfaces, the novelty of the funnel source is not only the rather simple adjustment of the inlet geometry of a conventional in-line ion source geometry. Key to the enhanced performance of our funnel source is the fact that the nanospray emitter is placed inside of a funnel that constitutes the inlet of a capillary instead of in front of the at cut capillary orice. This union of the funnel and the emitter represents a conceptual difference of great importance for the design of atmospheric pressure ion sources: the background gas is not only an imposition that has to be removed before the ions can be analysed in vacuum. On the contrary, the gas ow is the most efficient way to control ion transport at high pressures. Thus the ion motion in moving gases represents a highly interesting eld for future theoretical and experimental work, where our modelling approach could contribute to a better understanding of ion transport or to the development of better ion sources, gas phase reactors and high pressure ion

1866 | Analyst, 2014, 139, 1856–1867

Paper

optics.6,8,10,63 For instance, with an enhancement of electrospray ion currents in vacuum into the range of microamps, preparative mass spectrometry could gain technological relevance as a surface coating technique. The fact that we can relate the observed current limit to the capillary and not to the inlet section suggests that shapes could be found that have even better characteristics than the one presented. While the funnel shape could be further improved empirically, the model sketched in this paper can be used for a rational optimization. We show how the ion motion in the interface can be described by simple analytical relations. Together with a fast, precise and reliable CFD calculation of the gas ow this model forms a basis for further optimizations of ion sources. However, since we have seen that already very low gas velocities can strongly inuence the behaviour of the ions, high precision is required and thus CFD will always be the challenging part of such studies despite the simplicity of the geometries.

Acknowledgements The authors thank Dr Hartmut Schlichting for the discussion and the support with the current measurements, as well as Dr Ludger Harnau and Dr Peter B. O'Connor for fruitful discussions.

References 1 J. B. Fenn, M. Mann, C. K. Meng, S. F. Wong and C. M. Whitehouse, Science, 1989, 246, 64–71. 2 T. R. Covey, B. A. Thomson and B. B. Schneider, Mass Spectrom. Rev., 2009, 28, 870–897. 3 B. Domon and R. Aebersold, Science, 2006, 312, 212–217. 4 R. D. Smith, J. A. Loo, R. R. O. Loo, M. Busman and H. R. Udseth, Mass Spectrom. Rev., 1991, 10, 359–452. 5 D. E. Clemmer and M. F. Jarrold, J. Mass Spectrom., 1997, 32, 577–592. 6 G. E. Johnson, Q. Hu and J. Laskin, Annu. Rev. Anal. Chem., 2011, 4, 83–104. 7 J. Cyriac, T. Pradeep, H. Kang, R. Souda and R. G. Cooks, Chem. Rev., 2012, 112, 5356–5411. 8 I. Manisali, D. D. Y. Chen and B. B. Schneider, TrAC, Trends Anal. Chem., 2006, 25, 243–256. 9 I. Marginean, J. S. Page, A. V. Tolmachev, K. Tang and R. D. Smith, Anal. Chem., 2010, 82, 9344–9349. 10 R. T. Kelly, A. V. Tolmachev, J. S. Page, K. Tang and R. D. Smith, Mass Spectrom. Rev., 2010, 29, 294–312. 11 D. R. Zook and A. P. Bruins, Int. J. Mass Spectrom. Ion Processes, 1997, 162, 129–147. 12 A. El-Faramawy, K. W. M. Siu and B. A. Thomson, J. Am. Soc. Mass Spectrom., 2005, 16, 1702–1707. 13 I. Marginean, R. T. Kelly, D. C. Prior, B. L. LaMarche, K. Tang and R. D. Smith, Anal. Chem., 2008, 80, 6573–6579. 14 J. S. Page, I. Marginean, E. S. Baker, R. T. Kelly, K. Tang and R. D. Smith, J. Am. Soc. Mass Spectrom., 2009, 20, 2265–2272. 15 J. S. Page, R. T. Kelly, K. Tang and R. D. Smith, J. Am. Soc. Mass Spectrom., 2007, 18, 1582–1590.


View Article Online


Paper

16 G. T. T. Gibson, S. M. Mugo and R. D. Oleschuk, Mass Spectrom. Rev., 2009, 28, 918–936. 17 P. Mao, H.-T. Wang, P. Yang and D. Wang, Anal. Chem., 2011, 83, 6082–6089. 18 R. T. Kelly, J. S. Page, K. Tang and R. D. Smith, Anal. Chem., 2007, 79, 4192–4198. 19 T. Kim, H. R. Udseth and R. D. Smith, Anal. Chem., 2000, 72, 5014–5019. 20 T. Kim, K. Tang, H. R. Udseth and R. D. Smith, Anal. Chem., 2001, 73, 4162–4170. 21 L. Zhou, L. Zhai, B. Yue, E. Lee and M. Lee, Anal. Bioanal. Chem., 2006, 385, 1087–1091. 22 S. Wu, K. Zhang, N. K. Kaiser, J. E. Bruce, D. C. Prior and G. A. Anderson, J. Am. Soc. Mass Spectrom., 2006, 17, 772–779. 23 D. Prior, J. Price and J. Bruce, US Pat., 6455846, 2002. 24 G. L. Glish and R. M. Danell, US Pat., 6703611, 2004. 25 E. W. Sheehan and R. C. Willoughby, US Pat., 6914243, 2005. 26 R. C. Willoughby, E. W. Sheehan and D. Fries, presented in part at the Proceedings of the 2nd ASMS Conference on Mass Spectrometry and Allied Topics, Nashville, Tennessee, May 23–27, 2004. 27 R. Saf, M. Goriup, T. Steindl, T. E. Hamedinger, D. Sandholzer and G. Hayn, Nat. Mater., 2004, 3, 323–329. 28 R. C. Willoughby and E. W. Sheehan, US Pat., 6943347, 2005. 29 C. Jolliffe, G. Javahery, L. Cousins and S. Savtchenko, US Pat., 7659505, 2010. 30 R. Ramanathan, N. Raghavan, S. N. Comezoglu and W. G. Humphreys, Int. J. Mass Spectrom., 2011, 301, 127–135. 31 Z. Tak´ ats, J. M. Wiseman, B. Gologan and R. G. Cooks, Anal. Chem., 2004, 76, 4050–4058. 32 R. Wang, P. Allmendinger, L. Zhu, A. Gr¨ ohn, K. Wegner, V. Frankevich and R. Zenobi, J. Am. Soc. Mass Spectrom., 2011, 22, 1234–1241. 33 Y. Zhu and K. D. Nugent, US Pat., 8227750, 2012. 34 L. Zhou, B. Yue, D. V. Dearden, E. D. Lee, A. L. Rockwood and M. L. Lee, Anal. Chem., 2003, 75, 5978–5983. 35 A. M. Hawkridge, L. Zhou, M. L. Lee and D. C. Muddiman, Anal. Chem., 2004, 76, 4118–4122. 36 R. B. Dixon, D. C. Muddiman, A. M. Hawkridge and A. G. Fedorov, J. Am. Soc. Mass Spectrom., 2007, 18, 1909– 1913. 37 G. Robichaud, R. B. Dixon, A. S. Potturi, D. Cassidy, J. R. Edwards, A. Sohn, T. A. Dow and D. C. Muddiman, Int. J. Mass Spectrom., 2011, 300, 99–107. 38 R. B. Dixon and D. C. Muddiman, Rapid Commun. Mass Spectrom., 2007, 21, 3207–3212. 39 B. Lin and J. Sunner, J. Am. Soc. Mass Spectrom., 1994, 5, 873– 885. 40 M. Volný and F. Tureˇ cek, J. Mass Spectrom., 2006, 41, 124– 126. 41 B. B. Schneider, H. Javaheri and T. R. Covey, Rapid Commun. Mass Spectrom., 2006, 20, 1538–1544.


Analyst

42 N. Thontasen, G. Levita, N. Malinowski, Z. Deng, S. Rauschenbach and K. Kern, J. Phys. Chem. C, 2010, 114, 17768–17772. 43 S. Rauschenbach, R. Vogelgesang, N. Malinowski, J. W. Gerlach, M. Benyoucef, G. Costantini, Z. Deng, N. Thontasen and K. Kern, ACS Nano, 2009, 3, 2901–2910. 44 S. Rauschenbach, F. L. Stadler, E. Lunedei, N. Malinowski, S. Koltsov, G. Costantini and K. Kern, Small, 2006, 2, 540– 547. 45 D. A. Dahl, Int. J. Mass Spectrom., 2000, 200, 3–25. 46 A. D. Appelhans and D. A. Dahl, Int. J. Mass Spectrom., 2005, 244, 1–14. 47 J. Sesterhenn, Comput. Fluids, 2000, 30, 37–67. 48 M. S. Alexander, M. D. Paine and J. P. W. Stark, Anal. Chem., 2006, 78, 2658–2664. 49 J. F. De La Mora and I. G. Loscertales, J. Fluid Mech., 1994, 260, 155–184. 50 A. Wortmann, A. Kistler-Momotova, R. Zenobi, M. C. Heine, O. Wilhelm and S. E. Pratsinis, J. Am. Soc. Mass Spectrom., 2007, 18, 385–393. 51 S. Kahle, Z. Deng, N. Malinowski, C. Tonnoir, A. FormentAliaga, N. Thontasen, G. Rinke, D. Le, V. Turkowski, T. S. Rahman, S. Rauschenbach, M. Ternes and K. Kern, Nano Lett., 2011, 12, 518–521. 52 Z. Deng, N. Thontasen, N. Malinowski, G. Rinke, L. Harnau, S. Rauschenbach and K. Kern, Nano Lett., 2012, 12, 2452– 2458. 53 S. Rauschenbach, G. Rinke, N. Malinowski, R. T. Weitz, R. Dinnebier, N. Thontasen, Z. Deng, T. Lutz, P. M. de Almeida Rollo, G. Costantini, L. Harnau and K. Kern, Adv. Mater., 2012, 24, 2761–2767. 54 D. S. Wagner and R. J. Anderegg, Anal. Chem., 1994, 66, 706– 711. 55 V. Grill, J. Shen, C. Evans and R. G. Cooks, Rev. Sci. Instrum., 2001, 72, 3149–3179. 56 M. Volný, W. T. Elam, B. D. Ratner and F. Tureˇcek, Anal. Chem., 2005, 77, 4846–4853. 57 R. Wang and R. Zenobi, J. Am. Soc. Mass Spectrom., 2010, 21, 378–385. 58 M. Girod, X. Dagany, V. Boutou, M. Broyer, R. Antoine, P. Dugourd, A. Mordehai, C. Love, M. Werlich, J. Fjeldsted and G. Stafford, Phys. Chem. Chem. Phys., 2012, 14, 9389– 9396. 59 H. Schlichting and K. Gersten, Boundary-Layer Theory, Springer, Berlin, Heidelberg, 2000. 60 P. S. Babu, A. Goswami and V. S. Pandit, Phys. Plasmas, 2011, 18, 103117. 61 M. Bantscheff, M. Schirle, G. Sweetman, J. Rick and B. Kuster, Anal. Bioanal. Chem., 2007, 389, 1017–1031. 62 P. Mallick and B. Kuster, Nat. Biotechnol., 2010, 28, 695–709. 63 Z. Tak´ ats, J. M. Wiseman, B. Gologan and R. G. Cooks, Science, 2004, 306, 471–473.

Analyst, 2014, 139, 1856–1867 | 1867

View Article Online

Analyst


PAPER


Ultrasound-assisted magnetic solid-phase extraction based ionic liquid-coated Fe3O4@graphene for the determination of nitrobenzene compounds in environmental water samples Xiaoji Cao,*a Lingxiao Shen,b Xuemin Ye,a Feifei Zhang,b Jiaoyu Chenb and Weimin Mo*ab An ultrasound-assisted magnetic solid-phase extraction procedure with the [C7MIM][PF6] ionic liquid-coated Fe3O4-grafted graphene nanocomposite as the magnetic adsorbent has been developed for the determination of five nitrobenzene compounds (NBs) in environmental water samples, in combination with high performance liquid chromatography-photodiode array detector (HPLC-PDA). Several significant factors that affect the extraction efficiency, such as the types of magnetic nanoparticle and ionic liquid, the volume of ionic liquid and the amount of magnetic nanoparticles, extraction time, ionic strength, and solution pH, were investigated. With the assistance of ultrasound, adsorbing nitrobenzene compounds by ionic liquid and self-aggregating ionic liquid onto the surface of the Fe3O4-grafted graphene proceeded synchronously, which made the extraction achieved the maximum within 20 min using only 144 mL

Received 14th October 2013 Accepted 14th January 2014

[C7MIM][PF6] and 3 mg Fe3O4-grafted graphene. Under the optimized conditions, satisfactory linearities were obtained for all NBs with correlation coefficients larger than 0.9990. The mean recoveries at two spiked levels ranged from 80.35 to 102.77%. Attributed to the convenient magnetic separation, the Fe3O4-

DOI: 10.1039/c3an01937c

grafted graphene could be recycled many times. The proposed method was demonstrated to be feasible,

www.rsc.org/analyst

simple, solvent-saving and easy to operate for the trace analysis of NBs in environmental water samples.

1. Introduction Nitrobenzene compounds (NBs) are widely used as raw materials and intermediaries in rubber chemicals, pesticides, dyes, explosives and pharmaceuticals. NBs have been designated as priority pollutants by the Chinese Environmental Protection Agency and the United States Environmental Protection Agency due to their known or suspected carcinogenicity, mutagenicity, teratogenicity, or high acute toxicity.1–3 It is a challenge to determine NBs in environmental samples due to their low concentration and various interferences. Therefore, an appropriate sample pretreatment is indispensable. Hitherto, different pretreatment methods, including liquid–liquid extraction (LLE),3,4 liquid-phase microextraction (LPME),1 solid-phase extraction (SPE)5–8 and single drop microextraction (SDME)9,10 have been applied to extract NBs from environmental samples. However, these methods are oen tedious, unfriendly due to the

a

Research Center of Analysis and Measurement, Zhejiang University of Technology, Hangzhou, 310014, China. E-mail: [email protected]; [email protected]; Fax: +86 571 88320961; Tel: +86 571 88320085

b

College of Chemical Engineering and Material Science, Zhejiang University of Technology, Hangzhou, 310014, China

1938 | Analyst, 2014, 139, 1938–1944

usage of toxic organic solvents, or unstable and complicated extraction procedures.11,12 Ionic liquids (ILs) composed of various types of organic cations and anions have been extensively applied to analytical chemistry as environmentally benign solvents because of their low vapor pressure, excellent thermal stability, and good dissolving capacity of numerous compounds.12–19 The selfaggregation behavior of imidazolium-based ILs makes their successful application as adsorbents in magnetic mixed hemimicelles solid phase extraction, which combines the advantages of both ILs and magnetic materials.20,21 It is well known that the magnetic material plays an important role in magnetic mixed hemimicelles solid phase extraction because it not only affects the self-aggregation process of the ILs but also determines the loading capacity. Fe3O4-graed graphene (Fe3O4@G) possesses a large surface area and superparamagnetism, and thus is one of the considered alternative supported materials for ionic liquids aggregation. However, application of the ultrasound assisted magnetic solid-phase extraction (UAMSPE) technique with ionic liquid-coated Fe3O4@G (IL–Fe3O4@G) as an adsorbent has not been reported for the simultaneous determination of ve NBs in environmental samples until now.


View Article Online

Paper

Analyst


In this paper, the potential application of the UAMSPE with 1-heptyl-3-methylimidazolium hexauorophosphate [C7MIM] [PF6] ionic liquid-coated Fe3O4@G as an adsorbent to extract NBs from water samples was investigated. The predominant experimental factors affecting the extraction efficiency were studied. As a result, a convenient, solvent-saving method for the analysis of ve NBs in water samples was established by coupling the UAMSPE technique based on ionic liquid-coated Fe3O4@G to high performance liquid chromatography (HPLC).

2.

Experimental

2.1. Reagents and materials The pharmaceutical standards of 1,3,5-trinitrobenzene (purity 97.5%), 2,6-dinitrotoluene (purity 100%), 2-nitrotoluene (purity 99.0%), 3-nitrotoluene (purity 98.7%), and 4-nitrotoluene (purity 99.2%) were purchased from Accu Standard (New Haven, CT, USA). The working stock solution containing 10.00 mg L1 of each NB was prepared in HPLC-grade methanol and stored at 4 C in the refrigerator until use. The working solutions were prepared by diluting the working stock solution with methanol to the corresponding concentrations. 1-Tetradecyl-3-methylimidazolium tetrauoroborate [C14MIM][BF4], l-butyl-3-methylimidazolium perchlorate [C4MIM][ClO4], 1-hexyl-3-methylimidazolium hexauorophosphate [C6MIM][PF6] and 1-heptyl-3-methylimidazolium hexauorophosphate [C7MIM][PF6] were obtained from Chengjie Co., Ltd (Shanghai, China) and used as received. HPLC-grade methanol, ethanol and acetone were purchased from Merck (Darmstadt, Germany). Ultrapure water was prepared using a Milli-Q water purication system (18 MU, Millipore, Bedford, MA, USA). The industrial-grade graphene pulp (5%) was supplied by Ningbo Institute of Materials Technology & Engineering (Ningbo, China). Ammonium iron(II) sulfate hexahydrate (Fe(NH4)2(SO4)2$6H2O) and ammonium iron(III) sulfate dodecahydrate (FeNH4(SO4)2$12H2O) were obtained from Aladdin Chemistry Co., Ltd. The analytical-grade sodium chloride was purchased from Shanghai No.4 Reagent & H.V. Chemical Co., Ltd (Shanghai, China). The hydrochloric acid (analytical grade), sodium hydroxide (analytical grade) and dichloromethane (analytical grade) were supplied by Huadong Medicine Group Co., Ltd (Hangzhou, China). An N50-grade NdFeB magnet (50 mm 50 mm 10 mm) was used for magnetic separation, which was purchased from Guanneng Magnetic Co., Ltd (Yinzhou, Ningbo, China). Water samples, containing river water (the Grand Canal), lake water (the West Lake) and rain water, were all collected in the local area. All the water samples were stored at 4 C in the refrigerator until use.

The Fe3O4 nanoparticles were prepared based on the previous method.22 The synthesis of Fe3O4-graed multi-walled carbon nanotubes (Fe3O4@M) were performed according to the literature reported previously.23,24 The sizes and morphologies of the Fe3O4@G, Fe3O4 and Fe3O4@M magnetic nanoparticles (MNPs) were investigated by transmission electron microscopy (Tecnai G2 F30 S-Twin, Philips-FEI, Netherlands). And their phases were analyzed using an X0 Pert PRO X-ray diffraction instrument (PAN˚ alytical, Netherlands) with Cu Ka radiation (l ¼ 1.54056 A)

over the angular range from 10 to 80 at room temperature. Magnetic properties were analyzed using a Lake Shore 7410 vibrating sample magnetometer (Lake Shore Cryotronics, Inc., USA) at room temperature with a cycling the eld from 15 to 15 kOe. Images of the outer surface of the [C7MIM] [PF6] ionic liquid-coated Fe3O4@G were obtained using a Hitachi scanning electron microscope, model S-4700, equipped with a Thermo Noran Vantage ESI energy dispersive analytical system. 2.3. MSPE procedure

MNPs and ILs were added to 10 mL water displaced in a 20 mL glass vial and dispersed under ultrasound (KQ-500 DA ultrasonic water bath, Kunshan, Jiangsu, China) for 20 min. Then the IL–Fe3O4@G containing NBs was gathered to the bottom of the vial using an N50-grade NdFeB magnet and the supernatant was discarded. Next, the NBs were eluted from the IL–Fe3O4@G with 0.3 mL of methanol by ultrasound for 15 s. Finally, the Fe3O4@G nanoparticles were isolated easily from the eluent using the magnet, and the eluent was ltered through a 0.22 mm nylon lter membrane for the subsequent HPLC analysis. And Fe3O4@G was recycled by washing with 1 mL acetone and 1 mL ethanol respectively, twice. Illustration of the whole procedure is shown in Scheme 1. 2.4. Liquid–liquid extraction 500 mL of sample solution was extracted with 25 mL of dichloromethane two times. The extracts were combined and then evaporated to dryness. The residue was dissolved in 0.8 mL of dichloromethane and for the subsequent GC/MS analysis according to the Chinese Environmental Protection Agency standard HJ 592-2010.

2.2. Synthesis and characterization of MNPs The Fe3O4@G nanoparticles were synthesized by the in situ chemical coprecipitation of Fe2+ and Fe3+ in an alkaline solution in the presence of graphene pulp according to the literature method.22


Scheme 1

Schematic diagram for the IL–Fe3O4@G-MSPE procedure.

Analyst, 2014, 139, 1938–1944 | 1939

View Article Online


Analyst

Paper

Fig. 1 (A) TEM image of Fe3O4@G; (B) XRD pattern of Fe3O4@G; (C) magnetization curve of Fe3O4@G; (D) EDS image of IL–Fe3O4@G; (E) SEM image of IL–Fe3O4@G.

2.5. Determination by HPLC and GC/MS HPLC analysis was performed using a Waters 2695 HPLC system (Waters, Milford, MA, USA) equipped with a quaternary solvent delivery system, an auto-sampler with a 100 mL sample loop, a PDA detector and a data station running the Empower data soware. A CAPCELL PAK C18 column (4.6 mm i.d. 150 mm, 3 mm, Chuo-ku, Tokyo, Japan) was used for separation and was maintained at 30 C. The mobile phase was a methanol–water mixture (58/42, v/v) delivered at a ow rate of 0.6 mL min1. The injection volume was 10 mL and the detection wavelength was 254 nm. The extractant of the liquid–liquid extraction method was analyzed using a 7890A gas chromatography interfaced to a 7000B triple quadrupole mass spectrometer system (Agilent Technologies, Palo Alto, CA, USA) in the MS mode as described in the HJ 592-2010 standard. The temperatures of the EI ionization source and the transfer line were both set at 230 C.

3.


3.1. Characterization of the prepared Fe3O4@G and IL–Fe3O4@G The size and distribution of magnetite on the surface of graphene were examined by transmission electron microscopy (TEM). It can be seen from the TEM image shown in Fig. 1A that the Fe3O4 nanoparticles with a size of about 7–21 nm are well decorated on

1940 | Analyst, 2014, 139, 1938–1944

Fig. 2 (A) Effect of types of MNP. [C6MIM][PF6], 108 mL; MNPs, 3.0 mg; extraction time, 15 min; salt addition, 0%. (B) Effect of types of ionic liquid. ILs, 108 mL; Fe3O4@G, 3.0 mg; extraction time, 15 min; salt addition, 0%.

the surface of the graphene sheets. Some Fe3O4 nanoparticles are slightly aggregated due to the loading degree close to saturation, which matches up with that reported in other works.25 To conrm the integration of Fe3O4 particles with graphene, the phase of the synthesized Fe3O4@G was investigated by X-ray diffraction (XRD). As shown in Fig. 1B, all the characteristic diffraction peaks of the pure cubic spinel crystal Fe3O4 (JCPDS card 03-065-3107) at 2q ¼ 30.1 (220), 35.5 (311), 43.4 (400), 54.2 (422), 57.2 (511) and 62.8 (440) were observed, indicating that the material contains the Fe3O4 phase. The magnetic characteristics of the synthesized Fe3O4@G were evaluated by vibrating sample magnetometer at 300 K and the magnetization curve is illustrated in Fig. 1C. Neither remanence nor coercivity was observed in Fig. 1C, indicating that the synthesized Fe3O4@G magnetic particles are superparamagnetic, which ensures recycling of the magnetic adsorbent conveniently. And the saturation magnetization was found to be 35.93 emu g1, which is high enough for magnetic separation using an external magnet. The scanning electron microscopy (SEM) image of the IL– Fe3O4@G is shown in Fig. 1E. In contrast to Fe3O4@G, there is


View Article Online

Paper

Analyst

an additional mist-like layer on the surface of the IL–Fe3O4@G (see Fig. 1E), and nitrogen, uorine and phosphorus atoms were detected on the membrane by energy-dispersive X-ray spectroscopy (EDS) (as shown in Fig. 1D), which indicated that the [C7MIM][PF6] ionic liquid self-aggregated on the whole surface of the Fe3O4@G.


3.2. Optimization of extraction conditions

Fig. 3 (A) Effect of the volume of [C7MIM][PF6]. Fe3O4@G, 3.0 mg; extraction time, 15 min; salt addition, 0%. (B) Effect of the amount of Fe3O4@G. [C7MIM][PF6], 144 mL; extraction time, 15 min; salt addition, 0%.

Fig. 4 Effect of extraction time. [C7MIM][PF6], 144 mL; Fe3O4@G, 3.0 mg; salt addition, 0%.


In order to achieve a satisfactory extraction performance of the proposed magnetic solid-phase extraction based ionic liquidcoated Fe3O4@G (IL–Fe3O4@G-MSPE) procedure for the ve NBs, several parameters, including the types of MNPs and ILs, volume of IL, amount of MNPs, extraction time, ionic strength, and solution pH were examined and optimized. Ultrapure water samples spiked with 30 mg L1 of each NB were used for these optimization experiments. And all the experiments were performed in triplicate. 3.2.1. Screening of MNPs and ILs. The high loading capacity to ionic liquid aggregation is one of premises for achieving the desired extraction efficiency of the proposed IL–Fe3O4@G-MSPE procedure. In this study, three different MNPs, including Fe3O4@G, Fe3O4@M, and Fe3O4 NPs, were investigated in terms of their loading capability to ionic liquids. As illustrated in Fig. 2A, the integrated extraction efficiency of IL–Fe3O4@G was the most satisfactory, attributed to the strongest adsorptive capacity to the [C6MIM][PF6] ionic liquid used, which resulted from the large delocalized p-electron system and the large surface area of [email protected],26 In consideration of the low cost of the graphene pulp, the Fe3O4@G was employed in the following experiments. Choosing a suitable ionic liquid to decorate the Fe3O4@G surface is vital to improve the extraction efficiency of the proposed IL–Fe3O4@G-MSPE procedure. It is well known that the structure of the IL has a signicant inuence on its physicochemical properties, which might greatly affect the extraction efficiency of the targeted analytes.1 To obtain a high extraction efficiency, four kinds of ionic liquid including [C14MIM][BF4], [C4MIM][ClO4], [C6MIM][PF6] and [C7MIM][PF6] were investigated. As shown in Fig. 2B, the extraction efficiency of the ve NBs increased in the order of [C4MIM][ClO4], [C6MIM][PF6] and [C7MIM][PF6] due to their decreasing water miscibility, which facilitated the self-aggregation of the ILs on the Fe3O4@G surface. In the case of [C14MIM][BF4], the extraction efficiency was nearly zero. The reason may be attributed to the fact that the [C14MIM][BF4] ionic liquid is solid at room temperature, which prevents it selfaggregating effectively on the Fe3O4@G surface. Thus, [C7MIM][PF6] was selected for all of the subsequent experiments. 3.2.2. Effect of the volume of IL and the amount of Fe3O4@G. Considering that the quantity of IL aggregated on Fe3O4@G was another crucial factor related to the extraction efficiency, different volumes of [C7MIM][PF6] (from 0 to 180 mL) were evaluated. As shown in Fig. 3A, without adding [C7MIM][PF6], the extraction efficiency was very poor, indicating that pure Fe3O4@G is not an ideal adsorbent for NBs. When the [C7MIM][PF6] was added into the solution, the extraction efficiency was

Analyst, 2014, 139, 1938–1944 | 1941

View Article Online

Analyst

Paper Performance characteristics of the IL–Fe3O4@G-MSPE method combined with HPLC under the optimum conditions

Table 1


RSD (%) Analyte

Linearity/mg L1

r2

LOD/g L1

Intradaya

Interdayb

Recovery (%)

1,3,5-Trinitrobenzene 2,6-Dinitrotoluene 2-Nitrotoluene 3-Nitrotoluene 4-Nitrotoluene

15–3000 7.5–1500 15–3000 15–3000 15–3000

0.9990 0.9995 0.9993 0.9991 0.9990

1.35 2.97 3.55 4.57 3.95

1.19 2.18 2.23 1.93 1.36

1.21 4.08 2.26 1.28 1.64

96.80 2.61 100.12 1.86 95.71 4.89 98.30 2.67 88.16 0.75

a The intraday RSD was gained by analyzing a sample three times in one day. b The interday RSD was received by analyzing a sample once a day over three consecutive days

Typical chromatograms of spiked and blank samples obtained at wavelength 254 nm: (1) 1,3,5-trinitrobenzene, (2) 2,6-dinitrotoluene, (3) 2-nitrotoluene, (4) 3-nitrotoluene, (5) 4-nitrotoluene.

Fig. 6 Fig. 5

Effect of the reuse times of Fe3O4@G on the recoveries of NBs.

improved dramatically and achieved a maximum with 144 mL [C7MIM][PF6]. The reason could be that the loading capability of Fe3O4@G to [C7MIM][PF6] reached saturation at 144 mL. Further addition of [C7MIM][PF6] resulted in the formation of free micelles in the water solution, which would adsorb a part of NBs and lead to the loss of extraction efficiency. Therefore, 144 mL of [C7MIM][PF6] was selected for the further study. On the other hand, the amount of Fe3O4@G (from 1.0 mg to 9.0 mg) was also studied. As can be seen from Fig. 3B, the maximum extraction efficiency of the NBs was achieved when 3.0 mg of Fe3O4@G was used. When the amount of Fe3O4@G was more than 3.0 mg, the extraction efficiency decreased dramatically, resulting from the permanent adsorption of the excess Fe3O4@G NPs into the NBs. Finally, 3.0 mg was chosen in the following experiments.

Table 2

3.2.3. Effect of extraction time. The main role of ultrasonication is to disperse the [C7MIM][PF6] IL-coated Fe3O4@G extractant into the sample solution thoroughly and accelerate the mass transfer between the [C7MIM][PF6] aggregation and the target analytes.27 The ultrasound-assisted extraction time of target compounds was therefore investigated in the range of 10– 30 min. The results shown in Fig. 4 indicate that the extraction efficiency achieved a maximum at the extraction time of 20 min. The main reason may be that an adequate extraction time was needed to ensure the equilibrium of extraction. When the time was longer than 20 min, a part of the IL carrying NBs was desorbed from the surface of the Fe3O4@G MNPs, which would cause a decrease of the extraction efficiency. Consequently, the ultrasonication time of 20 min was enough.

Comparison of the proposed method with other methods for determination of NBs

Method

Extraction time/min

Sample volume/mL

Extraction solvent

Extraction solvent amount

LODs/mg L1

LLE SPE6 SDME9 LPME1 IL–Fe3O4@G-MSPE

12 90 120 30 20

500 500 1 8 10

Dichloromethane Methanol, ethyl acetate Micellar ionic liquid [C4MIM][PF6] [C7MIM][PF6]

50 mL 17 mL 6.5 mL 3 mL 144 mL

0.88–3.47 0.01–0.29 10.3 1.38 1.35–4.57

1942 | Analyst, 2014, 139, 1938–1944


View Article Online

Paper Table 3

Analyst Analytical results of water samplesa


Recovery (%) Analyte

Spiked level/mg L1

River water

Lake water

Rain water

1,3,5-Trinitrobenzene

0 30 300

nd 97.23 1.13 87.76 4.04

nd 84.59 3.95 98.98 3.11

nd 87.71 4.27 89.59 1.22

2,6-Dinitrotoluene

0 15 150

nd 96.01 1.24 97.34 4.37

nd 83.05 1.27 97.13 3.88

nd 101.07 1.79 83.13 1.13

2-Nitrotoluene

0 30 300

nd 102.77 1.10 98.53 0.02

nd 85.97 2.51 83.80 3.66

nd 80.35 1.45 80.86 0.55

3-Nitrotoluene

0 30 300

nd 89.81 2.11 99.99 0.92

nd 85.62 3.48 82.45 1.82

nd 91.88 1.77 86.11 2.64

4-Nitrotoluene

0 30 300

nd 94.76 2.81 99.68 0.04

nd 96.71 2.35 85.24 0.71

nd 84.57 1.85 85.93 0.42

a

nd: not detected.

3.2.4. Effect of ionic strength. The inuence of ionic strength on the extraction performance of the IL–Fe3O4@G-MSPE method was investigated by adding NaCl in the range of 0 to 8% (w/v) to the solution. The result showed that no obvious change was observed for the extraction efficiencies with the variation of NaCl concentration (data not shown), indicating that aggregation of the IL on the Fe3O4@G surface and the adsorption of the IL onto the NBs were driven by p–p interactions and hydrophobic interactions rather than electrostatic attraction interactions. Since the effect of salt addition was negligible and no salt was added in the subsequent experiments. 3.2.5. Effect of solution pH. Although the solution pH could change both the existing forms of the target compounds and the charges on the adsorbent surface, the extraction efficiency of the proposed IL–Fe3O4@G-MSPE procedure hardly varied over the whole pH range of 3–11, which indicated that p–p interactions and hydrophobic interactions between the ILcoated Fe3O4@G adsorbent and the NBs are very stable over a wide pH range and that the pH of the water sample does not need to be adjusted. 3.3. Evaluation of the method Under the optimum conditions, the linear range, limits of detection (LODs), relative standard deviations (RSDs) and recovery were evaluated and are shown in Table 1. As can be seen, good linearities were observed for all analytes with the coefficient of correlation (r2) ranging from 0.9990 to 0.9995. The LODs calculated at a signal-to-noise ratio (S/N) of 3 ranged from 1.35 to 4.57 mg L1; the extraction recoveries ranged from 88.16 0.75% to 100.12 1.86%; the intraday and interday RSDs varied from 1.19 to 2.23% and from 1.21 to 4.08%, respectively. The results demonstrate that the method shows potential for the detection of NBs in water samples.


3.4. Recycle of magnetic graphene Attributed to the superparamagnetism, the IL–Fe3O4@G adorbent can be readily dispersed in aqueous solution with ultrasonic assistance and isolated from the matrices easily by using an external magnet when necessary. Once the external magnetic eld is taken away, they can re-disperse rapidly and be used repeatedly. As shown in Fig. 5, the extraction recoveries of the ve NBs have no obvious uctuation even though the Fe3O4@G adsorbents have been recycled for ve runs. 3.5. Comparison of the proposed method with other methods To highlight the robust application of the IL–Fe3O4@G-MSPE method for the determination of NBs from water samples, the proposed method was compared to several published methods, such as LLE, SPE, SDME and LPME, as shown in Table 2. Compared to SPE, SDME and LPME, the proposed method considerably accelerated the sample preparation procedure because only 20 min was required to nish the ultrasound assisted extraction and 15 s was required for magnetic separation of the IL–Fe3O4@G adsorbents from the water samples. Meanwhile, the IL–Fe3O4@G-MSPE has a great advantage over LLE in the expenditure of organic solvent. Moreover, the Fe3O4@G MNPs could be recycled with an external magnet more than 5 times. 3.6. Analysis of real water samples In order to evaluate the matrix effects on the extraction efficiency, the proposed method was for the analysis of ve NBs in real environmental water samples including river, lake and rain water collected in Hangzhou. As shown in Table 3, all of the water samples were free of NB contamination. The typical

Analyst, 2014, 139, 1938–1944 | 1943

View Article Online

Analyst

chromatograms of the NBs in the spiked and the blank lake water are illustrated in Fig. 6. No matrix effect on extraction efficiency was found. The recoveries of the ve NBs in fortied river, lake and rain water samples ranged from 80.35 to 102.77% with RSDs in the range of 0.02 to 4.37%, which indicate that the proposed method gave satisfactory results.


4. Conclusions In this study, UAMSPE with [C7MIM][PF6] IL-coated Fe3O4@G as an adsorbent coupled to HPLC-PDA has been successfully applied for the simultaneous evaluation of ve NBs in environmental water samples. Under the optimized conditions, satisfactory extraction efficiencies and validation parameters including sensitivity, precision and recovery were obtained. By analyzing real water samples, no adverse matrix effect on the proposed preconcentration procedure was demonstrated. Compared to previous methods reported in literature, the advantages such as using a solvent-saving extraction with only 144 mL [C7MIM][PF6] adsorbent and 3 mg Fe3O4@G supported materials used, convenient magnetic separation without the need for any special instruments, and easy-to-recycle Fe3O4@G supported materials that can be reused over 5 times without loss of the extraction efficiency, make this technique a feasible alternative for routine screening of trace-level NBs in environmental samples.

Acknowledgements We gratefully acknowledge the nancial support from socially useful projects under the Zhejiang Province Science and Technology Department (2010C33074) and the project supported by the Zhejiang University of Technology Foundation (20070167).

References 1 D. M. Cha and B. Ma, Am. Lab., 2009, 41, 12–15. 2 H. K. Zhang, S. X. Liang and S. J. Liu, Anal. Bioanal. Chem., 2007, 387, 1511–1516. 3 Y. Zhang, W. Gao, K. Jing, C. B. Xu, N. B. Wang and Z. Y. Jiang, Liaoning Daxue Xuebao, Ziran Kexueban, 2010, 37, 282–284. 4 C. Zhao, J. Y. Wang, L. H. Chen and W. J. Lu, Lihua Jianyan, Huaxue Fence, 2012, 48, 834–835. 5 F. W. Xie, J. J. Shang, J. Z. Guo, Z. Ge and S. S. Zhang, J. Chromatogr. Sci., 2012, 50, 387–392.

1944 | Analyst, 2014, 139, 1938–1944

Paper

6 N. Liu, Shandong Huagong, 2012, 41, 48–51. 7 C. X. Ren, T. L. Xiang and J. H. Yang, Haiyang Huanjing Kexue, 2012, 31, 743–745. 8 X. T. Peng, X. Zhao and Y. Q. Feng, J. Chromatogr., A, 2011, 1218, 9314–9320. 9 C. Yao, P. Twu and J. L. Anderson, Chromatographia, 2010, 72, 393–402. 10 Y. F. Mu, L. L. Yang, E. Y. Hu and Y. Ji, Sepu, 2007, 25, 876– 880. 11 Y. C. Fan and S. L. Zhang, Eur. J. Chem., 2011, 2, 282–288. 12 C. Yao and J. L. Anderson, Anal. Bioanal. Chem., 2009, 395, 1491–1502. 13 X. J. Cao, J. F. Qiao, L. P. Wang, X. M. Ye, L. B. Zheng, N. Jiang and W. M. Mo, Rapid Commun. Mass Spectrom., 2012, 26, 740–748. 14 X. J. Cao, X. M. Ye, Y. B. Lu, Y. Yu and W. M. Mo, Anal. Chim. Acta, 2009, 640, 47–51. 15 X. J. Cao, L. X. Shen, X. M. Ye, F. F. Zhang, J. Y. Chen and W. M. Mo, Green Sustainable Chem., 2013, 3, 26–31. 16 R. Zhang, N. Li, C. L. Wang, Y. P. Bai, R. B. Ren, S. Q. Gao, W. Z. Yu, T. Q. Zhao and H. Q. Zhang, Anal. Chim. Acta, 2011, 704, 98–109. 17 M. H. Duan, M. Luo, C. J. Zhao, W. Wang, Y. G. Zu, D. Y. Zhang, X. H. Yao and Y. J. Fu, Sep. Purif. Technol., 2013, 107, 26–36. 18 M. Cruz-Vera, R. Lucena, S. C´ ardenas and M. Valc´ arcel, J. Chromatogr., A, 2009, 1216, 6459–6465. 19 J. Han, Y. Wang, Y. Liu, Y. F. Li, Y. Lu, Y. S. Yan and L. Ni, Anal. Bioanal. Chem., 2013, 405, 1245–1255. 20 Q. L. Zhang, F. Yang, F. Tang, K. Zeng, K. K. Wu, Q. Y. Cai and S. Z. Yao, Analyst, 2010, 135, 2426–2433. 21 J. H. Zhang, M. Li, M. Y. Yang, B. Peng, Y. B. Li, W. F. Zhou, H. X. Gao and R. H. Lu, J. Chromatogr., A, 2012, 1254, 23–29. 22 Q. H. Wu, G. Y. Zhao, C. Feng, C. Wang and Z. Wang, J. Chromatogr., A, 2011, 1218, 7936–7942. 23 X. D. Ren and Z. H. Xiong, Acta Chim. Sin., 2013, 71, 625– 633. 24 J. L. Gong, B. Wang, G. M. Zeng, C. P. Yang, C. G. Niu, Q. Y. Niu, W. J. Zhou and Y. Liang, J. Hazard. Mater., 2009, 164, 1517–1522. 25 W. N. Wang, Y. P. Li, Q. H. Wu, C. Wang, X. H. Zang and Z. Wang, Anal. Methods, 2012, 4, 766–772. 26 M. Sun, X. X. Ma, J. T. Wang, W. N. Wang, Q. H. Wu, C. Wang and Z. Wang, J. Sep. Sci., 2013, 36, 1478–1485. 27 Y. F. Zhang and H. K. Lee, J. Chromatogr., A, 2013, 1271, 56–61.


View Article Online

Analyst


PAPER


G-quadruplex DNAzymes-induced highly selective and sensitive colorimetric sensing of free heme in rat brain† Ruimin Li,ab Qin Jiang,ac Hanjun Cheng,ac Guoqiang Zhang,ab Mingming Zhen,ab Daiqin Chen,ab Jiechao Ge,*d Lanqun Mao,ac Chunru Wangab and Chunying Shu*ab Direct selective determination of free heme in the cerebral system is of great significance due to the crucial roles of free heme in physiological and pathological processes. In this work, a G-quadruplex DNAzymesinduced highly sensitive and selective colorimetric sensing of free heme in rat brain is established. Initially, the conformation of an 18-base G-rich DNA sequence, PS2.M (50 -GTGGGTAGGGCGGGTTGG-30 ), in the presence of K+, changes from a random coil to a “parallel” G-quadruplex structure, which can bind free heme in the cerebral system with high affinity through p–p stacking. The resulted heme/G-quadruplex complex exhibits high peroxidase-like activity, which can be used to catalyze the oxidation of colorless ABTS2 to green ABTSc by H2O2. The concentration of heme can be evaluated by the naked eye and determined by UV-vis spectroscopy. The signal output showed a linear relationship for heme within the concentration range from 1 to 120 nM with a detection limit of 0.637 nM. The assay demonstrated here was highly selective and free from the interference of physiologically important species such as dopamine (DA), 3,4-dihydroxyphenylacetic acid (DOPAC), ascorbate acid (AA), cysteine, uric acid (UA), glucose and

Received 28th October 2013 Accepted 27th January 2014

lactate in the cerebral system. The basal dialysate level of free heme in the microdialysate from the striatum of adult male Sprague-Dawley rats was determined to be 32.8 19.5 nM (n ¼ 3). The analytic

DOI: 10.1039/c3an02025h

protocol possesses many advantages, including theoretical simplicity, low-cost technical and instrumental

www.rsc.org/analyst

demands, and responsible detection of heme in rat brain microdialysate.

Introduction Cerebrovascular diseases and their consequences, such as strokes, are the second most common cause of death, aer ischemic heart diseases.1 Intracerebral hemorrhage (ICH) is a frequent cause of disability in adults, with a signicant amount of permanent brain damage arising weeks aer a stroke.2 A growing number of experimental evidence suggests that free heme released therein is related with brain damage, even the pathophysiology of ICH is complicated.3–6 Heme (Fe-protoporphyrin IX) is an iron-containing prosthetic group among a diverse group of proteins.7 Free heme, a protein-unbound form of heme, which can be released from hemoglobin following hemolysis, is pro-inammatory a

Beijing National Laboratory for Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, P. R. China. E-mail: [email protected]; Fax: +86-10-62561085; Tel: +86-10-62561085

b c

Key Laboratory of Molecular Nanostructure and Nanotechnology, P. R. China

Key Laboratory of Analytical Chemistry for Living Biosystems, P. R. China

d

Key Laboratory of Photochemical Conversion and Optoelectronic Materials, Technical Institute of Physics and Chemistry, Chinese Academy of Sciences, Beijing 100190, P. R. China † Electronic supplementary 10.1039/c3an02025h

information

(ESI)


available.

See

DOI:

and contributes to iron-derived reactive oxygen species, potentially causing oxidative damage to cells and tissues.8,9 Free heme level can be up-regulated under pathological conditions and leads to various inammatory lesions including vascular disorders, renal failure, and immunemediated disorders.4,10–12 In addition, recent studies have revealed that free heme is also responsible for the progression of cerebral malaria,13 colon cancer14 and Alzheimer's disease.15,16 Highly selective and sensitive detection of physiological free heme involved in the brain function is of great importance for understanding its nature involved in physiological and pathological events and therefore providing a platform for the diagnosis and therapy of related diseases.17,18 Up to date, a variety of techniques have been developed to satisfy this purpose, including mass spectrometry,19 absorption spectrophotometry,20,21 chemiluminescence,22 and uorescent spectrometry.23,24 However, these methods are either time-consuming, lacking specicity and selectivity, or depending on expensive instruments. In recent years, colorimetric biosensing has attracted much attention due to its low cost, simplicity, practicality and its potential application in eld analysis and point-of-care diagnosis.25–27 Therefore, colorimetric biosensing is very competitive.

Analyst, 2014, 139, 1993–1999 | 1993

View Article Online


Analyst

Paper

G-quadruplex DNAzymes are usually formed by G-rich nucleic acid sequences in the presence of alkali metal ions, which can easily combine with hemin,28–31 especially when it exists as an oxidized state in vitro.19 Considering the purpose of our study, heme is herein used as the general term for oxidized form. Most importantly, the resulting heme-G-quadruplex complexes possess peroxidase-like activity and exhibit catalytic activity towards the H2O2-mediated oxidation of 2,20 -azino-bis (3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS).32,33 In view of this unique property, the heme-G-quadruplex DNAzymes have shown great potential in colorimetric detection of various targets, including metal ions,34,35 small molecules,36,37 DNAs38,39 and proteins.40,41 Herein, we report a novel colorimetric assay for highly selective and sensitive sensing of cerebral free heme. The fundamentals of this sensing are essentially on the strength of in vivo microdialysis and peroxidase-like activity of G-quadruplex DNAzymes induced by free heme in the cerebral system, which can catalyze the oxidation of colorless ABTS2 to green ABTSc by H2O2 (Scheme 1). To the best of our knowledge, this is the rst example for direct selective sensing of cerebral free heme based on G-quadruplex DNAzymes, which may provide a facile strategy for monitoring brain chemistry in a primitive style.

Materials and methods

from Beijing Chemical Works. Glucose was purchased from Sinopharm Chemical Reagent Co., Ltd. Hemin (recognized as the oxidized state of heme in vitro), cysteine, dopamine (DA), ascorbate acid (AA), uric acid (UA) and 3,4-dihydroxyphenylacetic acid (DOPAC) were purchased from Sigma-Aldrich. L(+)-Lactate acid was obtained from Acros Organics. All reagents were used as received without further purication. The stock solution of oligonucleotide (100 mM) was prepared in 10 mM Tris-Ac buffer (pH 7.0) and accurately quantied using UV-vis absorption spectrum with the following extinction coefficients (260 nm, M1 cm1): (A) 15400, (G) 11500, (C) 7400, (T) 8700.36 The oligonucleotide solutions were diluted to required concentrations with the buffer prior to use. The stock solution of 1 mM heme was prepared in dimethyl sulfoxide (DMSO), and stored in darkness at 20 C. The concentration of H2O2 was accurately quantied using UV-vis absorption spectrum at 240 nm by the characteristic molar extinction coefficient (3 ¼ 43.6 dm3 M1 cm1). Articial cerebrospinal uid (aCSF) was prepared by mixing NaCl (126 mM), KCl (2.4 mM), KH2PO4 (0.5 mM), MgCl2 (0.85 mM), NaHCO3 (27.5 mM), Na2SO4 (0.5 mM), and CaCl2 (1.1 mM) into Milli-Q water. Aqueous solutions of DA, AA, UA, DOPAC, lactate, cysteine, and glucose were freshly prepared with aCSF prior to experiment. All aqueous solutions were prepared with Milli-Q water (18.2 MU cm1). Preparation of the G-quadruplex DNA solution modulated by K+

Materials 0

The puried G-rich oligonucleotide 5 -GTGGGTAGGGCG GGTTGG-30 (PS2.M), which could form a G-quadruplex structure in the presence of K+, is designed for the following study. All oligonucleotides were obtained from Sangon Biotech. Co. Ltd. (Beijing, China). 2,20 -Azinobis (3-ethylbenzothiozoline)-6sulfonic acid (ABTS) was purchased from Tokyo Chemical Industry Co. Ltd. Hydrogen peroxide (H2O2, 30%) was obtained

The PS2.M solution was heated at 88 C for 10 min to dissociate any intermolecular interaction, and gradually cooled to room temperature. Then KAc with an appropriate concentration was added into the DNA solution, allowing the DNA sequences to fold properly for 40 min, so they can form the quadruplex structures in the presence of K+. Colorimetric sensing of heme in aCSF All samples were prepared by dissolving the 1 mM heme stock solution directly into the aCSF. The aCSF was used to prepare heme solutions with different concentrations prior to each measurement. For the colorimetric detection of heme in aCSF, a reaction mixture was rst prepared by mixing 50.0 mL of different concentrations of heme in an aCSF, which was added to 430 mL of G-quadruplex DNA solution containing K+ (10 mM) in a vial. Aer 1.0 h, 10 mL of H2O2 (5 mM) and 10 mL of ABTS (50 mM) were added to the above mixture. The resulting mixtures were then incubated for 2.0 h (see ESI†) at 30 C and subsequently for UV-vis spectrometric measurements. The nal concentrations of heme in 500 mL of the resulting mixtures were 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 16, 20, 30, 40, 50, 60, 70, 80, 90, 100, 120, 150, 180 and 200 nM. In vivo microdialysis

G-quadruplex DNAzymes-based colorimetric sensing of cerebral free heme coupled with in vivo microdialysis.

Scheme 1

1994 | Analyst, 2014, 139, 1993–1999

Animal surgery and in vivo microdialysis were carried out according to the procedures reported.18 Briey, adult male Sprague-Dawley mice (250–300 g) obtained from Health Science Center, Peking University, were fed casually with food and water in a 12 : 12 h light–dark schedule. Based on standard stereotaxic procedures, the microdialysis guide cannula (BAS/MD-2250,


View Article Online

Paper


BAS) was implanted into the rat's striatum. The body temperature of the animal was kept at 37 C with a heating pad during the whole procedure of the surgery. A microdialysis probe was implanted into the rat's striatum and then perfused with aCSF at 1.0 mL min1 aer the rat recovered for at least 24 h. Subsequently perfusing ceaselessly for at least 90 min for the sake of equilibration, the microdialysate was collected for the colorimetric sensing. Colorimetric sensing of cerebral free heme For colorimetric detection of free heme in the brain microdialysate, 50 mL of the microdialysate sample from striatum was added into 430 mL of the G-quadruplex DNA solution containing K+ (10 mM), and then incubated for about 1.0 h. Aerwards, 10 mL of H2O2 (5 mM) and 10 mL of ABTS (50 mM) were added to the above mixture. For a quantitative assay of free heme in the striatum microdialysate, the resulting mixture was incubated for 2.0 h at 30 C for UV-vis spectrometric measurements. CD measurements Circular dichroism (CD) spectroscopy of PS2.M (1.0 mM) was measured on a J-815 spectropolarimeter under different conditions at room temperature. PS2.M solutions were prepared in 10 mM of Tris-Ac buffer (pH 7.0). Three scans (100 nm min1) from 225 to 325 nm at 0.2 nm intervals were accumulated and averaged. The background of the buffer solution was subtracted from the CD data. Measurement of UV-vis spectra All UV-vis absorption spectra were measured on a UV-2550 UVvis spectrophotometer (Shimadzu, Japan) equipped with a quartz cell (1 0.33 cm cross-section) at room temperature. The maximum absorption wavelength for the radical anion ABTSc (the oxidation product of ABTS2) is about 420 nm.

Results and discussion Principle of colorimetric sensing heme by G-quadruplex DNAzyme PS2.M (50 -GTGGGTAGGGCGGGTTGG-30 ), an 18-base G-rich DNA sequence, can form DNAzyme with high peroxidase-like activity.42 Since K+ has individual effects on the peroxidase like activity of G-quadruplex DNAzymes,43 K+-stabilized PS2.M was utilized to sense free heme in the cerebral system (see Scheme 1). In the presence of K+, the conformation of PS2.M changes from a random coil to a “parallel” G-quadruplex structure, which can bind free heme in the cerebral system with high affinity.29 Importantly, the resulting heme/G-quadruplex complex exhibits much higher peroxidase-like activity as compared to pristine heme, which can catalyze the oxidation of colorless ABTS2 to green ABTSc in the presence of H2O2.30,42 The presence of free heme in the cerebral system is therefore detected indirectly by the absorption enhancement of green ABTSc. The analytic protocol of heme based on the G-quadruplex structure was assessed by UV-vis absorption spectrum as described in the Experimental section.


Analyst

Fig. 1 shows the color and the absorption spectra of the H2O2–ABTS reaction at different cases. In the presence of PS2.M itself (curve (a)), the mixture of H2O2 and ABTS has no apparent absorption in the range of the wavelength 390–470 nm as well as that in the coexistence of PS2.M and K+ (curve (b)), which suggest that neither the random coil of PS2.M nor the “parallel” G-quadruplex structure of PS2.M induced by K+ have any peroxidase-like catalytic activity. In the presence of PS2.M and heme, the mixture of H2O2 and ABTS has weak absorption at about 420 nm (curve (c)), due to the weak peroxidase-like catalytic property of heme.36 However, the mixture of H2O2 and ABTS shows strong absorption at about 420 nm (curve (d)) when PS2.M, K+ and heme were added, suggesting that the heme/ G-quadruplex complex possesses enhanced peroxidase-like catalytic activity. These results have good accordance with that of the corresponding color changes. To explore the principle of enhanced peroxidase-like catalytic activity of heme/G-quadruplex complex toward H2O2– ABTS reaction, CD spectra of PS2.M under different conditions were measured. As shown in Fig. 2, the CD spectrum of PS2.M upon addition of K+ demonstrates a typical characteristic of parallel G-quadruplex structure, a positive peak near 265 nm and a negative peak near 240 nm.36 This phenomenon illustrates the structural transition of the 18-base G-rich DNA sequence PS2.M in the presence of K+ in comparison with that in the absence of K+. When heme, as a contrast, was added into the solution of PS2.M, the CD spectrum of the mixture doesn't show apparent change compared with that of PS2.M alone, which indicates that the heme itself can't promote the formation of G-quadruplex structure. Notably, CD spectrum of PS2.M in the coexistence of K+ and heme gives rise to a signicant change. The intensity near 290 nm has greatly reduced than that in the presence of K+ alone, which suggests the addition of heme is able to externally stack on the terminal G-tetrad owning to strong p–p interaction.44,45 These results indicate that heme and G-quadruplex structure formed by G-rich DNA sequence PS2.M in the presence of K+ are very important to the enhanced color reaction of H2O2–ABTS.

Fig. 1 Color and absorption spectra of the H2O2–ABTS reaction in different cases. (a) 0.1 mM H2O2 + 1 mM ABTS2 + 0.25 mM PS2.M; (b) 0.1 mM H2O2 + 1 mM ABTS2 + 0.25 mM PS2.M + 10 mM K+; (c) 0.1 mM H2O2 + 1 mM ABTS2 + 0.25 mM PS2.M + 100 nM heme; (d) 0.1 mM H2O2 + 1 mM ABTS2 + 0.25 mM PS2.M + 10 mM K+ + 100 nM heme.

Analyst, 2014, 139, 1993–1999 | 1995

View Article Online


Analyst

Paper

Fig. 2 CD spectra of PS2.M under different conditions. (1 mM PS2.M, 0.4 mM heme, 40 mM K+).

Optimization of experimental conditions The colorimetric biosensor for heme was established based upon G-quadruplex DNAzymes catalyzed the oxidation–reduction reaction of ABTS2 and H2O2. Therefore, any factors involved in this reaction may affect the sensitivity of the sensor, such as concentrations of H2O2, ABTS2, PS2.M and K+, temperature and pH of Tris-Ac buffer. H2O2 and ABTS are key components in this color reaction. Herein, the optimization of their concentrations is important for the assay of heme. The effect of the concentrations of H2O2 and ABTS were studied over the range of 0–400 mM and 0–2 mM, respectively. As shown in Fig. 3a, the absorption of the reaction mixture (100 nM heme, 1 mM ABTS2, 0.25 mM PS2.M, 10 mM K+, 30 C, pH 7.0) at 420 nm increased with the increase of H2O2 during the range of 0–80 mM, and kept a plateau when H2O2 varied from 80 to 120 mM, and then decreased with the increase of H2O2 concentration. The absorption of reaction mixture (100 nM heme, 100 mM H2O2, 0.25 mM PS2.M, 10 mM K+, 30 C, pH 7.0) at 420 nm increased with the increase of ABTS during the range of 0–0.8 mM, and then kept a plateau at the range of 0.8–2.0 mM (Fig. 3b). According to the above results, the optimal concentration of H2O2 and ABTS for the detection of heme is 100 mM and 1 mM, respectively. PS2.M and K+ are necessary to form G-quadruplex-based DNAzyme. The peroxidase-like catalytic activity of this DNAzyme is closely related with the concentrations of PS2.M and K+. The concentrations of PS2.M and K+ were investigated over the range of 0–0.4 mM and 0–25 mM, respectively, as demonstrated in Fig. 3c and d. The absorption of reaction mixture (100 nM heme, 100 mM H2O2, 1 mM ABTS2, 10 mM K+, 30 C, pH 7.0) at 420 nm increased rapidly with the increase of PS2.M at the range of 0–0.2 mM, reached a plateau at concentration of 0.2 mM, and kept constant even when the concentration of PS2.M increased to 0.4 mM. With the increase of K+, the absorption of reaction mixture (100 nM heme, 100 mM H2O2, 1 mM ABTS2, 0.25 mM PS2.M, 30 C, pH 7.0) at 420 nm increased sharply in the range of 0–7.5 mM and then kept a plateau at the concentration of K+ even as 25 mM. Based on above mentioned, 0.25 mM PS2.M and 10 mM K+ were used as optimal conditions for the detection of heme.

1996 | Analyst, 2014, 139, 1993–1999

Effect of different conditions on the absorption of the solution, including concentrations of H2O2 (a), ABTS2 (b), PS2.M (c) and K+ (d), temperature (e) and pH value of Tris-Ac buffer (f). The error bars represent the standard deviation of the three measurements. The optimal point in each curve was set as 100%. (a) 100 nM heme + 1 mM ABTS2 + 0.25 mM PS2.M + 10 mM K+, 30 C, pH 7.0; (b) 100 nM heme + 100 mM H2O2 + 0.25 mM PS2.M + 10 mM K+, 30 C, pH 7.0; (c) 100 nM heme + 100 mM H2O2 + 1 mM ABTS2 + 10 mM K+, 30 C, pH 7.0; (d) 100 nM heme + 100 mM H2O2 + 1 mM ABTS2 + 0.25 mM PS2.M, 30 C, pH 7.0; (e) 100 nM heme + 100 mM H2O2 + 1 mM ABTS2 + 0.25 mM PS2.M + 10 mM K+, pH 7.0; (f) 100 nM heme + 100 mM H2O2 + 1 mM ABTS2 + 0.25 mM PS2.M + 10 mM K+, 30 C. Fig. 3

The temperature and pH value of Tris-Ac buffer are crucial to the catalytic activity of G-quadruplex-based DNAzymes. In view of the sensitivity of the method, temperature and pH value of Tris-Ac buffer were optimized as well. Herein, the inuence of temperature in the range of 4–60 C and pH value of Tris-Ac buffer in the range of 4.5–10 on the reaction system were demonstrated in Fig. 3e and f, respectively. The absorption of the reaction mixture (100 nM heme, 100 mM H2O2, 1 mM ABTS2, 0.25 mM PS2.M, 10 mM K+, pH 7.0) at 420 nm increased with the temperature over the range of 4–25 C, kept a plateau when the temperature varied between 25 and 35 C, and declined rapidly at temperature above 40 C. The absorption of the reaction mixture (100 nM heme, 100 mM H2O2, 1 mM ABTS2, 0.25 mM PS2.M, 10 mM K+, 30 C) at 420 nm increased with the pH value of Tris-Ac buffer over the range of 4.5–7.0, and decreased with the increase of pH above 7.0. Therefore, 30 C and pH 7.0 of Tris-Ac buffer were selected as the optimal reaction environment. Selectivity and sensitivity Using the G-quadruplex formation of PS2.M in the presence of K+ as the probe, we studied the selectivity of the colorimetric method for heme sensing. Among the species that may


View Article Online


Paper

potentially interfere with the selectivity of heme sensing, physiologically important species such as DA, DOPAC, AA, cysteine, UA, glucose and lactate in the cerebral system were taken into account (Fig. 4). The existence of each of these species did not trigger obvious changes of the H2O2–ABTS reaction with G-quadruplex DNAzymes except that heme demonstrated signicant absorbance at 420 nm under the same conditions. These results substantially suggest that only the resulting heme/G-quadruplex complex can catalyze H2O2–ABTS reaction, which produces enhanced color reaction (green ABTSc). That is to say, the present detection method based on the rational design of G-quadruplex formation of PS2.M in the presence of K+ is very specic for cerebral heme sensing. To evaluate the sensitivity of the method, a series of heme in aCSF with different concentrations were added into the Tris-Ac buffer aqueous containing G-quadruplex DNAzymes to catalyze the H2O2–ABTS reaction. As shown in Fig. 5, the absorbance at 420 nm increased with the concentration of heme in aCSF and showed a linear response toward heme within a concentration range from 1 to 120 nM in Tris-Ac buffer solutions (pH 7.0, 30 C) containing 100 mM H2O2, 1 mM ABTS2, 0.25 mM PS2.M and 10 mM K+. The linear regression equation was A ¼ 0.0111C + 0.0472 (C: nM, R2 ¼ 0.999), with a detection limit of 0.637 nM (3s/slope). The enhanced color changes eventually form a straightforward basis for colorimetric sensing of heme.

Colorimetric sensing of heme in rat brain To explore the practicability of the as-established G-quadruplexbased assay for selectively sensing heme in the rat brain, the brain microdialysate (50 mL) was added into 430 mL of an Tris-Ac buffer aqueous containing formed G-quadruplex structure, which were incubated for 1 h. Aerwards, 10 mL of H2O2 (5 mM) and 10 mL of ABTS (50 mM) were added to the above mixture. For the quantitative assay of free heme in the striatum microdialysate, the resulting mixture was incubated for 2 h at 30 C for UV-vis spectrometric measurements. In view of heme

Analyst

The heme concentration dependent UV-vis absorption of ABTS–H2O2 system in the presence of G-quadruplex DNAzymes for colorimetric heme analysis at 420 nm. The inset shows a linear range of heme concentration from 1 nM to 120 nM. The error bars represent the standard deviation of the three measurements. Fig. 5

externally stacking on the terminal G-quadruplex structure through strong p–p interaction, the obtained heme/G-quadruplex DNAzyme exhibits high peroxidase-like activity, which can catalyze the oxidation of colorless ABTS2 to green ABTSc by H2O2. The presence of free heme in the cerebral system is therefore detected indirectly by the absorption of green ABTSc. It is important to note that the heme in the cerebral system is called free heme since no concentrated proteins exist in the cerebral microdialysis as a result of the use of microdialysis for in vivo sampling. According to the calibration curve described above A ¼ 0.0111C + 0.0472 (C: nM, R2 ¼ 0.999) (Fig. 5), the initial value of the basal level of heme in rat brain microdialysates was determined to be 32.8 19.5 nM (n ¼ 3), which was almost consistent with the reported values.46 These results substantially demonstrated that the colorimetric assay developed in this study would offer an effective way to directly selectively sense heme in the cerebral system. In order to demonstrate the universal applicability of this method, we also detect heme in rat serum samples (see ESI†). A comparison of various methods for heme estimation shows that the heme assay based on the high peroxidase-like activity of G-quadruplex/heme complexes is more sensitive than other methods and the low detection limit and high selectivity have made our approach a potential amplier for several biomarker events based on heme. In addition, this assay does not need expensive instruments, sophisticated material synthesis and toxic reagents. The labeled free G-rich nucleic acid sequences for aptamer-based heme analysis made the process simple and easy (Table 1).

Conclusions Fig. 4 Selectivity of heme detection. All measurements were performed in 0.5 mL of 10 mM Tris-Ac buffer solutions (pH 7.0, 30 C) containing 100 mM H2O2, 1 mM ABTS2, 0.25 mM PS2.M and 10 mM K+. The error bars represent the standard deviation of three measurements. The concentration was 10 nM for heme, 50 nM for DA, DOPAC, AA, cysteine, UA, glucose and lactate.


In the present work, a highly selective and sensitive colorimetric method for direct sensing free heme in the cerebral system was established based on the rational design of G-quadruplex DNAzymes. In the presence of K+, the conformation of PS.2M changes from a random coil to a “parallel” G-quadruplex Analyst, 2014, 139, 1993–1999 | 1997

View Article Online

Analyst Table 1

Paper Comparison of different methods for the determination of heme

Technique

Linear range (nM)

Detection limit (nM)

Reference

Mass spectrometry Fluorescent spectrometry

400–8000 10–500 310–2500 4–130 1–120

400 5 50 — 0.637

19 23 24 21 This work


Absorption spectrophotometry

structure, which can bind free heme in the cerebral system with high affinity. Importantly, the resulting heme/G-quadruplex complex exhibits high peroxidase-like activity, which can catalyze the oxidation of colorless ABTS2 to green ABTSc by H2O2. As a result, the content of free heme in the cerebral system is detected indirectly and easily by the absorption enhancement of green ABTSc. As far as we know, this is the rst example of direct sensing of heme in rat brain. In a word, the simple quantication of heme in the cerebral system herein not only offers a simple model to monitor brain chemistry, but also facilitates the understanding of the chemical essence involved in some physiological and pathological events. This study unambiguously provides a new analytical platform to comprehend brain chemistry by G-quadruplex structure.

Acknowledgements This work was supported by 973 Program (no. 2011CB302100), National Natural Science Foundation of China (no. 21121063, 51172244), and NSAF (no. 11076027).

References 1 F. A. Lara, S. A. Kahn, A. C. Da Fonseca, C. P. Bahia, J. P. Pinho, A. V. Graca-Souza, J. C. Houzel, P. L. de Oliveira, V. Moura-Neto and M. F. Oliveira, J. Cereb. Blood Flow Metab., 2009, 29, 1109. 2 S. Robinson, T. Dang, R. Dringen and G. Bishop, Redox Rep., 2009, 14, 228. 3 L. Chen, X. Zhang, J. Chen-Roetling and R. F. Regan, J. Neurosurg., 2011, 114, 1159. 4 R. Larsen, R. Gozzelino, V. Jeney, L. Tokaji, F. A. Bozza, A. M. Japiass´ u, D. Bonaparte, M. M. Cavalcante, A. Chora and A. Ferreira, Sci. Transl. Med., 2010, 2, 51. 5 T. N. Dang, G. M. Bishop, R. Dringen and S. R. Robinson, Glia, 2011, 59, 1540. 6 K. R. Wagner, F. R. Sharp, T. D. Ardizzone, A. Lu and J. F. Clark, J. Cereb. Blood Flow Metab., 2003, 23, 629. 7 S. Sassa, J. Clin. Biochem. Nutr., 2006, 38, 138. 8 R. Li, S. Saleem, G. Zhen, W. Cao, H. Zhuang, J. Lee, A. Smith, F. Altruda, E. Tolosano and S. Dor´ e, J. Cereb. Blood Flow Metab., 2009, 29, 953. 9 R. F. Regan, J. Chen and L. Benvenisti-Zarom, BMC Neurosci., 2004, 5, 1. 10 P. W. Buehler and F. D'Agnillo, Antioxid. Redox Signaling, 2010, 12, 275.

1998 | Analyst, 2014, 139, 1993–1999

11 E. Tolosano, E. Hirsch, E. Patrucco, C. Camaschella, R. Navone, L. Silengo and F. Altruda, Blood, 1999, 94, 3906. 12 S. Kumar and U. Bandyopadhyay, Toxicol. Lett., 2005, 157, 175. 13 A. Ferreira, J. Balla, V. Jeney, G. Balla and M. P. Soares, J. Mol. Med., 2008, 86, 1097. 14 S. I. Ishikawa, S. Tamaki, M. Ohata, K. Arihara and M. Itoh, Mol. Nutr. Food Res., 2010, 54, 1182. 15 H. Atamna and K. Boyle, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 3381. 16 H. Atamna and W. H. Frey, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 11153. 17 Y. Jiang, H. Zhao, Y. Lin, N. Zhu, Y. Ma and L. Mao, Angew. Chem., 2010, 122, 4910. 18 Q. Qian, J. Deng, D. Wang, L. Yang, P. Yu and L. Mao, Anal. Chem., 2012, 84, 9579. 19 J. R. Whiteaker, C. C. Fenselau, D. Fetterolf, D. Steele and D. Wilson, Anal. Chem., 2004, 76, 2836. 20 S. C. Liu, S. Zhai and J. Palek, Blood, 1988, 71, 1755. 21 N. T. Huy, T. Dai Thi Xuan, D. T. Uyen, M. Sasai, S. Harada and K. Kamei, Anal. Biochem., 2005, 344, 289. 22 T. Masuda and S. Takahashi, Anal. Biochem., 2006, 355, 307. 23 S. Pang, S. Liu and X. Su, Talanta, 2014, 118, 118. 24 Y. Shi, W. T. Huang, H. Q. Luo and N. B. Li, Chem. Commun., 2011, 47, 4676. 25 Y. Song, W. Wei and X. Qu, Adv. Mater., 2011, 23, 4215. 26 R. Li, M. Zhen, M. Guan, D. Chen, G. Zhang, J. Ge, P. Gong, C. Wang and C. Shu, Biosens. Bioelectron., 2013, 47, 502. 27 J. Sun, J. Ge, W. Liu, X. Wang, Z. Fan, W. Zhao, H. Zhang, P. Wang and S. T. Lee, Nano Res., 2012, 5, 486. 28 P. R. Majhi and R. H. Shafer, Biopolymers, 2006, 82, 558. 29 D. M. Kong, L. L. Cai, J. H. Guo, J. Wu and H. X. Shen, Biopolymers, 2009, 91, 331. 30 J. Kosman and B. Juskowiak, Anal. Chim. Acta, 2011, 707, 7. 31 X. Wang, Z. Ding, Q. Ren and W. Qin, Anal. Chem., 2013, 85, 1945. 32 T. Li, E. Wang and S. Dong, J. Am. Chem. Soc., 2009, 131, 15082. 33 L. Zhu, C. Li, Z. Zhu, D. Liu, Y. Zou, C. Wang, H. Fu and C. J. Yang, Anal. Chem., 2012, 84, 8383. 34 D. Zhang, M. Deng, L. Xu, Y. Zhou, J. Yuwen and X. Zhou, Chem. – Eur. J., 2009, 15, 8117. 35 H. Sun, X. Li, Y. Li, L. Fan and H. B. Kraatz, Analyst, 2013, 138, 856. 36 R. Li, C. Xiong, Z. Xiao and L. Ling, Anal. Chim. Acta, 2012, 724, 80.


View Article Online

Paper

42 43 44 45

P. Travascio, Y. Li and D. Sen, Chem. Biol., 1998, 5, 505. T. Li, E. Wang and S. Dong, Anal. Chem., 2010, 82, 7576. T. Li, S. Dong and E. Wang, Chem. – Asian J., 2009, 4, 918. S. Paramasivan, I. Rujan and P. H. Bolton, Methods, 2007, 43, 324. 46 S. Koga, S. Yoshihara, H. Bando, K. Yamasaki, Y. Higashimoto, M. Noguchi, S. Sueda, H. Komatsu and H. Sakamoto, Anal. Biochem., 2013, 433, 2.


37 M. Wang, Y. Han, Z. Nie, C. Lei, Y. Huang, M. Guo and S. Yao, Biosens. Bioelectron., 2010, 26, 523. 38 F. Du and Z. Tang, ChemBioChem, 2011, 12, 43. 39 S. Shimron, F. Wang, R. Orbach and I. Willner, Anal. Chem., 2011, 84, 1042. 40 Y. Zhang, B. Li and Y. Jin, Analyst, 2011, 136, 3268. 41 J. Zhang, Y. Chai, R. Yuan, Y. Yuan, L. Bai, S. Xie and L. Jiang, Analyst, 2013, 138, 4558.

Analyst


Analyst, 2014, 139, 1993–1999 | 1999

View Article Online

Analyst


CRITICAL REVIEW


Electroanalytical and surface plasmon resonance sensors for detection of breast cancer and Alzheimer's disease biomarkers in cells and body fluids Minghui Yang,a Xinyao Yi,a Jianxiu Wang*a and Feimeng Zhou*ab Cancer and neurological disorders are two leading causes of human death. Their early diagnoses will either greatly improve the survival rate or facilitate effective treatments or modalities. Detection of biomarkers in body fluids and some tissues (e.g., blood, urine and cerebrospinal fluids) is relatively non-invasive and provides useful chemical and biological information that is complementary to tomographic imaging (e.g., magnetic resonance imaging, positron emission tomography and X-ray computed tomography). Recent years have witnessed the contributions from and potential applications of bioanalytical methods for early detection of major diseases. In this review, we survey some recent developments of electroanalytical (as a representative label-based technique) and surface plasmon resonance (SPR) (as a representative label-free technique) biosensors for detection of biomarkers relevant to etiologies of breast cancer and Alzheimer's disease (AD). While breast cancer is representative of cancers of complexity (multiple biomarkers, false positives from tomographic scans, and a need for more effective early diagnostic methods), AD is the most prevalent neurological disorder that is also linked to multiple biomarkers. Both electroanalytical and SPR-based sensors have attractive features of sensitivity, portability, obviation of

Received 2nd November 2013 Accepted 21st January 2014

large sample volumes, and capability of multiplexed detection. Various sensing protocols developed in the past five years are reviewed, demonstrating the feasibility of both techniques for diagnostic purposes.

DOI: 10.1039/c3an02065g

Problems inherent in these two techniques that must be overcome before being clinically viable are also

www.rsc.org/analyst

discussed.

1. Introduction a

Key Laboratory of Resources Chemistry of Nonferrous Metals, Ministry of Education, College of Chemistry and Chemical Engineering, Central South University, Changsha, Hunan 410083, P.R. China. E-mail: [email protected]

b

Department of Chemistry and Biochemistry, California State University, Los Angeles, California 90032, USA. E-mail: [email protected]; Tel: +1 3233432390

Minghui Yang obtained his PhD degree in Analytical Chemistry from Hunan University in 2007. Aer post-doc training at the University of Maryland, Baltimore County and the University of Wisconsin, Milwaukee, he is currently a professor at Central South University. His research interests include biomaterials, biosensors and microuidics.

1814 | Analyst, 2014, 139, 1814–1825

Sensitive and selective detection of disease biomarkers is of great importance for early diagnosis, personalized treatment, and molecularly targeted therapy of major diseases.1–3 Three of the leading debilitating types of diseases are cardiovascular

Xinyao Yi obtained her B.S. degree in Applied Chemistry in 2009 from Central South University and is currently a PhD student at Central South University under the joint supervision of Drs Jianxiu Wang and Feimeng Zhou. Her current research mainly focuses on the development of new methods for the sensitive detection of biomarkers of neurodegenerative diseases by surface plasmon resonance.


View Article Online


Critical Review

Analyst

diseases, cancers, and neurological disorders, the latter two of which are difficult to diagnose in the early stages. In normal persons or patients whose cancers are in the early stage or those who suffer from mild cognition impairment, the expression level of many biomarkers in human uids (e.g., urine, blood, saliva, and cerebrospinal uids) and tissues (e.g., tumor cells) is usually at trace levels (e.g., pg mL1 range). Consequently, early diagnoses of these diseases require techniques and methods with high sensitivity and selectivity. Several reviews have already summarized the detection of various cancer biomarkers.4–8 Some reviews have also been published on the possible use of various biomarkers for assessment of the severity of neurological disorders.9,10 However, these reviews are rather broad and tend not to focus on biomarkers related to a specic type of cancer or neurological disorder. In this review, we chose to survey bioanalytical methods that could serve as alternative means for early diagnosis of breast cancer and neurological disorders. We chose these two major types of diseases based on the facts that both are closely related to aging as well as genetics, and environmental factors are also implicated in their etiologies.11,12 Bioanalytical techniques that have been applied to assays of the above-mentioned biomarkers can be classied into labelbased and label-free ones. Table 1 lists select techniques in each of the two categories. We should mention that, for biomarkers that contain inherent signaling moieties (e.g., uorophores or redox moieties), some of the techniques (e.g., uorescence or electrochemistry) can also be used in the label-free fashion. However, generally speaking, most biomarkers need to be tagged in order for these techniques to operate. Given the scope of a short critical review, we therefore further limit our surveys on

Jianxiu Wang obtained his PhD degree in Chemistry from Beijing University in 2002. He was trained as a post-doc at Graduate School, Chinese Academy of Sciences. He is currently a Professor of Chemistry at Central South University. His research interests include functionalized nanomaterials, disease biomarkers and biosensors.


the recent developments of electroanalytical and surface plasmon resonance (SPR) biosensors within the past ve years. Although electroanalytical and SPR-based sensors rely on a binding event between the biomarker and a surface-tethered capture molecule, the detection (or signal transduction) of these events differs considerably (cf. Table 1). Electrochemistry is a technique that studies the charge transfer process between the redox moiety (or tag) on the biomarker and the electrode, while SPR measures the refractive index change (adsorption of a biomarker) on the sensor chip. The former33,34 is a good representative of the label-based techniques, while the latter13,14 has features that are common among label-free techniques. Both electroanalytical15 and SPR biosensors13 possess promising characteristics for early clinical diagnosis in their sensitivity and speed. Their low cost and portability are also attractive for point-of-care assays.16–18 Another consideration in choosing to review these two types of sensors is that their detections occur at the solid–solution interface (i.e., heterogeneous biosensors). Specically, in both types of sensors, target biomarkers are captured by molecules (also referred to as probe molecules in the bioanalytical community or bait molecules in the medical eld) as varied as antibodies,19,20 aptamers,21 peptides22 and DNA.23 SPR and electrochemistry are also complementary. For example, SPR provides information about binding kinetics and affinity between a capture probe and a specic disease biomarker. Such information is important for the development of reliable and effective sensors including electrochemical biosensors.24,25 In electrochemical detection, a redox label or moiety has to be in close proximity to the electrode to generate detectable signals. On the other hand, a cleverly designed electrochemical sensor can yield higher sensitivity than many

Feimeng Zhou obtained his PhD degree in Chemistry from the University of Texas at Austin. He did his postdoctoral training at the Oak Ridge National Lab before becoming a professor at California State University, Los Angeles. He is currently a University Outstanding Professor of Chemistry and serves as the director of the Center for Research Excellence in Science and Technology (CREST) funded by the National Science Foundation. His awards include a Dreyfus Teacher-Scholar award and the 2012 Faculty Research Award from the 23-campus California State University system. He is also an adjunct professor in the College of Chemistry and Chemical Engineering at Central South University. His research interests include high-throughput biosensors and labelfree microarrays for disease diagnosis, studies of amyloid proteins, developments of coupled analytical techniques for biological and environmental applications, and nanomaterial synthesis, characterization, and applications.

Analyst, 2014, 139, 1814–1825 | 1815

View Article Online

Analyst Table 1

Critical Review Select label-based and label-free methods for cancer sample analyses


Label-based

Label-free

Method

Detection principle

Sensitivity

Limitation

Fluorescence98

Fluorescent light emission

High

Electrochemistry

Charge transfer

High

Electrogenerated chemiluminescence99 Chemiluminescence100

Charge-transfer-stimulated light emission Light emitted from chemical reactions Refractive index change/ mass loading Mass change

High

Static and dynamic quenching Nonspecic adsorption, sluggish charge transfer Complicated procedures

Moderate

Limited dynamic range

High Moderate–high

Impedance spectroscopy102

Conductivity

Moderate

Cantilever sensors103

Atomic force

High

Bulk refractive index change, nonspecic adsorption Viscoelastic/temperature effects, non-uniform sensing area Complicated modeling and data interpretation Limited dynamic range and semi-quantitative

Surface plasmon resonance Surface acoustic wave/quartz crystal microbalance101

SPR-based sensors. We should also mention that electrochemistry and SPR can be combined into a hybrid technique to study protein conformational and structural changes accompanying with charge transfer26,27 or to enhance the binding events with an applied potential.28 To help readers to assess the potential of electroanalytical and SPR sensors for breast cancer and AD biomarker detection, we have summarized some key performance criteria (e.g., dynamic ranges and detection limits) at the end of Sections 2 and 3. Improvements to be made and challenges to be overcome for implementing these sensors as viable diagnostic tools are critically reviewed in Section 4, Conclusions and outlook.

2.

Breast cancer biomarkers

Cancer is the second leading cause of mortality worldwide following cardiovascular disease. More than 100 types of cancers have been classied according to the type of cells initially affected. Breast cancer is the most common invasive cancer in females, accounting for 16% of all female cancers and 22.9% of invasive cancers in women in 2008.29 About 18.2% of all cancer deaths in both males and females worldwide are from breast cancers.30 It is a disease with several distinct subtypes.31 Among the various types of cancers, breast cancer is representative because, while its clinical diagnosis is easier than those of many other cancers, there still remain many challenges for early detection to guide effective therapies. The following issues are not exhaustive but certainly problematic: (1) there exist multiple proteins as possible biomarkers and relying on the analysis of a single biomarker for diagnosis is insufficient and unreliable,32 (2) current medical imaging methods are still prone to producing false-positive results and could be relatively insensitive to tiny tumors33 and (3) some of the existing diagnostic tools are still invasive.34 The common denition of breast cancer is whether or not it is human epidermal growth factor receptor 2- (HER2-) positive, estrogen receptor- (ER-) positive and/or progesterone receptor- (PR-) positive. Triple positive

1816 | Analyst, 2014, 139, 1814–1825

groups mean that the patients have a close correlation with the three biomarkers, while triple negative ones suggest that doctors cannot diagnose patients based on these biomarkers. Another biomarker for women with a family history of breast cancer is BRCA1, which has been proven to be extremely useful for preventive measures.35 The ve-year survival rate of breast cancers is substantially higher if detected early. During the past few decades, the treatment of breast cancer has been greatly advanced due to the discovery of these biomarkers that could direct more individualized therapies to different molecular subgroups. The breast cancer biomarkers discussed in this review include HER2, carbohydrate antigen 15-3 (CA 15-3), ER/PR, BRCA1/BRCA2 and p53.36,37 HER2 and CA 15-3 are membrane proteins and the rest of the above biomarkers are intracellular proteins. 2.1

HER2

The expression level of HER2 increases in approximately 20–30% of breast cancer tumors.38,39 Breast cancer patients who are positive to HER2 have increased HER2 concentrations in blood (15–75 ng mL1) compared to normal individuals (2–15 ng mL1).40 The HER2-positive breast cancer tumors tend to be more aggressive and fast-growing than the HER2-negative ones.29 Shim and co-workers developed a sandwich-type electrochemical immunosensor for the detection of HER2 and HER2overexpressing breast cancer cells.41 The immunosensor label was prepared via functionalization of gold nanoparticles (AuNPs) with hydrazine and an aptamer that is specic to HER2. Aer HER2 and the immunosensor label (AuNPs coated with the reductant hydrazine) were sequentially captured with the anti-HER2 antibody-modied electrode, the electrode was exposed to silver ions, which are reduced by hydrazine. The deposited silver could be clearly seen under a microscope and analyzed by square wave stripping voltammetry (SWSV), and the amount of HER2 or HER2-overexpressing cells was determined (Fig. 1). This method provides a simple alternative to


View Article Online

Critical Review

Analyst

the sandwich-type immunosensors, the label-free sensors are simpler and cost-effective. However, the sensitivity is lower than that of the sandwich-structured sensors.


2.2

Top: schematic representation of the immunosensor for detection of HER2 protein and HER2-overexpresing breast cancer cells. GCE ¼ glassy carbon electrode; DPB ¼ 2,5-bis(2-thienyl)-1Hpyrrole-1-(p-benzoic acid). Bottom: calibration curves of the immunosensor in the range of (A) 0.0001–0.05 ng mL1 and (B) 0.1–10 ng mL1. The insets show the corresponding square wave stripping voltammograms for HER2 with varied concentrations. (Adapted from ref. 41 with permission. Copyright© 2013 American Chemical Society.) Fig. 1

selectively stain breast cancer cells, thus making cancer cells easily detectable. Marrazza et al. constructed an electrochemical immunosensor for the detection of HER2.42 Protein A-modied magnetic beads were utilized for the immobilization of capture anti-HER2 antibodies and alkaline phosphatase (AP) was chosen as the label on the detection anti-HER2 antibody. The electrochemical signal was generated via oxidation of 1-naphthol, which was produced by AP-catalyzed hydrolysis of 1-naphthyl phosphate. Ugo's group developed an electrochemical immunosensor for HER2 based on nanoelectrode ensembles (NEEs) prepared in track-etch polycarbonate membranes.43 Aer construction of the sandwich immuno-structure on the NEEs, horseradish peroxidase (HRP)-catalyzed electrochemical current in response to H2O2 was recorded. The advantage of the NEEs is the highly improved signal-to-background current ratio. Consequently, the detection limit was two to three orders of magnitude lower than that obtained with conventional electrodes. However, the electrode preparation is complicated. Lee and co-workers fabricated a label-free electrochemical impedance aptasensor for HER2 by immobilizing the HER2specic single-stranded DNA aptamer onto the electrode.44 Aer capture of different concentrations of HER2, the impedance of the electrode was increased, and the increased resistance is proportional to the concentration of HER2. In comparison with


ER/PR

ER and PR are both endocrine receptors.45,46 Estrogen and progesterone are essential for the growth and development of mammary glands and have been associated with the promotion and growth of breast cancers.47 Both estrogen and progesterone exert many effects via their receptors, ER and PR. There are two kinds of ERs, ERa and ERb.48 Given the ill-dened role of ERb in human breast cancer development, ERa is more commonly used as the target receptor for clinical decisions. PR levels are primarily regulated in response to ER.49 Li and collaborators constructed an electrochemical biosensor for the detection of ER relying on the ability of exonuclease III (Exo III) to digest DNA duplexes.50 Two complementary DNA probes were designed, one for selective binding of ER and the other for the subsequent hybridization and intercalation of methylene blue (MB) as the redox reporter. Aer formation of a stable duplex onto the electrode, a high electrochemical current of MB was observed. When Exo III was introduced to digest the duplex, the electrochemical response was decreased because of the removal of MB from the electrode surface. However, in the presence of ER, the digestion of Exo III was blocked due to the formation of the ER–DNA complex, leading to an increased electrochemical signal. Such a method possesses high sensitivity and the detection limit was estimated to be 0.38 nM. Miyashita et al. reported an SPR-based competitive immunosensor for the detection of estrogen and measurement of ER-binding activity.51 The pre-immobilized estrogen–BSA conjugates were captured by anti-estrogen antibody, yielding SPR signals. When different concentrations of estrogen are present with the antibody solution, the binding of the antibody to the chip is inhibited, leading to the change in the response signal. The method is rapid and simple, with a detection limit of 200 pg mL1. 2.3

CA 15-3

CA 15-3 belongs to the large family of glycoproteins encoded by the MUC1 gene. CA 15-3 has been clinically analyzed for breast cancer diagnosis of at-risk women, whose CA 15-3 level is typically greater than 30 U mL1, the threshold value.52,53 Typically, high concentrations of CA 15-3 were observed in up to 80% of metastatic breast cancers.54 The biomarkers detected in the CA 15-3 assay are the soluble forms of MUC-1 protein, which is a transmembrane protein consisting of two subunits that form a stable dimer. Several papers have reviewed the structure and functions of MUC-1.55,56 Lin's group reported an SPR-based immunosensor for the detection of CA 15-3 with a detection limit of 0.025 U mL1.57 The SPR chip was modied with a gold/zinc oxide (Au/ZnO) nanocomposite to enhance the chip performance. The sensitivity of the method was increased by at least 2 fold over that of the traditional gold/chromium (Au/Cr) lm. Analyst, 2014, 139, 1814–1825 | 1817

View Article Online


Analyst

Li et al. developed a label-free electrochemical immunosensor for the detection of CA 15-3.58 Capture anti-CA 15-3 antibody was immobilized onto the carboxylic acid groupfunctionalized graphene surface. The electrode conductivity was decreased aer capture of CA 15-3, as the antigen is nonconductive. This label-free immunosensor has a low detection limit of 0.012 U mL1, which is signicantly lower than the aforementioned threshold value. By examining the direct electrochemistry of glucose oxidase (GOD), Yuan and co-workers designed a label-free immunosensor for CA 15-3.59 Electrodes modied with carbon nanotubes and core–shell organosilica@chitosan nanospheres were further coated with GOD to achieve direct electron transfer between GOD and the electrode. The formation of the antibody– antigen immuno-complex onto the electrode decreased the electrochemical current of GOD due to the increased spatial blocking and dielectric constant of the microenvironment around the GOD molecules. The current is inversely proportional to the concentration of CA 15-3 and a detection limit of 0.04 U mL1 was estimated. This method has a relatively high variability because of the difficulty in achieving direct electrochemistry of GOD. In addition, the immunosensor is rather complex to prepare. Breast cancer cells can also be detected by monitoring the overexpression of MUC-1 proteins on the cancer cell surface. For example, Liu and co-workers developed an immunoassay for the breast cancer cells (MCF-7) utilizing MUC-1 aptamer-modied magnetic beads to capture MCF-7.60 Quantum dot-modied silica nanoparticles were used as a detection label. Quantitative detection of MCF-7 cells was achieved by both photoluminescence and square wave voltammetric measurements. In another study, they proposed a strategy for the detection of MUC-1 protein and MCF-7 cells based on the electrochemiluminescence resonance energy transfer (ERET) between the MUC-1 aptamer-functionalized ruthenium complex and graphene oxide (GO).61 Furthermore, Li et al. reported on a sandwichtype immunosensor for the detection of MCF-7 cells based on the MUC-1 aptamer and HRP label.62

2.4

BRCA1/BRCA2

BRCA1 and BRCA2 are human genes, capable of producing tumor suppressor proteins. Genetically inherited mutations of the BRCA1 or BRCA2 gene put women at high risk of breast and ovarian cancers.63 BRCA1 and BRCA2 mutations account for about 20 to 25% of hereditary breast cancers and about 5 to 10% of all breast cancers.64,65 Wei and colleagues developed a sandwich-type electrochemical immunosensor for the BRCA1 assay. They immobilized capture anti-BRCA1 antibody onto a graphene-modied electrode and used HRP-functionalized silica nanomaterials as a detection label. Under the optimized conditions, the electrochemical immunosensor exhibited a wide linear range from 0.01 to 15 ng mL1 and a detection limit of 4.86 pg mL1 (Fig. 2).66 In another study, a label-free electrochemical immunosensor for the BRCA1 assay has been developed using a toluidine blue (TB)–mesoporous

1818 | Analyst, 2014, 139, 1814–1825

Critical Review

carbon nanosphere nanocomposite-modied electrode.67 The specic antibody–antigen immunoreaction on the electrode surface resulted in a decrease in the redox signal of TB. However, the possibility of distinguishing between wild-type and mutated BRCA1 has not been attempted. By using interdigitated gold nanoelectrodes, Valle et al. described the electrochemical impedance assay of BRCA1, and demonstrated the differentiation between the wild-type and mutated BRCA1 genes.68 2.5

p53 protein

p53 is a tumor suppressor protein and a transcription factor. The function of p53 involves elimination and inhibition of the proliferation of abnormal cells. In more than 50% of cancer cases, the p53 gene has been found mutated.69 In breast cancers, p53 mutation is associated with more aggressive cancer development and worse overall survival. Gasco et al. have reviewed the p53 pathway in breast cancers.70 Recently, our group designed an electrochemical biosensor for quantication of the wild-type p53 protein.71 Consensus DNA duplexes were modied onto the electrode to capture wild-type p53 protein and the electrochemical signal was then amplied using ferrocene-capped gold nanoparticle–streptavidin conjugates. This method affords the sensitivity and selectivity necessary for detecting wild-type p53 protein in normal and cancer cell lysates. However, such a protocol did not allow an accurate determination of the extent of p53 mutation. In our other study, simultaneous and label-free detection of wild-type and mutant p53 present in cancer cell lysates was carried out using a dual channel SPR instrument.72 One channel of the SPR chip was modied with a consensus double-stranded (ds-) DNA while the other with a monoclonal antibody (Fig. 3). The high affinity of the antibody to total p53 protein (wild-type and mutant combined) and the consensus ds-DNA to wild-type p53 protein results in remarkably low detection levels (10.6 and 1.06 pM for the wild-type and total p53, respectively). We found that for liver and colon cancers, the extent of p53 mutation (>60%) is substantially greater than control cell lines. This method is label-free and does not involve additional amplication steps to enhance the sensitivity. Such SPR-based assays can potentially serve as a viable alternative for facile and sensitive clinical analyses. It has been reported that mutation of p53 also occurs in breast cancer patients.70 Thus we envision that our methods should be extendable to breast cancer studies and diagnosis. A summary of the dynamic ranges and detection limits of electroanalytical and SPR-based sensors for detection of breast cancer biomarkers is shown in Table 2.

3. Biomarkers related to neurological disorders Neurodegenerative disorders are characterized by neuronal impairment that eventually leads to neuronal death.73 They are commonly linked to the pathological aggregation of misfolded proteins and accumulation of intracellular and


View Article Online


Critical Review

Fig. 2

Analyst

Schematic representation of the fabrication of the immunosensor. (Adapted from ref. 66 with permission. Copyright© 2011 Elsevier.)

extracellular amyloid inclusions or brils in the central nervous system.74 For neurological disorders, body uids are perhaps the only samples that are available clinically, as tissue samples are only accessible postmortem. Assays of body uids also offer biological and chemical information, which complements information obtained with imaging techniques (e.g., magnetic resonance imaging, positronemission tomography and X-ray computed tomography) about morphological changes in specic regions of an organ. Similarly, proteins and peptides are either overproduced or down-regulated or cannot be effectively cleared in neurological disorders. The major neurological disorders include Alzheimer's disease (AD), Parkinson's disease, Huntington's disease, and Creutzfeldt–Jacob disease.75–77 We decided to focus on the assays of clinical samples for biomarkers related to AD, given that AD is the most prevalent neurological disorder and the fact that the two major hypotheses (amyloid beta cascade and tau atrophy) have biomarkers (amyloid beta peptide for the former75,78 and the tau protein for the latter79) that interact with many proteins in an extra- and/or intracellular milieu. In addition to proteins and peptides, other biomarkers, such as phospholipids, nucleic acids, small molecules (including neurotransmitters), and metal ions, are also indicative of a neurodegenerative disorder. A more indepth survey of non-peptide or -protein biomarkers related to


neurological disorders is provided in another review paper by our group.80

3.1

Ab peptides

Ab peptides of 39–43 amino acid residues, proteolytically cleaved from the amyloid precursor protein (APP), are major constituents of neuritic plaques. Both postmortem analyses of the senile plaques from AD patient brain extracts and in vitro Ab peptide aggregation studies have been extensively carried out.75 Vestergaard et al. reported on the rst electrochemical assay of Ab aggregation with the detection limit estimated to be approximately 0.7 mg mL1 for both Ab(1–40) and Ab(1–42).81 However, the detection limit is not low enough for the Ab assay in human uids (Ab(1–42) is present at 500 pg mL1 in cerebrospinal uid (CSF)).82 Oh et al. performed a multiplexed electrochemiluminescence assay of Ab(1–40) and Ab(1–42) levels in plasma.83 The method possesses a large dynamic range, consumes a small amount of plasma (0.5 mL) and can be completed within a short period of time (shorter than that needed for an ELISA measurement). Along this line, Islam et al. developed a simple microuidic device coupled with cyclic voltammetric detection of very small quantities of Ab(1–42) peptide.84

Analyst, 2014, 139, 1814–1825 | 1819

View Article Online


Analyst

Critical Review

Top: schematic representation of the simultaneous SPR detection of wild-type and total p53 proteins by consensus DNA duplexes (a) and monoclonal antibody (b) in two separate fluidic channels covering a dextran-modified Au sensor chip. Bottom: SPR sensorgrams correspond to (A) injections of 50 mL of 0.266 nM wild-type p53 solution into fluidic channels covered with consensus double-stranded (ds-) DNA (curve a), single-stranded (ss-) DNA (curve b), and nonconsensus ds-DNA (curve c) and (B) injections of 50 mL of 0.106 nM wild-type p53 solution (curve a) and 0.106 nM IgG solution (curve b) into fluidic channels covered with the monoclonal antibody. The arrows indicate the time when the injections were made. (Adapted from ref. 72 with permission. Copyright© 2009 American Chemical Society.) Fig. 3

Lee et al. utilized a gold nanoparticle–antibody complex to amplify SPR detection of synthetic Ab(1–40) peptide in buffer solution.85 A detection limit of 1.0 fg mL1 was achieved. However, the extent of nonspecic adsorption and selectivity of the method were not examined. Recently, we have

Table 2

accomplished sensitive and simultaneous SPR quantication of Ab(1–40) and Ab(1–42) peptides in human CSF.86 The signal amplication was achieved by using a conjugate formed between streptavidin and a biotinylated antibody that is selective to the common N-terminus of the Ab peptides (Fig. 4). The

Dynamic ranges and detection limits of electroanalytical and SPR sensors for detection of breast cancer biomarkers

Biomarker

Technique

Dynamic range

Detection limit

Reference

HER2

Electroanalytical chemistry

0.037 pg mL1 6 ng mL1 Not specied 0.01 pg mL1

41 42 43 44

ER/PR

Electroanalytical chemistry SPR SPR Electroanalytical chemistry

0.1 pg mL1 to 10 ng mL1 0–15 ng mL1 Not specied 0.01 pg mL1 to 100 ng mL1 0.5–100 nM Not specied 1–40 U mL1 0.1–20 U mL1 0.1–160 U mL1 250–10 000 cells per mL 100–2500 cells per mL 100–10 000 000 cells 0.01–15 ng mL1 0.01–15 ng mL1 Not specied 2.2 pM to 5.6 nM 1.06 pM to 53.2 nM

0.38 nM Not specied 0.025 U mL1 0.012 U mL1 0.04 U mL1 85 cells per mL 30 cells per mL 100 cells 4.86 pg mL1 3.97 pg mL1 Not specied 2.2 pM 1.06 pM

50 51 57 58 59 60 61 62 66 67 68 71 72

CA 15-3

MCF-7


BRCA1


p53

Electroanalytical chemistry SPR

1820 | Analyst, 2014, 139, 1814–1825


View Article Online

Analyst


Critical Review

Top: schematic diagram showing the simultaneous SPR detection of Ab(1–40) and Ab(1–42). Fluidic channels are covered with the capture antibody for Ab(1–40) and Ab(1–42) respectively. Injection of Ab samples results in the attachment of Ab to the respective channel (Step 1), and injection of the detection conjugate that can recognize the common hydrophilic domain of Ab(1–40) and Ab(1–42) leads to signal amplification (Step 2). Bottom: SPR sensorgrams after injections of (a) 1.00 and (b) 50.00 nM of Ab(1–40) into a precoated channel with an Ab(1–40) capture antibody, (c) curve (a) + 30.00 nM of Ab(1–16) detection antibody, and (d) curve (a) + 30.00 nM of the detection conjugate. The arrow indicates the time when the injections were made and the flow rate was 10 mL min1. (Adapted from ref. 86 with permission. Copyright© 2010 American Chemical Society.) Fig. 4

concentration ratio between Ab(1–40) and Ab(1–42) in CSF samples from AD patients has been found to be almost twice as high as that from healthy donors. Using an array of Ag nanotriangles, Van Duyne's group developed a localized SPR for sensitive detection of amyloid-b derived diffusible ligands (ADDLs) in both CSF and brain extracts.87 The SPR biosensors can potentially serve as a viable alternative to gain a better understanding of AD pathology and for AD diagnosis.


3.2

Tau proteins

Tau proteins promote assembly and maintain structural integrity of microtubules, support axon outgrowth, and regulate transport of vesicles and organelles.88 Levels of tau have been found to be increased in CSF of AD individuals compared to those in age-matched controls, probably due to neuronal and axonal degeneration or accumulation of neurobrillary tangles

Analyst, 2014, 139, 1814–1825 | 1821

View Article Online


Analyst

Critical Review

Fig. 5 Setup and measuring principle of cells by microelectrode-based impedance spectroscopy. (a) Microelectrode sensor chips comprise of a ring-like culture reservoir for maintaining cells onto microelectrodes. (b) Signals from 60 transparent indium tin oxide (ITO) can be recorded at once. (c) Each of the squared electrodes have sizes of 40 mm 40 mm. (d) For recording impedance signals, a small alternating voltage (10 mV) is applied. The resulting alternating current is frequency-dependent and flows from a working electrode through or between the cells to an opposite counter electrode. (Adapted from ref. 90 with permission. Copyright© 2009 RSC Publishing.)

(NFTs).89 Jahnke et al. reported a microelectrode-based impedimetric array sensor for label-free detection of hyperphosphorylated tau in the human neuroblastoma cell line, SHSY5Y (Fig. 5).90 Such a method could offer fast and sensitive testing of the efficiency of tau kinase inhibitors in vitro. Vesterdgaard et al. reported on multi-spot localized SPR detection of tau protein with a detection limit of 10 pg mL1, which is lower than the typical concentration of tau (around 195 pg mL1 in CSF).91

3.3

Ab-binding proteins

It has been suggested that Ab aggregation and polymerization are reduced by a broad range of Ab-binding proteins, including b-trace/prostaglandin D2 synthase (b-trace),92 transthyretin (TTR),93 cystatin C (CysC)94 and apoE.95 Ab-binding proteins may play an important role in preventing amyloid formation or

Table 3

clearing preformed oligomers, which are believed to be more toxic than Ab brils. Therefore, the concentrations of Abbinding proteins in body uids are also indicative of the etiology of AD. As a powerful technique, SPR has also been utilized to detect the interaction between Ab and Ab-binding proteins to gain insight into the role of these proteins in the pathology of AD.92,96,97 Table 3 lists the dynamic ranges and detection limits of electroanalytical and SPR-based sensors for detection of the above-mentioned AD biomarkers.

4. Conclusions and outlook A number of electroanalytical and SPR biosensors have been successfully developed for the detection of breast cancer and neurological biomarkers. Owing to the exceedingly low levels of many biomarkers, a variety of signal amplication schemes

Dynamic ranges and detection limits of electroanalytical and SPR sensors for detection of select AD biomarkers

Biomarker

Technique

Dynamic range

Detection limit

Reference

Ab(1–42)


Ab(1–40)

SPR Electroanalytical chemistry

Not specied 10–3000 pg mL1 100–300 mM 0.02–150.00 nM Not specied 74.9–5344.4 pg mL1 1.0–109 pg mL1 0.02–150.00 nM Not specied Not specied

0.7 mg mL1 10 pg mL1 10 mM 3.5 pM 0.7 mg mL1 74.9 pg mL1 1.0 fg mL1 3.3 pM Not specied 10 pg mL1

81 83 84 86 81 83 85 86 90 91

SPR Tau

Electroanalytical chemistry SPR

1822 | Analyst, 2014, 139, 1814–1825


View Article Online


Critical Review

have been used. For such a purpose, functionalized nanomaterials are particularly powerful. However, most of the reported methods can only detect one analyte at a time. As mentioned in the Introduction, relying on the detection of a single biomarker is insufficient to reach a denite diagnostic conclusion of many cancers. Clinically, more reliable information can be gleaned from high-throughput analyses of multiple disease biomarkers. In this regard, multiplexed detection of different analytes can be conveniently performed with multichannel electrochemical cells with micro-fabricated and nanofabricated working electrodes. Such electrodes are becoming quite reliable, disposable, and convenient to use. Similarly, imaging SPR using multiple microuidics-based channels or preprinted microarray chips has matured and can be readily implemented for multiplexed sensing. We envision that more applications of these devices for simultaneous detection of multiple biomarkers will appear in the coming years. As stated earlier, both electroanalytical and SPR-based sensors operate with recognition or detection events occurring at the surface–sample solution interface (i.e., heterogeneous sensors). As a consequence, overcoming non-specic adsorption of interferents in complicated sample matrices (e.g., blood samples) is paramount, as uncontrolled adsorption can produce false-positive results and in severe cases can even completely “foul” the SPR chips or electrode surfaces. Fortunately, a number of antifouling molecules or lms (e.g., dextran, chitosan, polyethylene glycol, etc.) have proven to be effective in eliminating or signicantly reducing adsorption of interfering species. In this aspect, implantation of these molecules to SPR chips is generally more straightforward than to electrode surfaces, as the attachment of antifouling molecules to electrochemical sensors could potentially block or impede electron transfer between the redox label and the underlying electrode. Another problem inherent in heterogeneous biosensors is the variability from sensor to sensor. It is not difficult to imagine that reproducibility in fabricating sensors with uniform coverage of probe molecules and well controlled surface orientation is key to the quality of the ultimate clinical results. Validation of diagnoses of cancers and neurological disorders by different laboratories or hospitals requires that identical responses be obtained for the same samples. These two major challenges are perhaps the most serious roadblocks for electrochemical and SPR-based sensing to be accepted as standard clinical methods. We should mention that many analytical gures of merit can be improved by combining SPR with electrochemistry or coupling either with other important analytical techniques (e.g., HPLC and mass spectrometry). We envisage that more biomarker detections will be achieved with such hybrid techniques involving either SPR or electrochemistry. Finally, for most, if not all, of the existing bioanalytical methods to be commercially viable, many clinical trials and tests must be conducted to establish their robustness and reliability. Thus far, most of the published or even patented methods are evaluated only with limited sample pools for proofof-concept purposes. The ultimate goal is to apply these methods to clinical assays and point-of-care testing. With the ongoing intensive effort as well as the advancement of


Analyst

nanotechnology, fabrication methods (including reliable mass production of chips and disposable electrodes), and electronics (for reducing noise and enhancing robustness), we believe that simple, reliable, and cost-effective analytical devices will eventually be available for clinical diagnosis of various diseases.

Acknowledgements Partial support of this work from the National Key Basic Research Program of China (2014CB744502), the National Natural Science Foundation of China (no. 21105128, 21375150, and 21175156), the NIH (SC1NS070155-01 to FZ), and an NSF Grant (no. 1112105 to FZ) is gratefully acknowledged.

References 1 J. A. L. Coronell, P. Syed, K. Sergelen, I. Gyurjan and A. Weinhaeusel, J. Proteomics, 2012, 76, 102–115. 2 M. Stastna and J. E. Van Eyk, Proteomics, 2012, 12, 722–735. 3 V. Kulasingam and E. P. Diamandis, Nat. Clin. Pract. Oncol., 2008, 5, 588–599. 4 J. Li, S. Li and C. F. Yang, Electroanalysis, 2012, 24, 2213– 2229. 5 J. F. Rusling, C. V. Kumar, J. S. Gutkind and V. Patel, Analyst, 2010, 135, 2496–2511. 6 P. J. Reddy, S. Sadhu, S. Ray and S. Srivastava, Clin. Lab. Med., 2012, 32, 47–72. 7 S. K. Arya and S. Bhansali, Chem. Rev., 2011, 111, 6783–6809. 8 M. Yang, J. Wang and F. Zhou, in Functional Nanoparticles for Bioanalysis, Nanomedicine, and Bioelectronic Devices, Vol. 1, ed. M. Hepel and C. J. Zhong, 2012, vol. 1112, pp. 177–205. 9 K. Blennow, H. Hampel, M. Weiner and H. Zetterberg, Nat. Rev. Neurol., 2010, 6, 131–144. 10 P. Davidsson and M. Sjogren, Dis. Markers, 2005, 21, 81–92. 11 A. Jemal, F. Bray, M. M. Center, J. Ferlay, E. Ward and D. Forman, Ca-Cancer J. Clin., 2011, 61, 69–90. 12 M. Simonato, J. Bennett, N. M. Boulis, M. G. Castro, D. J. Fink, W. F. Goins, S. J. Gray, P. R. Lowenstein, L. H. Vandenberghe, T. J. Wilson, J. H. Wolfe and J. C. Glorioso, Nat. Rev. Neurol., 2013, 9, 277–291. 13 J. Homola, Surface Plasmon Resonance Based Sensors, Springer, Berlin, 2006. 14 J. Homola, Chem. Rev., 2008, 108, 462–493. 15 A. J. Bard and L. R. Faulkner, Electrochemical Methods. Fundamentals and Applications, John Wiley & Sons, New York, 2001. 16 M. J. Linman, A. Abbas and Q. Cheng, Analyst, 2010, 135, 2759–2767. 17 S. Szunerits, N. Maalouli, E. Wijaya, J.-P. Vilcot and R. Boukherroub, Anal. Bioanal. Chem., 2013, 405, 1435– 1443. 18 D. W. Kimmel, G. LeBlanc, M. E. Meschievitz and D. E. Cliffel, Anal. Chem., 2012, 84, 685–707. 19 M.-I. Mohammed and M. P. Y. Desmulliez, Lab Chip, 2011, 11, 569–595.

Analyst, 2014, 139, 1814–1825 | 1823

View Article Online


Analyst

20 Y. Ding, D. Li, B. Li, K. Zhao, W. Du, J. Zheng and M. Yang, Biosens. Bioelectron., 2013, 48, 281–286. 21 Q. Yuan, D. Lu, X. Zhang, Z. Chen and W. Tan, TrAC, Trends Anal. Chem., 2012, 39, 72–86. 22 H. Qi, X. Qiu, D. Xie, C. Ling, Q. Gao and C. Zhang, Anal. Chem., 2013, 85, 3886–3894. 23 F. Liu, H. Liu, M. Zhang, J. Yu, S. Wang and J. Lu, Analyst, 2013, 138, 3463–3469. 24 X. Yi, Y. Hao, N. Xia, J. Wang, M. Quintero, D. Li and F. Zhou, Anal. Chem., 2013, 85, 3660–3666. 25 L. Liu, N. Xia and J. Wang, RSC Adv., 2012, 2, 2200–2204. 26 J. Xiang, J. Guo and F. M. Zhou, Anal. Chem., 2006, 78, 1418– 1424. 27 S. Boussaad, J. Pean and N. J. Tao, Anal. Chem., 2000, 72, 222–226. 28 A. W. Peterson, L. K. Wolf and R. M. Georgiadis, J. Am. Chem. Soc., 2002, 124, 14601–14607. 29 S. Dent, B. Oyan, A. Honig, M. Mano and S. Howell, Cancer Treat. Rev., 2013, 39, 622–631. 30 H. Zhu, P. S. Dale, C. W. Caldwell and X. Fan, Anal. Chem., 2009, 81, 9858–9865. 31 A. Goldhirsch, E. P. Winer, A. S. Coates, R. D. Gelber, M. Piccart-Gebhart, B. Thuerlimann, H. J. Senn and M. Panel, Ann. Oncol., 2013, 24, 2206–2223. 32 A. Goldhirsch, W. C. Wood, A. S. Coates, R. D. Gelber, B. Thuerlimann, H. J. Senn and M. Panel, Ann. Oncol., 2011, 22, 1736–1747. 33 H. M. Linden and F. Dehdashti, Semin. Nucl. Med., 2013, 43, 324–329. 34 S. S. Raab, J. Swain, N. Smith and D. M. Grzybicki, Clin. Biochem., 2013, 46, 1180–1186. 35 E. Levy-Lahad and E. Friedman, Br. J. Cancer, 2007, 96, 11–15. 36 I. Diaconu, C. Cristea, V. Harceaga, G. Marrazza, I. Berindan-Neagoe and R. Sandulescu, Clin. Chim. Acta, 2013, 425C, 128–138. 37 A. M. Badowska-Kozakiewicz, M. Sobol, J. Patera and W. Kozlowski, Arch. Med. Sci., 2013, 9, 466–471. 38 K. D. Cole, H.-J. He and L. Wang, Proteomics: Clin. Appl., 2013, 7, 17–29. 39 D. J. Slamon, G. M. Clark, S. G. Wong, W. J. Levin, A. Ullrich and W. L. McGuire, Science, 1987, 235, 177–182. 40 J. A. Capobianco, W. Y. Shih, Q.-A. Yuan, G. P. Adams and W.-H. Shih, Rev. Sci. Instrum., 2008, 79, 076101. 41 Y. Zhu, P. Chandra and Y.-B. Shim, Anal. Chem., 2013, 85, 1058–1064. 42 Q. A. M. Al-Khafaji, M. Harris, S. Tombelli, S. Laschi, A. P. F. Turner, M. Mascini and G. Marrazza, Electroanalysis, 2012, 24, 735–742. 43 S. P. Mucelli, M. Zamuner, M. Tormen, G. Stanta and P. Ugo, Biosens. Bioelectron., 2008, 23, 1900–1903. 44 L. Chun, S.-E. Kim, M. Cho, W.-s. Choe, J. Nam, D. W. Lee and Y. Lee, Sens. Actuators, B, 2013, 186, 446–450. 45 L. Braun, F. Mietzsch, P. Seibold, A. Schneeweiss, P. Schirmacher, J. Chang-Claude, H. P. Sinn and S. Aulmann, Mod. Pathol., 2013, 26, 1161–1171. 46 E. Curtit, V. Nerich, L. Mansi, L. Chaigneau, L. Cals, C. Villanueva, F. Bazan, P. Montcuquet, N. Meneveau,

1824 | Analyst, 2014, 139, 1814–1825

Critical Review

47 48 49 50 51 52 53 54

55 56 57 58 59 60 61 62 63

64 65 66 67 68

69 70 71 72 73 74 75

S. Perrin, M.-P. Algros and X. Pivot, Oncologist, 2013, 18, 667–674. A. E. Obr and D. P. Edwards, Mol. Cell. Endocrinol., 2012, 357, 4–17. S. J. Neo, X. Su and J. S. Thomsen, Anal. Chem., 2009, 81, 3344–3349. B. M. Jacobsen and K. B. Horwitz, Mol. Cell. Endocrinol., 2012, 357, 18–29. S. Zhu, Y. Cao, Y. Xu, Y. Yin and G. Li, Int. J. Mol. Sci., 2013, 14, 10298–10306. M. Miyashita, T. Shimada, H. Miyagawa and M. Akamatsu, Anal. Bioanal. Chem., 2005, 381, 667–673. N. Samy, H. M. Ragab, N. Abd El Maksoud and M. Shaalan, Cancer Biomarkers, 2010, 6, 63–72. M. J. Duffy, Ann. Clin. Biochem., 1999, 36, 579–586. M. J. Duffy, C. Duggan, R. Keane, A. D. K. Hill, E. McDermott, J. Crown and N. O'Higgins, Clin. Chem., 2004, 50, 559–563. D. W. Kufe, Nat. Rev. Cancer, 2009, 9, 874–885. C. L. Hattrup and S. J. Gendler, Annu. Rev. Physiol., 2008, 70, 431–457. C.-C. Chang, N.-F. Chiu, D. S. Lin, Y. Chu-Su, Y.-H. Liang and C.-W. Lin, Anal. Chem., 2010, 82, 1207–1212. H. Li, J. He, S. Li and A. P. Turner, Biosens. Bioelectron., 2013, 43, 25–29. W. Li, R. Yuan, Y. Chai and S. Chen, Biosens. Bioelectron., 2010, 25, 2548–2552. X. Hua, Z. Zhou, L. Yuan and S. Liu, Anal. Chim. Acta, 2013, 788, 135–140. W. Wei, D. F. Li, X. H. Pan and S. Q. Liu, Analyst, 2012, 137, 2101–2106. X. Zhu, J. Yang, M. Liu, Y. Wu, Z. Shen and G. Li, Anal. Chim. Acta, 2013, 764, 59–63. A. Finch, K. Metcalfe, J. Lui, C. Springate, R. Demsky, S. Armel, B. Rosen, J. Murphy, L. Elit, P. Sun and S. Narod, Clin. Genet., 2009, 75, 220–224. A. R. Venkitaraman, Nat. Rev. Cancer, 2004, 4, 266–276. P. M. Campeau, W. D. Foulkes and M. D. Tischkowitz, Hum. Genet., 2008, 124, 31–42. Y. Cai, H. Li, B. Du, M. Yang, Y. Li, D. Wu, Y. Zhao, Y. Dai and Q. Wei, Biomaterials, 2011, 32, 2117–2123. H. Fan, Y. Zhang, D. Wu, H. Ma, X. Li, Y. Li, H. Wang, H. Li, B. Du and Q. Wei, Anal. Chim. Acta, 2013, 770, 62–67. A. Bonanni, I. Fern´ andez-Cuesta, X. Borris´ e, F. P´ erezMurano, S. Alegret and M. Valle, Microchim. Acta, 2010, 170, 275–281. A. J. Levine, Cell, 1997, 88, 323–331. M. Gasco, S. Shami and T. Crook, Breast Cancer Res., 2002, 4, 70–76. J. Wang, X. Zhu, Q. Tu, Q. Guo, C. S. Zarui, J. Momand, X. Z. Sun and F. Zhou, Anal. Chem., 2008, 80, 769–774. Y. Wang, X. Zhu, M. Wu, N. Xia, J. Wang and F. Zhou, Anal. Chem., 2009, 81, 8441–8446. D. C. Crowther, Hum. Mutat., 2002, 20, 1–14. L. M. Shaw, M. Korecka, C. M. Clark, V. M. Y. Lee and J. Q. Trojanowski, Nat. Rev. Drug Discovery, 2007, 6, 295–303. J. Hardy and D. J. Selkoe, Science, 2002, 297, 353–356.


View Article Online


Critical Review

76 J.-H. Kang, M. Korecka, J. B. Toledo, J. Q. Trojanowski and L. M. Shaw, Clin. Chem., 2013, 59, 903–916. 77 E.-M. Mandelkow and E. Mandelkow, Cold Spring Harbor Perspect. Med., 2012, 2, a006247. 78 D. J. Selkoe, Science, 2002, 298, 789–791. 79 J. Avila, J. J. Lucas, M. Perez and F. Hernandez, Physiol. Rev., 2004, 84, 361–384. 80 S. Chen, L. Zhang, Y. Long and F. Zhou, Electroanalysis, 2014, accepted. 81 M. Vestergaard, K. Kerman, M. Saito, N. Nagatani, Y. Takamura and E. Tamiya, J. Am. Chem. Soc., 2005, 127, 11892–11893. 82 H. Hampel, K. Buerger, S. J. Teipel, A. L. W. Bokde, H. Zetterberg and K. Blennow, Alzheimer's Dementia, 2008, 4, 38–48. 83 E. S. Oh, M. M. Mielke, P. B. Rosenberg, A. Jain, N. S. Fedarko, C. G. Lyketsos and P. D. Mehta, J. Alzheimers Dis., 2010, 21, 769–773. 84 K. Islam, Y.-C. Jang, R. Chand, S. K. Jha, H. H. Lee and Y.-S. Kim, J. Nanosci. Nanotechnol., 2011, 11, 5657–5662. 85 J.-H. Lee, D.-Y. Kang, T. Lee, S.-U. Kim, B.-K. Oh and J.-W. Choi, J. Nanosci. Nanotechnol., 2009, 9, 7155–7160. 86 N. Xia, L. Liu, M. G. Harrington, J. Wang and F. Zhou, Anal. Chem., 2010, 82, 10151–10157. 87 A. J. Haes, L. Chang, W. L. Klein and R. P. Van Duyne, J. Am. Chem. Soc., 2005, 127, 2264–2271. 88 E. Mandelkow, Nature, 1999, 402, 588–589. 89 M. Sjogren, H. Vanderstichele, H. Agren, O. Zachrisson, M. Edsbagge, C. Wikkelso, I. Skoog, A. Wallin, L. O. Wahlund, J. Marcusson, K. Nagga, N. Andreasen, P. Davidsson, E. Vanmechelen and K. Blennow, Clin. Chem., 2001, 47, 1776–1781. 90 H.-G. Jahnke, A. Rothermel, I. Sternberger, T. G. A. Mack, R. G. Kurz, O. Paenke, F. Striggow and A. A. Robitzki, Lab Chip, 2009, 9, 1422–1428.


Analyst

91 M. d. Vestergaard, K. Kerman, D.-K. Kim, H. M. Hiep and E. Tamiya, Talanta, 2008, 74, 1038–1042. 92 T. Kanekiyo, T. Ban, K. Aritake, Z.-L. Huang, W.-M. Qu, I. Okazaki, I. Mohri, S. Murayama, K. Ozono, M. Taniike, Y. Goto and Y. Urade, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 6412–6417. 93 A. L. Schwarzman, L. Gregori, M. P. Vitek, S. Lyubski, W. J. Strittmatter, J. J. Enghilde, R. Bhasin, J. Silverman, K. H. Weisgraber, P. K. Coyle, M. G. Zagorski, J. Talafous, M. Eisenberg, A. M. Saunders, A. D. Roses and D. Goldgaber, Proc. Natl. Acad. Sci. U. S. A., 1994, 91, 8368–8372. 94 S. A. Kaeser, M. C. Herzig, J. Coomaraswamy, E. Kilger, M.-L. Selenica, D. T. Winkler, M. Staufenbiel, E. Levy, A. Grubb and M. Jucker, Nat. Genet., 2007, 39, 1437– 1439. 95 W. J. Strittmatter, A. M. Saunders, D. Schmechel, M. Pericakvance, J. Enghild, G. S. Salvesen and A. D. Roses, Proc. Natl. Acad. Sci. U. S. A., 1993, 90, 1977– 1981. 96 M. Sugano, K. Yamauchi, K. Kawasaki, M. Tozuka, K. Fujita, N. Okumura and H. Ota, Clin. Chim. Acta, 2008, 388, 123– 129. 97 V. V. Shuvaev and G. Siest, FEBS Lett., 1996, 383, 9–12. 98 P. M. Goodwin, W. P. Ambrose and R. A. Keller, Acc. Chem. Res., 1996, 29, 607–613. 99 J. Li, S. Guo and E. Wang, RSC Adv., 2012, 2, 3579–3586. 100 Z. Lin, H. Chen and J.-M. Lin, Analyst, 2013, 138, 5182– 5193. 101 C. I. Cheng, Y.-P. Chang and Y.-H. Chu, Chem. Soc. Rev., 2012, 41, 1947–1971. 102 E. P. Randviir and C. E. Banks, Anal. Methods, 2013, 5, 1098– 1115. 103 B. N. Johnson and R. Mutharasan, Biosens. Bioelectron., 2012, 32, 1–18.

Analyst, 2014, 139, 1814–1825 | 1825

View Article Online

Analyst


PAPER


A two-photon “turn-on” fluorescent probe based on carbon nanodots for imaging and selective biosensing of hydrogen sulfide in live cells and tissues† Anwei Zhu,‡a Zongqian Luo,‡b Changqin Ding,a Bo Li,c Shuang Zhou,d Rong Wang*b and Yang Tian*a Determination of hydrogen sulfide (H2S) in live cells and tissues is still a challenge for evaluating the key roles that H2S plays in physiological and pathological processes. In this work, a “turn-on” two-photon fluorescent (TPF) sensor for H2S is developed, in which carbon nanodot (C-Dot) was employed as a two-photon fluorophore due to its large two-photon absorption cross-section (s), and AE-TPEA–Cu2+ complex [AE-TPEA ¼ N-(2-aminoethyl)-N,N,N0 -tris(pyridin-2-ylmethyl)ethane-1,2-diamine] was first designed as a specific receptor for H2S. The fluorescence of C-Dot conjugated with AE-TPEA (C-Dot-TPEA) was quenched upon the addition of Cu2+. Then, the fluorescence was restored after the addition of H2S, because Cu2+ could be released from TPEA binding site when H2S interacted with the Cu2+ ion. The designed C-Dot-TPEA–Cu2+ fluorescent sensor exhibited high specificity for H2S over biothiols, sulfurcontaining compounds, reactive oxygen species (ROS), and other biological interferences. Meanwhile, a broad linear range from 5 mM to 100 mM was obtained and the detection limit was achieved to 0.7 mM. In addition, the C-Dot-based TPF probe exhibited bright two-photon fluorescence, favourable photostability against light illumination and pH change, and low cytotoxicity. Accordingly, the

Received 7th November 2013 Accepted 20th January 2014

nanohybridized TPF sensor with high selectivity and sensitivity, as well as the fascinating properties of C-Dot themselves, successfully provided a new way for TPF imaging and biosensing of H2S in live cells

DOI: 10.1039/c3an02086j

and tissues. We believe this is the first report of TPF imaging and biosensing of H2S in live cells and

www.rsc.org/analyst

tissues using a specially engineered C-Dot-based nanosystem.

Introduction Over the past decades, two-photon uorescent (TPF) probes have attracted much attention for their applications in the eld of biomedical imaging and biosensing, because of the low background signal, deep tissue penetration depth, reduced photobleaching, and low phototoxicity associated with the use of near infrared two-photon excitation for TPF probes.1–5 Recently, carbon nanodots (C-Dot) have emerged as promising TPF probes due to

a

Department of Chemistry, Tongji University, Shanghai 200092, China. E-mail: [email protected]

b

Department of Chemistry, Shanghai Normal University, Shanghai 200234, China

c

Key Laboratory of Polar Materials and Devices, Ministry of Education, East China Normal University, Shanghai 200241, China d

Department of Histology and Embryology, School of Medicine, Tongji University, Shanghai 200092, China † Electronic supplementary information (ESI) available: NMR and MS spectra of AE-TPEA, apoptosis assay, and two-photon uorescence microscopy images of RAW264.7 cells, HEK 293T cells and lung cancer tissue slice. See DOI: 10.1039/c3an02086j

their unique properties such as high water solubility, favourable biocompatibility, and long-term photostability.6–12 However, exploration of C-Dot for two-photon imaging and sensing still remains at an early stage. In this work, we report the design and synthesis of a two-photon uorescent probe using C-Dot-based inorganic–organic nanohybrids for imaging and biosensing of physiological hydrogen sulde (H2S) in live cells and tissues. Although H2S is well known for the rotten egg smell, recent studies have demonstrated that H2S is the third gaseous transmitter, in addition to nitric oxide (NO) and carbon monoxide (CO).13,14 H2S contributes to various physiological processes, including relaxation of vascular smooth muscles, mediation of neurotransmission, regulation of inammation and O2 sensing, and it can also protect against ischemia/ reperfusion injury.15–19 However, disruption of the H2S level is linked to various diseases such as Alzheimer's disease, Down's syndrome, diabetes, and liver cirrhosis.20–23 So far, several elegant techniques have been developed for the detection of H2S, including colorimetric and electrochemical assays, gas chromatography, and metal-induced sulde precipitation.24–30 Fluorescence imaging and biosensing with spatial or

‡ These authors contributed equally.


Analyst, 2014, 139, 1945–1952 | 1945

View Article Online


Analyst

temporal resolution for typical cell-based or live tissue experiments have attracted great attention in recent years.31–38 However, most uorescent probes for H2S have been evaluated using onephoton microscopy and require relatively short excitation wavelengths,31–35 limiting their use in deep-tissue imaging because of the shallow penetration depth (400 mg mL 1 and for miconazole at >0.5 mg mL 1, which represent the minimum inhibitory concentrations of two drugs. At lower concentrations, the volatile metabolites are clearly changed, relative to untreated fungi, which can lead to either greater or diminished sensor array response (Fig. 5B). While the colorimetric sensor array does not provide a direct indication of what the changes in the VOCs may be (i.e., the component by component analysis provided by GC-MS, for


Fig. 5 Sensor array response to fungal volatiles under the influence of anti-fungal drugs. (A) Total array response (change in Euclidean distance after subtraction of background) under the influence of fungicidal drugs. Significant changes in array response are observed vs. clotrimazole or miconazole concentrations. (B) Colour difference maps vs. YPD growth medium control after 400 min as a function of fungicide concentration. Changes in the array response patterns appear to occur at $10 ng mL 1 clotrimazole and at $1 ng mL 1 for miconazole, which are presumably due to changes of metabolic states under the stress of the anti-fungal drug environment. The complete shutdown of volatile metabolites occurs at >400 mg mL 1 clotrimazole and at >0.5 mg mL 1 miconazole.

example), it does yield a rapid and simple indication that signicant metabolic changes are occurring. This could prove useful for rapid parallel screening of fungicidal effectiveness.

4 Conclusions In summary, we have used an optoelectronic nose to detect fungal VOCs and generate unique metabolic patterns that differentiate among twelve different fungal strains with high accuracy. The sensor array is also able to observe metabolic changes in growth patterns upon the addition of fungicides, which provides a facile screening tool for determining fungicide efficacy in real time.

Acknowledgements This work was supported through NIH GEI award U01ES016011. We thank Professor Lois L. Hoyer for generously providing

Analyst, 2014, 139, 1922–1928 | 1927

View Article Online


Analyst

Paper

fungal strains. We thank Jacqueline M. Rankin and John R.G. Sander for helpful suggestions and advice. KSS has nancial interests in iSense LLC, which is commercializing applications of colorimetric sensor arrays. No funding from iSense was involved in this research.

20

References

22

1 G. D. Brown, D. W. Denning, N. A. R. Gow, S. M. Levitz, M. G. Netea and T. C. White, Sci. Transl. Med., 2012, 4, 1–9. 2 G. D. Brown, D. W. Denning and S. M. Levitz, Science, 2012, 336, 647. 3 M. Nucci and K. A. Marr, Clin. Infect. Dis., 2005, 41, 521–526. 4 P. R. Murray and H. Masur, Crit. Care Med., 2012, 40, 3277– 3282. 5 B. M. Kuehn, JAMA, J. Am. Med. Assoc., 2013, 309, 219–221. 6 J. S. Klutts and B. Robinson-Dunn, J. Clin. Microbiol., 2011, 49, S39–S42. 7 S. F. Yeo and B. Wong, Clin. Microbiol. Rev., 2002, 15, 465– 484. 8 S. U. Morath, R. Hung and J. W. Bennett, Fungal Biol. Rev., 2012, 26, 73–83. 9 N. Sahgal, B. Monk, M. Wasil and N. Magan, Br. J. Dermatol., 2006, 155, 1209–1216. 10 J. M. Scotter, V. S. Langford, P. F. Wilson, M. J. McEwan and S. T. Chambers, J. Microbiol. Methods, 2005, 63, 127–134. 11 N. Sahgal and N. Magan, Sens. Actuators, B, 2008, 131, 117– 120. 12 K. Naraghi, N. Sahgal, B. Adriaans, H. Barr and N. Magan, Sens. Actuators, B, 2010, 146, 521–526. 13 N. P. Pont, C. A. Kendall and N. Magan, Mycopathologia, 2012, 173, 93–101. 14 F. Rock, N. Barsan and U. Weimar, Chem. Rev., 2008, 108, 705–725. 15 N. A. Rakow and K. S. Suslick, Nature, 2000, 406, 710–713. 16 K. S. Suslick, Curr. Opin. Chem. Biol., 2012, 16, 557–563. 17 S. H. Lim, L. Feng, J. W. Kemling, C. J. Musto and K. S. Suslick, Nat. Chem., 2009, 1, 562–567. 18 M. C. Janzen, J. B. Ponder, D. P. Bailey, C. K. Ingison and K. S. Suslick, Anal. Chem., 2006, 78, 3591–3600. 19 K. S. Suslick, D. P. Bailey, C. K. Ingison, M. Janzen, M. E. Kosal, W. B. McNamara III, N. A. Rakow, A. Sen,

23

1928 | Analyst, 2014, 139, 1922–1928

21

24 25 26

27

28 29

30 31 32

33

34

35 36

J. J. Weaver, J. B. Wilson, C. Zhang and S. Nakagaki, Quim. Nova, 2007, 30, 677–681. L. Feng, C. J. Musto, J. W. Kemling, S. H. Lim, W. Zhong and K. S. Suslick, Anal. Chem., 2010, 82, 9433–9440. L. Feng, C. J. Musto, J. W. Kemling, S. H. Lim and K. S. Suslick, Chem. Commun., 2010, 46, 2037–2039. H. Lin and K. S. Suslick, J. Am. Chem. Soc., 2010, 132, 15519– 15521. B. A. Suslick, L. Feng and K. S. Suslick, Anal. Chem., 2010, 82, 2067–2073. C. Zhang, D. P. Bailey and K. S. Suslick, J. Agric. Food Chem., 2006, 54, 4925–4931. C. Zhang and K. S. Suslick, J. Agric. Food Chem., 2007, 55, 237–242. J. R. Carey, K. S. Suslick, K. I. Hulkower, J. A. Imlay, K. R. Imlay, C. K. Ingison, J. B. Ponder, A. Sen and A. E. Wittrig, J. Am. Chem. Soc., 2011, 133, 7571–7576. P. J. Mazzone, X. F. Wang, Y. M. Xu, T. Mekhail, M. C. Beukemann, J. Na, J. W. Kemling, K. S. Suslick and M. Sasidhar, J. Thorac. Oncol., 2012, 7, 137–142. S. J. Haswell, Practical guide to chemometrics, CRC Press, 1992. W. Zhong and K. S. Suslick, Technometrics, 2014, 56, in press, http://publish.illinois.edu/wenxuan/les/2012/2012/ matrix_clust.pdf. B. Li, K. M. Kim and N. Altman, Ann. Statist., 2010, 38, 1094– 1121. P. Zeng and W. Zhong, Topics Appl. Stat., 2013, 55, 213–227. M. Desnos-Ollivier, M. Ragon, V. Robert, D. Raoux, J.-C. Gantier and F. Dromer, J. Clin. Microbiol., 2008, 46, 3237–3242. G. L. Newton, K. Arnold, M. S. Price, C. Sherrill, S. B. Delcardayre, Y. Aharonowitz, G. Cohen, J. Davies, R. C. Fahey and C. Davis, J. Bacteriol., 1996, 178, 1990–1995. D. L. Donoho, AMS Math Challenges Lecture, 2000, pp. 1–32, http://www-stat.stanford.edu/donoho/Lectures/AMS2000/ Curses.pdf. P. Belenky, D. Camacho and J. J. Collins, Cell Rep., 2013, 3, 350–358. R. D. Cannon, E. Lamping, A. R. Holmes, K. Niimi, K. Tanabe, M. Niimi and B. C. Monk, Microbiology, 2007, 153, 3211–3217.


View Article Online

Analyst


PAPER


Novel method of screening the oxidation and reduction abilities of photocatalytic materials† K. Katayama,*a Y. Takeda,a K. Shimaoka,a K. Yoshida,a R. Shimizu,a T. Ishiwata,a A. Nakamura,a S. Kuwahara,a A. Mase,b T. Sugitab and M. Morib Two analytical methods for the evaluation of photocatalytic oxidation and reduction abilities were developed using a photocatalytic microreactor; one is product analysis and the other is reaction rate analysis. Two simple organic conversion reactions were selected for the oxidation and reduction. Since the reactions were one-to-one conversions from the reactant species to the product species, the


product analysis was simply performed using gas chromatography, and the reactions were monitored in situ in the photocatalytic microreactor using the UV absorption spectra. The partial oxidation and reduction abilities for each functional group can be judged from the yield and selectivity, and the

DOI: 10.1039/c3an02167j www.rsc.org/analyst

corresponding reaction rate, while the total oxidation ability can be judged from the conversion. We demonstrated the application of these methods for several kinds of visible light photocatalysts.

Introduction Photocatalytic materials, based on titanium oxide (TiO2), have been commercialized for the purposes of self-cleaning walls in houses, toilets, interior walls in tunnels, etc. In principle, if a photocatalyst is irradiated with light, the hydroxyl radical is generated through the oxidation of water, and the superoxide anion is generated via the reduction of oxygen under aqueous or ambient conditions. The active species are in equilibrium with water, and hydrogen peroxide is generated. Such active species are called active oxygen species (AOS), and they induce the photocatalytic degradation of organic molecules. They are also subject to degradation due to direct oxidization by the holes generated at the surface of photocatalysts.1 In most of the applications of photocatalytic reactions, only the oxidation ability, especially total oxidation from organics to carbon dioxide and water, has attracted attention. However, photocatalysts also have the abilities of partial or selective oxidation,2–7 and reduction8–10 for various aromatic compounds.11,12 Selective oxidation of the alcohol functional group gave a high yield.13 The reaction mechanism of the reduction of nitro compounds was investigated in detail,8,14,15 and imine formation was reported aer reacting with alcohols.16 Furthermore, complicated reactions involving the

a

Department of Applied Physics, Chuo University, 1-13-27 Kasuga, Bunkyo, 112-8656, Tokyo, Japan

b

Graduate School of Engineering, Gunma University, 1-5-1 Tnejin-cho, Kiryu, Gunma 376-8515, Japan † Electronic supplementary information (ESI) available: Experimental setups are given in Fig. S1–S3. Examples of the tting of the UV spectra using the reference spectra of the reactants and products are given in Fig. S4. See DOI: 10.1039/c3an02167j


photocatalytically generated intermediate radical species were also observed.5,17,18 In recent years, the demand for visible light responsive materials has been increasing, and many papers on the development of new materials have been published. This is not only for organic decomposition, but also for generation of solar fuels and light harvesting applications.19–21 Although the development of materials has been intensively explored, the analytical methods for understanding the abilities of the materials have not changed. Typical methods of examining photocatalytic materials are the color degradation of methylene blue in aqueous solutions22,23 and the decomposition of acetaldehyde under ambient conditions,24 dened in the Japanese Industrial Standard (JIS). These methods give information on total oxidation ability. There is another promising method, where the oxidation of dimethylsulfoxide is observed by ion chromatography.25 Since methyl groups are replaced with hydroxyl radicals in a stepwise manner, the obtained products provide direct information on the hydroxyl radical activity. From our point of view, the analytical methods for photocatalysts are not sufficient to cover many recently developed materials, which may have other potential applications. The problems with the conventional analytical methods are as follows: (1) under aqueous or ambient conditions, various AOSs, depending on the experimental conditions (pH, humidity, coexisting species), are involved in photodecomposition, and it is difficult to determine which one is the actual cause of the reaction; (2) methylene blue is not a good test material for dye degradation because it is subject to multi-point reactions in the molecule at the same time, as we claried in a previous study,26 and the color degradation does not necessarily reect the total oxidation ability.

Analyst, 2014, 139, 1953–1959 | 1953

View Article Online


Analyst

We recently developed a spectroscopic monitoring system for photocatalytic reactions using a microreactor26 to reduce the reaction time, and successfully accomplished the quick monitoring of a photocatalytic reaction in a capillary. There are also several other reports on the decrease in reaction time by utilizing microreactors in photocatalytic reactions.27–32 We studied the reaction mechanisms for various dyes composed of different main structures, and we successfully performed a kinetic analysis and claried multi-step reactions, self-catalytic reactions, and more complicated reactions by monitoring the intermediate species involved in each reaction. The results suggested that typical analyses based on the quasi-rst-order or Langmuir–Hinshelwood mechanisms33,34 are not sufficient to clarify the reaction processes correctly. Furthermore, we proposed a new concept of a photocatalytic microreactor, in which no electrical resources are necessary to perform the photocatalytic reaction due to the use of a natural force such as capillary force or diffusion due to the concentration gradient, and successfully demonstrated its use in our previous paper.35 In this study, we propose a new methodology to clarify the actual abilities of photocatalytic materials, by obtaining the selective oxidation, reduction and total oxidation abilities and the corresponding reaction rates through simple screening methods using a photocatalytic microreactor. We performed some of the oxidation and reduction reactions in organic solvents, not in water, to avoid complex reactions involving various AOSs. By analysing the simple oxidation and reduction reactions, the electron or hole donation abilities can be understood. We applied two methods for these reactions; one is product analysis to obtain the conversion, yield and selectivity, giving information on the total oxidation, partial oxidation and reduction abilities, and the other is reaction rate analysis by monitoring the reaction directly inside the microreactor using UV spectra.

Experimental For the product analyses, photocatalytic reactions were performed using the automatic photocatalytic reactors we developed in our previous paper.35 The principle is briey explained as follows (Fig. S1†): photocatalytic microreactors were prepared by coating photocatalytic materials inside a capillary with both ends open. Six capillaries were bundled together and the bottoms were soaked in a reagent solution inside a test tube. Due to the capillary force, several centimeters of the solution entered the capillaries, and UV or visible light irradiation was used to induce the reaction. Since the reactant was decomposed by light irradiation, a concentration gradient formed inside the microreactors, driving the reactants beneath the capillary bundle into the microreactors due to diffusion, and vice versa for the products. The reaction then continued until the concentration gradient disappeared; thus, all of the reactants were reacted. The photocatalyst was coated inside a fused silica capillary (I.D. 1.1 mm, O.D. 1.4 mm) with a length of 6 cm. The capillary was transparent, and was treated using an alkaline solution before coating. The capillary was soaked in paste, lled with the

1954 | Analyst, 2014, 139, 1953–1959

Paper

solution three times, and heated to 450 C in a furnace for 1 h. Six capillaries were bundled together, and were placed in a test tube containing a reactant solution (1 mL). Due to the capillary force, the reactant solution typically rose a few centimeters above the solution surface. The test tubes were loaded onto a merry-go-round type photoreaction stage with 8 test tube holders (Fig. S2†), and a light was shone from the side at a distance of about 10 cm, typically for 18 h. The reaction time was decided in advance by checking the necessary reaction time for 1 mL of a reactant solution with 6 photocatalytic microreactors. A UV-LED with a wavelength of 365 nm and an intensity of 360 mW cm2 was used as a UV source, and a quasi sunlight source (XC-100, Solax) was used as a visible light source. Aer the reaction, the solution was analysed using gas chromatography (Shimadzu, GC-2014) and GC mass spectrometry (Agilent, 6890N(GC), 5975B (inert XL El/Cl MSD)). For the reaction rate analysis, a reaction monitoring system was used (Fig. S3†), which was developed in our previous paper for studying the dye degradation processes.26 Some modications were made to measure the UV spectra of simple organic molecules. The light source was a deuterium lamp (D2), and a UV spectrometer (Ocean optics, USB4000) covering a wavelength range of 220–400 nm was used. The capillary used for the photocatalytic microreactor was made of fused silica with a diameter of 500 mm. The solution was injected into the capillary using a syringe pump. Aer the ow stopped, we waited for ca. 30 seconds until adsorption equilibrium was reached. The photocatalytic reaction was initiated by light irradiation from the top side. The light sources were UV-LEDs at 365 nm and 450 nm for UV photocatalysts and visible light photocatalysts, respectively. Perpendicular to the light, a ber tip connected to the D2 lamp was placed at the reaction location. The other end of the ber tip was placed on the opposite side of the D2 lamp, and the absorption spectrum was monitored with a spectrometer. The intensity of the D2 lamp was adjusted so as not to induce photoreactions or photocatalytic reactions. As a typical photocatalyst, TiO2 powder (P25, average diameter (D): 21 nm) was used. It was mixed with water (3.33 mL) and acetylacetone (0.33 mL) in a mortar for 30 min, to prepare a TiO2 paste. A tungsten oxide coating solution (Toshiba materials, SJ9003-1, D: ca. 100 nm), nitrogen doped titanium oxide (N–TiO2, D: 22 nm), N–TiO2 doped with silicon (N–Si–TiO2, D: 11 nm), which is effective for reducing the defects in TiO2, and N–Si–TiO2 deposited with vanadium (V–N–Si–TiO2, D: 13 nm),36,37 for which enhanced photocatalytic ability was reported, were used as visible light photocatalysts. Powder samples were dispersed in a solution using the same method as for the TiO2 powder to prepare a paste. The method for the paste preparation is a general procedure for coating a nanoparticulate inorganic material on a glass. For all materials used here, it was conrmed that the thickness of the coating was 3 0.5 microns, but the porosities ranged from 40 to 60%. We investigated the organic conversion of various molecules due to photocatalytic reactions, and we selected the following reactions as candidates to show the oxidation and reduction


View Article Online


Paper

Analyst

abilities. To select the reactions, we placed importance on the simplicity of the reactions and easy monitoring of the UV spectra of the reactants and products. The oxidation of benzyl alcohol (BAlc) to benzaldehyde (BAld)2,5 and the reduction of BAld to BAlc were selected, and the reaction schemes are shown in reactions (1) and (2) (Scheme 1). They are the reverse reactions of each other, and they can be controlled simply by changing the solvent: acetonitrile for oxidation and ethanol for reduction. For the oxidation, holes are used as an oxidant. Electrons are used for reduction since ethanol is a good scavenger of photoexcited holes, which prevents the recombination of electrons.14,38 Initially, we chose these reactions because the product analyses by GC and the monitoring of the spectra were easy, as we only had to monitor BAlc and BAld. However, we selected another reduction reaction, the reduction of nitrobenzene (NB) to aniline (AN) (reaction (3), Scheme 1),8,14,15 due to several reasons which we will explain in the following section.

Results and discussion 1. Product analysis In the case of a typical photocatalyst, TiO2, the GC charts for the oxidation of BAlc and the reductions of BAld and NB before and aer reactions are shown in Fig. 1. The initial concentrations of the reactant solutions were 1 mM. As shown in Fig. 1, all the peak intensities for the reactants decreased aer the reactions. For the oxidation reaction, the peak for BAld at 21 min was the only product peak observed, except for the solvent peak. For the reduction reaction, the peak for the corresponding reduced species, BAlc, was observed at 32 min. Several other peaks were observed, corresponding to acetaldehyde and acetic acid due to the oxidation of ethanol. Since we did not observe any peaks except for those of the products and the solvent, no by-products were observed, and we can predict that the other products would be carbon dioxide and water. In the case of the reduction

The GC charts before and after the photocatalytic reactions: (a) oxidation of benzyl alcohol, (b) reduction of benzaldehyde, (c) reduction of nitrobenzene. The blue dotted line and the red dashed line correspond to before and after the reaction, respectively. The reaction time was 18 h for (1) and (2), and 30 min for (3).

Fig. 1

of NB, the peaks for aniline (AN) at 7.8 min and nitrosobenzene at 6.5 min were observed, and the latter is an intermediate species. For this reaction, the light irradiation was only applied for 30 min because unknown reactions proceeded at longer irradiation times. The conversions, yields, and selectivities are summarized in Table 1. The conversion values can be used as a guide for estimating how efficiently the photoexcited electrons and holes

Table 1 Conversions, yields and selectivities for the oxidation and reduction reactions in the photocatalytic microreactor: (1) oxidation of benzyl alcohol, (2) reduction of benzaldehyde, (3) reduction of nitrobenzene. The reaction time was 18 h for (1) and (2), and 30 min for (3)

Scheme 1

Reaction candidates for the photocatalytic oxidation and

reduction.


Entry

Conversion (%)

Yield (%)

Selectivity (%)

(1) (2) (3)

24 69 6

8 23 5

33 33 77

Analyst, 2014, 139, 1953–1959 | 1955

View Article Online

Analyst

Paper

were utilized, namely the total oxidation ability. The selectivities and yields are indicators of how many electrons and holes were used for the intended reactions.


2. Reaction rate analysis Each reaction was monitored in situ using UV absorption spectroscopy. The changes in the spectra with time for the oxidation and reduction reactions are shown in Fig. 2. For the oxidation reaction, 5 mM BAlc/acetonitrile (ACN) was used as a reactant, and the spectrum of BAlc with a peak at 270 nm gradually decayed and changed into the spectrum of BAld. It was noted that the reaction was almost saturated aer about 60 s. From the reference spectra for BAld and BAlc in ACN at certain concentrations, we could t each spectrum obtained during the reaction to a combination of these reference spectra, and we could obtain the concentrations of BAld and BAlc at each time. (A typical tted spectrum is shown in Fig. S4.†) The concentration changes during the reactions are shown in Fig. 3. An increase in the concentration of BAld and a decrease in the concentration of BAlc were conrmed and the change was gradually saturated aer about 60 s. It was noted that the sum of the BAld and BAlc concentrations remained at 5 mM. This indicated that this reaction proceeded with 100% selectivity until 60 seconds aer the reaction started. We suppose that the reaction was saturated because the product species covered the active reaction sites, since the reactivity was recovered by replacement of the solution. Furthermore, the concentrations of BAld and BAlc showed linear changes with time, and this indicated that this reaction is zeroth order, which can be explained by the rate analysis of the surface reaction (see Appendix). Since the reaction became saturated as it proceeded, the straight parts of the concentration curves were used to analyse the data. We successfully obtained a reaction rate constant of 0.011 mol L1 s1. Similarly, the reduction of BAld was monitored as shown in Fig. 2(b). The reduction did not proceed until 60 s aer the photoirradiation, and the concentration of BAld suddenly decreased aer that. We tried tting the spectrum in a similar way as for the oxidation reaction, but it was very difficult to obtain the exact concentration of BAlc because the absorption intensity of BAld is much stronger than that of BAlc for the same concentration, therefore the contribution of BAlc to the spectrum was difficult to deduce, and the error was very large for the smaller concentration of BAlc, as shown in Fig. 3(b). (It is noted that the absorbances of BAld and BAlc in Fig. 2(a and b) were different because of the different solvents.) Since it is still unclear why there was a lag time for the start of the reaction and the accuracy of the BAlc concentration was not good enough, we selected another reaction for the reduction reaction. Then, we used the conversion of NB to AN, which is a wellknown reduction reaction that proceeds via photocatalysis. The peak at 270 nm, corresponding to NB, gradually decayed and a new peak at 250 nm appeared. The reaction was saturated aer approximately 100 s. However, the spectrum of the product did not perfectly match that of AN, which shows a peak at approximately 240 nm.

1956 | Analyst, 2014, 139, 1953–1959

Absorption spectra for the oxidation and reduction reactions observed directly in a photocatalytic microreactor. (a) corresponds to the oxidation of benzyl alcohol. The spectra obtained 0, 10, 20, 30, 40, 50, 60 and 80 s after the reaction was started by UV irradiation are shown. The dotted lines indicate the reference spectra for benzyl alcohol and benzaldehyde. (b) corresponds to the reduction of benzaldehyde. The spectra obtained 0, 80, 100, 120, 140, 160 and 175 s after the reaction started are shown. The dotted lines indicate the reference spectra for benzyl alcohol and benzaldehyde. It is noted that the absorbances are different in (a) and (b) because of the different solvents used. (c) corresponds to the reduction of nitrobenzene. The spectra obtained 0, 20, 40, 60, 80 and 100 s after the reaction started are shown. The dotted line indicates the reference spectrum for aniline.

Fig. 2


View Article Online

Paper

Analyst


However, we found out that AN did not react when 2-propanol was used as a solvent, probably because the oxidized molecule, acetone, did not react with AN. The absorption spectra for the reduction of NB in 2-propanol are shown in Fig. 2(c), and we could successfully observe the pure reduction reaction of NB to AN. The spectra obtained during the reaction were fully tted to combinations of the AN and NB spectra (Fig. S4(c)†). The concentration change during the reaction is shown in Fig. 3(c). We observed that the reaction started aer 10–20 s, which was much shorter than the corresponding time for the reduction of BAld, and the lag time was reproducible as long as TiO2 was used. Neglecting the lag time, the concentration change corresponded to a 1st order reaction, as shown in Fig. 3(c), and a reaction rate of 0.032 s1 was obtained. (It is noted that this value is the reaction rate multiplied by the adsorption equilibrium constant. See Appendix.)

3. Application of the screening methods for visible light photocatalysts

Changes in the reactant and product concentrations with time for the oxidation and reduction reactions in the photocatalytic microreactors: (a) the oxidation reaction of benzyl alcohol, (b) the reduction reaction of benzaldehyde, (c) the reduction reaction of nitrobenzene. Fitting curves for BAlc and BAld in (a) and for NB in (c) were obtained by analysis of the 0th order and 1st order kinetics, respectively. Fig. 3

We suspected that AN was subject to photoreaction, and thus the change in the spectrum of AN on irradiation with UV light was monitored. Actually, the peak in the spectrum shied from 240 to 250 nm and the resulting spectrum agreed well with the spectrum of the product of the NB reduction. The product was an imine compound, as conrmed by the GC-MS analysis (N-ethylideneaniline). Thus, it was considered that AN was generated in the reduction of NB but that the generated AN was changed into the imine compound by reacting with the acetaldehyde generated via the oxidation of ethanol.


An evaluation of several visible light photocatalysts using the oxidation and reduction reactions was performed. The following photocatalysts were analysed: WO3 and N–TiO2 as typical visible light photocatalysts, N–TiO2 doped with silicon (N–Si–TiO2), which is effective for reducing the defects in TiO2,36 and N–Si–TiO2 deposited with vanadium (V–N–Si–TiO2), for which an enhanced photocatalytic activity was reported.37 The results for the oxidation reaction of BAlc are shown in Table 2. The corresponding results for TiO2 under UV irradiation of similar intensity are shown too. Overall, the oxidation abilities of the visible light photocatalysts were comparable to that of TiO2 under UV illumination, considering differences in the light intensity and the absorbance of the materials, etc. WO3 showed the highest conversion, but the selectivity was quite low. With regard to the Si doping, N–Si–TiO2 showed higher conversion and selectivity compared with N–TiO2. Vanadium deposition gave a lower efficiency for the oxidation. The results for the reduction reaction of BAld are shown in Table 3. Compared with TiO2, the reduction abilities of the visible light photocatalysts were quite low on the whole, and no reduction product was observed for WO3, which can be explained by the low conduction band position. For the visible light photocatalysts, N–Si–TiO2 showed the highest conversion and selectivity, and vanadium deposition lowered the reaction efficiency.

Table 2 Conversions, yields and selectivities for the oxidation of benzyl alcohol in the photocatalytic microreactor for various photocatalytic materials

Entry

Conversion (%)

Yield (%)

Selectivity (%)

TiO2 with UV WO3 N–TiO2 N–Si–TiO2 V–N–Si–TiO2

17 25 7.5 14 8.8

3.7 0.3 3.7 8.1 5.6

22 1.0 49 56 56

Analyst, 2014, 139, 1953–1959 | 1957

View Article Online

Analyst

Paper


Table 3 Conversions, yields and selectivities for the reduction of benzaldehyde in the photocatalytic microreactor for various photocatalytic materials

Entry

Conversion (%)

Yield (%)

Selectivity (%)

TiO2 with UV WO3 N–TiO2 N–Si–TiO2 V–N–Si–TiO2

51 8.3 14 18 12

17 0 2.4 5.2 1.6

17 0 17 29 13

Table 4 Reaction rates for the BAlc oxidation measured by direct UV

monitoring in the photocatalytic microreactor

Entry

Conversion reaction rate (103 mol L1 s1)

WO3 N–TiO2 N–Si–TiO2 V–N–Si–TiO2

0.440 0.298 0.214 0.160

A reaction rate analysis was performed for the visible light photocatalysts. Since they showed very low reduction abilities, we studied only the BAlc oxidation reaction, which proceeded with almost 100% selectivity even for the visible light photocatalysts. The changes in the reactant and product concentrations with time are shown in Fig. S5 in the ESI.† The obtained reaction rates are listed in Table 4. From the product analysis and reaction rate analysis, the oxidation ability of WO3 was high, although the reaction was not selective. There was a discrepancy between the product analysis and reaction rate analysis for the oxidation abilities of N–TiO2 and N–Si–TiO2. This is probably because N–TiO2 lost its reactivity during the long reaction time, while the initial reaction rate was large during the reaction rate analysis carried out within a few minutes. Basically, it can be said that Si doping in TiO2 is effective for enhancing both the oxidation and reduction reactivities due to the decrease in the internal defects. But we could not conrm the enhancement of the reaction efficiency by vanadium deposition, because both the oxidation and reduction reactions were deactivated. This is probably because the deposition would reduce the number of active reaction sites.

Conclusion We developed two analytical methods for the evaluation of photocatalytic materials using a photocatalytic microreactor; one method is product analysis and the other is reaction rate analysis. To perform a simple analysis of the photocatalytic abilities, we selected oxidation and reduction reactions of organic compounds, which are easily monitored by GC and absorption spectra due to the one-to-one reactions. We could establish the methods of product analysis and reaction rate analysis, and they were applied to several visible light photocatalysts. It is expected that this novel analytical procedure could accelerate the development of new materials which are very useful for various purposes.

1958 | Analyst, 2014, 139, 1953–1959

Appendix Whether the reaction is zeroth or rst order is determined by analysis of the rate equation for the surface reaction. We consider the following reactions at the surface: kA

! A þ s As 0 kA

kr

As ! Cs

(4)

(5)

where A and C are the reactant and product, respectively, and s is the active reaction site. The reaction rates for (4) and (5) are written as follows: r1 ¼ kA[A]qV kA0 qA

(6)

r 2 ¼ kr q A

(7)

where qV and qA are the vacancy ratio and the ratio occupied by A species on the surface, respectively, and [A] is the concentration of A in the liquid phase. It is assumed that the reaction rate of (6) is fast under equilibrium, and KA ¼ kA/kA0 is dened as an adsorption equilibrium constant. The reaction rate for the whole reaction is assumed to be r2, and it is written as follows: r2 ¼

kr KA ½A : 1 þ KA ½A

(8)

When KA[A] 1, r2 is approximated as krKA[A], and when KA[A] [ 1, it is approximated as kr. The latter condition was satised for the oxidation of BAlc, and the former was satised for the reduction of NB.

Notes and references 1 N. Serpone, E. Pelizzetti and H. Hidaka, Photocatalytic Purication and Treatment of Water and Air, Elsevier, London, 1993. 2 C. J. Li, G. R. Xu, B. H. Zhang and J. R. Gong, Appl. Catal., B, 2012, 115–116, 201–208. 3 P. Ciambelli, D. Sannino, V. Palma and V. Vaiano, Catal. Today, 2005, 99, 143–149. 4 K. Imamura, H. Tsukahara, K. Hamamichi, N. Seto, K. Hashimoto and H. Kominami, Appl. Catal., A, 2013, 450, 28–33. 5 S. Yurdakal and V. Augugliaro, RSC Adv., 2012, 2, 8375–8380. 6 J. T. Carneiro, A. R. Almeida, J. A. Moulijn and G. Mul, Phys. Chem. Chem. Phys., 2010, 12, 2744–2750. 7 O. S. Mohamed, A. M. Gaber and A. A. Abdel-Wahab, J. Photochem. Photobiol., A, 2002, 148, 205–210. 8 K. Imamura, T. Yoshikawa, K. Hashimoto and H. Kominami, Appl. Catal., B, 2013, 134–135, 193–197. 9 A. Maldotti, L. Andreotti, A. Molinari, S. Tollari, A. Penoni and S. Cenini, J. Photochem. Photobiol., A, 2000, 133, 129–133. 10 V. Brezov´ a, A. Blaˇzkov´ a, I. ˇ Surina and B. Havl´ınov´ a, J. Photochem. Photobiol., A, 1997, 107, 233–237. 11 Y. Shiraishi and T. Hirai, J. Photochem. Photobiol., C, 2008, 9, 157–170.


View Article Online


Paper

12 H. Yoshida, Photocatalytic Organic Synthesis, Springer, New York, 2010. 13 V. Augugliaro, T. Caronna, A. D. Paola, G. Marci, M. Pagliaro, G. Palmisano and L. Palmisano, TiO2-Based Photocatalysis for Organic Synthesis, Springer, New York, 2010. 14 S. O. Flores, O. Rios-Bernij, M. A. Valenzuela, I. Cordova, R. Gomez and R. Gutierrez, Top. Catal., 2007, 44, 507–511. 15 J. L. Ferry and W. H. Glaze, Langmuir, 1998, 14, 3551– 3555. 16 O. Rios-Berny, S. O. Flores, I. Cordova and M. A. Valenzuela, Tetrahedron Lett., 2010, 51, 2730–2733. 17 K. Selvam and M. Swaminathan, Tetrahedron Lett., 2010, 51, 4911–4914. 18 K. Selvam and M. Swaminathan, Catal. Commun., 2011, 12, 389–393. 19 R. Abe, J. Photochem. Photobiol., C, 2010, 11, 179–209. 20 S. I. Allakhverdiev, V. Thavasi, V. D. Kreslavski, S. K. Zharmukhamedov, V. V. Klimov, S. Ramakrishna, D. A. Los, M. Mimuro, H. Nishihara and R. Carpentier, J. Photochem. Photobiol., C, 2010, 11, 101–113. 21 K. Maeda, J. Photochem. Photobiol., C, 2011, 12, 237–268. 22 Y. Han, H. S. Kim and H. Kim, J. Nanomater., 2012, 2012, 1– 10. 23 S. I. In and R. Berg, Asian J. Chem., 2012, 24, 428–432. 24 Z. Liu, X. Zhang, S. Nishimoto, T. Murakami and A. Fujishima, Environ. Sci. Technol., 2008, 42, 8547–8551. 25 M. Mori, K. Tanaka, H. Taoda, M. Ikedo and H. Itabashi, Talanta, 2006, 70, 169–173.


Analyst

26 N. Tsuchiya, K. Kuwabara, A. Hidaka, K. Oda and K. Katayama, Phys. Chem. Chem. Phys., 2012, 14, 4734–4741. 27 N. Wang, Y. P. Zhang, L. Lei, H. L. W. Chan and X. M. Zhang, in Nems/Mems Technology and Devices, 2011, vol. 254, pp. 219–222. 28 Y. Matsushita, N. Ohba, S. Kumada, T. Suzuki and T. Ichimura, Catal. Commun., 2007, 8, 2194–2197. 29 Y. Matsushita, M. Iwasawa, T. Suzuki and T. Ichimura, Chem. Lett., 2009, 38, 846–847. 30 Y. Matsushita, S. Kumada, K. Wakabayashi, K. Sakeda and T. Ichimura, Chem. Lett., 2006, 35, 410–411. 31 R. Gorges, S. Meyer and G. Keisel, J. Photochem. Photobiol., A, 2004, 167, 95–99. 32 E. E. Coyle and M. Oelgemoller, Photochem. Photobiol. Sci., 2008, 7, 1313–1322. 33 N. M. Mahmoodi, M. Arami, N. Y. Limaee and N. S. Tabrizi, Chem. Eng. J., 2005, 112, 191–196. 34 I. K. Konstantinou and T. A. Albanis, Appl. Catal., B, 2004, 49, 1–14. 35 K. Katayama, Y. Takeda, K. Kuwabara and S. Kuwahara, Chem. Commun., 2012, 48, 7368–7370. 36 A. Mase, T. Sugita, M. Mori, S. Iwamoto, T. Tokutome, K. Katayama and H. Itabashi, Chem. Eng. J., 2013, 225, 440–446. 37 A. Maldotti and A. Molinari, in Photocatalysis, 2011, vol. 303, pp. 185–216. 38 A. I. Kryukov, S. Y. Kuchmy, S. V. Kulik and A. V. Korzhak, Theor. Exp. Chem., 1992, 28, 416–419.

Analyst, 2014, 139, 1953–1959 | 1959

View Article Online

Analyst


PAPER


Ion dynamics in a trapped ion mobility spectrometer† Diana Rosa Hernandez,a John Daniel DeBord,a Mark E. Ridgeway,b Desmond A. Kaplan,b Melvin A. Parkb and Francisco Fernandez-Lima*a In the present paper, theoretical simulations and experimental observations are used to describe the ion dynamics in a trapped ion mobility spectrometer. In particular, the ion motion, ion transmission and mobility separation are discussed as a function of the bath gas velocity, radial confinement, analysis time and speed. Mobility analysis and calibration procedure are reported for the case of sphere-like molecules

Received 22nd November 2013 Accepted 17th January 2014

for positive and negative ion modes. Results showed that a maximal mobility resolution can be achieved by optimizing the gas velocity, radial confinement (RF amplitude) and ramp speed (voltage range and

DOI: 10.1039/c3an02174b

ramp time). The mobility resolution scales with the electric field and gas velocity and R ¼ 100–250 can

www.rsc.org/analyst

be routinely obtained at room temperature.

Introduction Ion Mobility Spectrometry (IMS) has been widely used in the detection of chemical warfare agents, illicit drugs, explosives, and more recently in the separation and detection of biomolecules.1–9 Developed in the late 1970s, IMS permits the separation of ions in the gas-phase by their size-to-charge ratio (U/z) in the presence of a neutral gas. Initially known as plasma chromatography or ion chromatography, IMS has been coupled to mass spectrometry (MS) to obtain fast, complementary separations (i.e., U/z and m/z).10–13 In particular, IMS-MS offers several advantages over traditional MS, including separation of ions from mixtures based on the composition and charge state, the ability to separate geometric isomers, increase the dynamic range and discrimination against chemical noise.14–19 A variety of forms and designs of IMS analysers have been described in the literature.20 A leading force behind these efforts has been the drive for higher mobility separation and higher ion transmission. For example, variants of the basic, high-pressure dri tube incorporated entrance and exit ion funnels to increase ion transmission, thus turning the basic dri tube into a powerful analyzer.21 In a different approach, ion transmission has also been increased by using periodic-focusing DC ion guides.22,23 Other techniques have been used to add a degree of separation prior to MS analyses (e.g., eld asymmetric IMS,24 differential mobility spectrometer,3,4,25 segmented quadrupole a

Department of Chemistry and Biochemistry, Florida International University, Miami, USA. E-mail: fernandf@u.edu; Fax: +1 305 348 3772; Tel: +1 305 348 2037

b

Bruker Daltonics, Inc., Billerica, Massachusetts 01821, USA

† Electronic supplementary information (ESI) available: Detailed information on the structure of the sphere-like molecules considered here. In addition, ion trajectory simulations and mobility resolution dependence on the gas velocity are shown. See DOI: 10.1039/c3an02174b


dri cell,26 cylindrical dri tubes,27 and transient wave ion guide28). Several research groups have focused on achieving high resolution IMS separation (R > 50) as this factor mainly limits the information that can be experimentally derived.29–32 We have recently introduced a Trapped Ion Mobility Spectrometer (TIMS) that can be easily integrated into a mass spectrometer (MS) for IMS-MS analyses and which is capable of producing high resolution IMS separations.33,34 The most signicant breakthrough of TIMS technology lies in the fact that mobility separation can be tuned from low to high according to the analytical challenge. However, a detailed understanding of the ion dynamics in a TIMS cell, modes of operation, performance and limitations is necessary for the development of analytical applications. In the present paper, a detailed description of the ion dynamics inside a Trapped Ion Mobility Spectrometer (TIMS) is discussed. In particular, the inuence of the electrode geometry, bath gas ow, radial trapping potential, and ramp speed on the mobility resolution is modeled theoretically and compared with experimental values. The methodology to determine an accurate collision cross-section from TIMS measurements is also provided.

Experimental TIMS analyzer A TIMS cell consists of three main sections: entrance funnel, analyzer section and exit funnel (shown in the cross-section in Fig. 1). Ions are typically generated using an Electrospray Ion Source (Apollo II design, Bruker Daltonics Inc., MA) and introduced into the TIMS device via a glass ion transfer capillary. The TIMS sections are constructed from a set of segmented ring electrodes supported on PC boards (plate thickness of 1.6 mm).

Analyst, 2014, 139, 1913–1921 | 1913

View Article Online


Analyst

Each electrode is composed of four electrically isolated segments (two shown in the cross-section in Fig. 1 for each PC board). The basic design of the electrodes is the same throughout the three sections – only the inner diameter is varied from 26 to 8 mm and from 8 to 1 mm in the entrance and exit funnels, while kept constant at 8 mm in the analyzer section. In the entrance and exit funnel sections, the plates are spaced 1.5 mm from each other. However, in the analyzer section, plates are separated by an insulating gasket material (kapton, 0.125 mm thickness) which forms a gas tight seal. The length of the entrance funnel, analyzer section and exit funnel are 50, 46 and 15 mm, respectively. The segmented plate design has the advantage that it can be used to form a dipolar eld, as is typical in ion funnels, or a quadrupolar eld.35 In the entrance and exit funnel sections, the RF potential applied to the ion funnel is 180 out of phase between adjacent plates. This results in a pseudo-potential, which simply keeps the ions away from the funnel walls. The same RF waveform is used throughout the TIMS funnel. However, in the analyzer section, the phase of the RF potential does not alternate between adjacent plates but only between adjacent segments. Importantly, the purpose of the quadrupolar eld in the analyzer section is to conne (trap) the ions radially and avoid ion losses due to diffusion. The gas ow velocity in the analyzer region is regulated by the pressure difference between the entrance funnel (P1) and exit funnel (P2) regions. In practice, the pressure difference is regulated by changing the pumping impedance using a buttery valve. The operating pressure difference (few mbar) produces a cylindrically symmetric gas ow. Mobility separation in a TIMS device can be performed using different gases or mixtures in accordance with the analytical problem to be solved. In the examples presented herein, the separation was performed using nitrogen as the bath gas at ca. 300 K with typical entrance funnel and exit funnel pressures of P1 ¼ 1.0–2.6 and P2 ¼ 1.0 mbar, respectively.

TIMS-MS operation sequence The TIMS funnel can be operated in “transmission” mode or “IMS” mode. In transmission mode, the DC potentials on all the

Fig. 1 Cross-sectional view of the TIMS entrance funnel, analyser section and exit funnel.

1914 | Analyst, 2014, 139, 1913–1921

Paper

Time sequence of a mobility analysis in a TIMS analyser for positive ion mode. Notice that fill and trap times can be adjusted for the maximum mobility resolution and sensitivity and for kinetic measurements, respectively.

Fig. 2

funnel elements are set to continually push ions downstream – i.e. without a mobility separation. In practice, the analyzer entrance potential is set higher – i.e. more repulsive to the ions – than that of the analyzer exit. Similarly, the funnel entrance and deection plate are set to successively higher potentials when operating in transmission mode. In transmission mode the instrument produces conventional mass spectra. In IMS mode, the sequence begins with ion loading of the analyzer section (Fig. 2). The DC potentials on the deection plate, funnel entrance, and analyzer entrance are set to push ions orthogonally out of the gas stream from the capillary, through the entrance funnel, and into the analyzer section. However, the DC potential on the analyzer entrance is set below that of the exit so as to produce a retarding eld. The gas ow – from le to right in Fig. 1 – pushes the ions downstream with a “force” in accordance with the ions' mobility. Ions within a given mobility range thus become trapped by the DC electric eld which prevents the ions from progressing downstream, while the gas ow prevents the ions from returning upstream, and the RF quadrupolar pseudo-potential which prevents the ions from escaping radially. In practice, the potential proles are produced by a network of precision resistors operating as a voltage divider. Thus, to produce the given potential proles, voltages are applied to the deection plate, capillary exit, funnel entrance, analyzer entrance, and analyzer exit. With these few inputs to the resistor network, the complete potential prole is dened. At all times, the axial electric eld is kept under the low eld limit (E/p < 10 V cm1 torr1) throughout the TIMS device.36,37 The lling time can be tuned according to the number of ions (ESI beam current) coming from the capillary, with a typical lling time of about 10 ms. Ions trapped in the analyzer section will assume positions (or electrode number) depending on their mobilities. The DC electric eld strength, Ex, is


View Article Online


Paper

dependent on the position – highest eld strength nearer the exit end. Thus, the lowest mobility ions come to a “stop” in the highest eld strength region (nearer the exit end of the analyzer section) whereas the highest mobility ions assume positions in the lowest eld strength region (nearer the analyzer entrance). Following the trapping period, additional ions are prevented from entering the analyzer section by lowering the potential on the deection plate – i.e. by making the deection plate potential attractive. When the deection plate potential is attractive, ions exiting the capillary are accelerated towards the deection plate. Trapped ions in the analyzer section, however, are unperturbed by the deection plate potential variation. Mobility analysis of the trapped ions consists of reducing the strength of the retarding electric eld in the analyzer section stepwise over time. To reduce the retarding eld strength in the analyzer section, the potential at the analyzer entrance is increased at a constant rate – i.e. “scanned” – from its initial value to an end value – in this case, the analyzer exit potential. As the eld strength is reduced, the trapped ions shi to new positions in the analyzer – with ions of a given mobility always residing at a position having a given eld strength. As the retarding eld strength is reduced, ions of successively higher mobilities are pushed by the gas ow out of the analyzer region. Ions exiting the analyzer section are focused by the exit funnel towards the entrance ion optics of a mass spectrometer. In particular, for the examples shown here, a maXis Impact Q-UHR-ToF (Bruker Daltonics Inc., MA) was coupled to the TIMS analyzer. The measured ion current plotted over the course of the voltage ramp (i.e., ramp speed ¼ voltage/ramp time) provides the ion mobility spectrum from low ion mobilities to high ion mobilities. TIMS time sequences and voltages were controlled using the in-house soware, written in National Instruments Lab VIEW, and synchronized with the MS acquisition program. In particular, voltages were dened using National Instruments cards (NI-PXI-6704, NI-PXI-6289, NI-PXI-6361) and a 1 GHz digitizer (NI-PXI-5154) was used for the MS data acquisition. All cards were housed in a 1073e PXI chassis, operated by a computer PXIe extension bus. The TIMS-MS experiment was synchronized with the TOF start (extraction trigger). The digitizer trigger was internally routed through the PXI chassis and sent to the PXI6361. A digital pattern was then output at a rate of 10 MHz, which was triggered by the routed digitizer trigger. The digital pattern, dened by the user, consisted of three TTL triggers which were used for: (1) trapping ions in the analyzer section, (2) blocking ions from entering the analyzer section, and (3) signaling the start of the voltage ramp in the analyzer section for ion elution. The ramp voltage was controlled by the PXI-6289 analog output (AO). The ramp voltage range (Vramp) and time (Tramp) were dened in the soware and built as an array of output voltages. The output voltage of the PXI-6289 ( 10 V AO) was amplied using an in-house 15 amplier capable of 50 kHz ramp rates. The internally routed trigger was also used as the output clock of the PXI-6289, which results in the TIMS voltage steps synchronized with the TOF pulses. MS data were acquired from the digitizer for each ramp step (i.e., 1 record set at a time). Aer a full set of TIMS records were acquired from


Analyst

the digitizer, all of the PXI cards were reset and the next TOF trigger restarted the process again. This acquisition process (i.e., re-initializing the cards) was required due to the slow data transfer rate from the digitizer to the computer and to ensure proper timing synchronization during the entire TIMS-MS experiment. Materials and reagents Molecules used in this study were purchased from Sigma (St. Louis, MO) and used as received. Samples were electro-sprayed at a concentration of 1–10 mmolar in a 1 : 1 (v/v) water–methanol solution. An ESI Tuning Mix calibration standard (Tunemix, G2421A) was purchased from Agilent Technologies (Santa Clara, CA) and used as received.38 Details of the ESI tuning mix calibration structures can be found in the ESI.†

Theoretical Ion and gas dynamics in a TIMS cell To better understand the effect of the bath gas ow on the TIMS performance and operation, the pressure gradient in the TIMS cell was modelled (Fig. 3). A cylindrical geometry was used and the mass ow and energy equations were solved simultaneously for the case of nitrogen as the bath gas and using the P1 and P2 pressure values as boundary conditions for the calculations. The Fluent 6.2.16 (Fluent Inc., Lebanon, NH) soware was used.39 Inspection of Fig. 3 shows that the velocity prole is near homogenous in the central region (tens of m s1 for the pressure studied) and that the gas velocity decreases in magnitude near the wall of the analyser region. The end result is a parabolic prole as a consequence of the friction of the gas molecules near the walls. This parabolic ow prole determines the trapping conditions during TIMS operation; that is, in a TIMS device the basis for the mobility separation lies in compensation of the drag force (proportional to the gas velocity) with the electric force (increasing in magnitude from le to right with the electrode positions). Inhomogeneities in the gas prole could also inuence the ion transfer in funnel 2 but at this stage the gas velocity is much lower and has a smaller effect. To have a better understanding of the correlation between the gas velocity and the ion trajectories, the ion dynamics inside the TIMS cell were studied using an in-house routine incorporated into a Simion 3D (ref. 40) user developed program. In the ion simulations, an Elastic Hard Sphere Scattering model (EHSS) for the ion-neutral collisions was used.41 The bath gas temperature and velocity, electric eld gradient, RF amplitude and frequency, and collision cross-sections were used as simulation input parameters. Two cases were compared with the experimental observation in order to evaluate the bath gas velocity parabolic prole (see Fig. 4): one dimensional (vx) and three dimensional proles (vxyz parabolic). At low RF amplitude values, ions are poorly conned and diffuse toward the TIMS analyser walls and are neutralized. However, as the RF amplitude increases, ions are conned

Analyst, 2014, 139, 1913–1921 | 1915

View Article Online


Analyst

Paper

Fig. 4 Comparison of the experimental and theoretical relative ion abundance (m/z ¼ 622) as a function RF amplitude. Theoretical simulations were performed using one (vx ¼ 70 m s1) and three (vxyz, parabolic) dimensional gas profiles for the same conditions as shown in Fig. 3.

Fig. 3 Total pressure, gas velocity and velocity vector profiles for nitrogen in the TIMS analyser section. P1 and P2 were used as boundary conditions (P1 ¼ 2.6 and P2 ¼ 1.0 mbar). Scale: red: high and blue low.

towards the center of the TIMS analyser and higher ion abundances are observed. Comparison between the simulations shows that a better agreement is obtained when a three dimensional prole is used compared to a one-dimensional prole. In particular, this effect is more pronounced at lower RF amplitudes, where the ion cloud is more spread in the radial direction and a larger gas velocity gradient is experienced by the ions trapped in the cylindrical TIMS analyzer section. Similar conclusions can be obtained from the inspection of the ion clouds as a function of the RF amplitude in the TIMS analyzer section (Fig. 5). When no RF is applied, ions are conned in the axial direction by the electric and drag forces. In the radial direction, ions will diffuse over time and the ion population will decrease as they get neutralized by collisions with the walls. With the increase of the radial force (or RF

1916 | Analyst, 2014, 139, 1913–1921

amplitude), the ion cloud is pushed towards the center and ion connement occurs. The size of the ion cloud is dened by the radial and axial connement and by the number of ions (charges) that are conned. If the radial connement is too high, ions will spread in the axial direction to compensate for the coulombic repulsion of the ion cloud counterparts. That is, the higher the number of ions and the higher the RF amplitude the larger the axial spread of the ion cloud. Similar effects related to the number of ions trapped have been observed in other trapping devices (e.g., Paul trap, linear ion trap, ICR cell, etc.).42–46 In TIMS, the presence of the bath gas (higher pressures compared to other trapping devices, few mbar) induces larger changes on the ion trajectories by the ion-neutral collisions, ion-neutral collision being the dening parameter of the ion motion. Analogous to dri tube experiments, reduction of the bath gas temperature will reduce the ion diffusion and increase the mobility resolution (see the example in the ESI† for bath gas at ca. 80 K). Nevertheless, there are practical challenges associated with the use of lower bath gas temperatures due to the high gas ow used in the TIMS analyzer. TIMS separation The theoretical concept behind TIMS is the use of an electric eld to hold ions in place against a moving gas. The TIMS operation is analogous to that of parallel ow ion mobility analyzers,47,48 with the main difference that ions are also conned radially to guarantee higher transmission and sensitivity. In particular, the separation in a TIMS device can be described in the center of mass frame using the same principles as in a conventional dri tube.49,50 That is, in conventional dri cells ions are pushed through a stationary gas whereas in the TIMS analyzer ions are held in place against a moving gas. Since mobility separation is related to the number of ion-neutral collisions (or dri time in traditional cells), the mobility separation in a TIMS device depends on the bath gas dri velocity,


View Article Online

Paper

Analyst


where A is a calibration constant that can be experimentally determined using known standards, either internally or externally. That is, the elution voltage is a characteristic parameter of the ion mobility for the same bath gas and velocity and can be calculated from the total IMS time using eqn (2) (see Fig. 6). The elution voltage is independent of the voltage ramp characteristics; however, the width of the elution voltage peak changes with the strength of the radial connement (this is ultimately related to the mobility resolution as discussed later). For example, with the increase of the radial eld (RF amplitude) an increase in the elution voltage FWHM is observed. We attribute this increase to the spread of the ion cloud in the axial dimension, in good agreement with the simulations shown in Fig. 5. Noticed that the axial spread is limited by the width of the electric eld (DE) necessary to trap the ion cloud; that is, the elution voltage has a limited peak width. This trend can be observed for the case of m/z ¼ 622 in Fig. 6. Different ion clouds (e.g., mass, charge and size) will respond differently to the RF radial connement. That is, since larger molecules (assuming +1 charge for simplicity) will require a higher RF amplitude to be trapped, the axial spread will be observed at larger RF amplitudes. If the radial spread is large (close to DE limit), the ion-neutral collision (or diffusion) will lead to a decrease in the ion population. At the same time, this condition will impose a higher limit on the elution voltage FWHM, and consequently a lower limit in the mobility resolution.

Fig. 5 Ion trajectories as a function the RF amplitude theoretical relative on abundance (m/z ¼ 622) as a function RF in the TIMS analyser. Simulations were performed at the same conditions as Fig. 3 and 4 for m/z ¼ 322 and using an RF frequency of 890 kHz.

ion connement and ion elution parameters. For example, the mobility range (Ka–Kb) of an analysis is directly related to the ratio of the gas velocity (vg) and the axial electric eld range (Ea–Eb). The mobility of a given ion (i) can be extracted from the electric eld (Ex(i)) at which the ion package elutes. In practice, this can be related to the elution voltage (Velut(i)) at which this process occurs relative to the voltage applied to the last electrode (Vout) of the mobility section: Ki ¼ vg/Ex(i) ¼ A(1/(Vout Velut(i)) This journal is © The Royal Society of Chemistry 2014

(1)

(Top) Elution voltage (Velut) dependence on the fill/ramp time (20/150–20/450 ms) for m/z ¼ 622, 922 and 1222 at RF 250 Vpp. (Bottom) Dependence of the elution voltage peak width (FWHM) as a function of RF amplitude. Fig. 6

Analyst, 2014, 139, 1913–1921 | 1917

View Article Online


Analyst

Paper

(see eqn (1) and (2)). The mobility and collision cross-section values presented in Table 1 were obtained by calculating the calibration constant from mobility measurements performed in traditional dri tube IMS using the same nitrogen bath gas and temperature (ca. 300 K). For the case of negative ion values reported in Table 1, the same calibration constant A was considered. Notice that the procedure described here provides accurate mobility measurements from rst principles. Mobility values (K) were correlated with CCS (U) using the equation: 1=2 ð18pÞ1=2 z 1 1 1 760 T 1 U¼ (3) þ K P 273:15 N* 16 ðkB TÞ1=2 mI mb

Fig. 7 Dependence of the total IMS time on the ramp time as a function of m/z.

where z is the charge of the ion, kB is the Boltzmann constant, N* is the number density and mI and mb refer to the masses of the molecular ion and bath gas molecule, respectively. Mobility resolution

Mobility calibration

The mobility resolution in a TIMS device is dened by:

In a TIMS device, the total analysis time can be described as: Total IMS time ¼ Ttrap + (Velut/Vramp)Tramp + ToF ¼ To + (Velut/Vramp)Tramp

(2)

where, Ttrap is the thermalization/trapping time, ToF is the time aer the mobility separation, and Vramp and Tramp are the voltage range and time required to vary the electric eld, respectively. The elution voltage can be experimentally determined by varying the ramp time for a constant ramp voltage. This procedure also determines the time ions spend outside the separation region To (e.g., ion trapping and time-of-ight). Experimental results show the linear dependence of the total IMS time with the ramp time for all m/z's studied (see the example in Fig. 7). From the slope and the intercept of this graph the mobility values can be determined if the calibration constant A is known

R ¼ K/DK ¼ (Vout Velut)/DVelut

(4)

In a traditional dri tube, the resolution is dened by the device length, bath gas pressure and temperature and the applied electric eld.51,52 In a TIMS device, different mobilities are trapped at different electric eld strengths for the same velocity of the gas, and hence different mobility resolutions will be observed. In addition, as noticed in previous results, the radial connement may change the elution voltage FWHM (axial spread of the ion cloud). In a TIMS device, as a general rule the resolution mainly depends on three parameters: (a) the bath gas velocity, (b) the electric eld ramp speed (DVramp/Tramp) and (c) the RF connement. In the case of the bath gas velocity, an increase in the bath gas velocity requires an increase in the electric eld to

Mobility values of the ESI Tune Mix calibration standard. Negative values were obtained using the calibration constant A obtained from positive ion values

Table 1

Monoisotopic mass

Formula

Ko [cm2 V1 s1]

˚ 2] CCS [A

118.086255 322.048121 622.028960 922.009798 1221.990637 1521.971475 1821.952313 2121.933152 2421.913992 2721.894829 301.998139 601.978977 1033.988109 1333.968947 1633.949786 1933.930624 2233.911463 2533.892301 2833.873139

C2H3NO2(CH3)3[M]+1 P3N3(OCH3)6[M + H]+1 P3N3(OCH3CF2)6[M + H]+1 P3N3(OCH3CF2CF2)6[M + H]+1 P3N3(OCH3C2F4CF2)6[M + H]+1 P3N3(OCH3C3F6CF2)6[M + H]+1 P3N3(OCH3C4F8CF2)6[M + H]+1 P3N3(OCH3C5F10CF2)6[M + H]+1 P3N3(OCH3C6F12CF2)6[M + H]+1 P3N3(OCH3C7F14CF2)6[M + H]+1 C3N3(CF3)3[M]1 C3N3((CF2)2CF3)3[M]1 P3N3(OCH3CF2CF2)6[M + C2F3O2]1 P3N3(OCH3C2F4CF2)6[M + C2F3O2]1 P3N3(OCH3C3F6CF2)6[M + C2F3O2]1 P3N3(OCH3C4F8CF2)6[M + C2F3O2]1 P3N3(OCH3C5F10CF2)6[M + C2F3O2]1 P3N3(OCH3C6F12CF2)6[M + C2F3O2]1 P3N3(OCH3C7F14CF2)6[M + C2F3O2]1

1.920 1.376 1.013 0.835 0.740 0.638 0.590 0.520 0.469 0.444 1.909 1.187 0.776 0.710 0.609 0.573 0.498 0.465 0.399

116.2 151.9 202.4 243.8 274.1 317.2 342.5 388.3 429.5 447.4 109.6 172.8 261.9 285.4 332.1 352.5 405.2 433.6 505.1

1918 | Analyst, 2014, 139, 1913–1921


View Article Online


Paper

Analyst

velocity and trapping conditions. Simulations showed the effect of the radial connement on the ion cloud, which ultimately translates into the mobility resolution of the analyser. It was shown that the main parameters that dene the ion motion in a TIMS analyser are the dri gas velocity, the ion connement and the electric eld ramp speed. Comparison of theoretical and experimental values results that the parabolic gas velocity prole can increase the axial spread of the trapped ions; however, the radial connement counteracts the axial spread by conning the ions to the central region of the TIMS analyzer where a more homogenous gas speed is obtained. Experimental results showed that mobility values can be extracted using rst principles, analogous to dri tube designs; the main difference is that internal or external calibration standards are needed for accurate measurements of mobility values for each experimental condition (e.g., velocity of the gas and voltage ramp). Results are shown for high mobility resolution (R ¼ 100–250) over a wide mobility range.

Notes and references

Fig. 8 (Top) Mobility resolution as a function of the ramp time for the sphere-like ESI Tuning Mix standards for P1 ¼ 1.9 and 2.6 mbar and P2 ¼ 1.0 mbar, respectively. (Bottom) IMS profile for m/z ¼ 2722 with RF ¼ 300 Vpp, Tramp ¼ 450 ms and P1 ¼ 2.6 and P2 ¼ 1.0 mbar.

counteract the increased drag force of the ions. The latter results in an increase in the mobility resolution (see Fig. 8). In the case of the ramp speed, the narrower the DVramp and the larger the Tramp the higher the mobility resolution; that is, the slower the scan speed the higher the mobility resolution for a given bath gas velocity (see Fig. 8). For the conditions explored in the present work, mobility resolution values of 100–250 were routinely obtained. It should be noted that the ramp time can be increased for up to a few seconds in a TIMS device with small to no ion loss (current limitations are on the electronic acquisition of long IMS transients). For the case of RF connement, as discussed before, the lower the RF needed to conne the ions the smaller the axial spread for the same number of ions in the ion cloud (see the example in the ESI†). It should be noted that the number of ions can also be varied using the source conditions and the duration of the ll injection pulse, thus allowing for the highest resolution possible.

Conclusions Ion dynamics simulations showed the ion dynamics inside of a trapped ion mobility spectrometer as a function of the bath gas


1 P. Dwivedi, P. Wu, S. Klopsch, G. Puzon, L. Xun and H. Hill, Metabolic proling by ion mobility mass spectrometry (IMMS), Metabolomics, 2008, 4, 63–80. 2 R. Fernandez-Maestre, C. S. Harden, R. G. Ewing, C. L. Crawford and H. H. Hill, Chemical standards in ion mobility spectrometry, Analyst, 2010, 135, 1433–1442. 3 G. A. Eiceman, E. V. Krylov, N. S. Krylova, E. G. Nazarov and R. A. Miller, Separation of Ions from Explosives in Differential Mobility Spectrometry by Vapor-Modied Dri Gas, Anal. Chem., 2004, 76, 4937–4944. 4 E. V. Krylov, S. L. Coy, J. Vandermey, B. B. Schneider, T. R. Covey and E. G. Nazarov, Selection and generation of waveforms for differential mobility spectrometry, Rev. Sci. Instrum., 2010, 81, 024101. 5 W. Cheung, Y. Xu, C. L. Thomas and R. Goodacre, Discrimination of bacteria using pyrolysis-gas chromatographydifferential mobility spectrometry (Py-GC-DMS) and chemometrics, Analyst, 2009, 134, 557–563. 6 T. Koimtzis, N. J. Goddard, I. Wilson and C. L. Thomas, Assessment of the feasibility of the use of conductive polymers in the fabrication of ion mobility spectrometers, Anal. Chem., 2011, 83, 2613–2621. 7 V. Moll, V. Bocos-Bintintan, I. A. Ratiu, D. Ruszkiewicz and C. L. Thomas, Control of dopants/modiers in differential mobility spectrometry using a piezoelectric injector, Analyst, 2012, 137, 1458–1465. 8 W. Vautz, R. Slodzynski, C. Hariharan, L. Seifert, J. Nolte, R. Fobbe, S. Sielemann, B. C. Lao, R. Huo, C. L. Thomas and L. Hildebrand, Detection of metabolites of trapped humans using ion mobility spectrometry coupled with gas chromatography, Anal. Chem., 2013, 85, 2135–2142. 9 S. Prasad, H. Schmidt, P. Lampen, M. Wang, R. Guth, J. V. Rao, G. B. Smith and G. A. Eiceman, Analysis of bacterial strains with pyrolysis-gas chromatography/ differential mobility spectrometry, Analyst, 2006, 131, 1216–1225.

Analyst, 2014, 139, 1913–1921 | 1919

View Article Online


Analyst

10 P. R. Kemper and M. T. Bowers, A hybrid double-focusing mass spectrometer-high-pressure dri reaction cell to study thermal energy reactions of mass-selected ions, J. Am. Soc. Mass Spectrom., 1990, 1, 197–207. 11 C. Wu, W. F. Siems, G. R. Asbury and H. H. Hill, Electrospray Ionization High-Resolution Ion Mobility Spectrometry-Mass Spectrometry, Anal. Chem., 1998, 70, 4929–4938. 12 Y. Liu and D. E. Clemmer, Characterizing Oligosaccharides Using Injected-Ion Mobility/Mass Spectrometry, Anal. Chem., 1997, 69, 2504–2509. 13 M. F. Jarrold and V. A. Constant, Silicon cluster ions: Evidence for a structural transition, Phys. Rev. Lett., 1991, 67, 2994. 14 F. A. Fernandez-Lima, R. C. Blase and D. H. Russell, A study of ion-neutral collision cross-section values for low charge states of peptides, proteins, and peptide/protein complexes, Int. J. Mass Spectrom., 2010, 298, 111–118. 15 F. A. Fernandez-Lima, H. Wei, Y. Q. Gao and D. H. Russell, On the Structure Elucidation Using Ion Mobility Spectrometry and Molecular Dynamics, J. Phys. Chem. A, 2009, 113, 8221–8234. 16 F. A. Fernandez-Lima, C. Becker, A. M. McKenna, R. P. Rodgers, A. G. Marshall and D. H. Russell, Petroleum Crude Oil Characterization by IMS-MS and FTICR MS, Anal. Chem., 2009, 81, 9941–9947. 17 F. A. Fernandez-Lima, C. Becker, K. J. Gillig, W. K. Russell, S. E. Tichy and D. H. Russell, Ion Mobility-Mass Spectrometer Interface for Collisional Activation of Mobility Separated Ions, Anal. Chem., 2009, 81, 618–624. 18 C. Becker, F. A. Fern´ andez-Lima, K. J. Gillig, W. K. Russell, S. Cologna and D. H. Russell, A Novel Collision-Induced Dissociation (CID) Experiment for Ion Mobility-Mass Spectrometry, J. Am. Soc. Mass Spectrom., 2009, 20, 907–914. 19 F. A. Fernandez-Lima, C. Becker, K. Gillig, W. K. Russell, M. A. C. Nascimento and D. H. Russell, Experimental and Theoretical Studies of (CsI)nCs+ Cluster Ions Produced by 355 nm Laser Desorption Ionization, J. Phys. Chem. A, 2008, 112, 11061–11066. 20 A. B. Kanu, P. Dwivedi, M. Tam, L. Matz and H. H. Hill, Ion mobility-mass spectrometry, J. Mass Spectrom., 2008, 43, 1–22. 21 S. L. Koeniger, S. I. Merenbloom, S. J. Valentine, M. F. Jarrold, H. R. Udseth, R. D. Smith and D. E. Clemmer, An IMS-IMS Analogue of MS-MS, Anal. Chem., 2006, 78, 4161–4174. 22 K. J. Gillig and D. H. Russell, A periodic eld focusing ion mobility spectrometer, US Pat., The Texas A & M University System, United States, 2001, vol. WO0165589, p. 36. 23 K. J. Gillig, B. T. Ruotolo, E. G. Stone and D. H. Russell, An electrostatic focusing ion guide for ion mobility-mass spectrometry, Int. J. Mass Spectrom., 2004, 239, 43–49. 24 B. M. Kolakowski and Z. Mester, Review of applications of high-eld asymmetric waveform ion mobility spectrometry (FAIMS) and differential mobility spectrometry (DMS), Analyst, 2007, 132, 842–864. 25 E. G. Nazarov, R. A. Miller, G. A. Eiceman and J. A. Stone, Miniature differential mobility spectrometry using

1920 | Analyst, 2014, 139, 1913–1921

Paper

26

27

28

29

30

31

32

33

34

35 36

37

38

39 40 41

42 43

atmospheric pressure photoionization, Anal. Chem., 2006, 78, 4553–4563. Y. Guo, J. Wang, G. Javahery, B. A. Thomson and K. W. M. Siu, Ion Mobility Spectrometer with Radial Collisional Focusing, Anal. Chem., 2004, 77, 266–275. R. S. Glaskin, M. A. Ewing and D. E. Clemmer, Ion Trapping for Ion Mobility Spectrometry Measurements in a Cyclical Dri Tube, Anal. Chem., 2013, 85, 7003–7008. S. D. Pringle, K. Giles, J. L. Wildgoose, J. P. Williams, S. E. Slade, K. Thalassinos, R. H. Bateman, M. T. Bowers and J. H. Scrivens, An investigation of the mobility separation of some peptide and protein ions using a new hybrid quadrupole/travelling wave IMS/oa-ToF instrument, Int. J. Mass Spectrom., 2007, 261, 1–12. P. Dugourd, R. R. Hudgins, D. E. Clemmer and M. F. Jarrold, High-resolution ion mobility measurements, Rev. Sci. Instrum., 1997, 68, 1122–1129. S. I. Merenbloom, R. S. Glaskin, Z. B. Henson and D. E. Clemmer, High-Resolution Ion Cyclotron Mobility Spectrometry, Anal. Chem., 2009, 81, 1482–1487. P. R. Kemper, N. F. Dupuis and M. T. Bowers, A new, higher resolution, ion mobility mass spectrometer, Int. J. Mass Spectrom., 2009, 287, 46–57. R. C. Blase, J. A. Silveira, K. J. Gillig, C. M. Gamage and D. H. Russell, Increased ion transmission in IMS: A high resolution, periodic-focusing DC ion guide ion mobility spectrometer, Int. J. Mass Spectrom., 2011, 301, 166– 173. F. A. Fernandez-Lima, D. A. Kaplan and M. A. Park, Gas-phase separation using a trapped ion mobility spectrometer, Int. J. Ion Mobility Spectrom., 2011, 14, 93–98. F. A. Fernandez-Lima, D. A. Kaplan and M. A. Park, Integration of trapped ion mobility spectrometry with mass spectrometry, Rev. Sci. Instrum., 2011, 82, 126106. P. H. Dawson, Quadrupole Mass Sprectrometry and Its Applications, AIP Press, Woodbury, NY, 1995. L. A. Viehland and E. A. Mason, Gaseous ion mobility and diffusion in electric elds of arbitrary strength, Ann. Phys., 1978, 110, 287–328. B. T. Ruotolo, J. A. McLean, K. J. Gillig and D. H. Russell, The inuence and utility of varying eld strength for the separation of tryptic peptides by ion mobility-mass spectrometry, J. Am. Soc. Mass Spectrom., 2005, 16, 158– 165. L. A. Flanagan, Mass Spectrometry calibration using homogeneously substituted uorinated triazatriphosphorines, US Pat., Hewlett-Packard Company, Palo Alto, CA, United States, 1999, vol. 5872357, p. 19. CFD Fluent, Fluent 6.2.16, Fluent Inc., Lebanon, NH. Simion 3D, Scientic Instrument Services, Inc., Ringoes, NJ. L. Ding, M. Sudakov and S. Kumashiro, A simulation study of the digital ion trap mass spectrometer, Int. J. Mass Spectrom., 2002, 221, 117–138. C. Henry, Product Review: Building A Better Trap, Anal. Chem., 1998, 70, 533A–536A. K. J. Gillig, B. K. Bluhm and D. H. Russell, Ion motion in a Fourier transform ion cyclotron resonance wire ion guide


View Article Online


Paper

cell, Int. J. Mass Spectrom. Ion Processes, 1996, 157–158, 129–147. 44 E. N. Nikolaev, R. M. Heeren, A. M. Popov, A. V. Pozdneev and K. S. Chingin, Realistic modeling of ion cloud motion in a Fourier transform ion cyclotron resonance cell by use of a particle-in-cell approach, Rapid Commun. Mass Spectrom., 2007, 21, 3527–3546. 45 S. Kim, M. C. Choi, S. Kim, M. Hur, H. S. Kim, J. S. Yoo, G. T. Blakney, C. L. Hendrickson and A. G. Marshall, Modication of Trapping Potential by Inverted Sidekick Electrode Voltage During Detection To Extend TimeDomain Signal Duration for Signicantly Enhanced Fourier Transform Ion Cyclotron Resonance Mass Resolution, Anal. Chem., 2007, 79, 3575–3580. 46 T. M. Schaub, C. L. Hendrickson, S. Horning, J. P. Quinn, M. W. Senko and A. G. Marshall, High-Performance Mass Spectrometry: Fourier Transform Ion Cyclotron Resonance at 14.5 Tesla, Anal. Chem., 2008, 80, 3985–3990.


Analyst

47 J. Zeleny, On the ratio of velocities of the two ions produced in gases by R¨ ongten radiation, and on some related phenomena, Philos. Mag., 1898, 46, 120–154. 48 M. A. Park, Apparatus and method for parallel ow ion mobility spectrometry combined with mass spectrometry, US Pat., Bruker Daltonics, Inc., Billerica, MA, United States, 2010, vol. 7838826. 49 E. W. McDaniel and E. A. Mason, Mobility and diffusion of ions in gases, John Wiley and Sons, Inc., New York, 1973. 50 L. A. Viehland and E. A. Mason, Gaseous ion mobility in electric elds of arbitrary strength, Ann. Phys., 1975, 91, 499–533. 51 P. Watts and A. Wilders, On the resolution obtainable in practical ion mobility systems, Int. J. Mass Spectrom. Ion Processes, 1992, 112, 179–190. 52 G. F. Verbeck, B. T. Ruotolo, K. J. Gillig and D. H. Russell, Resolution equations for high-eld ion mobility, J. Am. Soc. Mass Spectrom., 2004, 15, 1320–1324.

Analyst, 2014, 139, 1913–1921 | 1921

View Article Online

Analyst


PAPER


A single-walled carbon nanotube thin film-based pH-sensing microfluidic chip Cheng Ai Li, Kwi Nam Han, Xuan-Hung Pham and Gi Hun Seong* A novel microfluidic pH-sensing chip was developed based on pH-sensitive single-walled carbon nanotubes (SWCNTs). In this study, the SWCNT thin film acted both as an electrode and a pH-sensitive membrane. The potentiometric pH response was observed by electronic structure changes in the semiconducting SWCNTs in response to the pH level. In a microfluidic chip consisting of a SWCNT pHsensing working electrode and an Ag/AgCl reference electrode, the calibration plot exhibited promising pH-sensing performance with an ideal Nernstian response of 59.71 mV pH1 between pH 3 and 11

Received 5th December 2013 Accepted 24th January 2014

(standard deviation of the sensitivity is 1.5 mV pH1, R2 ¼ 0.985). Moreover, the SWCNT electrode in the

DOI: 10.1039/c3an02195e

0.1 and 15 ml min1. The selectivity coefficients of the SWCNT electrode revealed good selectivity against common interfering ions.

www.rsc.org/analyst

microfluidic device showed no significant variation at any pH value in the range of the flow rate between

1. Introduction pH sensors are of great importance in real-time biomedical, clinical, in situ environmental monitoring and industrial process control applications.1–7 Among various detection methods, potentiometric detection using small electrodes is an excellent technique as it allows rapid, simple, and reliable pH measurement. Solid-state pH electrodes employing metal/metal oxides or polymer membranes as sensing layers have attracted much attention because of several advantages over conventional glass pH electrodes, i.e., robustness, cost-effective fabrication, no need for maintenance, exibility, possibility of miniaturization and compatibility with microuidic systems known as lab-on-a-chip and micro-total analytical systems.8–13 However, metal oxide-based pH electrodes require multiple fabrication steps to deposit several layers of metal on the substrate. In polymer membrane-based pH electrodes, solidcontact ion-to-electron transducers, such as conducting polymers or carbon nanostructured materials, are used between pH-sensitive polymer membranes and electrodes in order to achieve a stable potentiometric sensor response. Single-walled carbon nanotubes (SWCNTs) greatly improve potentiometric stability, which make them more attractive as transducers than conducting polymers.14 SWCNTs are well known for their outstanding electrical, chemical, thermal and mechanical properties and have been widely used as a promising electrode material in electroanalytical sensors.15 However, employing SWCNTs as pH-sensing materials in potentiometric pH sensors has not yet been explored.

Department of Bionano Engineering, Hanyang University, Ansan 425-791, South Korea. E-mail: [email protected]; Fax: +82-31-436-8148; Tel: +82-31-400-5202


In this article, microuidic pH-sensing chips were developed based on SWCNT lms. The SWCNT thin lms acted both as a pH-sensing membrane and a working electrode to receive the potentiometric signal. The microuidic device integrated with a SWCNT pH-sensing electrode and an Ag/AgCl reference electrode can provide a rapid, highly sensitive, and continuous pH determination with small sample consumption. The potentiometric pH response from the strip-type SWCNT electrodes fabricated on a exible substrate was also examined in a beaker system.

2.

Experimental

2.1. Materials and instruments A solution of SWCNTs (0.3 mg ml1) was purchased from Top Nanosys Co. (Sung Nam, South Korea). Ag/AgCl paste (C61003P7) was obtained from Gwent Electronic Materials Ltd (UK). Buffer solutions with various pH values (pH 2–11) were purchased from Samchun Pure Chemical Co. (Seoul, South Korea). Poly(dimethylsiloxane) (PDMS) and prepolymer components (Sylgard 184 silicone Elastomer Kit) were purchased from Dow Corning (Midland, MI, USA) and AZ4620 photoresist polymer was obtained from Clarivant Corporation (Somerville, NJ, USA). All other chemicals were of analytical grade, and aqueous solutions were prepared with deionized (DI) water. Potentiometric measurements were carried out with an electrochemical analyzer (CHI 660C, CH Instruments Inc., USA). The morphologies of the electrodes were characterized by microscopy (Olympus IX71) and eld emission-scanning electron microscopy (FE-SEM) (Hitachi S-4800, operating at 15 kV). Sample loading was performed by using a syringe pump (PHD2000, Harvard Apparatus, Holliston, MA) which was connected to the inlet port of the microuidic chip.

Analyst, 2014, 139, 2011–2015 | 2011

View Article Online


Analyst

Fig. 1

Paper

Schematic fabrication process of the microfluidic pH-sensing chip.

2.2. Fabrication of the microuidic pH-sensing chip A schematic of the fabrication procedure is illustrated in Fig. 1. Homogeneous SWCNT lms were prepared on glass substrates using a vacuum ltration method. A standard photolithography method was then used to generate photoresist polymer patterns on the SWCNT lms. The SWCNT lms pre-patterned with the photoresist polymer were treated with O2 plasma in the CCP system.16 Aer treatment, the remaining photoresist polymer on

the SWCNT lms was removed with an ethanol solution. For pH determination, potentiometric measurements were used to estimate the potential shis and were performed with a twoelectrode system in which the SWCNT electrode was the working electrode (WE), and the Ag/AgCl electrode was the pseudo reference electrode (RE). The reference electrode was fabricated by painting Ag/AgCl paste onto one of the SWCNT electrodes. A PDMS mold fabricated by so lithography was

Fig. 2 (A) FE-SEM image of SWCNT films. The scale bar is 200 nm. (B) Photo of the developed flexible strip-type pH sensor on a PET substrate. WE is the SWCNT working electrode and RE is the Ag/AgCl reference electrode. (C) Potential responses from a strip-type pH sensor at varied pH values. (D) Calibration plot of the SWCNT-based, strip-type pH sensor in different pH buffer solutions (n ¼ 3).

2012 | Analyst, 2014, 139, 2011–2015


View Article Online

Paper

Analyst

bonded to the electrode substrate using low power O2 plasma treatment.17 The microuidic chips contained a sensing chamber with a 2 mm width and 15 mm height, whereas the electrodes had 1 mm width and spacing.


3.


Homogeneous SWCNT lms were fabricated on the substrate with a thickness of 100 nm, which provided the most suitable conductivity and transparency. The average resistivity and transparency of the fabricated SWCNT lms were 400 ohm sq1 and 80%, respectively. In addition, the FE-SEM image of the SWCNT lms revealed that the ordered SWCNTs were wellconnected and formed a random and homogeneous lm on the substrate (Fig. 2A). Prior to integration into the microuidic device, the potentiometric response of the SWCNT sensing lm was examined using a exible strip-type pH sensor fabricated on a exible poly(ethylene terephthalate) (PET) substrate. As shown in Fig. 2B, the SWCNT lms exhibited a high degree of exibility with a negligible change in resistivity upon hard bending. The diameters of the working and reference electrode were 4 mm with 30 mm 2 mm connecting paths protected by a PET tape. Performance of a pH sensor is typically characterized by measuring the open circuit potential of the electrodes in solutions with various pH values. Fig. 2C illustrates the open circuit potential signal recorded aer immersing the SWCNT-based strip-type pH sensor in solutions with different pH values. The corresponding calibration plot (Fig. 2D) exhibits promising pHsensing performance with an ideal Nernstian response of 59.24 mV pH1 between pH 3 and 11 at 25 3 C (standard deviation of the sensitivity was 1.8 mV pH1, R2 ¼ 0.977). The SWCNT pH-sensitive electrode exhibited a nearly instantaneous response to varying pH solutions, yielding a steady-state signal within 5 s while a completely stabilized signal was observed within 30 s. The selectivity coefficients for the exible pH sensor on the SWCNT electrode were calculated in the presence of several interferents using the xed interference method (Table 1).18,19 For all cases, their selectivity coefficients revealed good selectivity against common interfering ions. The selectivity coefficients for K+, Na+, Li+ and Cl were acceptable compared to the literature results, which were less than 102 for an ion selective electrode.8 In previous reports, several carbon nanotube (CNT)-based pH sensors were demonstrated. CNT-modied glassy carbon electrodes were typically used, which required additional steps

for depositing CNTs on the electrodes. In potentiometric pH sensors, additional polymeric pH-sensitive membranes were prepared on the CNT-modied electrodes.12,20 In voltammetric pH sensors, the pH values were measured by the redox peak potentials of electrodes based on the electrochemical reaction of the electroactive groups on the SWCNTs.21,22 Moreover, the voltammetric pH sensors needed another electrode to make a three-electrode system. The open circuit potential between the SWCNT electrode and the Ag/AgCl reference electrode in different pH buffer solutions can be reasonably ascribed to the doping of the nanotube walls by H+ and OH ions that behave as electron acceptor and donor species, respectively. The change in the concentration of H+ and OH ions causes an electronic structure (rst transition S11) change in semiconducting SWCNTs, causing a respective downshi and up-shi of the Fermi level of SWCNTs in acidic and basic solutions.23,24 The Fermi level can be converted to the electrode potential relative to the standard hydrogen electrode reference with the following equation: E(electrochemical potential) ¼ F(work function)/e 4.44 V.25 More signicantly, the intensity of the rst transition S11 of semiconducting SWCNTs is reversibly tunable by changing the pH,23 which

Table 1 Selectivity coefficients of the flexible pH sensor based on the SWCNT electrode. The selectivity of the SWCNT pH-sensitive membrane was calculated using the fixed interference method

Interference ion (j)

Concentration (mol l1)

Selectivity coefficient (KH+–j)

K+ Na+ Li+ Cl

0.1 0.1 0.1 0.1

8.1 107 6.4 108 3.0 107 9.8 104


Performances of the pH-sensitive SWCNT electrode in the microfluidic device. (A) Potential responses of the SWCNT electrode in the microfluidic device from pH 11 to pH 2 at a flow rate of 5 ml min1. (B) Calibration plot of the microfluidic pH-sensing chip in different pH buffer solutions (n ¼ 3). Fig. 3

Analyst, 2014, 139, 2011–2015 | 2013

View Article Online


Analyst

offers the possibility for use as a pH-sensitive membrane. In addition, SWCNTs may behave as semiconductors or metals depending on helicity and diameter, but the physical separation of semiconducting and metallic SWCNTs has proven to be one of the most difficult challenges to overcome.26 The characteristics of SWCNTs are advantageous for the pH sensor in our study, as the simply patterned thin layer of SWCNTs can act both as a pH-sensitive membrane and a working electrode. Since an ion selective electrode response is size-independent, the pH sensor based on SWCNT electrodes successfully created in a beaker system can be integrated into the microuidic system using microfabrication technology without inuencing performance. Fig. 3A illustrates the potentiometric responses as the pH level of the sample decreased from 11 to 2 at a ow rate of 5 ml min1. The results show that the developed microuidic pH-sensing chip responds well to changes in the sample pH value. In addition, the microuidic device can continuously detect the pH with the performance of the SWCNT lm. Moreover, the open circuit potentials of the pH-sensitive SWCNT electrode with an Ag/AgCl reference electrode in the microuidic device greatly resemble the measurement in beaker solutions. The corresponding calibration plot (Fig. 3B) exhibits promising pH-sensing performance with an ideal

Paper

Nernstian response of 59.71 mV pH1 between pH 3 and 11 at 25 3 C (standard deviation of the sensitivity is 1.5 mV pH1, R2 ¼ 0.985). In addition, inuence of the ow rate on the performance of the microuidic pH-sensing chips was also investigated, and the results are shown in Fig. 4A. The range of the ow rate was examined between 0.1 and 15 ml min1. The SWCNT electrode in the microuidic device shows no signicant variation at any pH value. An average sensitivity of 58.53 mV pH1 was observed with a variation less than 2% over the ow rate range. Therefore, the SWCNT electrode can be operated in a ow analysis system without degrading the performance. Fig. 4B shows the open circuit potential of the SWCNT electrode in KNO3, NaNO3, NaCl solutions successively diluted from 10 mM to 0.1 M at pH 7 with a ow rate of 5 ml min1. The open circuit potentials of the pH-sensitive SWCNT electrode with an Ag/AgCl pseudo reference electrode in the microuidic device show small variations in the KNO3 and NaNO3 solutions over the wide concentration range of 10 mM to 0.1 M, which is less than 7 mV. In addition, the potential variation is also less than 7 mV at low NaCl concentration (10 mM to 0.1 mM), while it is slightly affected at high NaCl concentration (10 mM to 0.1 M) because of the Ag/AgCl pseudo reference electrode used in the microuidic device, for which the variation is lower than 26.5 mV.

4. Conclusions In the present study, a novel SWCNT lm-based pH sensor was developed based on a exible PET substrate or integrated into a microuidic device, and the performance was investigated. The SWCNT thin lm acted both as an electrode and a pH-sensitive membrane. Therefore, the fabrication process provides advantages of lower costs and simpler processes than other all-solidstate pH electrodes. A pair of miniature SWCNT working and Ag/AgCl reference electrodes generated electrical potentials in solutions through the electronic structure change in the semiconducting SWCNTs in response to the pH level. Moreover, the sensitivity of the miniaturized SWCNT-based pH sensors was in agreement with the Nernstian response. As a result, the proposed pH-sensitive SWCNT electrodes in the microuidic device are suitable for practical uses and will enable simultaneous multi-analyte detection of metabolic processes in biological cells when several electrodes are combined.

Acknowledgements This study was supported by National Research Foundation (NRF) grants funded by the Korean government (MSIP) (no. 2011-0015545, 2008-0061891, and 2013R1A1A2054887). We also acknowledge the nancial support of the Ministry of Knowledge Economy (MKE) through the industrial infrastructure program for fundamental technologies (N000600001). Fig. 4 (A) Relationship between the potential responses and flow rates at pH 4, 7, and 10. (B) Effects of interferents on the potential responses at pH 7 with a flow rate of 5 ml min1 ranging from 105 to 101 M.

2014 | Analyst, 2014, 139, 2011–2015

Notes and references 1 E. Lindner and R. P. Buck, Anal. Chem., 2000, 72, 336A.


View Article Online


Paper

2 E. Bitziou, D. O'Hare and B. A. Patel, Anal. Chem., 2008, 80, 8733. 3 C. E. Reimers, Chem. Rev., 2007, 107, 590. 4 A. J. Bandodker, V. W. S. Hung, W. Jia, G. Vald´ es-Ram´ırez, J. R. Windmiller, A. G. Martinez, J. Ram´ırez, G. Chan, K. Kerman and J. Wang, Analyst, 2013, 138, 123. 5 L. Li, A. L. Desouza and G. M. Swain, Analyst, 2013, 138, 4398. 6 S. Kim, T. Rim, K. Kim, U. Lee, E. Baek, H. Lee, C. K. Baek, M. Meyyappan, M. J. Deen and J. S. Lee, Analyst, 2011, 136, 5012. 7 J. Phair, L. Newton, C. McCormac, M. F. Cardosi, R. Leslie and J. Davis, Analyst, 2011, 136, 4692. 8 W. D. Huang, H. Cao, S. Deb, M. Chiao and J. C. Chiao, Sens. Actuators, A, 2011, 169, 1. 9 W. Y. Liao, C. H. Weng, G. B. Lee and T. C. Chou, Lab Chip, 2006, 6, 1362. 10 C. F. Lin, G. B. Lee, C. H. Wang, H. H. Lee, W. Y. Liao and T. C. Chou, Biosens. Bioelectron., 2006, 21, 1468. 11 I. A. Ges, B. L. Ivanov, D. K. Schaffer, E. A. Lima, A. A. Werdich and F. J. Baudenbacher, Biosens. Bioelectron., 2005, 21, 248. 12 G. A. Crespo, D. Gugsa, S. Macho and F. X. Rius, Anal. Bioanal. Chem., 2009, 39, 2371. 13 J. Noh, S. Park, H. boo, H. C. Kim and T. D. Chung, Lab Chip, 2011, 11, 664.


Analyst

14 M. Novell, M. Parrilla, G. A. Crespo, F. X. Rius and F. J. Andrade, Anal. Chem., 2012, 84, 4695. 15 L. Hu, D. S. Hecht and G. Gr¨ une, Chem. Rev., 2010, 110, 5790. 16 K. N. Han, C. A. Li, N. M. P. Bui, X. H. Pham, Y. H. Choa, E. K. Lee and G. H. Seong, Sens. Actuators, B, 2013, 177, 472. 17 K. N. Han, C. A. Li, N. M. P. Bui and G. H. Seong, Langmuir, 2009, 26, 598. 18 C. Macc` a, Electroanalysis, 2003, 15, 997. 19 Y. Umezawa, P. B¨ uhlmann, K. Umezawa, K. Tohda and S. Amemiya, Pure Appl. Chem., 2000, 72, 1851. 20 M. Kaempgen and S. Roth, J. Electroanal. Chem., 2006, 586, 72. 21 Z. Xu, X. Chen, X. Qu, J. Jia and S. Dong, Biosens. Bioelectron., 2004, 20, 579. 22 J. Weber, A. Kumar, A. Kumar and S. Bhansali, Sens. Actuators, B, 2006, 117, 308. 23 W. Zhao, C. Song and P. E. Pehrsson, J. Am. Chem. Soc., 2002, 124, 12418. 24 M. E. Itkis, S. Niyogi, M. E. Meng, M. A. Hamon, H. Hu and R. C. Haddon, Nano Lett., 2002, 2, 155. 25 S. Trasatti, Pure Appl. Chem., 1986, 58, 955. 26 C. B. Jacobs, M. J. Peairs and B. J. Venton, Anal. Chim. Acta, 2010, 662, 105.

Analyst, 2014, 139, 2011–2015 | 2015

View Article Online

Analyst


COMMUNICATION


Quantized double layer charging of Au130(SR)50 nanomolecules† Vijay Reddy Jupally, Jacob G. Thrasher and Amala Dass*

Received 28th November 2013 Accepted 20th January 2014 DOI: 10.1039/c3an02204h www.rsc.org/analyst

Quantized double layer (QDL) charging of a Au130(SR)50 nanomolecule is reported for the first time. In the past, QDL charging was known only for Au144(SR)60 and Au225(SR)75 nanomolecules, which intrigued much research in this field. Here, using differential pulse voltammetry, we demonstrate that Au130 shows QDL charging. Furthermore, 13 different oxidation–reduction waves corresponding to single electron charging events with a clear electrochemical gap of 450 mV are reported. For Au130(SR)50, both the QDL behaviour and electrochemical gap were observed. Considering the charging energy, the HOMO–LUMO gap of the nanomolecule is 200 mV. Further data analysis was performed to find the capacitance of Au130(SR)50. A comprehensive comparison with other magic sized nanomolecules was made to confirm the molecule-like to bulk metal transition with increasing size of gold nanomolecules. Gold nanomolecules with these electrochemical properties are of great interest in the field of catalysis.

Gold nanomolecules (99.0% purity) and glycyrrhetinic acid (130107, >98.0% purity) were purchased from Shanghai Winherb Medical Science Co., Ltd. (Shanghai, China). Standards of liquiritin apioside (11060921-20110609, 95.0% purity) and liquiritigenin (11111831-20111118, 98.0% purity) were bought from Shanghai Tauto Biotech Co., Ltd (Shanghai, China). Ranitidine (TB3UMRT, >98.0% purity, internal standard, IS) was purchased from Tokyo Chemical Industry, Japan. LC-grade methanol (MeOH) was obtained from Fisher Co. Ltd (Emerson, IA, USA). Acetic acid and other reagents and chemicals were of analytical grade and purchased from Beijing Chemical Works (Beijing, China). Wahaha puried water (Wahaha, Hangzhou, China) was used. 2.2. Instrumentation 2.2.1. UHPLC-PDA instrument and conditions. The ngerprinting analysis of tested samples was performed on a Waters Acquity UPLC H-Class system (Milford, MA, USA), which included a quaternary solvent delivery pump, an autosampler, a photodiode array (PDA) detector and a column heater equipped with a Waters Acquity UPLC HSS T3 column (2.1 mm 50 mm, 1.8 mm, Milford, MA, USA) at room temperature. The mobile phase consisted of (A) 0.5% acetic acid in water and (B) acetonitrile. Elution was performed in the gradient mode: 0 min, 80% A; 3 min, 60% A; 8 min, 20% A. The ow rate was 0.2 mL min1 and injection volume was 1.0 mL. The PDA detection wavelength was set at 248 nm.


View Article Online


Paper

2.2.2. UFLC-MS/MS instrument and conditions. The separation of ve targeted compounds was carried out on an ultrafast liquid chromatography (UFLC) system, which consisted of two LC-20ADXR pumps, a DGU-20 A3 degasser, an SIL-20AC auto-sampler and a CTO-20A column oven (Shimadzu, Japan). The UFLC separation was performed on a Phenomenex Kinetex C18 100A column (100 2.1 mm, 2.6 mm) by gradient elution using a mobile phase consisting of (A) water (containing 0.1% acetic acid) and (B) methanol (containing 0.1% acetic acid) with the following gradient procedure: 0 min 40% B, 5 min 90% B, 8.5 min 100% B and 8.51 min 60% B, with the ow rate of 0.3 mL min1. The injection volume was 1.0 mL. An Applied Biosystem 5500 QTRAP® hybrid triple quadrupole/linear ion trap (QqQLIT) mass spectrometer (Applied Biosystems/MDS Sciex, Foster City, CA, USA) equipped with Turbo V sources and a Turbo IonSpray interface was coupled to the UFLC system described above. Nitrogen was used as the nebulizer (GS1), heater (GS2) and curtain (CUR) gas, as well as the collision activation dissociation (CAD) gas. The mass spectrometer was calibrated semiannually using polypropylene glycol (PPG) for mass accuracy, and for practical applications, which was operated at unit resolution for Q1 and Q3 using the electrospray ionization (ESI) source in the negative ionization mode. Tandem MS analyses were performed in the MRM acquisition mode for simultaneous detection of all the targeted analytes, with two precursor-to-product ion transitions monitored for each compound. The source settings and instrument parameters for each MRM transition were optimized not only to maximize the generated deprotonated analyte molecule ([M–H]) of each targeted compound, but also to efficiently produce its characteristic fragment/product ions. The ion spray needle voltage was set at 4500 V and the source temperature was 550 C. Following optimization of the setting parameters, the ESI source was operated with the CUR, GS1 and GS2 set at 35, 50 and 50 psi, respectively. The compound-dependent instrumental parameters of two individual precursor-to-product ion transitions specic for each analyte including precursor ion, two product ions, declustering potential (DP), entrance potential (EP), collision energy (CE) and collision cell exit potentials (CXP) were optimized and are listed in Table 1. An 80 ms dwell time was used for each MRM transition in one single retention time window. To get synchronous supplementary conrmation of targeted analytes via the LC-QqQLIT-MS/MS system, an IDA experiment was carried out using the MRM mode to automatically trigger EPI scan (full-spectrum ion-trap MS/MS mode). The EPI scans were operated in the ESI mode over a mass range of m/z 100– 1000 for product ions at a scan rate of 20 000 Da s1 with dynamic ll in the linear ion trap and a step size of 0.12 Da. The CEs of EPI were set at three discontinuous energy levels of 15, 40 and 65 eV to provide rich EPI spectra. For the MRM-EPI analysis, the IDA criteria were set to allow the two most intense product ions of each analyte that exceeded 1000 counts per second (cps), with dynamic exclusion of the former ions for 15 s. Data acquisition and processing were performed using Analyst 1.6 soware (Applied Biosystems/MDS Sciex, Foster City, CA, USA).


Analyst

2.3. Standard solution preparation The appropriate amount of each standard and IS was weighed and dissolved in methanol to make 6 individual stock solutions. Then, each stock solution was mixed to prepare a nal mixed stock solution containing 5 standards and IS, which was stored in a refrigerator at 4 C. For analysis, the mixed solution was diluted by initial mobile phase to obtain a series of working solutions of different concentrations. 2.4. Sampling and sample preparation Ten batches of licorice samples cover the three species of Glycyrrhiza uralensis Fisch, Glycyrrhiza glabra L. and Glycyrrhiza inata Bat. Of them, 3 batches of fresh wild licorice samples (S1, S6, S8) were kindly supplied by the School of Traditional Chinese Medicine, Xinjiang Medical University (Xinjiang, China) with the origin of Xinjiang Autonomous Region, and other 7 batches of dried samples were collected from Inner Mongolia Autonomous Region (S2, S7), and Jiangxi (S3), Guizhou (S4), Gansu (S5, S9) and Hainan (S10) provinces in China. All samples were identied by Professor Bengang Zhang, Institute of Medicinal Plant Development, Chinese Academy of Medical Science and Peking Union Medical College (Beijing, China), and were sealed in plastic bags and stored at 20 C before analysis. The voucher specimens of these samples were deposited at the Institute of Medicinal Plant Development, Chinese Academy of Medical Science and Peking Union Medical College (Beijing, China). The dry licorice samples were powdered to a homogeneous size by a pulverizer and sieved through a 40-mesh sieve. An aliquot of 0.5 g of each sample was weighed into a 50 mL capped conical ask and 100 mL of IS solution (1 mg mL1) was added. The sample was macerated with 3 mL of methanol–water (50 : 50, v/v) for 10 h, and then was extracted in a KQ-500B ultrasonic water bath (Kunshan Ultrasonic Instrument Co. Ltd., Jiangsu, China) operating at 40 kHz with the output power of 500 W for 20 min at room temperature. The sample was centrifuged for 5 min at 8000 rpm using a centrifuge. The supernatant was transferred into a 10 mL volumetric ask, while the sediment was re-extracted with an additional 2 mL of methanol–water (50 : 50, v/v) twice, 5 min for each, and centrifuged as described above. All the supernatant was collected, combined into the volumetric ask to 10 mL by adding methanol–water (50 : 50, v/v). The solution was injected into the UHPLC and UFLC systems, respectively, aer 100-fold dilution and ltering through a 0.22 mm microporous membrane. All samples were stored at 4 C in the dark until analysis. 2.5. Similarity analysis of UHPLC-PDA chemical ngerprints To obtain integrated chromatographic proles and more composition information, the chemical ngerprints of tested licorice samples was established rstly by UHPLC-PDA. Then, these ngerprints were evaluated by similarity analysis, which came from the calculation on the correlative coefficient of original data of the whole chromatogram26,27 by professional soware named Similarity Evaluation System for Chromatographic

Analyst, 2014, 139, 1883–1894 | 1885

View Article Online

Analyst


Table 1

Paper LC-MS/MS parameters of the 5 targeted analytes and ranitidine

Analytes

Molecular formula

tRa (min)

m/z precursor [M–H]

DPb (V)

EPc (V)

m/z Productionsd

Ion ratioe

CEf (eV)

CXPg (V)

Dwell time (ms)

Ranitidine (IS) Liquiritin apioside Liquiritin Liquiritigenin Glycyrrhizic acid Glycyrrhetinic acid

C13H22N4O3S C26H30O13 C21H22O9 C15H12O4 C42H62O16 C30H46O4

0.70 2.08 2.16 3.44 5.81 7.26

313.3 549.2 417.2 255.3 821.4 469.3

110 120 100 100 80 100

10 10 10 10 10 10

170.2/111.9 135.0/255.0 255.1/135.0 119.0/134.9 351.1/253.3 355.3/409.2

0.31 2.88 0.30 0.82 0.39 0.60

19/35 49/43 27/41 35/21 55/43 62/62

12/11 15/15 25/35 16/18 18/18 17/17

80 80 80 80 80 80

a Retention time. b Declustering potential. c Entrance potential. d Values are given in the order of quantitative ion/qualitative ion. e Intensity of qualitative ion/intensity of quantitative ion. f Collision energy. g Cell exit potential.

Fig. 1 (A) Typical UHPLC-PDA fingerprint of licorice sample, and (B) reference fingerprint generated by professional software named Similarity Evaluation System for Chromatographic Fingerprint of Traditional Chinese Medicine (Version 2004A) using the median method.

Fingerprint of Traditional Chinese Medicine (Version 2004A) composed by the Chinese Pharmacopoeia Committee (Beijing, China). The aims were to describe the similarities and differences among these samples from different origins, further to explore the quality variations. 2.6. UFLC-MS/MS targeted analysis of active compounds 2.6.1. Method validation. The UFLC-MS/MS method was validated in terms of specicity, selectivity, linearity, sensitivity, precision (intra- and inter-day variability), repeatability, accuracy and matrix effect, in accordance with the U.S. FDA and International Conference on Harmonization (ICH) guidelines on analytical method validation.28,29 Licorice sample S6 was selected at random for this validation. 2.6.1.1. Specicity and selectivity. Known amounts of individual standards, which were almost equal to their contents in the sample, together with IS, were added into approximately 0.25 g of sample S6. The mixture was extracted and analyzed using the aforementioned procedures. Three replicates were performed for the test. The specicity and selectivity of the method was assessed by comparing the chromatograms of each standard, normal (unspiked with the standards, but with IS)

1886 | Analyst, 2014, 139, 1883–1894

and spiked samples through searching the characteristic transitions of the ve targeted compounds in the appropriate retention time, as well as two SRM transitions and monitoring of SRM ion ratio (intensity of qualitative ion/intensity of quantication ion). 2.6.1.2. Linearity and sensitivity. A nine-level series of mixed standard solutions containing appropriate concentrations of the ve standards (1, 10, 25, 50, 100, 200, 400, 800 and 1600 ng mL1 for liquiritin apioside, liquiritigenin, glycyrrhizic acid and glycyrrhetinic acid, and 0.727, 7.27, 18.18, 36.35, 72.7, 145.4, 290.8, 581.6 and 1163.2 ng mL1 for liquiritin) and IS were analyzed in duplicates to construct the calibration curves by plotting the peak area ratio (y) of individual standard to IS versus the concentration (x) of each analyte with least square linear regression of slope and intercept (y ¼ ax + b). Then, the linear range and the correlation coefficient (r) of each analyte were calculated. The sensitivity of the method was evaluated based on limits of detection (LOD) and quantication (LOQ) for each analyte, which were determined at a signal-to-noise ratio (S/N) of about 3 and 10, respectively. 2.6.1.3. Precision, repeatability and stability. Intra- and interday variation was selected to determine the precision of the established method. For intra-day variability test, the mixed standard solutions were analyzed for six replicates within one day, while for inter-day variability test, the solutions were examined in duplicates on three consecutive days. Variations were expressed by relative standard deviation (RSD) of the peak area of each standard. The repeatability of the analytical method was evaluated at three levels (0.1, 0.25 and 0.5 g) of the licorice samples S6, which were extracted and analyzed in triplicate as previously. The content (mg g1) of each compound in the licorice sample was determined from the corresponding calibration curve, and the RSD value of the content was calculated as a measurement of the repeatability of the established method. To assess the stability of the samples, the same sample (S6) solution was stored at 4 C in the dark and analyzed every 6 h over 2 days. The RSD value of the peak area was attained for this investigation. 2.6.1.4. Accuracy. The accuracy of the method was evaluated by recovery experiments. A known amount of individual standard and 100 mg of IS were spiked into approximately 0.25 g of the licorice sample S6. The mixture was extracted and analyzed


View Article Online


Paper

Analyst

Fig. 2 Full-scan product ion spectra and the proposed fragmentation schemes of (A) liquiritin apioside, (B) liquiritin, (C) liquiritigenin, (D) glycyrrhizic acid, (E) glycyrrhetinic acid, and (F) IS.

as previously. The detected content of each compound contained in the chosen sample was determined from their corresponding calibration curve. Three replicates were performed and the extraction recovery of each analyte was calculated using the following equation: Recovery (%) ¼ 100 (found amount original amount)/ spiked amount

2.6.1.5. Matrix effect. In developing an LC-MS/MS method using ESI source, the matrix components, especially in complex TCM matrices, may affect the ionization of targeted analytes, further compromising the quantitative analysis of them, as well as greatly affecting the method accuracy and reproducibility. When matrix components and targeted analytes enter the ion source at the same time, the ionization efficiency of the analyte might be inuenced by the co-eluting components, presenting as signal enhancement or suppression if reliable results are to be obtained. So, in this process, evaluation of the matrix effect is important and necessary. Standard addition is an effective method for providing favorable results even with variable matrices.30


In this study, approximately 0.25 g of the un-spiked licorice sample S6, the spiked licorice sample S6 with addition of known amounts of individual standards at low, middle and high levels (almost 0.8-, 1.0- and 1.2-fold their contents in sample S6), were extracted and analyzed as previously. Each sample was prepared three times and the solution was added with 100 mg of IS and was injected into the UFLC system aer 100-fold dilution and ltering through a 0.22 mm microporous membrane. The matrix effect (ME, %) was calculated using the following the formula:31 ME (%) ¼ 100 (A1 A2)/A3 where A1 is the peak area of the analyte in the spiked sample matrix, A2 is the peak area of the analyte in the un-spiked sample matrix and A3 is the peak area of the standard at the same concentration. When the value is equal to 100%, no matrix effect is observed, while a value over 100% indicates ionization enhancement, and a value lower than 100% suggests ionization suppression.32 2.6.2. Peak identication. The targeted analytes in the sample were clearly identied by the comparison of MS/MS spectra (MRM mode) based on two characteristic transitions and their retention times, as well as the selected reaction

Analyst, 2014, 139, 1883–1894 | 1887

View Article Online

Paper


Analyst

Typical LC-MS/MS MRM chromatograms of (A) standard solutions of each analyte (100 ng mL1) and IS (50 ng mL1), (B) blank solution, and (C) licorice sample (500 mg mL1) spiked with IS (10 ng mL1). Fig. 3

monitoring (SRM) ion ratio with those of the standard solution. For synchronous structural conrmation, all of the peaks of the targeted compounds were unambiguously identied through

1888 | Analyst, 2014, 139, 1883–1894

the comparison of their precursor and product ions in the EPI spectra with those of the standards. Then, the aims qualitative and quantitative analyses could be achieved at the same time.


View Article Online

Paper

Analyst Linear regression data, LOD and LOQ of the investigated compounds

Table 2

Analytes

Linear equationa

Linear range (ng mL1)

r

LOD (ng mL1)

LOQ (ng mL1)

Liquiritin apioside Liquiritin Liquiritigenin Glycyrrhizic acid Glycyrrhetinic acid

y ¼ 2.98 103x + 2.03 104 y ¼ 9.92 104x + 3.14e5 y ¼ 1.23 104x + 7.08 104 y ¼ 280x + 4.54 104 y ¼ 327x 2.70 103

1–1600 0.727–290.8 1–800 50–1600 1–1600

0.9997 0.9985 0.9992 0.9954 0.9995

0.3 0.03 0.5 0.05 0.3

1.0 0.1 1.2 0.16 1.0


a

x, analyte concentration (ng mL1); y, peak area.

3.


3.1. Establishment of UHPLC-PDA chemical ngerprints and similarity analysis As is known, chemical ngerprint could clearly character the component information of TCM samples. Here, the chemical ngerprint of tested licorice sample from various origins was established under the above-mentioned UHPLC-PDA conditions. Typical ngerprint of licorice sample, and reference ngerprint to describe the common information of the 10 chromatograms of samples generated by professional similarity evaluation soware were shown in Fig. 1. All compounds in the licorice samples could be well separated in 9.0 min, while, the peak responses/heights of the compounds in ngerprints of 10 batches of licorice samples revealed signicant variations, reecting the differences of chemical information. Then, the similarities between the entire chromatographic ngerprint of each licorice sample and the reference ngerprint were calculated by using this soware and presented as correlation coefcients of 0.934, 0.915, 0.914, 0.882, 0.577, 0.872, 0.888, 0.882, 0.409, and 0.553, respectively. The results, to a large extent, show the quality differences of licorice samples from various origins, as well as chemical compositions. So, in the next parts, an LC-MS/MS method was established for the analysis of the main active components, including two triterpene saponins compounds (glycyrrhizic acid and glycyrrhetinic acid), and three avonoids compounds (liquiritin, liquiritin apioside and liquiritigenin), to provide some quantitative information for exploring the quality differences of tested licorice samples.

3.2. UFLC-MS/MS targeted analysis of active compounds 3.2.1. Proposed fragmentation pathways of the ve targeted analytes. To achieve accurate qualitative and quantitative analysis of the analytes in licorice samples, the proposed

Table 3

fragmentation pathways of the ve targeted compounds were rst explored. The ve standards were characterized according to their mass spectra through manual tuning mode by the syringe pump continuous infusion analysis of individual standard solutions of each compound at 10, 50 or 100 ng mL1 and IS at 50 ng mL1 depending on the sensitivity of the compounds, to nd their optimum mass spectrometric behaviors and optimize the ESI source parameters. Experiments were carried out in both positive and negative ionization modes. Firstly, full-scan mass spectra ranged from m/z 50 to 1000 were recorded to select the most abundant m/z value. The results show that the compounds responses from the negative ionization mode show higher sensitivity and intensity to permit quantitative determination, and thus the deprotonated molecular ion [M–H] was measured and chosen as the precursor ion for each compound. As one of the most important mass spectrometer parameters affecting the ion response, DP was optimized to obtain the maximum sensitivity at the same time. Once the precursor ion was obtained, product ion scan acquisitions were carried out to select the most suitable product ions with high intensity and only the selected precursor ion was dissociated into the product ions. For this, dissociation with nitrogen was induced and different CEs were compared. By adjusting the CEs, several fragment ions of the analytes were observed in the product ion spectra and the predominant fragment ions were chosen in SRM transitions. The most abundant transition was commonly selected for quantication and the others for qualitation purposes. Product ions resulting from non-specic losses (such as H2O or CO2 losses) were avoided. Then, the most suitable CE and CXP for each precursor-to-product transition of the studied compounds were determined by observing the maximum response for the MS/MS monitoring fragment ions. All the MS/MS parameters have been summarized in Table 1, and most were in good agreement with

Intra- and inter-precision assay, accuracy and matrix effect of the five targeted compounds in licorice Precision (n ¼ 6)

Accuracy (n ¼ 6)

Analytes

Intra-day RSD (%)

Inter-day RSD (%)

Recovery (%)

RSD (%)

ME (%, n ¼ 9)

Liquiritin apioside Liquiritin Liquiritigenin Glycyrrhizic acid Glycyrrhetinic acid

1.15 1.81 1.47 1.45 4.56

3.95 0.81 2.60 1.42 3.62

99.87 97.33 97.80 100.40 99.70

2.21 4.27 0.31 0.74 1.05

102.91 96.74 97.21 99.95 99.87


Analyst, 2014, 139, 1883–1894 | 1889

View Article Online


Analyst

Paper

Fig. 4 EPI spectra for synchronous confirmation of five targeted analytes in licorice and IS based on MRM-IDA-EPI mode.

reported data.19 As liquiritin apioside and liquiritin have the same product ions under the above-described MS/MS conditions, the quantitative ions are selected differently to avoid some interference, although the intensity of a certain quantitative ion is not so high. Under the optimized parameters, fullscan product ion spectra and the proposed fragmentation schemes of each standard are shown in Fig. 2. 3.2.2. Optimization of sample extraction conditions. In order to get high recovery and quick extraction of the analytes and IS with no endogenous interference at the retention times, ultrasound-assisted solid–liquid extraction (USLE) method, a convenient, quick and cost-effective approach for extraction of analytes from complex matrices without the need of specialized instrumentation but offering high reproducibility, simple manipulation and low solvent consumption,33,34 was selected rst. Then, the extraction solvent (methanol, acetonitrile), ratio of methanol or acetonitrile/water (30, 50, 70 and 90%, v/v), solvent volume (1, 3 and 5 mL), and extraction time (10, 20 and

30 min), and extraction cycle (1, 2 and 3 times) were investigated. The sum numbers and responses of characteristic peaks in each chromatogram obtained under different conditions were compared. It was demonstrated that the afore-selected extraction conditions produced the best extraction efficiency for all the analytes and IS. 3.2.3. Optimization of UFLC conditions. To obtain reproducible separation and high sensitivity for targeted analytes within a reasonable separation time, UFLC conditions were systematically optimized. Different particle sizes and lengths of chromatographic columns were compared in the preliminary experiments, and the results show that Phenomenex Kinetex C18 100A column (100 2.1 mm, 2.6 mm) offers a satisfactory separation of the ve analytes and IS, as well as short analytical time, was selected as the analytical column in the present study. Next, to obtain high peak responses for each analyte and short analysis time for the whole operation, the suitable composition of mobile phase has also been concerned based on the

Table 4 Contents (mg g1) of five targeted compounds in tested licorice (n ¼ 3)

Contents (mg g1) (n ¼ 3)

Liquiritigenin

Glycyrrhizic acid

0.43 0.37

0.33 4.22

72.45 19.24

0.04 0.12

6.65 0.19 1.10 9.92 1.72

0.59 0.30 0.25 0.46 0.33

1.52 0.83 0.29 0.30 0.28

56.12 12.40 10.52 67.09 13.72

0.05 10.85 0.04 0.20 0.04

7.95 1.54 1.88 9.77 17.9

0.19 0.43 0.22 0.36 34.7

0.15 0.06 0.62 0.32 6.7

55.40 10.15 8.56 69.77 5.4

0.31 0.04 19.15 0.12 94.3

Species

Samples

Origin

Liquiritin apioside

Glycyrrhizae uralensis Fisch

S1 S2

Xinjiang Inner Mongolia Jiangxi Guizhou Gansu Xinjiang Inner Mongolia Xinjiang Gansu Hainan

11.43 23.31

Glycyrrhizae inata L.

Glycyrrhizae glabra Bat.

a

S3 S4 S5 S6 S7 S8 S9 S10 Mean C.Va. (%)

Liquiritin

Glycyrrhetinic acid

C.V% ¼ s/m 100; s is the standard deviation, m is the average value of the content.

1890 | Analyst, 2014, 139, 1883–1894


View Article Online


Paper

improved LC separation and ionization efficiency in LC-MS/MS protocols. The results manifested that there were no signicant differences between methanol–water and acetonitrile-water as the mobile phase. Taking the higher toxicity and price of acetonitrile than methanol into consideration, the binary mixture of methanol–water was preferred. According to the physical and chemical properties of targeted analytes, some mobile phase additives, such as formic acid and acetic acid, as well as different ratios of them to water, were assayed in the gradient program to attain high sensitivity and improve the peak shape. The data elucidated that the eluent methanol– water (containing 0.1% acetic acid) with a 0.3 mL min1 ow rate gave good separation and symmetric peak shape of the targeted analytes and were benecial for enhancing the ionization of them in the negative ionization mode. Owing to the long retention time and bad separation efficiency of some of the targeted analytes in the isocratic runs, gradient elution was employed in the UFLC analyses. When establishing an LC-MS/MS method, a suitable IS is required to eliminate the effects from matrix and the extraction efficiency. Usually, a radio-label is a preferred choice, however, it is not available during the period of method development. In this study, ranitidine, a readily available compound not contained in TCMs, was selected as the IS, which displayed high extraction efficiency (>80%) and did not interfere with the detection of the targeted analytes. Using the afore-described conditions, satisfactory separation of the targeted analytes and IS was achieved in 8.0 min by gradient elution, which was shorter than that of previous reports.18,19 Typical LC-MS/MS extract ions chromatograms (XIC) in the MRM mode of standard solutions of each analyte and IS (50 ng mL1) blank solution, and licorice sample spiked with IS are shown in Fig. 3. The ve analytes and IS show good separation and excellent repeatability, as the relative standard deviations (RSDs) for retention times were less than 0.2% for standards in solvent, and no more that 0.3% for the sample matrices. In addition, no carryover was observed between the runs, as demonstrated by the consecutive analysis of highconcentration standards and blanks. Although liquiritin apioside and liquiritin expressed close retention times (2.08 min and 2.16 min, respectively), this would not inuence the accurate quantitation based on their two precursor-to-product transitions in the MRM mode. 3.2.4. Method validation 3.2.4.1. Specicity and selectivity. By comparing the XIC for each MRM transition of the standards and IS, as well as them in the normal and spiked licorice samples, it could be found that no interference peaks appeared at the retention time of each analyte and IS in these samples. In addition, the responses of each MRM transition of the standards showed more signicant increase in the spiked licorice samples than that in the normal samples. The results expressed good specicity and selectivity of the established UFLC-MS/MS method. 3.2.4.2. Linearity and sensitivity. The linear regression equation, linearity ranges, together with the corresponding correlation coefficient (r) for the ve analytes are listed in Table 2. All of the analytes expressed satisfactory linearity


Analyst

(r $ 0.9954) over a relatively wide concentration range. The LOD and LOQ for the ve compounds were less than 0.03 and 0.1 ng mL1, respectively, and were much lower than those obtained using capillary electrochromatography with peak suppression diode array,13 high-performance liquid chromatography with ultraviolet,17 and ultra-performance liquid chromatography with photodiode array18 detection methods, indicating high sensitivity of the method with these chromatographic and MS/ MS conditions. 3.2.4.3. Precision, repeatability accuracy and stability. As shown in Table 3, the RSD values, which were used to express the intra- and inter-day precisions of the investigated compounds, ranged from 1.15 to 4.56% for the intra-day precision, and from 0.81 to 3.95% for the inter-day precision, illustrating good precision of the established method. By sampling licorice at two levels and analyzing for the targeted compounds, it could be seen that the RSD values of ve analytes ranged from 0.76 to 2.23%, revealing high repeatability of the method. The data listed in Table 3 show that the average recoveries of the ve targeted compounds were in the range of 97.33– 100.40% with RSD values from 0.31 to 4.27%, indicating that the proposed method is accurate and reproducible. Through analyzing the same licorice sample solution within 2 days, it was found to be stable as the RSD values of the peak area were lower than 2.56%. 3.2.4.4. Matrix effect. Matrix effects for each targeted compound were calculated for the licorice matrix and are summarized in Table 3. It could be seen that glycyrrhizic acid and glycyrrhetinic acid shows almost no matrix effects, while, liquiritin apioside displays slight signal enhancement, and signal suppression was observed for liquiritin and liquiritigenin. However, these poorly enhanced and suppressed signals would not interfere with the accurate determination of the ve targeted compounds by using the developed UFLC-MS/ MS method. 3.2.5. EPI spectra for synchronous conrmation of targeted compounds. The IDA criteria was employed for triggering the acquisition of EPI spectra to provide highly sensitive information about product ions from selected precursor ions for synchronous conrmation of targeted analytes, and further for accurate quantitation. Here, the MRM chromatograms of the ve targeted compounds were triggered by IDA to get their EPI spectra, which are shown in Fig. 4. Then, the EPI data were compared to the MS/MS product ion spectra in Fig. 2. It was clearly seen that the product ions (Fig. 2) for the ve compounds obtained through manual tuning by the syringe pump continuous infusion analysis of individual standard were almost correspondingly present in the EPI spectra obtained through the developed LC-MS/MS method (Fig. 4). These results have authenticated the reliability of the established method for determination of the ve targeted compounds. The above-described results have reected excellent performance of the developed UFLC-MS/MS method and elucidated that the method was an ideal alternative for simultaneous qualitative and quantitative analysis of the ve targeted compounds in licorice with regards to its satisfactory selectivity,

Analyst, 2014, 139, 1883–1894 | 1891

View Article Online


Analyst

sensitivity, precision, accuracy and few matrix effect, as well as strong qualitative and quantitative ability. 3.2.6. Quantitative analysis of the ve targeted compounds in real samples. The proposed UFLC-MS/MS method was practically applied to analyze 5 compounds in 10 batches of licorice samples from different origins in China. Owing to its good sensitivity, MRM scan mode signicantly decreased the levels of noise and accordingly enhanced the response of the analytes. Thus, the components at very low contents in licorice could be accurately detected. The ve targeted compounds were identied by comparison of their retention time, precursor and product ions, together with their ion ratio, obtained from LCMS/MS analysis with data of standards, and conrmed by the fragment ions produced in the MRM-IDA-EPI mode. The quantitative analyses were performed by mean of internal standard method and the results are summarized in Table 4. The results show that, in all licorice samples, glycyrrhizic acid was the most prevalent component with a mean content of 69.77 mg g1 in the range of 8.56–72.45 mg g1. Although the coefficient of variance (C.V%) of the content of glycyrrhizic acid was the smallest (5.4%) in the ve compounds, its contents in some samples, such as S2, S4, S5, S7, S9, S10, were all less than 20 mg g1 set in the Chinese Pharmacopeia,2 illustrating that these samples do not meet with the requirement and should be given more attention in practice. Glycyrrhetinic acid shows the most signicant variation of contents (range of 0.40–10.85 mg g1) in the tested samples with the biggest C.V% value (94.3%), which should be taken into consideration in quality control of licorice. The contents, as well as the mean values, of liquiritin were all smaller than 5 mg g1 set in the Chinese Pharmacopeia.2 Also, the C.V% value (34.7%) of the contents of liquiritin is bigger than that of liquiritin apioside, liquiritigenin, glycyrrhizic acid in licorice collected from different origins. Therefore, extra attention should be given to liquiritin. With regard to the species, it could be noticed that the contents of the ve targeted compounds in licorice samples from the same origins of Glycyrrhizae uralensis Fisch was the biggest, followed by Glycyrrhizae inata L. and Glycyrrhizae glabra Bat., such as S1, S6 and S8 from Xinjiang Autonomous Region, S2 and S7 from Inner Mongolia Autonomous Region, as well as S5 and S9 from Gansu Province. These results imply that Glycyrrhizae uralensis Fisch might be preferred in practical use. The above ndings are consistent with the results from similarity analysis of the UHPLC-PDA chemical ngerprints, and could be reected from the similarity values between the chromatographic ngerprint of each licorice sample and the reference ngerprint. The various differences of quality of licorice samples might be due to their different geographic setting, environmental conditions, cultivation methods, etc. which will be studied in future work. In addition, Chen et al.35 and Wei et al.36 have found that licorice is prone to contamination by various fungi and their toxic metabolisms-mycotoxins. Our studies have shown that this contamination has resulted in the decrease of the contents of the ve targeted active components in licorice,37,38 and further may have a negative impact on the quality, as well as the safety, of this TCM. Regarding the conditions under which

1892 | Analyst, 2014, 139, 1883–1894

Paper

fungi grow and mycotoxins are produced, the storage conditions, such as high humidity and temperature, should be in good control to avoid or reduce the contamination of licorice. Based on the current samples, we considered that the quality of licorice can be assessed, but the place of origin should be standardized or unied.

4. Conclusions Taking the signicant quality differences of licorice samples from various origins into account, a novel UFLC-ESI-MS/MS method based on MRM-IDA-EPI mode was established for the simultaneous qualitation and quantication of ve compounds in 10 batches of licorice. This method can quickly separate the ve targeted analytes in a short time with good specicity and is environmentally-friendly and inexpensive. The sample preparation procedure, UFLC and MS/MS conditions were optimized and excellent linearity, sensitivity, repeatability, precision, accuracy, and good recovery values (with RSDs below 5.0% in all cases) were obtained. The LODs and LOQs were sufficiently low allowing for identication and quantication of the ve targeted compounds in licorice, and the satisfactory results obtained from the real samples further conrmed the reliability and efficacy of the method. The proposed method is promising and can be regarded as a strong alternative tool for improving the quality control of licorice and also provides a model for qualitative and quantitative analysis of more complex TCM matrices.

Acknowledgements The authors thank the anonymous referees for powerful suggestions. Also, the authors are grateful for the support from the National Science Foundation of China (81274072) and PUMC Youth Fund and the Fundamental Research Funds for the Central Universities (3332013078).

References 1 R. A. Isbrucker and G. A. Burdock, Risk and safety assessment on the consumption of Licorice root (Glycyrrhiza sp.) its extract and powder as a food ingredient, with emphasis on the pharmacology and toxicology of glycyrrhizin, Regul. Toxicol. Pharm., 2006, 46, 167–192. 2 Chinese Pharmacopeia Commission. Pharmacopeia of the People's Republic of China, Version 2010, China Medical Science Press, Beijing, 2010, vol. 1, p. 80. 3 V. K. Gupta, A. Fatima, U. Faridi, A. S. Negi, K. Shanker, J. K. Kumar, N. Rahuja, S. Luqman, B. S. Sisodia, D. Saikia, M. P. Darokar and S. P. Khanuja, Antimicrobial potential of Glycyrrhiza glabra roots, J. Ethnopharmacol., 2008, 116, 377–380. 4 S. H. van Uum, Liquorice and hypertension, Neth. J. Med., 2005, 63, 119–120. 5 N. Mumoli and M. Cei, Licorice-induced hypokalemia, Int. J. Cardiol., 2008, 124, e42–e44.


View Article Online


Paper

6 Y. Fujisawa, M. Sakamoto, M. Matsushita, T. Fujita and K. Nishioka, Glycyrrhizin inhibits the lytic pathway of complement-possible mechanism of its anti-inammatory effect on liver cells in viral hepatitis, Micrrobiol. Immunol., 2000, 44, 799–804. 7 A. Ram, U. Mabalirajan, M. Das, I. Bhattacharya, A. K. Dinda, S. V. Gangal and B. Ghosh, Glycyrrhizin alleviates experimental allergic asthma in mice, Int. Immunopharmacol., 2006, 6, 1468–1477. 8 B. Liu and Y. Qi, Advances in studies on pharmacology of glycyrrhizic acid and glycyrrhetinic acid, World Notes Plant Med., 2006, 21, 100–104. 9 M. N. Asl and H. Hosseinzadeh, Hosseinzadeh. Review of pharmacological effects of Glycyrrhiza sp. and its bioactive compounds, Phytother. Res., 2008, 22, 709–724. 10 C. Fiore, M. Eisenhut, E. Ragazzi, G. Zanchin and D. Armanini, A history of the therapeutic use of liquorice in Europe, J. Ethnopharmacol., 2005, 99, 317–324. 11 H. Nagai, J. X. He, T. Akao and T. Tani, Antispasmodic activity of licochalcone A, a species-specic ingredient of Glycyrrhiza inata roots, J. Pharm. Pharmacol., 2007, 59, 1421–1426. 12 Y. Q. Liu, M. Y. Wang, H. M. Shi, Q. S. Sun and X. B. Li, Determination of ve active Constituents in Licorice and its processed products, Chin. Pharm. J., 2008, 43, 1268–1271. 13 X. J. Chen, J. Zhao, Q. Meng, S. P. Li and Y. T. Wang, Simultaneous determination of ve avonoids in licorice using pressurized liquid extraction and capillary electrochromatography coupled with peak suppression diode array detection, J. Chromatogr., A, 2009, 1216, 7329– 7235. 14 W. W. Tao, J. A. Duan, J. M. Guo, J. P. Li, Y. P. Tang, P. Liu and N. Y. Yang, Simultaneous determination of triterpenoid saponins in dog plasma by a validated UPLCMS/MS and its application to a pharmacokinetic study aer administration of total saponin of licorice, J. Pharmaceut. Biomed. Anal., 2013, 75, 248–255. 15 S. F. Cui, B. Q. Fu, F. S. C. Lee and X. R. Wang, Application of microemulsion thin layer chromatography for the ngerprinting of licorice, J. Chromatogr., B: Biomed. Appl., 2005, 828, 33–40. 16 T. Bo, K. A. Li and H. W. Liu, Fast determination of avonoids in Glycyrrhizae radix by capillary zone electrophoresis, Anal. Chim. Acta, 2002, 458, 345–354. 17 Y. C. Wang and Y. S. Yang, Simultaneous quantication of avonoids and triterpenoids in licorice using HPLC, J. Chromatogr., B: Biomed. Appl., 2007, 850, 392–399. 18 S. J. Zhou, J. L. Cao, F. Qiu, W. J. Kong, S. H. Yang and M. H. Yang, Simultaneous determination of ve bioactive components in Radix Glycyrrhizae by pressurised liquid extraction combined with UPLC-PDA and UPLC/ESI-QTOFMSconrmation, Phytochem. Anal., 2013, 24, 527–533. 19 Y. Zhou, M. K. Wang, X. Liao, X. M. Zhu, S. L. Peng and L. S. Ding, Rapid identication of compounds in Glycyrrhiza Uralensis by liquid chromatography/tandem mass spectrometry, Chin. J. Anal. Chem., 2004, 32, 174– 178.


Analyst

20 J. W. Hager and J. C. Yves Le Blanc, Product ion scanning using a Q-q-Qlinear ion trap (Q TRAP™) mass spectrometer, Rapid Commun. Mass Spectrom., 2003, 17, 1056–1064. 21 C. L. Huang, B. Guo, X. Y. Wang, J. Li, W. T. Zhu, B. Chen, S. Ouyang and S. Z. Yao, A generic approach for expanding homolog-targeted residue screening of sulfonamides using a fast matrix separation and class-specic fragmentationdependent acquisition with a hybrid quadrupole-linear ion trap mass spectrometer, Anal. Chim. Acta, 2012, 737, 83–98. 22 B. Wen, L. Ma, S. D. Nelson and M. S. Zhu, High-throughput screening and characterization of reactive metabolites using polarity switching of hybrid triple quadrupole linear ion trap mass spectrometry, Anal. Chem., 2008, 80, 1788–1799. 23 W. Yang, M. Ye, M. Liu, D. Z. Kong, R. Shi, X. W. Shi, K. R. Zhang, Q. Wang and L. T. Zhang, A practical strategy for the characterization of coumarins in Radix Glehniae by liquid chromatography coupled with triple quadrupolelinear ion trap mass spectrometry, J. Chromatogr., A, 2010, 1217, 4587–4600. 24 J. G. Huang, L. Q. Si, Z. Z. Fan, L. Hu, J. Qiu and G. Li, In vitro metabolic stability and metabolite proling of TJ0711 hydrochloride, a newly developed vasodilatory b-blocker, using a liquid chromatography-tandem mass spectrometry method, J. Chromatogr., B: Biomed. Appl., 2011, 879, 3386–3392. 25 M. Z. Peng, L. Liu, M. Y. Jiang, C. L. Liang, X. Y. Zhao, Y. N. Cai, H. Y. Sheng, Z. Y. Ou and H. Luo, Measurement of free carnitine and acylcarnitines in plasma by HILICESI-MS/MS without derivatization, J. Chromatogr., B: Biomed. Appl., 2013, 932, 12–218. 26 W. J. Kong, Y. L. Zhao, X. H. Xiao, J. B. Wang, H. B. Li, Z. L. Li, C. Jin and Y. Liu, Spectrum-effect relationships between ultra performance liquid chromatography ngerprints and anti-bacterial activities of, Rhizoma coptidis, Anal. Chim. Acta, 2009, 634, 279–285. 27 W. J. Kong, J. B. Wang, Q. C. Zang, X. Y. Xing, Y. L. Zhao, W. Liu, C. Jin, Z. L. Li and X. H. Xiao, Fingerprint-efficacy study of articial Calculus bovis in quality control of Chinese materia medica, Food Chem., 2011, 127, 1342–1347. 28 US Food and Drug Administration, Analytical Procedures and Methods Validation, 2000, http://www.fda.gov/downloads/ Drugs/Guidances/ucm122858.pdf. 29 International conference on harmonization (ICH). Q2(R1): Text on validation of analytical procedures, 2005, http:// www.ich.org/leadmin/Public_Web_Site/ICH_Products/ Guidelines/Quality/Q2_R1/Step4/Q2_R1__Guideline.pdf. 30 P. J. Taylor, Matrix effects: the Achilles heel of quantitative high-performance liquid chromatography-electrospraytandem mass spectrometry, Clin. Biochem., 2005, 38, 328–334. 31 Y. R. Jin, Y. F. Du, X. W. Shi and P. W. Liu, Simultaneous quantication of 19 diterpenoids in Isodon amethystoides by high-performance liquid chromatography-electrospray ionization tandem mass spectrometry, J. Pharmaceut. Biomed. Anal., 2010, 53, 403–411. 32 A. Kruve, A. Kunnaps, K. Herodes and I. Leito, Matrix effects in pesticide multiresidue analysis by liquid chromatography-mass spectrometry, J. Chromatogr., A, 2008, 1187, 58–66.

Analyst, 2014, 139, 1883–1894 | 1893

View Article Online


Analyst

33 C. Wang and Y. Zuo, Ultrasound-assisted hydrolysis and gas chromatography- mass spectrometric determination of phenolic compounds in cranberry products, Food Chem., 2011, 128, 562–568. 34 W. J. Kong, X. Y. Xing, X. H. Xiao, J. B. Wang, Y. L. Zhao and M. H. Yang, Multi-component analysis of bile acids in natural Calculus bovis and its substitutes by ultrasoundassisted solid-liquid extraction and UPLC-ELSD, Analyst, 2012, 137, 5845–5853. 35 J. Chen, L. Yang, F. Cai, M. H. Yang and W. W. Gao, Fungi associated with Glycyrrhiza uralensis and their capability of mycotoxin production, Mycosystema, 2010, 29, 335–339.

1894 | Analyst, 2014, 139, 1883–1894

Paper

36 R. W. Wei, F. Qiu, W. J. Kong, J. H. Wei, M. H. Yang, Z. L. Luo, J. P. Qin and X. J. Ma, Co-occurrence of aatoxin B1, B2, G1, G2 and ochrotoxin A in Glycyrrhiza uralensis analyzed by HPLC-MS/MS, Food Control, 2013, 32, 216–221. 37 S. J. Zhou, J. L. Cao, W. J. Kong, M. H. Yang, A. F. Logrieco, S. H. Yang and L. Wan, Contamination effect of Aspergillus carbonarius on the inherent quality of Glycyrrhiza uralensis, Anal. Methods, 2013, 5, 5528–5533. 38 S. J. Zhou, W. J. Kong, J. L. Cao, A. F. Logrieco, S. H. Yang and M. H. Yang, Effect of Aspergillus avus contamination on the internal quality of Glycyrrhiza uralensis, World Mycotoxin J., 2014, 7, 83–89.


View Article Online

Analyst


PAPER


Analytical approaches for quantification of a Nrf2 pathway activator: overcoming bioanalytical challenges to support a toxicity study Hermes Licea Perez,*a Venkatraman Junnotula,a Dana Knecht,a Hong Nie,b Yolanda Sanchez,b Jeffrey C. Boehm,b Catherine Booth-Genthe,b Hongxing Yan,b Roderick Davisb and James F. Callahanb Activation of the Nrf2 stress pathway is known to play an important role in the defense mechanism against electrophilic and oxidative damage to biological macromolecules (DNA, lipids, and proteins). Chemical inducers of Nrf2 such as sulforaphane, dimethyl fumarate (Tecfidera®), CDDO-Me (bardoxolonemethyl), and 3-(dimethylamino)-4-((3-isothiocyanatopropyl)(methyl)amino)cyclobut-3-ene-1,2-dione (a synthetic sulforaphane analogue; will be referred to as 1) have the ability to react with Keap1 cysteine residues, leading to activation of the Antioxidant Response Element (ARE). Due to their electrophilic nature and poor matrix stability, these compounds represent great challenges when developing bioanalytical methods to evaluate in vivo exposure. 1 like SFN reacts rapidly with glutathione (GSH) and nucleophilic groups in proteins to form covalent adducts. In this work, three procedures were developed


to estimate the exposure of 1 in a non-GLP 7 day safety study in rats: (1) protein precipitation of blood samples with methanol containing the free thiol trapping reagent 4-fluoro-7-aminosulfonylbenzofurazan (ABD-F) to measure GSH- and N-acetylcysteine conjugated metabolites of 1; (2) an Edman degradation

DOI: 10.1039/c3an02216a

procedure to cleave and analyze N-terminal adducts of 1 at the valine moiety; and (3) treatment with

www.rsc.org/analyst

ammonium hydroxide to measure circulating free- and all sulfhydryl bound 1.

1. Introduction The nuclear factor-E2 related factor 2 (Nrf2) pathway (see Fig. 1) is believed to play an essential role in activating cellular defense mechanisms against oxidative stress by activation of a large set of phase II metabolizing enzymes, such as glutathione S-transferase and quinone reductase [NAD(P)H: (quinone-acceptor) oxidoreductase 1].1–3 As a key transcription factor, the levels of Nrf2 are highly regulated in cells.4 Nrf2 binds to a dimeric protein, Keap1, which targets it for Cullin-3 mediated ubiquitin degradation.3,5 Under stress conditions, the Keap1/Cul3-dependent degradation of Nrf2 is disrupted leading to accumulation of Nrf2 in the cytoplasm which is followed by translocation to the nucleus where it complexes with other proteins (such as MafK), binds Antioxidant Response Elements (AREs) of DNA and ultimately leads to induction of the phase II gene machinery.3,5 There are two approaches for activation of the Nrf2 pathway (see Fig. 1): (1) targeting the direct inhibition of Nrf2 binding to the Kelch

domain of Keap1 (ref. 6 and 7) and (2) disrupting the Keap1/ Cul3/Nrf2 complex by reversible-covalent modication of reactive cysteine residues including Cys151 in the BTB domain of Keap1.8–15 Alkylation of reactive cysteine residues including Cys151 is considered to be the mechanism of action of many natural and synthetic Nrf2 activators including sulforaphane (SFN),8–13 CDDO-Me (bardoxolone-methyl),14 and dimethyl fumarate (Tecdera®).15 This mechanism has been evaluated in drug discovery programs to target diseases such as chronic

a

Bioanalytical Science and Toxicokinetics, Platform Science and Technology, GlaxoSmithkline Pharmaceuticals, 709 Swedeland Road, King of Prussia, PA 19406, USA. E-mail: [email protected]; Tel: +1 610 270 5116

b

Respiratory Stress & Repair DPU, GlaxoSmithkline Pharmaceuticals, 709 Swedeland Road, King of Prussia, PA 19406, USA

1902 | Analyst, 2014, 139, 1902–1912

Fig. 1

Targeting Nrf2 activation.


View Article Online


Paper

Analyst

kidney disease, chronic obstructive pulmonary disease (COPD), asthma, multiple sclerosis (MS) and Parkinson's disease.16 SFN is a molecule containing an isothiocyanate functional group and is generated by the enzyme myrosinase from the natural product, glucoraphanin, when cruciferous vegetables such as broccoli, brussels sprouts or cabbages9 are damaged (e.g. by chewing). Animal experimental models suggest that SFN may have anti-cancer and antimicrobial activity9–11 and several clinical trials are in progress12,13 including a phase II trial for prostate cancer.13 SFN is a liquid and therefore its oral formulations are limited. For this reason, SFN has mostly been administered, in the form of its natural product precursor, the glucosinolate glucoraphanin, as an extract from three-day old broccoli sprouts.12 Evgen Pharma has developed a stable synthetic formulation, Sulforadex®, containing SFN as a non-cytotoxic agent against tumor proliferation and also designed to target and destroy cancer stem cells, the underlying cause of tumor recurrence and metastasis.17 The company reported successful completion of a rstin-man clinical trial; and a trial in prostate cancer patients is intended for 2014.17 3-(Dimethylamino)-4-((3-isothiocyanatopropyl)(methyl)amino)cyclobut-3-ene-1,2-dione (1, see Fig. 2) is a synthetic isothiocyanate (SFN analog) with in vitro potency similar to SFN and as a crystalline solid compound, it provides the potential for improved developability properties compared to SFN itself. Compound 1, like SFN, readily undergoes conjugation with glutathione (GSH) and nucleophilic groups in other proteins.

Fig. 2

The GSH conjugated metabolite of 1 (referred to as conjugate 2 in this publication), as many other GSH conjugates,18 is expected to be formed in the liver and further metabolized in the kidney by gamma-glutamyltransferase and dipeptidases, enzymes that catalyze the sequential removal of the glutamyl and glycyl moieties, respectively, to form a cysteine S-conjugate. Cysteine conjugate metabolites of 1 are then transported back to the liver and acetylated by intracellular N-acetyl-transferases, to form a mercapturic acid derivative or N-acetylcysteine (NAC) S-conjugate (referred to as conjugate 3 in this publication). In this work, three procedures were developed to estimate the exposure of 1 in a non-GLP 7 day safety study in rats: (1) protein precipitation of blood samples upon collection (separate aliquots) with methanol containing free thiol trapping reagent 4-uoro-7-aminosulfonylbenzofurazan (ABD-F) to measure conjugates 2 and 3; (2) an Edman degradation procedure of blood samples to analyze N-terminal adducts of 1 at the valine moiety in hemoglobin; (3) treatment of blood samples with ammonium hydroxide leading to the formation of a thiourea derivative 4 allowing analysis of “total measurement” (free 1 plus sulydryl bound 1, e.g. bound to GSH (conjugate 2), N-acetylcysteine (conjugate 3), and cysteine in hemoglobin). These methodologies enabled a successful bioanalytical support of a non-GLP 7 day rat safety study with oral doses of 1 at 3, 30 or 100 mg per kg per day. The study was successfully completed and evaluated. These procedures may be applied to similar compounds containing isothiocyanate functional groups such as SFN.

Preparation of compounds 1 to 4.


Analyst, 2014, 139, 1902–1912 | 1903

View Article Online

Analyst

2.

Experimental


2.1. Chemicals and reagents SFN ($90% (HPLC), synthetic liquid), ABD-F, ammonium formate, acetonitrile, methanol, dimethylformamide (DMF), 3,4-diethoxycyclobut-3-ene-1,2-dione dichloromethane, triethylamine, glutathione (GSH), N-acetyl-cysteine (NAC), carbon disulde (CS2), tosyl chloride (TsCl), hexane, dioxane, isopropanol, nitrophenyl isothiocyanate, triuoroacetic acid (TFA) and tert-butyl methyl ether (MTBE) were purchased from SigmaAldrich (St. Louis, MO, USA). Formic acid was purchased from Alfa Aesar (Ward Hill, MA, USA). Compounds 1–4 and the alkylated N-terminal peptide 1–VHLTPEEK were prepared by the Respiratory Stress & Repair DPU at GlaxoSmithKline (King of Prussia, USA). tert-Butyl (3-(methylamino)propyl)carbamate was obtained from Advanced ChemBlocks Inc., (Burlingame, CA, USA). Ethanol was purchased from Decon labs, Inc. (King of Prussia, PA, USA). VHLTPEEK was prepared by 21 Century Biochemicals (Marlborough, MA, USA). Rat blood was obtained from Bioreclamation Inc. (East Meadow, NY, USA).

2.2. Equipment An Eppendorf 5810R centrifuge with a rotor capacity of four 96well plates (Brinkmann Instrument, Westbury, NY, USA) and a Mettler UMX2 balance (Columbus, OH, USA) were used. ArcticWhite LLC 96-well round 2 mL polypropylene plates, ArctiSeal silicone mats with a PTFE lm (Bethlehem, PA, USA), a VWR Economy Incubator model 1500E (Radnor, PA, USA) and a Barnstead Lab Line Titer Plate Shaker (Radnor, PA, USA) were used. Waters one-milliliter plastic plates (Milford, MA, USA) along with Arctiseal mats (Bethlehem, PA, USA) were used for sample introduction into the LC-MS/MS. An ACQUITY™ UPLC integrated system from Waters (Milford, MA, USA) consisting of a sample manager combined with a sample organizer, capable of holding ten 96-deep well plates, and a binary solvent manager were used. A triple quadrupole mass spectrometer API-4000 (Applied Biosystems/MDS-Sciex, Concord, Ontario, Canada) and Waters Xevo-TQS (Waters co, MA, USA) were used. A Quadra 4 SPE™ Liquid Handling Workstation from TOMTEC (Connecticut, USA) was used for liquid–liquid extraction (LLE). The Biotage V-10™ solvent evaporation system (Biotage AB, Sweden) was used in the synthesis of 1–VHLTPEEK.

2.3. Preparation of alkylated N-terminal peptide 1–VHLTPEEK VHLTPEEK (20 mg) was dissolved in 2 mL of pyridine–water (50 : 50). The solution was warmed to 40 C and the pH was adjusted to 9.0 by adding three drops of 1 M NaOH. Then 11.18 mg 1 was added. The reaction was stirred at 40 C for 5.5 h. The solvent was evaporated using a Biotage V-10™ and puried by Gilson LC using a Sunre C18 (19 mm 100 mm; 5 mm). The mobile phase composition was 0.1% TFA in water (A) and 0.1% TFA in acetonitrile (B). The separation was performed using a linear gradient from 10% B to 50% B in 15 min and the ow rate was kept at 18 mL min1. The desired product was

1904 | Analyst, 2014, 139, 1902–1912

Paper

characterized by LC-MS (m/z 1206.4 [M + H]+) and the overall yield was 9.3 mg (24.5%). 2.4. Preparation of compounds 1 to 4 Compound 1 was prepared in four steps in 34% overall yield from the commercially available 3,4-diethoxycyclobut-3-ene-1,2dione and tert-butyl (3-(methylamino)propyl)carbamate as depicted in Fig. 2. Treatment of 1 with GSH and in the presence of sodium hydroxide (pH 7.8) provides conjugate 2. Treatment of compound 1 with NAC in the presence of sodium hydroxide (pH 7.8) provides conjugate 3. Treatment of 1 with ammonia provides compound 4 as described in Fig. 2. These compounds were characterized by 1H NMR and the following characteristic data were recorded. 1H NMR (400 MHz, DMSO-d6) d 3.60–3.95 (m, 4H), 3.27–3.41 (m, 3H), 3.21 (s, 6H), 1.75–2.17 (m, 2H) for 1; 1 H NMR (400 MHz, deuterium oxide) d 2.02–2.14 (m, 4H), 2.46 (m, 2H), 3.18 (s, 3H), 3.22–3.27 (m, 6H), 3.50 (m, 1H), 3.68–3.89 (m, 8H), for 2; 1H NMR (400 MHz, deuterium oxide) d 1.97 (s, 3H), 2.04 (m, 2H), 3.18 (s, 3H), 3.24 (s, 6H), 3.51 (m, 1H), 3.70– 3.89 (m, 5H) for 3; and 1H NMR (400 MHz, DMSO-d6) d 7.60 (br. s., 1H), 6.97 (br. s., 1H), 3.67 (br. s., 2H), 3.38 (br. s., 2H), 3.21 (s, 6H), 3.13 (s, 3H), 1.81 (br. s., 2H) for 4. 2.5. Internal standard (hemoglobin adduct of SFN) preparation An aliquot of SFN (100 mL of 1 mgmL1 acetonitrile) was added to 1.9 mL fresh rat blood. The above solution was incubated at 37 C for 24 hours under constant and gentle mixing. Aer incubation, the solution was diluted 10-fold with water and used as the internal standard. 2.6. Analysis of conjugates 2 and 3 Stock solutions of conjugates 2 and 3 were individually prepared in 50/50 (v/v) acetonitrile–water solution at a concentration of 0.5 mg mL1. These stock solutions were stored at 4 C. Separate working solutions (WSs) at 0.5 mM were prepared fresh on the day of analysis for 3 (WS-A) and 2 (WS-B), respectively, in 10 mM ammonium acetate (native pH). Aliquots of each of the working solutions (50 mL) were combined and added to 150 mL of 10 mM ammonium acetate (native pH) to make the working solution (WS1) containing 3 and 2 conjugates at a concentration of 100 mM. Two additional working solutions (WS2 and WS3) at 10 and 1 mM were prepared by diluting WS1 10- and 100- fold, respectively, with 10 mM ammonium acetate (native pH). These working solutions were used to make calibration standards in rat blood containing both analytes at 10 000, 5000, 2500, 1000, 500, 250, 100, 50, 25, and 10 nM 3 and 2 conjugates in rat blood. The quality control (QC) samples were prepared in rat blood at 8000, 400, and 30 nM 3 and 2 conjugates. Due to the limited stability of these analytes the samples were processed immediately following preparation as described below. Aliquots (50 mL) of calibration and QC samples were transferred to uniquely labeled MATRIX Screwtop Trakmate™ tubes (1.4 mL) with Screwtop Trakmate caps. An aliquot (400 mL) of internal standard solution (100 ng mL1 of N-acetylcysteine


View Article Online


Paper

conjugate of SFN) in a cold methanol solution containing ABD-F at 1 mg mL1 was added to all tubes with the exception of the blanks, which instead received 400 mL of 1 mg mL1 ABD-F solution in methanol. The tubes were capped and vortex-mixed for approximately 3 min. Aer vortex-mixing, the tubes were centrifuged for approximately 5 min at approximately 1000 g. Aer centrifugation, 350 mL of the supernatant was transferred to uniquely-labeled 1.4 mL MATRIX Screwtop Trakmate™ tubes with Screwtop Trakmate™ caps, and 20 mL of 10% formic acid in water was added to all tubes and briey vortex-mixed. The samples were analyzed by LC-MS/MS. 2.7. LC-MS/MS analysis of 3 and 2 conjugates The analytical column used was an Imtakt Cadenza CD-C18 50 mm 2 mm, with 3 mm particle size. The column temperature was held at 30 C and the sample compartment was at ambient temperature. Mobile phase A consisted of 0.1% formic acid in water and acetonitrile was used as mobile phase B. Mobile phase B was held at 10% until 0.3 min, followed by a linear gradient from 10% B to 80% B for 2.0 min, aer which the system was returned to the initial conditions. The total run time, including sample loading, was approximately 3.0 min and the ow rate was maintained at 0.7 mL min1 throughout the run. A typical injection volume of 4 mL in a 10 mL loop (partial loop injection mode) of a Waters® ACQUITY UPLC® H-Class System (SM-FTN) was used. The UPLC weak wash was a solution of 10/90 acetonitrile–water (v/v) and the strong wash consisted of 0.1% formic acid in a solution of 40 : 40 : 20 acetonitrile : isopropanol : water (v/v). A Sciex API-4000 with a TurboIonspray (TIS) interface was operated in positive ionization mode. The instrument was optimized for conjugates 2, 3 and the internal standard by infusing a 10 ng mL1 solution of the analytes in acetonitrile– water (50/50 v/v) at 0.5 mL min1 through an auxiliary Agilent pump 1100 series (Palo Alto, CA, USA) directly connected to the mass spectrometer. The MRM transitions of m/z 417 / 254, m/z 561 / 254, and 341 / 114 were chosen for conjugates 2, 3, and the internal standard, respectively. Dwell times of 125 ms were used for the analytes and the internal standard. The optimized mass spectrometric conditions included the following MS conditions: TIS source temperature, 650 C; TIS voltage, 5500 V; curtain gas, 40 psi (nitrogen); nebulizer gas (GS1), 80 psi (zero air); turbo gas (GS2), 80 psi (zero air); collision energy, 42 eV for 2 and 25 eV for 3; declustering potential 60 eV. 2.8. Analysis of valine adducts in hemoglobin Stock solutions of 1–VHLTPEEK were prepared in water at a concentration of 1 mM. These stock solutions were stored at 4 C. Additional working solutions of 1–VHLTPEEK at 100, 20, and 4 mM were prepared fresh on the day of analysis in rat blood. These solutions were used to make calibration standards in rat blood at 100, 50, 20, 10, 4, 2, 1, 0.4, 0.2 and 0.1 mM 1–VHLTPEEK in rat blood. The QC samples were prepared in rat blood at 80, 5, and 0.3 mM 1–VHLTPEEK. Aliquots of 20 mL of calibration and QC blood samples were transferred to uniquely labeled MATRIX Screwtop Trakmate™


Analyst

tubes (1.4 mL) with Screwtop Trakmate caps. A 20 mL aliquot of internal standard solution (alkylated blood with SFN, see Section 2.5) was added to all wells with the exception of the blanks, which instead received 20 mL of water only. A 200 mL aliquot of methanol was added to all samples to precipitate blood proteins. The tubes were capped and vortex-mixed for approximately 3 min. Aer vortex-mixing, the samples were centrifuged for approximately 5 min at approximately 3000 g to concentrate the samples on the bottom of the tube. The samples (containing the supernatant and precipitate) were dried down using a steady stream of nitrogen at 45 C. A 100 mL aliquot of concentrated TFA was then added to all dried samples followed by incubation at 65 C for 20 min under constant vortex mixing. Aer incubation, the samples were dried using a steady stream of nitrogen at 45 C to remove TFA. An aliquot of acetonitrile (400 mL) was added to all tubes and followed by vortex-mixing of samples for approximately 5 min. The samples were then centrifuged for approximately 5 min to remove the solid residue at approximately 3000 g and 300 mL of supernatant was transferred to clean tubes. The samples were analyzed by LC-MS/MS. 2.9. LC MS/MS analysis of valine adducts in hemoglobin The analytical column used was an ACQUITY HSS-T3, 2.1 mm 50 mm with 1.8 mm particle size from Waters Co. The column temperature was held at 65 C and the sample compartment at 4 C. Mobile phase A consisted of 0.1% formic acid in water and mobile phase B was acetonitrile. Mobile phase B was held at 5% until 0.1 min, followed by a linear gradient from 5% B to 45% B for 1.2 min and then a steep gradient from 45% B to 95% B from 1.2 min to 1.21 min. The system was held at 95% B until 1.70 min to remove late eluting substances from the analytical column, aer which the system was returned to the initial conditions. The total run time, including sample loading, was approximately 2.0 min and the ow rate was maintained at 0.7 mL min1 throughout the run. A typical injection volume of 1 mL in a Waters® ACQUITY UPLC® I-Class System (SM-FTN) was used. A Waters Xevo-TQS (Waters Co., MA, USA) with a TIS interface was operated in the positive ionization mode. The instrument was optimized for Edman derivatives of 1–valine and sulforaphane–valine by infusing the corresponding derivatized solutions at 2 ng mL1 in acetonitrile–water (50/50, v/v) at a ow rate of 300 mL min1 through an Agilent pump 1100 series (Palo Alto, CA, USA). The MRM transitions of m/z 353 / 254 and m/z 277 / 178 were used for the Edman derivatives of 1–valine and sulforaphane–valine, respectively. The optimized mass spectrometric conditions for both analytes were used as follows: desolvation temperature, 600 C; source temperature, 150 C; cone voltage, 5 V; cone gas ow 150 L h1; desolvation gas ow, 900 L h1; nebulizer gas, 7 bar; and collision energy, 20 eV. 2.10. Total measurement (1 bound to the sulydryl moiety and free fraction of 1) Stock solutions of 1 were prepared in water at a concentration of 3.95 mM. These stock solutions were stored at 4 C. Additional

Analyst, 2014, 139, 1902–1912 | 1905

View Article Online


Analyst

working solutions of 1 at 100, 20, and 4 mM were prepared fresh on the day of analysis in rat blood. The WSs were used to make calibration standards in rat blood at 100, 50, 20, 10, 4, 2, 1, 0.4, 0.2, and 0.1 mM 1. The QC samples were prepared in rat blood at 80, 5, and 0.3 mM 1. MTBE (0.5 mL) was added to each well of the 2 mL ArcticWhite 96-well PTFE coated plate. The plate was sealed with the ArcticSeal PTFE coated mat and vortex-mixed in an inverted position for approximately 3 min. Subsequently, the MTBE was discarded and the plate was le to dry in a chemical hood. This wash step was used to remove any plastic residue from the plates and plate seals. Aliquots (20 mL) of calibration and QC blood samples were transferred to the washed ArcticWhite 96-well plate. A 120 mL aliquot of internal standard solution (freshly prepared 1 mg mL1 of nitrophenyl isothiocyanate in ammonium hydroxide) was added to all wells with the exception of the blanks, which instead received 120 mL of concentrated ammonium hydroxide containing no internal standard solution. The plate was sealed with the ArcticSeal mat and vortexmixed for approximately 2 min. Aer vortex-mixing, the plate was incubated for approximately 20 min at approximately 65 C. Aer incubation, 1 mL ethyl acetate was added to each well and vortex-mixed for 3 min. The plate was then centrifuged for approximately 5 min at approximately 3000 g. An aliquot (50 mL) of DMF was added to clean silanized glass inserts (into the 2 mL 96-well collection plate). Using a liquid handler (Tomtech), 825 mL of the ethyl acetate supernatant was transferred to silanized glass inserts containing 50 mL of DMF. The ethyl acetate residue was removed using a steady stream of nitrogen at 45 C approximately to the level corresponding to 50 mL DMF. It is recommended to stop the drying process before DMF is completely evaporated. An aliquot of a 100 mL of water– acetonitrile solution (1/1; v/v) was then added to each well. The plate with glass inserts was capped and vortex-mixed for approximately 1 min and then centrifuged for 2 min at approximately 1000 g. The samples were analyzed by LC-MS/MS.

2.11. LC-MS/MS analysis of compound 4 (1-thiourea derivative) The analytical column used was an ACQUITY HSS-T3, 2.1 mm 50 mm with 1.8 mm particle size from Waters Co. The column temperature was held at 65 C and the sample compartment was at 4 C. Mobile phase A consisted of 0.1% formic acid in water and acetonitrile was used as mobile phase B. Mobile phase B was held at 10% until 0.1 min, followed by a linear gradient from 10% B to 45% B for 1.2 min and then a steep gradient from 45% B to 95% B from 1.2 min to 1.21 min. The system was held at 95% B until 1.70 min to remove late eluting substances from the column, aer which the system was returned to the initial conditions. The total run time, including sample loading, was approximately 2.0 min and the ow rate was maintained at 0.7 mL min1 throughout the run. A typical injection volume of 1 mL in a Waters® ACQUITY UPLC® I-Class System (SM-FTN) was used. A Waters Xevo-TQS (Waters Co., MA, USA) with a TIS interface was operated in the positive ionization

1906 | Analyst, 2014, 139, 1902–1912

Paper

mode. The instrument was optimized for the thiourea derivatives of 1 and nitrophenyl isothiocyanate (internal standard), by infusing the corresponding derivatized solutions at 2 ng mL1 in acetonitrile–water (50/50, v/v) at a ow rate of 300 mL min1 through an Agilent pump 1100 series (Palo Alto, CA, USA). The MRM transitions of m/z 271 / 254 and m/z 198 / 135 were chosen for the thiourea derivatives of 1 and nitrophenyl isothiocyanate, respectively. The optimized mass spectrometric conditions for both analytes were used as follows: desolvation temperature, 600 C; source temperature 150 C, cone voltage, 5 V; cone gas ow 150 L h1; desolvation gas ow, 900 L h1; nebulizer gas, 7 bar; collision energy for the thiourea derivatives of 1 and nitrophenyl isothiocyanate were 20 eV and 30 eV, respectively.

2.12.

Toxicity study

A 7 day oral toxicity study was conducted at the GSK Safety Assessment Facility in accordance with the GSK Policy on the Care, Welfare and Treatment of Laboratory Animals and was reviewed by the Institutional Animal Care and Use Committee at GSK. The objective of the study was to characterize the toxicity and toxicokinetics (TK) of 1 following oral administration to male Crl:WI(Hans) rats [Charles River Laboratories, Inc., Stone Ridge, NY] for 7 days. Rats were group-housed (up to 3 rats per cage) in clear plastic cages with ALPHA-dri™ bedding (Shepherd Specialty Papers, Inc., Kalamazoo, MI). The rats were in the range of 10 to 12 weeks of age and weighed approximately 200 to 400 g at the initiation of dosing. Rats were offered 5002 Certied Rodent Diet (PMI Nutrition International, Richmond, IN) ad libitum. Filtered tap water (supplied by Aqua Pennsylvania, Inc. and periodically analyzed) was available ad libitum from an automatic watering system. Compound 1 was administered orally, once daily for up to 7 days at 3, 30 or 100 mg per kg per day. The vehicle used was 0.5% hydroxpropyl methylcellulose (K15M with 0.1% (w/v) polyoxyethylene sorbitan monooleate 80™). A serial sampling scheme was used (3 animals per dose group) at 0.25, 0.5, 1, 4, 8, and 24 h aer dosing on days 1 and 7. The same animals provided samples on days 1 and 7. Approximately 0.2 mL blood was collected from each rat per time point in uniquely labelled tubes containing EDTA and split into two aliquots (aliquots A and B). For aliquotA, a 50 mL sample of whole blood was used for analysis of 2 and 3. The remaining blood, aliquot-B was used for analysis of total measurement and hemoglobin valine adducts. Aliquot A was precipitated immediately upon collection with internal standard solution (N-acetylcysteine conjugate of SFN) in a solution of methanol containing ABD-F and further processed in an identical manner to the calibration and QC samples.

2.13.

Data analysis

MS data were acquired and processed (integrated) using the proprietary soware application Analyst™ (Version 1.4.2, Applied Biosystems/MDS-Sciex, Canada) for analysis of conjugates 2 and 3 and MassLynx 4.1 (Waters Co., MA, USA) for analysis of the Edman- and thiourea-derivatives.


View Article Online


Paper

Analyst

Calibration plots of the analyte/internal standard peak area ratio versus analyte concentrations were constructed and a weighted 1/x2 linear regression was used. Concentrations of the investigated analytes in validation samples were determined from the appropriate calibration line and used to calculate the bias and precision of the method with an in-house LIMS (Study Management System, SMS2000, version 2.3, GlaxoSmithKline). Toxicokinetic analysis was performed by noncompartmental pharmacokinetic analysis using WinNonlin™ (WNL), Version 6.1 (Pharsight Corp., USA).

3.


3.1. Challenges during method development The objective was to develop rugged and sensitive LC-MS/MS methods allowing determination of the exposure of 1 in a rat safety assessment study. The data generated were crucial for decision making in the further progression of 1 into development. The major challenge faced during method development was associated with the poor stability and low recovery of 1 in rat blood. Approximately 90% loss of 1 was observed when blood was fortied and then immediately extracted using protein precipitation. Compound 1 reacts rapidly with GSH and nucleophilic groups in proteins (adducts), making it challenging to develop a method for determination of free 1 in blood. All attempts to establish a method for quantication of free 1 in rat blood were unsuccessful, with the results showing lack of reproducibility indicating that analysis of free 1 would be unreliable. Since 1 like SFN contains an isothiocyanate functional group, it was decided to evaluate the method available for determination of SFN in biological matrices. Ye et al.19 described a method for total analysis of isothiocyanates and dithiocarbamates using a reaction with 1,2-benzenedithiol to produce 1,3-benzodithiole-2-thione that can be quantied through photodiode array integration. Although this cyclocondensation reaction has been highly useful for analyzing plant material and urine samples, the determination of isothiocyanates in biological matrices has been limited by assay sensitivity and selectivity. Our efforts on developing a similar method for analysis of 1 using this reagent and LC-MS/MS as a detector were also unsuccessful. Due to the importance of having methodologies available to support a rodent non-GLP toxicology study, numerous unconventional bioanalytical approaches were evaluated to estimate the exposure of 1 in rodents.

to react with GSH and cysteine in proteins. It is important to clarify that ABD-F is used in this procedure to alkylate free thiols (thiol scavengers) in rat blood to prevent thiol exchange reactions and it does not form any derivative with conjugates 2 and 3. In addition, formic acid was used to stabilize these conjugates during storage in the autosampler. Based on QC data analysis, conjugates 2 and 3 are stable in the described process extract solution for at least 24 hours at ambient temperature.

3.3. Edman degradation procedure to cleave and analyze N-terminal adducts of 1 Another approach evaluated during method development in the intention to assess exposure of 1 in the safety study utilized the Edman degradation procedure20 to determine adducts formed on the N-terminal valine. The Edman degradation procedure is typically used to determine the peptide amino acid sequence from the N-terminus (see Fig. 3). This technique employs the reaction of the N-terminal a-amino group with phenyl isothiocyanate at a slightly basic pH to give a phenylthiocarbamyl (PTC) derivative.20 The PTC derivative is then treated with a strong acid (e.g., TFA) to cleave the peptide at the last peptide bond leading to the formation of a phenylthiohydantoin (see Fig. 3). Compound 1 is expected to react with N-terminal valine in hemoglobin in a similar way, and if treated with TFA a thiohydantoin derivative is expected to be formed thus allowing determination of the fraction of 1 that reacted with this moiety in hemoglobin. This is in principle the same technique that has been used for determination of exposure of many reactive industrial chemicals.21–23 In these cases, industrial chemicals and/or their metabolites react with N-terminal valine in hemoglobin and a modied Edman reagent pentauorophenyl isothiocyanate is used to cleave the valine adduct and subsequently analyzed by LC-MS/MS or GC-MS/MS.23

3.2. Analysis of conjugates 2 and 3 Considering that conjugate 2 exists in equilibrium with free 1, in order to avoid any potential shi in the equilibrium during sample handling, it was decided to process blood samples immediately upon collection. Conjugates 2 and 3 were analyzed the same day of sample collection by protein precipitation using ice-cold methanol solution. It was observed that the recovery and consistency of the assay were improved by adding ABD-F at 1 mg mL1 to the methanol solution. ABD-F is well documented


Application of the Edman degradation procedure for analysis of 1 in rat blood.

Fig. 3

Analyst, 2014, 139, 1902–1912 | 1907

View Article Online


Analyst

The cleavage of adducted valine with 1 was optimized (see Section 2.8) for a 20 mL aliquot rat blood. Since the reaction requires an aqueous free environment, methanol was used to precipitate blood proteins facilitating the drying process. A lower recovery was observed when separating the supernatant from the solid material and therefore the TFA treatment was performed aer drying the whole fraction containing both the solid material and the supernatant. A 100 mL aliquot of concentrated TFA was used to dissolve the extract and to cleave the valine adduct. Incubation at 65 C for 20 min was optimum for the cleavage to be performed. Aer incubation, the samples were dried down to remove TFA and re-dissolved in acetonitrile. A centrifugation step was employed to remove some solid residue before LC-MS/MS analysis. An adducted peptide “1–VHLTPEEK” with an amino acid sequence the same as for the eight N-terminal residues in hemoglobin was used as the reference standard for preparation of calibration standards and QC samples for this assay. This approach has been used before for analysis of acrylamine adducts of hemoglobin.24 A similar peptide alkylated with labeled 1 (13C- or deuterium labeled) would be an ideal internal standard for this assay; however due to time constraints in supporting the safety study this synthesis was not prioritized. Instead, alkylated hemoglobin with SFN was used as internal standard. SFN was allowed to react with blood proteins to form adducts with all nucleophilic groups in blood including valine adducts. Valine adducts are known to be stable and the Edman procedure is reproducible;21–24 therefore a treated solution may be used as the internal standard. The approach has been used in the past for determination of hemoglobin adducts of industrial chemicals due to difficulties with the synthesis of alkylated hemoglobin.25,26 The Edman procedure described in this work may be applied to similar compounds containing isothiocyanate functional groups such as SFN.

3.4. Total measurement (bound to the sulydryl moiety and free fraction of 1) Since 1 is expected to undergo an equilibrium reaction with GSH and cysteine in haemoglobin, we sought to shi the equilibrium towards the formation of free 1 by changing the solution pH to basic. An aliquot of ammonium hydroxide was added to a neat solution containing equal quantities of 1, and the two conjugates, 2 and 3. The solution was injected repeatedly into the LC-MS operated in full scan mode. The levels of free 1 did not increase as expected; however, a new peak increased with a retention time close to that for 2 (see Fig. 4). Aer an hour of treatment, only traces of the investigated analytes were observed and a peak with m/z ¼ 271 corresponding to the thiourea derivative of 1, compound 4, was detected. This observation led to the idea of using ammonium hydroxide for quantication of 1 in its free form plus the fraction bound to the sulydryl moiety in hemoglobin (this fraction is referred to as the total measurement). To enable this, it was necessary to estimate the recovery of thiourea derivatives when these analytes are fortied into blood samples. For that reason, it was decided to synthesize the thiourea derivative from 1. Addition of

1908 | Analyst, 2014, 139, 1902–1912

Paper

Fig. 4 Formation of the thiourea derivative from 1, 2 and 3, after treatment with ammonium hydroxide. LC-MS total ion chromatograms (TICs) at time t ¼ 0 h (top chromatogram) and after storage in the autosampler for an hour (bottom chromatogram) under ambient conditions.

ammonium hydroxide to blood samples, containing 1, and the conjugates, 2 and 3, converted all sulydryl conjugates and the free 1 into the thiourea derivative 4 to a high extent (see Fig. 5). As mentioned earlier, approximately 90% of 1 is lost when rat blood is fortied with this compound. The treatment with ammonium hydroxide recovers 1 in the form of a thiourea derivative to a high extent (in molar quantities). For that reason, 1 is added to the preparation of calibration standards for quantication of total measurement (bound to the sulydryl moiety and free fraction of 1). The thiourea derivative 4 was extracted by LLE with ethyl acetate before LC-MS/MS. It is important to mention that this thiourea derivative is also lost during the drying process. To overcome this issue, an aliquot (50 mL) of DMF was added to collection silanized glass inserts (into the 2 mL 96-well collection plate). The ethyl acetate phase was then evaporated to approximately the level corresponding to 50 mL DMF to avoid complete drying. The formation of the thiourea derivative from SFN was also observed to a high extent, therefore it was initially evaluated as the internal standard for the analysis of total 1 assay; however, this derivative had a shorter retention time than that for 4. For that reason, nitrophenyl isothiocyanate had a similar retention time to 4 and was used as the internal standard. 3.5. Summary of TK data from the 7 day rat safety study The methodologies described above enabled successful bioanalytical support of a non-GLP 7 day rat safety study with doses


View Article Online


Paper

Analyst

Fig. 5 Recovery of the thiourea derivative 4 from blood samples fortified with 1, 2 and 3 after treatment with ammonium hydroxide. Formation of conjugate 3 is a result of inter-organ synthesis by gamma-glutamyltransferase and dipeptidases in the kidney and acetylation by intracellular N-acetyl-transferases in the liver.

Table 1

Toxicokinetic parameters for conjugate 2 Male (n ¼ 3) Dose (mg per kg per day)

Parameter

Period

3

30

AUC0–t (h mM) mean [range]

Day 1 Day 7 Day 1 Day 7

0.969 [0.857–1.18] 0.623 [0.55–0.713] 0.511 [0.309–0.717] 0.335 [0.229–0.522]

13.9 7.74 4.54 2.19

Cmax (mM) mean [range]

Table 2

100 [11.1–15.8] [5.75–9.03] [2.76–6.05] [1.77–2.69]

33.0 [28.5–38.4] 15.4 [11.1–18.4] 9.75 [7.48–11.4] 2.38 [1.14–3.21]

Toxicokinetic parameters for conjugate 3 Male (n ¼ 3) Dose (mg per kg per day)

Parameter

Period

3

30

100



0.310 [0.216–0.437] 0.950 [0.102–2.27] 0.095 [0.064–0.118] 0.101 [0.037–0.049]

3.34 [1.53–6.60] 1.40 [0.775–2.57] 0.419 [0.144–0.966] 0.175 [0.063–0.275]

3.40 [2.20–4.12] 2.63 [0.778–5.45] 0.379 [0.295–0.517] 0.285 [0.117–0.590]


Table 3

Toxicokinetic parameters for the hemoglobin valine adduct of 1 Male (n ¼ 3) Dose (mg per kg per day)

Parameter

Period

3

30

100



18.1 [15.7–20.9] 50.5 [49.3–51.5] 0.982 [0.908–1.10] 2.44 [2.36–2.55]

317 [308–327] 567 [545–578] 17.1 [15.4–20.0] 29.0 [26.3–31.0]

730 [573–914] 1140 [1090–1230] 46.7 [34.4–67.1] 50.9 [45.9–56.8]



Analyst, 2014, 139, 1902–1912 | 1909

View Article Online

Analyst

Paper

Table 4 Toxicokinetic parameters for the total measurement of 1 bound to the sulfhydryl moiety and free fraction of 1

Male (n ¼ 3) Dose (mg per kg per day) Parameter

Period

3



16.0 24.6 5.74 6.54



Fig. 6

[15.7–16.3] [22.0–27.6] [5.32–5.97] [4.88–7.82]

30

100

375 [350–388] 345 [305–398] 91.9 [78.9–98.8] 71.8 [41.0–89.3]

698 [517–917] 611 [530–677] 201 [129–295] 59.8 [57.0–64.1]

Concentration profiles of conjugates 2 and 3 for days 1 and 7 after oral administration of 1.

Fig. 7 Concentration profiles of valine adducts and total measurement for days 1 and 7 after oral administration of 1.

1910 | Analyst, 2014, 139, 1902–1912


View Article Online


Paper

of 3, 30 or 100 mg per kg per day with 1. Analysis of free 1 was not performed due to poor stability of the compound and high variability in the assay. Alternatively, the total measurement, hemoglobin valine adduct of 1, and adducts 2 and 3 were used as surrogate markers for evaluation of the systemic exposure of 1. Following oral administration of 1 at doses of 3, 30 or 100 mg per kg per day for 7 days to the rats, total measurement and valine adduct concentrations were quantiable during the entire 24 hour sampling period aer dosing on days 1 and 7 except the 3 mg kg1 day 1 dose group with total measurement concentrations quantiable for up to 8 h post dose. Concentrations of conjugates 2 and 3 were also quantiable for up to 8 h post dose for all dose levels on days 1 and 7 except the 3 mg kg1 day 7 dose group with concentrations quantiable for up to 4 h post dose. The maximum blood concentrations aer dosing were observed at 0.25 to 1 h for total measurement, 0.5 to 8.0 h for valine adduct, 0.25 to 8.0 h for conjugate 3 and 0.25 to 0.5 h for conjugate 2. A summary of TK parameters and blood concentration proles on days 1 and 7 is presented in Tables 1–4 and shown graphically in Fig. 6 and 7. Systemic exposure (AUC0–t and Cmax values) based on the total measurement increased the dose proportionally from 3 to 100 mg per kg per day. For the 33.3-fold increase in dose, the mean AUC0–t and Cmax values increased 43.6- and 35.0-fold, respectively, on day 1, and 24.8- and 9.1-fold, respectively, on day 7. Following 7 days of repeated dosing, there was no marked (>2-fold) change in the systemic exposure of total measurement from day 1 to day 7 for any of the dose levels, except for the mean Cmax values at 100 mg per kg per day, which decreased 3.4-fold from day 1 to day 7. The systemic exposure based on the data from valine adducts in hemoglobin was also pronounced and the AUC0–t and Cmax values increased proportionally with dose from 3 to 100 mg per kg per day. For the 33.3-fold increase in dose, the mean AUC0–t and Cmax values increased 40.3- and 47.6-fold, respectively, on day 1 and 22.6- and 20.9-fold, respectively, on day 7. On both days 1 and 7, the concentrations of valine adduct generally were constant throughout the 24 hour sampling period. There was a trend toward higher systemic exposure on day 7 than that on day 1. AUC0–t ratios of total measurement/ valine adducts ranged from 0.780 to 1.24 on day 1 and 0.431 to 0.690 on day 7. AUC0–t ratios of total measurement/conjugate 3 and total measurement/conjugate 2 were variable and overlapped between day 1 and day 7, ranging from 12.2 to 681 and 13.6 to 56.5, respectively.

4. Conclusion A multiple component semi-automated sample preparation methodology in a 96-well plate format for determination of in vivo exposure of 1 in rat blood was developed. These methodologies enabled a successful bioanalytical support of a non-GLP 7 day rat safety study with oral doses of 1 at 3, 30, or 100 mg per kg per day. Even though all three methodologies described in this publication provided good understanding on the in vivo exposure of 1, the total measurement methodology is advised for assessing the exposure of similar compounds bearing an


Analyst

isothiocyanate group due to the following reasons: (1) the simplicity of the methodology and (2) these compounds most likely exist mainly as sulydryl conjugates in vivo, which are well recovered with the total measurement methodology.

Acknowledgements The authors would like to acknowledge Lee Abberley and Christopher Evans from the department of Drug Metabolism and Pharmacokinetics at GlaxoSmithKline for their review of this work.

References 1 K. Itoh, T. Chiba, S. Takahashi, T. Ishii, K. Igarashi, Y. Katoh, T. Oyake, N. Hayashi, K. Satoh, I. Hatayama, M. Yamamoto and Y.-I. Nabeshima, Biochem. Biophys. Res. Commun., 1997, 236, 313. 2 E. Kansanen, H.-K. Jyrkkanen and A.-L. Levonen, Free Radicals Biol. Med., 2012, 52, 973. 3 E. Kansanen, S. M. Kuosmanen, H. Leinonen and A.-L. Levonenn, Redox Biol., 2013, 1, 45. 4 N. Li, J. Alam, M. I. Venkatesan, A. Eiguren-Fernandez, D. Schmitz, E. Di Stefano, N. Slaughter, E. Killeen, X. Wang, A. Huang, M. Wang, A. H. Miguel, A. Cho, C. Sioutas and A. E. Nel, J. Immunol., 2004, 173, 3467. 5 Z. Sun, S. Zhang, J. Y. Chan and D. D. Zhang, Mol. Cell. Biol., 2007, 27, 6334. 6 L. Hu, S. Magesh, L. Chen, L. Wang, T. A. Lewis, Y. Chen, C. Khodier, D. Inoyama, L. J. Beamer, T. J. Emge, J. Shen, J. E. Kerrigan, A. T. Kong, S. Dandapani, M. Palmer, S. L. Schreiber and B. Munoz, Bioorg. Med. Chem. Lett., 2013, 23, 3039. 7 D. Marcotte, W. Zeng, J.-C. Hus, A. McKenzie, C. Hession, P. Jin, C. Bergeron, A. Lugovskoy, I. Enyedy, H. Cuervo, D. Wanga, C. Atmanene, D. Roecklin, M. Vecchi, V. Vivat, J. Kraemer, D. Winkler, V. Hong, J. Chao, M. Lukashev and L. Silvian, Bioorg. Med. Chem., 2013, 21, 4011. 8 T. W. Kensler, T. J. Curphey, Y. Maxiutenko and B. D. Roebuck, Drug Metab. Drug Interact., 2000, 17, 3. 9 A. Yanaka, J. W. Fahey, A. Fukumoto, M. Nakayama, S. Inoue, S. Zhang, M. Tauchi, H. Suzuki, I. Hyodo and M. Yamamoto, Cancer Prev. Res., 2009, 2, 353. 10 J. W. Fahey, X. Haristoy, P. M. Dolan, T. W. Kensler, I. Scholtus, K. K. Stephenson, P. Talalay and A. Lozniewski, Proc. Natl. Acad. Sci. U. S. A., 2002, 99, 7610. 11 M. V. Galan, A. A. Kishan and A. L. Silverman, Dig. Dis. Sci., 2004, 49, 1088. 12 http://clinicaltrials.gov/ct2/results?term¼Sulforaphane. 13 http://clinicaltrials.gov/ct2/show/NCT01228084. 14 http://www.clinicaltrials.gov/show/NCT01351675. 15 M. Meissner, E. M. Valesky, S. Kippenberger and R. Kaufmann, JDDG: Journal der Deutschen Dermatologischen Gesellscha, 2012, 10, 793. 16 A. J. Wilson, J. K. Kerns, J. F. Callahan and C. J. Moody, J. Med. Chem., 2013, 56, 7463. 17 http://www.evgen.com/news/Evgen-ASCO-2013-Newsletter.pdf.

Analyst, 2014, 139, 1902–1912 | 1911

View Article Online

Analyst

23 M. T¨ ornqvist, J. Mowrer, S. Jensen and L. Ehrenberg, Anal. Biochem., 1986, 154, 255. 24 H. W. Vesper, M. Ospina, T. Tunde, L. Ingham, A. Smith, J. G. Gray and G. L. Myers, Rapid Commun. Mass Spectrom., 2006, 20, 959. 25 A. Kautiainen and M. T¨ ornqvist, Int. Arch. Occup. Environ. Health, 1991, 63, 27. 26 H. L. P´ erez, D. Segerb¨ ack and S. Osterman-Golkar, Chem. Res. Toxicol., 1999, 12, 869.


18 C. A. Hinchman and N. Ballatori, J. Toxicol. Environ. Health, 1994, 41, 387. 19 L. Ye, A. T. Dinkova-Kostova, K. L. Wade, Y. Zhang, T. A. Shapiro and P. Talalay, Clin. Chim. Acta, 2002, 316, 43. 20 P. Edman and G. Begg, Eur. J. Biochem., 1967, 1, 8. 21 S. Osterman-Golkar, L. Ehrenberg, D. Segerb¨ ack and I. H¨ allstr¨ om, Mutat. Res., 1976, 34, 1. 22 L. Ehrenberg and S. Osterman-Golkar, Teratog., Carcinog., Mutagen., 1980, 1, 105.

Paper

1912 | Analyst, 2014, 139, 1902–1912


View Article Online

Analyst


PAPER


A novel fluorescent ‘off-on-off’ probe for relay recognition of Zn2+ and Cu2+ derived from N,Nbis(2-pyridylmethyl)amine† Yameng Liu, Qiang Fei, Hongyan Shan, Minghui Cui, Qing Liu, Guodong Feng and Yanfu Huan* A metal-to-metal relay recognition was discovered with sequence specificity via a fluorescence ‘off-on-off’ phenomenon. We present here the synthesis of a new N,N-bis(2-pyridylmethyl)amine-based fluorescent sensor termed NBPA and its application as a selective relay probe for sensing of free trace zinc and copper ions. The addition of Zn2+ to NBPA causes a drastic enhancement of fluorescent intensity by the formation of a 1 : 1 NBPA–Zn2+ complex whereas other cations have no response. Moreover, Cu2+ can induce fluorescence quenching in the NBPA–Zn2+ system by the formation of a 1 : 1 NBPA–Cu2+ complex confirmed by Job's Plot. The signal change of the sensor is based on the chelation-enhanced fluorescence (CHEF) effect of NBPA–Zn2+ with the inhibition of the photoinduced electron transfer (PET) effect and the paramagnetic nature of Cu2+ to the NBPA–Zn2+ system. Under optimized conditions, the fluorescence intensity is linear to the concentration of Zn2+ and Cu2+ separately, exhibiting good

Received 2nd December 2013 Accepted 2nd January 2014

linearities from 1.0 106 M to 8.0 106 M and 1.0 106 M to 1.0 105 M (R2 ¼ 0.9996 and R2 ¼ 0.9914) with detection limits of 7.89 107 M and 1.27 107 M, respectively. Furthermore, the rapid,

DOI: 10.1039/c3an02230g

selective feature enables the proposed sensor to be promising in monitoring Zn2+ and Cu2+ in biological

www.rsc.org/analyst

and environmental research with good accuracy and precision.

Introduction Serving as the second and the third richest heavy metals in human bodies,1 zinc and copper carry irreplaceable weight in our existence. The signicance of zinc is not conned to chemical engineering, for instance, manufacturing batteries,2 galvanization and fabricating alloy.3 It appears that zinc also possesses an extraordinary range of biological functions comprising aspects of brain work and pathology, immune function, mammalian reproduction,4 gene expression, and enzyme regulation,5 etc. Still, its overload in an organism induces apoptosis, epilepsy, ischemia, Alzheimer's disease6 and the like. In a parallel development, copper is instrumental in biological processes with regard to acting as a catalytic cofactor for a variety of enzymes4 which provide energy for biochemical reactions, transform melanin for pigmentation of the skin, assist the formation of cross-links in collagen and elastin, thereby maintaining and repairing connective tissues.7 Mitochondrial abnormalities and neurodegenerative diseases may have their roots in copper deciency. Surplus Cu2+ triggers the formation of reactive oxygen radicals in eukaryotes and

College of Chemistry, Jilin University, Changchun 130023, People's Republic of China. E-mail: y[email protected]; Fax: +86 43188499805; Tel: +86 13504326897 † Electronic supplementary 10.1039/c3an02230g

information

1868 | Analyst, 2014, 139, 1868–1875

(ESI)

available.

See

DOI:

prokaryotes which damage the intracellular membranes, proteins and nucleic acids.8 Consequently, the recognition and quantication of zinc and copper are signicant goals from both biological and environmental points of view. Small-molecule uorescent sensors possessing a sight of merits like high sensitivity, simplicity9 and tunability10 have sprung up as a powerful tool for conquering the drawbacks of electrochemical, photometric determination and electrothermal atomic absorption spectrophotometry which are costly, require pretreatment of the samples, and are not appropriate for large-scale monitoring.11 Reports for the detection of zinc and copper ions at trace quantity levels by enhancement or quenching of uorescence12,13 have been proposed recently and in the past few years. Nevertheless, on the one hand, the fact that zinc ions are relatively inert and possess a closed shell with a 3d10 electronic conguration has enabled some uorescent sensors to bind competitively with biologically abundant Ca2+, Mg2+ (see ref. 1) and toxic Cd2+.14–16 On the other hand, it is also of utmost consideration that only a few uorescent sensors capable of detecting two metal ions17 by relay recognition have been put forward with commendable selectivity. In consequence, there is still room for designing a novel uorescent sensor with high affinity and effective multiple discriminating ability. Prevalent mechanisms on cations binding to uorescent sensors include FRET,18 CHEF,19 ICT, PET, excimers20 and so on.


View Article Online


Paper

Among them, chelation-enhanced uorescence (CHEF) provides uorescence enhancements with some or no spectral shis. Sensors are always designed containing an organic uorophore as a signaling unit linked with a binding unit. The N,O donors are among the most widely used binding units.21 N,NBis(2-pyridylmethyl)amine (DPA) has one aliphatic and two aromatic nitrogen atoms able to coordinate to metals and it is a versatile tridentate ligand having been reported that can complex with a range of transition metals.22 In these papers, uorescent sensors based on DPA for Zn2+ and Cu2+ have been mentioned on many occasions.13,23,24 So, the DPA framework seems to be an ideal model to construct uorescent sensors for metal ions. Herein, in addition to introducing a novel DPA-based sensor, entitled N-(2-benzooxazol-2-ylphenyl)-2-(bispyridin-2-ylmethylamino)acetamide (NBPA), for Zn2+ with high sensitivity and selectivity alone, we discovered that NBPA–Zn2+ was in a position to identify Cu2+ in succession. NBPA showed a uorescent response toward Zn2+ directly and Cu2+ for the second step with remarkably high sensitivity and selectivity. The distinct feature was utilized to quantify Zn2+ and Cu2+ levels conveniently for healthcare and environmental monitoring over other competing ions including alkaline metals, alkaline-earth metals, transition metals, halogens, etc.

Experimental Chemicals and apparatus Distilled water was employed throughout the experiments. All the materials for synthesis were purchased from commercial suppliers and used without further purication. The solutions of cations were prepared from the corresponding chloride or nitrate salts and the solutions of anions were prepared from the corresponding sodium salts. Absorption spectra were recorded using a UV-2550 spectrophotometer (Shimadzu Corporation, Japan). Fluorescence measurements were performed on an RF5301PC Spectrophotouorometer (Shimadzu Corporation, Japan). All pH measurements were made with a Sartorius PB-10 digital pH meter. NMR spectra were recorded using a Varian Mercury YH-300 spectrometer operated at 300 MHz. ESI MS spectra were obtained using a Q-Trap2000 (Applied Biosystems Corporation, American) without using the LC part. Fluorescence decay curves were performed using an Edinburgh Instruments F900 apparatus. All measurements were carried out at room temperature (298 K). All the uorescent determinations were put into effect with an entrance slit and exit slit of 2.5 nm and 5 nm, respectively.

Analyst

We preserved both the solid and the ltrate. It then underwent 12 h of oven drying and the solid obtained was dissolved in ethanol and decolorized with acticarbon. CH2Cl2 played the part of extractant for the acquired ltrate. Both the extraction of CH2Cl2 and ethanol ltrates were evaporated under reduced pressure. Compound 1 (Scheme 1) was puried by alumina column chromatography (mineral ether/CH2Cl2 ¼ 2 : 1). 1H NMR (CDCl3, 300 MHz) d(ppm): 8.07 (dd, J ¼ 1.6 Hz, J ¼ 2.0 Hz, 1H), 7.68–7.44 (m, 1H), 7.54–7.59 (m, 1H), 7.28–7.36 (m, 1H), 6.77–8.82 (m, 2H), 6.0 (s, 2H). ESI-MS m/z: (1 + H)+ calc. for C13H11N2O, 211.2; found, 211.2. Synthesis of compound 2 0.4 ml (5.3 mmol) of 2-chloroacetyl chloride in 5 ml of dry CH2Cl2 was added dropwise to a solution of 0.63 g (3.0 mmol) of 1 in 20 ml of anhydrous CH2Cl2 at 0 C. Aer being stirred at room temperature for 12 h, the solvent was removed under reduced pressure to give product 2 (Scheme 1) as white solid with purication by alumina column chromatography (mineral ether/CH2Cl2 ¼ 1 : 1). 1H NMR (CDCl3, 300 MHz) d (ppm): 8.85 (d, J ¼ 8.4 Hz, 1H), 8.27 (d, J ¼ 8.0 Hz, 1H), 7.77–7.80 (m, 1H), 7.63–7.66 (m, 1H), 7.58 (t, J ¼ 4.0 Hz, 1H), 7.40–7.46 (m, 2H), 7.28–7.32 (m, 1H), 4.34 (s, 2H). ESI-MS m/z: (2 + H)+ calc. for C15H12N2O2Cl, 287.7; found, 287.1. Synthesis of NBPA To a solution of 2 (0.2860 g, 1.0 mmol) dissolved in acetonitrile (30 ml) was added N,N-bis(2-pyridylmethyl)amine (DPA) (0.18 ml, 1.0 mmol), potassium carbonate (0.2760 g, 2.0 mmol) and potassium iodide (0.0050 g, 0.03 mmol). Aer being stirred and reuxed for 8 h under nitrogen atmosphere, the mixture was cooled to room temperature and the volatiles were removed under reduced pressure to obtain a brown oil, which was puried by alumina column chromatography (CH2Cl2/CH3OH ¼ 100 : 2) to afford NBPA (Scheme 1). Yield: 0.1225 g (27.3%). 1H NMR (CDCl3, 300 MHz) d (ppm): 8.93 (d, J ¼ 3.8 Hz, 1H), 8.50 (d, J ¼ 8.4 Hz, 2H), 8.28 (d, J ¼ 7.2 Hz, 1H), 7.95 (d, J ¼ 6.8 Hz, 2H), 7.80 (t, J ¼ 3.6 Hz, 1H), 7.62 (d, J ¼ 7.2 Hz, 1H), 7.54 (m, 1H), 7.40 (m, 4H), 7.22 (s, 1H), 7.08 (s, 2H), 4.08 (s, 4H), 3.55 (s, 2H). 13C NMR (CDCl3, 300 MHz) d (ppm): 58.981, 60.550, 76.303, 76.621,

Synthesis of compound 1 Compounds 1 and 2 were synthesized according to a reported method.25 In sufficient polyphosphoric acid a mixture consisting of (2.7428 g, 15.9 mmol) 2-aminobenzoic acid and (2.1826 g, 20.0 mmol) o-hydroxyaniline was heated to 220 C. It gradually developed a blackish green color. Aer stirring for 4 h, the mixture was cooled to 100 C and poured into an ice water mixture. The mixture was ltered and then washed with water. This journal is © The Royal Society of Chemistry 2014

Scheme 1

The structure and synthesis route of the proposed probe

NBPA.

Analyst, 2014, 139, 1868–1875 | 1869

View Article Online


Analyst

Paper

Fig. 2 Job's plot for the determination of the stoichiometry of NBPA and Zn2+ ([NBPA + Zn2+] ¼ 10 mM) in CH3CN–buffer solution (1 : 1, v/v).

Fig. 3 Job's plot for the determination of the stoichiometry of NBPA– Zn2+ and Cu2+ ([NBPA–Zn2+ + Cu2+] ¼ 10 mM) in CH3CN–buffer solution (1 : 1, v/v).

The UV-vis absorption and fluorescence emission spectra of the probe: (a) absorption spectra: (1) NBPA; (2) NBPA–Zn2+; (3) NBPA– Zn2+–Cu2+; (4) NBPA–Cu2+; (b) emission spectra (lem ¼ 443 nm): (1) NBPA; (2) NBPA–Zn2+; (3) NBPA–Zn2+–Cu2+; (4) NBPA–Cu2+; buffer solution (HAc–NaAc), 0.10 M; pH, 6.3; buffer solution : CH3CN ¼ 1 : 1; entrance slit, 2.5 nm; exit slit, 5 nm; t, 10 min; NBPA, 1.0 105 M; Zn2+, 1.0 105 M; Cu2+, 1.0 105 M; (c) effect of reaction time on the fluorescence intensity with the addition of Zn2+ (t ¼ 146 s) and Cu2+ (t ¼ 822 s) to NBPA (t ¼ 0–146 s); t, 0–1200 s; NBPA, 1.0 105 M. Fig. 1

76.939, 110.325, 113.029, 119.014, 120.440, 121.796, 122.739, 122.865, 124.393, 125.139, 128.170, 132.258, 136.041, 138.218, 140.743, 148.711, 149.001, 157.643, 161.251, 170.322. ESI-MS m/z: (NBPA + H)+ calc. for C27H24N5O2 450.5; found, 450.4; m/z: (M + Na)+ calc. for C27H23N5O2Na 472.4; found, 472.3. Preparation of solutions for measurements A stock solution of NBPA (10.0 mM) was prepared by dissolving the required amount of NBPA (44.9 mg, 0.1 mmol) in

1870 | Analyst, 2014, 139, 1868–1875

Fig. 4 Fluorescence spectra of NBPA (1.0 105 M) in different concentrations of Zn2+, 1.0 106 to 8.0 106 M; other conditions as in Fig. 1. Inset: the plot of the fluorescence intensity vs. the concentration of Zn2+.

acetonitrile and diluting it to the mark in a 10 ml volumetric ask. Further dilutions were made to prepare 1.0 mM solution by diluting 1.0 ml stock solution to 10 ml with acetonitrile. Tap water, lake water and solid samples were obtained from our laboratory tap water supply, South Lake in Changchun and


View Article Online


Paper

Analyst

Fluorescence spectra of NBPA–Zn2+ (1.0 105 M) in different concentrations of Cu2+, 1.0 106 to 1.0 105 M; other conditions as in Fig. 1. Inset: a plot of the fluorescence intensity vs. the concentration of Cu2+. Fig. 5

Fluorescence intensity of NBPA/NBPA–Zn2+ (1.0 105 M) upon the addition of 1 equiv. Zn2+/Cu2+ in the presence of background anions in CH3CN–buffer solution (1 : 1, v/v); background ions, 1.0 105 M; other conditions as in Fig. 1. Fig. 8

Scheme 2

Proposed binding modes of NBPA with Zn2+ and Cu2+ in

solutions. 5

Fig. 6 Fluorescence intensity of NBPA (1.0 10 M) upon the addition of 1 equiv. different metal ions (red bar) and the addition of 1 equiv. Zn2+ in the presence of background cations (green bar) in CH3CN–buffer solution (1 : 1, v/v); background ions, 1.0 105 M; other conditions as in Fig. 1.

1

H NMR spectra in DMSO-d6 upon addition of (a) NBPA, (b) NBPA + 1.0 equiv. Zn2+, (c) NBPA + 1.0 equiv. Zn2+ + 1.0 equiv. Cu2+. Fig. 9

Fig. 7 Fluorescence intensity of NBPA–Zn2+ (1.0 105 M) upon the addition of 1 equiv. different metal ions (red bar) and the addition of 1 equiv. Cu2+ in the presence of background cations (green bar) in CH3CN–buffer solution (1 : 1, v/v); background ions, 1.0 105 M; other conditions as in Fig. 1.


campus soil respectively. No further purication was adopted before tap water and lake water were applied. The soil collected was dried in a hot oven for 12 h at 60 C. We stirred 10 g

Analyst, 2014, 139, 1868–1875 | 1871

View Article Online

Analyst Table 1

Paper Recovery study of Zn2+ in real samples

Sample

Zn2+ added/mol L1

Zn2+ found/mol L1

Recovery (%)

RSD (%)

Tap water

0 3.00 5.00 7.00 0 3.00 5.00 7.00 0 3.00 5.00 7.00 0 2.00

Not detected 2.98 106 5.15 106 7.02 106 Not detected 2.97 106 4.86 106 6.92 106 Not detected 2.92 106 5.12 106 6.92 106 5.04 106 6.83 106

— 99.3 103 100.3 — 99 97.2 98.9 — 97.3 102.4 98.9 — 90.5

— 1.9


Lake water

Soil sample

Nutriment sample

106 106 106 106 106 106 106 106 106 106

— 1.03

— 2.62

—

desiccated soil with 50 ml distilled water for 12 h to acquire the soluble melt salts. Aer ltration of the mixture, the soil sample solution was obtained. A tea sample (0.5020 g) that originated in Fujian Province of China and a nutriment product sample (0.5000 g) (aer removing the surface icing) produced by Shanghai Goldpartner Biotech Co., Ltd. were digested by a wet method respectively. Excess acids were evaporated before ltration. Finally, the tea sample and nutriment sample were diluted to 25 ml and 100 ml, respectively.

Cu2+ induced a further red shi of the new band whose peak transferred to 363 nm according with NBPA–Cu2+. In the uorescence emission spectrum (Fig. 1b), the emission intensity in the presence of Zn2+ had a sharp increase at 443 nm. Then, it was quenched as a consequence of the addition of copper ions. The emission spectra were recorded from 360 to 550 nm from excitation at 353 nm.


Fluorescence Job's method was applied to determine the stoichiometry of the NBPA–Zn2+ complex by keeping the sum of the initial concentration of zinc and NBPA at 1.0 105 M and the molar ratio of Zn2+ changing from 0.1 to 0.9. The uorescence of NBPA in the absence (F0) and presence (F) of Zn2+ was determined, respectively. A plot of (F F0)/F0 versus the molar fraction of zinc ion was obtained (Fig. 2). It showed that the (F F0)/F0 value went through a maximum at a molar fraction of about 0.5, indicating a 1 : 1 stoichiometry of Zn2+ to NBPA in the complex. Binding analysis between Cu2+ and NBPA–Zn2+ suggested that the complex formation also had a 1 : 1 stoichiometry

Absorption and emission spectra The UV-vis absorption and the uorescence emission spectra were investigated (Fig. 1) to show a remarkable Zn2+-induced uorescence enhancement and Cu2+-induced uorescence quenching. They were studied in the absence and presence of Zn2+ or Cu2+ in a solution of NBPA (1.0 105 M). In the absorption spectrum (Fig. 1a), Zn2+ caused the appearance of a new band centered at 353 nm along with red shis from the other three original peaks of NBPA. Aerwards, the addition of

Zn2+ and Cu2+ binding behaviors

Table 2

Recovery study of Cu2+ in real samples

Sample

Cu2+ added/mol L1

Cu2+ founded/mol L1

Recovery (%)

RSD (%)

Tap water

0 3.33 106 5.00 106 7.00 106 0 3.33 106 5.00 106 7.00 106 0 3.33 106 5.00 106 7.00 106 0 2.00 106

Not detected 3.25 106 4.99 106 7.19 106 Not detected 3.35 106 5.09 106 6.95 106 Not detected 3.30 106 5.18 106 6.90 106 0.52 106 2.44 106

— 97.6 99.8 102.7 — 100.6 101.8 99.3 — 99.1 103.6 98.6 — 96

— 2.56

Lake water

Soil solution

Tea sample

1872 | Analyst, 2014, 139, 1868–1875

— 1.25

— 2.75

—


View Article Online

Paper


(Fig. 3). More direct evidence was obtained by the ESI mass spectra of NBPA–Zn2+ (Fig. S-1†) and NBPA–Zn2+–Cu2+ (Fig. S-2†). In Fig. S-1,† the ion at m/z 514.3 corresponded to the molecular ion peak of NBPA–Zn2+ (calcd ¼ 514.5). While in Fig. S-2,† with 1 equiv. Cu2+ added to the NBPA–Zn2+ complex, the molecular ion peaks corresponding to NBPA–Cu2+ in the 511.0–515.0 Da mass range with 1 Da mass difference, due to the isotopic distributions of carbon and copper,26,27 were clearly observed, which indicated the substitution of Cu2+ for Zn2+.

Analyst

in the range of 1.0 106 M to 8.0 106 M. When it came to copper, the linear relationship ranged from 1.0 106 M to 1.0 105 M with a correlation coefficient of 0.9957 (R2 ¼ 0.9914) (Fig. 5). They had detection limits of 7.89 107 M and 1.27 107 M (3s) respectively for Zn2+ and Cu2+, which indicated the inherent high sensitivity of NBPA. According to the formulae as follows,28 a¼

a 1 ¼ 1 a K½Mnþ

Effects of experimental conditions For the sake of a steadier as well as a higher level of uorescence intensity, conditions like solvents, pH values, buffer solutions, kinetics of the complex compound, temperatures were explored. We rst considered the uorescence response of NBPA to Zn2+ in different common laboratory organic solvents (Fig. S-3†). In mixed solvent of 1 : 1 acetonitrile/H2O, NBPA showed the best uorescence response to Zn2+. The solvent composition test was performed by measuring different ratios of CH3CN to H2O in solution (Fig. S-4a†). The content of CH3CN ranged from 10 to 90% and there was a platform between 50 and 80%. As shown in Fig. S-4b,† the uorescence intensity of NBPA displayed weak signals within the pH range from 2.30 to 10.30 (mediated by NaOH and a mixture of acid comprising H3PO4, CH3COOH, H3BO3). Also, the uorescence intensity of the NBPA–Zn2+ complex displayed a maximum signal toward Zn2+ under pH 6.30. It was also available over a relatively wide pH range. The observed pKa values with a signicant response between 5 and 8.3 varied due to the effect of protonation and deprotonation of the functional groups, NH– and O–, of NBPA showing the transition between the cationic and anionic species. Hence buffer solutions like potassium acid phthalate–sodium hydroxide, acetic acid–sodium acetate, PBS were selected and compared. Fig. S-4c† indicates acetic acid–sodium acetate buffer solution fostered a higher uorescence intensity at pH 6.30. In Fig. 1c, Zn2+ was added at 146 s and the uorescence intensity reached a maximum at 181 s. During 822–920 s, Cu2+ accomplished the task of quenching. Therefore, uorescence determinations were made in 10 min according to the kinetic curve that showed either the zinc complex or the copper complex was formed within 3 min. Moreover, it seemed that temperature (Fig. S-5†) had no signicant interference on the uorescence response. By comprehensive consideration of the inuence of every factor, it was decided to implement the uorescent determination in 50% acetic acid–sodium acetate (0.1 M, pH ¼ 6.3) aqueous buffer solution (CH3CN : H2O ¼ 1 : 1, v/v). Zn2+ and Cu2+ titration The uorescence spectra of NBPA in the presence of different concentrations of Zn2+ in CH3CN–buffer solution (1 : 1, v/v) were recorded in Fig. 4. When excited at 353 nm, free NBPA exhibited weak uorescence intensity. The addition of different concentrations of Zn2+ led to a signicant uorescence emission enhancement peaking at 443 nm. There was a good linear relationship with a correlation coefficient of 0.9998 (R2 ¼ 0.9996) between the uorescence intensity and the concentration of Zn2+ This journal is © The Royal Society of Chemistry 2014

Ft F Ft F0

the association constant K could be calculated. Here, F0 and Ft are the limiting uorescence intensities in the absence of metal ions and completely complexed with cations. F is the corresponding uorescence intensity of the probe actually measured when contacting with a metal ion solution of a given concentration. An association constant KZn–NBPA value of 1.1 105 M was then calculated from the uorescence titration experimental data in Fig. 4. Also, the association constant KCu–NBPA was calculated to be 2.0 106 M with the same method. Therefore, it was more likely that Cu2+ would take the place of Zn2+ with the addition of Cu2+ to NBPA–Zn2+. Furthermore, the Stern–Volmer quenching constant (Fig. S-6†) Ka ¼ 7.27 106 L mol1 was given on the basis of the Stern–Volmer equation.29 We determined the quenching rate kq ¼ 7.9 107 mol L1 s1 according to Fig. S-6† and Fig. 1c.30,31 Zn2+ and Cu2+ binding selectivity A signicant feature of the sensor was its high selectivity toward Zn2+ and Cu2+ over other competitive species. The selectivity of the synthesized receptor NBPA was evaluated by observing changes in their uorescence emission spectra (Fig. S-7 and S-8†). Interference experiments were performed by mixing NBPA (1.0 105 M) with 1 equiv. of cations like Ag+, Al3+, As3+, Ba2+, Be2+, Ca2+, Cd2+, Cr6+, Cs+, Fe3+, Hg2+, K+, Li+, Mg2+, Mn2+, Mo6+, Na+, Ni2+, Pb2+, Rb+, Sr2+, V5+, Cu2+/Zn2+ (Fig. 6 and 7) or anions such as Ac, Br, Cl, CO32, F, NO3, SCN, SO42 (Fig. 8) respectively before Zn2+ or Cu2+ was added. It was noticeable that the miscellaneous competitive metal ions did not lead to signicant interference. As for the interference caused by the addition of Cu2+ to Zn2+–NBPA in Fig. 6, ascorbic acid was used to mask Cu2+. Samples containing both zinc ions and copper ions could be addressed by detecting the concentration of Cu2+ rst with NBPA–Zn2+ before adding ascorbic acid to mask Cu2+ for detecting Zn2+ with NBPA. The gures indicated that the receptor was highly selective and sensitive for zinc and copper ions. Sensing mechanism Different proposed binding modes of NBPA with Zn2+ and Cu2+ are described in Scheme 2. 1H NMR studies were performed to understand the details of the binding modes of NBPA with Zn2+ and Cu2+. For example, Xu et al.32 reported a sensor which coordinated with Zn2+ bound to the imidic acid nitrogen in the form of imidic acid. They also mentioned another binding

Analyst, 2014, 139, 1868–1875 | 1873

View Article Online


Analyst

mode formation which had been ruled out with Zn2+ bound to carbonyl oxygen resulting in large upeld shis of the resonance of the NH proton in 1H NMR studies. In our work, the NH proton (d ¼ 12.45 ppm) of NBPA underwent an upeld shi (d ¼ 9.20 ppm) by the addition of 1 equiv. Zn2+ with a large shi (Fig. 9). Zn2+ was believed to coordinate with carbonyl oxygen. The binding of the carbonyl oxygen with Zn2+ blocked the imidic acid structure and then shis the NH resonance upeld. The dissimilar binding modes between these two sensors is rooted in the different uorophores connected. In another example, Cu2+ may combine with the NBPA–Zn2+ system in the same manner of a corresponding report33 without the imidic acid structure as well. The enhancement of the uorescence intensity in the rst step was attributed to the selective CHEF mechanism mediated with the PET effect. Binding with Zn2+ blocked the PET quenching of the singlet excited state of the uorophore. Therefore, a drastic uorescence enhancement was observed. Upon the addition of copper ions, Zn2+ was replaced by Cu2+ owing to its lower association constant. Aer coordination with Cu2+, the paramagnetic effect from spin-orbit coupling of Cu2+ induced the uorescence quenching.34 Fluorescence decay curves for NBPA in the presence of Zn2+, 2+ Cu , Zn2+–Cu2+ were obtained in Fig. S-9† which showed the typical decay curves with sZn ¼ 2.99 ns, sCu ¼ 3.09 ns, sZn/Cu ¼ 3.20 ns, respectively. As shown, the time constants of NBPA– Zn2+ and NBPA–Zn2+–Cu2+ for the PET effect were 1.74 ns and 1.62 ns.35 For the CHEF effect, according to Fig. S-4d,† the rise time constant was 0.69 s caused by the addition of Zn2+. Similarly, the decay time constant was 5.23 s caused by the addition of Cu2+. Applications in real samples The practical applications of the designed sensor were evaluated by determination of Zn2+ and Cu2+ in tap water, lake water and soil solution samples. To keep a stable system at a given pH and a given solvent ratio, 3 ml of the prepared solution samples were added to every 12 ml acetic acid–sodium acetate buffer solution. The results obtained with the proposed sensor were in good agreement with the concentration of metal ions in the systems. As shown in Tables 1 and 2, it can be conrmed that recovery studies of Zn2+ and Cu2+ based on NBPA were satisfactory with RSD values ranging from 1.03 to 2.75%. Samples where Zn2+ and Cu2+ ions exist naturally, including a tea sample and nutriment sample (30 ml in every 1.5 ml buffer solution), were detected along with the recovery studies. The sensor synthesized in this paper showed possible feasibility for real sample testing with high selectivity and sensitivity.

Conclusions In conclusion, we have developed an off-on-off uorescent probe for Zn2+ and Cu2+ by combining the strong affinity of the N,N-bis(2-pyridylmethyl)amine unit to zinc and copper. The proposed probe exhibits a uorescence enhancement response toward Zn2+ and the following NBPA–Zn2+ system can also quantify trace amounts of Cu2+ by means of the decrease of the

1874 | Analyst, 2014, 139, 1868–1875

Paper

uorescence intensity with high sensitivity and selectivity. Moreover, its uorescent response toward Zn2+ and Cu2+ can be used for quantication of Zn2+ or Cu2+ in aqueous samples. Besides, it exhibits excellent selectivity over competing with other common cations and anions, which could meet the selectivity requirements for practical applications.

Acknowledgements The authors are grateful for the nancial support from the Basic Research Foundation of Jilin University (no. 201103096, 201103102), the Science-Technology Development Project of Jilin Province of China (no. 201105008, 20126018, 20130206014GX) and the State Major Project for Science and Technology Development, China (no. 2011YQ14015001).

Notes and references 1 F. A. Abebe and E. Sinn, Tetrahedron Lett., 2011, 52, 5234– 5237. 2 Q. Lai, H. Zhang, X. Li, L. Zhang and Y. Cheng, J. Power Sources, 2013, 235, 1–4. 3 S. Cai, T. Lei, N. Li and F. Feng, Mater. Sci. Eng., C, 2012, 32, 2570–2577. 4 K. Jobe, C. H. Brennan, M. Motevalli, S. M. Goldup and M. Watkinson, Chem. Commun., 2011, 47, 6036– 6038. 5 M. Hagimoria, N. Mizuyamab, Y. Yamaguchia, H. Sajic and Y. Tominagad, Talanta, 2011, 83, 1730–1735. 6 G. Sivaraman, T. Anand and D. Chellappa, Analyst, 2012, 137, 5881. 7 D. Maity, A. K. Manna, D. Karthigeyan, T. K. Kundu, S. K. Pati and T. Govindaraju, Chem.–Eur. J., 2011, 17, 11152–11161. 8 B. Hotzer, R. Ivanov, S. Altmeier, R. Kappl and G. Jung, J. Fluoresc., 2011, 21, 2143–2153. 9 Y. Cai, X. Meng, S. Wang, M. Zhu, Z. Pan and Q. X. Guo, Tetrahedron Lett., 2013, 54, 1125–1128. 10 K. Shen, X. Yang, Y. Cheng and C. Zhu, Tetrahedron, 2012, 68, 5719–5723. 11 Z. Dong, Y. Guo, X. Tian and J. Ma, J. Lumin., 2013, 134, 635– 639. 12 M. Shellaiah, Y. Wu and H. Lin, Analyst, 2013, 138, 2931. 13 D. Maity, V. Kumar and T. Govindaraju, Org. Lett., 2012, 14, 6008–6011. 14 R. Azadbakht, H. Keypour, H. A. Rudbari, A. H. Mohammad Zaheri and S. Menati, J. Lumin., 2012, 132, 1860–1866. 15 H. Song, S. Rajendiran, E. Koo, B. K. Min, S. K. Jeong, T. D. Thangadurai and S. Yoon, J. Lumin., 2012, 132, 3089– 3092. 16 S. Kotha, D. Goyal (nee Bansal), S. Banerjee and A. Datta, Analyst, 2012, 137, 2871. 17 R. Pandey, P. Kumar, A. K. Singh, M. Shahid, P. Li, S. K. Singh, Q. Xu, A. Misra and D. S. Pandey, Inorg. Chem., 2011, 50, 3189–3197. 18 K. Baek, M. S. Eom, S. Kim and M. S. Han, Tetrahedron Lett., 2013, 54, 1654–1657.


View Article Online


Paper

19 C. Gao, X. Jin, X. Yan, P. An, Y. Zhang, L. Liu, H. Tian, W. Liu, X. Yao and Y. Tang, Sens. Actuators, B, 2013, 176, 775–781. 20 P. A. Panchenko, Y. V. Fedorov, O. A. Fedorova and G. Jonusauskas, Dyes Pigm., 2013, 98, 347–357. 21 V. Kumar, A. Kumar, U. Diwan and K. K. Upadhyay, Dalton Trans., 2013, 42, 13078. 22 V. Tiwow, G. A. Lawrance, M. Maeder and P. Jensen, J. Coord. Chem., 2011, 64(20), 3637–3651. 23 Y. Cai, X. Meng, S. Wang, M. Zhu, Z. Pan and Q. X. Guo, Tetrahedron Lett., 2013, 54, 1125–1128. 24 J. Li, C. F. Zhang, Z. Z. Ming, G. F. Hao, W. C. Yang and G. F. Yang, Tetrahedron, 2013, 69, 4743–4748. 25 M. Chen, X. Lv, Y. Liu, Y. Zhao, J. Liu, P. Wang and W. Guo, Org. Biomol. Chem., 2011, 9, 2345. 26 A. Yazici, N. Un¨ u¸s , A. Altindal, B. Salih and O. Bekaro˘ glu, Dalton Trans., 2012, 41, 3773–3779. 27 G. Feng, Y. Dai, H. Jin, P. Xue, Y. Huan, H. Shan and Q. Fei, Sens. Actuators, B, 2014, 193, 592–598.


Analyst

28 T. T. Wei, J. Zhang, G. J. Mao, X. B. Zhang, Z. J. Ran, W. Tan and R. Yu, Anal. Methods, 2013, 5, 3909– 3914. 29 L. R. Gouvea, D. A. Martins, D. G. J. Batista, M. N. C. Soeiro, S. R. W. Louro, P. J. S. Barbeira and L. R. Teixeira, BioMetals, 2013, 26, 813–825. 30 M. Liang, A. Kaintz, G. A. Baker and M. Maroncelli, J. Phys. Chem. B, 2012, 116(4), 1370–1384. 31 S. Yuan, Y. Zhang, R. Lu and A. Yu, J. Photochem. Photobiol., A, 2013, 260, 39–49. 32 Z. Xu, X. Liu, J. Pan and D. R. Spring, Chem. Commun., 2012, 48, 4764–4766. 33 J. T. Hou, K. Li, K. K. Yu, M. Y. Wu and X. Q. Yu, Org. Biomol. Chem., 2013, 11, 717. 34 L. Qu, C. Yin, F. Huo, Y. Zhang and Y. Li, Sens. Actuators, B, 2013, 183, 636–640. 35 Q. Wang, B. Yang, Y. Zhang, H. Xia, T. Zhao and H. Jiang, J. Alloys Compd., 2013, 581, 801–804.

Analyst, 2014, 139, 1868–1875 | 1875

View Article Online

Analyst


PAPER


A capillary electrophoresis-based immobilized enzyme reactor using graphene oxide as a support via layer by layer electrostatic assembly Zhengri Yin, Wenwen Zhao, Miaomiao Tian, Qian Zhang, Liping Guo and Li Yang* A novel capillary electrophoresis (CE)-based immobilized enzyme reactor (IMER) using graphene oxide (GO) as a support was developed by using a simple and reliable immobilization procedure based on layer by layer electrostatic assembly. Using trypsin as a model enzyme, the performance of the fabricated CE-based IMERs was evaluated. Various conditions, including trypsin concentration, trypsin coating time, number of trypsin layers and buffer pH, were investigated and optimized. The Michaelis constant Km (0.24 0.02 mM) and the maximum velocity Vmax (0.32 0.04 mM s1) were determined using the CE-based IMERs, and the values are consistent with those obtained using free trypsin, indicating that enzyme immobilized via the proposed approach does not cause a significant structural change of the enzyme or any reduction of enzyme activity. The presented CE-based IMERs exhibit excellent reproducibility with RSD less than 2.8% over 20 runs, and still remain 79.5% of the initial activity after five days with more than 100 runs. Using the proposed CE-based IMERs, the digestion of angiotensin was completed within 3 min,


while quite a number of trypstic peptides were observed for BSA on-line digestion with an incubation time of 30 min. As identified by MS analysis, the online digestion products of BSA using the present CEbased IMER are comparable with those obtained using free trypsin digestion for 12 h incubation. It is

DOI: 10.1039/c3an02241b

indicated that the present immobilization strategy using GO as a support is reliable and practicable for

www.rsc.org/analyst

accurate on-line analysis and characterization of peptides and proteins.

1. Introduction Since enzyme immobilization has been revealed as a powerful tool to improve almost all enzyme properties, such as stability, activity, specicity, selectivity and reusability, immobilized enzyme reactors (IMERs) have been applied widely in chemical and biological assays.1–7 Usually IMERs are integrated with a separation and identication system for on-line separation and detection of substrates and products of enzyme reactions, thus fast, efficient, high-throughput and automated enzymatic analysis can be achieved.8–10 Among a variety of separation techniques, capillary electrophoresis (CE) offers several advantages, such as high efficiency, sensitivity, fast analysis, low sample volume requirement and so on.11–16 By combining with IMERs, CE can be applied not only as a separation tool with high performance but also as a versatile platform for on-line enzyme studies. Over the past few decades, CE-based IMERs, in which IMERs are fabricated on capillaries (or microuidic chips), have attracted intense research interest, representing a promising miniature approach over a wide range of application

Faculty of Chemistry, Northeast Normal University, 5268 Renmin Road, Changchun, Jilin, 130024, P. R. China. E-mail: [email protected]; Fax: +86-431-85099762; Tel: +86-431-85099762


of enzyme assays including enzyme activity, peptide mapping in proteomics, inhibition screening and diagnostics.17–23 Effort has been made to prepare CE-based IMERs, which can be assigned to three different approaches: (i) immobilizing enzymes on the surface of a capillary leading to an open tubular enzyme reactor; (ii) immobilizing enzymes on beads or membranes that are entrapped in a dened area of a capillary network; (iii) immobilizing enzymes on monoliths formed in situ in a capillary. For either of the approaches, developing a new enzyme support in CE-based IMERs remains an important research aspect and challenging work. Several factors have to be considered for searching a suitable material of enzyme support, such as binding capacity of enzymes to the capillary, improvement of activity and stability of enzymes, ease-to-operate immobilization procedures, high-efficient separation and sensitive detection of substrates and products.24–26 Recently, along with the rapid development in nano-science, nano-structured materials have emerged as supports for enzyme immobilization.27–31 It has been demonstrated that the enzymes immobilized on the nano-structured materials have some advantages over the bulk solid substrates due to their large surface areas and good biocompatibility. Generally, surface modication or functionalization is required in order to efficiently immobilize enzymes onto the nano-structured materials, which could be a labored work and could reduce

Analyst, 2014, 139, 1973–1979 | 1973

View Article Online


Analyst

reproducibility and accuracy of enzyme assay. As one of the most studied sheet-based materials, graphene oxide (GO) has shown several advantages such as ease of synthesis, large surface area to mass ratio, and surface functionalities for induced-t interactions for enzyme binding,32–38 thus making it a potential synthetic support for enzyme immobilization. In particular, since the GO sheet is enriched with oxygen-containing groups, it is possible to immobilize enzymes without any surface modication or any coupling reagents. To date, there has been no application of GO in CE-based IMERs, however, few recent signicant advances, which have made in the GO-based nanobiocatalytic systems,32,33,36,37 show promise to use GO as a support for efficient immobilization of various enzymes, such as lipases, esterase, protease, etc. In this work, we reported a novel CE-based IMER using GO as an enzyme immobilization support, which was fabricated with a simple and reliable immobilization procedure based on layer by layer (LBL) electrostatic assembly. Using trypsin as a model enzyme, the performance of the activity of the CE-based IMERs was investigated to demonstrate the feasibility and accuracy of the present method for on-line enzyme assay. Analysis of on-line trypsin digestion of peptide (angiotensin) and protein (BSA) was also studied using the fabricated CEbased IMERs.

2. 2.1

Schematic drawing of the strategy to fabricate CE-based IMERs using GO as an enzyme support.

Fig. 1

Experimental Chemicals

Poly(diallyldimethylammonium chloride) (PDDA) (20%, w/w in water, Mw ¼ 200 000–350 000) was purchased from Jing Chun Reagent Inc. (Shanghai, China). GO dispersion (1 mg mL1) was purchased from XF NANO Inc (Nan Jing, China). N-a-Benzoyl-Larginine ethyl ester hydrochloride (BAEE) and N-a-benzoyl-Larginine (BA) were purchased from Alfa Aesar (Lancs, UK). Angiotensin (HPLC purity >98%) (Asp-Arg-Val-Tyr-Ile-His-ProPhe) was synthesized by Shanghai Science Peptide Biological Technology Co. (Shanghai, China). Trypsin TPCK treated from bovine pancreas and bovine serum albumin (BSA) was purchased from Sigma Chemical Co. (Mt. Louis, MO). Other reagents were of analytical grade and used without further purication. All the solutions were ltered using a 0.22 mm membrane lter prior to use. 2.2

Paper

The charge polarity was reversed aer adsorption of a layer of negative-charged GO, which was achieved by injection of 1 mg mL1 GO dispersion solution (50 mbar for 20 s) and remained in the capillary for 30 min. A single-layer IMER was then developed by injecting the trypsin enzyme solution (1 mg mL1 in 50 mM Tris–HCl buffer at pH 8.5) into the capillary and remained in the capillary for 30 min. Because the pI value of trypsin is about 10.5, trypsin should be positively charged at pH 8.5 and can be absorbed on the negative-charged GO layer by electrostatic assembly coating. Between the steps, the capillary was ushed with deionized water for 5 min to wash out any unreacted reagent. To fabricate multi-layer IMERs, the procedure for coating GO and trypsin was repeated. The modication of the capillary at each step was characterized by scanning electron microscopy (SEM) images, which were recorded using an XL30ESEM-FEG SEM Microscope (SEI).

Preparation of the CE-based IMERs

GO, which was immobilized on the inner surface of the capillary, has been proved as a stable stationary phase for open-tubular capillary electrochromatography.39–41 The CEbased IMERs using GO as an enzyme support were developed using LBL electrostatic assembly, as shown schematically in Fig. 1. Prior to modication, an untreated capillary was successively rinsed with 0.1 M NaOH for 30 min and deionized water for 10 min. Once preconditioned, PDDA solution was injected into the capillary by pressure at 50 mbar for 20 s, resulting in an about 2 cm long plug of PDDA solution. The plug was then stayed in the capillary for 1 h to create positivecharged coating on the inner wall of the 2 cm long capillary.

1974 | Analyst, 2014, 139, 1973–1979

2.3

Enzyme assay using the CE-based IMERs

Enzyme assay using the fabricated IMER in a capillary column (25 mm i.d., 365 mm o.d.) was performed in a CE apparatus (CL1020, Beijing Cailu Science Apparatus, China) with a UV detector. The total length of the capillary was 40 cm and the length between the detection window and the outlet was 8 cm. The CE running buffer was 20 mM Tris–HCl buffer at pH 8.5. As shown in Fig. 1, the IMER with the length of 2 cm was set at the inlet of the capillary. Prior to analysis, the IMER capillary was lled with the running buffer and was equilibrated at 200 V cm1 until a stable current and baseline were achieved. Substrate solutions were injected into the IMER capillary at


View Article Online

Paper

240 V cm1 for 3 s. Aer incubation by suspending the column in buffer, an electric potential of 240 V cm1 was applied to separate the substrate and products. The reacted substrate was determined by measuring the peak height of the product, which was detected by UV absorption at a wavelength of 214 nm.


2.4

Digestion of angiotensin and BSA

1 mg mL1 angiotensin and 10 mg mL1 BSA were digested using the CE-based IMERs. Angiotensin solution was directly injected into the IMER capillary at 240 V cm1 for 3 s. BSA was rst denatured into 50 mM Tris–HCl buffer containing 8 M urea for 1 h at 37 C, and then the sample was diluted with the same buffer to the concentration of urea less than 1 M and was stored at 4 C prior to use. On-line digestion of angiotensin or BSA was then analyzed using the CE-based IMERs. Aer incubation by suspending the column in buffer, digested products were separated by applying 240 V cm1 electric potential and detected by UV absorption at a wavelength of 214 nm. For comparison, the digestion of BSA was also carried out using free trypsin. The in-solution digestion was performed by adding 0.5 mg free trypsin into the denatured BSA solution and the mixed solution was incubated at 37 C for 12 h. Aer adding 20 mL of formic acid to stop the reaction, the digested solution was then ready for CE analysis. 2.5

ESI-MS conditions and data analysis

Digestion of the standard BSA sample using either free trypsin or the CE-based IMER was identied from peptide ngerprint mass spectra. To collect the eluent from the CE-based IMERs for MS analysis, the cathode end of the capillary was placed inside a stainless steel needle using a coaxial liquid-sheath-ow conguration (three-way connection). The sheath ow buffer was 50 mM Tris buffer (pH 8.0) with a ow rate of 2 mL min1 controlled by a digital syringe pump (Jiashan Ruichueng Electronic Tec. Co., Ltd., China). The eluent was collected aer 5 min CE running for 10 min, and then was subsequently 1 : 1 (v/v) diluted with a 0.3 wt% TFA in water. Aer desalting on Milipore ZipTip C18 tips with 10 mL of 0.1 wt% FA/50 wt% ACN as the eluting buffer, the sample was introduced directly to the MS. For MS analysis of the offline trypsin digestion, the digested sample was directly desalted and sent to the MS. The LTQ XL linear ion trap mass spectrometer (Thermo, USA) was operated in positive ionization mode. The ESI(+) source parameters are: capillary voltage 3.0 kV, sample cone voltage 35 V, extraction cone voltage 3 V, radio frequency lens voltage 450 V, desolvation temperature 120 C and desolvation gas 250 L h1. Signals were recorded in a m/z range of 600–2000 at a 1.0 s scan time. The peak list from the MS spectrum was exported to peptide mass ngerprint for protein identication using the MASCOT search engine (http://www.matrixscience.com) with the SwissProt database. Up to 1 missed cleavage in trypsin digestion was allowed. The peptide tolerance was set to 2.0 Da. Peptide masses searched were monoisotopic. Entries with a MASCOT MOWSE score correlative to p < 0.05 were identied as signicant hits.


Analyst

3. 3.1

Results and discussion Performance of the CE-based IMERs

To show the modication process on the inner surface of the capillary, we present in Fig. 2 the SEM images of (a) a PDDA coated capillary end, (b) a PDDA–GO coated capillary end, (c) a single-layer IMER and (d) a double-layer IMER. The surface of a bare capillary was very smooth aer rinsing with 0.1 M NaOH for 30 min (SEM image not shown). Aer the bare capillary was modied with PDDA, a series of small hills or mounds of less varying depth, width, and shape on the inner surface of the capillary were observed, as shown in Fig. 2a. Obviously, the surface area of the inner wall was greatly increased aer coating with PDDA, and the new surface was quite uniform and well dened. When GO was coated onto the PDDAcolumn, it was observed that the surface of the PDDA-column was then covered by a layer of the GO sheet (Fig. 2b). Such a sheet-layer-like structure was quite identical to the TEM image of GO dispersion, which is shown in the inset gure of Fig. 2b. The SEM image clearly indicates successful modication of GO onto the capillary wall. In addition, our results showed that GO can maintain a sheet-layer-like structure when coated onto the capillary via electrostatic assembly. Aer immobilization of trypsin, the surface of either the single-layer or the doublelayer IMER is also uniform and well-dened and presents a similar sheet-layer-like structure as the PDDA–GO column (Fig. 2c and d). Several key factors were studied to optimize the performance of the CE-based IMERs. The conditions for fabrication of CEbased IMERs are essential to the enzyme loading capacity and IMER activity. Shown in Fig. 3(a)–(c) are the effects of trypsin concentration, trypsin coating time and trypsin layer numbers on the enzymatic activity, respectively. Each data point is an averaged result of three replicated analysis. For those experiments, BAEE with a concentration of 0.5 mM was used as the substrate. The CE running buffer was 20 mM Tris–HCl buffer at

SEM images of (a) a PDDA-coated capillary, (b) a PDDA–GO capillary, (c) a single-trypsin-layer IMER and (d) a double-trypsin-layer IMER. The inset image in (b) is the TEM image of GO dispersion.

Fig. 2

Analyst, 2014, 139, 1973–1979 | 1975

View Article Online


Analyst

Paper

Fig. 3 Effect of the (a) trypsin concentration, (b) trypsin coating time and (c) trypsin layer numbers on the enzymatic activity of the IMER. BAEE with a concentration of 0.5 mM was used as the substrate. The buffer pH value was kept at 8.5. Conditions for fabrication of the IMER in each figure are (a) double-layer IMER and 30 min trypsin coating time for each layer, (b) double-layer IMER and 1 mg mL1 trypsin, (c) 1 mg mL1 trypsin and 30 min trypsin coating time for each layer. Without incubation after injection of BAEE, the product was separated and detected by UV absorption at a wavelength of 214 nm. The CE running buffer was 20 mM Tris–HCl buffer at pH 8.5 and the electric field strength was set at 240 V cm1.

pH 8.5 and the electric eld strength was set at 240 V cm1. As presented in Fig. 3(a), the peak height of the product BA rst increases sharply as the trypsin concentration is increased to 1 mg mL1, and then remains almost constant as the concentration is further increased, indicating that the enzyme loading capacity as well as the activity of the fabricated CE-based IMER reach the maximum. As the trypsin coating time is increased, the amount of enzyme that can interact with the GO layer and can immobilize on the inner surface of the capillary increases, resulting in the sharply increased peak height of the product BA in the coating time of 10–30 min (Fig. 3(b)). For the coating time larger than 30 min, the peak height of BA only increases slowly which could be attributed to saturation of trypsin that covers the GO layer. Similarly, increasing the trypsin layers also could increase the enzyme loading amount thus increasing the activity of the IMER, as shown in Fig. 3(c). Considering the reactor activity as well as the fabrication time, a two-layertrypsin format, a trypsin concentration of 1 mg mL1 and a coating time of 30 min for each layer were chosen for fabrication of the CE-based IMERs. Fig. 4 shows the effect of buffer pH on the trypsin cleavage reaction of BAEE in the pH range of 6.5–9.5 using the CE-based IMER. For comparison, we also present in the gure the result using free trypsin. 1 mM BAEE was used as the substrate for the

Fig. 4 Effect of buffer pH on the relative activity using a CE-based IMER and free trypsin. The conditions for fabrication of the CE-based IMER were the double-trypsin-layer, 1 mg mL1 trypsin and 30 min trypsin coating time for each layer. Other conditions were the same as those in Fig. 3.

1976 | Analyst, 2014, 139, 1973–1979

tests and 20 mM Tris–HCl buffer at different pH values was used as the running buffer. For either of the plots in Fig. 4, the relative activity of trypsin was presented, corresponding to the BA peak height normalized to its maximum value. The maximum activity of the CE-based IMER was observed at pH 8.5, which was shied by 1 unit towards the alkaline compared to that of the free enzyme. Such a shi may be attributed to the alteration of the microenvironment of the enzyme aer immobilization on the inner surface of the capillary. In this study, the buffer pH was kept at the optimal value of 8.5 for experiments using the IMER, and 7.5 for experiments using free trypsin. The digestion of BAEE by trypsin using the CE-based IMER was determined by measuring the BA peak height as a function of the concentration of the substrate BAEE, as shown in Fig. 5. Each data point in the gure was carried out using the same IMER and under the same experimental conditions. By nonlinear regression of the Michaelis–Menten diagrams, the Michaelis constant (Km) and the maximum velocity (Vmax) were determined to be 0.24 0.02 mM and 0.32 0.04 mM s1, respectively. For comparison, the Km and Vmax values were also measured using free trypsin, and the values were determined to be 0.18 0.03 mM and 0.39 0. 03 mM s1, respectively. It can be seen that the Km and Vmax values using the CE-based IMER and free trypsin are close, indicating that immobilization of trypsin via the proposed approach does not cause a signicant

Michaelis–Menten diagram of trypsin on the IMER. The buffer pH value was kept at 8.5. Other conditions were the same as those in Fig. 4.

Fig. 5


View Article Online

Paper

Analyst


autolysis of trypsin in solution. Regarding the batch-to-batch reproducibility, ve freshly prepared IMERs were tested under the optimized conditions, and an average value of three runs for each IMER was provided. The results give a good batch-to-batch reproducibility with an RSD of 6.8%. Our results show excellent intraday and interday stability and batch-to-batch reproducibility of the fabricated CE-based IMERs, implying that the present immobilization strategy using GO as an enzyme support is reliable and practicable for accurate on-line enzyme assay.

Reproducibility over 20 runs per day in three days using the same CE-based IMER. The inset figures show the change of enzyme activity after 5 days using the CE-based IMER (blue) and free trypsin (red). The measured BA peak height in the first day was set as 100% enzyme activity. For the results of free trypsin, 1 mM BAEE was used as a substrate, the trypsin solution (1 mg mL1) was placed in 20 mM Tris– HCl buffer (pH 7.5) at 4 C for storage and the enzyme assay was carried out at 25 C. Enzyme activity was assayed at regular intervals. Other conditions were the same as those in Fig. 4. Fig. 6

structural change of the enzyme or any reduction of accessibility of the substrate to the active sites of the immobilized trypsin. 3.2

Reproducibility and stability of the CE-based IMERs

The reproducibility and stability of the CE-based IMERs were investigated by sampling and analyzing more than 20 times per day in consecutive 5 days. Aer each day's test, the same IMER was kept in the running buffer at 4 C for the next day's test. The activity of the IMER keeps relatively constant in 20 runs during the same day. As shown in Fig. 6 from the results of the rst three days, the RSD for 20 runs in each day is less than 2.8%. The activity of the IMER decreases slightly day by day, but still remains 79.5% of initial activity aer ve days with more than 100 runs, as shown in the inset gure of Fig. 6. The decrease of enzyme activity of the IMER might be caused by the small amount of enzyme release from the IMER during rinse and separation procedures. In the inset gure of Fig. 6, we also present the results using free trypsin. The results show that the activity of free trypsin decreases rapidly and can only maintain 21.4% of initial activity aer ve days, which could be due to

3.3 Angiotensin and BSA digestion and analysis on the CEbased IMERs To show the feasibility of the fabricated CE-based IMERs for online analysis of trypsin digestion of peptides and proteins, angiotensin and BSA were used as the substrates for enzyme assay. Fig. 7a shows the electropherograms for digestion of 1 mg mL1 angiotensin on the IMER with different incubation times. Peak 1 and peak 2 in each electropherogram refer to UV absorption of angiotensin (Asp-Arg-Val-Tyr-Ile-His-Pro-Phe) and the digested product (Val-Tyr-Ile-His-Pro-Phe), respectively. It can be seen that, as increasing the incubation time from 0–3 min, the peak of the substrate gradually decreases while that of the corresponding product gradually increases. The peak of angiotensin almost disappears and only the product peak remains in the electropherogram with 3 min incubation, indicating that the digestion of angiotensin is completed. Fig. 7b presents electropherograms for digestion of 1 mg mL1 BSA on the CE-based IMER, with incubation times of 10 min and 30 min. The result recorded using free trypsin with 12 h incubation time is also shown in Fig. 7b for comparison. A bunch of peaks appearing within the migration time of 3–8 min correspond to various peptide products from BSA digestion. With an incubation time of only 10 min, the electropherogram for digestion of BSA using the CE-based IMER shows fewer peptide peaks than that recorded using free trypsin with 12 h incubation time. On the other hand, as the incubation time is increased to 30 min, the number and shapes of digestion products are comparable with the results obtained using free trypsin digestion for 12 h. To further demonstrate the feasibility of the fabricated CE-based IMER for enzyme assay, we performed MS analysis of the eluent from the IMER column (with

Fig. 7 (a) Electropherograms for digestion of 1 mg mL1 angiotensin on the CE-based IMER with different incubation times. Analyte peaks in order of elution: peak 1, angiotensin and peak 2, digested product. (b) Electropherograms for digestion of 1 mg mL1 BSA on the CE-based IMER with incubation times of 10 min and 30 min, and that recorded using free trypsin with 12 h incubation time.


Analyst, 2014, 139, 1973–1979 | 1977

View Article Online

Analyst


Table 1

Paper Identified peptide masses and sequence from trypsin digestion of BSA using the fabricated CE-based IMER and the free trypsin

Mass (Da)

Position

Missed cleavage

Peptide sequence

CE-based IMER

657 711 2434 3578 1361 976 926 2044 1632 2315 700 648 688 846 921 1577 1385 1566 1478 816 2700 1666 840 659 724

24–28 29–34 45–65 45–75 89–100 123–130 161–167 168–183 184–197 184–204 198–204 223–228 236–241 242–248 249–256 267–280 286–297 347–359 421–433 452–459 460–482 469–482 483–489 490–495 581–587

1 0 0 1 0 0 0 1 0 1 0 0 0 1 0 0 0 0 0 1 1 0 0 0 0

R.RDTHK.S K.SEIAHR.F K.GLVLIAFSQYLQQCPFDEHVK.L K.GLVLIAFSQYLQQCPFDEHVKLVNELTEFAK.T K.SLHTLFGDELCK.V R.NECFLSHK.D K.YLYEIAR.R R.RHPYFYAPELLYYANK.Y K.YNGVFQECCQAEDK.G K.YNGVFQECCQAEDKGACLLPK.I K.GACLLPK.I R.CASIQK.F K.AWSVAR.L R.LSQKFPK.A K.AEFVEVTK.L K.ECCHGDLLECADDR.A K.YICDNQDTISSK.L K.DAFLGSFLYEYSR.R K.LGEYGFQNALIVR.Y R.SLGKVGTR.C R.CCTKPESERMPCTEDYLSLILNR.L R.MPCTEDYLSLILNR.L R.LCVLHEK.T K.TPVSEK.V K.CCAADDK.E

O

30 min incubation time) and compared with that from free trypsin digestion (with 12 h incubation time). The identied peptides from trypsin digestion are listed in Table 1. For digestion using the CE-based IMER, 20 peptides were identied with 31% coverage of the BSA sequence, which is very comparable with that obtained using free trypsin (23 peptides, 34% coverage of the BSA sequence). The results indicate that the present CE-based IMERs can be used for online digestion of peptides and proteins for efficient analysis and characterization of proteins.

4. Summary In this work, a novel CE-based IMER using GO as an enzyme support was developed using a simple and reliable immobilization procedure based on layer-by-layer electrostatic assembly, which can easily load multiple layers of enzymes, thus increase the capacity of the IMER. SEM images clearly show successful modication of the capillary surface, and indicate that the surface of either single-layer or double-layer IMER is uniform and well-dened and present the similar sheet-layer-like structure as the PDDA–GO coating column. Using trypsin as the model enzyme and BAEE as the substrate, the performance of the fabricated CE-based IMERs was evaluated. Various conditions which are essential for fabrication of the IMERs, including trypsin concentration, trypsin coating time and numbers of trypsin layers, were investigated to optimize the enzyme activity. The maximum activity of the IMER was observed at the buffer pH of 8.5, which was shied by 1 unit towards the alkaline compared to the free enzyme; however, the tendency of

1978 | Analyst, 2014, 139, 1973–1979

O O O O O O O O O O O O O O O O O O O

Free trypsin

O O O O O O O O O O O O O O O O O O O O O O O

dependence of the enzyme activity on the buffer pH is identical to the IMER and free trypsin. The Michaelis constant and the maximum velocity of BAEE determined using the CE-based IMER (0.24 0.02 mM and 0.32 0.04 mM s1) were close to those obtained using free trypsin, indicating that the enzyme immobilized via the proposed approach does not cause a signicant structural change of the enzyme or any reduction of enzyme activity. Run-to-run and batch-to-batch reproducibility as well as stability of the IMERs were investigated. The RSD over 20 reduplicate runs is less than 2.8%, and that of ve batches is 6.8%. The IMER can still remain 79.5% of the initial activity aer ve days with more than 100 runs. Such good reproducibility and stability ensure accurate on-line enzyme assay using the present method. Finally, analysis of on-line trypsin digestion of peptide or protein was investigated using the CE-based IMERs. Both the CE assay and the MS analysis give comparable results of BSA digestion using the CE-based IMER and free trypsin, indicating the potential valuable application of our approach using GO as an enzyme support to fabricate IMERs for efficient on-line analysis and characterization of proteins.

Acknowledgements This work is supported by the National Natural Science Foundation of China (Grant no. 21175018) and Jilin Provincial Science and Technology Development Foundation (Grant no. 20120431).

References 1 Y. Asanomi, H. Yamaguchi, M. Miyazaki and H. Maeda, Molecules, 2011, 16, 6041–6059.


View Article Online


Paper

2 J. Ma, L. Zhang, Z. Liang, Y. Shan and Y. Zhang, TrAC, Trends Anal. Chem., 2011, 30, 691–702. 3 A. Monzo, E. Sperling and A. Guttman, TrAC, Trends Anal. Chem., 2009, 28, 854–864. 4 R. A. Sheldon, Adv. Synth. Catal., 2007, 349, 1289–1307. 5 G. Ghafourifar, A. Fleitz and K. C. Waldron, Electrophoresis, 2013, 34, 1804–1811. 6 T. Wang, J. Ma, G. Zhu, Y. Shan, Z. Liang, L. Zhang and Y. Zhang, J. Sep. Sci., 2010, 33, 3194–3200. 7 X. Wang, K. Li, E. Adams and A. V. Schepdael, Electrophoresis, 2014, 35, 119–127. 8 G. Massolini and E. Calleri, J. Sep. Sci., 2005, 28, 7–21. 9 Y.-l. Nie and W.-h. Wang, Chromatographia, 2009, 69, S5–S12. 10 C. I. C. Silvestre, P. C. A. G. Pinto, M. A. Segundo, M. L. M. F. S. Saraiva and J. L. F. C. Lima, Anal. Chim. Acta, 2011, 689, 160–177. 11 C. Cao, W. Zhang, L. Fan, J. Shao and S. Li, Talanta, 2011, 84, 651–658. 12 M. Shanmuganathan and P. Britz-McKibbin, Anal. Chim. Acta, 2013, 773, 24–36. 13 C. Chang, G. Xu, Y. Bai, C. Zhang, X. Li, M. Li, Y. Liu and H. Liu, Anal. Chem., 2013, 85, 170–176. 14 M. Geiger, A. L. Hogerton and M. T. Bowser, Anal. Chem., 2012, 84, 577–596. 15 Y. Fan and G. K. E. Scriba, J. Pharm. Biomed. Anal., 2010, 53, 1076–1090. 16 X. Hai, B.-f. Yang and A. Van Schepdael, Electrophoresis, 2012, 33, 211–227. 17 L. Blanes, M. F. Mora, C. L. Do Lago, A. Ayon and C. D. Garcia, Electroanalysis, 2007, 19, 2451–2456. 18 S.-S. Park, S. I. Cho, M.-S. Kim, Y.-K. Kim and B.-G. Kim, Electrophoresis, 2003, 24, 200–206. 19 J. Qiao, L. Qi, H. Yan, Y. Li and X. Mu, Electrophoresis, 2013, 34, 409–416. 20 B. M. Simonet, A. Rios and M. Valcarcel, Electrophoresis, 2004, 25, 50–56. 21 Z. Tang, T. Wang and J. Kang, Electrophoresis, 2007, 28, 2981–2987. 22 J. Iqbal, Anal. Biochem., 2011, 414, 226–231.


Analyst

23 Y. Li, R. Wojcik and N. J. Dovichi, J. Chromatogr. A, 2011, 1218, 2007–2011. 24 Y. Fang, X. Huang, P. Chen and Z. Xu, BMB Rep., 2011, 44, 87–95. 25 P. Jochems, Y. Satyawali, L. Diels and W. Dejonghe, Green Chem., 2011, 13, 1609–1623. 26 T. Xie, A. Wang, L. Huang, H. Li, Z. Chen, Q. Wang and X. Yin, Afr. J. Biotechnol., 2009, 8, 4724–4733. 27 H.-G. Hicke, M. Becker, B.-R. Paulke and M. Ulbricht, J. Membr. Sci., 2006, 282, 413–422. 28 Y. Zhang, F. Gao, S.-P. Zhang, Z.-G. Su, G.-H. Ma and P. Wang, Bioresour. Technol., 2011, 102, 1837–1843. 29 M. Soleimani, A. Khani and K. Najafzadeh, J. Mol. Catal. B: Enzym., 2012, 74, 1–5. 30 M. L. Verma, C. J. Barrow and M. Puri, Appl. Microbiol. Biotechnol., 2013, 97, 23–39. 31 S. Zhao, X. Ji, P. Lin and Y.-M. Liu, Anal. Biochem., 2011, 411, 88–93. 32 H. Bao, Q. Chen, L. Zhang and G. Chen, Analyst, 2011, 136, 5190–5196. 33 B. Jiang, K. Yang, Q. Zhao, Q. Wu, Z. Liang, L. Zhang, X. Peng and Y. Zhang, J. Chromatogr. A, 2012, 1254, 8–13. 34 L. Jin, K. Yang, K. Yao, S. Zhang, H. Tao, S.-T. Lee, Z. Liu and R. Peng, ACS Nano, 2012, 6, 4864–4875. 35 H. Razmi and R. Mohammad-Rezaei, Biosens. Bioelectron., 2013, 41, 498–504. 36 M. Sakata, A. Funatsu, S. Sonoda, T. Ogata, T. Taniguchi and Y. Matsumoto, Chem. Lett., 2012, 41, 1625–1627. 37 G. Xu, X. Chen, J. Hu, P. Yang, D. Yang and L. Wei, Analyst, 2012, 137, 2757–2761. 38 Y. Zhang, J. Zhang, X. Huang, X. Zhou, H. Wu and S. Guo, Small, 2012, 8, 154–159. 39 C. Wang, S. de Rooy, C.-F. Lu, V. Fernand, L. Moore, Jr, P. Berton and I. M. Warner, Electrophoresis, 2013, 34, 1197– 1202. 40 X. Liu, X. Liu, M. Li, L. Guo and L. Yang, J. Chromatogr. A, 2013, 1277, 93–97. 41 X. Liu, X. Liu, X. Liu, L. Guo, L. Yang and S. Wang, Electrophoresis, 2013, 34, 1869–1876.

Analyst, 2014, 139, 1973–1979 | 1979

View Article Online

Analyst


PAPER


A highly selective turn-on fluorescent probe for Al(III) based on coumarin and its application in vivo† Hongde Xiao,a Kun Chen,a Nannan Jiang,b Dandan Cui,b Gui Yin,*a Jie Wanga and Ruiyong Wang*b In this study, a turn-on coumarin-based fluorescent probe, 7-hydroxy-6-[(2-hydroxy-naphthalen-1ylmethylene)-amino]-4-methyl-chroman-2-one (CN), was developed for detecting Al3+ in aqueous systems. The binding ratio of CN–Al3+ complexes was determined from the Job plot and ESI-MS data to

Received 5th December 2013 Accepted 23rd January 2014

be 1 : 1. The binding constant (Ka) of Al3+ binding to CN was calculated to be 9.55 104 M1 from a Benesi–Hildebrand plot and the detection limit was evaluated to be as low as 0.10 mM (LOD ¼ 3s/slope).

DOI: 10.1039/c3an02247a

CN could be used as an effective fluorescent probe for detecting Al3+ in living HeLa cells. Moreover, CN

www.rsc.org/analyst

could also be applied in the in vivo detection of Al3+ in zebrafish.

Introduction Fluorescent probes have attracted considerable interest due to their high sensitivity, selectivity, rapid response rate and ease of manipulation.1 Turn-on probes, in particular, have a relatively higher sensitivity in comparison with turn-off probes due to the lack of background signal. However, most of the reported turnon probes are single-emission, in other words they monitor only a single wavelength in the uorescence region.2 The possible errors associated with photo-transformation, sensor concentration, and environmental effects3 cannot be ignored. Some excited-state intramolecular proton-transfer (ESIPT) probes4 can exhibit double-emission, and minimize these negative effects to a great extent.5 Furthermore, ESIPT compounds can result in a large Stokes shi caused by the transformation of the excited enol form (E*) to the excited proton-transferred keto form (K*), diminishing the self-absorption to a great extent.6 Also, the relatively longer excitation wavelengths are favorable for biological experiments. Aluminium is the third most abundant element in the earth's crust. Its derivatives are widely used in food additives, water treatments, and medicines.7 The interaction of aluminium ions with biological species leads to harmful and destructive effects.8 According to the World Health Organization (WHO), the average daily human intake of aluminium is approximately 3–10 mg. The tolerable weekly aluminium dietary intake in the human body is estimated to be 7 mg kg1 body weight.9 Over consumption of a

State Key Laboratory of Analytical Chemistry for Life Science, School of Chemistry and Chemical Engineering, Nanjing University, Nanjing 210093, People's Republic of China. E-mail: [email protected]

b

School of Life Science, Nanjing University, Nanjing, 210093, China. E-mail: wangry@ nju.edu.cn † Electronic supplementary 10.1039/c3an02247a

information

1980 | Analyst, 2014, 139, 1980–1986

(ESI)

available.

See

DOI:

aluminium inuences the absorption of calcium in the bowel, resulting in osteomalacia, colic, gastrointestinal problems, anemia, and headaches.10 Decreased liver and kidney function can also be caused by aluminium toxicity.11 More terribly, the toxicity of aluminium, causing damage to the central nervous system, is suspected to be responsible for neurodegenerative diseases such as Alzheimer's disease and Parkinsonism. Moreover, aluminium ions from acidic soils are toxic to plant roots12 and kill sh in acidied water.13 Compared to other transitionmetal ions, however, it has always been troublesome to detect Al3+ due to its poor coordination ability, strong hydration ability, and the lack of spectroscopic characteristics.14 Although many Al3+ uorescent probes have been reported,15 most of them suffer from some limits, such as working well only in anhydrous systems and therefore being unsuitable for practical applications, poor selectivity or sensitivity, weak binding ability, quenched interference caused by Cu2+ and Fe3+ because of chelation enhanced uorescent quenching (CHEQ), tedious synthetic methods or even having to be heated for hours aer mixing with Al3+ to detect uorescent signals. Therefore, it is of great importance to develop an effective Al3+ uorescent probe that can be easily prepared. In this work, we developed a highly selective, sensitive and quickly responsive turn-on uorescent probe 7-hydroxy-6[(2-hydroxy-naphthalen-1-ylmethylene)-amino]-4-methyl-chroman2-one (CN) for Al3+ which worked well in aqueous systems. Strong yellow-green uorescence can be observed under excitation at 365 nm. In uorescence titration experiments, the uorescence intensity did not change when the concentration of Al3+ exceeded 1 equiv. Moreover, CN acted well in the biological experiments. Intracellular Al3+ was detected, with use of a confocal uorescence microscope, in contaminated living cells. We also applied CN to the in vivo detection of Al3+ employing a living zebrash.


View Article Online

Paper

Experiment Equipments


1

H NMR and 13C NMR spectra were recorded on a Bruker Ultrashield 300 MHz NMR spectrometer. Chemical shis were expressed in ppm (in DMSO-d6 or CD3OD; TMS as internal standard) and coupling constants ( J) in Hz. Mass spectroscopy was carried out using Micromass GC-TOF and Agilent Technologies 6540 UHD Accurate-Mass Q-TOF LC/MS instruments. Fluorescence spectra were measured using an Hitachi Fluorescence spectrophotometer-F-4600. The uorescence lifetime was measured with a FLS 920 time-resolved spectroscope (Edinberge). Fluorescent images were captured with an Olympus FV-1000 laser scanning confocal uorescence microscope. Chemicals Methanol was HPLC grade. All of the other reagents were of analytical grade unless otherwise noted. Al(NO3)3$9H2O, LiClO4, NaCl, KCl, MgCl2, CaCl2, Fe(NO3)2$6H2O, FeCl3$6H2O, CoCl2$6H2O, Ni(NO3)2$6H2O, AgBF4, Cr(NO3)3$9H2O, Cd(NO3)2$4H2O, ZnCl2, Hg(ClO4)2$3H2O, CuCl2$2H2O, Pb(NO3)2, Mn(NO3)2, GaCl3 were employed in the selective experiment. For all aqueous solutions, deionized water puried with a MilliQ Advantage Water Purication System was used. Synthesis of 7-hydroxy-4-methyl-coumarin 4-Methylbenzenesulfonic acid (5 g) was added into a solution of resorcinol (11.0 g, 0.1 mol) in toluene (40 mL), and then acetoacetic ester (13.0 g) was added drop wise. The reaction mixture was reuxed for 3 h and monitored by TLC until the resorcinol was completely consumed. Aer cooling, the solution was poured into ice water (100 mL), the yellow precipitate was collected by ltration and washed with cold water. A pale yellow solid was obtained by recrystallization with ethanol (11.3 g, yield: 64%). M.p. 185.0–186.0 C. 1H NMR (300 MHz, CD3OD), 7.61 (d, 1H, J ¼ 8.9 Hz), 6.83 (d, 1H, J ¼ 10 Hz), 6.70 (s, 1H), 6.10 (s, 1H), 2.42 (s, 3H). TOF MS EI: calculated 176.0, found 176.0.

Analyst

was ushed with hydrogen three times to get rid of air. The hydrogenation reaction was performed under a hydrogen atmosphere (6.0 MPa) at 80 C with stirring at 300 r.p.m. for 5 h. Then, the autoclave was cooled to the ambient temperature and ushed twice with nitrogen. Pd–C was ltered to obtain the ethanol solution and the next step was conducted without further treatment. Synthesis of compound 2 Compound 2 was synthesized according to our previous procedures.16 Synthesis of the probe CN 2-Hydroxy-naphthalene-1-carbaldehyde (516 mg, 3 mmol) and a drop of AcOH were added to the ethanol solution of 6-amino-7hydroxy-4-methyl-chroman-2-one obtained from the hydrogenation. Then the solution was reuxed for 3 h. Aer completion of the reaction, the obtained yellow precipitate was ltered and washed several times with cold ethanol to give CN (905 mg) as a pure yellowish solid in 87% yield. 1H NMR (300 MHz, DMSO-d6) d 15.88 (d, J ¼ 8.5 Hz, 1H), 10.58 (s, 1H), 9.52 (d, J ¼ 8.5 Hz, 1H), 8.40 (d, J ¼ 8.4 Hz, 1H), 7.88 (s, 1H), 7.81 (d, J ¼ 9.4 Hz, 1H), 7.69 (d, J ¼ 7.7 Hz, 1H), 7.50 (t, J ¼ 7.4 Hz, 1H), 7.28 (t, J ¼ 7.4 Hz, 1H), 6.82 (d, J ¼ 9.3 Hz, 1H), 6.66 (s, 1H), 3.19 (d, J ¼ 6.3 Hz, 1H), 2.93 (dd, J ¼ 15.8, 5.5 Hz, 1H), 2.60 (dd, J ¼ 15.9, 6.6 Hz, 1H), 1.29 (d, J ¼ 6.9 Hz, 3H). 13C NMR (75 MHz, DMSO-d6) d 176.29, 168.46, 150.50, 149.52, 148.92, 137.90, 134.15, 129.46, 128.42, 126.45, 126.24, 124.92, 123.52, 120.30, 120.01, 116.31, 108.42, 104.36, 36.76, 28.96, 20.68. ESI-HRMS: calculated for (C21H17NO4 H) , m/z 346.1079, found (M H), m/z 346.1087 (Scheme 1). Cell culture HeLa cells were cultured in Dulbecco's modied Eagle's medium (DMEM) containing 10% fetal bovine serum and antibiotics (100 units per mL penicillin and 100 mg mL1 streptomycin), maintained at 37 C in a humidied atmosphere of 95% air and 5% CO2.

Synthesis of 7-hydroxy-4-methyl-6-nitro-2H-chromen-2-one A solution of concentrated nitric acid (4.90 mL, 0.0775 mol) in concentrated sulphuric acid (5.1 mL) was added to a stirred solution of 7-hydroxy-4-methyl-coumarin (13.64 g, 0.0775 mol) in concentrated sulphuric acid (30 mL) at such a rate as to keep the temperature below 5 C. Aer warming to 20 C, the reaction mixture was poured into a stirred mixture of ice cold water. The yellow solid that separated was ltered and washed with hot ethanol thoroughly to obtain a pure product. 1H NMR (300 MHz, DMSO-d6) d 11.98 (s, 1H), 8.27 (s, 1H), 6.97 (s, 1H), 6.32 (d, J ¼ 0.8 Hz, 1H), 2.40 (d, J ¼ 0.8 Hz, 3H).

Cell imaging A fresh stock of HeLa cells was seeded into a glass bottom dish with a density of 1 105 cells per dish, and incubated for 24 h. Subsequently, the cells were exposed to 10 mM CN solution (900 mL DMEM mixed with 100 mL of 100 mM CN solution in ethanol) for 10 min at room temperature. The solution was then removed, and the cells were washed with PBS (2 mL 3) to clear the CN molecules attached to the surface of the cells. Aerwards, the cells were incubated for 15 min with 10 mM Al(NO3)3 DMEM

Synthesis of compound 1 Compound 1 was obtained from the hydrogenation carried out in a 100 mL stainless steel autoclave. 7-Hydroxy-4-methyl-6nitro-2H-chromen-2-one (663 mg, 3 mmol), Pd–C (10%, 70 mg) and EtOH (40 mL) were added into the autoclave. The autoclave


Scheme 1

Synthesis of probe CN.

Analyst, 2014, 139, 1980–1986 | 1981

View Article Online

Analyst

Paper

solution at room temperature. The culture medium was removed, and the treated cells were washed three times with PBS (2 mL 3) before observation. Fluorescence imaging was performed by confocal laser scanning microscopy (Olympus, FV-1000; lex ¼ 458 nm; uorescent signals were collected at 480– 580 nm). The images were captured using a photomultiplier.


Fluorescence detection of Al3+ in zebrash Zebrash were kept at 28 C and maintained at optimal breeding conditions. For mating, male and female zebrash were maintained in one tank at 28 C on a 12 h light/12 h dark cycle, and then the spawning of eggs was triggered by giving light stimulation in the morning. Almost all the eggs were fertilized immediately. All the stages of zebrash were maintained in E3 embryo media (15 mM NaCl, 0.5 mM KCl, 1 mM MgSO4, 1 mM CaCl2, 0.15 mM KH2PO4, 0.05 mM Na2HPO4, 0.7 mM NaHCO3, 5–10% methylene blue; pH 7.5). For the imaging experiment, ve-day-old zebrash were incubated with 10 mM CN solution (900 mL DMEM mixed with 100 mL of 100 mM CN solution in ethanol) for 20 min at room temperature. The solution was then removed, and the cells were washed with PBS (2 mL 3) to clear the CN molecules attached to the surface of the zebrash. Aerwards, the zebrash were incubated for 20 min with 40 mM Al(NO3)3 DMEM solution at room temperature. The culture medium was removed, and the treated cells were rinsed three times with PBS (2 mL 3) before observation. The treated zebrash were imaged by confocal microscopy.

Results and discussion Fluorescence response of the probe CN to different metal ions When 10 equiv. of various metal ions (Li+, Na+, K+, Mg2+, Ca2+, Fe2+, Fe3+, Co2+, Ni2+, Ag+, Cr3+, Cd2+, Zn2+, Hg2+, Cu2+, Pb2+, Mn2+, Al3+, Ga3+) were added to 50 mM ethanol–Tris–HCl buffer (v/v, 1/9, pH ¼ 7.3) solutions of CN, it was found that Fe3+ produced a color change from light yellow to light orange, which was probably caused by the chromogenic reaction of phenolic hydroxyl with Fe3+ (Fig. S1†). No change in color was observed upon the addition of other metal ions. The absorbance spectra of CN in the presence and absence of Al3+ are shown in Fig. S3a.† The molar extinction coefficient of CN was evaluated to be 1.98 104 L mol1 cm1 (in the presence of Al3+) and 1.80 104 L mol1 cm1 (in the absence of Al3+) respectively (lmax ¼ 442 nm). Under excitation at 365 nm with a hand-held UV lamp, a solution of the probe CN was nearly non-uorescent (Fig. 1). Aer the addition of Al3+ and Ga3+, yellow-green and yellow uorescence can be observed respectively. For the

Fig. 1 Changes in the fluorescence of CN (50 mM) in ethanol–Tris– HCl buffer (v/v, 1/9, pH ¼ 7.3) solutions upon addition of various metal ions (10 equiv.) under excitation at 365 nm (with a hand-held UV lamp). Mix means CN with all kinds of metal ions.

1982 | Analyst, 2014, 139, 1980–1986

Fluorescence spectra of CN (10 mM) in ethanol–Tris–HCl buffer (v/v, 1/9, pH ¼ 7.3) in the presence of different metal ions upon excitation at 442 nm. Insert: normalized fluorescence spectra of CN upon addition of Al3+ and Ga3+.

Fig. 2

addition of Ga3+, the uorescence intensity was relatively weaker in contrast to that with the addition of Al3+. And the additions of other metal ions made no difference. To explore the utility of CN as an ion-selective uorescent probe more accurately, the uorescence spectra of CN upon addition of various metal ions were measured. As shown in Fig. 2, CN showed nearly no uorescence upon excitation at 442 nm (excitation spectra in the presence and absence of Al3+ are shown in Fig. S3b†). Treatment with Al3+ induced a great increase in uorescence intensity. Two emission peaks at 504 nm and 529 nm can be detected. Moreover, the addition of Ga3+ also resulted in an obvious enhancement of uorescence with an emission peak at 535 nm. Thus, the 6 nm red-shi between Al3+ and Ga3+ caused uorescent emissions, the different shapes of the emission peaks and the various uorescence intensities are distinguishable enough for the probe to discriminate between Al3+ and Ga3+. The additions of other metal ions, Li+, Na+, K+, Mg2+, Ca2+, Fe2+, Fe3+, Co2+, Ni2+, Ag+, Cr3+, Cd2+, Zn2+, Hg2+, Cu2+, Pb2+, Mn2+ in our experiment, did nothing to the uorescence of the probe. And particularly, Cr3+, which cannot be differentiated from Al3+ by many reported probes, still caused no change.

Fluorescence measurement The quantitative uorescence titration experiment was carried out using 10 mM CN in the presence of different concentrations of Al3+ from 0 to 5.0 equiv. in ethanol–Tris–HCl buffer (v/v, 1/9, pH ¼ 7.3) solution. As shown in Fig. 3, CN exhibited remarkable uorescence enhancements (130 fold) at 504 nm and 529 nm respectively, with the quantum yields increasing from 0.003 to 0.51 (using quinoline as the reference with a known F value of 0.55 in 0.05 M H2SO4). This enhancement can be ascribed to the cation-induced inhibition of the ESIPT process. The ESIPT process leads to very weak uorescence.17 If the hydroxyls are coordinated with metal ions, metal coordination will remove this proton and enhance the uorescence.


View Article Online


Paper

Fluorescence spectra of CN (10 mM) in ethanol–Tris–HCl buffer (v/v, 1/9, pH ¼ 7.3) solution in the presence of different concentrations of Al3+ (0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.5, 2.0, 2.5, 3.0, 4.0, 5.0 equiv.) upon excitation at 442 nm. Insert: fluorescence intensity changes at 529 nm. Fig. 3

Analyst

Fig. 5 Fluorescence intensities at 529 nm of CN (1 mM) in ethanol– Tris–HCl buffer (v/v, 1/9, pH ¼ 7.3) solution as a function of [Al3+]. The excitation wavelength was 442 nm and the observed wavelength was 529 nm.

Job plot of the CN–Al3+ complexes in ethanol–Tris–HCl buffer (v/v, 1/9, pH ¼ 7.3) solution, keeping the total concentration of CN and Al3+ at 10 mM. The monitored wavelength was 516 nm.

Fig. 6 Fig. 4 Benesi–Hildebrand plot of CN with Al3+ in ethanol–Tris–HCl buffer (v/v, 1/9, pH ¼ 7.3). The excitation wavelength was 442 nm and the observed wavelength was 529 nm.

The ESIPT process, and a signicant emission enhancement can be observed.18 The luminescence lifetime of CN + Al3+ was 2.75 ns in 529 nm (Fig. S4†). Aer the concentrations of Al3+ increased up to 1.0 equiv., the uorescence intensities did not change. The association constant (Ka) was calculated to be 9.55 104 M1 from a Benesi–Hildebrand plot19 (Fig. 4, R2 ¼ 0.99886), demonstrating CN's strong binding capacity to Al3+. To determine the detection limit of CN to Al3+, a uorescence titration was conducted, employing 1 mM CN in the presence of different concentrations of Al3+ from 0 to 1.0 equiv. in ethanol– Tris–HCl buffer (v/v, 1/9, pH ¼ 7.3) solution (Fig. 5, R2 ¼ 0.998). The detection limit was evaluated to be as low as 0.10 mM. Investigation of the binding mode In order to explore the binding stoichiometry between CN and Al3+, a Job plot experiment was conducted. As shown in Fig. 6,


the uorescence intensities at 529 nm were plotted against the molar fractions of Al3+ keeping the total concentration constant (10 mM). The maximum uorescence intensity, when the molar fraction of Al3+ was 0.5, indicated a 1 : 1 stoichiometry for CN– Al3+ complexes. In addition, the ESI-MS (negative mode) data (Fig. S2†) of the complex showed m/z ¼ 967.00 for CN with Al3+, corresponding to [2(CN 2H+ + Al3+ + MeOH + NO3 + H2O) H+] (calculated m/z ¼ 967.20), and conrmed a 1 : 1 stoichiometry for the CN–Al3+ complex. The proposed structure of the CN–Al3+ complex is shown in Scheme 2. Competition experiment and the effect of pH To examine the selectivity for Al3+ in a complex background of potentially competing species, the uorescence enhancement of CN with Al3+ was investigated in the presence of other metal ions. When CN was treated with 1 equiv. of Al3+ in the presence of various metal ions (K+, Na+, Ca2+ were 50 equiv., others were

Analyst, 2014, 139, 1980–1986 | 1983

View Article Online

Analyst


Scheme 2

Paper

Proposed structure of the CN–Al3+ complex.

Moreover, to test the application extent of CN as an Al3+ probe, the uorescent properties of CN in the absence and presence of Al3+ were investigated at different pH values. As shown in Fig. 8, CN showed nearly no uorescence in the pH range of 2–12. In the presence of 1 equiv. Al3+, the uorescence was not inuenced by an acidic or neutral environment with a pH range of 2–7. However, as the alkalinity increased, the uorescence intensity decreased greatly because the formation of Al(OH)3 reduced the concentration of Al3+ in solution. When the pH value increased to 8.6, no enhanced signal could be detected. To exclude the possibility of the decreased uorescence being caused by the decomposition of the probe, we examined the stability of CN under a basic environment. The probe was kept in the solution (pH ¼ 7.2 and 12) for 24 h, and no new compound was detected by thin-layer chromatography (TLC), indicating that CN was stable under a basic environment. Thus, the decreased uorescence should be ascribed to the transformation of Al(OH)3. Imaging of intracellular Al3+

Fig. 7 Competition experiment of CN towards Al3+ (1 equiv.) in the absence or presence of various metal ions (K+, Na+, Ca2+ were 50 equiv., others were 10 equiv.). Excitation wavelength was 442 nm, emission wavelength was 529 nm.

10 equiv.), the uorescence enhancements caused by Al3+ was retained with all the other kinds of metal ions (Fig. 7). Particularly, Fe3+ and Cu2+ also did not quench the uorescence of CN enhanced by Al3+, which is a common phenomenon for many of the reported Al3+ turn-on probes. Furthermore, the crosssensitivity to K+, Na+, Ca2+ (50 equiv.) indicated that the probe would perform well in living cells in which the activities of K+, Na+ and Ca2+ could be quite high.

Fig. 8 Fluorescence intensity of CN (10 mM) at various pH values in ethanol–Tris–HCl buffer (v/v, 1/9, pH ¼ 7.3) solution, in the absence and presence of Al3+ (1 equiv.). Excitation wavelength was 442 nm, emission wavelength was 529 nm.

1984 | Analyst, 2014, 139, 1980–1986

Then, the potential of the application of CN for sensing Al3+ in living cells by bioimaging was investigated. The images of cells were obtained using a confocal uorescence microscope. When HeLa cells were incubated with CN (10 mM) for 20 min at room temperature, nearly no uorescence could be observed (Fig. 9a). Aer the treated cells were incubated with Al3+ (20 mM) in the culture medium for 20 min at room temperature, a bright yellow-green uorescence was observed in the HeLa cells (Fig. 9d). Moreover, the merged picture (Fig. 9f) conrms that the uorescence signals are only located in the intracellular area. Moreover, the cell morphology remained in good condition aer intake of CN, indicating the great cytocompatibility and low toxicity of the probe. Accordingly, the CN probe is cell membrane permeable and could be applied in the detection of Al3+ within living cells.

Fig. 9 Fluorescence images of Al3+ in HeLa cells using the CN probe upon excitation at 458 nm. (a) Fluorescent image, (b) bright field image and (c) merged image of HeLa cells incubated with the CN probe (10 mM) for 20 min. (d) Fluorescent image, (e) bright field image and (f) merged image of HeLa cells incubated with Al3+ (20 mM) for 20 min after treating with the CN probe.


View Article Online


Paper

Fig. 10 Fluorescence images of Al3+ in zebrafish using the CN probe upon excitation at 458 nm. (a) Fluorescent image, (b) bright field image and (c) merged image of zebrafish incubated with the CN probe (10 mM) for 20 min. (d) Fluorescent image, (e) bright field image and (f) merged image of CN-loaded zebrafish incubated with Al3+ (40 mM) for 20 min.

Imaging of Al3+ in zebrash A further exploratory effort was made to determine if the CN probe could be used for imaging Al3+ in zebrash. As shown in Fig. 10, ve-day-old zebrash, treated with the CN probe (10 mM) for 20 min at room temperature, showed very weak uorescence (Fig. 10a). Moreover, the CN-loaded zebrash treated with 40 mM Al3+ give a strong yellow-green uorescence (Fig. 10d). These in vivo studies indicated that the CN probe is suitable for monitoring the distribution of Al3+ in living bodies.

Conclusions In summary, an easily synthesized coumarin-based Al3+ selective uorescent probe, CN, was synthesized and characterized. The binding mode was proved to be 1 : 1 and the detection limit was calculated to be as low as 0.10 mM. CN could be applied in the detection of Al3+ in living cells, and studies towards the in vivo detection of Al3+ were also carried out successfully using zebrash.

Acknowledgements This work was supported by the National Basic Research Program of China (no. 2014CB846004) and the Jiangsu Province Science Technology Innovation and Achievements Transformation Special Fund (BY2012149).

Notes and references 1 (a) R. Azadbakht and J. Khanabadi, Inorg. Chem. Commun., 2013, 30, 21–25; (b) Y. M. Yang, Q. Zhao, W. Feng and F. Y. Li, Chem. Rev., 2013, 113, 192–270. 2 (a) X. B. Huang, Y. Dong, Q. W. Huang and Y. X. Cheng, Tetrahedron Lett., 2013, 54, 3822–3825; (b) Q. Y. Cao, Y. M. Han, H. M. Wang and Y. Xie, Dyes Pigm., 2013, 99, 798–802; (c) C. J. Gao, X. J. Jin, X. H. Yan, P. An, Y. Zhang, L. L. Liu, H. Tian, W. S. Liu, X. J. Yao and Y. Tang, Sens. Actuators, B, 2013, 176, 775–781.


Analyst

3 S. Banthia and A. Samanta, J. Phys. Chem. B, 2006, 110, 6437– 6440. 4 (a) N. Singh, N. Kaur, R. C. Mulrooney and J. F. Callan, Tetrahedron Lett., 2008, 49, 6690–6692; (b) K. Tiwari, M. Mishra and V. P. Singh, RSC Adv., 2013, 3, 12124–12132; (c) S. Goswami, S. Paul and A. Manna, RSC Adv., 2013, 3, 10639–10643. 5 W. H. Ding, W. Cao, X. J. Zheng, D. C. Fang, W. T. Wong and L. P. Jin, Inorg. Chem., 2013, 52, 7320–7322. 6 (a) K. Ando, S. Hayashi and S. Kato, Phys. Chem. Chem. Phys., 2011, 13, 11118; (b) C. C. Hsieh, C. M. Jiang and P. T. Chou, Acc. Chem. Res., 2010, 43, 1364; (c) G. Y. Li, G. J. Zhao, Y. H. Liu, K. L. Han and G. Z. He, J. Comput. Chem., 2010, 31, 2109; (d) J. N. Li, M. Pu, D. C. Fang, M. Wei, J. He and D. E. Evans, J. Mol. Struct., 2012, 1015, 106; (e) G. Gui, Z. Lan and W. Thiel, J. Am. Chem. Soc., 2012, 134, 852. 7 E. Oliveira, H. M. Santos, J. L. Capelo and C. Lodeiro, Inorg. Chim. Acta, 2012, 381, 203–211. 8 G. Frantzios, B. Galatis and P. Apostolakos, New Phytol., 2000, 145, 211–214. 9 B. Valeur and I. Leray, Coord. Chem. Rev., 2000, 205, 3–40. 10 (a) G. D. Fasman, Coord. Chem. Rev., 1996, 149, 125; (b) P. Nayak, Environ. Res., 2002, 89, 101. 11 (a) C. S. Cronan, W. J. Walker and P. R. Bloom, Nature, 1986, 324, 140; (b) G. Berthon, Coord. Chem. Rev., 2002, 228, 319. 12 C. Stephen, W. Smith, S. Heng, H. E. Heidepriem, D. A. Abell and T. M. Monro, Langmuir, 2011, 27, 5680–5685. 13 N. E. W. Alstad, B. M. Kjelsberg, L. A. Vøllestad, E. Lydersen and A. B. S. Pol´ eo, Environ. Pollut., 2005, 133, 333. 14 (a) Y. Zhao, Z. Lin, H. Liao, C. Duan and Q. Meng, Inorg. Chem. Commun., 2006, 9, 966–968; (b) K. Soroka, R. S. Vithanage, D. A. Phillips, B. Waker and P. K. Dasgupta, Anal. Chem., 1987, 59, 629–636. 15 (a) X. Sun, Y. W. Wang and Y. Peng, Org. Lett., 2012, 14, 3420– 3423; (b) A. Sahana, A. Banerjee, S. Lohar, B. Sarkar, S. K. Mukhopadhyay and D. Das, Inorg. Chem., 2013, 52, 3627–3633; (c) S. Kim, J. Y. Noh, K. Y. Kim, J. H. Kim, H. K. Kang, S. W. Nam, S. H. Kim, S. Park, C. Kim and J. H. Kim, Inorg. Chem., 2012, 51, 3597–3602; (d) S. Guha, S. Lohar, A. Sahana, Ar. B. erjee, D. A. San, M. G. Babashkina, M. P. Mitoraj, M. Bolte, Y. Garcia, S. K. M. adhyay and D. Das, Dalton Trans., 2013, 42, 10198– 10207; (e) Y. K. Jang, U. C. Nam, H. L. Kwon, I. H. Hwang and C. Kim, Dyes Pigm., 2013, 99, 6–13; (f) S. B. Maity and P. K. Bharadwaj, Inorg. Chem., 2013, 52, 1161–1163; (g) J. Y. Jung, S. J. Han, J. H. Chun, C. M. Lee and J. Y Yoon, Dyes Pigm., 2012, 94, 423–426; (h) L. Y. Wang, H. H. Li and D. R. Cao, Sens. Actuators, B, 2013, 181, 749–755; (i) S. B. Maity and P. K. Bharadwaj, Inorg. Chem., 2013, 52, 1161–1163; (j) J. Y. Jung, S. J. Han, J. H. Chun, C. M. Lee and J. Y Yoon, Dyes Pigm., 2012, 94, 423–426; (k) L. Y. Wang, H. H. Li and D. R. Cao, Sens. Actuators, B, 2013, 181, 749–755; (l) B. Jisha, M. R. Resmi, R. J. Maya and R. Luxmi Varma, Tetrahedron Lett., 2013, 54, 4232– 4236. 16 L. Jiang, L. Wang, M. Guo, G. Yin and R. Y. Wang, Sens. Actuators, B, 2011, 156, 825–831.

Analyst, 2014, 139, 1980–1986 | 1985

View Article Online

Analyst

18 M. M. Henary, Y. G. Wu and C. J. Fahrni, Chem.–Eur. J., 2004, 10, 3015–3025. 19 H. A. Benesi and J. H. Hildebrand, J. Am. Chem. Soc., 1949, 71, 2703–2707.


17 (a) J. S. Wu, W. M. Liu, J. C. Ge, H. Y. Zhang and P. F. Wang, Chem. Soc. Rev., 2011, 40, 3483–3495; (b) M. M. Henary and C. Fahrni, J. Phys. Chem. A, 2002, 106, 5210–5220.

Paper

1986 | Analyst, 2014, 139, 1980–1986


View Article Online

Analyst


PAPER


Development of a single aptamer-based surface enhanced Raman scattering method for rapid detection of multiple pesticides† Shintaro Pang,a Theodore P. Labuzab and Lili He*a The objective of this study was to develop a simple and rapid method that could detect and discriminate four specific pesticides (isocarbophos, omethoate, phorate, and profenofos) using a single aptamerbased capture procedure followed by Surface Enhanced Raman Spectroscopy (SERS). The aptamer is a single stranded DNA sequence that is specific to capture these four pesticides. The thiolated aptamer was conjugated onto silver (Ag) dendrites, a nanostructure that can enhance the Raman fingerprint of pesticides, through Ag–thiol bonds. It was then backfilled with 6-mercaptohexanol (MH) to prevent nonspecific binding. The modified SERS platform [Ag–(Ap + MH)] was then mixed with each pesticide solution (P) for 20 min. After capturing the pesticides, the Ag–(Ap + MH)–P complex was analyzed under a DXR Raman microscope and TQ Analyst software. The results show that the four pesticides can be captured and detected using principal component analysis based on their distinct fingerprint Raman peaks. The limits of detection (LODs) of isocarbophos, omethoate, phorate, and profenofos were 3.4 mM


(1 ppm), 24 mM (5 ppm), 0.4 mM (0.1 ppm), and 14 mM (5 ppm) respectively. This method was also validated successfully in apple juice. These results demonstrated the super capacity of aptamer-based

DOI: 10.1039/c3an02263c

SERS in rapid detection and discrimination of multi-pesticides. This technique can be extended to detect

www.rsc.org/analyst

a wide range of pesticides using specific aptamers.

Introduction The development of rapid detection techniques for pesticide residual analysis has become a hot topic in recent years due to increased application of pesticides and fear of their health deteriorating effects.1,2 This eld of analysis is termed “QuEChERS” for nding methods that include making the process to be quick, easy, cheap, effective, rugged, and safe.3 Traditional/currently-used methods mostly apply chromatography (i.e. GC or LC) coupled with MS.4–6 Despite its sensitivity and capability to detect multiple residues quantitatively, this method carries several disadvantages including extensive sample preparation (i.e. extraction, ltration, etc.), need for technical expertise and high cost. Many alternative methods, such as ELISA,7–9 radioimmunoassay10,11 and multiarray biosensor methods,12 have been developed for rapid and sensitive detection of pesticide

a

Department of Food Science, University of Massachusetts, Amherst, MA, USA. E-mail: [email protected]; Tel: +1 413 545 5847

b

Department of Food Science and Nutrition, University of Minnesota, St. Paul, MN 55108, USA † Electronic supplementary information (ESI) available: Supporting gures that detail some of the secondary steps that were taken into account in order to develop the single aptamer based SERS method for detecting multiple pesticides. Fig. S1–S7. See DOI: 10.1039/c3an02263c


residues. In these cases, a recognition element captures the target analyte, and in doing so, the receptor–analyte complex produces a biochemical signal (i.e. change in pH, color, radioactivity, charge potential, etc.) that can be qualitatively or quantitatively determined by an appropriate transducer (i.e. signal probe).13 However, these methods are not without drawbacks. For example, the recognition element might exhibit broad specicity, potentially capturing substrates that are unrelated. The signals produced by the transducer are also oen secondary, meaning that the output can only tell us if something was captured. This can be disadvantageous as the analyst will not be sure if the signal came from the target analyte or from some other compound that triggered a similar signal. Surface-enhanced Raman scattering (SERS) is a powerful spectroscopic technique utilizing nanotechnology and Raman spectroscopy that can detect traces of closely adsorbed molecules on metallic nanostructures (oen gold or silver).14 Even though SERS methods can be conditioned to be sensitive enough to detect a single molecule, SERS alone will not singlehandedly separate compounds present in a sample. This might prove to be disadvantageous, especially when detecting trace amounts of a target in a complex matrix like food. In this case, Raman signals from the target analyte will be drowned out by the signal of other ingredients/compounds present, thus making it impossible to detect anything.

Analyst, 2014, 139, 1895–1901 | 1895

View Article Online


Analyst

To overcome the non-specic nature of SERS, aptamers can be deployed as suitable capture agents. Aptamers are single stranded oligonucleotides that can be synthesized in vitro to capture target molecules. They have become increasingly popular as a capture agent because they are adaptable to various targets, convenient in screening, reproducible for synthesis, versatile in labelling, immobilizing, signalling and regenerating.15,16 In addition, recent technological advancements have made it faster and cheaper to create new aptamers, particularly through the SELEX (Systemic Evolution of Ligands by Exponential Enrichment) protocol.17 An aptamer based SERS method has been previously evaluated on proteins in liquid foods.18 However, the problem of nonspecic binding by other food components has not been solved. The objective of this study was to evaluate the feasibility of an aptamer-based SERS technique for pesticide detection in complex liquid food (e.g. apple juice) using a single aptamer that was previously synthesized to be specic to not one, but four commercially available pesticides (isocarbophos, omethoate, profenofos, and phorate).19 The optimization of aptamer conjugation to eliminate the nonspecic binding in apple juice as well as the feasibility of multi-detection was emphasized and discussed. To the best of our knowledge, very few studies have been published on aptamer-based SERS for small molecule detection.20,21 No similar study has been reported for detection in a complex food matrix.

Experimental section Materials All the chemicals were of analytical reagent grade and were purchased through Fisher Scientic unless otherwise noted. The multiple pesticide binding aptamer (SS2-55) reported previously19 with the sequence 50 -AAG CTT GCT TTA TAG CCT GCA GCG ATT CTT GAT CGG AAA AGG CTG AGA GCT ACG C-30 with disulde (S–S) modication at the 50 end was purchased through Eurons Operon MWG (Ebersberg, Germany). Apple Juice (Stop N Shop, MD, USA) was purchased from a local grocery store. Preparation of dendritic silver (Ag) nanoparticles Silver (Ag) dendrites were synthesized in accordance with previously published methods.22,23 Briey, a zinc metal plate was rinsed with 1 M HCl to remove any metal oxides forming on the outer layer, then rinsed with double distilled water and dried. The zinc plate was then immersed in a 200 mM AgNO3 (aq) solution for exactly 1 min, which produced nanoparticle size diameters of z50 nm. Aer incubation, the silver dendrites formed on the surface of the zinc plate were gently peeled off and washed with double distilled water several times. SEM images obtained from FEI Magellan 400 (FEI, Oregon, USA) conrmed that consistent nanoparticle sizes of z50 nm in diameter were crystallized and form a dendritic structure (Fig. 1). This unique structure can provide locally consistent enhancement factors and has been previously demonstrated as a sensitive and reliable SERS substrate.22,24,25

1896 | Analyst, 2014, 139, 1895–1901

Paper

Fig. 1 SEM image of (A) silver dendrites and (B) a thiolated aptamer (S2-55) conjugated onto silver dendrites.

Conjugation of deprotected aptamer onto Ag dendrites The step by step sequence for developing this single aptamerbased SERS method is illustrated in Fig. 2. The aptamer was rst dissolved in a 1 TE buffer (pH 7.4) to give a stock concentration of 129 mM. 100 ml of the aptamer stock solution was then added to 10 mM Tris(2-carboxyethyl) phosphine hydrochloride (TCEP) (i.e. the concentration ratio was 1 : 77) and incubated for 1 hour in order to reduce the disulde (S–S) to thiol (SH) groups, which have stronger binding affinity to the Ag surface. Then, 30 ml of the aptamer solution was mixed with 370 ml of 1 M phosphate buffer (pH 7), aer which 100 ml (z400 mg) Ag was added. This mixture was homogenized by mixing them together briey with a micropipette tip. The mixture was then incubated at room temperature for 4 hours on a Nutating mixer (Fisher Scientic) at the speed of 24 rpm to allow the aptamer to conjugate onto the Ag through Ag–thiol binding interactions. Addition of a blocker molecule (i.e. 6-mercaptohexanol) onto the silver–aptamer complex (Ag–Ap) In order to eliminate all forms of non-specic binding interactions with Ag, 40 mM 6-mercaptohexanol (MH) was incubated with Ag–Ap for 1 hour, aer which it was rinsed 3 times with double distilled water to fabricate our modied Ag dendrites [Ag–(Ap + MH)]. Detection of multiple specic pesticides using aptamer-based SERS The pesticides (isocarbophos, omethoate, phorate, and profenofos) were spiked in a buffer containing 300 mM NaCl, 50 mM KCl, 10 mM MgCl2, and 50 mM Tris–HCl, pH 8.3. 5 ml of Ag–(Ap + MH) was added to 400 ml of spiked buffer solution, stirred and incubated for 20 minutes. Aer incubation, the mixture was centrifuged at 6000 g for 1 min to allow the modied Ag dendrites and the captured pesticides [Ag–(Ap + MH)–P] to settle to the bottom. The supernatant was then removed and Ag–(Ap + MH)–P was quickly rinsed with double distilled water twice before being deposited onto a glass slide to dry. If the pure pesticide solution is tested, this washing and drying step can be minimized. But if the detection is in a complex matrix, i.e. apple juice, a quick wash step is necessary to remove food components. Raman instrumentation Aer drying, the samples were analyzed immediately using a DXR Raman Spectro-microscope (Thermo Scientic,


View Article Online


Paper

Analyst

The schematic illustration of the development of the single aptamer-based SERS method for the detection of four specific pesticides (isocarbophos, omethoate, phorate, and profenofos).

Fig. 2

Madison, WI) with the following conditions: a 10 confocal microscope objective (3 mm spot diameter and 5 cm1 spectral resolution), 780 nm excitation wavelength, 5 mW laser power and 50 mm slit width for 2 s integration time. OMNIC™ soware version 9.1 was used to control the Raman instrument. Eight spots were selected randomly for each sample within the range of 100–3300 cm1. Data analysis SERS spectral data were analyzed using TQ Analyst soware (version 8.0) from Thermo Scientic. The SERS spectra obtained from the multiple spots from each sample were averaged and compared against other samples. Second derivative transformation and smoothing were applied at times in order to reduce spectral noise and to separate overlapping bands. The variances of spectral data between spots and samples were then assessed using principal component analysis (PCA). This method focuses on a multidimensional data set to the most dominant features while removing random variation so that the principal components can be used to capture the variation between spectra. This discriminant analysis is thus useful in evaluating SERS spectra for variance within a class and between classes. Generally speaking, if two data clusters (classes) do not overlap, then it means they are signicantly different at the p ¼ 0.05 level. The limit of detection, LOD, was determined to be the lowest concentration of the data cluster that can be separated from the negative control in the PCA plot. Partial least square (PLS), a multivariate analysis model, was also applied to see if there was a linear relationship between the obtained spectral peak intensities and the actual spiked concentrations.

Results and discussion Optimization of aptamer conjugation with Ag In order to maximize Ag–Ap conjugation, various concentrations of aptamers up to 0.512 mM were initially added to 100 ml Ag in water as the solvent. The SERS spectra were then analyzed


to pick out the concentration with the highest aptamer peak (i.e. highest amount of aptamers bound to Ag).26 Interestingly, the aptamer peak intensities started to decrease aer 0.128 mM (Fig. S1†). This could be explained due to the intermolecular electrostatic repulsion from the negatively charged ssDNA molecules. In order to minimize the intermolecular electrostatic repulsion, 1 M phosphate buffer (pH 7) was used as a medium to increase the ionic strength, thus electrostatically shielding the charged oligonucleotides.27 This caused the aptamer peak intensities to increase signicantly at higher aptamer concentrations up to 5 mM (Fig. 3), thus proving the need for a higher level of salts to increase the surface coverage. Other media such as Tris buffers, chloride and nitrate salts were also tested to evaluate their effectiveness in increasing the

Fig. 3 Raman spectra showing the concentration optimization of a thiolated aptamer (S2-55) on 100 ml Ag dendrites in 1 M phosphate buffer, pH 7. As the thiolated aptamer concentration increased, the aptamer peaks, notably around 1330 cm1, increased up to 5 mM. At the same time, the Raman intensities of other competing molecules in the buffer decreased. At 5 mM, the buffer peak 970 cm1 is no longer reflected on the Raman spectra, suggesting complete displacement by the thiolated aptamer.

Analyst, 2014, 139, 1895–1901 | 1897

View Article Online

Analyst

Paper


aptamer optimization, but they brought other problems (i.e. clumping/reaction of Ag nanoparticles) and subsequent reduction in SERS peak intensities. The incubation time to optimize the aptamer surface coverage was also evaluated (up to 24 hours) and it was found that aer 4 hours, the coverage was optimized (Fig. S2†). This result was consistent with other ndings that showed thiol-derivative oligonucleotides having an optimized surface coverage aer 4 hours.27,28

Optimization of the MH concentration and incubation time with Ag–Ap Since aptamers are relatively large molecules, maximizing their surface coverage on Ag does not ensure elimination of nonspecic binding to Ag because steric hindrance can occur. The small, empty spaces between the aptamers can become nonspecic binding sites for smaller molecules (e.g. pesticides, salts, and food matrix). Furthermore, although Ag can bind specically to thiolated aptamers through Ag–thiol bonds, Ag will also interact non-specically with the N1 groups present on the A, T, G, and C ring functional groups of the aptamer.27,29 This particular interaction will likely make them incapable of capturing their target agents since their 3D conformation will be unable to change. In order to eliminate non-specic binding and to ensure that the aptamer is free to change its 3D conformation during target capture, a small blocker molecule (i.e. 6-mercaptohexanol) was introduced.30 Since this molecule is relatively very small, it does not interfere with the 3D conformational change that occurs when the aptamer captures its target. In addition, its thiol end has a strong binding affinity to Ag, thus enabling displacement of any non-specic binding of aptamer or other molecule (e.g. TCEP) that might be present in the Ag–Ap complex. Inadvertently, studies have shown that high concentrations and longer incubation times of MH can even displace thiolated oligonucleotides;30,31 thus it was essential to optimize the time/ concentration of added MH to fully cover the empty surface area and to displace non-specic binding, but at the same time, to minimize displacement of the thiolated aptamer. Varying concentrations (0–1 mM) of MH and varying incubation times (0–1 h) were tested and their Raman spectra were analyzed to monitor the appearance of MH peaks and aptamer peak intensities. As the MH concentration and incubation time increased, the Raman peaks attributed to MH increased, but at the same time, the peaks attributed to the aptamer decreased, albeit at a lower extent. The optimized time/concentration was determined to be 40 mM for 1 h because the Raman peaks from MH were visible while the aptamer peak intensities did not drop drastically (Fig. 4). This observation was in line with another study that showed that when MH was introduced into thiolated oligonucleotides conjugated to gold nanoparticles, the oligonucleotides started to become displaced aer using 10 mM of MH for 10 min and a signicant displacement was observed when 100 mM MH was incubated for 10 min.30 One visual test to see if the nanoparticles had been fully covered with the aptamer and MH was by adding them into the capture buffer solution. When bare Ag dendrites were added to

1898 | Analyst, 2014, 139, 1895–1901

Fig. 4 Raman spectra of Ag–Ap and Ag–(Ap + MH). 40 mM MCH was incubated for 1 h with Ag–Ap and analyzed to record the Raman spectra for Ag–(Ap + MH). The Raman peaks at 680 and 1080 cm1 are both attributed to the capture of MH. The decrease in aptamer peaks (e.g. 1330 cm1) is inevitable due to the strong binding dissociation nature of MH.

the buffer, they aggregated immediately. When Ag–Ap was added to the buffer, it remained dispersed for a few minutes, but slowly started to aggregate as time went on. However, when the fully modied Ag dendrites [Ag–(Ap + MH)] were added to the buffer, they remained dispersed for as long as 20 min without any aggregation, thus proving that the Ag dendrites had been fully covered by the aptamer and MH (data not shown). Detection of multiple pesticides using Ag–(Ap + MH) The chemical structures of the four target pesticides are shown in Fig. 5A. Pesticides at varying concentrations (0–0.5 mM) individually or as a mixture were initially spiked in water before Ag–(Ap + MH) was added. However, when the Raman spectra were recorded, there was no noticeable capture of any of the pesticides. Furthermore, among the 8 replicates tested for each sample, the aptamer peaks were very inconsistent (data not shown). This could be due to the instability of the aptamers in a medium that does not have cations, as they are being introduced into the buffers used to select the aptamers.19 In addition, some other aptamer papers have reported the dependency of cations to form stable 3D conformations to ensure the capture of their target agents.32,33 Therefore, the aptamer selection buffer was then employed as a capture buffer. Raman spectra show a huge difference in pesticide capture peaks when the capture buffer was used. As shown in Fig. 5B, each pesticide capture produces distinct peaks at different Raman shis, signifying the capture of the target pesticides. Fig. 5B highlights the most noticeable difference between each pesticide and the control. Their full spectra are shown in the ESI (Fig. S3†). Because each pesticide produces distinct Raman peaks, the type of pesticides being captured can be discriminated based on the peak intensities at various Raman shis. Furthermore, by


View Article Online

Paper

Analyst


are specic to this aptamer produced signicantly different Raman spectral peaks from each other and the control (i.e. no pesticide) (Fig. 6). This shows that the Raman peak changes attributed to the pesticides can be statistically quantied and differentiated. The four pesticides were also added together as a mixture and measured to nd out if they could all be detected at the same time (Fig. 5C). By looking at each distinct Raman peak from each pesticide capture, it was found to be correlated with a Raman peak that appeared in the mixture. Thus, this method provides a method that can identify four specic pesticides with one sample. On the other hand, when acetamiprid, a pesticide not specic to this aptamer was introduced, no visible pesticide capture peaks were seen (data not shown), thus validating the aptamer's high specicity.19 The great advantage of this technique is not only that it can measure multiple target analytes, but it can also simultaneously validate that the captured molecules are the target analytes based on their specic signature peaks at certain Raman peak shis. This unique “selfvalidation” method ensures great accuracy of this technique, superior to other sensor techniques that are based on color, uorescence or electrochemical signals and begins to satisfy many of the requirements of “QueChERS”. Determination of the limit of detection and quantitative capability of this method

Fig. 5 (A) Chemical structure of the four pesticides that is specific to the aptamer being used. (B) Second derivative Raman spectra of an isocarbophos capture peak and the control between 1220 and 1170 cm1; a profenofos capture peak and the control between 1110 and 1060 cm1; a phorate capture peak and the control between 545 and 510 cm1; an omethoate capture peak and the control between 425 and 375 cm1. The control was the modified Ag dendrites [Ag– (Ap + MH)]. The spiked concentration for the four pesticides was 0.5 mM. All samples were kept in a capture buffer with a 20 min incubation period. The raw spectra yielded similar trends (Fig. S4†). (C) A second derivative Raman spectrum reflecting the capture of all four pesticides (isocarbophos, profenofos, phorate, and omethoate) by Ag–(Ap + MH). 125 mM of each pesticide was added. Distinct Raman peaks were produced at different Raman shifts that correlated with each individual pesticide capture peak.

looking at the principal component analysis score, the whole spectrum can be analyzed to see if these peaks are signicantly different from each other. In this case, all four pesticides that


Fig. 7A shows the relationship between the phorate capture peaks and the phorate concentration. The limit of detection (LOD) was determined to be 0.4 mM (0.1 ppm) by PCA (Fig. 7B). The LODs of other pesticides were also determined and are shown in Fig. S5.† Compared with the LOD of phorate, the LODs of isocarbophos (3.5 mM ¼ 1 ppm), omethoate (24 mM ¼ 5 ppm) and profenofos (14 mM ¼ 5 ppm) were much higher, although the reported binding dissociation constants were similar for these four pesticides. One hypothesis is that the phorate molecules aer being captured by the aptamers were positioned at a distance closer to the surface of Ag–(Ap + MH) compared with other molecules or the interaction between phorate and the aptamer resulted in a specic charge transfer mechanism that enhanced the Raman scattering of phorate. More studies are needed to understand this phenomenon. The LOD of pesticides obtained from this method was also compared with

Fig. 6 PCA plot comparing the second derivative Raman spectra of each pesticide (0.5 mM) captured by modified Ag dendrites [Ag–(Ap + MH)] as well as the control. This result shows significant differences between each pesticide capture, proving the detection and discrimination of the four pesticides.

Analyst, 2014, 139, 1895–1901 | 1899

View Article Online


Analyst

Paper

(A) Second derivative Raman spectra of the phorate capture peak between 645 and 600 cm1 when 0, 0.04 mM, 0.4 mM, 4 mM and 40 mM phorate were exposed to the modified Ag dendrites [Ag–(Ap + MH)]. This result shows an increase in peak intensity around 625 cm1 as the phorate concentration increases; (B) PCA plot for Raman spectra of phorate at 0 mM-control (B), 0.04 mM (+), 0.4 mM (O), 4 mM (,) and 40 mM (C). A significant difference was seen beginning at 0.4 mM. Fig. 7

the currently used methods4–6 (i.e. chromatography). The sensitivity of this SERS method was found to be slightly lower. Nevertheless, the LODs mainly depend on the binding dissociation constants of the aptamer. Aptamers that are specic to smaller molecules tend to have a lower binding affinity due to the limited binding sites available for the aptamer to bind to the molecule.34 Thus, this method can be further improved by selecting and applying an aptamer of higher binding dissociation constants. The quantitative capability of this method to phorate was also evaluated. The partial least square line depicting the relationship between the calculated phorate concentration based on the phorate capture peak and the actual spiked phorate concentration added shows a linear relationship within the range of 0– 3.8 mM (10 ppm) with a root mean square error coefficient of 0.9628 (Fig. S6†). However, not all pesticides had such a linear response. It was found that, at high concentrations (i.e. 0.5 mM), the captured omethoate produced the most profound peak intensities among the four pesticides. However, at lower concentrations (i.e. 4.8–48 mM), the captured omethoate produced little enhanced peaks compared to the other pesticides, suggesting that Raman peak intensities are not linearly correlated with the omethoate concentration. Little is known about this relationship and more studies will be needed to understand the molecular mechanism behind this phenomenon.

Method validation in apple juice to ensure elimination of nonspecic binding and applicability in the food matrix When Ag was exposed to liquid foods (e.g. apple juice), nonspecic binding interaction occurred rampantly, causing multiple Raman peaks to form that easily drowned the target agent peaks. In order to validate the elimination of non-specic binding, the prepared complex [Ag–(Ap + MH)] was incubated with apple juice by adding 100 ml of apple juice into 900 ml of optimized Ag complex in buffer. The spectrum in apple juice showed no signicant changes in peaks compared with the one in buffer (Fig. 8A), proving that non-specic binding was

1900 | Analyst, 2014, 139, 1895–1901

Fig. 8 (A) Raman spectra of modified Ag dendrites [Ag–(Ap + MH)] mixed with (from top to bottom) capture buffer, apple juice (diluted 1 : 10) (AJ), and phorate (Ph) (0.5 mM) spiked in AJ. Spectral results show no differences between Ag–(Ap + MH) exposed to buffer and AJ, suggesting that nonspecific binding has been eliminated. However, when Ag–(Ap + MH) was exposed to Ph & AJ, a huge spectral change was observed, proving the capture of phorate. (B) A second derivative Raman spectrum reflecting the capture of all four pesticides (isocarbophos, profenofos, phorate, and omethoate) by Ag–(Ap + MH) in AJ. The obtained results were similar to the experiments done without AJ (Fig. 5C).

eliminated. When phorate was added to apple juice, intense phorate capture peaks appeared, signifying the capture of phorate in apple juice. Analysis of the other pesticides spiked with apple juice was also performed and the results showed similar trends as phorate (Fig. S7†). Furthermore, a mixture of all four pesticides yielded results similar to the same experiment done without apple juice (Fig. 8B), although the sensitivity for each pesticide was different. This may be due to a change in the incubation environment (i.e. pH, the presence of other solutes) that deterred from the optimized condition. It is therefore worthy to note that an optimized dilution of the apple juice was needed to produce the most intense phorate capture peaks. At higher apple juice concentrations, the food components might alter the optimized condition for the Ag complex to function. Nevertheless, this method can be easily applied to other food and beverage samples for more effective capture and analysis.

Conclusions In summary, a single aptamer based SERS method for detecting multiple specic pesticides in complex liquid food (i.e. apple juice) was established. The elimination of nonspecic binding


View Article Online


Paper

was achieved by optimization of aptamer conjugation and with the blocker molecule MH to ll the empty spaces. The detection, discrimination, and validation of multiple pesticides using an aptamer-based SERS method were achieved based on their distinct Raman ngerprint peaks. Furthermore, the simultaneous detection of multiple pesticides was shown using the SERS spectra obtained. The total analytical time for measuring six samples was 40 min. The developed aptamer SERS method shows great potential to analyze pesticide residues in apple juice. Further experiments are needed to explore the application in other food matrices.

Acknowledgements This project was supported by the Agriculture and Food Research Initiative Program of the US Department of Agriculture (USDA-AFRI, grant no.: 2012-67017-30194).

Notes and references 1 B. Jin, L. Xie, Y. Guo and G. Pang, Food Res. Int., 2012, 46, 399–409. 2 D. Liu, W. Chen, J. Wei, X. Li, Z. Wang and X. Jiang, Anal. Chem., 2012, 84, 4185–4191. 3 M. Anastassiades, S. Lehotay, and D. ˇ Stajnbaher, in 18th Annual waste testing and quality symposium proceeding, Arlington, VA, 2002, pp. 231–241. 4 J. Lee and H. K. Lee, Anal. Chem., 2011, 83, 6856–6861. 5 M. A. Mart´ınez-Uroz, M. Mezcua, N. Belmonte Valles and A. R. Fern´ andez-Alba, Anal. Bioanal. Chem., 2012, 402, 1365–1372. 6 J. S. Punzi, M. Lamont, D. Haynes and R. L. Epstein, Outlooks Pest Manage., 2005, 16, 131–137. 7 L. Wang, Q. Zhang, D. Chen, Y. Liu, C. Li, B. Hu, D. Du and F. Liu, Anal. Lett., 2011, 44, 1591–1601. 8 Z.-L. Xu, H. Deng, X.-F. Deng, J.-Y. Yang, Y.-M. Jiang, D.-P. Zeng, F. Huang, Y.-D. Shen, H.-T. Lei, H. Wang and Y.-M. Sun, Food Chem., 2012, 131, 1569–1576. 9 X. Hua, L. Wang, G. Li, Q. Fang, M. Wang and F. Liu, Anal. Methods, 2013, 5, 1556. 10 C. D. Ercegovich, R. P. Vallejo, R. R. Gettig, L. Woods, E. R. Bogus and R. O. Mumma, J. Agric. Food Chem., 1981, 29, 559–563. 11 J. C. Hall, R. J. A. Deschamps and K. K. Krieg, J. Agric. Food Chem., 1989, 37, 981–984. 12 Y. Guo, J. Tian, C. Liang, G. Zhu and W. Gui, Microchim. Acta, 2013, 180, 387–395.


Analyst

13 S. White, in Handbook of Food Analysis, Second Edition-3 Volume Set, ed. L. Nollet, CRC Press, 2004, pp. 2133–2148. 14 H. Im, K. C. Bantz, N. C. Lindquist, C. L. Haynes and S.-H. Oh, Nano Lett., 2010, 10, 2231–2236. 15 J.-W. Chen, X.-P. Liu, K.-J. Feng, Y. Liang, J.-H. Jiang, G.-L. Shen and R.-Q. Yu, Biosens. Bioelectron., 2008, 24, 66–71. 16 Y. Dong, Y. Xu, W. Yong, X. Chu and D. Wang, Crit. Rev. Food Sci. Nutr., 2013, DOI: 10.1080/10408398.2011.642905. 17 A. D. Ellington and J. W. Szostak, Nature, 1992, 355, 850– 852. 18 L. He, E. Lamont, B. Veeregowda, S. Sreevatsan, C. L. Haynes, F. Diez-Gonzalez and T. P. Labuza, Chem. Sci., 2011, 2, 1579. 19 L. Wang, X. Liu, Q. Zhang, C. Zhang, Y. Liu, K. Tu and J. Tu, Biotechnol. Lett., 2012, 34, 869–874. 20 M. Li, J. Zhang, S. Suri, L. J. Sooter, D. Ma and N. Wu, Anal. Chem., 2012, 84, 2837–2842. 21 F. Barahona, C. L. Bardliving, A. Phifer, J. G. Bruno and C. a. Batt, Ind. Biotechnol., 2013, 9, 42–50. 22 L. He, M. Lin, H. Li and N.-J. Kim, J. Raman Spectrosc., 2010, 41, 739–744. 23 B. D. J. Fang, H. You, P. Kong, Y. Yi and X. Song, Cryst. Growth Des., 2007, 7, 864–867. 24 L. He, T. Rodda, C. L. Haynes, T. Deschaines, T. Strother, F. Diez-Gonzalez and T. P. Labuza, Anal. Chem., 2011, 83, 1510–1513. 25 K. G. M. Laurier, M. Poets, F. Vermoortele, G. De Cremer, J. A. Martens, H. UjI-i, D. E. De Vos, J. Hoens and M. B. J. Roeffaers, Chem. Commun., 2012, 48, 1559–1561. 26 L. He, B. D. Deen, A. H. Pagel, F. Diez-Gonzalez and T. P. Labuza, Analyst, 2013, 138, 1657–1659. 27 T. M. Herne and M. J. Tarlov, J. Am. Chem. Soc., 1997, 119, 8916–8920. 28 K. A. Peterlinz, R. M. Georgiadis, T. M. Herne and M. J. Tarlov, J. Am. Chem. Soc., 1997, 119, 3401–3402. 29 P. Sandstrom, M. Boncheva and B. Akerman, Langmuir, 2003, 19, 7537–7543. 30 S. Park, K. a. Brown and K. Hamad-Schifferli, Nano Lett., 2004, 4, 1925–1929. 31 A. Wijaya, S. B. Schaffer, I. G. Pallares and K. HamadSchifferli, ACS Nano, 2009, 3, 80–86. 32 P. Jayapal, G. Mayer, A. Heckel and F. Wennmohs, J. Struct. Biol., 2009, 166, 241–250. 33 W. S. Ross and C. C. Hardin, J. Am. Chem. Soc., 1994, 116, 6070–6080. 34 M. McKeague and M. C. Derosa, J. Nucleic Acids, 2012, 2012, 748913.

Analyst, 2014, 139, 1895–1901 | 1901

View Article Online

Analyst


COMMUNICATION


Symmetric cyanovinyl-pyridinium triphenylamine: a novel fluorescent switch-on probe for an antiparallel G-quadruplex† Hai Lai,‡a Yijie Xiao,‡a Shengyong Yan,‡a Fangfang Tian,a Cheng Zhong,a Yi Liu,a Xiaocheng Weng*ab and Xiang Zhou*ab

Received 8th December 2013 Accepted 20th January 2014 DOI: 10.1039/c3an02269b www.rsc.org/analyst

In this study, we present a fluorescent switch-on probe based on a cyanovinyl-pyridinium triphenylamine (CPT) derivative that exhibited a 190-fold increase in fluorescence upon binding to G-quadruplexforming oligonucleotide 22AG. This probe showed specificity and selectivity towards an antiparallel G-quadruplex, indicating its promising potential in G-quadruplex imaging.

G-quadruplexes, formed from the folding of guanine (G)-rich nucleic acid sequences by Hoogsteen base-pairing, are of signicant interest among non-canonical nucleic acids owing to their applications in cancer and other diseases.1,2 These G-rich regions are found to be over-represented in telomeres or transcriptional start sites, indicating their multiple biological functions.3 In recent years, the studies on the polymorphism and folding topology of G-quadruplexes have shown that they are suitable candidates for G-quadruplex imaging.4–8 Recently Balasubramanian et al. reported the direct evidence of endogenous G-quadruplex formation and its stabilization using small-molecule ligands.9,10 Therefore, the recognition of G-quadruplexes has received signicant attention for better understanding their functionality and localization in vivo.11–13 Numerous uorescent probes for G-quadruplexes have been developed.14–16 An ideal uorescent probe should exhibit both high selectivity and appropriate uorescence intensity when binding to its target.17 Triphenylamine-based dyes are the common uorescent molecules used in photoelectric materials, uorescence sensing, and cellular imaging.18,19 Moreover, the triphenylamine moiety usually plays the role of donor in various systems and can form a D–p–A system to increase the uorescence.20 Some triphenylamine derivatives have demonstrated

the ability of interaction with AT regions, indicating their potential for sensing nucleic acid structures.21,22 Herein, we design two cyanovinyl-pyridinium triphenylamine (CPT) derivatives for the purpose of G-quadruplex recognition. CPT1 is designed by linking a cationic pyridinium unit to the triphenylamine core via a cyanovinyl group, whereas CPT2 possesses an additional functional branch compared to CPT1. The triphenylamine core acts as the electron donor owing to its electron-rich properties, and N-methylated heterocycles are chosen because of their electron-accepting as well as DNA binding properties.23 The pyridinium triphenylamine (Py-Tp) compounds are non-uorescent in the free state because of the molecular motions around the vinyl bond.24 Furthermore, the pyridylacetonitrile motifs are expected to introduce aggregation-induced emission (AIE) properties to the molecule.25,26 We envision that the molecule would bind to the G-quadruplex and regain its uorescence owing to the conformational change (Fig. 1). The compounds CPT1 and CPT2 were prepared from commercially available materials using simple coupling and successive methylation procedures. Fig. 2 shows that with gradual addition of 22AG DNA to the CPT2 solution, the uorescence intensity increased gradually and eventually reached to 190-fold at the DNA concentration of 10 mM. CPT2 showed negligible uorescence emission in the buffered solution without any oligonucleotide, whereas CPT2 (10 mM) exhibited

a

College of Chemistry and Molecular Sciences, Wuhan University, Wuhan 430072, Hubei, P. R. China. E-mail: [email protected]; [email protected]

b

The State Key Laboratory of Natural and Biomimetic Drugs, State, Key Laboratory of Applied Organic Chemistry, † Electronic supplementary 10.1039/c3an02269b

information

(ESI)

‡ These three authors contributed equally to this work.

1834 | Analyst, 2014, 139, 1834–1838

available.

See

DOI:

Fig. 1

Structures of CPT1 and CPT2.


View Article Online


Communication

Analyst

Fig. 2 Fluorescence spectra of CPT2 (10 mM) upon titration with prefolded 22AG quadruplex DNA in 10 mM Tris (pH 7.4) and 50 mM KCl. [22AG]/mM: (1) 0, (2) 0.5, (3) 1.0, (4) 1.5, (5) 2.0, (6) 3.0, (7) 4.0, (8) 5.0, (9) 6.0, (10) 7.0, (11) 8.0, (12) 9.0, and (13) 10.0. lex is 521 nm.

Fig. 3 Fluorescence intensity enhancement (I/I0) of CPT2 (10 mM) at 620 nm plotted against the 22AG DNA in solution containing (1) 10 mM Tris (pH 7.4) and 50 mM NaCl; (2) 10 mM Tris (pH 7.4); and (3) 10 mM Tris (pH 7.4) and 50 mM KCl.

high uorescence emission at 620 nm aer adding a prefolded 22AG DNA quadruplex in the presence of K+. Consistent with the change in the uorescence intensity, the absorption band of CPT2 showed a bathochromic shi of approximately 20 nm aer adding 22AG DNA, and exhibited signicant hypochromism (nearly 28%, Fig. S1†), both indicating the strong interaction between CPT2 and a G-quadruplex structure of 22AG DNA. The DFT optimization in the excited states of CPT1 and CPT2 at the TD-B3LYP/6-31G(d) level using the Gaussian 09 program27 showed that both compounds have a twisted excited state geometry where the donor, diphenyl amino group, and the acceptor groups are perpendicular to each other (see the ESI†). This phenomenon is also known as the twisted intramolecular charge transfer (TICT) state, which predicts the uorescence quenching and uorescence recovery in highly polar and less polar solvents or a constrained environment, respectively.28 However, CPT1 did not show any signicant uorescence enhancement aer adding 22AG DNA. Its absorption band only showed a 10 nm red-shi (Fig. S2†). The CD spectrum also reects that the CPT1 molecule interacts with 22AG DNA; however, the effect is less pronounced than CPT2 (Fig. S3†) because the monocharged compound has a lower affinity towards DNA, and it is relatively harder to bind to the DNA groove. These results indicated that only CPT2 has great potential as the G-quadruplex uorescent probe. The uorescence measurements were also performed in 10 mM Tris–HCl buffer with K+ or Na+, or without any high concentration of alkali metals (Fig. 3). It has been reported that the G-quadruplex formed in the presence of K+ is more stable than that in the presence of Na+.29 The binding of CPT2 to the K+ induced G-quadruplex was indeed found to be strong, indicating the higher uorescence enhancement in the buffered solution with KCl (190-fold) than that with NaCl (90-fold). Interestingly, in the absence of metal ions (curve 2, Fig. 3), the CPT2-22AG system still shows a uorescence turn-on up to 160-fold. Aer adding 22AG DNA without prefolding, the absorption spectra showed 15 nm red-shi and approximately

30% hypochromism (Fig. S4†), which is similar to the result of absorption spectra in the presence of K+ (Fig. 2b). This might indicate the potential of CPT2 to induce 22AG into the G-quadruplex. To validate the assumption, the structural changes in 22AG DNA caused by CPT2 were studied using CD measurements. In the absence of alkali metal ions, the CD signal did not show any characteristic G-quadruplex-formation of 22AG DNA alone (Fig. 4, curve 1a). Aer adding CPT2 to the 22AG DNA solution, two apparent CD bands gradually appeared. Both negative and positive peaks located at 265 and 295 nm, respectively, conrm an antiparallel quadruplex formation. This transformation indicates that CPT2 could bind to 22AG DNA and induces a topological change towards antiparallel quadruplex formation. In the presence of K+, 22AG DNA exhibited two positive peaks at


Fig. 4 CD spectra for 22AG DNA (10 mM) with or without CPT2. (a) Solution buffered with 10 mM Tris (pH 7.4); (b) solution buffered with 10 mM Tris (pH 7.4) and 50 mM KCl. [CPT2]/mM: (1) 0; (2) 10; (3) 20; (4) 50; and (5) 100.

Analyst, 2014, 139, 1834–1838 | 1835

View Article Online


Analyst

265 and 290 nm without adding CPT2, conrming the mixed type of G-quadruplex structures. It was reported that K+ induces a mixed population of both parallel and antiparallel structures.30 Aer adding CPT2, the CD spectra changed from mixed to antiparallel types (Fig. 4b), indicating specic selectivity towards the antiparallel structure, and the other topological structures of G-quadruplex could adapt an antiparallel structure. Meanwhile, in the presence of Na+, the 22AG DNA maintained its original antiparallel structure aer adding CPT2 (Fig. S5†), which further proves the antiparallel selectivity of CPT2. Moreover, to determine the binding stoichiometry of the Ru-complex with 22AG DNA, continuous variation analysis was carried out (Job plot, Fig. S12†). The intersection points obtained in the Job plot demonstrated 1 : 1 binding stoichiometries, indicating one CPT2 per G-quadruplex. The CD melting studies also suggested that CPT2 molecules could stabilize the 22AG DNA G-quadruplex by a Tm of 8 C (Fig. S6†). Note that though the continuous variation analysis implies 1 : 1 binding stoichiometries, the higher thermal stabilization may be attributed to the higher equivalents of CPT2 used in the study. To further prove the potential of CPT2 as a G-quadruplex specic probe, the interactions between CPT2 and other types of DNA conrmations were investigated. From the uorescence titration experiments (Fig. 5a), the uorescence intensity of

Fig. 5 (a) Fluorescence titration of CPT2 with various G-quadruplexes and CTDNA/ssDNA/dsDNA. Solution contains 10 mM CPT2, 10 mM Tris (pH 7.4) and 50 mM KCl. lex is 521 nm and the fluorescence intensity is recorded at 620 nm. (b) Fluorescence distinction of CPT2 with different DNAs under irradiation of UV light, the solution is buffered with 10 mM CPT2, 10 mM Tris (pH 7.4) and 50 mM KCl. From left to right, DNAs are as follows: ssDNA, CTDNA, H-RAS, 22AG, G3T3, c-kit* and c-myc. Concentrations of different DNAs are all 10 mM.

1836 | Analyst, 2014, 139, 1834–1838

Communication

CPT2 to a complementary oligonucleotide of 22AG DNA (ssDNA) and calf thymus DNA (CTDNA) was found to be very low compared to other G-quadruplex DNAs. The uorescence enhancement of ssDNA and CTDNA was only 10- and 17-fold, respectively, at 10 mM DNA concentration, which is less than one-tenth compared to G-quadruplex-forming 22AG DNA. Similarly, the uorescence enhancement of dsDNA was 45-fold in the presence of K+, which is one-fourth compared to 22AG DNA. Moreover, CPT2 also showed different uorescence responses towards other DNA G-quadruplex sequences and RNA G-quadruplexes (NRAS and TERRA). The uorescence enhancement of 22AG, G3T3, H-RAS, and c-kit DNAs was higher than other oligonucleotides. In the presence of 50 mM KCl, different oligonucleotides showed three main conformations of G-quadruplexes. Therefore, the topologies of G-quadruplexes with CPT2 were studied using the CD studies that revealed great selectivity of CPT2 towards the topology of the G-quadruplex. The oligonucleotides G3T3, H-RAS, and c-kit were added to the CPT2 solution, similar to the 22AG DNA, and the structure of these mixed-type G-quadruplexes eventually changed to the antiparallel form and the uorescence enhancement was found to be signicant (Fig. 5 and S7†). However, the oligonucleotides src1, ckit, c-myc, NRAS and TERRA, which formed parallel structures in the presence of K+, cannot be changed to other topological forms even at a CPT2 concentration of 100 mM. The uorescence enhancements were found to be less than 50-fold. Notably, although TBA could form an antiparallel structure, its uorescence enhancement was only found to be 22-fold. This may account for the relatively shorter sequence of TBA, the loop of which may not be large enough to bind the CPT2 molecule. In addition, a 8TG sequence could form a four strand intermolecular quadruplex and the inter 1 + 3 sequence could form a (1 + 3) intermolecular quadruplex.31,32 The uorescence enhancement of 8TG and inter 1 + 3 was only 19- and 18-fold, respectively, indicating that CPT2 had low affinity toward intermolecular quadruplexes. These ndings indicate that CPT2 has the potential to be used as a uorescent switch-on probe for antiparallel G-quadruplexes. Moreover, in the absence of K+, none of the sequences showed a typical G-quadruplex structure at rst (Fig. S8†); however, aer the interaction with CPT2, 22AG, G3T3, H-RAS, and c-kit, all the sequences showed antiparallel structures. There is a clear distinction between antiparallel G-quadruplex DNA and other DNAs, which could be visualized by the naked eye (Fig. 5b) under UV-light irradiation. These observations were found to be consistent with the results of the titration experiments. The MD simulation shows that CPT2 binds strongly to G-quadruplex DNA than CPT1 and is more stable in the antiparallel conrmation (Fig. S9 and S10†). According to the physical meaning of the score in docking, the number presents the negative logarithmic of the dissociation equilibrium constant for the ligand–DNA system. Therefore, the binding strength of CPT2 to 143D is stronger than that of CPT2 to 1KF1, indicating that CPT2 binds selectively to the antiparallel G-quadruplex structure. As shown in Fig. S9,† CPT2 binds in the groove of the anti-parallel G-quadruplex structure (143D) by way of bending its molecule and tting into the shape of the groove.


View Article Online


Communication

Hydrogen binding and van der Waals forces are the dominant driving forces of this binding process, and hydrogen binding also helps to x the position and conguration of the ligand. The p–p stacking force plays a major role when CPT2 binds to the parallel G-quadruplex structure (1KF1). However, this p–p stacking force is crippled by the imperfect overlap of CPT2 to G-quadruplex base-pairs. Considering the differences between CPT1 and CPT2, the CPT2 molecule is bigger than CPT1 and possesses more hydrogen-binding acceptor sites. The hydrogen-binding was observed in CPT2's binding model, whereas no hydrogenbinding was observed in CPT1's binding model. The molecular modelling calculation was found to be consistent with the above mentioned results, and supported that CPT2 could be the probe for the antiparallel G-quadruplex.

Conclusions In conclusion, we synthesized a novel symmetric cyanovinylpyridinium triphenylamine derivative that exhibited a 190-fold increase in uorescence upon binding to the G-quadruplex of 22AG DNA. The CD studies showed that the CPT2 molecule could induce the 22AG DNA sequence to adapt an antiparallel structure. The molecular modelling experiments also indicated that CPT2 could adapt an antiparallel structure, which is more stable than antiparallel G-quadruplex DNA. Furthermore, CPT2 showed a high selectivity towards antiparallel G-quadruplexes, which was observed by the naked eye under UV light. These characteristics of the CPT2 molecule make it a promising uorescent probe for sensing antiparallel G-quadruplexes, and it may provide useful information for future studies.

Acknowledgements This work was supported by the National Basic Research Program of China (973 Program) (2012CB720600, 2012CB720603), the National Science Foundation of China (no. 91213302, 21072115, 21272181), the National Grand Program on Key Infectious Disease (2012ZX10003002-014), the Program for Changjiang Scholars and Innovative Research Team in University (IRT1030), and the Fundamental Research Funds for the Central Universities.

References 1 J. L. Huppert, Chem. Soc. Rev., 2008, 37, 1375–1384. 2 S. Burge, G. N. Parkinson, P. Hazel, A. K. Todd and S. Neidle, Nucleic Acids Res., 2006, 34, 5402–5415. 3 M. L. Bochman, K. Paeschke and V. A. Zakian, Nat. Rev. Genet., 2012, 13, 770–780. 4 X. Cang, J. Sponer and T. E. Cheatham 3rd, Nucleic Acids Res., 2011, 39, 4499–4512. 5 Q. Wang, L. Ma, Y.-H. Hao and Z. Tan, Anal. Chem., 2010, 82, 9469–9475. 6 J.-Q. Liu, C.-Y. Chen, Y. Xue, Y.-H. Hao and Z. Tan, J. Am. Chem. Soc., 2010, 132, 10521–10527.


Analyst

7 K.-W. Zheng, Z. Chen, Y.-H. Hao and Z. Tan, Nucleic Acids Res., 2010, 38, 327–338. 8 Z. Chen, K.-W. Zheng, Y.-H. Hao and Z. Tan, J. Am. Chem. Soc., 2009, 131, 10430–10438. 9 E. Y. Lam, D. Beraldi, D. Tannahill and S. Balasubramanian, Nat. Commun., 2013, 4, 1796. 10 G. Biffi, D. Tannahill, J. McCafferty and S. Balasubramanian, Nat. Chem., 2013, 5, 182–186. 11 C. Zhang, H. H. Liu, K. W. Zheng, Y. H. Hao and Z. Tan, Nucleic Acids Res., 2013, 41, 7144–7152. 12 L. Yuan, T. Tian, Y. Chen, S. Yan, X. Xing, Z. Zhang, Q. Zhai, L. Xu, S. Wang, X. Weng, B. Yuan, Y. Feng and X. Zhou, Sci. Rep., 2013, 3, 1811. 13 V. Gabelica, R. Maeda, T. Fujimoto, H. Yaku, T. Murashima, N. Sugimoto and D. Miyoshi, Biochemistry, 2013, 52, 5620– 5628. 14 M. Nikan, M. D. Antonio, K. Abecassis, K. McLuckie and S. Balasubramanian, Angew. Chem., Int. Ed., 2013, 52, 1428–1431. 15 J. Alzeer, B. R. Vummidi, P. J. C. Roth and N. W. Luedtke, Angew. Chem., Int. Ed., 2009, 48, 9362–9365. 16 J. Mohanty, N. Barooah, V. Dhamodharan, S. Harikrishna, P. I. Pradeepkumar and A. C. Bhasikuttan, J. Am. Chem. Soc., 2013, 135, 367–376. 17 B. R. Vummidi, J. Alzeer and N. W. Luedtke, ChemBioChem, 2013, 14, 540–558. 18 B. Dumat, G. Bordeau, A. I. Aranda, F. Mahuteau-Betzer, Y. El Harfouch, G. Metge, F. Charra, C. Fiorini-Debuisschert and M. P. Teulade-Fichou, Org. Biomol. Chem., 2012, 10, 6054– 6061. 19 J. Wang, C. He, P. Y. Wu, J. Wang and C. Y. Duan, J. Am. Chem. Soc., 2011, 133, 12402–12405. 20 Z. Yang, N. Zhao, Y. Sun, F. Miao, Y. Liu, X. Liu, Y. Zhang, W. Ai, G. Song, X. Shen, X. Yu, J. Sun and W. Y. Wong, Chem. Commun., 2012, 48, 3442–3444. 21 A. I. Aranda, S. Achelle, F. Hammerer, F. Mahuteau-Betzer and M.-P. Teulade-Fichou, Dyes Pigm., 2012, 95, 400– 407. 22 D. Blaise, B. Guillaume, F. P. Elodi, M. B. Florence, S. Nicolas, M. Germain, F. D. Celine, C. Fabrice and T.-F. Marie-Paule, J. Am. Chem. Soc., 2013, 135(34), 12697– 12706. 23 J. Bunkenborg, N. I. Gadjev, T. Deligeorgiev and J. P. Jacobsen, Bioconjugate Chem., 2000, 11, 861–867. 24 C. Allain, F. Schmidt, R. Lartia, G. Bordeau, C. FioriniDebuisschert, F. Charra, P. Tauc and M. P. Teulade-Fichou, ChemBioChem, 2007, 8, 424–433. 25 Y. Hong, J. W. Y. Lamab and B. Z. Tang, Chem. Soc. Rev., 2011, 40, 5361–5388. 26 E. P´ erez-Guti´ errez, M. J. Percino, V. M. Chapela, M. Cer´ on, J. L. Maldonado and G. Ramos-Ortiz, Materials, 2011, 4, 562–574. 27 M. J. Frisch, G. W. Trucks, H. B. Schlegelet al., Gaussian 09, Revision A.02, Gaussian, Inc., Wallingford CT, 2009. 28 Z. R. Grabowski, K. Rotkiewicz and W. Rettig, Chem. Rev., 2003, 103, 3899–4032.

Analyst, 2014, 139, 1834–1838 | 1837

View Article Online

Analyst

31 P. B. Arimondo, J. F. Riou, J. L. Mergny, J. Tazi, J. S. Sun, T. Garestier and C. Helene, Nucleic Acids Res., 2000, 28, 4832–4838. 32 K. T. Kim and B. H. Kim, Chem. Commun., 2013, 49, 1717– 1719.


29 P. L. Tran, J. L. Mergny and P. Alberti, Nucleic Acids Res., 2011, 39, 3282–3294. 30 K. W. Lim, S. Amrane, S. Bouaziz, W. Xu, Y. Mu, D. J. Patel, K. N. Luu and A. T. Phan, J. Am. Chem. Soc., 2009, 131, 4301– 4309.

Communication

1838 | Analyst, 2014, 139, 1834–1838


View Article Online

Analyst


PAPER


Surface plasmon resonance sensing properties of a 3D nanostructure consisting of aligned nanohole and nanocone arrays Mohamadreza Najiminaini,abc Erden Ertorer,de Bozena Kaminska,b Silvia Mittlerd and Jeffrey J. L. Carson*ac Molecular surface plasmon resonance (SPR) sensing is one of the most common applications of an array of periodic nanoholes in a metal film. However, metallic nanohole arrays (NHAs) with low-hole count have lower resolution and SPR sensing performance compared to NHAs with high-hole count. In this paper, we present a compact three-dimensional (3D) plasmonic nanostructure with extraordinary optical transmission properties benefiting from surface plasmon matching and enhanced localized surface plasmon coupling. The 3D nanostructure consisted of a gold film containing a NHA with an underlying cavity and a gold nanocone array (NCA) at the bottom of the cavity. Each nanocone was aligned with the nanohole above and the truncated apex of each nanocone was in close proximity (100 nm) to the gold film. The NHA–NCA structures outperformed conventional NHA structures in terms of bulk sensitivity and Figure of Merit


(FOM). Furthermore, the NHA–NCA structure with 525 nm periodicity was capable of sensing streptavidin

DOI: 10.1039/c3an02332j

down to 2 nM exhibiting a 10-fold increase in streptavidin sensitivity compared to conventional NHA structures. The sensitivity and performance of the 3D nanostructure can be further improved by exploiting

www.rsc.org/analyst

multiplexing methods in combination with stable light sources and detection systems.

Introduction Surface Plasmon Resonance (SPR) sensing has been recognized as one of the most important applications of metallic nanostructures and has been exploited for the detection of molecules.1,2 The surface plasmon (SP), or physically exact surface plasmon polariton (SPP), is a collective oscillation of the free electron gas at the interface between a metal and a dielectric; it is a longitudinal wave propagating along the metal-dielectric interface producing evanescent elds in the metal and in the dielectric.3 The SPR sensing principle is based on a change in the coupling conditions between an incident light beam and the SP due to a change in the surrounding dielectric material, e.g. resulting from material adhesion onto the metal, changing the refractive index architecture.4 Localized surface plasmon resonance (LSPR), on the other hand, is the collective oscillation of the free electron gas in a metal nanoparticle leading to an

a

Imaging Program, Lawson Health Research Institute, St. Joseph's Health Care, London, ON, Canada. E-mail: [email protected]

b

The School of Engineering Science, Simon Fraser University, Burnaby, BC, Canada

c

The Department of Medical Biophysics, Schulich School of Medicine and Dentistry, University of Western Ontario, London, ON, Canada d

The Department of Physics and Astronomy, University of Western Ontario, London, ON, Canada

e

Biomedical Engineering Program, Faculty of Engineering, University of Western Ontario, London, ON, Canada

1876 | Analyst, 2014, 139, 1876–1882

absorption band whose location also depends on the refractive index architecture around the nanoparticle.5 A periodic array of nanoholes in a metal lm allows wave vector matching between an incident light on a nanohole array (NHA) and the SP waves existing at the interface between a metal and a dielectric. This effect results in an interaction of the light with the SP waves.6 The excitations of SP waves by an incident light on NHAs generate transmission based SPRs or so called extraordinary optical transmissions (EOTs). The EOTs of a NHA depend greatly on the dielectric functions of the metal and the dielectric materials and geometrical parameters of the NHA (hole spacing, hole arrangement and hole size).7 However, an individual nanohole can produce transmission based LSPR, which is associated to an interaction of the incident light with an LSP around the nanohole. Both an individual nanohole and a NHA have enhanced electric eld intensity around each nanohole at the corresponding resonance wavelengths. Both LSPR of a single nanohole and SPRs of a NHA have been extensively exploited in LSPR or SPR-based detection of chemicals and molecules.2,6 However, in the context of the NHA literature, SPR sensing performance of a NHA has been evaluated by bulk sensitivity (nm per refractive index unit (RIU)) and gure of merit (FOM (1/RIU)). The bulk sensitivity is well known as SPR peak shi (nm) of a NHA per RIU and the FOM is recognized as the bulk sensitivity over the peak width (FWHM) of the SPR. Nevertheless, the FOM terminology for nanoparticles and a signal nanohole is dened differently and it is


View Article Online


Paper

calculated based on LSPR peak shis per RIU. In this context, we use performance metrics associated with the NHA and name all resonances SPR despite some of them being local. Kretschmann-based SPR sensing remains the most sensitive method and has motivated several groups to study methods to improve the bulk sensitivity and FOM of NHAs.8–14 Based on recent studies, SPR performance of NHAs was improved by narrowing the resonance peak width, incorporating a smooth metal surface, index matching the substrate and the sensing material, and using design strategies to increase the electric eld near the nanoholes. For example, it was reported for a bulk SPR sensing experiment (where the environmental refractive index is changed) that the bulk sensitivity of a NHA in a gold lm reached Kretschmann-like performance when the substrate was index-matched to the analyte, the gold NHA was fabricated with an ultra-smooth surface, and elliptical hole shapes were used.9 Further improvements to the bulk sensitivity and FOM of metallic NHAs were achieved by fabricating the NHA with dimensions larger than the excited SP propagation length, which constructively enhanced the excited SP interactions thereby improving the transmission efficiency of the resonance and narrowing the resonance peak width.15 Various transmission resonances of a NHA are associated with different SP excitation modes based on the relationship between the scattering order of the holes, similar to Bragg diffraction in a crystal.6 The transmission resonances are therefore classied with a number combination (x, y) regarding the plane of the NHA. The bulk sensitivity of various transmission resonance modes of a NHA are different and the maximum bulk sensitivity has been achieved for the (1, 0) transmission resonance of a NHA.13 For example, the (1, 0) resonances of large NHAs were reported to have bulk sensitivity as high as 600 nm per refractive index unit (nm/RIU),8,16,17 while the bulk sensitivity and FOM of a NHA with low hole count (i.e. less than 100) for the (1, 0) resonance was 286 nm/RIU and 4.1 (1/RIU), respectively.16 Some attempts to improve bulk sensitivity of a low-hole-count NHA were tested by incorporating Bragg reectors into the structure.12,18 Our group recently observed that a low-hole-count NHA with a large integrated cavity provided a 1.6-fold improvement in bulk sensitivity compared to a NHA structure without a cavity, due to dynamic SP matching between the top and the bottom of the NHA.14 The matching between SP propagating along the top and the bottom of the NHA were very similar due to the symmetric gold structure in the case with the cavity and therefore enhanced the transmission resonance by multiple fold.19 Threedimensional (3D) metallic nanostructures consisting of co-registered nanoholes and nanodisks have recently been reported and shown to exhibit transmission resonances due to coupled LSP interaction between nanoholes and nanodisks through enhanced electric elds.20 In bulk SPR sensing, the 3D metallic nanostructures had a higher bulk sensitivity than low-hole-count NHAs with cavities, but the resonance peak width was broader.20 Building upon these prior results, the objective here was to design a new 3D nanostructure with improved bulk sensitivity and FOM, where transmission resonances resulted from the coupling and interaction of both LSP and SPP. The approach was to devise a structure with enhanced electric eld near each nanohole by This journal is © The Royal Society of Chemistry 2014

Analyst

incorporating dynamic-SP matching and one at metal surface. We fabricated a 3D metallic nanostructure consisting of a NHA in a gold lm with an underlying cavity carrying a gold nanocone array (NCA) at its bottom (Fig. 1). The gold NCA was responsible for generating strong LSP interactions between each nanohole and nanocone as well as providing for coupling between the SPP and the LSP of the NHA and NCA, respectively. We performed both simulations and experimental studies to determine the transmission resonance properties of the NHA– NCA structures. Various NHA–NCA structures with different geometries were fabricated and tested as bulk SPR sensors. The NHA–NCA structures were functionalized with a binary mixture of OH-terminated and biotinylated thiols and evaluated as molecular SPR sensors for streptavidin. We employed an electron beam lithography and lioff fabrication process to create the NHA in an 80-nm-thick gold lm on a Pyrex substrate. Extensive description of the NHA fabrication methodology has been reported previously.21 Prior to 80 nm gold deposition and the lioff process, a 6-nm-thick Ti adhesion layer was evaporated on to the substrate using electron beam deposition. Aer fabrication, the NHA sample was immersed in Ti etchant (TFTN, Transene company, Inc.) for 70 seconds to remove the Ti adhesion layer and create a 250-nm-deep cavity within the Pyrex substrate beneath the NHA in the metal lm. A 150-nm-thick gold lm was deposited by electron beam evaporation on top of the freestanding gold NHA and resulted in a thickening of the gold lm to 230 nm and formation of a NCA at the bottom of the cavity. A schematic of the fabrication processes before and aer gold deposition is shown in Fig. 1(a). We fabricated various 3D nanostructures with different hole periodicities and diameters. All devices had a 250-nm-deep cavity and 150-nmtall nanocones. The nanoholes and nanocones had periodicities between 375 nm and 550 nm in increments of 25 nm. The ratio between overall nanohole area to lled gold area (i.e. square of hole periodicity) was 0.096 for all 3D nanostructures. The lowest and highest hole diameters for the 3D nanostructures were 131 nm and 192 nm, respectively. The periodicities were selected such that the main transmission resonances were located between 400 nm and 850 nm to match the spectral range of the characterization setup. The area of each array was about 25 mm2 and each array was repeated in a 2D grid of 7 by 7 with 10 mm spacing between adjacent arrays to minimize surface plasmon crosstalk.22 Fig. 1(b) displays an SEM image of a NHA–NCA structure with 500-nm periodicity, a 230-nm-thick gold lm, 150nm-tall gold nanocones, and a 250-nm-deep cavity. The SEM images revealed that each nanocone was right circular in geometry and had a truncated apex. The base diameter of each nanocone was similar to the nanohole diameter measured prior to deposition of the 150-nm-thick layer of gold. The diameter of the truncated apex of each nanocone was approximately half of the diameter of the base. The diameter of the truncated nanocone apex was dependent on the gold deposition rate (data not shown). The diameter of each nanohole at the top and the bottom surfaces of the gold lm were similar in diameter to the base of each gold nanocone. In addition, the nanohole diameter decreased from both the top and the bottom surfaces to the middle of the gold lm. Analyst, 2014, 139, 1876–1882 | 1877

View Article Online


Analyst

Paper

(a) Schematic of a NHA in a gold film with a large cavity etched into the Pyrex substrate (top panel). A schematic of a NHA in a gold film with an underlying gold NCA, where each nanocone was axially aligned with the hole above and was formed by deposition of gold through the NHA from the top surface (bottom panel). (b) SEM image of a NHA in a 230-nm-thick gold film and a 150-nm-tall gold NCA with 500-nm periodicity. The cavity depth was 250 nm. The largest and smallest diameters of the holes and cones were 175 nm and 90 nm, respectively.

Fig. 1

We employed Finite Difference Time Domain (FDTD) simulations (Lumerical Inc., Vancouver, Canada) to compute the optical transmission properties of the NHA–NCA structures. In the simulation model, we used an incident plain wave propagating in the direction normal to the sandwich structure. Periodic and perfectly matched layer boundary conditions were used in xy directions and at the z boundaries, respectively. The minimum mesh size was three nanometers. Fig. 2(a) displays optical transmission spectra of NHA–NCA structures with 500nm periodicity for a range of nanohole and nanocone diameters. The gold NCA had two main LSP absorption resonances, at 678 nm and 847 nm, related to the truncated apex and the base,

respectively as seen from the electric eld distribution for each absorption peak (Fig. 2(b)). The highest electric eld intensity (hot spot) was found at the perimeter of the truncated apex of the nanocone, which had a dipolar mode due to the excitation of the LSP. The dipolar mode stems from strong electric eld distributions (hot spots) on the two opposite sides of the nanocone perimeter (data not shown). Based on previous work on similar structures,23,24 it has been shown that a strong electric eld is conned to the tip of the nanocone. As a result, the absorption peak at 678 nm from the truncated surface of the gold nanocone was due to the vertical LSP excitations of the nanocone, which eventually resulted in the LSP absorption at

Fig. 2 (a) Simulated optical transmission spectra of a 150-nm-thick gold NCA with periodicity of 500 nm. Each nanocone had an apex diameter of 87.5 nm and a base diameter of 175 nm. The substrate had a refractive index of n ¼ 1.474 (Pyrex, at l ¼ 587 nm) and the surrounding material had a refractive index of n ¼ 1.330 (water, at l ¼ 587 nm) (blue curve). Simulated optical transmission spectrum of a 230-nm-thick gold NHA membrane and a 150-nm-thick gold NCA at the bottom of the 250-nm-deep cavity (green curve). The nanohole had a diameter of 90 nm (midpoint in film) and a diameter of 175 nm (top and bottom of film). (b) Electric field distribution of a nanocone at absorption wavelengths of 678 nm (top left) and 847 nm (bottom left). Electric field distribution of a combined nanohole and nanocone at transmission resonances located at 621 nm (top right panel) and 739 nm (bottom right panel).

1878 | Analyst, 2014, 139, 1876–1882


View Article Online


Paper

the apex. Moreover, the simulation intensity maps for wavelengths between 550 nm and 780 nm revealed the presence of a conned electric eld at the apex of the nanocone. Nevertheless, amplitudes of energy connement at the apex of the nanocone were not the same for different wavelengths. The variations in the localized electric eld intensity at the apex of the nanocone were likely due to size-dependence of vertical LSP excitations of the nanocone as well as the wavelength-dependent absorption properties of gold. The simulated optical transmission of the NHA–NCA structure demonstrated the existence of two transmission resonance peaks that were related to the (1, 0) and (1, 1) SP excitation modes. The electric eld maps at both resonances showed pronounced LSP coupling between the apex of the nanocone and the bottom aperture of the nanohole as shown in Fig. 2(b) (right panel). The LSP coupling between the nanocone and nanohole was stronger for the (1, 0) resonance than the (1, 1) resonance, which was due to the deeper SP decay length from the metal lm into the dielectric at longer wavelengths.3 In addition, LSP absorption was apparent above 800 nm due to the base of the nanocone. Furthermore, the NHA–NCA was responsible for SP-light coupling and an evanescent transfer of SP-light energy to the other side of the metal lm. As a result, each of the (1, 0) and the (1, 1) resonances had LSP and SPP characteristics. Also, due to inherent SPP properties of the resonances of the 3D nanostructure, both the (1, 0) and the (1, 1) spectral resonance positions depended greatly on the spacing between the nanoholes. However, we observed that the (1, 0) resonance was further divided into two resonances when the periodicity was larger than 500 nm. The reason behind this phenomenon was the LSP interaction between nanohole and nanocone, which became dominant and generated two individual resonance peaks. To measure the optical transmission of the 3D gold nanostructures, we employed an inverted microscope (TE300, Nikon, Japan) equipped with a photometer, monochromator, and photomultiplier-based detector (PTI, Birmingham, NJ). The incident illumination (Halogen lamp, 12 V, 100 W, Ushio Inc., Tokyo, Japan) was focused on to each 3D nanostructure array using the bright eld condenser. The illumination angle was controlled to within 6 of normal incidence. The transmitted light was collected with a 20 objective (NA ¼ 0.45; model# 93150, Nikon). A region of interest was selected with the aperture controls on the photometer. For each spectral scan, the region of interest was chosen to allow the transmitted light from one set of 25 identical 3D nanostructures through to the monochromator and detector. This methodological step resulted in average transmission spectra representative of many NHA–NCAs and served to enhance the signal-to-noise ratio. Transmission spectra of 3D nanostructures with 475-nm periodicity and 250-nm cavity-depth were measured when each liquid refractive index standard between 1.30 and 1.39 in increments of 0.03 was applied to the top surface (Fig. 3(a)). Because of capillary forces, the liquid lled the cavity. Similar to simulation results, the (1, 0) and the (1, 1) transmission resonances were observed in the measured spectra. However, an additional weak transmission resonance was observed


Analyst

between the (1, 1) and (1, 0) resonances and was likely related to the (0, 1) SP excitation mode. The weak resonance was apparent in the measured spectra due to the wide range of incident illumination angles (6 ) compared to normally incident illumination used in the simulations. In a similar manner to the simulation results, the 3D nanostructure arrays with 525-nm and 550-nm periodicities revealed a double resonance near the expected position of the (1, 0) resonance. The double resonance may have been related to a stronger LSP interaction between nanocone and nanohole at longer wavelengths, where the second resonance at longer wavelength was associated with the LSP interaction between each nanohole and nanocone. Using the transmission spectra from each set of 3D nanostructure arrays, estimates of resonance peak width and bulk sensitivity were determined. The bulk sensitivity was estimated from the slope of the line of best t of the resonance position versus refractive index. For the 3D nanostructures with 475 nm periodicity, the measured (1, 0) and (1, 1) resonance peak widths (FWHM) were 40 nm and 56 nm, respectively. Also, the bulk sensitivities for the (1, 0), the (1, 1) and the (0, 1) resonances were 440 nm/RIU, 280 nm/RIU and 460 nm/RIU, respectively. The FOM for the (1, 0) and the (1, 1) resonances were 11 1/RIU and 5 1/RIU, respectively. The bulk sensitivities related to the (1, 0) resonance of the 3D nanostructures are shown in Fig. 3(b) as a function of periodicity. As expected from earlier experimental and theoretical analysis (see ref. 13 and 14), the bulk sensitivity of the 3D nanostructures increased with increasing periodicity. The maximum bulk sensitivity (593 nm/ RIU) and FOM (12.90 1/RIU) was found for the 3D nanostructures with 550-nm periodicity. Therefore, for bulk SPR sensing, the 3D nanostructure had a larger than 2-fold higher bulk sensitivity and in excess of 3-fold greater FOM compared to reported results (see ref. 12 and 16) for conventional NHAs with low hole count. Functionalized 3D nanostructures were evaluated in SPR sensing of streptavidin in PBS buffer. The gold surfaces of the 3D nanostructures were functionalized with a binary mixture of an OH-terminated and biotinylated thiol to offer biotin as recognition sites for streptavidin in the right amount and to allow streptavidin binding without steric hindrance (see schematic diagram in Fig. 4(a)). Functionalization was achieved by immersing each device into an anhydrous ethanol solution composed of 0.45-mM 11-mercaptoundecanol (Sigma-Aldrich, Ontario, Canada) and 0.05-mM biotinylated thiol (NanoScience Chemicals, Phoenix, AZ, USA) for 24 hours to create a selfassembled monolayer (SAM) with biotin moieties. The device was then rinsed with ethanol and dried with nitrogen. Finally, the sample was placed in PBS buffer, streptavidin (Rockland Inc., PA, USA) was added. Measurements were performed within 5 minutes. The (1, 0) resonance peak shis of each 3D nanostructure array in response to 2-nM, 20-nM, 200-nM, 2000-nM, and 3500nM streptavidin were measured and are shown in Fig. 4(b). In order to determine a precise (1, 0) resonance position from each spectrum, we employed Gaussian curve tting on the (1, 0) resonance peak. All measurements were repeated in

Analyst, 2014, 139, 1876–1882 | 1879

View Article Online


Analyst

Paper

(a) Optical transmission spectra of a NHA–NCA structure taken when one of four refractive index standards (1.30, 1.33, 1.36, and 1.39 at l ¼ 589 nm) was placed on top of the surface. The NHA and NCA periodicity was 475 nm. The cavity beneath the NHA was 250 nm deep. The apex and base of each nanocone had diameters of 84 nm and 170 nm, respectively. The gold film was 230 nm thick. The height of each nanocone was 150 nm. (b) Measured bulk sensitivity of the NHA–NCA structure as a function of NHA and NCA periodicity. Bulk sensitivity was estimated by the slope of the wavelength of the (1, 0) resonance peak versus the refractive index of the medium above the top surface as shown in panel (a).

Fig. 3

triplicate and the results are reported in Fig. 4(b) as mean 1 standard deviation (SD). Furthermore, the detection condence was calculated for each 3D nanostructure based on the mean value from triplicate PBS measurements (blue solid line in Fig. 4(b)) 3 times the average noise (blue dashed lines in Fig. 4(b)). The limit of detection (LOD) of a 3D nanostructure was the lowest detectable concentration of streptavidin where there was enough certainty (i.e. above detection condence). Considering all devices, the (1, 0) resonance peak red-shied as the concentration of streptavidin increased. The largest spectral resonance shi was observed when the concentration was increased from 200 nM to 2000 nM. However, the size of the resonance shi in response to streptavidin was not consistent across all devices. For example, the (1, 0) resonance shi between PBS base line and 3500-nM streptavidin was largest for the 3D nanostructure with 375-nm periodicity. This effect may have been due to the smaller nanohole and nanocone dimensions for the 375-nm periodicity structure, which resulted in more signicant changes in LSP coupling between nanocones and nanoholes due to the smaller size of nanohole and nanocone. However, 3D nanostructures with larger periodicities (i.e. 500 nm and 525 nm) had larger (1, 0) resonance shis for the lowest concentrations of streptavidin. Moreover, the lowest concentration of streptavidin (2 nM) was only detectable with the 3D nanostructure with 525-nm periodicity, where the detection certainty of the LOD was calculated at 95% condence interval.25 This nding substantiates the observation of higher bulk-SPR sensing for devices with larger periodicities compared to devices with smaller periodicities (e.g. see ref. 13). Previously, large conventional NHAs with 650-nm periodicity and 400-mm2 area were used to detect streptavidin binding by monitoring the spectral SPR shi.17 The average resonance peak shi for high concentrations of streptavidin (20 mM) was found to be between 3.5 nm and 4 nm. In another study, a multiplexed

1880 | Analyst, 2014, 139, 1876–1882

arrangement of NHAs, each with an area of approximately 40 mm2 and providing a variety of resonance positions, were shown to improve the limit of streptavidin detection down to 25 nM with the use of a stable laser source and cooled CCD camera.10 In comparison with these previous studies, our 3D nanostructure was capable of sensing a 10-fold lower streptavidin concentration (2 nM) without the use of a laser or a cooled CCD camera. Therefore, it is expected that the detection limit of streptavidin with our 3D nanostructure could be reduced below 2 nM by implementing a stable light source and a cooled detector. Based on the bulk sensitivity measures, the NHA–NCA structures offer improved performance compared to conventional NHAs. Improved performance is likely due to a combination of factors. First, the NHA–NCA structure benets from dynamic surface plasmon matching between the top and bottom surfaces of the gold NHA membrane due to the presence of the cavity.14 Second, the bottom surface of the NHA membrane has an ultra-smooth metal surface, which resulted from the polished surface of the Pyrex substrate (with rough˚ prior to etching. However, the top ness parameter, Ra < 15 A) surface roughness of the NHA was about 1.9 nm as reported previously in ref. 21. Third, the highly localized electric eld between the nanohole and the nanocone due to the LSP interactions enable high sensitivity of this region of the device to changes in refractive index of the intermediate dielectric materials. Fourth, there were no Ti adhesion layers in the NHA– NCA structure to reduce optical performance, which resulted in higher bulk sensitivity of the NHA–NCA structure as reported for NHAs in previous studies.21,26 One disadvantage of the NHA– NCA structure is the lower transmission efficiency at the resonances compared to the resonance transmission efficiency of a conventional NHA. According to experimental results, the resonance transmission efficiency for the NHA–NCA was almost 42-fold lower than the NHA structure prior to forming a NCA.


View Article Online


Paper

Analyst

(a) Schematic of a nanohole and adjacent nanocone showing the functionalized surfaces. Streptavidin molecules bind to biotinylated thiol. (b) The SPR response of functionalized NHA–NCA structures to streptavidin. Each panel represents the SPR peak position for 25 identical nanostructure arrays with a periodicity indicated in the top left corner versus the streptavidin concentration in solution. The red error bars represent 1 standard deviation computed from three different measurements for each data point. The solid blue line represents the base line from PBS measurements (mean value among triplicate PBS measurements) and the dashed blue lines represents 3 times the standard deviation of the PBS measurements representative of the detection threshold. Fig. 4

Lower transmission was a consequence of the optical absorption properties of the nanocone, a thicker NHA gold lm, smaller nanohole size, and nanocone shadowing effects on each incident nanohole.20 In conclusion, a 3D nanostructure consisting of a metallic lm with a NHA, a cavity beneath the NHA, and a NCA at the bottom of the cavity was designed, fabricated, and tested through simulations and experiments. Experimental measurements of the NHA–NCA structure revealed transmission resonances due to LSP interactions between the nanocones and nanoholes and SPP interactions related to the NHA. A series of NHA–NCA structures were studied in bulk SPR sensing and found to have higher bulk sensitivity and FOM compared to conventional NHAs. Furthermore, NHA–NCA structures were employed for the detection of streptavidin. Compared to conventional NHA structures, one of the NHA–NCA structures was able to reliably detect streptavidin down to 2 nM, which exceeded the sensitivity of multiplexed detection methods


(25 nM).10 In future work, a multiplexed SPR sensor based on an array of NHA–NCA structures can be designed to improve the LOD for biomolecules over the integration method used to evaluate NHA–NCA structures described in this work.

Acknowledgements The authors thank Dr Todd Simpson for his technical support at the Western Nanofabrication Facility at the University of Western Ontario (UWO). We also would like to thank Dr Hao Jiang for helpful discussion about the biosensing results. This project was funded by grants from the Natural Sciences and Engineering Research Council of Canada (NSERC) to Dr Bozena Kaminska and Dr Jeffrey J. L. Carson. Dr Mohamadreza Najiminaini was supported by the MITACS program. Dr Silvia Mittler would like to thank the NSERC BiopSys Strategic Network for nancial contributions. Analyst, 2014, 139, 1876–1882 | 1881

View Article Online

Analyst

Paper


References 1 J. N. Anker, W. P. Hall, O. Lyandres, N. C. Shah, J. Zhao and R. P. Van Duyne, Biosensing with plasmonic nanosensors, Nat. Mater., 2008, 7, 442–453. 2 T. Rindzevicius, Y. Alaverdyan, B. Sepulveda, T. Pakizeh, ˜ de M. Kall, R. Hillenbrand, J. Aizpurua and F. J. GarcAa Abajo, Nanohole Plasmons in Optically Thin Gold Films, J. Phys. Chem. C, 2007, 111, 1207–1212. 3 H. Raether, Surface plasmons on smooth and rough surfaces and on gratings, Springer-Verlag, 1988. 4 C. Valsecchi and A. G. Brolo, Periodic Metallic Nanostructures as Plasmonic Chemical Sensors, Langmuir, 2013, 29, 5638–5649. 5 T. Jensen, L. Kelly, A. Lazarides and G. Schatz, Electrodynamics of Noble Metal Nanoparticles and Nanoparticle Clusters, J. Cluster Sci., 1999, 10, 295–317. 6 T. W. Ebbesen, H. J. Lezec, H. F. Ghaemi, T. Thio and P. A. Wolff, Extraordinary optical transmission through sub-wavelength hole arrays, Nature, 1998, 391, 667–669. 7 F. Przybilla, A. Degiron, J.-Y. Laluet, C. Genet and T. W. Ebbesen, Optical transmission in perforated noble and transition metal lms, J. Opt. A: Pure Appl. Opt., 2006, 8, 458. 8 A. A. Yanik, M. Huang, A. Artar, T. Chang and H. Altug, Integrated nanoplasmonic–nanouidic biosensors with targeted delivery of analytes, Appl. Phys. Lett., 2010, 96, 021101–021103. 9 G. A. Cervantes Tellez, S. Hassan, R. N. Tait, P. Berini and R. Gordon, Atomically at symmetric elliptical nanohole arrays in a gold lm for ultrasensitive refractive index sensing, Lab Chip, 2013, 13, 2541–2546. 10 H. Im, A. Lesuffleur, N. C. Lindquist and S. Oh, Plasmonic Nanoholes in a Multichannel Microarray Format for Parallel Kinetic Assays and Differential Sensing, Anal. Chem., 2009, 81, 2854–2859. 11 A. Lesuffleur, H. Im, N. C. Lindquist, K. S. Lim and S. Oh, Laser-illuminated nanohole arrays for multiplex plasmonic microarray sensing, Opt. Express, 2008, 16, 219–224. 12 N. C. Lindquist, A. Lesuffleur, H. Im and S. Oh, Sub-micron resolution surface plasmon resonance imaging enabled by nanohole arrays with surrounding Bragg mirrors for enhanced sensitivity and isolation, Lab Chip, 2009, 9, 382– 387. 13 L. Pang, G. M. Hwang, B. Slutsky and Y. Fainman, Spectral sensitivity of two-dimensional nanohole array surface plasmon polariton resonance sensor, Appl. Phys. Lett., 2007, 91, 123112–123113. 14 M. Najiminaini, F. Vase, B. Kaminska and J. J. L. Carson, Effect of surface plasmon energy matching on the sensing

1882 | Analyst, 2014, 139, 1876–1882

15

16

17

18

19

20

21

22

23

24

25

26

capability of metallic nano-hole arrays, Appl. Phys. Lett., 2012, 100, 063110–063114. F. Przybilla, A. Degiron, C. Genet, T. Ebbesen, F. de L´ eonP´ erez, J. Bravo-Abad, F. J. Garc´ıa-Vidal and L. Mart´ınMoreno, Efficiency and nite size effects in enhanced transmission through subwavelength apertures, Opt. Express, 2008, 16, 9571–9579. J. Henzie, M. H. Lee and T. W. Odom, Multiscale patterning of plasmonic metamaterials, Nat. Nanotechnol., 2007, 2, 549– 554. A. De Leebeeck, L. K. S. Kumar, V. de Lange, D. Sinton, R. Gordon and A. G. Brolo, On-Chip Surface-Based Detection with Nanohole Arrays, Anal. Chem., 2007, 79, 4094–4100. R. Gordon and P. Marthandam, Plasmonic Bragg reectors for enhanced extraordinary optical transmission through nano-hole arrays in a gold lm, Opt. Express, 2007, 15, 12995–13002. A. Krishnan, T. Thio, T. J. Kim, H. J. Lezec, T. W. Ebbesen, P. A. Wolff, J. Pendry, L. Martin-Moreno and F. J. GarciaVidal, Evanescently coupled resonance in surface plasmon enhanced transmission, Opt. Commun., 2001, 200, 1–7. M. Najiminaini, F. Vase, B. Kaminska and J. J. L. Carson, A three-dimensional plasmonic nanostructure with extraordinary optical transmission, Plasmonics, 2012, 8, 217–224. M. Najiminaini, F. Vase, B. Kaminska and J. J. L. Carson, Optical resonance transmission properties of nano-hole arrays in a gold lm: effect of adhesion layer, Opt. Express, 2011, 19(27), 26186–26197. F. Vase, M. Najiminaini, B. Kaminska and J. J. L. Carson, Effect of surface plasmon cross-talk on optical properties of closely packed nano-hole arrays, Opt. Express, 2011, 19, 25773–25779. A. V. Goncharenko, M. M. Dvoynenko, H. Chang and J. Wang, Electric eld enhancement by a nanometer-scaled conical metal tip in the context of scattering-type near-eld optical microscopy, Appl. Phys. Lett., 2006, 88, 104101. A. Mohammadi, F. Kaminski, V. Sandoghdar and M. Agio, Fluorescence Enhancement with the Optical (Bi-) Conical Antenna, J. Phys. Chem. C, 2010, 114, 7372–7377. D. Armbruster, M. Tillman and L. Hubbs, Limit of Detection (LOD)/Limit of Quantitation (LOQ): Comparison of the Empirical and the Statistical Methods Exemplied with GC-MS Assays of Abused Drugs, Clin. Chem., 1994, 40, 1233–1238. M. Najiminaini, F. Vase, B. Kaminska and J. J. L. Carson, Nano-hole array structure with improved surface plasmon energy matching characteristics, Appl. Phys. Lett., 2012, 100, 043105.


View Article Online

Analyst


COMMUNICATION


Received 2nd January 2014 Accepted 24th January 2014

Versatile chiral chromatography with mixed stationary phases of water-impregnated silica gel and reversed-phase packing† Satsuki Takahashi and Tetsuo Okada*

DOI: 10.1039/c4an00003j www.rsc.org/analyst

A novel chiral chromatographic scheme is proposed, which requires no organic syntheses in stationary phase preparation. A normal phase chiral chromatography stationary phase is prepared by simple passage of an aqueous solution of an appropriate chiral selector through e.g. a silica gel column. For this scheme, a mixed-bed of bare silica gel and octadecylsilanized silica gel (ODS) provides much better separation performance than a silica gel column. Isolation of silica gel particles in the column is important for successful chiral separation based on the present scheme.

Introduction Chiral liquid chromatographic separation is oen performed with a stationary phase, on which a selector (CS) is chemically bonded.1–7 Polysaccharides including cellulose,4,5 cyclodextrin and its derivatives,6 protein,7 etc. have been successfully employed as CSs for chiral chromatography. Since the applicability of CSs is usually limited to a particular class of enantiomers, the stationary phase should be carefully selected to attain required separation; i.e. a given stationary phase may be effective for chiral separation of particular compounds but may not work at all for separating other enantiomers.8,9 It should therefore be very useful if we can introduce a given CS into a common stationary phase in a simple way that requires no heavy organic syntheses. We reported chiral ice chromatography, in which an ice containing CS was used as a chiral liquid chromatographic stationary phase.10 An advantage of this method over usual chiral chromatography is that the stationary phase can be prepared by simple freezing of an aqueous solution of CSs and, thus, no organic syntheses are necessary for stationary phase preparation. Some enantiomers have been successfully separated with the Department of Chemistry, Tokyo Institute of Technology, Meguro-ku, Tokyo 152-8551, Japan. E-mail: [email protected]; Fax: +81-3-5734-2612; Tel: +81-3-57342612 † Electronic supplementary information (ESI) available: Temperature dependent chiral separation of hexobarbital and bisnaphthol, DSC of water in silica gel pores, and chiral chromatograms with various CSs. See DOI: 10.1039/c4an00003j

1830 | Analyst, 2014, 139, 1830–1833

b-cyclodextrin (bCD)-incorporated ice stationary phase. The ice stationary phase has, however, intrinsic analytical disadvantages, e.g. poor separation efficiency and limited durability, which hinder the wide applications of this method.11,12 Water molecules on the surface of ice have rather high mobility, which causes sintering of ice particles. The effective surface area of the ice stationary phase is rapidly diminished thereby, and separation also becomes poorer. Column storage suffers from the same problem. Even though the column is stored at low temperature, sintering proceeds and the column efficiency becomes lower during storage. Chiral recognition of bCD in the ice stationary phase occurs not on the surface of ice but in the liquid phase coexistent with ice.10 If the liquid phase can be stably retained in the stationary phase, a solid support does not have to be ice. Porous silica gel can, for example, be used for this purpose. However, it has been found that water-impregnation in silica gel pores results in unexpectedly poor separation, and, therefore, chiral separation has not been conrmed with such stationary phases. In the present communication, we propose a new idea to dramatically improve the separation efficiency of water-impregnated silica gel stationary phase and report chiral separation with the stationary phase prepared in this simple and versatile way.

Experimental section The chromatographic system was composed of a Shimadzu degasser unit Model DGU-20A, a Shimadzu HPLC pump Model LC30AD, a Rheodyne injection valve equipped with a 20 mL sample loop, a JASCO circular dichroism detector Model CD2095 Plus, and a Huber low-temperature bath ModelCC-410wl. The stationary phases were Daiso SP-1000-10 silica gel (10 mm in diameter, mean pore size 100 nm), Wakosil 10SIL (silica gel, 10 mm in diameter, mean pore size 6 nm), and Wakosil 10C18 (10 mm in diameter). The stationary phase was packed in a stainless HPLC column (4.6 mm i.d. 75 mm) by the slurry packing method.13 An aqueous solution containing CS (ca. 100 mL) passed through the column, and then the column was


View Article Online


Communication

Analyst

rinsed with hexane. This procedure allowed the replacement of the aqueous solution in the extra particle space, but the aqueous phase remained in the pores of the stationary phase. The mobile phase was hexane containing THF or chloroform. The amount of an aqueous solution retained in the stationary phase was determined by the Karl-Fischer titration of column effluents rinsed by isopropanol with a Metrohm Model 870 KF Titrino Plus. For chromatographic measurements under frozen conditions, the water-impregnated column was immersed in partially frozen acetonitrile (42 C) and then the temperature was raised to the operation point. Differential scanning calorimetric (DSC) measurements were conducted with a DSC Perkin Elmer Model DSC8500 equipped with a cooling unit.

Results and discussion bCD forms an inclusion complex in an aqueous phase by accommodating a guest molecule in its hydrophobic cavity. Chiral recognition with bCD should therefore occur in an aqueous bCD solution.14 In chiral ice chromatography, a salt was added to the ice stationary phase containing bCD to allow the development of an aqueous liquid phase supported by the ice matrix. The simplest way to retain an aqueous solution in the stationary phase is its impregnation in the pores of a solid support, such as silica gel. This is the same scenario as the rst partition liquid chromatography invented by Martin and Synge.15 However, we have found that simple water-impregnation in silica gel leads to poor separation. Very weak retention was for example conrmed for hexobarbital with silica gel containing an aqueous bCD solution as shown in Fig. 1A. Bulk extraction experiments have revealed the moderate partition of this analyte to an aqueous phase (Kd ¼ 0.2–0.5, depending on the salt concentration in an aqueous phase).10 If the entire separation system is in equilibrium, the retention of hexobarbital should be higher, suggesting that all of the aqueous phase does not act for solute partition. Severe peak broadening occurred for 1,10 -bis-2-naphthol (bisN) even when the peak appeared immediately aer the unretained peak as shown in Fig. S1;† the theoretical plate number was smaller than 15 for this solute. The poor separation efficiency can be interpreted assuming the wide-range connection of the aqueous phase in the extra particle space as schematically illustrated in Fig. 2A. The formation of a massive aqueous phase in the column hinders the effective solute diffusion in the mobile phase, reduces the effective interfacial area, and causes multi-isolated streamlines that are mutually not accessible. It is thus important to prevent aqueous solutions impregnated in silica gel pores from forming a wide-range network in order to attain better separation based on the present idea. We here propose the utilization of a mixed column bed, in which stationary phase beads of entirely different surface properties are incorporated. Since water-impregnated silica gel beads are highly hydrophilic, the contact of the beads causes the water connection in the extra particle space. The incorporation of octadecylsilanized silica gel (ODS) beads is expected to allow the connement of aqueous solution near silica beads (Fig. 2B). Even though the aqueous solution partially came out of the


Fig. 1 Chiral separation of hexobarbital. Column, 4.6 mm i.d. 75 mm. Stationary phase, 1 mM bCD + 75 mM KCl impregnated (A) silica gel, (B) 33% silica gel/ODS and (C) 25% silica gel/ODS. Mobile phase, 3% THF in hexane. Temperature, 20 C.

pores, it should stay in the vicinity of the silica gel beads and would not be spread over a wide range. Fig. 1 demonstrates the successful chiral separation of the hexobarbital enantiomers based on the present idea; mixed stationary phases of 1 : 2 and 1 : 3 (silica : ODS) containing bCD as CS were employed. Although the 1 : 3 stationary phase gives smaller retention due to the reduction of the aqueous phase volume supported in the pores of silica gel, the enantiomers are separated well; the theoretical plate number increased up to ca. 500, corresponding to ca. 1000 for a 15 cm long usual column. In the present case, the retention factor (k) of an analyte is given by10 Kd Vwater 1 þ K1 ½CD þ K1 K2 ½CD2 k ¼ kads þ Vmob where kad represents the retention by the adsorption of a solute, Kd (¼[s]water /[s]mob, the ratio of the solute concentrations in the water phase to that in the mobile phase) is the partition

Fig. 2 Schematic representations of water-impregnated silica gel beads and their spatial isolation by mixed packing with ODS beads.

Analyst, 2014, 139, 1830–1833 | 1831

View Article Online


Analyst

coefficient of the solute, [CD] is the concentration of bCD in the water phase, Vwater and Vmob are the volumes of the water and mobile phases in the column, and K1 and K2 are the association constants for 1 : 1 (s-CD) and 1 : 2 (s-CD2) complexes between a solute (s) and bCD, respectively. For an analyte forming only a 1 : 1 complex, the last term in the right side of the above equation can be omitted. Fig. 3 shows the amount of water retained in a 4.6 75 mm column as a function of a weight ratio of silica gel in the mixed silica/ODS stationary phase. The water amount linearly decreases with decreasing silica gel ratio in the stationary phase. The amount of water retained in the ODS stationary phase is ca. onesixth as small as that in silica gel. The aqueous phase is thus mostly retained in silica gel. The above equation suggests that the retention factor depends on Vwater at constant [CD] as long as kad is constant. A change in the amount of the ODS beads packed in the column may cause a variation of kad. However, it has been conrmed that the adsorption of a solute on the ODS stationary phase is so weak in the present mobile phase condition that ODS is expected to act as the inert spacer to keep silica gel beads not in contact with one another. A decrease in the silica gel fraction in the column packing reduces Vwater and, in turn, solute retention. Therefore, the 25% silica gel column gives lower retention than the 33% counterpart as shown in Fig. 1. Temperature is also an important parameter governing chiral separation. Fig. 4 and 5 show the temperature dependence of various separation parameters obtained for the enantioseparation of hexobarbital and bisN, respectively. Silica gel with large pores (Daiso SP) was employed to examine freezing effects on separation parameters such as the separation ratio (a) and resolution (Rs). Chromatograms measured at various temperatures are shown in Fig. S2 and S3.† In both compounds, the retention increases with decreasing temperature due to the exothermic nature of bCD complexation. A difference in the retention time between hexobarbital enantiomers becomes larger as the temperature decreases; interestingly, the temperature dependence of the bCD complexation of one enantiomer appears very small as shown in Fig. 4. A separation ratio for hexobarbital enantiomers increases from 1.7 at 20 C to 3.0 at 15 C as a result. However, lowering the temperature simultaneously reduces the diffusivity of an analyte and leads to band

Fig. 3 Change in the amount of water retained in stationary phase pores with the weight ratio of silica gel to the entire column packing.

1832 | Analyst, 2014, 139, 1830–1833

Communication

Fig. 4 Temperature dependence of (A) retention times and (B) separation ratio, a, and resolution, Rs, for the enantiomers of hexobarbital. Column, 4.6 mm i.d. 75 mm. Stationary phase, 1 mM bCD + 75 mM KCl impregnated 33% Daiso SP/ODS. Mobile phase, 3% THF in hexane.

broadening. Thus, the temperature affects the separation in two opposite ways, i.e. a difference in retention increases but peak broadening becomes severe with decreasing temperature. The maximum resolution is therefore produced in the middle temperature range around 0 C. In contrast, the separation ratio for bisN only slightly increases with decreasing temperature, and the resolution does not show clear temperature dependence. However, it should be noted that the enantiomers of bisN are not separated at room temperature with this stationary phase as shown in Fig. S2.† DSC measurements have indicated that water in the pores of Daiso SP silica gel is frozen at ca. 20 C; the melting point of formed ice is ca. 2 C (Fig. S4†). It is known that the melting

Fig. 5 Temperature dependence of (A) retention times and (B) separation ratio, a, and resolution, Rs, for the enantiomers of bisN. Column, 4.6 mm i.d. 75 mm. Stationary phase, 1 mM bCD + 75 mM KCl impregnated 33% Daiso SP/ODS. Mobile phase, 1% THF in hexane.


View Article Online

Communication

Analyst

point of water in a small space is lowered and the melting point depression (DT) can be explained by the Gibbs–Thomson effect.16


DT ¼

2T0 gSL VL DHr

where T0 is the melting point (273 K), gSL is the interfacial tension between ice and liquid water (ca. 31.7 mJ m2), VL is the molar volume of liquid water (1.8 105 m3 mol1), DH is the melting enthalpy of ice (6.01 kJ mol1), and r is the radius of the space, in which water is conned, respectively. According to this equation, the melting point of ice is lowered by ca. 1 K in r ¼ 50 nm (in Daiso SP silica gel) and 17 K in the 6 nm pores in Wakosil 10SIL. As described in the Experimental section, the stationary phase was frozen at 42 C and the temperature was raised to the operation point for low temperature operations. Thus, water in the pore of Daiso SP silica gel remains frozen at a temperature lower than 2 C. It is not clear whether the presence of ice is favorable or unfavorable for chiral separation because a clear discontinuity is not found in any retention data in Fig. 4 and 5. However, the temperature is still an important experimental parameter and should be optimized for desired separation. In addition, from a practical point of view, the low temperature operation is advantageous because the column durability is enhanced. The aqueous phase retained in the silica gel pores is likely to be lost during chromatographic operation for a long time period, reducing the separation performance. This column degradation is largely improved by low temperature operation. The column can be repeatedly used over several days. The present scheme was applied to other water-soluble CSs. Table 1 summarizes selected separation data obtained with bCD, 3-hydroxyl-propyl bCD, or L-proline + bCD as CSs at the room temperature. The corresponding chromatograms are shown in Fig. S5 and S6 in the ESI.† Since low temperature is not necessary for efficient chiral separation with the present scheme, silica gel with small pores (Wakosil 10SIL) was employed with ODS. As shown above, though bCD was effective for separation of the enantiomers of hexobarbital and bisN, these were not separated with hydroxyl-propyl bCD. In contrast, bCD did not separate the enantiomers of 3-hydroxy avonon, which were resolved with OHPrbCD though base-line separation was not attained. Thus, the separation selectivity of the stationary phase can easily be varied by changing a loaded CS. The same packed column can be repeatedly used for the preparation of different stationary phases in the present scheme.

Table 1 Selected separation data with a 1 : 2 Wakosil 10SIL-ODS mixed stationary phase impregnated with various CSs

CS

Analyte

a

Rs

Mobile phasea

bCDe

2HFlb Ibuprofen 2HFl 3HFlc BisNd

1.39 1.17 1.22 1.11 1.43

2.38 0.88 1.58 0.98 2.67

3% THF

OHPrbCD f Pro g

A mixed silica gel and ODS column bed has allowed us to prepare chiral stationary phases in a versatile and simple way. This approach has high applicability, i.e. various chiral stationary phases can be prepared without anchoring CSs on a solid support. Also, a similar approach can be used for lipophilic CSs, which can be impregnated in the pores of ODS. Dynamic coating has been known as an effective way to prepare chiral stationary phases.17,18 One of the important differences between dynamic coating and our method is that a partition process is involved in the separation mechanism in the present case. We can manage an extent of the solute partition by changing various experimental parameters such as the volume of the liquid phase, temperature, pH, ionic strength, and type of solvent. Also, the concentration of CS can be changed in the present method, though it is difficult in dynamic coating chromatography to vary the CS concentration.

Acknowledgements This work has been supported by SENTAN from the Japan Science and Technology Agency.

Notes and references 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

3% THF 10% CHCl3

Mobile phases were prepared in hexane. 2-Hydroxy avonon. 3-Hydroxy avonon. d 1,10 -Bis-2-naphthol. e 10 mM bCD. f 10 mM 3-hydroxyl-propyl bCD. g 10 mM L-proline containing bCD. a

Conclusions

17

b

c


18

R. Healey and A. Ghanem, Chirality, 2013, 25, 314–323. H. Lu and G. Chen, Anal. Methods, 2011, 3, 488–508. B. Kasprzyk-Hordern, Chem. Soc. Rev., 2010, 39, 4466–4503. Y. Okamoto and Y. Kaida, J. Chromatogr. A, 1994, 666, 403– 419. G. M. Peng, S. Q. Wu, W. G. Zhang, J. Fan, S. R. Zhang, G. Xu and X. Q. Xu, Anal. Sci., 2013, 29, 637–642. Z. Juvancz and J. Szejtli, TrAC, Trends Anal. Chem., 2002, 21, 379–388. J. Haginaka, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2008, 875, 12–19. H. Y. Aboul-Enein and I. Ali, Chromatographia, 2000, 52, 679– 691. C. V. Hoffmann, R. Pell and M. Laemmerhofer, Anal. Chem., 2008, 80, 8780–8789. T. Shamoto, Y. Tasaki and T. Okada, J. Am. Chem. Soc., 2010, 132, 13135–13137. Y. Tasaki and T. Okada, Anal. Chem., 2006, 78, 4155–4160. Y. Tasaki and T. Okada, Anal. Chem., 2009, 81, 890–897. P. R. Bristow, P. N. Brittain, C. M. Riley and B. F. Williams, J. Chromatogr., 1977, 131, 57–64. K. A. Connors, Chem. Rev., 1997, 97, 1325–1357. A. J. P. Martin and R. L. M. Synge, Biochem. J., 1941, 35, 91. S. J¨ ahnert, F. V. Ch´ avez, G. E. Shaumann, A. Schreiber, M. Sch¨ onhoff and G. H. Findenegg, Phys. Chem. Chem. Phys., 2008, 10, 6039–6051. T. Shinbo, T. Yamaguchi, K. Nishimura and M. Sugiura, J. Chromatogr., 1987, 405, 145–153. M. H. Hyun, H. J. Koo, J. S. Jin and W. Lee, J. Liq. Chromatogr. Relat. Technol., 2000, 23, 2669–2682.

Analyst, 2014, 139, 1830–1833 | 1833

View Article Online

Analyst


PAPER


Binding assay for low molecular weight analytes based on reflectometry of absorbing molecules in porous substrates† Milena Stephan,a Corinna Kramer,b Claudia Steinemb and Andreas Janshoff*a Small molecule sensing is of great importance in pharmaceutical research. While there exist well established screening methods such as EMSA (electrophoretic motility shift assay) or biointeraction chromatography to report on successful binding interactions, there are only a few techniques that allow studying and quantifying the interaction of low molecular weight analytes with a binding partner directly. We report on a binding assay for small molecules based on the reflectivity change of a porous transparent film upon

Received 3rd January 2014 Accepted 14th January 2014

immobilisation of an absorbing substance on the pore walls. The porous matrix acts as a thin optical transparent film to produce interference fringes and accumulates molecules at the inner wall to amplify

DOI: 10.1039/c4an00009a

the sensor response. The benefits and limits of the assay are demonstrated by investigating the binding

www.rsc.org/analyst

of biotin labelled with an atto dye to avidin physisorbed within an anodic aluminium oxide membrane.

1. Introduction Protein interactions with small molecules (M # 500 Da) play an important role in regulatory biological processes, making small molecules potential candidates for future drugs. The action of the most prospective small molecule drugs is based on their ability to form stable complexes with therapeutic targets. Therefore, characterizing the binding kinetics of small molecule–protein complexes is important in pharmaceutical research.1 However, quantifying the affinity of the respective binding partners is rather challenging due to the intrinsically low concentration at which the complexes are formed. Standard kinetic affinity methods may be categorized according to the phase state of the binding partners as heterogeneous or homogeneous. In heterogeneous measurements, one of the binding partners is coupled to a sensor surface, whereas homogeneous methods do not require the immobilisation of any of the binding partners for affinity measurements and are, in general, preferred over heterogeneous methods in protein-small molecule studies. Perhaps the most prominent technique for homogeneous small molecule interaction studies is the electrophoretic mobility shi assay (EMSA).2 In EMSA the analyte is incubated with a xed concentration of a radio-labelled probe and varying amounts of protein. The solutions are run on a native

a

Department of Physical Chemistry, Georg-August University, Göttingen, Germany. E-mail: [email protected]

b

Department of Organic and Biomolecular Chemistry, Georg-August University, Göttingen, Germany † Electronic supplementary 10.1039/c4an00009a

information

(ESI)


available.

See

DOI:

polyacrylamide gel to separate the bound from the free analyte.2 While EMSA is a very sensitive technique, the probes also have several disadvantages, including safety and environmental problems associated with their radioactivity. Homogeneous methods generally do not allow the direct observation of binding events. Something that is more easily accomplished with heterogeneous affinity measurements is uorescence resonance energy transfer (FRET). FRET exploits the energy transfer between two chromophores for sensing purposes. Binding reactions may be sensed with FRET by coupling one interacting agent with a ‘donor’ and its complementary binding partner with an ‘acceptor’, whose absorption spectrum overlaps with the donor's emission spectrum.3 However, using FRET for small molecule sensing can be problematic, since the required labels are oen quite large in comparison to the molecules to be investigated and could interfere with the binding activity of both partners.3 Most heterogeneous label-free methods such as SPR require the immobilisation of the small interaction partner on the surface of a sensor for sensitive detection since it is easier to monitor the adsorption of the larger binding partner if the detection principle relies on a change in refractive index, for instance. The optical or acoustic signal as a function of the compound concentration can be analysed to obtain both kinetic rates and thermodynamic binding constants. Coupling of a small molecule to a sensor surface, however, may inuence its binding ability due to steric reasons.1 Various workarounds to increase sensing sensitivity have been devised also with a focus on surface plasmon resonance (SPR) spectroscopy. Murthy et al. used a strategy to enlarge the detectable mass by adding micelles4 that display the epitopes of interest to the sensor surface. Alternatively, surface-enhanced uorescence spectroscopy can be used to boost the detection level

Analyst, 2014, 139, 1987–1992 | 1987

View Article Online


Analyst

substantially.5 Usually, however, competition experiments readily circumvent the problem of detection sensitivity due to the small molecular mass of the analyte. In competition experiments the analyte of interest competes with a larger and thus detectable analyte for the immobilised ligand. Moreover, there have been efforts to overcome the problem of small molecule detection by improving the sensitivity of biosensors in general.6–11 In 2010, Guo et al. reported the development of a photonic crystal biosensor for real-time biomolecular binding detection.12 They used a one-dimensional photonic crystal structure in a totalinternal-reection geometry, creating a sensor that functions as a Fabry–Perot resonator. Using their sensor assembly, they were able to detect the binding of D-biotin (M ¼ 244 Da) to streptavidin covalently attached to a surface at D-biotin concentrations as small as 1 mM, but were not able to determine interaction kinetics with their method.12 Prior to this work, Piehler et al. used reectometric interference spectroscopy (RIfS) to investigate the binding of biotin to a polymerised streptavidin network. RIfS allows determination of the optical thickness (refractive index physical thickness) of a transparent lm deposited on a reective surface by recording the reectivity spectrum of the sample. When the sample is irradiated with white light at the right angle, an interference pattern will arise in the reectivity spectrum originating from the superposition of partial light-beams reected at the different interfaces of the substrate. The modulation of the interference pattern depends on the path length the light travels inside the transparent lm, in other words its optical thickness. Piehler et al. were able to measure kinetic curves with their system, but the curves displayed a very large dri, affecting the accuracy of the measurement requiring polymerisation of one binding partner.13 Here, we demonstrate a new binding assay for low molecular weight analytes, whose sensing principle is based on the change in reectivity of a thin porous lm upon immobilisation of an absorbing substance. The accumulation of an absorbing agent on the walls of a porous transparent lm leads to a detectable change of its extinction coefficient. To our knowledge, this concept has not been used for biomolecular interaction studies before. The potential of the method will be demonstrated by applying the assay to quantify the binding of Atto488-Biotin to avidin physisorbed on anodic aluminium oxide (AAO) substrates. The association of avidin with biotin is among the strongest known non-covalent protein ligand interactions. The equilibrium dissociation constant KD of the biotin–avidin complex was found to be 1 1015 M in solution.14,15 The high affinity of avidin for biotin has led to their use as capture elements in a multitude of biotechnological applications and gave rise to many commercial products.16 Proteins and peptides, nucleic acids and aptamers have been modied with biotin and immobilised to an avidin surface for use in biological sensing and immunoassays.17,18 While there are many studies which characterized the binding of avidin to biotin, there were not many successful attempts to try the reverse.16

2.

Experimental

2.1. Materials Common chemicals as well as avidin from egg white and Atto488-Biotin were acquired from Sigma Aldrich (Deisenhofen,

1988 | Analyst, 2014, 139, 1987–1992

Paper

Germany). Aluminium with 99.999% purity was purchased from GoodFellow (Coraopolis, Pennsylvania, USA). 2.2. Preparation of Anodic Aluminium Oxide (AAO) membranes 2.2.1. Aluminium metal preparation and electrochemical polishing. Aluminium was annealed at 500 C overnight to increase the self-organisation of the AAOs to be fabricated. The surface of the substrates was polished using a solution made from concentrated acids (H2O/H3PO4/H2SO4, 1/1/1). The aluminium plates were warmed to 70 C, and a constant voltage of 25 V was applied. The electrochemical dissolution of aluminium was continued until the surface became reective (usually 3–5 min). The process was repeated until the surface was as close to mirrorlike as possible, but was done at least 3 times for aluminium supports meant for AAO growth. 2.2.2. AAO membrane growth. The electrochemically polished aluminium plates were placed in an anodization chamber, lled with 0.3 M oxalic acid and cooled under stirring to 1.0 C. The rst anodisation step was carried out at 40 V for 3 h to pre-texture the surface. The oxide layer was dissolved using 5 vol% H3PO4 solution over 2–3 h. Anodisation was repeated under the same conditions for 75 min, resulting in 3 mm thick porous membranes. AAOs produced in 0.3 M oxalic acid at 40 V provide an interpore distance of 100 nm and a diameter of 25 nm. The pores were widened to diameters of 60 nm through slow dissolution of the pore material in 5 vol% of H3PO4 for 50 min. Functionalisation of the substrate was not necessary. Prior to measurement, the AAO chip was treated with oxygen plasma for 1 min. 2.3. Performing small molecule binding experiments 2.3.1. Set-up. The binding studies were carried out with a home-made RIfS set-up. This type of instrumental assembly has been described in the literature previously.19,20 A tungsten halogen lamp (LS-1, Ocean Optics, Dunedin, Florida, USA) is used as the light source, from which radiation is coupled into a y-shaped optical bre (also purchased from Ocean Optics), and guided to a measurement chamber holding a transducer chip, which is irradiated from the top perpendicular to the surface (Fig. 1). The optical bre consists of six illuminating bres placed concentrically around one collecting bre, resulting in a sensing area of 1 mm2. The reected light is gathered by the collecting optical bre and guided to a spectrometer (Nanocalc2000-UV/VIS, Ocean Optics). The measurement chamber was built to allow the exchange of uids during experiments. For this purpose, a ow channel was milled (3 10 1 mm3; volume 30 ml) into an acrylic glass cover. A continuous ow was generated via a peristaltic pump (Perimax, Spetec, Erding, Germany). The ow rate was usually set to 1.2 ml min1. 2.3.2. Adsorption of avidin within AAO membranes. A chip was placed in the measurement chamber of the RIfS set-up and a baseline was recorded in buffer (20 mM HEPES, 100 mM Na2SO4, pH ¼ 7.0). Aer approximately 5 min, a 0.1 mg ml1 solution of avidin was added and allowed to circulate until the


View Article Online


Paper

Analyst

Schematic drawing of the RIfS setup: light from a tungsten halogen lamp is coupled into a y-shaped optical fibre. The radiation is guided to a flow cell containing the sample. The light reflected by the sample is collected again by the fibre and guided to the spectrometer. Fig. 1

adsorption reaction of avidin to the pore walls of the substrate reached equilibrium (Fig. 3). Buffer was own through the system until the measured signal stabilized again and the RIfS measurement was terminated. 2.3.3. Monitoring of biotin–avidin interaction. The evaluation soware was set to track the reectivity values of 501 nm radiation (the absorption maximum of Atto488-Biotin). A baseline was recorded 10 min before Atto488-Biotin was added in increasing concentrations (0.5 nM–15 nM) and the system was le to equilibrate aer each analyte addition (Fig. 4A). Eventually, the system was rinsed with buffer for 30 min. To determine the equilibrium dissociation constant, measurements were done in triplicate.

Fig. 2 (A) SEM image of the AAO chip (scale bar: 500 nm). (B) Scheme of the sensing concept. Avidin (red dots) adsorbs on the pore walls of the AAO chip and causes a red shift of the interference spectrum as shown in (C) (blue and red curves). When the dye-labelled biotin (green dots leading to the green beam) binds to the immobilised avidin, the reflectivity of the system is attenuated at the absorption maximum of the Atto dye, which is shown as a drop in the spectrum in (C) (the green curve).

t12 ¼

where n1 indicates the refractive index of the ambient medium (water) and n2 is the refractive index of the AAO membrane. n2 becomes a complex quantity n2 ¼ neff ikeff, if an absorbing substance is present in the porous lm. neff and keff may be calculated according to an effective medium model published by Garahan et al.21 The model predicts the optical properties of nanoporous thin lms with horizontally aligned cylindrical nanopores with a specied diameter and porosity.21 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 neff 2 ¼ (2) A þ A2 þ B 2 ; 2

2.4. Data processing 2.4.1. Spectrum analysis. All reectivity spectra recorded during avidin immobilisation and biotin binding were evaluated with RIfS soware to track the changes in optical thickness. Additionally, the reectivity changes for 501 nm radiation, the absorption maximum of the employed uorescent dye, were read-out. The RIfS evaluation was executed using wavelengths ranging from 550 nm to 700 nm, in order to avoid the analysis being compromised by the absorption effects of the atto dye. A typical reectivity spectrum is shown in Fig. 2C. 2.4.2. Inuence of light-absorbing molecules on reectivity. Filling the pores with an absorbing medium such as a liquid with dissolved dyes changes the reectivity of the effective medium comprising the porous matrix (alumina oxide) and the liquid in the pores. It is, however, not possible to detect small amounts of absorbing molecules dissolved in buffer (mM–nM regime) as detailed below. Therefore, adding a solution with dissolved dyes to the porous alumina oxide below 1 mM remains almost invisible to the reectometric sensor. Only molecules accumulating at the pore walls change the reectivity of the thin lm signicantly enough to be detected by a change in reectivity. The Fresnel coefficients for normal incidence are r12 ¼

n1 n2 ; n1 þ n2


(1)

2n1 ; n1 þ n2

keff 2 ¼

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 A þ A2 þ B 2 ; 2

with A ¼ f (nd2 kd2) + (1 f )(nm2 km2)

(3)

B ¼ 2ndkd f + 2nmkm(1 f )

(4)

and

f denotes the lling factor of domains (pores) present in the substrate, nm and km are the real and imaginary parts of the refractive index of the substrate material, while nd and kd are the real and imaginary parts of the refractive index of the material forming the domains, respectively. Having calculated the refractive index of the porous layer and the Fresnel coefficients of the system, its reectivity R may be calculated by 2 r12 þ t12 t21 r23 e2id R¼ ; (5) 1 r21 r23 e2id with the phase difference d d¼

2p dn2 l

(6)

between the light-beam reected on the surface of the substrate and the light-beam refracted into the AAO lm. r12, r21, r23, t12, and

Analyst, 2014, 139, 1987–1992 | 1989

View Article Online


Analyst

Paper

t21 are the typical Fresnel coefficients (t: transmission, r: reection) between the media numbered in the subscripts (1: aqueous solution, 2: AAO membrane, 3: alumina substrate).22 We found that for a dye such as Atto488 with an extinction coefficient of 8.7 104 M1 cm1, and a thickness of alumina oxide of 3000 nm, a concentration of >1 mM in a porous layer of the given geometry and material would be necessary to cause an appreciable change in reectivity that is detectable by our sensor 4p set-up.22 Note that a ¼ kd , with a being the extinction coeffil ax cient (I ¼ I0e ), kd being the imaginary part of the refractive index and l being the corresponding wavelength (here 590 nm). In other words, for a signal change as high as the one measured in our experiment, many more molecules would have to be present in each pore than simply those in the lling solution. This leads to the conclusion that we are able to see the reectivity changes at very low concentrations since the labelled compound accumulates on the pore walls of the AAO lm. The nding agrees well with the concept of the experiment, since the pore walls are functionalised with avidin accumulation of labelled biotin and thus absorbing dye molecules is not only wanted but also expected. 2.4.3. Equilibrium dissociation constant. The reaction of Atto488-Biotin binding to avidin can be described as B þ Asurf # BAsurf ; where Asurf represents the avidin adsorbed in the porous lm, B represents the free biotin in solution and BAsurf represents the bound complex. The equilibrium dissociation constant KD can be obtained from the corresponding Langmuir equation23 ½B KD DR ¼ ; ½B 1þ KD DRmax

(7)

where DRmax represents the change in reectivity obtained when all binding sites on the surface are occupied and DR is the steady state signal for any given biotin concentration. KD is obtained by tting the parameters KD and DRmax of eqn (7) to a graph with DR as a function of concentration.

3.


As a sensing principle for this small molecule binding assay we used a variation of reectometric interference spectroscopy (Fig. 2). If a substance adheres to the walls of a porous AAO lm, its refractive index changes. The refractive index may be inuenced in two ways. First, the deposition of the material alters the real part of the refractive index of the porous lm, increasing its refractive number and thus its optical thickness, which becomes visible as a red-shi of the interference pattern in the reectivity spectrum of the lm (Fig. 2B and C: blue to red curve). This effect was used to measure the adsorption of avidin in the porous layer. Secondly, another way to inuence the refractive index of the porous lm is to deposit a small dye-labelled agent on the pore walls. This leads to a change of the extinction coefficient of the porous lm. In other words, the reectivity of the AAO lm is attenuated at the absorption maximum of the immobilised

1990 | Analyst, 2014, 139, 1987–1992

agent, which becomes visible as a drop in the interference pattern of the reectivity spectrum (Fig. 2B and C: red vs. green curve). We utilised this phenomenon to sense the binding of Atto488-Biotin to immobilised avidin as it is more sensitive than monitoring a shi in optical thickness. Fig. 3 depicts the changes in optical thickness during the deposition of avidin in the porous matrix. The wavelength range used to calculate DOT is indicated by the brackets in Fig. 2C. The adsorption of avidin on the pore walls of the AAO lm is clearly visible in this graph as an increase in OT of 48 nm immediately aer protein addition at point ‘a’. Avidin adsorbs on AAOs via electrostatic interaction without the help of any surface functionalisation at a pH higher than 3.0.24 The average surface concentration of avidin hG(t)i as a function of time is at low values of time proportional to the bulk concentration cb of avidin and the ratio between pore radius Rpore and thickness of Rpore cb t.24 the AAO membrane d leading to hGðtÞif 2d Excess protein in solution was removed from the system causing the measured signal to drop by 5 nm (point ‘b’). Since avidin was not covalently attached to the pore walls, the continuous ow of liquid through the measurement chamber might affect desorption of avidin molecules, which could be the cause of the constant dri of 1.3 nm h1 seen in the graph. Once the signal stabilized, Atto488-Biotin was added in increasing overall concentration. Fig. 4A displays the change in reectivity recorded at 501 nm, close to the adsorption maximum of the attached uorescent dye Atto488 (as indicated in Fig. 2C). The original data shown in grey were smoothed by averaging over 5 data adjacent points and t with a monoexponential decay to illustrate the suitability of the Langmuir description. All further data analyses were done with the original data. Once it was conrmed that the reectivity at the wavelength is stable, the addition of labelled biotin in very low concentrations was started. The binding of Atto488-Biotin to avidin in the AAO membrane is clearly visible in the graph as a decrease of its reectivity (Fig. 4A). An adsorption isotherm was measured by increasing the concentration of Atto488-Biotin stepwise (concentrations are shown in the graph) and waiting for equilibrium to establish aer each successive addition (see ESI, Fig. 1†).

Fig. 3 Change of optical thickness during addition of avidin to an AAO membrane. (a) Avidin addition. The adsorption of avidin on the walls of the porous film leads to an increase of the optical thickness. (b) Rinsing with buffer.


View Article Online


Paper

Analyst

Fig. 4 (A) Change in reflectivity at 501 nm due to adsorption of Atto488-Biotin to an avidin-coated AAO membrane. The original data are shown in grey (smoothed) and the corresponding monoexponential fit in red. Atto488-Biotin was added at an overall concentration of 2 nM. (B) Adsorption isotherm; change in reflectivity as a function of biotin concentration (solid circles). The parameters of the Langmuir equation were fitted to the data (red curve), yielding a KD value of 1.2 0.3 nM for the avidin–Atto488-Biotin interaction.

In order to prove that the labelled biotin binds to the physisorbed avidin and not the pore wall itself, Atto488-Biotin in a concentration of 100 nM was allowed to circulate over an untreated AAO chip (data not shown). No signicant change of its reectivity was observable upon addition of the labelled compound, indicating that the observations during the actual experiment can safely be attributed to the specic binding of biotin to avidin. Biotin without a label did not produce any noticeable change in reectivity (ESI, Fig. 2†). The dissociation constant KD of the avidin–biotin complex can be determined from adsorption isotherm measurements as shown in graph B of Fig. 4 by tting the parameters of the Langmuir equation (Experimental section, eqn (7)) to the acquired data. A value of 1.6 0.6 nM was found (the mean value of three independent measurements). This value differs signicantly from the dissociation constant of the complex in solution (KD ¼ 1 1015 M), but is comparable with other surface based techniques, which also reported values in the nanomolar regime.14–16 There are two possible explanations for the difference of the KD value determined in solution and the KD values determined with surface-based sensors. In 2000, Ben-Tal et al. reported that for affinity measurements on surfaces, an entropy loss upon adsorption of one interacting agent on a surface needs to be considered, resulting from the loss of translational degrees of freedom of the said interaction partner that decrease the free energy of binding of the complex by around 1.5 kBT per degree of freedom.25 This accounts for higher KD values determined from surface based techniques compared to those measured in solution. This argument is further strengthened by a study published by Zhao and Reichert that showed the binding kinetics of streptavidin interacting with immobilised biotin differ, if biotin is directly attached to the surface or a spacer is present.26 Another interesting study was conducted by Sheehan and Whitman in 2005.27 They calculated the detection limits for nano- and microscale biosensors and were able to show a signicant dependence on surface geometry and analyte ow. They deduced that a at sensor surface is least suitable for low analyte concentration sensing and furthermore that even with ideal geometry, detection schemes based on static or conventional microuidic ow This journal is © The Royal Society of Chemistry 2014

systems are unlikely to exceed the femtomolar range. We have to come to the same conclusion as Ben-Tal et al. that dissociation constants found for surface-based reactions are usually considerably higher than those measured in solution. The dynamic range of the sensor scheme lies in between 0.1 and 10 nM for biotin to avidin. The lower concentration regime is limited by the signal to noise ratio that is inherently related to the differences in refractive index and absorbance of the substrate and adsorbent.22

4. Conclusion On the basis of the avidin–biotin model system for high affinity small molecule interaction, we were able to demonstrate the high sensitivity of a new sensing approach which allows for the direct observation and quantication of small molecule binding events. Using uorescently labelled biotin, we were able to quantify its binding to surface immobilised avidin by sensing the change in reectivity of the transducer chip at the absorption maximum of the uorescent dye. Even though, the technique also relies on labelled compounds for detection purposes, which might be considered as a drawback, it offers advantages compared to other small molecule binding assays. The labelling itself is not restricted to rather bulky uorescent dyes. The only criterion the label would have to meet is that its absorption maximum lies in between the medium UV and the near infrared spectrum to produce an appreciable absorption peak in the reectivity spectrum. This allows for the use of smaller compounds than the conventional uorescent dyes. Another advantage compared to assay formats based on uorescence is that the measurements are not notably affected by photobleaching. Furthermore, employing an anodic alumina oxide membrane as a transducer chip, a well established substrate for biosensing, opens the possibility of various applications for this binding assay, since a library of surface functionalisation is readily available. Yet another advantage of this sensor compared to other techniques based on planar surfaces is that it offers the opportunity to directly observe binding events and to measure equilibrium dissociation constants in the low nanomolar regime with a very simple and inexpensive instrumental

Analyst, 2014, 139, 1987–1992 | 1991

View Article Online

Analyst

arrangement. In the future, it is conceivable to combine the AAO chip with other more sensitive optical transducers.28

Acknowledgements


We gratefully acknowledge the invaluable help of Michaela Klingebiel by preparing AAO membranes. Financial support was granted by the DFG through grant SFB 937 (A08).

References 1 J. Bao and S. N. Krylov, Volatile kinetic capillary electrophoresis for studies of protein-small molecule interactions, Anal. Chem., 2012, 84, 6944–6947. 2 A. E. Dahlberg, C. W. Dingman and A. C. Peacock, Electrophoretic characterization of bacterial polyribosomes in agarose–acrylamide composite gels, J. Mol. Biol., 1969, 41, 139–147. 3 R. Laine, D. W. Stuckey, H. Manning, S. C. Warren, G. Kennedy, D. Carling, C. Dunsby, A. Sardini and P. M. W. French, Fluorescence lifetime readouts of troponin-c-based calcium fret sensors: a quantitative comparison of cfp and mtfp1 as donor uorophores, PLoS One, 2012, 7, e49200. 4 B. N. Murthy, N. H. Voelcker and N. Jayaraman, Evaluation of a-D-mannopyranoside glycolipid micelles–lectin interactions by surface plasmon resonance method, Glycobiology, 2006, 16, 822–832. 5 S. Ekgasit, F. Yu and W. Knoll, Fluorescence intensity in surface-plasmon eld-enhanced uorescence spectroscopy, Sens. Actuators, B, 2005, 104, 822–832. 6 J. N. Anker, W. P. Hall, O. Lyandres, N. C. Shah, J. Zhao and R. P. Van Duyne, Biosensing with plasmonic nanosensors, Nat. Mater., 2008, 7(6), 442–453, DOI: 10.1038/nmat2162. 7 R. Miller, V. B. Fainerman, A. V. Makievski, J. Kr¨ agel, D. O. Grigoriev, V. N. Kazakov and O. V. Sinyachenko, Dynamics of protein and mixed protein/surfactant adsorption layers at the water/uid interface, Adv. Colloid Interface Sci., 2000, 86(1–2), 39–82. 8 A. K. Naik, M. S. Hanay, W. K. Hiebert, X. L. Feng and M. L. Roukes, Towards single-molecule nanomechanical mass spectrometry, Nat. Nanotechnol., 2009, 4(7), 445–450, DOI: 10.1038/nnano.2009.152. 9 V. C. Ozalp, Dual-polarization interferometry for quantication of small molecules using aptamers, Anal. Bioanal. Chem., 2012, 402(2), 799–804, DOI: 10.1007/ s00216-011-5499-9. 10 F. Patolsky, G. Zheng and C. M. Lieber, Nanowire sensors for medicine and the life sciences, Nanomedicine, 2006, 1(1), 51– 65, DOI: 10.2217/17435889.1.1.51. 11 H. Habib Qazi, A. B. Bin Mohammad and M. Akram, Recent progress in optical chemical sensors, Sensors, 2012, 12(12), 16522–16556, DOI: 10.3390/s121216522.

1992 | Analyst, 2014, 139, 1987–1992

Paper

12 Y. Guo, J. Y. Ye, C. Divin, B. Huang, T. P. Thomas, J. R. Baker and T. B. Norris, Real-time biomolecular binding detection using a sensitive photonic crystal biosensor, Anal. Chem., 2010, 82, 5211–5218. 13 J. Piehler, A. Brecht and G. Gauglitz, Affinity detection of low molecular weight analytes, Anal. Chem., 1996, 68, 139–143. 14 N. M. Green, Avidin and streptavidin, Methods Enzymol., 1990, 184, 51–67. 15 M. Wilchek and E. A. Bayer, Introduction to avidin–biotin technology, Methods Enzymol., 1990, 184, 5–13. 16 G. R. Broder, R. T. Ranasinghe, C. Neylon, H. Morgan and P. L. Roach, Kinetics and thermodynamics of biotinylated oligonucleotide probe binding to particle-immobilized avidin and implications for multiplexing applications, Anal. Chem., 2011, 83, 2005–2011. 17 J. W. Chung, J. M. Park, R. Bernhardt and J. C. Pyun, Immunosensor with a controlled orientation of antibodies by using neutravidin-protein a complex at immunoaffinity layer, J. Biotechnol., 2006, 126, 325–333. 18 N. Haddour, J. Chauvin, C. Gondran and S. Cosnier, Photoelectrochemical immunosensor for label-free detection and quantication of anti-cholera toxin antibody, J. Am. Chem. Soc., 2006, 128(30), 9693–9698. 19 G. Gauglitz, A. Brecht, G. Kraus and W. Nahm, Chemical and biochemical sensors based on interferometry at thin (multi-) layers, Sens. Actuators, B, 1993, 11, 21–27. 20 A. Janshoff, K. P. S. Dancil, C. Steinem, D. P. Greiner, V. S. Y. Lin, C. Gurtner, K. Motesharei, M. J. Sailor and M. R. Ghadiri, Macroporous p-type silicon Fabry–Perot layers. fabrication, characterization, and applications in biosensing, J. Am. Chem. Soc., 1998, 46, 12108–12116. 21 A. Garahan, L. Pilon and J. Yin, Effective optical properties of absorbing nanoporous and nanocomposite thin lms, J. Appl. Phys., 2007, 101, 014320. 22 O. Stenzel, The Physics of Thin Film Optical Spectra: An Introduction, Springer Series in Surface Sciences, 2010, pp. 107–110 23 H. J. Butt, K. Graf and M. Kappl, Physics and Chemistry of Interfaces, Wiley-VCH Verlag GmbH and Co KGaA, Weinheim, 1st edn, 2003. 24 T. D. Lazzara, I. Mey, C. Steinem and A. Janshoff, Benets and limitations of porous substrates as biosensors for protein adsorption, Anal. Chem., 2011, 83, 5624–5630. 25 N. Ben-Tal, B. Honig, C. K. Bagdassarian and A. Ben-Shaul, Association entropy in adsorption processes, Biophys. J., 2000, 79, 1180–1187. 26 S. Zhao and W. M. Reichert, Inuence of biotin lipid surfacedensity and accessibility of avidin binding to the tip of an optical ber, Langmuir, 1992, 8, 2785–2791. 27 P. E. Sheehan and L. J. Whitman, Detection limits for nanoscale biosensors, Nano Lett., 2005, 5, 803–807. 28 M. A. Cooper, Optical biosensors in drug discovery, Nat. Rev. Drug Discovery, 2002, 1(7), 515–528, DOI: 10.1038/nrd838.


View Article Online

Analyst


PAPER


Detection of theophylline utilising portable electrochemical sensors Tiancheng Wang, Edward P. Randviir and Craig E. Banks* The electrochemical oxidation of theophylline (TP) is investigated utilising screen-printed electrodes. Through thorough investigation of pH, we propose a reaction mechanism, finding that the oxidation of TP is stable over a wide pH range, in particular under acidic conditions. Conversely under alkaline conditions, theophylline fouls the electrode surface. The screen-printed carbon sensors are applied towards the electroanalytical sensing of TP with a remarkable amount of success in aqueous solution at

Received 10th January 2014 Accepted 25th January 2014

physiological pH. The screen-printed sensors have been shown to be applicable to the detection of TP at unharmful, medicinally relevant (55–110 mM), and toxic concentrations in aqueous media at

DOI: 10.1039/c4an00065j

physiological pH. Thus this work presents a proof-of-concept approach towards TP detection utilising

www.rsc.org/analyst

sensors commonly implemented in point-of-care applications.

Introduction Theophylline (TP) is a commonly utilised drug clinically for treatment of respiratory diseases such as asthma due to its bronchodilatory effects.1 However TP exhibits a narrow safety range2 which means technologies are required which have the ability to monitor the levels of TP within the body. Currently the clinical practise requires the clinic to regularly monitor plasma by taking blood samples from patients, less chronic conditions may develop.3 Such blood samples are fairly large (in excess of 25 mL and normally more than one sample is taken at once) and thus patients may consider providing these regular blood samples as problematic. Nevertheless theophylline still has to be monitored due to its toxicological effects; thus scaling down the detection technology for use on-site, with small sample sizes ( 0.2 or |DAbs| < 0.2. The Identity and XOR gates operating in parallel were realized in a ow system (m-Slide III 3in1 Flow Kit; ibidi GmbH) outlined in Scheme 2. The solutions containing biochemicals representing Inputs A and B with the variable binary logic values 0 and 1, as well as the constant composition of the “machinery” represented by mixed NADH/NAD+ were pumped through ow devices containing immobilized enzymes (see details in the ESI†) biocatalyzing chemical transformations mimicking logic operations. The ow design of the biochemical device allowed “clocking” (temporal control) and spatial separation of the reaction steps. While the Identity gate realized in a single ow unit performed a simple biocatalytic transformation of PNPP to PNP yielding an optically readable signal, the XOR gate was composed of two ow-through units: one modied with G6PDH that reduced NAD+ when G6P was available; another modied with LDH that oxidized NADH in the presence of Pyr, Scheme 2. Therefore, the optical absorbance corresponding to the NADH


View Article Online

Communication

Analyst

(A and B) Optical responses of the Identity and XOR parts of the CNOT gate, respectively, upon application of input combinations: (a) 0, 0; (b) 0, 1; (c) 1, 0 and (d) 1, 1. The insets show the optical outputs obtained after 400 s of the flow system operation.


Fig. 1

Scheme 2 Experimental realization of the biocatalytic CNOT gate in the flow device (abbreviations and processes are explained in the text and technical details are given in the ESI†).

concentration decreased in the presence of Pyr and absence of G6P (input combination 1, 0) and increased in the absence of Pyr and presence of G6P (input combination 0, 1). Since the NADH concentration and the corresponding absorbance were changed in different directions (decreasing and increasing), the Output B value 1 was dened as the absolute value of the absorbance change. The ow cells modied with G6PDH and LDH were optimized (balanced) in such a way that the simultaneous presence of Pyr and G6P (input combination 1, 1) resulted in no changes in the concentration of NADH, meaning logic value 0 for Output B. Balancing of the biocatalytic reaction rates was achieved by optimization of input concentrations for Pyr and G6P for the specic activity of the enzymes immobilized in the ow units. Obviously, the absence of Pyr and G6P (input combination 0, 0) did not result in any reaction and maintained the NADH absorbance resulting in logic 0 for Output B. These biocatalytic processes performed in the ow system allowed operation of the XOR part of the CNOT gate. It should be noted that the absorbance measurements for the solutions reacted in the ow device were performed vs. the “machinery” solution containing NADH/NAD+ applied to the reference channel of the spectrophotometer, thus reecting absorbance difference in the reacting solutions rather than their full absorbance. Fig. 1 shows the experimental data obtained upon application of the input signals in different logic combinations. Fig. 1(A) shows the absorbance changes observed in Output A (from the Identity gate). In the absence of PNPP only minor absorbance changes were observed, meaning an Output A logic 0 value. Each time application of PNPP (regardless of the presence or absence of any other species) resulted in the absorbance increase at lmax ¼ 420 nm reecting the formation of PNP and resulting in the output logic 1; thus, Input A was directly copied to Output A. Fig. 1(B) shows the XOR gate performance where only unbalanced input signals 0, 1 and 1, 0 resulted in the absorbance changes (output 1), while the absence of the reacting Pyr and G6P (inputs 0, 0) or their balanced application (inputs 1, 1) resulted in no absorbance changes (output 0).


The rst results for the CNOT gate realized in a biocatalytic ow system are promising for integrating the new reversible gate into complex biomolecular logic networks. The logic processing of biomolecular signals in the reversible mode will be particularly benecial for biomedical/biosensing applications where each combination of the output signals corresponds to the unique pattern of the input signals, thus allowing restoration of the original input values. Application of logically reversible gates in the analysis of biomedically important biomarker signaling on various physiological dysfunctions, similarly to previously reported injury diagnostics,40–43 is feasible. Technological realization of the information processing systems in uidic devices allows “clocking” (temporal control) and spatial separation of the various steps of multistage biochemical processes, thus providing novel options for their sophistication and functional exibility.

Acknowledgements This work was supported by the National Science Foundation (award CBET-1066397).

References 1 Unconventional Computation. Lecture Notes in Computer Science, ed. C. S. Calude, J. F. Costa, N. Dershowitz, E. Freire and G. Rozenberg, Springer, Berlin, 2009, vol. 5715. 2 Molecular and Supramolecular Information Processing – From Molecular Switches to Unconventional Computing, ed. E. Katz, Willey-VCH, Weinheim, 2012. 3 A. P. De Silva, S. Uchiyama, T. P. Vance and B. Wannalerse, Coord. Chem. Rev., 2007, 251, 1623–1632. 4 K. Szacilowski, Chem. Rev., 2008, 108, 3481–3548. 5 Biomolecular Computing – From Logic Systems to Smart Sensors and Actuators, ed. E. Katz, Willey-VCH, Weinheim, 2012. 6 Y. Benenson, Nat. Rev. Genet., 2012, 13, 455–468. 7 E. Katz and V. Privman, Chem. Soc. Rev., 2010, 39, 1835–1857. 8 E. P´ erez-Inestrosa, J.-M. Montenegro, D. Collado, R. Suau and J. Casado, J. Phys. Chem. C, 2007, 111, 6904–6909. 9 J. Cervera and S. Mafe, ChemPhysChem, 2010, 11, 1654–1658. 10 P. Remon, R. Ferreira, J.-M. Montenegro, R. Suau, E. PerezInestrosa and U. Pischel, ChemPhysChem, 2009, 10, 2004– 2007.

Analyst, 2014, 139, 1839–1842 | 1841

View Article Online


Analyst

11 P. Rem´ on, M. Hammarson, S. Li, A. Kahnt, U. Pischel and J. Andr´ easson, Chem. – Eur. J., 2011, 17, 6492–6500. 12 R. Orbach, F. Remacle, R. D. Levine and I. Willner, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 21228–21233. 13 W. Sun, C. H. Xu, Z. Zhu, C. J. Fang and C. H. Yan, J. Phys. Chem. C, 2008, 112, 16973–16983. 14 W. Sun, Y.-R. Zheng, C.-H. Xu, C.-J. Fang and C.-H. Yan, J. Phys. Chem. C, 2007, 111, 11706–11711. 15 D. B. Liu, W. W. Chen, K. Sun, K. Deng, W. Zhang, Z. Wang and X. Y. Jiang, Angew. Chem., Int. Ed., 2011, 50, 4103–4107. 16 G. J. Brown, A. P. de Silva and S. Pagliari, Chem. Commun., 2002, 2461–2463. 17 D. Margulies, G. Melman and A. Shanzer, J. Am. Chem. Soc., 2006, 128, 4865–4871. 18 J. Zhou, M. A. Arugula, J. Hal´ amek, M. Pita and E. Katz, J. Phys. Chem. B, 2009, 113, 16065–16070. 19 V. Privman, M. A. Arugula, J. Hal´ amek, M. Pita and E. Katz, J. Phys. Chem. B, 2009, 113, 5301–5310. 20 R. Baron, O. Lioubashevski, E. Katz, T. Niazov and I. Willner, Angew. Chem., Int. Ed., 2006, 45, 1572–1576. 21 Y. F. Wang, J. W. Sun, X. C. Zhang and G. Z. Cui, Adv. Sci. Lett., 2011, 4, 383–390. 22 V. Privman, J. Comput. Theor. Nanosci., 2011, 8, 490–502. 23 M. Pita, M. Kr¨ amer, J. Zhou, A. Poghossian, M. J. Sch¨ oning, V. M. Fernandez and E. Katz, ACS Nano, 2008, 2, 2160–2166. 24 A. Poghossian, K. Malzahn, M. H. Abouzar, P. Mehndiratta, E. Katz and M. J. Sch¨ oning, Electrochim. Acta, 2011, 56, 9661–9665. 25 M. Kr¨ amer, M. Pita, J. Zhou, M. Ornatska, A. Poghossian, M. J. Sch¨ oning and E. Katz, J. Phys. Chem. C, 2009, 113, 2573–2579. 26 R. Pei, E. Matamoros, M. Liu, D. Stefanovic and M. N. Stojanovic, Nat. Nanotechnol., 2010, 5, 773–777. 27 S. Roy and M. Prasad, Opt. Eng., 2010, 49, 065201.

1842 | Analyst, 2014, 139, 1839–1842

Communication

28 J. P. Klein, T. H. Leete and H. Rubin, Biosystems, 1999, 52, 15–23. 29 R. Landauer, IBM J. Res. Dev., 1961, 5, 183–191. 30 T. Toffoli, Int. J. Theor. Phys., 1982, 21, 165–175. 31 E. Fredkin and T. Toffoli, Int. J. Theor. Phys., 1982, 21, 219– 253. 32 C. H. Bennett, IBM J. Res. Dev., 1973, 17, 525–532. 33 E. Katz, J. Wang, M. Privman and J. Hal´ amek, Anal. Chem., 2012, 84, 5463–5469. 34 M. W. Toepke, V. V. Abhyankar and D. J. Beebe, Lab Chip, 2007, 7, 1449–1453. 35 K. Scida, B. L. Li, A. D. Ellington and R. M. Crooks, Anal. Chem., 2013, 85, 9713–9720. 36 J. L. O'Brien, G. J. Pryde, A. G. White, T. C. Ralph and D. Branning, Nature, 2003, 426, 264–267. 37 C. Monroe, D. M. Meekhof, B. E. King, W. M. Itano and D. J. Wineland, Phys. Rev. Lett., 1995, 75, 4714–4717. 38 M. Siomau and S. Fritzsche, Eur. Phys. J. D, 2010, 60, 417– 421. 39 V. Privman, J. Zhou, J. Hal´ amek and E. Katz, J. Phys. Chem. B, 2010, 114, 13601–13608. 40 L. Hal´ amkov´ a, J. Hal´ amek, V. Bocharova, S. Wolf, K. E. Mulier, G. Beilman, J. Wang and E. Katz, Analyst, 2012, 137, 1768–1770. 41 J. Zhou, J. Hal´ amek, V. Bocharova, J. Wang and E. Katz, Talanta, 2011, 83, 955–959. 42 J. Hal´ amek, V. Bocharova, S. Chinnapareddy, J. R. Windmiller, G. Strack, M.-C. Chuang, J. Zhou, P. Santhosh, G. V. Ramirez, M. A. Arugula, J. Wang and E. Katz, Mol. BioSyst., 2010, 6, 2554–2560. 43 J. Hal´ amek, J. R. Windmiller, J. Zhou, M.-C. Chuang, P. Santhosh, G. Strack, M. A. Arugula, S. Chinnapareddy, V. Bocharova, J. Wang and E. Katz, Analyst, 2010, 135, 2249–2259.


View Article Online

Analyst www.rsc.org/analyst


The Royal Society of Chemistry is the world's leading chemistry community. Through our high impact journals and publications we connect the world with the chemical sciences and invest the profits back into the chemistry community.

IN THIS ISSUE ISSN 0003-2654 CODEN ANALAO 139(8) 1803–2026 (2014) Cover See M. Pauly et al., pp. 1856–1867. Image reproduced by permission of M. Pauly from Analyst, 2014, 139, 1856.

Inside cover See Guilherme N. M. Ferreira et al., pp. 1847–1855. Image reproduced by permission of Guilherme N. M. Ferreira from Analyst, 2014, 139, 1847.

CRITICAL REVIEW 1814 Electroanalytical and surface plasmon resonance sensors for detection of breast cancer and Alzheimer's disease biomarkers in cells and body fluids Minghui Yang, Xinyao Yi, Jianxiu Wang* and Feimeng Zhou* Cancer and neurological disorders are two leading causes of human death.

COMMUNICATIONS 1826 Quantized double layer charging of Au130(SR)50 nanomolecules Vijay Reddy Jupally, Jacob G. Thrasher and Amala Dass* Quantized double layer (QDL) charging of a Au130(SR)50 nanomolecule is reported for the first time.


Analyst, 2014, 139, 1805–1813 | 1805


View Article Online

View Article Online

COMMUNICATIONS 1830


Versatile chiral chromatography with mixed stationary phases of water-impregnated silica gel and reversed-phase packing Satsuki Takahashi and Tetsuo Okada* A mixed silica and ODS stationary phase promises versatile functional separation.

1834 Symmetric cyanovinyl-pyridinium triphenylamine: a novel fluorescent switch-on probe for an antiparallel G-quadruplex Hai Lai, Yijie Xiao, Shengyong Yan, Fangfang Tian, Cheng Zhong, Yi Liu, Xiaocheng Weng* and Xiang Zhou* We describe a probe based on a cyanovinyl pyridinium triphenylamine (CPT) derivative, which showed fluorescent switch-on properties toward an antiparallel G-quadruplex.

1839 An enzyme-based reversible CNOT logic gate realized in a flow system Fiona Moseley, Jan Hala ´mek,* Friederike Kramer, Arshak Poghossian, Michael J. Scho ¨ ning and Evgeny Katz* A logically reversible CNOT gate was realized in a fluidic system aiming at unconventional biomolecular information processing and biosensing/biomedical applications.

1843 A turn-on fluorescent aptasensor for adenosine detection based on split aptamers and graphene oxide Yunfeng Bai, Feng Feng,* Lu Zhao, Zezhong Chen, Haiyan Wang and Yali Duan A simple, sensitive and selective turn-on fluorescent aptasensor for adenosine detection was developed based on target-induced split aptamer fragment conjunction and different interactions of graphene oxide and the two states of the designed aptamer sequences.


Analyst, 2014, 139, 1805–1813 | 1807

View Article Online

PAPERS 1847


Conformational and mechanical changes of DNA upon transcription factor binding detected by a QCM and transmission line model Jorge de-Carvalho, Roge ´rio M. M. Rodrigues, Brigitte Tome ´, S´ılvia F. Henriques, Nuno P. Mira, Isabel Sa ´-Correia and Guilherme N. M. Ferreira* A novel QCM analytical methodology based on the transmission line model detects DNA bending upon specific binding of transcription factors.

1856 A hydrodynamically optimized nano-electrospray ionization source and vacuum interface M. Pauly,* M. Sroka, J. Reiss, G. Rinke, A. Albarghash, R. Vogelgesang, H. Hahne, B. Kuster, J. Sesterhenn, K. Kern and S. Rauschenbach* The gas flow is the most efficient way of controlling ion motion at ambient conditions. The hydrodynamic optimization of a capillary inlet can yield up to 100% ion transmission to vacuum.

1868 A novel fluorescent ‘off-on-off’ probe for relay recognition of Zn2+ and Cu2+ derived from N,N-bis(2-pyridylmethyl)amine Yameng Liu, Qiang Fei, Hongyan Shan, Minghui Cui, Qing Liu, Guodong Feng and Yanfu Huan* A metal-to-metal relay recognition was discovered with sequence specificity via a fluorescence ‘off-on-off’ phenomenon.

1876 Surface plasmon resonance sensing properties of a 3D nanostructure consisting of aligned nanohole and nanocone arrays Mohamadreza Najiminaini, Erden Ertorer, Bozena Kaminska, Silvia Mittler and Jeffrey J. L. Carson* Molecular surface plasmon resonance (SPR) sensing is one of the most common applications of an array of periodic nanoholes in a metal film.

1808 | Analyst, 2014, 139, 1805–1813


View Article Online

PAPERS 1883


Simultaneous targeted analysis of five active compounds in licorice by ultra-fast liquid chromatography coupled to hybrid linear-ion trap tandem mass spectrometry Weijun Kong, Jing Wen, Yinhui Yang, Feng Qiu, Ping Sheng and Meihua Yang* A UFLC-ESI-MS/MS method based on MRM-IDA-EPI mode was developed for targeted identification and quantification of active compounds in licorice.

1895 Development of a single aptamer-based surface enhanced Raman scattering method for rapid detection of multiple pesticides Shintaro Pang, Theodore P. Labuza and Lili He* A single aptamer-based SERS method is developed for the detection and discrimination of four specific pesticides (isocarbophos, omethoate, phorate, and profenofos).

1902 Analytical approaches for quantification of a Nrf2 pathway activator: overcoming bioanalytical challenges to support a toxicity study Hermes Licea Perez,* Venkatraman Junnotula, Dana Knecht, Hong Nie, Yolanda Sanchez, Jeffrey C. Boehm, Catherine Booth-Genthe, Hongxing Yan, Roderick Davis and James F. Callahan Activation of the Nrf2 stress pathway is known to play an important role in the defense mechanism against electrophilic and oxidative damage to biological macromolecules (DNA, lipids, and proteins).

1913 Ion dynamics in a trapped ion mobility spectrometer Diana Rosa Hernandez, John Daniel DeBord, Mark E. Ridgeway, Desmond A. Kaplan, Melvin A. Park and Francisco Fernandez-Lima* In the present paper, theoretical simulations and experimental observations are used to describe the ion dynamics in a trapped ion mobility spectrometer.


Analyst, 2014, 139, 1805–1813 | 1809

View Article Online

PAPERS 1922 Identification of pathogenic fungi with an optoelectronic nose


Yinan Zhang, Jon R. Askim, Wenxuan Zhong, Peter Orlean and Kenneth S. Suslick* A disposable colorimetric sensor array capable permits rapid differentiation and identification of 12 pathogenic fungi in 3 h with >98% accuracy, based on their metabolic profiles of emitted volatiles.

1929 Innovative fabrication of a Au nanoparticledecorated SiO2 mask and its activity on surfaceenhanced Raman scattering Liang-Yih Chen, Kuang-Hsuan Yang, Hsiao-Chien Chen, Yu-Chuan Liu,* Ching-Hsiang Chen and Qing-Ye Chen SEM image of SERS-active Au NPs sonoelectrochemically deposited on a SiO2 mask array with high reproducibility.

1938 Ultrasound-assisted magnetic solid-phase extraction based ionic liquid-coated Fe3O4@graphene for the determination of nitrobenzene compounds in environmental water samples Xiaoji Cao,* Lingxiao Shen, Xuemin Ye, Feifei Zhang, Jiaoyu Chen and Weimin Mo* A magnetic solid-phase extraction based on ionic liquidcoated Fe3O4-grafted graphene has been developed for the determination of nitrobenzenes in water.

1945 A two-photon “turn-on” fluorescent probe based on carbon nanodots for imaging and selective biosensing of hydrogen sulfide in live cells and tissues Anwei Zhu, Zongqian Luo, Changqin Ding, Bo Li, Shuang Zhou, Rong Wang* and Yang Tian* A “turn-on” two-photon fluorescent sensor for H2S is developed, in which C-Dot is employed as a two-photon fluorophore and AE-TPEA–Cu2+ complex is first designed as a specific receptor for H2S.

1810 | Analyst, 2014, 139, 1805–1813


View Article Online

PAPERS 1953


Novel method of screening the oxidation and reduction abilities of photocatalytic materials K. Katayama,* Y. Takeda, K. Shimaoka, K. Yoshida, R. Shimizu, T. Ishiwata, A. Nakamura, S. Kuwahara, A. Mase, T. Sugita and M. Mori A methodology for understanding the photocatalytic abilities of materials is presented. The conversion of simple organic molecules was monitored in situ in photocatalytic microreactors.

1960 Trace vapour detection at room temperature using Raman spectroscopy Alison Chou,* Babak Radi, Esa Jaatinen, Saulius Juodkazis and Peter M. Fredericks A miniaturized flow-through system consisting of a gold coated silicon substrate based on enhanced Raman spectroscopy has been used to study the detection of vapour from model explosive compounds.

1967 Methanol-induced conformation transition of gland fibroin monitored by FTIR spectroscopy and terahertz spectroscopy Chao Yan, Bin Yang* and Zhicheng Yu THz-TDS presents great potential as a complementary approach in monitoring methanol-induced conformation transition of gland fibroin.

1973 A capillary electrophoresis-based immobilized enzyme reactor using graphene oxide as a support via layer by layer electrostatic assembly Zhengri Yin, Wenwen Zhao, Miaomiao Tian, Qian Zhang, Liping Guo and Li Yang* Using graphene oxide as an enzyme support, we developed a novel CE-based microreactor via layer-by-layer electrostatic assembly, which can be used for accurate on-line analysis and characterization of peptides and proteins.


Analyst, 2014, 139, 1805–1813 | 1811

View Article Online

PAPERS 1980 A highly selective turn-on fluorescent probe for Al(III) based on coumarin and its application in vivo


Hongde Xiao, Kun Chen, Nannan Jiang, Dandan Cui, Gui Yin,* Jie Wang and Ruiyong Wang* The probe was applied in the detection of Al3+ in living HeLa cells.

1987 Binding assay for low molecular weight analytes based on reflectometry of absorbing molecules in porous substrates Milena Stephan, Corinna Kramer, Claudia Steinem and Andreas Janshoff* Small molecule sensing is of great importance in pharmaceutical research.

1993 G-quadruplex DNAzymes-induced highly selective and sensitive colorimetric sensing of free heme in rat brain Ruimin Li, Qin Jiang, Hanjun Cheng, Guoqiang Zhang, Mingming Zhen, Daiqin Chen, Jiechao Ge,* Lanqun Mao, Chunru Wang and Chunying Shu* Direct selective determination of free heme in the cerebral system is of great significance due to the crucial roles of free heme in physiological and pathological processes.

2000 Detection of theophylline utilising portable electrochemical sensors Tiancheng Wang, Edward P. Randviir and Craig E. Banks* Screen-printed carbon electrodes are utilised to determine concentrations of theophylline under physiological conditions, and display a remarkably versatile linear range.

1812 | Analyst, 2014, 139, 1805–1813


View Article Online

PAPERS


2004 Characterization of micro- and nanocapsules for self-healing anti-corrosion coatings by high-resolution SEM with coupled transmission mode and EDX V.-D. Hodoroaba,* D. Akcakayiran, D. O. Grigoriev and D. G. Shchukin Micro/nanocapsules for self-healing coatings were characterized by high-resolution SEM with coupled transmission mode and EDX.

2011 A single-walled carbon nanotube thin film-based pH-sensing microfluidic chip Cheng Ai Li, Kwi Nam Han, Xuan-Hung Pham and Gi Hun Seong* A novel microfluidic pH-sensing chip integrated with a pH-sensing SWCNT electrode and an Ag/AgCl reference electrode was developed. The SWCNT thin film acted both as an electrode and a pH-sensitive membrane.

2016 Non-competitive aptamer-based quenching resonance energy transfer assay for homogeneous growth factor quantification Kari Kopra,* Markku Syrja ¨npa ä ¨, Pekka Ha ¨nninen and Harri Ha ¨rma ¨ A direct non-competitive quenching resonance energy transfer (QRET) assay enables a universal single-label protein detection platform utilized by aptamers.


Analyst, 2014, 139, 1805–1813 | 1813

Pt- and Pd-decorated MWCNTs for vapour and gas detection at room temperature.

Kinetics of rhodopsin at room temperature measured by picosecond spectroscopy.

Trace detection of perchlorate in industrial-grade emulsion explosive with portable surface-enhanced Raman spectroscopy.

Photo-induced enhanced Raman spectroscopy for universal ultra-trace detection of explosives, pollutants and biomolecules.

Corneal storage at room temperature.

Line coupling effects in the isotropic Raman spectra of N2: a quantum calculation at room temperature.

Detection of Sphingomyelin Clusters by Raman Spectroscopy.

Soft x-ray absorption spectroscopy of metalloproteins and high-valent metal-complexes at room temperature using free-electron lasers.

Continuous temperature-dependent Raman spectroscopy of melamine and structural analog detection in milk powder.

Ultimate sensing resolution of water temperature by remote Raman spectroscopy.

Polymorph Discrimination using Low Wavenumber Raman Spectroscopy.

Visualizing cell state transition using Raman spectroscopy.

Trace aerosol detection and identification by dynamic photoacoustic spectroscopy.

Non-invasive detection of nasopharyngeal carcinoma using saliva surface-enhanced Raman spectroscopy.

Surface enhanced Raman spectroscopy detection of biomolecules using EBL fabricated nanostructured substrates.

Rolling-circle amplification detection of thrombin using surface-enhanced Raman spectroscopy with core-shell nanoparticle probe.

Separation and enhanced detection of anesthetic compounds using solid phase micro-extraction (SPME)-Raman spectroscopy.

Label-free Detection for a DNA Methylation Assay Using Raman Spectroscopy.

Rapid detection of acetamiprid in foods using surface-enhanced Raman spectroscopy (SERS).

Rapid detection of nasopharyngeal cancer using Raman spectroscopy and multivariate statistical analysis.

Label-free detection of native proteins by surface-enhanced Raman spectroscopy using iodide-modified nanoparticles.

In situ detection and identification of hair dyes using surface-enhanced Raman spectroscopy (SERS).

Rapid detection of Listeria monocytogenes in milk using confocal micro-Raman spectroscopy and chemometric analysis.

Label-free detection of serum proteins using surface-enhanced Raman spectroscopy for colorectal cancer screening.