Author’s Accepted Manuscript Extended-gate field-effect transistor (eg-fet) with molecularly imprinted polymer (mip) film for selective inosine determination Zofia Iskierko, Marta Sosnowska, Piyush Sindhu Sharma, Tiziana Benincori, Francis D’Souza, Izabela Kaminska, Krzysztof Fronc, Krzysztof Noworyta

PII: DOI: Reference:

www.elsevier.com/locate/bios

S0956-5663(15)30240-2 http://dx.doi.org/10.1016/j.bios.2015.06.073 BIOS7810

To appear in: Biosensors and Bioelectronic Received date: 17 April 2015 Revised date: 12 June 2015 Accepted date: 27 June 2015 Cite this article as: Zofia Iskierko, Marta Sosnowska, Piyush Sindhu Sharma, Tiziana Benincori, Francis D’Souza, Izabela Kaminska, Krzysztof Fronc and Krzysztof Noworyta, Extended-gate field-effect transistor (eg-fet) with molecularly imprinted polymer (mip) film for selective inosine determination, Biosensors and Bioelectronic, http://dx.doi.org/10.1016/j.bios.2015.06.073 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Corrected manuscript: Biosensors and Bioelectronics

Extended-gate field-effect transistor (EG-FET) with molecularly imprinted polymer (MIP) film for selective inosine determination Zofia Iskierko,1 Marta Sosnowska,1 Piyush Sindhu Sharma,1 Tiziana Benincori,2 Francis D’Souza,3 Izabela Kaminska,4 Krzysztof Fronc,4 Krzysztof Noworyta1,* 1

Institute of Physical Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, 01-224 Warsaw, Poland

2

Dipartimento di Scienza ed Alta Tecnologia, Università degli Studi dell'Insubria, Via Valleggio, 11-22100 Como, Italy 3

Department of Chemistry, University of North Texas, Denton, TX 76203-5017, USA

4

Institute of Physics, Polish Academy of Sciences, 32/46 Al. Lotników, 02-668 Warsaw, Poland

*Corresponding author phone: +48 22 343 3217 and e-mail: [email protected]

Abstract A novel recognition unit of chemical sensor for selective determination of the inosine, renal disfunction biomarker, was devised and prepared.

For that purpose, inosine-templated

molecularly imprinted polymer (MIP) film was deposited on an extended-gate field-effect transistor (EG-FET) signal transducing unit. The MIP film was prepared by electrochemical polymerization of bis(bithiophene) derivatives bearing cytosine and boronic acid substituents, in the presence of the inosine template and a thiophene cross-linker.

After MIP film

deposition, the template was removed, and was confirmed by UV-visible spectroscopy. Subsequently, the film composition was characterized by spectroscopic techniques, and its morphology and thickness were determined by AFM. The finally MIP film-coated extendedgate field-effect transistor (EG-FET) was used for signal transduction. This combination is not widely studied in the literature, despite the fact that it allows for facile integration of electrodeposited MIP film with FET transducer. The linear dynamic concentration range of the chemosensor was 0.5 to 50 M with inosine detectability of 0.62 M. The obtained detectability compares well to the levels of the inosine in body fluids which are in the range 0 to 2.9 µM for patients with diagnosed diabetic nephropathy, gout or hyperuricemia, and can reach 25 µM in certain cases. The imprinting

factor for inosine, determined from piezomicrogravimetric experiments with use of the MIP film-coated quartz crystal resonator, was found to be 5.5. Higher selectivity for inosine with respect to common interferents was also achieved with the present molecularly engineered sensing element. The obtained analytical parameters of the devised chemosensor allow for its use for practical sample measurements.

Keywords:

extended-gate

field-effect

transistor,

molecularly

imprinted

polymer,

chemosensor, inosine, piezomicrogravimetry.

1.

Introduction

For several decades now, ion-sensitive field-effect transistors (ISFETs) have been applied for chemosensing (Janata 2004). These ISFETs have been devised using metal oxide field-effect transistors (MOSFETs). After the first successful application of ISFETs (Bergveld 1970), various reports have described similar devices for determination of different analytes of interest (Jimenez-Jorquera et al. 2010; Lee et al. 2009). As an improvement to isolate FET from the chemical environment, an extended-gate field-effect transistor (EG-FET) was devised (Batista et al. 2006; Chen et al. 2011; Chi et al. 2000 ; Yin et al. 2000). In this transistor, the recognition or chemically sensitive unit was deposited on surface of the gate extending from FET. The experimental setup prepared that way showed major advantage of flexibility in the gate shape (Chi et al. 2000 ). Advantageously, without a need of the use of expensive instruments and reagents, minute changes in potential at the gate surface due to the presence of charged biomolecules/analytes, it was possible to convert into detectable electric signals. Additionally, the stability of FET characteristics to the ambient environment is greatly improved, and more importantly, packing and transportation of such a setup for field applications is rather easy. In various applications sensitivity provided by the EG-FET transduction was appreciable (Casalini et al. 2013 ; Chen et al. 2011; Selvanayagam et al. 2002). Typically, however, selectivity was poor. Therefore, there is still a need to improve selectivity in this device. For selective determination of the analyte of interest in the presence of interferences nowadays a synthetic receptor recognition unit is often used. One of the well established procedures for preparation of synthetic receptors is molecular imprinting (Haupt and Mosbach 2000; Malitesta et al. 2012; Sharma et al. 2012a; Sharma et al. 2012b). This procedure consists in impressing molecular cavities in a polymer matrix with template molecules. The

cavity shape and size as well as orientation of recognition sites generated in these cavities correspond to the binding sites of the template molecule. The analyte itself, or its close analogue, is selected as a template. These synthetic receptors showed selectivity very similar to recognition units based on biological molecules (Malitesta et al. 2012; Sharma et al. 2012b). Inosine (Scheme 1a) is a purine nucleoside composed of hypoxanthine and D-ribose. It is a major degradation product of adenosine with potential immuno-modulatory and neuroprotective effects. It is used as a drug to relieve symptoms of many diseases (Hasko et al. 2004). Moreover, it has been identified as a potential early-warning biomarker of renal disfunction (Xia et al. 2009), as well as of gout and asymptomatic hyperuricemia (Zhao et al. 2005). Its increased level have also been observed in critically ill patients with sepsis (Grum et al. 1985). Together with its nucleotides, inosine plays an important role in human body. It correlates with the sclerosis symptoms, on the one hand (Amorini et al. 2009), and it protects organisms against inflammation (Buckley et al. 2005; Schneider and Klein 2005), on the other. Furthermore, inosine is one of the biomarkers in diabetic nephrophaty (Xia et al. 2009). It is estimated that the possibility of death due to renal disease is 17 times higher in diabetics than in nondiabetics (Susztak and Bottinger 2006).

In addition, diabetic

nephrophathy is associated with considerably increased risk and mortality of cardiovascular disease (Xia et al. 2009). The inosine concentration in blood for healthy patients is reported to be in the range from 0 to 0.75 µM (Xia et al. 2009; Zhao et al. 2005). On the other hand, inosine level in blood for patients with developed diabetic nephropathy was as high as 2.4 µM (Xia et al. 2009). Also for patients with diagnosed gout or asymptomatic hyperuricemia the blood concentration of this compound reached 2.4 – 2.9 µM (Zhao et al. 2005). Furthermore, urinary inosine level can reach level of 25 µM for person after physical exercise(Stathis et al. 2005). Considering the above issues, development of a rapid and efficient diagnostic tool for early inosine determination in humans is important from the clinical analysis point of view. Up to now, inosine is determined using mainly flow analytical techniques, such as capillary electrophoresis (Kong et al. 2003; Terzuoli et al. 1999), or high performance liquid chromatography (HPLC) (Mei et al. 1996; Zhao et al. 2005), or techniques using enzymatic reactions (Park and Kim 1999; Watanabe et al. 1986; Watanabe et al. 1984; Yao 1993). Enzymatic reaction systems combined with flow-injection analysis (FIA) (Park and Kim 1999) and systems with enzymes immobilized directly on electrodes (Watanabe et al. 1984; Yao 1993) or membranes (Watanabe et al. 1986) are used for this purpose.

However,

chromatographic techniques are time-consuming, expensive, and a qualified personnel is needed to operate them. Reproductibility of techniques based on enzymatic reactions is usually low and, moreover, their sensitivity is not satisfactory for real biological systems. Therefore, many other techniques for selective inosine determination have been developed. These techniques are mainly based on oxidation of inosine and their nucleotides on different carbon electrodes, such as ultramicroelectrodes with carbon fibres,(Cavalheiro et al. 2000) carbon paste electrodes modified with La(OH)3 nanowires (Liu et al. 2006), pyrolytic carbon electrodes coated with one-dimensional carbon nanorods (Goyal et al. 2008), or glassy carbon electrodes modified with 3-amine-5-mercapto-1,2,4-triazole (Revin and John 2012). These electrochemical techniques have sufficient sensitivities, allowing determination of inosine in biological systems (Revin and John 2012). However, their drawback is instability caused by chemical reactions of oxidation products of purine nucleobasis on electrode surfaces (Oliveira-Brett et al. 2003). Additionally, these systems are usually not selective, and they require application of a relatively high potential. Although, various reports described the application of FET based chemosensors, combination of EG-FET transduction and MIP based recognition unit is still not well studied. Moreover, this combination offers numerous advantages. Firstly, the selectivity of MIP together with amplification offered by the FET leads to development of highly sensitive chemosensors selective toward chosen analytes. What is more, the EG-FET concept offers much easier way to integrate MIP recognition films than classical FETs as it does not require difficult and costly processing methods. The EG-FET can be easily adapted to work in both organic and water solutions, which is important from the point of view of sensor fabrication and practical application.

Moreover, the EG-FETs offer an excellent possibility of

chemosensor miniaturization. Considering these points, we report here on novel chemosensor capable of selective sensing of this analyte. In this chemosensor, thin inosine molecularly imprinted polymer (MIP) film and the EG-FET plays the role of a recognition and signal transduction unit, respectively. The chemical recognition unit introduces so much desired selectivity into the sensor, whereas the EG-FET provides the sensitivity of the integrated chemosensor device. The MIP was deposited in the form of a thin film on an Au-coated extended gate surface by potentiodynamic electropolymerization of functional monomers, vis. the bithiophene derivative bearing the boronic acid substituent (7) and bis(bithiophene) derivatized with cytosine (8) (Huynha et al. 2015) in the presence of the inosine template (1) and the 2,4,5,2’,4’,5’-hexa(thiophen-2-yl)-3,3’-bithiophene (6) cross-linker.(Sannicolo and Benincori

2015). Although, there is one literature account on preparation of the MIP in a form of nanospheres for inosine extraction, and then its slow release (Kusunoki and Kobayashi 2010), there are no attempts of devising an MIP recognition unit selective to inosine.

2.

Materials and Methods

2.1

Reagents and chemicals

Inosine 1, adenosine 2, guanosine 3, glucose 4, thymine 5, and the 2,2’-bithiophene-5-boronic acid 7 functional monomer, as well as acetonitrile were purchased from Sigma-Aldrich. The tetra-n-butylammonium perchlorate [(TBA)ClO4] supporting electrolyte, was supplied by Fluka. Hydrochloric acid (HCl), isopropanol, and methanol were from CHEMPUR. The 2(cytosin-1-yl)ethyl p-bis(2,2’-bithien-5-yl)methylbenzolate 8 functional monomer, was prepared according to the procedure described in Supporting Information and in literature (Huynha et al. 2015). The 2,4,5,2’,4’,5’-hexa(thiophen-2-yl)-3,3`-bithiophene 6 cross-linking monomer was synthesized in the University of Milan (Sannicolo and Benincori 2015).

2.2

Instrumentation

An AUTOLAB computerized electrochemistry system of Eco Chemie (Utrecht, The Nederlands), equipped with expansions cards of the PGSTAT 12 potentiostat and the FRA2 frequency response analyzer and controlled by the GPES 4.9 software of the same manufacturer, was used for the potentiodynamic deposition of thin MIP-inosine films. The UV–visible spectra were recorded with 0.1-nm resolution by using a UV 2501-PC recording spectrophotometer of Shimadzu Corp (Tokyo, Japan). Atomic force microscopy (AFM) imaging with TappingTM mode was performed with the use of a Multimode 8 microscope under control of the Nanoscope V controller, of Bruker. For this imaging, the films were deposited on glass slides sputtered with thin Au layers on Ti underlayers. A model EQCM 5610 and EQCM 5710 quartz crystal microbalance, controlled by the EQCM 5710-S2 software, all of the Institute of Physical Chemistry (Warsaw, Poland), were used to perform electropolymerization and the piezomicrogravimmetric (PM) experiments under FIA conditions, respectively. The resonant frequency change was measured with 1-Hz resolution using a 14-mm diameter, AT-cut, plano-plano quartz crystal resonators (Au-QCRs) of 10-MHz resonant frequency with 5-mm diameter and 100-nm thick Au film electrodes over a Ti underlayer film on both sides. However, only one Au-QCR side was wetted by the

working solution and the Au film electrode of this side was used both as the working electrode and the substrate of the MIP film. The resonators were cleaned for 5 min with methanol before electropolymerization. A Keithley 2636A dual-channel system source meter instrument, (1 fA, 10 A pulse). (Keithley Instruments, Inc., Cleveland, Ohio, USA) and a CD4007UB MOSFET system was applied for measurements of the EG-FET characteristics during inosine additions. For all the measurements, Au working electrode coated with the MIP film and the reference electrode were immersed in the aqueous solution.

The working electrode was connected to the

MOSFET gate (Scheme S2). The Pt grid was used as the pseudo-reference electrode and connected to the source meter. The distance between the working and reference electrode was fixed at 10 mm. Source and drain of the commercial MOSFET were electrically connected to the second channel of the source-meter allowing for changes of drain voltage and measurements of drain current. Polarization-modulated IR reflection-absorption spectroscopy (PM-IRRAS) spectra were recorded using a Vertex80v IR spectrophotometer with the PMA50 accessory equipped with the liquid nitrogen-cooled HgCdTe (MCT) detector of Bruker. All quantum-chemical calculations were performed on the four-dual processor workstation of Intel with the Gaussian 09 software installed (Gaussian, Inc., CT, USA). Structures of the monomers and the pre-polymerization complexes were optimized using the density functional theory (DFT) with the B3LYP functional and 3-21G* basis set. 3.

Results and discussion

3.1

Quantum-chemical modeling of the pre-polymerization complex structure

Structure of the pre-polymerization complex (Scheme 1b) of the template 1 with the functional monomers 6 and 7 was optimized (Scheme 1c). Thermodynamic parameters were defined according to the density functional theory (DFT) with the B3LYP/6-31G* functional and basis set, respectively. Results of the calculations proved that stability of the complex of the 1 : 1 : 1 stoichiometry of 1 : 7 : 8, was highest. The Gibbs free energy calculated for this complex was G = -135.6 kJ mol-1, indicating formation of a stable pre-polymerization complex. The attempts to optimize complex with stoichiometry 1 : 2 : 1 resulted in complexes with lower G. Therefore, the former stoichiometry was chosen for MIP preparation. The G calculated separately for complexes of 1 and 8, and 1 and 7 was -87.7 kJ mol-1 and -54.3 kJ mol-1 respectively. Therefore, interaction between inosine and 8 is stronger than for inosine and 7.

Moreover, molecular modeling showed multiple interactions within the complex. On the one side, the calculation results show formation of hydrogen bonds between vicinal diols of the ribose moiety of inosine (atoms O1 and O2) and oxygen atoms O3 and O4 of boronic acid moiety of 7. On the other side, the cytosine moiety of 8 forms another pair of hydrogen bonds between atom N1 from -NH2 group and N2 from heteroaromatic ring of cytosine and nitrogen atom N3 an oxygen atom O5 of inosine. These interactions decided on selectivity of the molecular cavities then imprinted. 3.2

Deposition of the thin MIP-inosine films on different electrode surfaces

Thin MIP-inosine films were prepared by electrochemical polymerization under potentiodynamic conditions over the potential range of 0.50 to 1.40 V vs. Ag pseudoreference electrode, at the potential scan rate of 50 mV/s. The acetonitrile solution of inosine and functional monomers 7 and 8 was used for the electropolymerization. The template to functional monomers molar ratio was 1 : 1 : 1, as suggested by the quantum-chemical calculations. Concentrations of inosine and both functional monomers were 0.1 mM. To provide accessible molecular cavities in the resulting MIP, an excess of the cross-linking monomer 6 was used (the template-to-cross-linker molar ratio was 1 : 4). Conductivity was provided by the 0.1 M tetra-n-bytylammonium perchlorate, (TBA)ClO4 supporting electrolyte. Thickness of the MIP-inosine film, deposited on the Au-QCR (Figure 1), or on the Au-coated glass slides (data not shown), was controlled by the number of the potential cycles applied. The growth of the MIP film was manifested by the simultaneously recorded corresponding frequency decrease (Figure 1b). Noticeably, the anodic peak at ∼1.30 V (Figure 1a) indicated electropolymerization and growth of the conducting polymer film, as evidenced by the frequency decrease (Figure 1b). In the subsequent cycle, decrease at ∼1.0 V, followed by the increase in the dynamic resistance change with the anodic potential increase (Figure 1c) at around 1.20 V, are most plausibly due to the egress of cations and ingress of anions of the supporting electrolyte, substantiating the rigidity variations of the MIP film. Interestingly, growth in current was observed with each potentiodynamic cycle (Figure 1a), which indicates that the growing polymer is conducting. Non-imprinted polymer (NIP) control film was deposited from the template-free solution using the same electropolymerization procedure. Total frequency change observed during five deposition cycles reaches -8.8 kHz for MIP and -8.9 kHz for NIP. As the changes of dynamic resistance are negligible, one can calculate the total mass of the polymer deposited on the electrode from the Saurebrey

equation.(Tsionsky et al. 2004) The mass reaches 11.3 µg for the MIP and 11,5 µg for the NIP. Taking into account estimated polythiophene density of 1.3 g cm-3 one can assess thickness of the deposited film to be 444 nm and 449 nm, for MIP and NIP respectively. 3.3

Extraction of the inosine template

The inosine template was extracted from the MIP-inosine film before determination of inosine. For that, liquid-solid extraction with 0.1 M HCl at 35oC was performed. The completeness of the extraction was confirmed by the UV-vis spectroscopy measurement (Supporting Information Fig. S1). Repeated extraction of MIP film with 0.1 M HCl resulted in complete removal of inosine from its molecular cavities (Fig. S1, curves 1, 2 and 3). The inosine removal from the film was also supported by observation of PM-IRRAS spectra of the Inosine-MIP, extracted inosine-MIP and NIP (Supporting Information Fig. S2). The vibration band characteristic for -C-O- vibration of ribose was observed at 1150 cm-1. This band substantially decreases after extraction of MIP and is absent in NIP spectrum. 3.4

Characterization of the MIP-inosine film by AFM

For surface characterization of MIP-inosine film, AFM imaging was performed (Figure 2). Apparently, the film was relatively rough, and being composed of small clusters 20 to 60 nm in diameter. The film thickness was 209  5 nm and its roughness, Ra, was 2 nm. For comparison, the NIP film was AFM imaged. Apparently, thickness of this film was 334  7 nm, and its average roughness, Ra, was 3 nm. Apparently, the NIP film was thicker and slightly rougher than the MIP film. The measured MIP and NIP film thickness is lower than that estimated from EQCM measurements during films deposition. This indicates that electrochemically deposited films are denser than expected on the basis of the density of pristine polythiophene.

3.5

Determination of inosine at the MIP-inosine film-coated EG-FET

The EG-FET sensing system is composed of two parts. The sensing part was made of a thin MIP-inosine film with surface area of 21 mm2, deposited on the Au-glass slide (100 nm Au on 15 nm Ti). This part was electrically connected to the gate of a commercial MOSFET device (Scheme S2 in supporting information). From the point of view of electronics the MIP film coated Au-glass slide is in fact simple extension of the classical field effect transistor gate. Therefore, all changes of potential at the Au-glass slide are transferred to the transistor gate leading to changes of source-drain current, exactly as if the potential changes would take

place on the gate itself. In order to drive transistor gate open or closed there is also need to apply certain voltage across the solution/polymer/Au-glass interfaces. For that purpose the additional electrode is also immersed into the studied solution, which is subsequently polarized by using the source meter allowing opening, or closing, of the transistor gate (Scheme S2). At constant gate voltage, the drain voltage voltage (Vd) is changed in time and the resulting drain current (Id) is recorded. When sorption of analyte is taking place in the MIP film-coated Au-glass slide it modifies effective gate potential leading to change in drain current value which is proportional to amount of the analyte sorbed in the film. Therefore, the operation of the EG-FET is very similar to that of a conventional MOSFET, except that an additional sensing structure in the form of an extended gate was immersed in the aqueous test solution. Figure 3 shows the Id vs. Vd characteristic of the devised EG-FET. The saturation region current can be expressed as

where µo is the electron mobility in the channel, λ is the length modulation factor of the channel, Cox is the oxide capacity per unit area, W/L is the width-to-length ratio, Vref and Vds is the voltage applied to the reference electrode and the drain-source voltage, respectively. As expected the measured current was higher the higher was the gate voltage applied to the reference electrode in the range of 1.5 to 3.0 V (Fig. 3a). Importantly, addition of inosine to the solution at constant gate voltage led to a pronounced decrease of Id,max (Fig. 3b). Inosine binding to the imprinted cavities of the MIP film caused a surface potential change of the MIP recognition film. This change resulted in the decrease of the effective voltage applied to the gate. In effect, the current flowing through the channel decreased because of the decrease of the electron density in the gate region at the enhancement-mode n-MOSFET. This measured change in the drain current was used as the analytical signal for inosine determination. For optimizing the gate voltage at which the drain current was the highest for a given inosine concentration different gate voltages were applied (Fig. 4a). As expected, higher gate voltage resulted in higher Id,

max

(Fig. 4a). Moreover, sensitivity of the chemosensor was

higher, the higher was gate voltage (Table 1). However, the apparent imprinting factor, which is the primary criterion to the prove imprinting, was drastically lower the lower was the gate voltage (Table 1). Apparently, the higher gate voltage promoted non-specific binding of the analyte in the polymer film. Furthermore, the detection limit of the chemosensor increased at

the gate voltages above 2.5 V, indicating an increased noise. Therefore, relatively low gate voltage of 1.5 V was selected as the optimum voltage for further measurements.

Table 1. Analytical parameters of the EG-FET MIP-inosine chemosensor for different gate voltages. Analytical parameters

Gate voltage , V 2.0 V 2.5 V

1.5 V Limit of detection, LOD, (M) at S/N = 3 Sensitivity, MIP (A·M-1) Sensitivity, NIP (A·M-1) Apparent imprinting factor (AIF)

3.0 V

(0.620.01)

(0.110.04)

(1.850.17)

(3.440.10)

(2.30.2) × 10-3 (81.2) × 10-5 (29.01.8)

(0.320.03) (0.0600.006) (5.350.04)

(0.990.08) (0.2610.026) (3.800.07)

(1.640.15) (0.5120.051) (3.200.03)

Figure S3 presents the time dependence of the maximal drain current for different inosine concentrations determined with the use of the MIP-inosine film or the NIP film based EG-FET chemosensors under the batch conditions. Both chemosensors exhibited a similar decrease of the drain current after inosine addition. However, the response of the NIP filmcoated chemosensor was markedly lower. As expected, sensitivity of the MIP-film coated EG-FET was markedly higher (Table 1) than that of the NIP film to inosine because of selective pre-concentration of inosine in the imprinted molecular cavities.

Therefore,

performance of the MIP chemosensor was excellent over a concentration range from 0.5 to 50 µM with detectability and sensitivity sufficient to work in biological conditions (Table 1). The biological levels of inosine are known to be in the range 0 to 2.9 µM for patients with diagnosed diabetic nephropathy, gout or hyperuricemia, and 25 µM for person after physical exercises. Advantageously, the developed chemosensor was highly selective with respect to inosine (Fig. 4b). That is, sensitivity to inosine was 90 times that to glucose, and over 100 times that to thymine, guanosine or adenosine.

Worth mentioning, the adenosine and

guanosine functionality is very similar to that of inosine, and glucose is a common interference present in biological samples.

This high selectivity confirms formation of

selective molecular cavities in the MIP film. These results prove suitability of the devised chemosensor for selective determination of inosine. 3.6

Piezomicrogravimetric determination of inosine

The above results confirmed that the MIP-film coated EG-FET can be applied as an analytical tool to monitor concentration of the inosine analyte. In order to get deeper insight into

binding of the MIP film, the EG-FET chemosensor results were compared with those obtained with an alternative piezoelectromicrogravimetry (PM) transduction. For preparation of the recognition film for the PM chemosensor, the same electropolymerization conditions were used as those described above for the MIP and NIP film deposition on the extended gate of the FET. To this end, Au-QCRs were used. The experiments were recorded under FIA conditions (not shown). After injection of inosine to the carrier solution the resonant frequency decreased. This behavior confirms inosine binding by MIP. The resonant frequency nearly returned to its original value as the carrier solution eluted the inosine analyte from the MIP film. reversibly.

Advantageously, the analyte was bound

The calibration curve for inosine was constructed by injecting the analyte

solutions of different concentrations. The measured frequency decrease was linear with the analyte concentration up to at least 50 mM and its sensitivity was -1.060.04 Hz mM-1 and R2 = 0.9919 (Fig 5). Similarly, the resonant frequency changes, due to the analyte binding by the NIP film under the FIA conditions, were also used to construct the calibration plots (curve 2 in Fig. 5). From the slope of the plot (-0.210.07 Hz mM-1), the sensitivity to inosine for the NIP film was determined. From the ratio of the sensitivity of the MIP-inosine (curve 1 in Fig. 5) and that of NIP (curve 2 in Fig. 5) to inosine, the imprinting factor was determined to be 4.90.5. It is important to note that piezoelectromicrogravimetry measurements are directly proportional to mass changes related to the analyte binding, while EG-FET measurements are affected also by the charge of sorbed species. Integration of the MIP film with the PM transducer for determination of inosine proved suitability of the devised MIP for integration with different transduction platforms. However, higher detectability was obtained for the EG-FET (0.620.01 µM) than that for the PM chemosensor (0.850.15 mM). The higher detectability of the EG-FET chemosensor may originate from its sensitivity to charge changes rather than to the mass changes. Furthermore, the EG-FET sensor allows for signal amplification via the gate voltage adjustment. The higher detectability of the EG-FET chemosensor and its ease of use led us to conclude that the EGFET chemosensor is better suited for practical purposes. As it has been mentioned, inosine determination in body fluids requires higher detectability,(Stathis et al. 2005) than that obtained for PM chemosensor.

4.

Conclusions

We have successfully devised, fabricated, and tested selective EG-FET chemosensors for inosine determination. The chemosensor recognition film was prepared via deposition of the MIP film by potentiodynamic electropolymerization on the extended gate of the field-effect transistor. Composition of the most stable pre-polymerization complex was determined using basis of quantum-chemical calculations. Subsequently, well-adhering, rigid films of MIP and NIP were deposited on the EG-FET and PM signal transducing units. The detectability for the EG-FET was higher than for the PM chemosensor. EG-FET chemosensor limit of detection (LOD) reached 0.620.01 µM, which is suitable for studies of the biological samples, where level of inosine is reported to be from 0 to 2.9 µM for patients with diagnosed diabetic nephropathy, gout or hyperuricemia, and 25 µM for person after physical exercises (Stathis et al. 2005).

The imprinting factor determined by the PM

measurements was 4.9 while practical sensitivity ratio of the MIP to NIP coated EG-FET chemosensor was as high as 29. Furthermore, the MIP film-coated EG-FET chemosensor allowed for advantageous flexibility during measurements through gate voltage adjustments. This procedure allowed for obtaining either increased detectability or selectivity, depending on requirments.

Importantly, selectivity of the developed MIP film-coated EG-FET

chemosensor was very high with respect to inosine structural analogues and interferences, including thymine, adenosine, guanosine and glucose. Acknowledgements Special thanks are due to the National Science Center of Poland for financial support through the project OPUS NCN 2011/01/B/ST5/03796 to K.N. and Z.I., project OPUS DEC2012/07/B/ST5/02080 (to IK and ZF), the US-National Science Foundation (grant no. 140188 to F.D.) and the Fondazione Cariplo (grant No 2001-0417 to T.B.). References Amorini, A.M., Petzold, A., Tavazzi, B., Eikelenboom, J., Keir, G., Belli, A., Giovannoni, G., Pietro, V.D., Polman, C., D'Urso, S., Vagnozzi, R., Uitdehaa, B., Lazzarino, G., 2009. Increase of uric acid and purine compounds in biological fluids of multiple sclerosis patients. Clin. Biochem. 42, 1001-1006. Batista, P.D., Mulato, M., Graeff, C.F.d.O., Fernandez, F.J.R., Marques, F.d.C., 2006. SnO2 extended gate field-effect transistor as pH sensor. Braz. J. Phys. 36, 478-481. Bergveld, P., 1970. Development of an ion-sensitive solid-state device for neurophysiological measurements. IEEE Trans. Biomed. Eng. 17, 70-71. Buckley, S., Barsky, L., Weinberg, K., Warburton, D., 2005. In vivo inosine protects alveolar epithelial type 2 cells against hyperoxia-induced DNA damage through MAP kinase signaling. Am. J. Physiol. Lung Cell. Mol. Physiol. 288, L569-L575.

Casalini, S., Leonardi, F., Cramer, T., Biscarini, F., 2013 Organic field-effect transistor for label-free dopamine sensing. Org. Electron. 14 156-163. Cavalheiro, E.T.G., El-Noura, K.A., Brajter-Toth, A., 2000. Possibilities of the use of fast scan voltammetry in simultaneous determination of purines at carbon fiber ultramicroelectrodes. J. Braz. Chem. Soc. 11, 512-515. Chen, C.-P., Ganguly, A., Lu, C.-Y., Chen, T.-Y., Kuo, C.-C., Chen, R.-S., Tu, W.-H., Fischer, W.B., Chen, K.-H., Chen, L.-C., 2011. Ultrasensitive in Situ Label-Free DNA Detection Using a GaN Nanowire-Based Extended-Gate Field-Effect-Transistor Sensor. Anal. Chem. 83, 1938–1943. Chi, L.-L., Chou, J.-C., Chung, W.-Y., Sun, T.-P., Hsiung, S.-K., 2000 Study on extended gate field effect transistor with tin oxide sensing membrane. Mater. Chem. Phys. 63, 19–23. Goyal, R.N., Gupta, V.K., Chatterjee, S., 2008. Simultaneous determination of adenosine and inosine using single-wall carbon nanotubes modified pyrolytic graphite electrode. Talanta 76, 662-668. Grum, C.M., Simon, R.H., Dantzker, D.R., Fox, I.H., 1985. Evidence for adenosine triphosphate degradation in critically-ill patients. Chest J. 88(5). Hasko, G., Sitkovsky, M.V., Szabo, C., 2004. Immunomodulatory and neuroprotective effects of inosine. Trends Pharmacol. Sci. 25, 152-157. Haupt, K., Mosbach, K., 2000. Molecularly imprinted polymers and their use in biomimetic sensors. Chem. Rev. 100, 2495-2504. Huynha, T.-P., Wojnarowicz, A., Sosnowska, M., Srebnik, S., Benincorid, T., Sannicolò, F., D'Souza, F., Kutner, W., 2015. Cytosine derivatized bis(2,2′-bithienyl)methane molecularly imprinted polymer for selective recognition of 6-thioguanine, an antitumor drug. Biosens. Bioelectronics 70, 153–160. Janata, J., 2004. Thirty Years of CHEMFETs – A Personal View. Electroanalysis 16, 18311835. Jimenez-Jorquera, C., Orozco, J., Baldi, A., 2010. ISFET based microsensors for environmental monitoring. Sensors 10, 61-83. Kong, Y., Zheng, N., Zhang, Z., Gao, R., 2003. High-performance capillary zone electrophoretic assay for markers of diabetic nephropathy in plasma and urine. J. Chromatogr. A 987, 477-483. Kusunoki, T., Kobayashi, T., 2010. Molecular Imprinting Micropolymerbeads Having Cooperative Effect of both Surfactant and Inosine Template. J. Appl. Polym. Sci. 117, 565– 571. Lee, C.-S., Kim, S.K., Kim, M., 2009. Ion-sensitive field-effect transistor for biological sensing. Sensors 9, 7111-7131. Liu, L., Song, J.-f., Yu, P.-f., Cui, B., 2006. A novel electrochemical sensing system for inosine and its application for inosine determination in pharmaceuticals and human serum. Electrochem. Commun. 8, 1521. Malitesta, C., Mazzotta, E., Picca, R.A., Poma, A., Chianella, I., Piletsky, S.A., 2012. MIP sensors – the electrochemical approach. Anal. Bioanal. Chem. 402, 1827-1846. Mei, D.A., Gross, G.J., Nithipatikom, K., 1996. Simultaneous determination of adenosine, inosine, hypoxanthine, xanthine, and uric acid in microdialysis samples using microbore column high-performance liquid chromatography with a diode array detector. Anal. Biochem. 238, 34-39. Oliveira-Brett, A.M., Silva, L.s.A., Farace, G., Vadgama, P., Brett, C.M.A., 2003. Voltammetric and impedance studies of inosine-5'-monophosphate and hypoxanthine. Bioelectrochemistry 59, 49-56.

Park, I.-S., Kim, N., 1999. Simultaneous determination of hypoxanthine, inosine and inosine 5'-monophosphate with serially connected three enzyme reactors. Anal. Chim. Acta 394, 201210. Revin, S.B., John, S.A., 2012. Selective determination of inosine in the presence of uric acid and hypoxanthine using modified electrode. Anal. Biochem. 421, 278-284. Sannicolo, F., Benincori, T., 2015. to be published. to be published. Schneider, S., Klein, H.H., 2005. Inosine improves islet xenograft survival in immuncompetent diabetic mice. Eur. J. Med. Res. 10, 283-286. Selvanayagam, Z.E., Neuzil, P., Gopalakrishnakone, P., Sridhar, U., Singh, M., Ho, L.C., 2002. An ISFET-based immunosensor for the detection of -Bungarotoxin. Biosens. Bioelectron. 17, 821-826. Sharma, P.S., D’Souza, F., Kutner, W., 2012a. Molecular imprinting for selective chemical sensing of hazardous compounds and drugs of abuse. TrAC-Trends Anal. Chem. 34, 59-77. Sharma, P.S., Pietrzyk-Le, A., D’Souza, F., Kutner, W., 2012b. Electrochemically synthesized polymers in molecular imprinting for chemical sensing. Anal. Bioanal. Chem. 402, 3177-3204. Stathis, C.G., Carey, M.F., Snow, R.J., 2005. The influence of allopurinol on urinary purine loss after repeated sprint exercise in man. Metab. Clin. Exp. 54, 1269-1275. Susztak, K., Bottinger, E.P., 2006. Diabetic nephropathy: A frontier for personalized medicine katalin. J. Am. Soc. Nephrol. 17, 361-367. Terzuoli, L., Porcelli, B., Setacci, C., Giubbolini, M., Cinci, G., Carlucci, F., Pagani, R., Marinello, E., 1999. Comparative determination of purine compounds in carotid plaque by capillary zone electrophoresis and high-performance liquid chromatography. J. Chromatogr. B 728, 185-192. Tsionsky, V., Daikhin, L., Urbakh, M., Gileadi, E., 2004. Looking at the Metal/Solution Interface with the Electrochemical Quartz Crystal Microbalance: Theory and Experiment. Marcel Dekker, New York. Watanabe, E., Endo, H., Hayashi, T., Toyama, K., 1986. Simultaneous determination of hypoxanthine and inosine with an enzyme sensor. Biosensors 2, 235-244. Watanabe, E., Tokimatsu, S., Toyama, K., 1984. Simultaneous determination of hypoxanthine, inosine, inosine-5′-phosphate and adenosine-5′-phosphate with a multielectrode enzyme sensor. Anal. Chim. Acta 164, 139-146. Xia, J.-F., Liang, Q.-L., Liang, X.-P., Yi-MingWangb, P.H., Li, P., Luo, G.-A., 2009. Ultraviolet and tandem mass spectrometry for simultaneous quantification of 21 pivotal metabolites in plasma from patients with diabetic nephropathy. J. Chromatogr. B 877, 1930– 1936. Yao, T., 1993. Enzyme electrode for the successive detection of hypoxanthine and inosine. Anal. Chim. Acta 281, 323-326. Yin, L.-T., Chou, J.-C., Chung, W.-Y., Sun, T.-P., Hsiung, S.-K., 2000. Separate structure extended gate H+-ion sensitive field effect transistor on a glass substrate. Sens. Actuators B 71, 106-111. Zhao, J.Y., Liang, Q., Luo, G., Wang, Y., Zuo, Y., Jiang, M., Yu, G., Zhang, T., 2005. Purine metabolites in gout and asymptomatic hyperuricemia: Analysis by HPLC–electrospray tandem mass spectrometry. Clin. Chem. 51, 1742-1744.

a

H2N

H N

N

N O

O

N

N O

NH

HO

NH2

N

N

N

HO

O

O

H3C

O OH

O

HO

H N

N

N

N

OH

HO

N

HO HO

OH

HO

OH

HO

2

1

H O

H

OH

3

4

OH

5

NH2 N O N

O O HO S

S

S

B OH

S

S S

S

S

S

S

S

S

S

S

6

7

8

S

b

S

HO

1

HO

S

O

OH N

HO

7

S N

OH

S

S

N

B

N H

O

H N

O H

O N

O

8

N H

c N1 O5

N3

N2

O1

O2

O3

O4

Scheme 1. (a) Structural formulas of the template, vis., inosine 1, interferants, namely adenosine 2, guanosine, 3, glucose 4, thymine 5, as well as the cross-linking monomer, vis., 2,4,5,2’,4’,5’-hexa(thiophen-2-yl)-3,3`-bithiophene 6, and functional monomers, vis., 2,2’bithiophene-5-boronic acid 7 and 2-(cytosin-1-yl)ethyl p-bis(2,2’-bithien-5yl)methylbenzolate 8. (b) Structural formula of the pre-polymerization complex of 1 with functional monomers 7 and 8 and (c) the B3LYP/6-31G* optimized structure of the prepolymerized inosine 1 complex with functional monomers 7 and 8.

Figure captions: Figure 1. Curves of potential dependence of (a) current, as well as the (b) resonant frequency and (c) dynamic resistance changes simultaneously recorded for thin MIP-inosine film deposition by potentiodynamic electropolymerization at the 5-mm diameter electrode of the 10-MHz Au-QCR. The electropolymerization was carried using acetonitrile solution of 0.1 mM 1, 0.1 mM 7, 0.1 mM 8, 0.4 mM 6, and 0.1 M (TBA)ClO4. The potential scan rate was 50 mV s-1. Figure 2. The 1×1 µm2 area AFM images of (a) the MIP-inosine, and (b) the non-imprinted polymer (NIP) films. Figure 3. The drain current against drain voltage characteristics of the MIP-inosine EG-FET chemosensor for (a) different applied gate voltages, as well as (b) before and after addition of 50 µM inosine with the constant applied gate voltage of 1.5 V. Figure 4. Calibration plots for inosine constructed for the MIP-inosine chemosensor, (a) for different gate voltages and (b) for different concentrations of (1) inosine, (2) adenosine, (3) glucose, (4) thymine, and (5) guanosine. Inset shows comparison of the calibration plots for the MIP and the NIP film EG-FET. Figure 5. Calibration plots for inosine recorded at the (1) inosine extracted MIP film and (2) the NIP film coated 10-MHz Au-QCRs under flow-injection analysis conditions. The carrier solution flow rate was 35 µL min-1 and the injected sample volume was 200 L. Relative standard deviation for MIP film-coated resonator is 3.7%, while that for NIP film-coated is 33%.

Current , mA Dynamic Resistance ,  Frequency , kHz

8

4

a

0 -2

b

0 2 4 6 8 10

c

0

-3

-6 0,6

0,8

1,0

1,2

1,4

Potential, V vs. Ag pseudo-reference electrode

Figure 1.

Figure 2.

a

3.0

0.4 0.3 0.2

2.5 V

0.1 2.0 V

0.0

1.5 V

0

1

2

3

Drain voltage, V

Figure 3.

3.5

3.0 V

4

5

Drain current, mA

Drain current, mA

0.5

b

before inosine addition

2.5 2.0 after inosine addition

1.5 1.0 0.5 0.0 0

1

2

3

Drain voltage, V

4

5

1.5 V

-5 2.0 V -10 -15

Concentration, M 0

2

-20

4

6

8

10 0.005

NIP

0.000 -0.005

-25

-0.010

MIP

-30 0

5

-0.015 -0.020

2.5 V

10

3.0 V 15

Concentration, M

Figure 4.

20

b Current change, A

0

Drain current change, A

Current change, A

a

4

3

0.0

2

5

-0.5

-1.0 1

-1.5 0

10

20

30

Concentration, M

40

50

Frequency change, Hz

0

2

-10 -20 -30 -40

1

-50 -60

0

10

20

30

40

50

Inosine concentration, mM Figure 5.

Highlights

    

A novel chemical sensor for selective determination of the inosine was devised. Molecularly imprinted polymer recognition unit was electrodeposited on transducer. The EG-FET transducing element allowed for fine tuning of the sensor Inosine detectability of the sensor reached 0.62 M and its imprinting factor 29. Performance of the piezomicrogravimetric inosine chemosensor was compared.