Lab on a Chip PAPER
Cite this: Lab Chip, 2015, 15, 1717
Microbial trench-based optofluidic system for reagentless determination of phenolic compounds† David Sanahuja,ab Pablo Giménez-Gómez,b Núria Vigués,a Tobias Nils Ackermann,b Alfons Eduard Guerrero-Navarro,a Ferran Pujol-Vila,a Jordi Sacristán,b Nidia Santamaria,b María Sánchez-Contreras,c María Díaz-González,b Jordi Masa and Xavier Muñoz-Berbel*b Phenolic compounds are one of the main contaminants of soil and water due to their toxicity and persistence in the natural environment. Their presence is commonly determined with bulky and expensive instrumentation (e.g. chromatography systems), requiring sample collection and transport to the laboratory. Sample transport delays data acquisition, postponing potential actions to prevent environmental catastrophes. This article presents a portable, miniaturized, robust and low-cost microbial trench-based optofluidic system for reagentless determination of phenols in water. The optofluidic system is composed of a poly(methyl methacrylate) structure, incorporating polymeric optical elements and miniaturized discrete auxiliary components for optical transduction. An electronic circuit, adapted from a lock-in amplifier, is used for system control and interfering ambient light subtraction. In the trench, genetically modified bacteria are stably entrapped in an alginate hydrogel for quantitative determination of model phenol catechol. Alginate is also acting as a diffusion barrier for compounds present in the sample. Additionally, the superior refractive index of the gel (compared to water) confines the light in the lower level of the chip. Hence, the optical readout of the device is only altered by changes in the trench. Catechol molecules (colorless) in the
Received 10th December 2014, Accepted 2nd February 2015 DOI: 10.1039/c4lc01446d www.rsc.org/loc
sample diffuse through the alginate matrix and reach bacteria, which degrade them to a colored compound. The absorbance increase at 450 nm reports the presence of catechol simply, quickly (~10 min) and quantitatively without addition of chemical reagents. This miniaturized, portable and robust optofluidic system opens the possibility for quick and reliable determination of environmental contamination in situ, thus mitigating the effects of accidental spills.
Introduction Phenolic compounds are common contaminants in the environment due to many industrial activities (petroleum refining, dye manufacturing, textiles, etc.) and the use of numerous pesticides.1,2 Standard methods established by the US Environmental Protection Agency (EPA methods 420.1, 420.2, 420.4, 528, 604, etc.)3 or the International Organization for Standardization (ISO 6439:1990, ISO 8165-1:1992, ISO 81652:1999)4 for the determination of phenols include different spectrophotometric and chromatographic procedures. The
a
Department of Genetics and Microbiology Universitat Autonòma de Barcelona (UAB), Bellaterra, Barcelona, Spain b Instituto de Microelectrónica de Barcelona, IMB-CNM (CSIC), Spain. E-mail: [email protected] c Department of Biology, Universidad Autónoma de Madrid, Madrid, Spain † Electronic supplementary information (ESI) available. See DOI: 10.1039/ c4lc01446d
This journal is © The Royal Society of Chemistry 2015
most common spectrophotometric method is based on the detection of the colored product formed by reaction of phenols with 4-aminoantipyrine. Although this method is very simple and sensitive, it is not suitable for all phenolic compounds. Furthermore it cannot identify the different kinds of phenols, requiring, for this purpose, the use of sophisticated chromatographic chemical techniques.5 The expensive and bulky instrumentation required makes chromatographic methods well-suited for laboratory analysis but not for rapid in situ measurements. To this end, the development of portable and low cost lab-on-a-chip/microfluidic devices based on microbial bioassays appears as an interesting alternative to be considered. Unlike the classical chromatographic methods which can only quantify concentration of pollutants, bacterial bioassays provide information about the bioavailable fraction of toxic compounds and can be employed as early warning screening methods.6,7 Although a great variety of microbial biosensors have been designed and tested for phenolic compound determination,8–11 their integration into microfluidic
Lab Chip, 2015, 15, 1717–1726 | 1717
Paper
architectures has been poorly evaluated.12,13 Zhao and Dong12 reported a microfluidic device for acute toxicity sensing based on the bioluminescent Vibrio fischeri bacteria. In this case, bacteria are introduced into the fluidic system and mixed with the sample for toxicity assessment. Bioluminescence is then recorded externally with a photomultiplier module. In an attempt to improve the performance of this system, Yoo et al.13 immobilized bioluminescent bacteria on a microfluidic bio-MEMS chip for toxicity assessment. Bacteria are immobilized in the chip structure by entrapment in a polyvinyl alcohol/styrylpyridinium matrix after UV irradiation. UV radiation is very energetic and may compromise bacterial viability, making this immobilization method not suitable for cells. In this case, again, luminescence is determined with an external luminometer. Thus, microfluidic systems for microbial determination of phenolic species rely on poor physiological entrapments and fluidic structures without transduction capacities, which limit their application in situ. Considering recent advances in optofluidics,14–16 the aim of this work is to develop a low-cost, reagentless, portable, trench-based optofluidic system for fast in situ determination of phenolic compounds in water samples. Trench-based structures have been recently reported as interesting elements to retain cells free of hydrodynamic shear stress in microfluidic systems.17,18 Here, the trench is filled with an alginate/microbial hydrogel, which leads to the system with additional functionalities. Concretely, alginate19 is an ideal material for cell entrapment. Its use for the encapsulation of cells, drugs, enzymes, etc. has been widely reported in many different applications due to its excellent biocompatibility, biodegradability and its ability to form gels in the presence of certain cations (i.e. Ca2+) under extremely mild reaction conditions (water, room temperature) and without compromising cell viability.20–22 In addition to this, homogeneous and transparent calcium alginate gels, susceptible to be constituent materials of optical elements, can be prepared by controlling the gelation process.23,24 The development of gel-based optical waveguides, for example, has attracted increasing attention recently due to the great demand for biocompatible optical components in different research fields, including biochemical sensing.25–27 Finally, alginate has a porous matrix which acts as a diffusion barrier, allowing or impeding molecule diffusion depending on its size and change. For the properties described above, optical and physical properties of different calcium alginate gels as well as their ability to encapsulate genetically modified Escherichia coli PtomAB Amp 100 microorganisms will be evaluated. These bacteria can be used for reagentless determination of catechol since they are able to degrade mono- or bi-aromatic hydrocarbons to 2-hydroxymuconic semialdehyde (2-HMS). As this compound is yellow (brown) colored, the degradation of the colorless catechol is easily followed by measuring the increase in absorbance at 405 nm.28 Alginate gels incorporating these microorganisms are finally loaded in the trenchbased optofluidic system and evaluated as both sensing and optical components.
1718 | Lab Chip, 2015, 15, 1717–1726
Lab on a Chip
Materials and methods Trench-based optofluidic structure design and fabrication The trench-based architecture is illustrated in Fig. 1. This structure was composed of 5 individual poly(methyl methacrylate) (PMMA) layers containing optical and fluidic elements (Fig. 1A). Most relevant structures for optical transduction were incorporated in the 3 mm thick layer (2) (Fig. 1B). These included: a polymeric plano-convex converging lens (curvature radius = 1.7 mm, thickness = 3 mm) for emitted light focusing through the trench to the photo-detector; a floating lightguide (polymeric rectangular structure, 1.7 mm wide, 1.4 mm high, 3 mm thick) fixed by the edges to the PMMA structure and surrounded by air (see Fig. S1†) which, acting as air mirrors, improved light confinement and guidance from emitter to detector; elements for stable clamping and positioning of both light source (royal blue light emitting diode, LED, Philips Lumileds, The Netherlands) and photo-detector (Osram, Germany). The emitter and photo-detector were facing each other at a distance of 1.6 cm, which coincided with the optical path length of the system. The polymeric lens and lightguide architecture were theoretically optimized using TracePro simulation software (Lambda Research, Littleton, MA, USA). Optical simulations were performed in a 3D model of the optofluidic system, considering the optical properties of constituent materials (refractive index; nPMMA = 1.49, nAlginate = 1.39 and nair = 1). For consistency with experimental data, the light source and photo-detector geometry and performance were in agreement with supplied specifications (e.g. divergent LED emission pattern). In each optical simulation, a minimum of 40 000 rays were always considered. In Fig. 2, optical performance of the system without (A and B) and with (C and D) polymeric optical elements was analyzed and compared. As shown, the plano-convex converging lens was correcting LED emission divergence, enhancing total measured irradiance. At the same time, the floating lightguide helped in focusing light from the sample to the photo-detector, improving light collection. According to simulations, total irradiance magnitude reaching the photodetector when incorporating both optical elements (total irradiance = 0.05 W) was 5 times higher than that obtained when these elements were not present in the system (total irradiance = 0.01 W). This result validated the presence of these elements in the final design of the system. Apart from that, layer (2) also included a 120 μL trench (10 mm wide, 4 mm high, 3 mm thick) for in situ microbial/alginate hydrogel formation. A 120 μL fluidic reservoir (10 mm wide, 4 mm high, 3 mm thick) connected by two fluidic channels (700 μm wide, 6 mm long, 3 mm high) to the fluidic inlet and outlet is incorporated in layer (3). Layers (1) and (4), 1 mm thick each, sealed both reservoirs and channels and included some elements to facilitate LED and photo-detector exchange, as well as connection cable positioning. Layer (4) contained two holes (1 mm diameter), coinciding with the fluidic inlet/outlet present in layer (3), for fluidic connection between layers.
This journal is © The Royal Society of Chemistry 2015
Lab on a Chip
Paper
Fig. 1 Layer-by-layer illustration of the trench-based architecture indicating the most relevant optical and fluidic elements (A). Plan view of layer (2) with details of optical elements present in the system (B). Cross section of the assembled optofluidic system (C). Image illustrating the final embedded PMMA structure (D).
Fig. 2 Ray-tracing simulations (A and C) and total irradiance maps (B and D) at the detector surface for trench-based analytical systems with (C and D) or without (A and B) polymeric optical elements (i.e. plano-convex lens and floating lightguide). Ray color illustrates ray intensity, being red ones much more intense than blue ones.
Finally, layer (5), with a thickness of 5 mm, incorporated two fitting threads (6 mm diameter) for connection with highpressure fittings.
This journal is © The Royal Society of Chemistry 2015
All layers had corner holes (3 mm diameter) for layer assembly using screws. A cross section of the assembled trench-based optofluidic system showing the disposition of
Lab Chip, 2015, 15, 1717–1726 | 1719
Paper
the trench, optical elements and the fluidic channel is illustrated in Fig. 1C. Optofluidic system fabrication was performed on PMMA by using a CO2-laser system (Epilog Mini 24 Laser) and solvent assisted bonding. Optimal ablation conditions (applied power, laser emission frequency, ablation speed and number of cycles or repetitions) providing optical quality structures (i.e. spatial resolution, high transparency in the ablated region and low roughness of the walls) were established according to previous results (unpublished data). Optimal conditions were 20% power, 15% speed (both expressed as a percentage of the maximum magnitude available), 5000 Hz and one cycle for 1 mm thick layers, 70% power, 10% speed, 5000 Hz and one cycle for 3 mm thick layers and 70% power, 10% speed, 5000 Hz and two cycles for 5 mm thick layers. After fabrication, all layers were cleaned by sonication with distilled water for 30 min. Layers (1) and (2), as well as layers (3), (4) and (5), were permanently bonded after pretreatment of clean surfaces with methacrylic acid and pressure (5 kN for 40 min) using a 2-column precision hydraulic press IJP/O/Weber). After that, royal blue LED (maximum emission at 450 nm) and a wide-range photo-detector (350– 900 nm) were introduced in the optofluidic structure, which was finally assembled with screws using polydimethylsiloxane (PDMS) toric junctions to avoid fluid leakage. An image of the final optofluidic structure is illustrated in Fig. 1D. An electronic circuit was used for system control and interfering ambient light subtraction (for design and fabrication details S2†). It was based on a simplification of a lock-in amplifier where information was at the level of the modulus of the received signal and the phase was ruled out. Chemical reagents Sodium alginate, anhydrous calcium chloride, calcium sulphate, calcium carbonate, potassium hexacyanoferrate, potassium di-hydrogen phosphate, di-potassium hydrogen phosphate and sodium chloride were purchased from SigmaAldrich (St. Louis, MO, USA). Ringer's medium was 0.9% NaCl in distilled water. All solutions were prepared in distilled water as received without further purification Alginate gel preparation and physicochemical characterization Calcium-induced alginate hydrogels containing different concentration of sodium alginate (from 0.5 to 2.0% w/v) and calcium chloride (between 0 and 200 mM) were prepared in distilled water. The gelling kinetics, small-molecule diffusion and optical properties of the gels were determined with the optical setup illustrated in Fig. S3.† This setup consisted of a PMMA structure containing a 120 μL reservoir for alginate hydrogel formation and two 230 μm multimode optical fibers (Thorlabs, Dachau, Germany). Optical fibers were located on both sides of the PMMA structure (optical path = 1.6 cm) and connected to a broadband halogen lamp HL-2000-FHSA (Ocean Optics, USA) and to an USB2000 microspectrometer
1720 | Lab Chip, 2015, 15, 1717–1726
Lab on a Chip
(Ocean Optics, USA) for optical interrogation of the hydrogel. SpectraSuite software (Ocean Optics, USA) was used for data acquisition. In all cases, 90 μL of sodium alginate solution were introduced in the PMMA reservoir already containing 30 μL of calcium chloride and carefully mixed for gelation. After 10 min of gelation, transmittance spectra of alginate hydrogels were recorded between 300 and 900 nm taking water as the reference. Absorbance spectra were the average of three consecutive measurements performed at an integration time of 300 ms. Absorbance spectra of non-gelled alginate samples were also obtained under identical experimental conditions for comparison. In gelling kinetics assays, light intensity at 450 nm was recorded over time every 1 s (integration time = 300 ms). Each recorded spectrum was the average of 3 consecutive measurements. Finally, the diffusion of hydrophilic molecules in the alginate matrix was studied using ferricyanide as the model analyte. 20 μL of 40 mM ferricyanide solutions (in distilled water) were dispensed over alginate gels prepared in the PMMA reservoir and light intensity at 420 nm (maximum of the ferricyanide absorption band) was recorded (same experimental parameters as those in the gelling kinetics). Hydrogel texture and homogeneity was evaluated after each experiment by visual inspection. Escherichia coli PtomAB Amp 100 cultivation and characterization. Genetically modified Escherichia coli (E. coli) bacteria with a plasmid containing PtomAB gene and ampicillin resistance gene (for mutants' selection) were supplied by Grupo de Investigación Rizosfera at the Universidad Autónoma de Madrid. This plasmid allowed for the selective degradation of mono- and bi-aromatic phenolic compounds to 2-hydroxymuconic semialdehyde (2-HMS), a colored compound with an absorption band at 419 nm, through the enzyme toluene o-monooxygenase.29 Bacterial growth was performed aerobically in Luria– Bertani (LB) medium containing 100 mg mL−1 ampicillin (Sigma-Aldrich, St. Louis, MO, USA) at 37 °C for 16–18 h. Overnight cultures were centrifuged (Centrifuge 5804R, Eppendorf) at 10 100 × g for 15 min and re-suspended in Ringer's medium to a final concentration of 1 × 1010 colony forming units per mL (CFU mL−1). Bacterial concentration was determined by optical density at 600 nm (Smartspec™ Plus spectrophotometer, Bio-rad) by considering the relationship 1 A.U. = 3 × 108 CFU mL−1. Exact concentrations were determined by cell culturing on agar plates and colony counting. Microbial degradation of phenolic compounds was determined by monitoring 2-HMS production over time by performing the following protocol. First, 200 μL samples containing bacteria (between 0 and 1 × 109 CFU mL−1) and catechol (used as model phenolic compound; concentration range of 0–8% w/v) in Ringer's medium were introduced in 96-well plates. Absorbance at 405 nm (illustrative of 2-HMS production) was recorded for 20 h using an ELISA plate reader (Multiskan Ex Primary EIA V. 2.3, Thermo Electron
This journal is © The Royal Society of Chemistry 2015
Lab on a Chip
Corporation) with control software (Ascent Software V.2.6. for Multiskan, Thermo Scientific).
Microbial trench-based optofluidic system preparation and sensing principle Reagentless phenolic compound determination with the microbial trench-based optofluidic system was performed using catechol as the model molecule. Previous to catechol determination, microbial-alginate matrices were prepared in situ in the optofluidic system. Concretely, 30 μL of 100 mM calcium chloride and 90 μL of 0.5% alginate containing 1 × 109 CFU mL−1 genetically modified E. coli were introduced into the reservoir, mixed and left to gel for 10 min. After that, the optofluidic system was embedded and the LED and detector were connected to the electronic control circuit (already connected to a PC for data acquisition). The final architecture of the trench-based optofluidic system is illustrated in Fig. 3A. Water samples containing different catechol concentrations were then infused in the system and monitored over time (measurements recorded each 300 ms) for 30 min. After each catechol concentration, the alginate matrix was removed from the system and prepared again to avoid contamination. The sensing principle is schematized in Fig. 3B. Catechol samples infused in the optofluidic system diffused through the alginate hydrogel matrix (in the trench) until reaching genetically modified bacteria. These microorganisms
Paper
selectively metabolize catechol to produce 2-HMS, a colored molecule with an absorption band at 419 nm. Enzymatically mediated 2-HMS production is monitored for 30 min, providing quantitative values of catechol concentration in the sample. Alginate hydrogels were used to confine the light in the lower level of the chip, thus improving their optical performance.
Results and discussion Physicochemical characterization of alginate matrices Alginate hydrogels were the key elements of the optofluidic system described in this work, acting, at the same time, as entrapping matrix for bacteria, diffusion barrier and optical elements for light confining. For this reason, physicochemical properties of a wide variety of calcium alginate hydrogels were studied in order to select a gel that fulfils the required properties, concretely long-term stability, high homogeneity, high transmittance in the visible range, high diffusion rates and microbial entrapping capacity without compromising bacterial integrity or functionality. Alginate is one of the polymers most widely employed in the immobilization of living organisms due to its biocompatibility and soft polymerization conditions.20–22 Furthermore, alginate hydrogels are mostly transparent and present a refractive index of 1.37–1.40 (ref. 30) which ensures light confinement when surrounded by water. However, alginate solutions above 4% IJw/v) present a brown-orange color, which
Fig. 3 Picture of the trench-based optofluidic system (A). Schematic illustration of optofluidic system sensing principle (B).
This journal is © The Royal Society of Chemistry 2015
Lab Chip, 2015, 15, 1717–1726 | 1721
Paper
may interfere with optical measurements. For this reason, transmittance spectra of gelled and non-gelled alginate samples between 0.5 and 2.0% alginate were obtained using the optical setup described in the experimental section. As illustrated in Fig. 4A, non-gelled alginate samples with a concentration up to 2% IJw/v) presented high transmittances (above 90% transmittance) from 400 to 900 nm, independent of the concentration. After gelling (Fig. 4B), although maintaining values above 80% in most of the visible spectrum (from 500 to 900 nm), transmittance decreased below 70% in the region between 350 and 500 nm. As before, similar transmittance spectra were obtained independent of the alginate concentration in the gel. The reason for that wavelength-dependent transmittance drop after gelling is still controversial. From all possible causes, dispersion seemed to be the most plausible one. According to this, alginate hydrogel may be able to selectively disperse the light in this wavelength range, something that could not be done before gelling. Light dispersion may be due to several causes, for instance the formation of small-size scattering centers. In this sense, one possibility may be the formation of small porous structures, below 500 nm,
Lab on a Chip
which, acting as scattering centers, may disperse the light below 450 nm, without interacting with longer wavelengths. According to these results, optical properties of alginate hydrogels may vary during gel formation, interfering with optical measurements. Thus, although high transmittances were recorded in most of the visible light spectrum, it may be ensured that optical properties of the hydrogel were not changing during phenolic compound determination. To this end, several hydrogels containing different alginate and calcium concentrations were prepared. Gel formation was monitored over time by following light intensity changes at 450 nm (reporting light scattering center formation). Fig. 4C and E illustrate light intensity changes due to gel formation for samples containing different calcium and alginate concentrations, respectively. In most of the cases, light intensity randomly decreased with time until reaching a plateau at around 10 min of incubation. Hence, after 10 min, hydrogels should be optically stable and should not interfere with optical measurements. Thus, 10 min was selected as the ideal incubation time for gel formation. However, after 10 min of gelation, not all hydrogels exhibited texture and homogeneity
Fig. 4 Physicochemical characterization of alginate hydrogels. Transmittance spectra of non-gelled (A) and gelled (B) alginate samples. Gelling kinetic plots (at 450 nm) for samples containing 0.5% alginate and calcium chloride concentrations from 0 to 200 mM (C) and alginate concentrations from 0.5 to 2.0% and 100 mM calcium chloride (E). Intensity magnitude (at 450 nm) of 0.5% alginate hydrogels containing different calcium chloride concentrations (D) and hydrogels containing 100 mM calcium chloride and alginate concentrations from 0.5 to 2.0% (F) after 10 min gelling. Error bars represent the standard deviation from the mean (95% confidence, n = 3). See experimental section for details.
1722 | Lab Chip, 2015, 15, 1717–1726
This journal is © The Royal Society of Chemistry 2015
Lab on a Chip
suitable for the final application. For this reason, an in-depth analysis of gel formation kinetics was performed. Considering calcium concentration, both visual inspection and gel formation kinetics (Fig. 4C) confirmed that a minimum concentration of calcium around 20 mM was necessary for hydrogel formation (below 20 mM light intensity remained constant for the duration of the experiment). For calcium concentrations higher than 20 mM, hydrogel texture tremendously varied. Samples containing between 20 and 40 mM calcium present viscous fluid texture, quite far from that of a gel. Above 100 mM it was possible to obtain homogeneous hydrogels. This observation was in accordance with light intensity magnitude recorded in the plateau. As shown in Fig. 4D, light intensity magnitude in the plateau decreased when increasing calcium concentration until stabilizing at around 100 mM, where homogeneous and stable hydrogels were obtained. Thus, intensity magnitude in the plateau may be used as a quantitative indicator of the degree of gel formation. According to this parameter, 100 mM calcium concentrations were necessary to obtain homogeneous and stable gels. In the case of alginate, no significant differences were observed when comparing gel formation kinetics (Fig. 4E) or light intensity magnitude in the plateau (Fig. 4F) of hydrogels containing alginate concentrations between 0.5 and 2.0%. However, visual inspection of these hydrogels confirmed that those containing higher concentrations of alginate were more rigid and less homogeneous, coinciding with already reported data.31 This result suggested the use of low alginate concentration (below 1%) in this application. On the other hand, optimal performance of the system also required suitable diffusion of molecules through the alginate matrix until reaching bacteria, which may determine the response time of the system. To this end, diffusion kinetics through hydrogel matrices containing different alginate concentrations (from 0.5 to 2.0%) was evaluated by following the protocol detailed in the experimental section. Ferricyanide, with a wide absorption band at 420 nm, was used as the model molecule. Diffusion kinetic results are illustrated in Fig. 5. In all cases, a progressive absorbance increase at 420 nm was recorded over time (Fig. 5A). To ensure that this was due to ferricyanide (and not to changes in the gel structure), absorbance spectra were recorded along the experiment, taking the spectrum of the hydrogel before ferricyanide inoculation as the reference. Fig. 5B shows how the ferricyanide absorption band, with a maximum at 420 nm, increased with time due to diffusion. Two different diffusion tendencies were observed depending on the alginate concentration. Low concentrations (below 1% alginate) present slow diffusion patterns and ferricyanide required around 250 s to achieve an absorbance magnitude of around 1. On the other hand, above 1% alginate, ferricyanide diffuse faster, requiring less than 150 s to cross the hydrogel. The reason for that may be again the heterogeneity of high alginate concentration matrices, combining gelled and non-gelled areas that favor fast ferricyanide diffusion.
This journal is © The Royal Society of Chemistry 2015
Paper
Fig. 5 Ferricyanide diffusion plots in alginate hydrogels. Absorbance variation (at 420 nm) with time illustrating ferricyanide diffusion into 0.5, 0.7, 1.0, 1.5 and 2.0% alginate hydrogels (A). Representation of the variation of absorption spectra of 0.5% alginate hydrogels during ferricyanide diffusion (B). The alginate matrix spectrum after 10 min of incubation is used as the reference. Error bars represent the standard deviation from the mean (95% confidence, n = 3).
According to experimental data, alginate hydrogels with 0.5% alginate and 100 mM calcium were selected to be implemented in the trench-based optofluidic system for being the most homogeneous, repetitive ones and also presenting excellent optical and diffusional properties. Escherichia coli PtomAB Amp 100 characterization. Phenolic compound determination was based on the use of genetically modified E. coli bacteria capable to metabolize monoand bi-aromatic phenolic compounds to 2-HMS, a degradation product with a wide absorption band around 419 nm (Fig. S4†). Thus, in this case, phenolic compounds could be selectively determined by monitoring absorption changes without the need of any additional chemical/biological compound. Bacterial activity was evaluated as detailed in the experimental section, using catechol as the model molecule. Four bacterial concentrations, from 0 to 1 × 109 CFU mL−1, and eight catechol concentrations (between 0 and 8% w/v) were analyzed for assay optimization (Fig. 6A–C). In all cases, absorbance magnitude increased with time for the duration of the experiment due to catechol degradation/metabolism mechanisms. According to experimental data, degradation kinetics depended on bacterial concentration. Below 1 × 107 CFU mL−1, similar degradation kinetics was obtained independent of bacterial concentration or even the presence or the absence of bacteria (Fig. S5†). Considering this, catechol degradation kinetics below 1 × 107 CFU
Lab Chip, 2015, 15, 1717–1726 | 1723
Paper
Lab on a Chip
Fig. 6 Bacterial metabolism characterization. Representation of the variation of absorbance at 405 nm with time for several catechol samples containing 0 CFU mL−1 (A), 1 × 108 CFU mL−1 (B) or 1 × 109 CFU mL−1 (C) genetically modified bacteria. Each plot is the result of three independent experiments. Error bars represent the standard deviation (95% confidence). Degradation kinetic plots, where the slope (absorbance at 405 nm variation with time) is represented against catechol concentration for samples containing several E. coli concentrations (D). Slope was determined by considering the first 10 min of the absorbance versus time plot.
mL−1 may be attributed to catechol photo-oxidation. Above 1 × 108 CFU mL−1, faster and bacterial concentrationdependent catechol degradation kinetics was recorded (Fig. 6B and C), suggesting a microbial-mediated degradation mechanism. These results indicated that a minimum concentration of 1 × 108 CFU mL−1 was necessary for microbialmediated selective determination of catechol. Catechol concentration was determined through catechol degradation kinetic plots using the slope of the absorbance versus time graphics (first 10 min) as the analytical signal. Degradation kinetic plots were advantageous for allowing quick (10 min) and reliable quantification of catechol concentration without interference of biomass scattering. As shown in Fig. 6D, the slope magnitude of the degradation curve increased with increasing catechol concentration until 1%. After that, it started decreasing to values similar to those recorded by samples without bacteria (around 8% catechol). According to these data, catechol was toxic for these genetically modified bacteria which started dying when catechol concentration in the sample exceeded a concentration magnitude above 1% catechol. Thus, with this method, reagentless determination of catechol was only possible below 1% concentrations.
Determination of catechol in water samples using the microbial trench-based optofluidic system Before catechol determination in water samples, microbial/ hydrogel formation interference in optical measurements and bacterial viability in the hydrogel matrix were analyzed. In the evaluation of gel formation interference, light intensity
1724 | Lab Chip, 2015, 15, 1717–1726
variation during microbial/alginate gel formation was analyzed and compared with alginate gels without bacteria. Fig. 7A illustrates light intensity variation with time during 0.5% alginate hydrogel formation containing 1 × 109 CFU mL−1. As shown, light intensity decreased progressively during gel formation until stabilizing after 8 min gelling, coinciding with previous data. In order to evaluate microbial stability in the hydrogel, microbial/alginate hydrogels prepared as before were stored in the fridge (4 °C) in Ringer's medium or dried for 19 days. Bacterial viability was regularly evaluated by cell counting. No significant differences in the number of viable cells were obtained after 19 days for wet samples, whereas in dried samples the number of viable bacteria drastically decreased with time (Fig. 7B). Hence, alginate hydrogels stored in wet conditions are suitable for long-term microbial entrapment. For the reagentless determination of phenolic compounds in water samples, 0.5% alginate matrices containing 1 × 109 CFU mL−1 genetically modified bacteria were prepared in the trench-based optofluidic system and analyzed as detailed in the experimental section. After 10 min of hydrogel formation, the trench-based optofluidic system was embedded and LED intensity was adjusted (with a resistance) to provide an initial intensity around 1100–1200 counts. LED stability was evaluated before measurement by monitoring its intensity for around 1 h after addition of water in the optofluidic system. In all cases, variability measurements below 1% were obtained (Fig. S6†). Water samples containing catechol concentrations from 0 to 2% were then inoculated in the system and absorbance variation was monitored for 1 h with the electronic control system already described. In Fig. 7C,
This journal is © The Royal Society of Chemistry 2015
Lab on a Chip
Paper
Fig. 7 Representation of the variation of light intensity magnitude at 450 nm during gel formation for hydrogels containing 0.5% alginate and 1 × 109 CFU mL−1 E. coli (A). Representation of the variation of the number of viable cells with time for microbial/hydrogels stored at 4 °C either wet or dried (B). Representation of the variation of absorbance magnitude with time for several catechol samples containing 1 × 109 CFU mL−1 genetically modified bacteria (C). Degradation kinetic plots, where the slope (absorbance variation with time) is represented against catechol concentration (D). Slope was determined by considering the first 10 min of the absorbance versus time plot. Each plot is the result of three independent experiments. Error bars represent the standard deviation (95% confidence).
absorbance magnitude, taking water as the reference, is represented with time for the five catechol concentrations under study. As before, microbial-mediated catechol degradation (Fig. 7D) was dependent on catechol concentration and allowed for reagentless and quantitative determination of catechol concentration in a wide range of concentration (up to 1% catechol). In summary, this microbial trench-based optofluidic system represents a cheap, simple and robust alternative for quick (10 min), sensitive, reagentless and quantitative determination of catechol in water samples. The robustness, simplicity and reagentless nature of the bioassay (for the use of genetically modified microorganisms) open the possibility for in situ determination of phenolic compound contamination without the need of sample transport or pretreatment.
Conclusions Reagentless determination of catechol was achieved with a microbial trench-based optofluidic system controlled by a miniaturized portable electronic circuit adapted from a lockin-amplifier, capable to monitor light intensity changes, subtracting, at the same time, ambient light interference. An alginate hydrogel was incorporated into the trench for microbial entrapment and light confinement from the LED emitter to the photodetector, both located on opposite sides of the trench. Integrated optical elements such as PMMA
This journal is © The Royal Society of Chemistry 2015
convergent lenses, floating lightguides and fluidic elements (i.e. inlet, outlet, channels) were also included in the PMMA optofluidic structure for optimal optical transduction and fluid management. Alginate hydrogels containing 0.5% alginate and 100 mM calcium demonstrated suitable optical (high transmittance in the visible range), structural (high homogeneity and stability) and diffusional (high diffusion of small molecules) properties to be implemented in the optofluidic system. Additionally, their properties did not change by incorporation of genetically modified bacteria (up to 1 × 109 CFU mL−1) capable of selectively metabolizing catechol (used as model molecule). With this system, catechol concentrations from 0 to 1% IJw/v) could be determined without the need of additional reagents after only 10 min of reaction through the formation of a colored bacterial degradation product. This miniaturized, portable, robust and simple trenchbased optofluidic system, providing information on relevant contaminants after short incubation times and without the need of additional reagents, opens the possibility of in situ determination of environmental pollutants, thus mitigating the effects of accidental spills.
Acknowledgements This work was supported in part by the Spanish R & D National Program (MINECO Project TEC2011-29045-C04-01/04). X. M.-B. also wants to acknowledge the “Ramón y Cajal” program
Lab Chip, 2015, 15, 1717–1726 | 1725
Paper
from the Spanish Government. P. G.-G. is grateful to the Spanish MEC for the award of a research studentship from FPI Program.
References 1 J. Michalowicz and W. Duda, Pol. J. Environ. Stud., 2007, 16, 347–362. 2 S. Wasi, S. Tabrez and M. Ahmad, Environ. Monit. Assess., 2013, 185, 2585–2593. 3 United States Environmental Protection Agency, http://www. epa.gov/. 4 International Organization for Standardization, http://www. iso.org/. 5 J. Martin, D. Camacho-Munoz, J. L. Santos, I. Aparicio and E. Alonso, Anal. Bioanal. Chem., 2014, 406, 3709–3716. 6 M. Park, S. L. Tsai and W. Chen, Sensors, 2013, 13, 5777–5795. 7 T. T. Xu, D. M. Close, G. S. Sayler and S. Ripp, Ecol. Indic., 2013, 28, 125–141. 8 H. J. Shin, Appl. Microbiol. Biotechnol., 2012, 93, 1895–1904. 9 C. Liu, D. M. Yong, D. B. Yu and S. J. Dong, Talanta, 2011, 84, 766–770. 10 X. M. Liu, K. J. Germaine, D. Ryan and D. N. Dowling, Sensors, 2010, 10, 1377–1398. 11 L. A. Su, W. Z. Jia, C. J. Hou and Y. Lei, Biosens. Bioelectron., 2011, 26, 1788–1799. 12 X. Y. Zhao and T. Dong, Int. J. Environ. Res. Public Health, 2013, 10, 6748–6763. 13 S. K. Yoo, J. H. Lee, S. S. Yun, M. B. Gu and J. H. Lee, Biosens. Bioelectron., 2007, 22, 1586–1592. 14 Y. C. Seow, S. P. Lim and H. P. Lee, Lab Chip, 2012, 12, 3810–3815.
1726 | Lab Chip, 2015, 15, 1717–1726
Lab on a Chip
15 Y. C. Seow, S. P. Lim and H. P. Lee, Microfluid. Nanofluid., 2011, 11, 451–458. 16 Y. Yang, L. K. Chin, J. M. Tsai, D. P. Tsai, N. I. Zheludev and A. Q. Liu, Lab Chip, 2012, 12, 3785–3790. 17 I. K. Dimov, G. Kijanka, Y. Park, J. Ducree, T. Kang and L. P. Lee, Lab Chip, 2011, 11, 2701–2710. 18 T. M. O'Connell, D. King, C. K. Dixit, B. O'Connor, D. Walls and J. Ducree, Lab Chip, 2014, 14, 3629–3639. 19 S. N. Pawar and K. J. Edgar, Biomaterials, 2012, 33, 3279–3305. 20 K. Y. Lee and D. J. Mooney, Prog. Polym. Sci., 2012, 37, 106–126. 21 W. R. Gombotz and S. F. Wee, Adv. Drug Delivery Rev., 2012, 64, 194–205. 22 S. M. Selimoglu and M. Elibol, Crit. Rev. Biotechnol., 2010, 30, 145–159. 23 T. Andersen, J. E. Melvik, O. Gaserod, E. Alsberg and B. E. Christensen, Carbohydr. Polym., 2014, 99, 249–256. 24 C. K. Kuo and P. X. Ma, Biomaterials, 2001, 22, 511–521. 25 S. T. Parker, P. Domachuk, J. Amsden, J. Bressner, J. A. Lewis, D. L. Kaplan and F. G. Omenetto, Adv. Mater., 2009, 21, 2411–2415. 26 A. Jain, A. H. J. Yang and D. Erickson, Opt. Lett., 2012, 37, 1472–1474. 27 M. Choi, J. Choi, S. Kim, S. Nizamoglu, S. K. Hahn and S. H. Yun, Nat. Photonics, 2013, 7, 987–994. 28 L. A. Edghill, A. D. Russell, M. J. Day and J. R. Furr, J. Appl. Microbiol., 1999, 87, 91–98. 29 D. C. Yee, J. A. Maynard and T. K. Wood, Appl. Environ. Microbiol., 1998, 64, 112–118. 30 O. Esteban, F. Marva and J. C. Martinez-Anton, Opt. Mater., 2009, 31, 696–699. 31 Y. Maki, K. Ito, N. Hosoya, C. Yoneyama, K. Furusawa, T. Yamamoto, T. Dobashi, Y. Sugimoto and K. Wakabayashi, Biomacromolecules, 2011, 12, 2145–2152.
This journal is © The Royal Society of Chemistry 2015
Copyright of Lab on a Chip - Miniaturisation for Chemistry & Biology is the property of Royal Society of Chemistry and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Cite this: Lab Chip, 2015, 15, 1717
Microbial trench-based optofluidic system for reagentless determination of phenolic compounds† David Sanahuja,ab Pablo Giménez-Gómez,b Núria Vigués,a Tobias Nils Ackermann,b Alfons Eduard Guerrero-Navarro,a Ferran Pujol-Vila,a Jordi Sacristán,b Nidia Santamaria,b María Sánchez-Contreras,c María Díaz-González,b Jordi Masa and Xavier Muñoz-Berbel*b Phenolic compounds are one of the main contaminants of soil and water due to their toxicity and persistence in the natural environment. Their presence is commonly determined with bulky and expensive instrumentation (e.g. chromatography systems), requiring sample collection and transport to the laboratory. Sample transport delays data acquisition, postponing potential actions to prevent environmental catastrophes. This article presents a portable, miniaturized, robust and low-cost microbial trench-based optofluidic system for reagentless determination of phenols in water. The optofluidic system is composed of a poly(methyl methacrylate) structure, incorporating polymeric optical elements and miniaturized discrete auxiliary components for optical transduction. An electronic circuit, adapted from a lock-in amplifier, is used for system control and interfering ambient light subtraction. In the trench, genetically modified bacteria are stably entrapped in an alginate hydrogel for quantitative determination of model phenol catechol. Alginate is also acting as a diffusion barrier for compounds present in the sample. Additionally, the superior refractive index of the gel (compared to water) confines the light in the lower level of the chip. Hence, the optical readout of the device is only altered by changes in the trench. Catechol molecules (colorless) in the
Received 10th December 2014, Accepted 2nd February 2015 DOI: 10.1039/c4lc01446d www.rsc.org/loc
sample diffuse through the alginate matrix and reach bacteria, which degrade them to a colored compound. The absorbance increase at 450 nm reports the presence of catechol simply, quickly (~10 min) and quantitatively without addition of chemical reagents. This miniaturized, portable and robust optofluidic system opens the possibility for quick and reliable determination of environmental contamination in situ, thus mitigating the effects of accidental spills.
Introduction Phenolic compounds are common contaminants in the environment due to many industrial activities (petroleum refining, dye manufacturing, textiles, etc.) and the use of numerous pesticides.1,2 Standard methods established by the US Environmental Protection Agency (EPA methods 420.1, 420.2, 420.4, 528, 604, etc.)3 or the International Organization for Standardization (ISO 6439:1990, ISO 8165-1:1992, ISO 81652:1999)4 for the determination of phenols include different spectrophotometric and chromatographic procedures. The
a
Department of Genetics and Microbiology Universitat Autonòma de Barcelona (UAB), Bellaterra, Barcelona, Spain b Instituto de Microelectrónica de Barcelona, IMB-CNM (CSIC), Spain. E-mail: [email protected] c Department of Biology, Universidad Autónoma de Madrid, Madrid, Spain † Electronic supplementary information (ESI) available. See DOI: 10.1039/ c4lc01446d
This journal is © The Royal Society of Chemistry 2015
most common spectrophotometric method is based on the detection of the colored product formed by reaction of phenols with 4-aminoantipyrine. Although this method is very simple and sensitive, it is not suitable for all phenolic compounds. Furthermore it cannot identify the different kinds of phenols, requiring, for this purpose, the use of sophisticated chromatographic chemical techniques.5 The expensive and bulky instrumentation required makes chromatographic methods well-suited for laboratory analysis but not for rapid in situ measurements. To this end, the development of portable and low cost lab-on-a-chip/microfluidic devices based on microbial bioassays appears as an interesting alternative to be considered. Unlike the classical chromatographic methods which can only quantify concentration of pollutants, bacterial bioassays provide information about the bioavailable fraction of toxic compounds and can be employed as early warning screening methods.6,7 Although a great variety of microbial biosensors have been designed and tested for phenolic compound determination,8–11 their integration into microfluidic
Lab Chip, 2015, 15, 1717–1726 | 1717
Paper
architectures has been poorly evaluated.12,13 Zhao and Dong12 reported a microfluidic device for acute toxicity sensing based on the bioluminescent Vibrio fischeri bacteria. In this case, bacteria are introduced into the fluidic system and mixed with the sample for toxicity assessment. Bioluminescence is then recorded externally with a photomultiplier module. In an attempt to improve the performance of this system, Yoo et al.13 immobilized bioluminescent bacteria on a microfluidic bio-MEMS chip for toxicity assessment. Bacteria are immobilized in the chip structure by entrapment in a polyvinyl alcohol/styrylpyridinium matrix after UV irradiation. UV radiation is very energetic and may compromise bacterial viability, making this immobilization method not suitable for cells. In this case, again, luminescence is determined with an external luminometer. Thus, microfluidic systems for microbial determination of phenolic species rely on poor physiological entrapments and fluidic structures without transduction capacities, which limit their application in situ. Considering recent advances in optofluidics,14–16 the aim of this work is to develop a low-cost, reagentless, portable, trench-based optofluidic system for fast in situ determination of phenolic compounds in water samples. Trench-based structures have been recently reported as interesting elements to retain cells free of hydrodynamic shear stress in microfluidic systems.17,18 Here, the trench is filled with an alginate/microbial hydrogel, which leads to the system with additional functionalities. Concretely, alginate19 is an ideal material for cell entrapment. Its use for the encapsulation of cells, drugs, enzymes, etc. has been widely reported in many different applications due to its excellent biocompatibility, biodegradability and its ability to form gels in the presence of certain cations (i.e. Ca2+) under extremely mild reaction conditions (water, room temperature) and without compromising cell viability.20–22 In addition to this, homogeneous and transparent calcium alginate gels, susceptible to be constituent materials of optical elements, can be prepared by controlling the gelation process.23,24 The development of gel-based optical waveguides, for example, has attracted increasing attention recently due to the great demand for biocompatible optical components in different research fields, including biochemical sensing.25–27 Finally, alginate has a porous matrix which acts as a diffusion barrier, allowing or impeding molecule diffusion depending on its size and change. For the properties described above, optical and physical properties of different calcium alginate gels as well as their ability to encapsulate genetically modified Escherichia coli PtomAB Amp 100 microorganisms will be evaluated. These bacteria can be used for reagentless determination of catechol since they are able to degrade mono- or bi-aromatic hydrocarbons to 2-hydroxymuconic semialdehyde (2-HMS). As this compound is yellow (brown) colored, the degradation of the colorless catechol is easily followed by measuring the increase in absorbance at 405 nm.28 Alginate gels incorporating these microorganisms are finally loaded in the trenchbased optofluidic system and evaluated as both sensing and optical components.
1718 | Lab Chip, 2015, 15, 1717–1726
Lab on a Chip
Materials and methods Trench-based optofluidic structure design and fabrication The trench-based architecture is illustrated in Fig. 1. This structure was composed of 5 individual poly(methyl methacrylate) (PMMA) layers containing optical and fluidic elements (Fig. 1A). Most relevant structures for optical transduction were incorporated in the 3 mm thick layer (2) (Fig. 1B). These included: a polymeric plano-convex converging lens (curvature radius = 1.7 mm, thickness = 3 mm) for emitted light focusing through the trench to the photo-detector; a floating lightguide (polymeric rectangular structure, 1.7 mm wide, 1.4 mm high, 3 mm thick) fixed by the edges to the PMMA structure and surrounded by air (see Fig. S1†) which, acting as air mirrors, improved light confinement and guidance from emitter to detector; elements for stable clamping and positioning of both light source (royal blue light emitting diode, LED, Philips Lumileds, The Netherlands) and photo-detector (Osram, Germany). The emitter and photo-detector were facing each other at a distance of 1.6 cm, which coincided with the optical path length of the system. The polymeric lens and lightguide architecture were theoretically optimized using TracePro simulation software (Lambda Research, Littleton, MA, USA). Optical simulations were performed in a 3D model of the optofluidic system, considering the optical properties of constituent materials (refractive index; nPMMA = 1.49, nAlginate = 1.39 and nair = 1). For consistency with experimental data, the light source and photo-detector geometry and performance were in agreement with supplied specifications (e.g. divergent LED emission pattern). In each optical simulation, a minimum of 40 000 rays were always considered. In Fig. 2, optical performance of the system without (A and B) and with (C and D) polymeric optical elements was analyzed and compared. As shown, the plano-convex converging lens was correcting LED emission divergence, enhancing total measured irradiance. At the same time, the floating lightguide helped in focusing light from the sample to the photo-detector, improving light collection. According to simulations, total irradiance magnitude reaching the photodetector when incorporating both optical elements (total irradiance = 0.05 W) was 5 times higher than that obtained when these elements were not present in the system (total irradiance = 0.01 W). This result validated the presence of these elements in the final design of the system. Apart from that, layer (2) also included a 120 μL trench (10 mm wide, 4 mm high, 3 mm thick) for in situ microbial/alginate hydrogel formation. A 120 μL fluidic reservoir (10 mm wide, 4 mm high, 3 mm thick) connected by two fluidic channels (700 μm wide, 6 mm long, 3 mm high) to the fluidic inlet and outlet is incorporated in layer (3). Layers (1) and (4), 1 mm thick each, sealed both reservoirs and channels and included some elements to facilitate LED and photo-detector exchange, as well as connection cable positioning. Layer (4) contained two holes (1 mm diameter), coinciding with the fluidic inlet/outlet present in layer (3), for fluidic connection between layers.
This journal is © The Royal Society of Chemistry 2015
Lab on a Chip
Paper
Fig. 1 Layer-by-layer illustration of the trench-based architecture indicating the most relevant optical and fluidic elements (A). Plan view of layer (2) with details of optical elements present in the system (B). Cross section of the assembled optofluidic system (C). Image illustrating the final embedded PMMA structure (D).
Fig. 2 Ray-tracing simulations (A and C) and total irradiance maps (B and D) at the detector surface for trench-based analytical systems with (C and D) or without (A and B) polymeric optical elements (i.e. plano-convex lens and floating lightguide). Ray color illustrates ray intensity, being red ones much more intense than blue ones.
Finally, layer (5), with a thickness of 5 mm, incorporated two fitting threads (6 mm diameter) for connection with highpressure fittings.
This journal is © The Royal Society of Chemistry 2015
All layers had corner holes (3 mm diameter) for layer assembly using screws. A cross section of the assembled trench-based optofluidic system showing the disposition of
Lab Chip, 2015, 15, 1717–1726 | 1719
Paper
the trench, optical elements and the fluidic channel is illustrated in Fig. 1C. Optofluidic system fabrication was performed on PMMA by using a CO2-laser system (Epilog Mini 24 Laser) and solvent assisted bonding. Optimal ablation conditions (applied power, laser emission frequency, ablation speed and number of cycles or repetitions) providing optical quality structures (i.e. spatial resolution, high transparency in the ablated region and low roughness of the walls) were established according to previous results (unpublished data). Optimal conditions were 20% power, 15% speed (both expressed as a percentage of the maximum magnitude available), 5000 Hz and one cycle for 1 mm thick layers, 70% power, 10% speed, 5000 Hz and one cycle for 3 mm thick layers and 70% power, 10% speed, 5000 Hz and two cycles for 5 mm thick layers. After fabrication, all layers were cleaned by sonication with distilled water for 30 min. Layers (1) and (2), as well as layers (3), (4) and (5), were permanently bonded after pretreatment of clean surfaces with methacrylic acid and pressure (5 kN for 40 min) using a 2-column precision hydraulic press IJP/O/Weber). After that, royal blue LED (maximum emission at 450 nm) and a wide-range photo-detector (350– 900 nm) were introduced in the optofluidic structure, which was finally assembled with screws using polydimethylsiloxane (PDMS) toric junctions to avoid fluid leakage. An image of the final optofluidic structure is illustrated in Fig. 1D. An electronic circuit was used for system control and interfering ambient light subtraction (for design and fabrication details S2†). It was based on a simplification of a lock-in amplifier where information was at the level of the modulus of the received signal and the phase was ruled out. Chemical reagents Sodium alginate, anhydrous calcium chloride, calcium sulphate, calcium carbonate, potassium hexacyanoferrate, potassium di-hydrogen phosphate, di-potassium hydrogen phosphate and sodium chloride were purchased from SigmaAldrich (St. Louis, MO, USA). Ringer's medium was 0.9% NaCl in distilled water. All solutions were prepared in distilled water as received without further purification Alginate gel preparation and physicochemical characterization Calcium-induced alginate hydrogels containing different concentration of sodium alginate (from 0.5 to 2.0% w/v) and calcium chloride (between 0 and 200 mM) were prepared in distilled water. The gelling kinetics, small-molecule diffusion and optical properties of the gels were determined with the optical setup illustrated in Fig. S3.† This setup consisted of a PMMA structure containing a 120 μL reservoir for alginate hydrogel formation and two 230 μm multimode optical fibers (Thorlabs, Dachau, Germany). Optical fibers were located on both sides of the PMMA structure (optical path = 1.6 cm) and connected to a broadband halogen lamp HL-2000-FHSA (Ocean Optics, USA) and to an USB2000 microspectrometer
1720 | Lab Chip, 2015, 15, 1717–1726
Lab on a Chip
(Ocean Optics, USA) for optical interrogation of the hydrogel. SpectraSuite software (Ocean Optics, USA) was used for data acquisition. In all cases, 90 μL of sodium alginate solution were introduced in the PMMA reservoir already containing 30 μL of calcium chloride and carefully mixed for gelation. After 10 min of gelation, transmittance spectra of alginate hydrogels were recorded between 300 and 900 nm taking water as the reference. Absorbance spectra were the average of three consecutive measurements performed at an integration time of 300 ms. Absorbance spectra of non-gelled alginate samples were also obtained under identical experimental conditions for comparison. In gelling kinetics assays, light intensity at 450 nm was recorded over time every 1 s (integration time = 300 ms). Each recorded spectrum was the average of 3 consecutive measurements. Finally, the diffusion of hydrophilic molecules in the alginate matrix was studied using ferricyanide as the model analyte. 20 μL of 40 mM ferricyanide solutions (in distilled water) were dispensed over alginate gels prepared in the PMMA reservoir and light intensity at 420 nm (maximum of the ferricyanide absorption band) was recorded (same experimental parameters as those in the gelling kinetics). Hydrogel texture and homogeneity was evaluated after each experiment by visual inspection. Escherichia coli PtomAB Amp 100 cultivation and characterization. Genetically modified Escherichia coli (E. coli) bacteria with a plasmid containing PtomAB gene and ampicillin resistance gene (for mutants' selection) were supplied by Grupo de Investigación Rizosfera at the Universidad Autónoma de Madrid. This plasmid allowed for the selective degradation of mono- and bi-aromatic phenolic compounds to 2-hydroxymuconic semialdehyde (2-HMS), a colored compound with an absorption band at 419 nm, through the enzyme toluene o-monooxygenase.29 Bacterial growth was performed aerobically in Luria– Bertani (LB) medium containing 100 mg mL−1 ampicillin (Sigma-Aldrich, St. Louis, MO, USA) at 37 °C for 16–18 h. Overnight cultures were centrifuged (Centrifuge 5804R, Eppendorf) at 10 100 × g for 15 min and re-suspended in Ringer's medium to a final concentration of 1 × 1010 colony forming units per mL (CFU mL−1). Bacterial concentration was determined by optical density at 600 nm (Smartspec™ Plus spectrophotometer, Bio-rad) by considering the relationship 1 A.U. = 3 × 108 CFU mL−1. Exact concentrations were determined by cell culturing on agar plates and colony counting. Microbial degradation of phenolic compounds was determined by monitoring 2-HMS production over time by performing the following protocol. First, 200 μL samples containing bacteria (between 0 and 1 × 109 CFU mL−1) and catechol (used as model phenolic compound; concentration range of 0–8% w/v) in Ringer's medium were introduced in 96-well plates. Absorbance at 405 nm (illustrative of 2-HMS production) was recorded for 20 h using an ELISA plate reader (Multiskan Ex Primary EIA V. 2.3, Thermo Electron
This journal is © The Royal Society of Chemistry 2015
Lab on a Chip
Corporation) with control software (Ascent Software V.2.6. for Multiskan, Thermo Scientific).
Microbial trench-based optofluidic system preparation and sensing principle Reagentless phenolic compound determination with the microbial trench-based optofluidic system was performed using catechol as the model molecule. Previous to catechol determination, microbial-alginate matrices were prepared in situ in the optofluidic system. Concretely, 30 μL of 100 mM calcium chloride and 90 μL of 0.5% alginate containing 1 × 109 CFU mL−1 genetically modified E. coli were introduced into the reservoir, mixed and left to gel for 10 min. After that, the optofluidic system was embedded and the LED and detector were connected to the electronic control circuit (already connected to a PC for data acquisition). The final architecture of the trench-based optofluidic system is illustrated in Fig. 3A. Water samples containing different catechol concentrations were then infused in the system and monitored over time (measurements recorded each 300 ms) for 30 min. After each catechol concentration, the alginate matrix was removed from the system and prepared again to avoid contamination. The sensing principle is schematized in Fig. 3B. Catechol samples infused in the optofluidic system diffused through the alginate hydrogel matrix (in the trench) until reaching genetically modified bacteria. These microorganisms
Paper
selectively metabolize catechol to produce 2-HMS, a colored molecule with an absorption band at 419 nm. Enzymatically mediated 2-HMS production is monitored for 30 min, providing quantitative values of catechol concentration in the sample. Alginate hydrogels were used to confine the light in the lower level of the chip, thus improving their optical performance.
Results and discussion Physicochemical characterization of alginate matrices Alginate hydrogels were the key elements of the optofluidic system described in this work, acting, at the same time, as entrapping matrix for bacteria, diffusion barrier and optical elements for light confining. For this reason, physicochemical properties of a wide variety of calcium alginate hydrogels were studied in order to select a gel that fulfils the required properties, concretely long-term stability, high homogeneity, high transmittance in the visible range, high diffusion rates and microbial entrapping capacity without compromising bacterial integrity or functionality. Alginate is one of the polymers most widely employed in the immobilization of living organisms due to its biocompatibility and soft polymerization conditions.20–22 Furthermore, alginate hydrogels are mostly transparent and present a refractive index of 1.37–1.40 (ref. 30) which ensures light confinement when surrounded by water. However, alginate solutions above 4% IJw/v) present a brown-orange color, which
Fig. 3 Picture of the trench-based optofluidic system (A). Schematic illustration of optofluidic system sensing principle (B).
This journal is © The Royal Society of Chemistry 2015
Lab Chip, 2015, 15, 1717–1726 | 1721
Paper
may interfere with optical measurements. For this reason, transmittance spectra of gelled and non-gelled alginate samples between 0.5 and 2.0% alginate were obtained using the optical setup described in the experimental section. As illustrated in Fig. 4A, non-gelled alginate samples with a concentration up to 2% IJw/v) presented high transmittances (above 90% transmittance) from 400 to 900 nm, independent of the concentration. After gelling (Fig. 4B), although maintaining values above 80% in most of the visible spectrum (from 500 to 900 nm), transmittance decreased below 70% in the region between 350 and 500 nm. As before, similar transmittance spectra were obtained independent of the alginate concentration in the gel. The reason for that wavelength-dependent transmittance drop after gelling is still controversial. From all possible causes, dispersion seemed to be the most plausible one. According to this, alginate hydrogel may be able to selectively disperse the light in this wavelength range, something that could not be done before gelling. Light dispersion may be due to several causes, for instance the formation of small-size scattering centers. In this sense, one possibility may be the formation of small porous structures, below 500 nm,
Lab on a Chip
which, acting as scattering centers, may disperse the light below 450 nm, without interacting with longer wavelengths. According to these results, optical properties of alginate hydrogels may vary during gel formation, interfering with optical measurements. Thus, although high transmittances were recorded in most of the visible light spectrum, it may be ensured that optical properties of the hydrogel were not changing during phenolic compound determination. To this end, several hydrogels containing different alginate and calcium concentrations were prepared. Gel formation was monitored over time by following light intensity changes at 450 nm (reporting light scattering center formation). Fig. 4C and E illustrate light intensity changes due to gel formation for samples containing different calcium and alginate concentrations, respectively. In most of the cases, light intensity randomly decreased with time until reaching a plateau at around 10 min of incubation. Hence, after 10 min, hydrogels should be optically stable and should not interfere with optical measurements. Thus, 10 min was selected as the ideal incubation time for gel formation. However, after 10 min of gelation, not all hydrogels exhibited texture and homogeneity
Fig. 4 Physicochemical characterization of alginate hydrogels. Transmittance spectra of non-gelled (A) and gelled (B) alginate samples. Gelling kinetic plots (at 450 nm) for samples containing 0.5% alginate and calcium chloride concentrations from 0 to 200 mM (C) and alginate concentrations from 0.5 to 2.0% and 100 mM calcium chloride (E). Intensity magnitude (at 450 nm) of 0.5% alginate hydrogels containing different calcium chloride concentrations (D) and hydrogels containing 100 mM calcium chloride and alginate concentrations from 0.5 to 2.0% (F) after 10 min gelling. Error bars represent the standard deviation from the mean (95% confidence, n = 3). See experimental section for details.
1722 | Lab Chip, 2015, 15, 1717–1726
This journal is © The Royal Society of Chemistry 2015
Lab on a Chip
suitable for the final application. For this reason, an in-depth analysis of gel formation kinetics was performed. Considering calcium concentration, both visual inspection and gel formation kinetics (Fig. 4C) confirmed that a minimum concentration of calcium around 20 mM was necessary for hydrogel formation (below 20 mM light intensity remained constant for the duration of the experiment). For calcium concentrations higher than 20 mM, hydrogel texture tremendously varied. Samples containing between 20 and 40 mM calcium present viscous fluid texture, quite far from that of a gel. Above 100 mM it was possible to obtain homogeneous hydrogels. This observation was in accordance with light intensity magnitude recorded in the plateau. As shown in Fig. 4D, light intensity magnitude in the plateau decreased when increasing calcium concentration until stabilizing at around 100 mM, where homogeneous and stable hydrogels were obtained. Thus, intensity magnitude in the plateau may be used as a quantitative indicator of the degree of gel formation. According to this parameter, 100 mM calcium concentrations were necessary to obtain homogeneous and stable gels. In the case of alginate, no significant differences were observed when comparing gel formation kinetics (Fig. 4E) or light intensity magnitude in the plateau (Fig. 4F) of hydrogels containing alginate concentrations between 0.5 and 2.0%. However, visual inspection of these hydrogels confirmed that those containing higher concentrations of alginate were more rigid and less homogeneous, coinciding with already reported data.31 This result suggested the use of low alginate concentration (below 1%) in this application. On the other hand, optimal performance of the system also required suitable diffusion of molecules through the alginate matrix until reaching bacteria, which may determine the response time of the system. To this end, diffusion kinetics through hydrogel matrices containing different alginate concentrations (from 0.5 to 2.0%) was evaluated by following the protocol detailed in the experimental section. Ferricyanide, with a wide absorption band at 420 nm, was used as the model molecule. Diffusion kinetic results are illustrated in Fig. 5. In all cases, a progressive absorbance increase at 420 nm was recorded over time (Fig. 5A). To ensure that this was due to ferricyanide (and not to changes in the gel structure), absorbance spectra were recorded along the experiment, taking the spectrum of the hydrogel before ferricyanide inoculation as the reference. Fig. 5B shows how the ferricyanide absorption band, with a maximum at 420 nm, increased with time due to diffusion. Two different diffusion tendencies were observed depending on the alginate concentration. Low concentrations (below 1% alginate) present slow diffusion patterns and ferricyanide required around 250 s to achieve an absorbance magnitude of around 1. On the other hand, above 1% alginate, ferricyanide diffuse faster, requiring less than 150 s to cross the hydrogel. The reason for that may be again the heterogeneity of high alginate concentration matrices, combining gelled and non-gelled areas that favor fast ferricyanide diffusion.
This journal is © The Royal Society of Chemistry 2015
Paper
Fig. 5 Ferricyanide diffusion plots in alginate hydrogels. Absorbance variation (at 420 nm) with time illustrating ferricyanide diffusion into 0.5, 0.7, 1.0, 1.5 and 2.0% alginate hydrogels (A). Representation of the variation of absorption spectra of 0.5% alginate hydrogels during ferricyanide diffusion (B). The alginate matrix spectrum after 10 min of incubation is used as the reference. Error bars represent the standard deviation from the mean (95% confidence, n = 3).
According to experimental data, alginate hydrogels with 0.5% alginate and 100 mM calcium were selected to be implemented in the trench-based optofluidic system for being the most homogeneous, repetitive ones and also presenting excellent optical and diffusional properties. Escherichia coli PtomAB Amp 100 characterization. Phenolic compound determination was based on the use of genetically modified E. coli bacteria capable to metabolize monoand bi-aromatic phenolic compounds to 2-HMS, a degradation product with a wide absorption band around 419 nm (Fig. S4†). Thus, in this case, phenolic compounds could be selectively determined by monitoring absorption changes without the need of any additional chemical/biological compound. Bacterial activity was evaluated as detailed in the experimental section, using catechol as the model molecule. Four bacterial concentrations, from 0 to 1 × 109 CFU mL−1, and eight catechol concentrations (between 0 and 8% w/v) were analyzed for assay optimization (Fig. 6A–C). In all cases, absorbance magnitude increased with time for the duration of the experiment due to catechol degradation/metabolism mechanisms. According to experimental data, degradation kinetics depended on bacterial concentration. Below 1 × 107 CFU mL−1, similar degradation kinetics was obtained independent of bacterial concentration or even the presence or the absence of bacteria (Fig. S5†). Considering this, catechol degradation kinetics below 1 × 107 CFU
Lab Chip, 2015, 15, 1717–1726 | 1723
Paper
Lab on a Chip
Fig. 6 Bacterial metabolism characterization. Representation of the variation of absorbance at 405 nm with time for several catechol samples containing 0 CFU mL−1 (A), 1 × 108 CFU mL−1 (B) or 1 × 109 CFU mL−1 (C) genetically modified bacteria. Each plot is the result of three independent experiments. Error bars represent the standard deviation (95% confidence). Degradation kinetic plots, where the slope (absorbance at 405 nm variation with time) is represented against catechol concentration for samples containing several E. coli concentrations (D). Slope was determined by considering the first 10 min of the absorbance versus time plot.
mL−1 may be attributed to catechol photo-oxidation. Above 1 × 108 CFU mL−1, faster and bacterial concentrationdependent catechol degradation kinetics was recorded (Fig. 6B and C), suggesting a microbial-mediated degradation mechanism. These results indicated that a minimum concentration of 1 × 108 CFU mL−1 was necessary for microbialmediated selective determination of catechol. Catechol concentration was determined through catechol degradation kinetic plots using the slope of the absorbance versus time graphics (first 10 min) as the analytical signal. Degradation kinetic plots were advantageous for allowing quick (10 min) and reliable quantification of catechol concentration without interference of biomass scattering. As shown in Fig. 6D, the slope magnitude of the degradation curve increased with increasing catechol concentration until 1%. After that, it started decreasing to values similar to those recorded by samples without bacteria (around 8% catechol). According to these data, catechol was toxic for these genetically modified bacteria which started dying when catechol concentration in the sample exceeded a concentration magnitude above 1% catechol. Thus, with this method, reagentless determination of catechol was only possible below 1% concentrations.
Determination of catechol in water samples using the microbial trench-based optofluidic system Before catechol determination in water samples, microbial/ hydrogel formation interference in optical measurements and bacterial viability in the hydrogel matrix were analyzed. In the evaluation of gel formation interference, light intensity
1724 | Lab Chip, 2015, 15, 1717–1726
variation during microbial/alginate gel formation was analyzed and compared with alginate gels without bacteria. Fig. 7A illustrates light intensity variation with time during 0.5% alginate hydrogel formation containing 1 × 109 CFU mL−1. As shown, light intensity decreased progressively during gel formation until stabilizing after 8 min gelling, coinciding with previous data. In order to evaluate microbial stability in the hydrogel, microbial/alginate hydrogels prepared as before were stored in the fridge (4 °C) in Ringer's medium or dried for 19 days. Bacterial viability was regularly evaluated by cell counting. No significant differences in the number of viable cells were obtained after 19 days for wet samples, whereas in dried samples the number of viable bacteria drastically decreased with time (Fig. 7B). Hence, alginate hydrogels stored in wet conditions are suitable for long-term microbial entrapment. For the reagentless determination of phenolic compounds in water samples, 0.5% alginate matrices containing 1 × 109 CFU mL−1 genetically modified bacteria were prepared in the trench-based optofluidic system and analyzed as detailed in the experimental section. After 10 min of hydrogel formation, the trench-based optofluidic system was embedded and LED intensity was adjusted (with a resistance) to provide an initial intensity around 1100–1200 counts. LED stability was evaluated before measurement by monitoring its intensity for around 1 h after addition of water in the optofluidic system. In all cases, variability measurements below 1% were obtained (Fig. S6†). Water samples containing catechol concentrations from 0 to 2% were then inoculated in the system and absorbance variation was monitored for 1 h with the electronic control system already described. In Fig. 7C,
This journal is © The Royal Society of Chemistry 2015
Lab on a Chip
Paper
Fig. 7 Representation of the variation of light intensity magnitude at 450 nm during gel formation for hydrogels containing 0.5% alginate and 1 × 109 CFU mL−1 E. coli (A). Representation of the variation of the number of viable cells with time for microbial/hydrogels stored at 4 °C either wet or dried (B). Representation of the variation of absorbance magnitude with time for several catechol samples containing 1 × 109 CFU mL−1 genetically modified bacteria (C). Degradation kinetic plots, where the slope (absorbance variation with time) is represented against catechol concentration (D). Slope was determined by considering the first 10 min of the absorbance versus time plot. Each plot is the result of three independent experiments. Error bars represent the standard deviation (95% confidence).
absorbance magnitude, taking water as the reference, is represented with time for the five catechol concentrations under study. As before, microbial-mediated catechol degradation (Fig. 7D) was dependent on catechol concentration and allowed for reagentless and quantitative determination of catechol concentration in a wide range of concentration (up to 1% catechol). In summary, this microbial trench-based optofluidic system represents a cheap, simple and robust alternative for quick (10 min), sensitive, reagentless and quantitative determination of catechol in water samples. The robustness, simplicity and reagentless nature of the bioassay (for the use of genetically modified microorganisms) open the possibility for in situ determination of phenolic compound contamination without the need of sample transport or pretreatment.
Conclusions Reagentless determination of catechol was achieved with a microbial trench-based optofluidic system controlled by a miniaturized portable electronic circuit adapted from a lockin-amplifier, capable to monitor light intensity changes, subtracting, at the same time, ambient light interference. An alginate hydrogel was incorporated into the trench for microbial entrapment and light confinement from the LED emitter to the photodetector, both located on opposite sides of the trench. Integrated optical elements such as PMMA
This journal is © The Royal Society of Chemistry 2015
convergent lenses, floating lightguides and fluidic elements (i.e. inlet, outlet, channels) were also included in the PMMA optofluidic structure for optimal optical transduction and fluid management. Alginate hydrogels containing 0.5% alginate and 100 mM calcium demonstrated suitable optical (high transmittance in the visible range), structural (high homogeneity and stability) and diffusional (high diffusion of small molecules) properties to be implemented in the optofluidic system. Additionally, their properties did not change by incorporation of genetically modified bacteria (up to 1 × 109 CFU mL−1) capable of selectively metabolizing catechol (used as model molecule). With this system, catechol concentrations from 0 to 1% IJw/v) could be determined without the need of additional reagents after only 10 min of reaction through the formation of a colored bacterial degradation product. This miniaturized, portable, robust and simple trenchbased optofluidic system, providing information on relevant contaminants after short incubation times and without the need of additional reagents, opens the possibility of in situ determination of environmental pollutants, thus mitigating the effects of accidental spills.
Acknowledgements This work was supported in part by the Spanish R & D National Program (MINECO Project TEC2011-29045-C04-01/04). X. M.-B. also wants to acknowledge the “Ramón y Cajal” program
Lab Chip, 2015, 15, 1717–1726 | 1725
Paper
from the Spanish Government. P. G.-G. is grateful to the Spanish MEC for the award of a research studentship from FPI Program.
References 1 J. Michalowicz and W. Duda, Pol. J. Environ. Stud., 2007, 16, 347–362. 2 S. Wasi, S. Tabrez and M. Ahmad, Environ. Monit. Assess., 2013, 185, 2585–2593. 3 United States Environmental Protection Agency, http://www. epa.gov/. 4 International Organization for Standardization, http://www. iso.org/. 5 J. Martin, D. Camacho-Munoz, J. L. Santos, I. Aparicio and E. Alonso, Anal. Bioanal. Chem., 2014, 406, 3709–3716. 6 M. Park, S. L. Tsai and W. Chen, Sensors, 2013, 13, 5777–5795. 7 T. T. Xu, D. M. Close, G. S. Sayler and S. Ripp, Ecol. Indic., 2013, 28, 125–141. 8 H. J. Shin, Appl. Microbiol. Biotechnol., 2012, 93, 1895–1904. 9 C. Liu, D. M. Yong, D. B. Yu and S. J. Dong, Talanta, 2011, 84, 766–770. 10 X. M. Liu, K. J. Germaine, D. Ryan and D. N. Dowling, Sensors, 2010, 10, 1377–1398. 11 L. A. Su, W. Z. Jia, C. J. Hou and Y. Lei, Biosens. Bioelectron., 2011, 26, 1788–1799. 12 X. Y. Zhao and T. Dong, Int. J. Environ. Res. Public Health, 2013, 10, 6748–6763. 13 S. K. Yoo, J. H. Lee, S. S. Yun, M. B. Gu and J. H. Lee, Biosens. Bioelectron., 2007, 22, 1586–1592. 14 Y. C. Seow, S. P. Lim and H. P. Lee, Lab Chip, 2012, 12, 3810–3815.
1726 | Lab Chip, 2015, 15, 1717–1726
Lab on a Chip
15 Y. C. Seow, S. P. Lim and H. P. Lee, Microfluid. Nanofluid., 2011, 11, 451–458. 16 Y. Yang, L. K. Chin, J. M. Tsai, D. P. Tsai, N. I. Zheludev and A. Q. Liu, Lab Chip, 2012, 12, 3785–3790. 17 I. K. Dimov, G. Kijanka, Y. Park, J. Ducree, T. Kang and L. P. Lee, Lab Chip, 2011, 11, 2701–2710. 18 T. M. O'Connell, D. King, C. K. Dixit, B. O'Connor, D. Walls and J. Ducree, Lab Chip, 2014, 14, 3629–3639. 19 S. N. Pawar and K. J. Edgar, Biomaterials, 2012, 33, 3279–3305. 20 K. Y. Lee and D. J. Mooney, Prog. Polym. Sci., 2012, 37, 106–126. 21 W. R. Gombotz and S. F. Wee, Adv. Drug Delivery Rev., 2012, 64, 194–205. 22 S. M. Selimoglu and M. Elibol, Crit. Rev. Biotechnol., 2010, 30, 145–159. 23 T. Andersen, J. E. Melvik, O. Gaserod, E. Alsberg and B. E. Christensen, Carbohydr. Polym., 2014, 99, 249–256. 24 C. K. Kuo and P. X. Ma, Biomaterials, 2001, 22, 511–521. 25 S. T. Parker, P. Domachuk, J. Amsden, J. Bressner, J. A. Lewis, D. L. Kaplan and F. G. Omenetto, Adv. Mater., 2009, 21, 2411–2415. 26 A. Jain, A. H. J. Yang and D. Erickson, Opt. Lett., 2012, 37, 1472–1474. 27 M. Choi, J. Choi, S. Kim, S. Nizamoglu, S. K. Hahn and S. H. Yun, Nat. Photonics, 2013, 7, 987–994. 28 L. A. Edghill, A. D. Russell, M. J. Day and J. R. Furr, J. Appl. Microbiol., 1999, 87, 91–98. 29 D. C. Yee, J. A. Maynard and T. K. Wood, Appl. Environ. Microbiol., 1998, 64, 112–118. 30 O. Esteban, F. Marva and J. C. Martinez-Anton, Opt. Mater., 2009, 31, 696–699. 31 Y. Maki, K. Ito, N. Hosoya, C. Yoneyama, K. Furusawa, T. Yamamoto, T. Dobashi, Y. Sugimoto and K. Wakabayashi, Biomacromolecules, 2011, 12, 2145–2152.
This journal is © The Royal Society of Chemistry 2015
Copyright of Lab on a Chip - Miniaturisation for Chemistry & Biology is the property of Royal Society of Chemistry and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
