Anal. Chem. 2001, 73, 1296-1300
Microseparation Chips for Performing Multienzymatic Dehydrogenase/Oxidase Assays: Simultaneous Electrochemical Measurement of Ethanol and Glucose Joseph Wang,* Madhu Prakash Chatrathi, and Baomin Tian
Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, New Mexico 88003
This report describes a new “lab-on-a-chip” protocol integrating on-line precolumn biocatalytic reactions of multiple (oxidase and dehydrogenase) enzymes and substrates with effective capillary electrophoresis microseparations and amperometric detection. The operation of the new oxidase/dehydrogenase reaction/separation microchip is illustrated for the simultaneous measurement of glucose and ethanol in connection to the corresponding glucose oxidase and alcohol dehydrogenase reactions, respectively. The enzymatic reactions generate hydrogen peroxide and NADH species that are separated (on the basis of their different charges) and detected amperometrically at the end-column thick-film detector. A driving voltage of 2000 V results in peroxide and NADH migration times of 74 and 230 s, respectively. Operating the goldcoated carbon detector at +1.0 V allows simultaneous anodic detection of both reaction products. Factors influencing the reaction, separation, and detection processes are examined and optimized. The applicability of the new multienzyme assay to wine samples is illustrated. Microfabricated fluidic devices, integrating the sample preparation process and the measurement step onto microchip platforms, are of considerable recent interest.1,2 Such technology offers great promise for developing versatile and miniaturized analytical microsystems with a high degree of automation, rapid analysis times, and negligible consumption of reagents. Separation microchips have been combined with chemical derivatization reactions3,4 but rarely to enzymatic processes. Because of their specificity and catalytic (amplification) properties, enzymes have found widespread use in bioanalysis. Similarly, the use of enzymes can impart higher selectivity into capillary electrophoresis (CE) microchips and expand their scope toward redox-inactive substrates. Regnier’s group5 described the coupling of conventional CE systems with on-capillary biocatalytic reactions. Yet, such systems lack the (1) Figeys, D.; Pinto, D. Anal. Chem. 2000, 71, 330A. (2) Jakeway, S.; de Mallo, A. J.; Russell, E. L. Fresenius J. Anal. Chem. 2000, 366, 525. (3) Jacobson, S. C.; Hergenroder, R.; Moore, A. W., Jr.; Ramsey, J. M. Anal. Chem. 1994, 66, 4127. (4) Fluri, K.; Fitzpatrick, G.; Chiem, N.; Harrison, D. J. Anal. Chem. 1996, 68, 4285. (5) Wu, D.; Regnier, F. Anal. Chem. 1993, 65, 2029.
1296 Analytical Chemistry, Vol. 73, No. 6, March 15, 2001
versatility of microfluidic chip devices, particularly when pre- or postcolumn enzymatic reactions are concerned. Ramsey’s group described microchip separation devices for performing enzyme (galactosidase or acetylcholinesterase) inhibition assays.6,7 Very recently, we demonstrated a micromachined reaction/separation chip for the simultaneous measurement of glucose, ascorbic acid (AA), and uric acid (UA), based on the coupling of the glucose oxidase (GOx) enzymatic reaction and electrophoretic separation of the peroxide product from the common AA and UA interferences.8 However, to our knowledge, there are no reports on the simultaneous use of multiple enzymes on a single microchip platform and of on-chip dehydrogenase-based reactions. In this paper, we exploit the versatility of microfluidic devices for carrying out simultaneously multiple oxidase- and dehydrogenase-based precolumn reactions, in connection with a fast electrophoretic separation and amperometric detection of the liberated NADH and hydrogen peroxide. There are over 200 dehydrogenases and 100 oxidases. Many of these enzymes catalyze specifically the reactions of clinically and biotechnologically important analytes (e.g., lactate, glucose, cholesterol, amino acids, urate, pyruvate, glutamate, alcohol, and hydroxybutyrate) to generate electrochemically detectable (NADH and hydrogen peroxide) products. The new concept of on-chip oxidase/dehydrogenase bioassays is examined and demonstrated in the following sections using a simple and yet important (glucose/alcohol) model system. The simultaneous measurement of sugars and alcohols is of great importance for the biotechnology and food industries. The new on-chip assay of glucose and alcohol relies on their precolumn reactions with glucose oxidase and alcohol dehydrogenase [(ADH), and its NAD+ cofactor], respectively, and electrophoretic separation and amperometric detection of the resulting peroxide and NADH products. The layout of this glucose/alcohol microchip, shown in Figure 1, is similar to that used by Ramsey for precolumn chemical reactions.3 Both enzymatic reactions are thus carried out simultaneously within the same precolumn reactor (having a wider channel than the separation one, to allow sufficient reaction time). Since the reaction products possess a different charge (6) Hadd, A. G.; Raymond, D. E.; Halliwell, J. W.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1997, 69, 3407. (7) Hadd, A. G.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1999, 71, 5206. (8) Wang, J.; Chatrathi, M. P.; Tian, B.; Polsky, R. Anal. Chem. 2000, 72, 25142518. 10.1021/ac001205t CCC: $20.00
© 2001 American Chemical Society Published on Web 02/08/2001
Figure 1. Layout of the reaction/separation/detection microchip with precolumn oxidase/dehydrogenase enzymatic reactions, electrophoretic separation, and anodic detection of the peroxide/NADH products. See text for exact dimensions and details.
[neutral (peroxide) vs negative (NADH)] and are readily oxidized, they can be separated and detected at moderate anodic potentials using the same working electrode amperometric detector. Other attractive advantages of electrochemical detectors for CE microchips, including high sensitivity and selectivity, compatibility with micromachining technologies, and miniaturization of both the detector and control instrumentation, were discussed earlier.8-10 While the attractive on-chip combination of oxidases and dehydrogenases is demonstrated below for the testing of glucose and ethanol, other combinations of these enzymes and/or parallel (multichannel) operations can be used for monitoring simultaneously numerous analytically important substrates. Operating conditions and benefits of the new oxidase/dehydrogenase reactor/separation microchip are reported below. EXPERIMENTAL SECTION Reagents. ADH, glucose, β-nicotinamide adenine dinucleotide [sodium salt (β-NAD+)], GOx, and β-nicotinamide adenine dinucleotide, reduced form [sodium salt (β-NADH)], were obtained from Sigma. Ethanol was purchased from Quantum Chemical Co. (Tuscola, IL). The gold atomic absorption standard solution (1000 mg/L) was purchased from Aldrich. All chemicals were used without further purification. The running buffer (pH 7.8) consisted of 20 mM phosphate buffer. Stock solutions were prepared daily and filtered with a 0.45-µm filter (Gelman Acrodisc). Sample solutions were prepared by diluting the corresponding stock solutions with the electrophoresis buffer. The wine samples were purchased at a local store and diluted in the electrophoresis buffer. Apparatus. The glass chip used in this study was fabricated by Alberta Microelectronic Co. (AMC, model MC-BF4-001, Edmonton, ON, Canada) based on our custom design. The layout is different from that employed previously10 in that it permits precolumn reactions (Figure 1). The chip consisted of a reagent (GOx + ADH + NAD+) reservoir, a sample (glucose + ethanol) reservoir, a running buffer (RB) reservoir, and an waste (B) reservoir. A reaction chamber (200 µm wide and 3.6 mm long) was connected through 50-µm-wide, 20-µm-deep channels to the reagent and sample reservoirs at one side and to a four-way injection cross at the other side. The injection cross was followed by a 74-mm-long, 50-µm-wide separation microchannel. A Plexiglas holder was fabricated for holding the separation chip and housing the detector and the reservoirs. A short pipet (9) Woolley A. T.; Lao K.; Glazer A. N.; Mathies R. A. Anal. Chem. 1998, 70, 684. (10) Wang, J.; Tian, B.; Sahlin, E. Anal. Chem. 1999, 71, 5436.
tip was inserted into each of the holes on the glass chip for solution contact between the channel on the chip and corresponding reservoir on the chip holder. The amperometric detector was located in the detection reservoir (at the channel outlet side) and consisted of a Ag/AgCl wire reference electrode, a platinum wire counter electrode, and a gold-modified screen-printed carbon working electrode. The working electrode was placed at the channel outlet; the distance between its surface and the channel outlet (50 µm) was controlled by a plastic screw and a thin-layer spacer. Electrical contact with the solutions was achieved by placing platinum wires into each of the reservoirs. A “homemade” power supply, containing multiple voltage terminals, was used for applying the selected driving voltage (between 0 and +4000 V) to a given reservoir and for switching between the “reaction/ injection” and “separation” modes. The screen-printed working electrodes were fabricated with a semiautomatic printer (model TF 100, MPM, Franklin, MA). The Acheson carbon ink Electrodag 440B (49AB90) (Acheson Colloids, Ontario, CA) was used for printing electrode strips. Details of the printing processes were described previously.10 A working electrode strip (0.4 in. × 1.333 in.) consisted of a carbon-ink line and its silver contact, printed onto the ceramic substrate. The active working electrode area (0.30 × 2.50 mm) was defined by a layer of insulator. The carbon working electrode was coated with gold by applying a square-wave potential pulse between -0.2 and +0.75 V vs Ag/AgCl (with pulse width of 0.6 s) for 20 min using a 0.1 M NaCl/1.5% HCl solution containing 300 ppm Au(III). Amperometric detection was performed with an electrochemical analyzer 621 (CH Instruments, Austin, TX) connected to a Pentium 166-MHz computer with 32 MB RAM. The electropherograms were recorded with a time resolution of 0.1 s while applying +1.0 V detection potential (vs Ag/AgCl wire). Sample injections were performed after stabilization of the baseline. To minimize the background noise associated with the detection of NADH, the raw data of electropherograms were digitally filtered using a 17-point least-squares smoothing. Electrophoresis Procedure. To perform separations, the “reagent” reservoir was filled with 80 µL of the GOx/ADH/NAD+ reagent solution (usually 10 units/mL GOx, 10 units/mL ADH, and 10 mM NAD+) while the “sample” reservoir was filled with 80 µL of glucose/ethanol mixture. The other two “buffer” reservoirs were filled with 75-µL volume of the electrophoresis buffer. The detection/waste reservoir at the channel outlet side was filled with the electrophoresis buffer solution. After an initial filling of the reaction chamber, a potential of +2500 V was applied to both sample and reagent reservoirs for 3 s with the detection reservoir grounded and the other reservoirs floating. The use of the same voltage assured a constant mixing volume ratio. The biocatalytic reactions of glucose and ethanol with the GOx and ADH enzymes, respectively, occurred upon mixing the reagent and sample streams. Separation of these products was usually performed by applying +2000 V to the running buffer reservoir with the detection reservoir grounded and the other reservoirs floating. To minimize adsorption of the enzymes onto the walls of the capillary (while retaining their biocatalytic activity), the pH of the buffer (7.8) was chosen to be above the pI of the enzymes used. Furthermore, the channels were treated between groups of runs Analytical Chemistry, Vol. 73, No. 6, March 15, 2001
1297
Figure 2. Electropherograms for (A) a mixture containing 5 × 10-5 M hydrogen peroxide (a) and 5.0 × 10-4 M NADH (b); (B) 4.0 × 10-4 M glucose; (C) 2.0 × 10-2 M ethyl alcohol; and (D) a mixture of 4.0 × 10-4 M glucose (a) and 2.0 × 10-2 M ethanol (b). Gold-plated screen-printed electrode held at +1.0 V. Injection potential, +2500 V; separation potential, +2000 V; injection time, 3 s. Electrophoresis buffer, 20 mM phosphate buffer (pH 7.8). Reagent solution, 10 units/ mL GOx (B), 10 units/mL ADH, and 1 × 10-2 M NAD+ (C), and 10 units/mL GOx, 10 units/mL ADH, and 1 × 10-2 M NAD+ (D).
by rinsing with 0.1 M sodium hydroxide and deionized water for 20 and 5 min, respectively. The initial loading (of the arms of the reagent and sample reservoirs) and filling of the reaction chamber was accomplished by applying a voltage of +2500 V for 60 s to the reagent and sample reservoirs (with the detection reservoir grounded and other reservoirs floating). All experiments were carried out at room temperature. Safety Considerations. The high-voltage power supply (and related electrical connections) should be handled with extreme care to avoid electrical shock. RESULTS AND DISCUSSION Microchip CE enzymatic assays can be carried using three different forms, involving the use of precolumn, postcolumn, or on-column reactions. The former is particularly attractive for performing mixed oxidase/dehydrogenase reactions, in view of the different charge of the reaction products [neutral (peroxide) vs negative (NADH)]. Using the layout of Figure 1, the new bioassay thus relies on precolumn reactions of glucose and ethanol with GOx and ADH/NAD+, respectively, followed by electrophoretic separation and detection of the peroxide and NADH products. Figure 2A displays a typical electropherogram for a sample containing 5 × 10-5 M hydrogen peroxide and 5 × 10-4 M NADH. Both compounds are readily oxidized and can be detected by holding the gold-coated carbon at +1.0 V. As expected, the mobility of the anionic NADH is lower than that of the neutral peroxide species. Well-defined peaks, with favorable signal-to-noise characteristics, are thus observed at migration times of 74 (peroxide) and 230 (NADH) s. Electropherograms B and C (of Figure 2) display the corresponding response for injections of glucose and ethanol solutions in connection with the GOx and ADH/NAD+ precapillary reactions, respectively. Such injections of the individual substrate solutions result in well-defined oxidation peaks, at the same migration times as those observed for the corresponding pure products (A). Note the small negative peak (at ∼130 s) associated with the change of ionic strength (within 1298 Analytical Chemistry, Vol. 73, No. 6, March 15, 2001
Figure 3. Hydrodynamic voltammograms for a mixture of 1.0 × 10-3 M glucose (A) and 2.0 × 10-2 M ethanol (B). Reagent solution, 10 units/mL GOx, 10 units/mL ADH, and 3 × 10-2 M NAD+. Other conditions, as in Figure 2.
the NAD+ zone). A typical electropherogram for a glucose/ethanol sample mixture (in the presence of both enzymes) is shown in Figure 2D. This electropherogram indicates convenient simultaneous measurement of glucose (a) and ethanol (b) with a total time of ∼4 min (using a separation potential of +2000 V). Note the similarity of the migration times (A vs D). Yet, some tailing of the glucose peak (associated with a change in the ionic strength) is observed in the presence of ethanol/ADH/NAD+. High overvoltages are commonly required for the anodic detection of hydrogen peroxide and NADH at conventional electrodes.11 Hydrodynamic voltammograms were constructed for assessing the effect of the detector potential. Figure 3 shows such hydrodynamic voltammograms for glucose (A) and ethanol (B), obtained by changing the potential of the screen-printed working electrode over the +0.4 to +1.20 V range, while injecting a mixture of these substrates. The peroxide and NADH products display similar voltammetric profiles, with oxidation starting above +0.6 V, a gradual rise in the current between +0.6 and +1.0 V, and a leveling off thereafter. All subsequent work employed a potential of +1.0 V that offered the most favorable signal/background characteristics. Note also the similarity of the limiting currents, despite of the different levels of hydrogen peroxide and NADH (1 vs 20 mM, respectively). The influence of the separation voltage upon the multienzyme assay is examined in Figure 4, using a 4 × 10-4 M glucose (a)/2 × 10-2 M alcohol (b) mixture. As expected, increasing the voltage between 1000 (A) and 3000 (B) V, in steps of 500 V, decreases the migration times for both reaction products. While pushing the products faster through the separation channel, higher field strengths also shorten the reaction time of the reactants and hence impair the efficiency of the enzymatic reactions. (Note that the enzymatic reactions proceed also in the separation column.) The driving voltage has a different effect upon the plate number for glucose and ethanol. While the ethanol plate number decreases dramatically from 4475 to 640 upon raising the voltage from 1000 to 3000 V, that of glucose changes only mildly (from 1800 to 1200) over the same range. Such trends reflect the effect of the voltage upon both the enzymatic reaction and the separation process and, hence, the different rates of the GOx and ADH reactions (with the latter being the slower one). In the absence of any preceding (11) Wang, J. Analytical Electrochemistry; Wiley: New York, 2000; Chapter 6.
Figure 4. Influence of separation voltage upon the response for a mixture containing 4 × 10-4 M glucose (a) and 2.0 × 10-2 M ethanol (b). Separations performed using (A) +1000, (B) +1500, (C) +2000, (D) +2500, and (E) +3000 V. Other conditions, as in Figure 2.
Figure 5. Influence of the concentration of (A) GOx, (B) ADH, and (C) NAD+ on the response. Sample mixture containing 1.0 × 10-3 M glucose (A) and 2.0 × 10-2 M ethanol (B, C). NAD+ was fixed at 3 × 10-2 M for (B) while ADH was fixed at 10 units/mL. Other conditions, as in Figure 2.
reaction, the plate number is expected to increase with the voltage. The initial baseline slope, observed at voltages higher than 2000 V, indicates an incomplete isolation from such high fields. Such change in the background is common to the present end-column detection system.10 A voltage of 2000 V was used for all subsequent work, as it provided the most favorable balance between speed, sensitivity, resolution, and isolation from the detection circuitry. The successful operation of the new enzyme/CE microchip requires proper attention to the levels of the individual bioreagents. Figure 5 examines the influence of the GOx (A) and ADH (B) activities, as well as of the NAD+ concentration (C), upon the response for a glucose/ethanol mixture. The glucose peak increases rapidly with GOx activity up to ∼5 units/mL, then more slowly, and nearly levels off above 10 units/mL (A). The alcohol signal increases very rapidly with the ADH activity between 5 and 10 units/mL, and with the NAD+ concentration up to 2 × 10-2 M, and levels off at higher levels of the enzyme and cofactor. All subsequent work employed GOx and ADH activities of 10 units/ mL and a NAD+ concentration of 1 × 10-2 M. The latter was selected as optimal in view of the larger baseline slope at higher NAD+ concentrations. When coupled with the nanoliter volume of the reagent solution, such levels correspond to a negligible consumption over repetitive runs. We also examined the effect of NAD+ upon the detection of glucose (i.e., possible cross reaction with the peroxide product) and found no such interaction even at an excess (3 × 10-2 M) level of NAD+ (not shown).
Figure 6. Precision of repetitive glucose/ethanol runs. Electropherograms for a mixture containing 4.0 × 10-4 M glucose and 2.0 × 10-2 M ethanol. The alternatively recorded electropherograms are shown. Other conditions, as in Figure 1.
Good precision is another attractive feature of the new bienzymatic/separation microchip protocol. Figure 6 displays a series of six representative electropherograms obtained during repetitive assays of a 7 × 10-4 M glucose - 2 × 10-2 M ethanol solution mixture. These measurements are a part of a series of 12 successive runs that resulted in a highly reproducible response. (Such repetitive runs were performed without an intermittent NaOH treatment.) Relative standard deviations (RSDs) of 3.7 and 3.1% were estimated for the peak currents for glucose and ethanol, respectively. The migration times are also very reproducible with RSD of less than 0.5%. Such good precision reflects the high reproducibility of the reaction/separation/detection processes and indicates a negligible electrode fouling or enzyme adsorption onto the channel walls. Note that dehydrogenase-based biosensors are often prone to a gradual surface passivation (due to accumulation of NADH oxidation products).12 Apparently, the ultrasmall sample volumes of the microchip platform, and the corresponding minimal amount of NADH, lead to a negligible accumulation of reaction products and, hence, eliminate this fouling problem. Similar advantages were illustrated also for other deactivating compounds, e.g., phenolic contaminants.13 The design of the present microsystem permits rapid (5-10 s) replacement of the detector strip in case of surface fouling. The amperometric detector displays a well-defined concentration dependence. Such dependence was examined by recording the electropherograms for sample mixtures containing increasing levels of glucose and ethanol in steps of 2 × 10-4 and 5 × 10-3 M, respectively (not shown; conditions as in Figure 2). Defined peaks, proportional to the analyte concentration, were observed. The resulting calibration plots were highly linear up to 8 × 10-4 M glucose (sensitivity of 13.6 nA/mM) and up to 2.5 × 10-2 M ethanol (sensitivity, 0.2 nA/mM; correlation coefficients, 0.998 and 0.955, respectively). The sensitive response is coupled to a low noise level and, hence, to detection limits of 1.0 × 10-5 M glucose and 5.4 × 10-4 M ethanol (based on three standard deviations of the noise in assays of a mixture containing 1 × 10-4 M glucose and 2.5 × 10-3 M ethanol; not shown; conditions as in Figure 2). The ability to measure rapidly and simultaneously glucose and ethanol is of obvious relevance to wine analysis. Such capability (12) Lobo, M. J.; Miranda, A. J.; Tunon, P. Electroanalysis 1997, 9, 191. (13) Wang, J.; Chatrathi, M. P.; Tian, B. Anal. Chim. Acta 2000, 416, 9.
Analytical Chemistry, Vol. 73, No. 6, March 15, 2001
1299
Figure 7. Electropherograms for (A) a mixture of (a) 7.0 × 10-4 M glucose and (b) 2.0 × 10-2 M ethanol, and for different wine samples (diluted 100-fold): (B) California red wine sample, (C) Spanish red wine sample, and (D) Mexican red wine. Other conditions, as in Figure 2.
of the microchip is illustrated in Figure 7 for assays of different red wine samples from California (B), Spain (C), and Mexico (D). Also shown is a typical electropherogram for a 7 × 10-5 M glucose/2 × 10-2 M ethanol sample mixture. The wine samples were analyzed without any sample preparation, with the exception
1300 Analytical Chemistry, Vol. 73, No. 6, March 15, 2001
of a 100-fold dilution, and yielded well-defined glucose and ethanol peaks. Note the similarity of the migration times of all sample and standard solutions. In conclusion, we have demonstrated a new biochip strategy for performing alcohol/sugar testing more rapidly, easily, and economically. While the concept of oxidase/dehydrogenase separation chips has been illustrated in connection to alcohol and sugar, it can be readily expanded to other enzymes/substrates pairs. For example, we are currently examining a similar “lab-ona-chip” protocol for the simultaneous measurement of the critical care” metabolites lactate and glucose (in connection with the GOx/LDH/NAD+ reagent). We are also exploring a multichannel operation, involving different combinations of enzymes, for highthroughput measurements of numerous substrates. ACKNOWLEDGMENT This project was supported by the National Institute of Health (NIH Grant RO1 RR14173-02).
Received for review December 31, 2000. AC001205T
October
11,
2000.
Accepted
