Chemiluminometric Sensor for Simultaneous ... - ACS Publications

Chemiluminometric Sensor for Simultaneous Determination of l-Glutamate and l-Lysine with Immobilized Oxidases in a Flow Injection System. Nobutoshi Ki...
0 downloads 0 Views 78KB Size
Anal. Chem. 2002, 74, 1269-1274

Chemiluminometric Sensor for Simultaneous Determination of L-Glutamate and L-Lysine with Immobilized Oxidases in a Flow Injection System Nobutoshi Kiba,* Takao Miwa, Masaki Tachibana, Kazue Tani, and Hitoshi Koizumi

Department of Applied Chemistry and Biotechnology, Faculty of Engineering, Yamanashi University, Kofu 400-8511, Japan

A chemiluminometric flow-through sensor for simultaneous determination of L-glutamate (Glu) and L-lysine (Lys) in a single sample has been developed. Immobilized uricase, immobilized peroxidase, support material, coimmobilized glutamate oxidase/peroxidase, support material, and coimmobilized lysine oxidase/peroxidase were packed sequentially in a transparent PTFE tube, and the tube was placed in front of a photomultiplier tube as a flow cell. A three-peak recording was obtained by one injection of the sample solution. The peak height of the first peak was due to the concentrations of urate and other reductants in the sample; the immobilized uricase was used to decompose urate, and the hydrogen peroxide produced was decomposed with a luminol-hydrogen peroxide reaction by immobilized peroxidase. The peak heights of the second and third peaks were free from the interferences from the reductants and were dependent only on the concentrations of Glu and Lys, respectively. Calibration graphs for Glu and Lys were linear at 401000 and 50-1200 nM, respectively. The sampling rate was 11/h without carryover. The sensor was stable for two weeks. The sensor system was applied to the simultaneous determination of Glu and Lys in serum. Chemiluminescence enzyme sensors based on the integration of reaction and detection in a flow cell have received much attention in wide set of application fields, owing to the inherent advantages such as sensitivity, ease of operation, high sample throughput, and suitability for miniaturization in analytical chemistry.1-3 In conventional flow injection (FI) systems with immobilized enzyme reactors, a sample plug from a reactor disperses into the carrier stream and is diluted with reagent solution because enzyme reactions (reactors) and continuous detection in a conventional flow cell take place in different modules that are isolated in space; these make the peak broader and therefore decrease the sensitivity and lower the possible sampling rate. The sensitivity of the sensors rely on the efficient coimmobilization of oxidases and peroxidase (POx) on an appropriate (1) Valcarcel, M.; Luque de Castro, M. M. Analyst 1993, 118, 593-600. (2) Zhang, X. R.; Baeyenes, W. R. G.; Garcia-Campana, A. M.; Ouyang, J. Trends Anal. Chem. 1999, 18, 384-391. (3) Aboul-Enein, H. Y.; Stefan, R.-I.; van Staden, J. F.; Zhang, X. R.; GarciaCampana, A. M.; Baeyens, W. R. G. Crit. Revi. Anal. Chem. 2000, 30, 271289. 10.1021/ac011013d CCC: $22.00 Published on Web 02/14/2002

© 2002 American Chemical Society

support; POx catalyzes a chemiluminescence reaction of luminol with hydrogen peroxide formed in the oxidase reaction. Fiberoptic chemiluminometric flow-through sensors with POx and oxidases such as glutamate oxidase (GOx),4 lysine oxidase (LOx),4 xanthine oxidase,4 glucose oxidase,5 choline oxidase,6 and lactate oxidase7 have been developed for the determination of L-glutamate (Glu), L-lysine (Lys), xanthine, D-glucose, choline, and L-lactate, with lower detection limits of 10-6 M levels. For the coimmobilization of the enzymes, preactivated nylon membrane4,6,7 and polyvinylidene membrane5 were used. The sensors were not highly sensitive because of their small binding capacities; the more enzymes are immobilized, the less is the inherent activity of each immobilized enzyme. For efficient coimmobilization of two or three enzymes, polymer beads have been used. A sensitive trienzyme sensor for branched chain amino acids was constructed from transparent PTFE tube containing leucine dehydrogenase, NADH oxidase, and POx coimmobilized on tresylated hydrophilic vinyl polymer beads that were coiled spirally in front of a photomultiplier tube.8 A uricase/POx coimmobilized on the polymer beads in a flow cell was used to develop a sensor for uric acid in a closedloop FI system.9 Lower detection limits for the sensors were 10-8 M levels. The sensors have been applied to the assay of a few real samples,8,9 because the luminol-H2O2 reaction is seriously subjected to interference from reducing substances, which depress results by interacting with the H2O2 and the intermediates of the luminol reaction.10-14 So far, considerable attention has been focused on the development of flow-through chemiluminescence sensors for the determination of a single component. Sensors for simultaneous (4) Sphon, U.; Preuschoff, F.; Blankenstein, G.; Janasek, D.; Kula, M.-R.; Hacker, A. Anal. Chim. Acta 1995, 303, 109-120. (5) Blum, L. J. Enzyme Microb. Technol. 1993, 15, 407-411. (6) Lapp, H.; Spohn, U.; Janasek, D. Anal. Lett. 1996, 29, 1-17. (7) Berger A.; Blum, L.J. Enzyme Microb. Technol. 1994, 16, 979-984. (8) Kiba, N.; Tachibana, M.; Tani, K.; Niwa, T. Anal. Chim. Acta 1998, 375, 65-70. (9) Kiba, N.; Suzuki, K.; Miwa, T.; Tachibana, Koizumi, H.; Tani, K. Anal. Sci. 2000, 16, 1203-1205. (10) Tabata, M.; Fukunaga, C.; Ohyabu, M.; Murachi, T. J. Appl. Biochem. 1984, 6, 251-258. (11) Tabata, M.; Murachi, T. J. Biolumin. Chemilumin. 1988, 2, 63-67. (12) Bostick, D. T.; Hercules, D. M. Anal. Chem. 1975, 47, 447-452 (13) Landay, M. J.; Nuttall, S. L.; Maxwee, S. R. J.; Thorpe, G. H. G. Ann. Clin. Biochem. 1998, 35, 533-544. (14) Cui, H.; Meng, R.; Jiang, H. Y.; Sun, Y. G.; Lin, X. Q. Luminescence 1999, 14, 175-179.

Analytical Chemistry, Vol. 74, No. 6, March 15, 2002 1269

Figure 1. Schematic depiction of the PTFE tube packed with immobilized enzyme beads and polymer beads without enzymes. The tube (1.5 mm o.d., 1.0 mm i.d) was coiled spirally and set in front of a photomultiplier tube as the flow cell. I, immobilized uricase (UC); II, immobilized peroxidase (POx); III, polymer beads without enzymes (supports) (UPB); IV, coimmobilized glutamate oxidase and peroxidase (GOx/POx); V, coimmobilized lysine oxidase and peroxidase (LOx/POx).

determination of multisubstrates, as well as chemiluminometric methods that combine immobilized enzyme reactors with an FI system, are now desired by application fields such as food processing, fermentation control, and diagnosis.15-17 The aim of this work was the development of a sensitive and selective flow-through sensor for the simultaneous determination of Glu and Lys free from interference from reducing substances such as urate and ascorbate; urate is reductive and one of the major interfering components in serum for a luminol-H2O2 chemiluminescence reaction.10-14 The sensor was based on the integration of oxidase reactions and chemiluminescence detection reactions, which involved no separation process. The novelty of this method was in the simultaneous determination in a singleline flow injection system with a single detector. Glu is an important ingredient of various foods, used as a flavor enhancer and present in many biological fluids and pharmaceuticals. Lys is an essential amino acid and widely used in the food industry, agriculture, and medicine because a lysine-disbalanced diet is implicated in several diseases. The simultaneous measurement of these amino acids is necessary in clinical and food chemistry. In this work, four kinds of immobilized enzymes were used. GOx and LOx, respectively, and POx were coimmobilized covalently to the vinyl polymer beads to obtain coimmobilized enzymes such as GOx/POx and LOx/POx. Uricase (UC) and POx were immobilized separately. The immobilized enzymes and untreated polymer beads (support materials for the immobilized enzymes, UPB) were alternately packed into a PTFE tube and these were aligned as UC|POx|UPB|GOx/POx|UPB| LOx/POx. The tube was used as a flow cell in a single-line FI system using luminol solution as a carrier solution. The sensor system was applied to the simultaneous determination of Glu and Lys in serum. EXPERIMENTAL SECTION Chemicals. Luminol (5-amino-2,3-dihydro-1,4-phthalazinedione, free acid), tresyl chloride (2,2,2-trifluoroethanesulfonyl chloride), L-amino acids, and all other chemicals were from a commercial source (Sigma or Aldrich) and were used as supplied without further purification. POx (EC 1.11.1.7, from Arthromyces ramosus, 250 units/mg) and UC (EC 1.7.3.3, from Arthrobacter globiformis, 30 units/mg) were obtained from Suntory (Osaka, Japan) and Asahi Kasei (Tokyo, Japan), respectively. GOx (EC 1.4.3.11, from Streptomyces sp, 20 units/mg) and LOx (L-lysine R-oxidase, EC 1.4.3.14, from Trichoderma viride, 20 units/mg) were from Yamasa (Chiba, Japan). Hydrophilic vinyl polymer beads (TSKgel Toyopearl HW(15) Valcarcel, M.; Luque de Castro, M. D. Analyst 1984, 109, 413-419. (16) Schmit, R. D. Flow Injection Analysis (FIA) Based on Enzymes or Antibodies; GBF Monograph 14; VCH: Weinheim, 1991. (17) Chen, R. L. C.; Matsumoto, K. J. Flow Injection Anal. 1995, 12, 167-175.

1270

Analytical Chemistry, Vol. 74, No. 6, March 15, 2002

65F) was purchased from Tosoh (Tokyo, Japan). The beads were sieved to obtain a 50 ( 10 µm sieve fraction. A stock solution (1 mM) of each amino acid was prepared by dissolving in 0.1 M sodium carbonate solution. Standard urate solutions were prepared daily by dissolving uric acid in 0.05 M sodium carbonante-0.1 M sodium hydrogen carbonate buffer (pH 9.0). A stock solution (10 mM) of luminol was prepared by dissolving in 0.1 M sodium carbonate and the resultant mixture stored for 3 days in a refrigerator (4 °C) to obtain stability before use.10,19 The solution was stable enough to be used for two months. The solution was diluted 50-fold with 0.05 M sodium carbonate0.1 M sodium hydrogen carbonate buffer (pH 9.0) before use and stored in an amber glass bottle for 12 h at room temperature. Preparation of Immobilized Enzymes. Four kinds of immobilized enzymes were prepared: immobilized UC, immobilized POx, coimmobilized GOx/POx, and coimmobilized LOx/POx. The beads (1 g) were washed with dry acetone and suspended in 10 mL of dry acetone-pyridine (1:1 v/v). With magnetic stirring, 1 mL of tresyl chloride was added dropwise to the suspension for 10 min. The reaction was continued for 10 min. The beads were washed with acetone and then 0.1 M phosphate buffer (pH 7.0). The tresylate-beads (two parts for UC, each one part for POx, GOx/POx, and GOx/POx) were suspended in 5 mL of 0.1 M phosphate buffer (pH 7.0) containing UC (8 mg), POx (3 mg), GOx/POx (2 mg:1 mg) or GOx/POx (2 mg:1 mg) at room temperature for 4 h with occasional shaking. After the immobilization, the activity of UC in the solution was assayed by measuring the initial rate of uric acid consumption at 293 nm with uric acid as a substrate. POx, GOx, and GOx activities were measured at 500 nm with the 4-aminoantipyrine method. UC and POx were immobilized at 59 and 100% yields, respectively. For coimmobilized GOx/POx, GOx and POx were immobilized with 87 and 94% yields, respectively, and for LOx/POx, LOx and POx were immobilized with 74 and 98% yields, respectively. The immobilized UC, POx, support (UPB), GOx/POx, UPB, and LOx/POx, respectively, were packed in turn into a transparent PTFE tube (1.5 mm o.d., 1.0 mm i.d., 73 cm length, limiting pressure 50 kg/ cm2, Valuqua, Tokyo) by a slurry packing method. The end of the tube was closed with a ceramic frit (pore size 500 nm, GL Sciences, Tokyo, Japan) as shown in Figure 1. The tube was coiled spirally and used as a flow cell. The flow cell was washed with 0.1 M Tris-HCl buffer (pH 8.0) to saturate the free sites. When not in use, the cell was washed with 0.1 M phosphate buffer (pH 7.0) and stored in a refrigerator. The design of a chemiluminometric flow-through cell of packed immobilized enzymes for simultaneous measurements involves problems such as sensitivity, resolution, operational stability, and (18) Kiba, N.; Ito, S.; Tachibana, M.; Tani, K.; Koizumi, H. Anal. Sci. 2001, 17, 929-933. (19) Koerner, C. A.; Nieman, T. A. Anal. Chem. 1986, 58, 116-119.

does not interfere in subsequent measurements because the emission is instantaneous. The enzyme reaction schemes are as follows:

urate + O2 + H2O

Figure 2. FI system for simultaneous determination of L-glutamate and L-lysine. A, 200 µM luminol in carbonate buffer (pH 9.0); B, pump (0.1 mL/min); C, injection valve; D, flow-through cell; E, photomultiplier tube; F, recorder. All connecting tubing (0.2 mm i.d.) was made of PTFE.

pressure drop. For sensitivity and resolution, small particles of enzyme support are advisable because dispersion causes a decrease in peak height (sensitivity) and an increase in peak width (resolution). The smaller the particles, the sharper the width of the sample plug. On the other hand, when small particles were used, the pressure drop increased greatly. In this work, the size of the support was inevitably limited by the maximum tolerable pressure drop because the limiting pressure of the PTFE tube used was 50 kg/cm2. Particles of 50 ( 10 µm were used owing to the low tolerable pressure drop for the tube. Measurement and Apparatus. The FI system used in this study is shown in Figure 2. It was assembled with a piston pump (model L-6000, Hitachi, Tokyo, Japan) for the luminol solution, a six-way valve (model SVM-6M2L, alkali-proof, dead volume 4 µL, Sanuki, Tokyo, Japan) with a 2-µL loop, a luminometr (model S-3400, applied voltage of photomultiplier tube; 900 V, Soma, Tokyo, Japan), and a recorder (model FBR251A, TOA, Tokyo, Japan). The flow rate of the luminol solution was 0.1 mL/min, and all measurements were performed at room temperature (∼25 ( 2 °C). The height of the recorded FI peaks was manually determined in relative units. Determination of L-Glutamate and L-Lysine in Serum. Fresh human serum (2 µL) was diluted with 400 nM urate in 0.05 M sodium carbonate-0.1 M sodium hydrogen carbonate buffer (pH 9.0) to 2 mL and filtered through an ultrafiltration membrane (USY-1, molecular weight cutoff: 10000, Advantec, Tokyo, Japan). The filtrate was injected via the valve. The results obtained by the present method were compared with those obtained with HPLC (column, TSKgel SCX (150 × 60 mm i.d.); mobile phase, citrate buffer (gradient); postcolumn derivatization with ninhydrin). RESULTS AND DISCUSSION Basis of the Proposed Method. Four kinds of immobilized enzymes were used. Immobilized UC and POx have been used in a single-line FI system to eliminate the interferences from urate and other reductans.18 The immobilized UC decomposes urate in sample plug, in which urate is converted quantitatively into allantoin with a concomitant formation of H2O2 and then the nonenzymatical decomposition of the H2O2 by the reaction of other reducing compounds takes place simultaneously. Subsequently, POx-catalyzed luminol reaction takes place in the immobilized POx, emitting luminescence light. The luminescence intensity is affected by the interaction between the unreacted reductants and radical intermediates of the luminol reaction.14 The luminescence

f allantoin + CO2 + H2O2 UC

luminol + 2H2O2 + OH- f POx aminophthalate + N2 + 3H2O + hv (∼420 nm)

The sample plug moves down from the UPB zone to coimmobilized GOx/POx in which GOx catalyzes the oxidative deamination of Glu with concomitant production of H2O2. The H2O2 participates in a luminol reaction catalyzed by POx with emission of light; the light intensity is proportional to the Glu concentration in the sample plug. Finally, Lys was oxidized and determined with coimmobilized POx/LOx. The enzyme reaction schemes are as follows: L-glutamate

L-lysine

+ O2 + H2O f GOx ketoglutarate + NH3 + H2O2

+ O 2 + H 2O f LOx ketoaminocaproate + NH3 + H2O2

Thus, tree-peak recording was obtained by one shot of the sample solution. The peak height of the first peak depends on the concentration of urate and other reductants and heights for the second and third peaks are based only on the concentrations of Glu and Lys, respectively. Characterization of the Sensor. To evaluate the effect of the pH on the activities and the stabilities of the immobilized enzymes, 0.1 M Tris-0.1 M HCl buffer (pH 8.0-9.0) and 0.05 M sodium carbonate-0.1 M sodium hydrogen carbonate buffer (pH 9.010.5) were used. The effect of the pH on the activities of the immobilized enzymes UC, POx, GOx/POx, and LOx/POx for urate, H2O2, Glu, and Lys, respectively, was studied by using the system shown in Figure 2. The maximum sensitivities were achieved at pH 9.5, 10.0, 9.0, and 9.8 for urate, H2O2, Glu, and Lys, respectively. For the stability test for two weeks, the luminol solutions buffered at pH 9.0, 9.5, and 10.0, respectively, were used. The sensor was used for 5 h in a day and each standard solution containing a substrate was injected every 6 min. The flow cell was washed with 0.1 M phosphate buffer (pH 7.0) and stored in a refrigerator when not in use. The decrease of the peak height obeyed almost first-order kinetics, and the operational stabilities of the enzymes at pH 9.0 were higher than those at higher pH buffers. The half-life periods of immobilized enzymes UC, POx, GOx/POx, and LOx/POx at pH 9.0 were 39, 45, 40, and 37 days, respectively. For the preparation of a sensitive sensor containing coimmobilized GOx/POx and LOx/POx, the protein mass ratios between GOx (20 units/mg) or LOx (20 units/mg) and POx (250 units/ mg) in the enzyme solutions were varied between 1:1 and 9:1, while keeping the protein mass constant (3 mg). The optimum mass ratios were in the range 2:1-3:1 (activity ratios 1:6-1:4). Analytical Chemistry, Vol. 74, No. 6, March 15, 2002

1271

Figure 4. Effects of flow rate on the resolution values and peak widths (in mL) at half-height for the peaks. Rs1 ) td1/(WI + WII); Rs2 ) td2/(WII + WIII). WIV, WIIV, and WIIIV are peak widths (in mL) at halfheight for the first peak, for the second peak, and for the third peak, respectively.

Figure 3. FI response for a standard mixture of urate, Glu and Lys (150 nM each). First peak, urate; second peak, Glu; third peak, Lys. WI, WII, and WIII (in mm) are the peak widths at half-height for the first, the second, the third peaks, respectively, td1 and td2 (in mm) are the distances between the first peak and the second peak and between the second peak and the third peak, respectively.

Estimation of interference from urate was performed by using a flow cell packed with UPB (15 cm) instead of immobilized UC (15 cm) shown in Figure 1 and by preparing a series of mixtures (Glu and Lys, 150 nM each) containing from 50 to 600 nM urate. The peak height was compared to a 150 nM Glu and Lys standard solution containing no urate. Urate concentrations above 150 nM depressed the responses by reducing the H2O2 and by interacting with intermediates of luminol. To avoid the interference from urate, the immobilized UC and POx were aligned in the flow cell in turn. The amounts of UC (15 cm length) and POx (5 cm length) were sufficient to decompose urate or H2O2 up to 1 µM under the conditions shown in Figure 2. By injecting a mixture of Glu (200 nM) and urate (1 µM), negative error, which is evidence for incomplete decomposition of urate with immobilized UC, and positive error due to the incomplete decomposition of the H2O2 with immobilized POx were not observed. Figure 3 shows the FI response obtained from a standard mixture (urate, Glu, and Lys, 150 nM each) under the conditions shown in Figure 2. The conversion efficiencies for urate, Glu, and Lys were 100, 43, and 21%, respectively, immediately after the preparation of the immobilized enzymes; the peak area of urate peak was compared with that of the H2O2 peak and each peak area of the Glu and Lys peaks was compared with that of the urate 1272 Analytical Chemistry, Vol. 74, No. 6, March 15, 2002

Figure 5. FI responses for three injections of the serum sample. Second peak, Glu; third peak, Lys. Serum was diluted 1000-fold with 400 nM urate in carbonate buffer (pH 9.0) and filtered through an ultrafiltration membrane.

peak. By comparing the peak height of each peak, urate:Glu:Lys ) 100:37:15. The flow rate of the luminol solution was changed from 0.05 to 0.3 mL/min to obtain the optimal resolutions between two peaks. The resolutions (Rs1 and Rs2) between two peaks were defined as Rs1 ) td1/(WI + WII) and Rs2 ) td2/(WII + WIII), respectively, where td1 is the distance (mm) between the urate peak and Glu peak and td2 is the distance between the Glu peak and Lys peak. WI, WII, and WIII were the peak widths (mm) at half-height for the urate, Glu, and Lys peaks, respectively, as shown in Figure 3. With increasing flow rate, the resolutions decreased because peak widths at half-height for urate, Glu, and Lys peaks in milliliter units (WIv, WIIv, WIIIv) increased, as shown in Figure 4. Also, the resolutions were influenced by the injection volume. Maximum injection volume was limited by the volume of the solution held in the UPB (i.e., the dead space in the UPB). The resolutions decreased with an increase in the total injection volume (the loop volume plus the dead volume (4 µL) in the valve); at 10

Figure 6. Comparison between the present method and chromatography for the determination of Glu and Lys in serum.

µL, Rs1 and Rs2 values were 1.01and 0.54, respectively. A flow rate of 0.1 mL/min and a loop of 2 µL (total injection volume was 6 µL) were selected; under the conditions, Rs1 and Rs2 were 1.45 and 0.83, respectively, and the maximum sample throughput was 11/h. Higher resolution between the urate and Glu peaks was required because in sera for healthy subjects urate concentration ranges between about 0.7 and 3 times that of Glu, and for gout and hyperuricemia, higher urate concentrations are observed.20,21 The 30-cm length of UPB between immobilized POx and coimmobilized Glu/POx permitted the assay of the sample in which urate was contained at even 5 times the quantity of Glu. The sensor exhibited constant peak heights in the luminol solutions ranging in concentration from 80 to 400 µM. Simultaneous Determination of Urate, Glu, and Lys. After characterization of the sensor system, the peak heights were measured by changing the concentrations of urate, Gly, and Lys, keeping the concentration of these substrates at 1:1:1. Linear responses of the calibration curves were observed from 20 to 1200 nM urate with a correlation coefficient of 0.999 (15 data points), from 40 to 1300 nM Glu with 0.995, and from 50 to 1500 nM Lys with 0.990. The slopes of the graphs for urate and Glu were 6.68 and 2.44 times that of Lys, respectively. The detection limits (signal-to-noise ratio) were 5 nM urate, 20 nM Glu, and 40 nM Lys. A mixture (400 nM urate, 150 nM Glu, 150 nM Lys) was repeatedly analyzed during two weeks. The system was used for analyses of 30 sample solutions for 3 h in a day. The peak heights for urate, Glu, and Lys had within-day relative standard deviations (RSDs) of 0.66, 0.83, and 0.97% and day-to-day RSDs of 0.83, 1.5, and 2.1%, respectively. Interference from other Reductants. The height of the urate peak was not dependent only on the concentration of urate in the serum sample because the H2O2 produced in the immobilized UC is consumed for reacting with other reductants such as ascorbate and glutathione. When ascorbate or glutathione contained 1.2 times the amount of urate in a standard solution, a urate peak was not observed and the peak hight of Glu peak was affected. To decompose the excess reductants by H2O2 from the UC reaction, serum was diluted with 400 nM urate in 0.05 M sodium carbonate-0.1M sodium hydrogen carbonate buffer (pH 9.0), (20) Nagele, U.; Ziegenhorn, J.; Kolose, S. In Methods of Enzymatic Analysis; Bergmeyer, H. U., Ed.; VHC: Weinheim, 1985; Vol. VII, p 140. (21) Gemba, T. Bunseki 1976, 433-440.

Table 1. Repeatability of the Peak Height (mm) for Ten Replicate Injections no.

first peak

Glu peak

Lys peak

1 2 3 4 5 6 7 8 9 10 average RSD, %

338 309 323 307 319 335 327 323 316 319 322 3.1

259 261 257 272 270 262 261 264 263 264 263 1.8

206 208 210 213 212 213 218 210 215 210 211 1.6

since the fresh serum of healthy subjects20-22 contains 140-410 µM urate, 120-580 µM ascorbate, and 1-2 µM glutathione. Application. The present method was applied to the simultaneous determination of Glu and Lys in serum from normal adult subjects, containing 140 µM Glu and 154 µM Lys; the concentrations of urate,20 ascorbate,23 and glutathione,24 which were obtained by enzymatic methods, were 161, 598, and 1 µM, respectively. Figure 5 shows three responses obtained from the serum; the serum was diluted with 400 nM urate in the carbonate buffer (pH 9.0) and filtered. The filtrate was then injected into the system. The RSDs for 10 replicate injections were 1.8% for Glu and 1.6% for Lys, as shown in Table 1. For a comparison, determinations were carried out by HPLC; n ) 10, range 136-205 µM Glu and 105-199 µM Lys (144-480 µM urate and 201-511 µM ascorbate). The results are shown in Figure 6. The respective calculated linear regressions and correlation coefficients were y ) 0.991x + 0.876 and 0.987 for Glu, and y ) 0.980x - 0.581 and 0.992 for Lys. The sensor gave slightly bigger values for Lys because of the lower specificity of LOx to other amino acids in serum.25 CONCLUSIONS A sensor for simultaneous determination of Glu and Lys was designed to utilize the principle of a chemiluminometric flow (22) Griffith, O. W. Anal. Biochem. 1980, 106, 203-205. (23) Beutler, H.-O. In Methods of Enzymatic Analysis; Bergmeyer, H. U., Ed.; VHC: Weinheim, 1985; Vol. VI, p 376. (24) Griffith, O. W. Methods of Enzymatic Analysis; Bergmeyer, H. U., Ed.; VHC: Weinheim, 1985; Vol. VIII, p 521. (25) Kusakabe, H.; Kodama, K.; Kuninaka, A.; Yoshino, H.; Misono, H.; Soda, K. J. Biol. Chem. 1980, 255, 976-981.

Analytical Chemistry, Vol. 74, No. 6, March 15, 2002

1273

injection system with the integration of coimmobilized GOx/POx and LOx/POx and to avoid interference from reducing compounds such as urate and ascorbate by using immobilized UC and POx. The sensor was stable enough to permit the measurements of more than 400 samples. The sample throughput was 11/h. This work showed one possibility of simultaneous determination of oxidase substrates by a single-line flow injection system. Besides GOx, LOx and UC, alcohol oxidase, amino acid oxidase, cholesterol oxidase, choline oxidase, glycerophosphate oxidase, putorescine oxidase, sarcosine oxidase, and xanthine oxidase will be stable enough in alkaline media to permit to be used in immobilized forms in the flow cell. Therefore, the integration of

1274

Analytical Chemistry, Vol. 74, No. 6, March 15, 2002

oxidase reactions and the POx-catalyzed H2O2-luminol luminescence reaction in a flow cell holds great promise for development of simultaneous FI determination systems for the substrates. ACKNOWLEDGMENT The authors acknowledge the Grant-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology, Japan (10650795, 1355231) for partial support of the research. Received for review September 19, 2001. Accepted December 5, 2001. AC011013D