Anal. Chem. 1984, 56, 1876-1880
eraging signals from more pulses per dalton would produce a smoothing effect, but the pulses/dalton is limited by scan time, mass scanning range, and LC sample elution time considerations. A time-of-flightmass analyzer is more compatible with a pulsed ionization source and would also give better sensitivity. As a continuous ionization source, primary ion bombardment provides more characteristic fragment peaks without losing molecular weight information. Mass spectra as well as LC chromatograms are stable and reproducible. Our detection limits for positive ions at the present time correspond typically to a few nanograms of sample in both techniques. But detection limits and sensitivities depend strongly on the properties of the compound and may be quite different between LDMS and SIMS. For example, in LDMS the sensitivity for adenosine is substantially lower than for the other three nucleosides. But in SIMS, sensitivity for adenosine is much higher than for the other nucleosides. The sample deposition efficiency from the thermosprayer was determined by recovering a known sample introduced on the belt under normal operating conditions into 5 mL of H20. Its UV absorbance was measured with a spectrophotometer. The amount of sample recovered was quantitated by comparing to a set of standards manually spread on the belt with a syringe and then recovered by the same procedure. The average deposition efficiency from our current thermosprayer configuration is only 20-40%. We have found that using the thermospray device to introduce sample at atmospheric pressure can give deposition efficiencies of at least 80%. Some modification of the instrument is planned so that the sample introduction system plus a solvent scrubber belt cleanup system can be placed in an extended chamber under ambient pressure. The sensitivity of the system should improve significantly from the enhanced deposition efficiency and de-
creased background expected to result from this modification. Registry No. Adenosine, 58-61-7;histidine, 71-00-1;cytidine, 65-46-3; uridine, 58-96-8. LITERATURE CITED (1) Hardin, E. D.; Fan, T. P.; Vestal, M. L. Anal. Chem. 1983, 56, 2. (2) Benninghoven, A.; Jaspers, D.; Sichtermann, W. Appl. Phys. 1976, 1 1 , 35. (3) Benninghoven, A.; Sichtermann, W. K. Anal. Chem. 1978, 50, 1180. (4) Eicke, A.; Sichtermann, W.; Benninghoven, A. Org Mass Spectrom, 1980, 15, 289. (5) Rinehart, K. L., Jr. Science 1982, 218, 254. (6) Barber, M.; Bordoii, R. S.; Sedgwick, R. D.; Tyler, A. N. J . Chem. Soc., Chem. Commun. 1981, 7 , 325. (7) Vastoia, F. J.; Pirone, A. J. A&. Mass Specfrom. 1968, 107. (8) Posthumus. M. A.; Kistemaker, P. G.: Menzeiaar, H. L. C. Anal. Chem. 1978, 50, 985. (9) Heinen, H. J. Int. J . Mass Specfrom. Ion Phys. 1981, 38, 309. (10) Van Breeman, R. B.; Snow, M.; Cotter, R. J. Int. J . Mass Specfrom. Ion Phys. 1983, 4 9 , 35. (11) Zakett, D.; Schoen, A. E.; Cooks, R. G.; Hemberger, P. H. J . Am. Chem. SOC. 1981, 103, 1295. (12) Hiilenkarnp, F. "Laser Induced Ion Formation from Organic Solids"; I n Proceedings of 2nd International Conference on Ion Formation from Organic Solids; Benninghove, A., Ed.; Springer-Veriag: New York, 1983; Springer S e r b in Chemical Physics 25; p 190. (13) Schueier, 8.; Krueger, F. R. Org. Mass Specfroom. 1979, 14 (E), 439. (14) Ens, W.; Standing, K. G.; Chait, 8. T.; Field, F. H. Anal. Chem. 1981, 53, 1241. (15) Benninghoven, A.; Eicke, A.; Junack, M.; Sichtermann, W. Org. Mass Specfrom. 1980, 15, 459. (16) Smith, R. D.; Burger, J. E.; Johnson, A. L. Anal. Chem. 1981, 53, 1603. (17) Benninghoven, A. "Secondary Ion Mass Spectrometry of Organic Compounds"; In Proceedings of 2nd International Conference on Ion Formation from Organic Solids"; Benninghoven, A., Ed.; Springer-Verlag: New York, 1983; Springer Series in Chemical Physics 25; p 64. (18) Benninghoven, A. Surf. Sci. 1973, 35, 427. (19) Hardin, E. D.; Vestal, M. L. Anal. Chem. 1981, 53, 1492.
RECEIVED for review March 8,1984. Accepted April 16,1984. This work was supported by the Institute of the General Medical Sciences (NIH) under Grant GM 29451 and the Robert A. Welch Foundation.
Sequential Determination of Glutamate-Oxalacetate Transaminase and Glutamate-Pyruvate Transaminase Activities in Serum Using an Immobilized Bienzyme-Poly(viny1 chloride) Membrane Electrode Kunio Kihara,* Eiki Yasukawa, and Sachio Hirose Central Research Laboratory, Mitsubishi Petrochemical Co., Ltd., Wakaguri, Ami, Inashiki-gun, Ibaraki 300-03, Japan
Oxalacetate decarboxylase (EC 184.108.40.206) and pyruvate oxldase (EC 220.127.116.11) are colmmoblllred by adsorption on a wet poly(v1nyl chloride) membrane. The activities of glutamateoxalacetate transamlnase (GOT) and glutamate-pyruvate transamlnase (OPT) In serum are sequentially determlned by a blenryme sensor conslstlng of the lmmoblllred oxalacetate decarboxylase-pyruvate oxldase-poly( vinyl chloride) membrane and a hydrogen peroxlde electrode. The assay of GOT and OPT actlvltles requlres 1 4 mln and has a preclslon of ca. 4%. The sequentlal determlnatlon by use of the lmmoblllred blenzyrne-pofy(vlny1 chloride) membrane has advantages In speed, slmpllclty and preclslon.
alacetate transaminase (GOT) is of clinical significance with elevated values in serum indicating myocardial, hepatic, and jaundice diseases (I). In addition, the comparison with the activities of GPT and GOT, as well as the level of each activity of them, is very important in diagnosis. Many spectrophotometric and fluorimetric methods exist for the assay of these enzymes (2-6), all based on a measurement of the decrease in absorbance of NADH as a result of the following reaction sequence: for GOT L-aspartate
The determination of activities of transaminase such as glutamate-pyruvate transaminase (GPT) and glutamateox-
+ NADH + H+ pH 7.8
0003-2700/84/0356-1876$01.50/00 1984 American Chemical Society
+ NADH + H+
ANALYTICAL CHEMISTRY, VOL. 56, NO. 11, SEPTEMBER 1984
+ L-glutamate (3)
The sequential determination of GOT and GPT activities in the same sample of serum is difficult by these methods. Recently, the few determinations of enzyme activities, not substrates, such as GPT (7), a-amylase (8), and lactate dehydrogenase (LDH) (9) in biological fluids are monitored by an electrochemical method of which the advantage is convenient and inexpensive. Mizutani e t al. have described a sequential determination of L-lactate and LDH as an example of substrate and enzyme, based on immobilization of a single enzyme on a carrier (IO). Here a widely applicable enzyme sensor consisting of a bienzyme poly(viny1 chloride) (PVC) membrane and polarographic hydrogen peroxide electrode for sequential GOT and GPT activities in serum is proposed. The following two enzymes, oxalacetate decarboxylase (OAC) and pyruvate oxidase (POP), were effectively coimmobilized onto a porous membrane which has been prepared by PVC ( I I , I 2 ) . GOT activity could be measured by the bienzyme sensor based on hydrogen peroxide produced by the enzymatic reactions with the use of immobilized oxalacetate decarboxylase and pyruvate oxidase as a result of the following reaction: oxalacetate pyruvate
+ O2+ phosphate
C 0 2 + H202 (6)
Oxalacetate is enzymatically produced by GOT in serum as shown in eq 1and GPT in serum enzymatically produces the pyruvate as shown in eq 3. These immobilized enzymes catalyze the decarboxylation of oxalacetate to produce pyruvate and carbon dioxide and the oxidation of pyruvate to produce acetyl phosphate, carbon dioxide, and hydrogen peroxide as described above. Thus, the immobilized bienzyme-PVC membrane and its biosensor system make it possible to detect GOT activity as well as GPT, which in serum has been determined by a pyruvate oxidase-PVC membrane electrode in a previous paper
(7). This paper describes the bienzyme sensor consisting of the immobilized bienzyme-PVC membrane and a hydrogen peroxide electrode for the assay of GOT activity in human serum and the GOT and G P T activities in the same sample of serum were sequentially determined by the sequential addition of substrates, which were L-aspartate and a-ketoglutarate, and L-alanine.
EXPERIMENTAL SECTION Materials. The enzymes used were pyruvate oxidase (EC 18.104.22.168, from Pediococcus sp., 21 IU mg-'; Toyo Jozo Co., Tokyo, Japan), oxalacetate decarboxylase (EC 22.214.171.124, from Pseudomonas sp., 110 IU mg-'; Toyo Jozo Co., Tokyo, Japan), glutamate oxalacetate transaminase (GOT, EC 126.96.36.199, from pig heart, 200 IU mg-l; Boehringer Mannheim GmbH), and glutamate pyruvate transaminase (GPT, EC 188.8.131.52, from pig heart, 140 IU mg-'; Boehringer Mannheim GmbH). Human control serum was purchased from Japan Travenol Co. (Tokyo, Japan), and L-aspartate, L-alanine, a-ketoglutarate, lithium pyruvate, thiamine pyrophosphate (TPP), flavine adenine dinucleotide (FAD), 3(N-morpho1ino)propanesulfonicacid (MOPS), and tricine were purchased from Nakarai Chemical Co. (Kyoto, Japan). Poly(viny1 chloride) (PVC, mol wt 48400)was purchased from Kanegafuchi Kagaku Co. (Tokyo, Japan). Other reagents were analytical or laboratory grade materials. Deionized water was used for all procedures.
Flgure 1. Schematic diagram of the bienzyme membrane sensor system: (1) blenzyme-PVC membrane; (2) cellulose acetate membrane; (3) O-ring; (4) hydrogen peroxide electrode; (5) microsyringe; (6) cell; (7), (a), (9), and (10) peristaltic pumps; (11)substrate (1) Solution for GOT (L-aspartate (250 mM), a-ketoglutarate (10 mM), FAD (0.1 mu), TPP (1 mM), MgC12 (10 mM), KH2P0, (1 mM), KCI (100 mM), and MOPS (pH 7.3, 50 mM)); (12) substrate (11) solution for GPT (Lalanine (1600 mM), a-ketoglutarate (10 mM) and MOPS (pH 7.3, 50 mM)); (13) converter and controller; (14) recorder; (15) digital meter.
Enzyme Assay. Pyruvate oxidase activity was determined by the method of Harger et al. (13), oxalacetate decarboxylase and GOT activity by the method of Horton and Kornberg (14), and GPT activity by the method of Wroblewski and LaDue (4). Coimmobilization of Oxalacetate Decarboxylase and Pyruvate Oxidase on a Wet Poly(viny1 chloride) Membrane. A PVC/dimethylformamide solution containing 8 wt % PVC was cast on a glass plate (10 x 7 cm) and immersed in methanol at room temperature for 4 h. A wet PVC membrane (thickness ca. 40 Km) obtained was washed with an excess of distilled water (11). Coimmobilization of oxalacetate decarboxylase and pyruvate oxidase on the wet PVC membrane was carried out under the optimum conditions by the adsorption method as follows: to 0.6 mL of an enzyme solution (0.5 mg mL-' oxalacetate decarboxylase, 4.5 mg mL-l pyruvate oxidase, 50 mM phosphate buffer (pH 7.0), 5 mM TPP, 50 mM MgC12,5 mM FAD), the wet PVC membrane (diameter 8 mm) was added and incubated for 24 h at 4 "C. The oxalacetate decarboxylase-pyruvate oxidase-PVC (the immobilized bienzyme-PVC) membrane obtained was washed with a large volume of phosphate buffer and stored in a phosphate buffer at 4 "C. The oxalacetate decarboxylase and pyruvate oxidase activities of the immobilized bienzyme-PVC membrane were 0.36 IU cm-2 and 0.21 IU cm-2, respectively. Assembly of the Sensor. Figure 1 is a schematic diagram of the sensor system. The sensor consisting of the immobilized bienzyme-PVC membrane, cellulose acetate membrane, and a combined platinum electrode (Polarographic electrode, TOA Electronics, Ltd., Tokyo, Japan) was inserted in a small cell (1.0 mL) which was filled with solution that was magnetically stirred. The electrode was maintained at +0.75 V vs. Ag. The current from the sensor was converted to voltage by a 10 M Q resistance and displayed on a recorder (Type 3066, Yokogawa Electric Works Ltd., Tokyo, Japan). The cellulose acetate membrane has such a small pore size that only molecules the size of hydrogen peroxide and water can contact the probe. This excludes the potentially interfering substances from being oxidized at the probe. Procedures. The procedure employed for the determination of GOT activity using the immobilized bienzyme-PVC one was the same as described previously for GPT activity (7). When 50 pL of a GOT standard solution (780IU L-') or control serum was injected into the substrate (I) solution in the cell, a slope of potential differencesbased on current increases against time was measured. The slopes obtained from various concentrations of GOT standard solutions were plotted against GOT activities (12-1700 IU L-l) determined by the method of Wroblewski and LaDue (4) to provide a calibration graph. The sequential determination of GOT and GPT activities was carried out as follows: After the GOT activity was measured by the addition of the substrate (I) solution (No. 11in Figure 1)and serum (50 pL) in the cell as described previously, 250 p L of another substrate (11) solution (No. 12 in Figure 1) including L-alanine (1600 mM) was fed into the cell. The two straight lines were obtained on graph paper with current increase vertically and time and the latter rate (mz)of two horizontally. The former rate (ml) straight lines refer to GOT activity and GOT and GPT activities,
ANALYTICAL CHEMISTRY, VOL. 56, NO. 11, SEPTEMBER 1984
Figure 2. Effect of pyruvate concentrations in the GOT standard solution on the response curves of GOT. Experiments were carried out under standard condtions: GOT standard solution (50 pL, 780 I U L-') and pyruvate Concentratlons (1) 0.0, (2) 0.5, (3) 1.0, and (4) 2.0 mM L-'.
Flgure 3. Effect of pH of the substrate (I) solution on differential current. Experiments were carried out under the standard conditions except for pH; MOPS (50 mM, pH 6.5-8.0) and tricine (50 mM, pH 8.3).
respectively. Therefore, the GPT activity was obtained as a function of the difference of ml and m2. The immobilized bienzyme-PVC membrane in the cell was washed by the substrate (1) solution after measurement and kept in the same solution.
RESULTS AND DISCUSSION Determination of Glutamate-Oxalacetate Transaminase by a Bienzyme Sensor. In order to measure the GOT activity in serum, the response of the sensor was examined. The standard enzyme solution (GOT; 50 pL, 780 IU L-l) was injected into the cell; after the substrate (I) solution including L-aspartate (250 mM), a-ketoglutarate (10 mM) as substrates of the enzymatic reaction was fed into the cell of the sensor. As shown in Figure 2, the current increased with increasing reaction time, initially linearly for more than 5 min. The current increase indicated that pyruvate was produced by a bienzyme immobilized on the PVC membrane and GOT in the standard solution from aspartate and a-ketoglutarate in the substrate (I) solution (eq 1,5, and 6) and consequently that the current based on a production of HzOz increased linearly. The activity of GOT defined as 1pmol of hydrogen peroxide produced per unit time and unit volume of the enzyme solution was represented as a function of a slope of the current-time response, that is, differential current. The experiments were carried out under the standard conditions as cited in the Experimental Section. Endogenous pyruvate in serum must be solved in the determination of transaminase such as GOT and GPT. Figure 2 shows the effect of concentrations of pyruvate in GOT standard solution (780 IU L-l) on the response curve of the bienzyme sensor. The initial output increased with increasing the concentrations of pyruvate in GOT standard solutions. However, no effect of pyruvate concentrations on the slope of the linear region was observed after 30 s, because the production of hydrogen peroxide produced from endogenous pyruvate by the enzymatic reaction and the diffusion of hydrogen peroxide from the solution in the cell to the hydrogen peroxide electrode, via the bienzyme-PVC membrane, were in equilibrium and the output based on endogenous pyruvate added to the GOT standard solution was constant. No endogenous pyruvate in serum interfered with the determination of GOT activity, the same as for GPT activity previously (7). A linear relationship (Yl = (8.34 X 10-3)X1+ 0.45; correlation coefficient 0.99, nine assays) was obtained between differential current ( Yl, nA min-l) within the initial linear region from 30 to 60 s and the various activities of GOT (Xl, IU L-l) standard solutions below 1700 IU L-l. A wide range of the GOT activities, even the high activities found in he-
1-Aspartate concentration trnMi")
Flgure 4. Effect of L-aspartate concentration of the substrate (I) solution on differential current. Experiments were carried out under the standard conditons; L-aspartate concentration (5-300 mM L-l) and GOT standard solution (50 pL, 780 IU L-').
patitis diseases (-1500 IU L-l), can be determined by this electrochemical method. The effect of pH of the substrate (I) solution on the output of the bienzyme sensor is shown in Figure 3. The optimum pH of the bienzyme sensor was in the range from 7.3 to 7.8. The pH stability of pyruvate oxidase immobilized on the wet PVC membrane decreased sharply above pH 7.5, where the poor reproducibility of the bienzyme sensor was observed. Therefore, the buffer at pH 7.3 was used for further works. L-Aspartate, a-ketoglutarate, and phosphate ion as substrates and TPP as a coenzyme of the sequent and enzymatic reactions (eq 1, 5, and 6) are necessary. The effect of concentrations of L-aspartate in the substrate (I) solution on the output of the sensor is shown in Figure 4, when 50 MLof the GOT standard solution (780 IU L-l) was injected into the sensor. In Figure 4, the output of the sensor increased gradually and a steady state was obtained at the concentration of 150 mM of L-aspartate. Enough L-aspartate concentration of 250 mM for enzymatic reactions was employed for further works. The other effects of concentrations of a-ketoglutarate, phosphate ion and TPP on the output of the bienzyme sensor were the almost same as those given in a previous paper (7). Simultaneous Determination of GOT and GPT Activities by the Bienzyme Sensor. The response curve of the bienzyme sensor is shown in Figure 5, when a standard enzyme solution (GOT: 574 IU L-l, GPT; 429 IU L-l) was used. The composition of connected straight lines based on
ANALYTICAL CHEMISTRY, VOL. 56, NO. 11, SEPTEMBER 1984
! 401 0
" ? -
FI ure 5. Response curve of the blenzyme sensor for GOT (574 I U L- ) and GPT (429 I U L-') assay.
hydrogen peroxide produced by the enzymatic reactions was observed. A slope (m,) of the former straight line in Figure 5 refers to the GOT activity as described in this paper. Another slope (m2) of the latter straight line obtained by the addition of the substrate (11)solution involving L-alanine was based on hydrogen peroxide produced by GOT and GPT in the sample solution, via eq 1, 5, and 6 and eq 3 and 6, respectively. Therefore, the GPT activity was measured as a function of the difference of ml and m2 by the computerized and controlled system of the bienzyme sensor. The assay of GOT and GPT activities required less than 4 min. The GOT/GPT standard solutions for the GPT calibration curve were prepared by the addition of various amounts of GPT in the GOT standard solution (GOT; 780 IU L-l). A linear relation ( Yz = (1.21 X 10-2)Xz 0.46, correlation coefficient 0.99, eight assays) was obtained between the difference of m, and m2 (YznA m i d ) and the GOT activity (Xz, IU L-l) below 1600 IU L-l. Also, the relationship between the different current and GOT activity described in this paper could be employed to determine GOT activity in the sequential determination system. In the sequential determination by the bienzyme sensor, the measurable ranges of GOT and GPT activities were from 12 to 1700 IU L-l and from 8 to 1600 IU L-l, respectively. Minimum detectable values of GOT and GPT activities corresponded to 0.6 nA min-l of bienzyme sensor system. This indicates that the bienzyme sensor can be applicable to the determination of GOT and GPT activities in serum. The reproducibilities of GOT and GPT activities were examined each ten times using two kinds of the control sera: one of which was lower activities of GOT (38.0 IU L-l) and GPT (30.8 IU L-l) and another of which was higher ones of GOT (222 IU L-l) and GPT (201 IU L-l). The GOT and GPT activities in sera were reproducible with 3.9% and 2.4% of the coefficient of variation for lower and higher activities of GOT and with 2.4% and 2.3% of the coefficient of variation for lower and higher activities of GPT by the bienzyme sensor, respectively. The bienzyme sensor system was applied to the sequential determination of GOT and GPT activities in sera (25 assays) which were prepared by the appropriate addition of GOT and GPT in control serum. As shown in Figures 6 and 7, a good agreement was obtained between the GOT and GPT activities determined individually by the conventional (spectrophotometric) method ( 4 ) and that obtained sequentially by the bienzyme sensor proposed here. The durability of the bienzyme membrane electrode was investigated by repeated assays of sequential GOT and GPT activities in control sera. The apparent enzyme activity of the bienzyme membrane electrode retained 70% of the initial activity after 400 assays over 14 days. Moreover the immobilized enzyme retained 80% for POP and 85% for OAC of
Y=O.987X+6.8 rz0.99 11-25
Conventional method (l.U,L-')
Flgure 7. Relation between a spectrophotometric (conventional) method and the bienzyme sensor (proposed) method in the determination of GPT activity.
their initial activity after storage at 4 "C for 6 months. The relatively close bienzyme immobilization on the wet PVC membrane led to profound effects that consecutive multistep enzymatic reactions for the GOT assay and parallel multistep enzymatic ones for the GOT and GPT assay proceeded efficiently due to the extremely high local concentration of intermediates and the product (H,O,) in the membrane environment. In addition, inherent properties of the wet PVC membrane ( I I , 1 2 ) such as rapid diffusion of substrates had advantages in the sequential enzymatic reactions. Therefore, the sequential determination of the GOT and GPT activities in serum using the bienzyme-PVC membrane electrode could be performed. Registry No. GOT, 9000-97-9; GPT, 9000-86-6; EC 184.108.40.206, 9024-98-0;EC 220.127.116.11,9001-96-1; poly(viny1 chloride), 9002-86-2; hydrogen peroxide, 7722-84-1.
LITERATURE CITED (1) Rietz, B.; Gullbault, G. G. Anal. Chim. Acta 1975, 77, 191-198. (2) Karmen, A.; Wroblewskl, F.; LaDue, J. S. J . Clin. Invest. 1955, 3 4 , 126-131. (3) Karmen, A. J . Clln. Invest. 1955, 3 4 , 131-133. (4) Wroblewskl, F.; LaDue, J. S. f f o c . SOC.Exp. B i d . M e d . 1956, 9 7 , 569-573. (5) Reitman, S.; Frankel, S. Am. J . Clln. fathol. 1957, 2 8 , 56-61. (6) Rietz, B.; Gullbault, G. G. Clln. Chem. (Winston-Salem, N.C.) 1975, 27, 1544-1546. (7) Khara, K.; Yasukawa, E.; Hayashi, M.; Hirose, S. Anal. Chim. Acta 1984, 1 5 9 , w a e . (8) Yoshio, F.; Osawa, H.; Harada, K. Clln. Chem. (Winston-Salem, N . C . ) 1981, 2 7 , 1098. (9) Mimoura, N.; Yamada, S.; Karube, I . ; Kubo, 1.; Suzuki, S Anal. Chlm. Acta 1982, 132, 355-357.
Anal. Chem. 1984, 56,1880-1884
(10) Mizutani, F.; Sasaki, K.; Shimura, Y. Anal. Chem. 1983, 55, 35-38. (11) Hlrose, S.; Yasukawa, E.; Nose, T. J . Appl. Po/ym. Sci. 1981, 26, 1039- 1046. (12) Hlrose, S.; Yasukawa, E.; Hayashl, M.; Vleth, W. R. J . Membr. Sci. 1982, 11, 177-105. (13) Harger, L. P.; Geller, D. M.; Llpmann, F. Fed. Proc., Fed. Am. Soc. Exp. Biol. 1954, 13, 11-15.
(14) Hotion, A. A.; Kornberg, H. L. Biochim. Biophys. Acta 1984, 89, 381-383.
for review January 277 1984. Accepted April 23, 1984.
Externally Buffered Enzyme Electrode for Determination of Glucose Neil Cleland and Sven-Olof Enfors* Department of Biochemistry and Biotechnology, Royal Institute of Technology, S-100 44 Stockholm, Sweden
A new type of electrode for In situ analysis, the externally buffered enzyme electrode, is presented. I n this system an lmmobllized enzyme Is Immersed In a buffer flow in such a way that the enzyme Is confined to a chamber, with an electrochemical sensor to one side and a dlalysls membrane facing the sample solutlon to the other. While constant chemical conditions are maintained lnslde the enzyme chamber, the buffer flow allows the electrode’s measuring range to be varied through aiteratlons in the buffer flow rate. The system has been applled to glucose determination by uslng glutaraldehyde-Immobilizingglucose oxidase and an amperometric oxygen electrode. Llnear response has been extended from 5 g/L to 150 g/L in phosphate buffer. Havlng an oxygen stablllzation system, the electrode can be used in completely anaeroMc medla. I n this case it has been used in 8 cell-free medium from acetone-butanol fermentation and In corn steep ilquor-based penicillin medium. The electrode Is characterlzed with respect to several Important parameters and the conditlons lnslde the enzyme chamber are dlscussed.
Ever since Clark and Lyons presented the first enzyme electrode in 1962 (I)many workers have been occupied with the development of enzyme electrodes in general (2, 3). Glucose electrodes, due to their great potential applicability, have been subject to particular attention ( 4 , 5 ) . Since there is a strong need for a rapid continuous monitoring of blood glucose in diabetic patients, much work has been focused on development of clinical glucose electrodes for in vivo use (6,
7). For fermentation applications, however, only a few reports on sugar sensors have appeared to our knowledge. One describes an enzyme thermistor device for sucrose analysis (8) and another an enzyme electrode of the self-contained type for glucose analysis (9). The sensors mentioned represent two different concepts: (a) The analytical enzyme reactor in which sample is withdrawn from the medium and pumped to the sensor stie which is outside the fermentation vessel. The sample can be diluted or treated in different ways prior to contact with the sensor. There is a drawback in that the sample must be withdrawn from the fermentation medium, since when the microorganisms grow, they consume substrate fairly rapidly and a t a varying rate during the course of operation. TO prevent alterations in substrate concentration during trans0003-2700/84/0356-1880$01 S O / O
port, continuous dialysis or filtration of the sample is necessary. (b) The enzyme electrode, which can be immersed in the sample and thus has the advantage of being able to measure in situ provided that it is sterilizable. Since in this case the sample is not diluted, the enzyme electrode is severely restricted upward in its linear measuring range which is determined by the intrinsic enzymatic properties, i.e., the apparent Michaelis K,,, of the immobilized enzyme preparation. Also, reaction products may in time build up to detrimental levels inside the enzyme membrane. The enzyme electrode can be said to be sample buffered, since the only buffer capacity available is that of the sample. In this study, a new type of enzyme electrode has been developed, in which some of the advantages of the two aforementioned types are combined: the externally buffered enzyme electrode. I t is of type b but incorporates a flowthrough system so that the enzyme chamber is continuously washed with a buffer solution. Another feature of the enzyme electrode is that is contains a Pt anode for electrolytic oxygen production. This system has been described earlier (9), as well as the use of a Pt anode (though not for O2 production) in clinical p 0 2 and pC02 measurement (10). Though demonstrated here for the case of a glucose electrode, the external buffer concept should be generally applicable to enzyme electrodes.
EXPERIMENTAL SECTION Solutions and Reagents. Unless stated otherwise, the buffer used both for samples and flow-through buffer was 0.025 M Na phosphate buffer of pH 6.0. The penicillin medium had the following composition (g/L): lactose, 10; corn steep liquor (Fermenta AB, Strangnas, Sweden) 30; (NH4)2S04,2; CaC03,5; KH2P04,0.5. The butanol medium contained originally the following compounds: glucose, 40 g/L; yeast extract (Difco Lab., Detroit, MI), 2 g/L; tryptone (Difco Lab., Detroit, MI), 3 g/L; (NH4)S04,2 g/L; KH2P04,2 g/L; K2HP04,2; CoCl,, 1.3 mg/L; Na2Se0,, 90 wg/L; MgSO4.7Hz0,0.1 g/L; CaCl,, 10 mg/L; FeS04.7H2010 mg/L; NazMoO4.2H20,2 mg/L; MnS04.H20,2 mg/L. The medium was kept anaerobic during measurement wth the glucose electrode by bubbling with nitrogen and as at the end of the fermentation with Clostridium acetobutylicum (when all glucose had been consumed) found to contain (g/L): glucose, 0.0; ethanol, 0.61; acetone, 1.84; butanol, 8.33; acetate, 0.54. Stock glucose solution used for additions was 200 g/L. All chemicals were analytical grade. Enzymes and Immobilization. The enzymes and amounts used were 2.5 mg of glucose oxidase (glucose, oxygen oxidoreductase, EC 18.104.22.168) from Aspergillus niger (Worthington, 0 1984 American Chemical Society