- Email: [email protected]
Potentiometric flow injection determination of amylase activity by using hexacyanoferrate(III)-hexacyanoferrate(II) potential buffer
Talanta 45 (1998) 565 – 573
Potentiometric flow injection determination of amylase activity by using hexacyanoferrate(III)-hexacyanoferrate(II) potential buffer Hiroki Ohura a, Toshihiko Imato b,*, Yasukazu Asano c, Sumio Yamasaki a a
Department of Industrial Chemistry, Faculty of Engineering, Kyushu Sangyo Uni6ersity, Matsugadai, Higashi-ku, Fukuoka 813, Japan b Department of Chemical Sciences and Technology, Faculty of Engineering, Kyushu Uni6ersity, Hakozaki, Higashi-ku, Fukuoka 812, Japan c DKK Corporation, Kitamachi, Kichijoji, Musashino, Tokyo 180, Japan Received 3 April 1997; received in revised form 15 August 1997; accepted 18 August 1997
Abstract A highly sensitive potentiometric flow injection determination of amylase activity was carried out, utilizing a redox reaction of hexacyanoferrate(III) in alkaline media with reducing sugar as product of the enzymatic hydrolysis reaction of starch with amylase. The analytical method is based on the potential change detection of a flow-through type redox electrode detector due to the composition change of a [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer solution, which is caused by the redox reaction with the product of the enzymatic reaction. A linear relationship exists between the potential change (peak height) and the activity of amylase. Amylase of a wide activity range from 2.5 ×10 − 2 to 1.2 × 10 − 4 U ml − 1 can be determined by the changing the concentrations of the [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer from 10 − 3 to 10 − 5 M. The lower detection limit of amylase activity is 6.0× 10 − 5 U ml − 1. The sampling rate and relative standard deviation are 15 h − 1 and 0.9% (n= 5) for 3.8 × 10 − 3 U ml − 1 of amylase. The present method was successfully applied to determine amylase activity in real samples (commercial digestive medicines) with an accuracy of 4% compared with analytical results obtained using the present method with those achieved using the conventional titration method. © 1998 Elsevier Science B.V. Keywords: Flow-injection; Amylase activity; [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer; Digestive medicine
1. Introduction An accurate and rapid method for determination of amylase activity in digestive medicines has * Corresponding author. Fax: +81 92 6424134; e-mail: [email protected]
been requested from process and quality control in the pharmaceutical industry. Some standard official testing methods [1] for amylase activity, regulated by the pharmacopoeia of Japan are based on an iodide–thiosulphate titration modified Somogyi method and a spectrophotometric method, where the amount of product of
0039-9140/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved. PII S 0 0 3 9 - 9 1 4 0 ( 9 7 ) 0 0 3 0 3 - 2
566
H. Ohura et al. / Talanta 45 (1998) 565–573
the enzyme reaction of starch with amylase in unit reaction time is determined. The enzyme reactions are carried out at 37°C, pH 5.0 for 10 min. One of these methods is conducted using the procedure illustrated in Scheme 1. The same procedure is required for blank solutions as well as the sample solution. The official methods provide good accuracy but are cumbersome as well as time-consuming requiring about 40 min for the determination of one sample. A promising approach for a more rapid and simpler operation is the application of flow-injection analysis (FIA) techniques [2]. Spectrophotometric determination of fungel a-amylase activity using flow injection analysis has been reported for Aspergillus oryzae fermentations [3]. The method is based on the spectrophotometric measurement of residual starch by formation of an iodine complex, after degradation of starch by the enzyme. Nikolelis et al. [4] has reported the flow injection determination method of amylase using an enzymatic catalyzed reaction yielding glucose as a product, and an electrochemical oxygen detector with a three-electrode amperometric system. This method has been applied to real samples such as blood serum — for amylase. The limit of detection for amylase activity is 0.357 U ml − 1. We have reported a highly sensitive potentiometric flow injection determination of reducing sugars using a [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer solution [5]. The analytical method is based on the detection of change in the composition of potential buffer solution due to a reaction of reducing sugar with [Fe(CN)6]3 − using a flowthrough type redox electrode. The advantages of the present potentiometric method are (1) the electrode potential is very stable since the electrode is immersed in a well-defined potential buffer, (2) samples with a wide concentration range are determinable by appropriately selecting the concentration of the buffer solution. For example, the detection limit was 1×10 − 7 M for glucose and the measurable concentration range of reducing sugars was wide from 1× 10 − 7 M to ca. 0.1 M by appropriately changing the concentration of the potential buffer solution and injection volume [7]. If our FIA method is adapted to determine the reducing sugar generated by the
enzyme reaction of amylase with starch, activity of amylase is expected to be determined rapidly and with high sensitivity. In this paper, we describe an application of our FIA method to determine amylase activity and demonstrate the usefulness of this method for real samples of digestive medicines.
2. Experimental
2.1. Reagents and preparation of solutions Soluble starch was purchased from Wako and was used after drying. Biodiastase 1000, which is the commercial name of a-amylase (EC 3.2.1.1) from Aspergillus oryzae, was donated from Amano Pharmaceutical, and was used as the standard amylase reagent without purification. Digestive medicines containing biodiastase produced by five pharmaceutical companies in Japan, A, B, C, D and E, were purchased from local drug stores. Other reagents were of analytical grade and used as received. A 1% (w/v%) of starch solution was substrate for amylase and prepared according to the procedure of the Japanese Pharmacopoeia Official Testing Method [1]: Dried soluble starch (1 g) was added to ca. 25 ml of 0.4 N sodium hydroxide solution until the solution became a paste-like solution. After heating the solution for 3 min and cooling to room temperature, the solution was neutralized with a 2 N hydrochloric acid solution and was then adjusted to pH 5.0 by acetate buffer solution. The solution was finally made up 100 ml in a volumetric flask with deionized water. Concentration (molarity) of the starch solution was defined as the weight of starch per litre divided by the molecular weight of the monomer unit (glucose). A standard solution of a-amylase was prepared by dissolving 10 mg of biodiastase 1000 with 500 ml of a 0.1 M acetate buffer (pH 5.0) solution containing 6 mM CaCl2 and 20 mM NaCl. The addition of CaCl2 and NaCl to the amylase solution stabilizes enzyme activity. Decrease in activity by self-decomposition was less than 3% in 3 days when stored in a refrigerator. Amylase activity was determined using the official
H. Ohura et al. / Talanta 45 (1998) 565–573
method shown in Scheme 1. One unit of amylase activity was defined as the amount of enzyme that yielded reducing sugar equivalent to 1 mg of reducing sugar from starch min − 1 at pH 5.0 and 37°C g − 1 of solid-amylase. Activity of standard a-amylase solution was 9.2× 103 U g − 1 (0.184 U ml − 1) and was used as a stock solution. The sample solutions were prepared by dissolving 0.473, 0.561, 0.350, 0.463 and 0.325 g, which are the weights of one tablet of digestive medicines obtained from A, B, C, D and E pharmaceutical, respectively, with 500 ml of the same buffer solution used in the above standard solution. Amylase activity, U g-tablet − 1 (U ml − 1), of the digestive medicines from A, B, C, D and E pharmaceutical, respectively, were determined using the official method as being 1.8× 102 (0.168), 1.1× 102 (0.248), 4.1×102 (0.288), 5.5×102 (0.256) and 0.98 ×102 (0.064), respectively. These solutions were stored in a refrigerator as stock solutions. In the FIA method, the stock solution was diluted with the above buffer solution so as to measure activity using the flow system. A stock potential buffer solution consists of 0.1 M [Fe(CN)6]3 − , 0.1 M [Fe(CN)6]4 − and NaOH was prepared according to the procedure previously described [5]. For the sake of simplicity, this buffer is abbreviated to 0.1 M [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer solu-
Scheme 1. Procedure of the official method (modified Somogyi method).
567
tion. Buffer solutions of different concentrations were prepared by serial dilution of the stock solution with deionized water.
2.2. Apparatus The flow-injection apparatus consisted of a peristaltic pump (Model Minipuls 2, Gilson), a six-way valve (HPV-6, GLC Science) with a loop for sample injection (sample volume 140 ml), a flow-through type redox electrode detector which has a platinum-plate electrode and a silver/silver chloride reference electrode (DKK) and a potentiometer (IOC 10, DKK) for measuring the potential of the electrode detector. The potential signal of the detector was fed to a recorder (Model EPR221E, Toa Electronics). The manifold was constructed with Teflon tubing (0.5 mm i.d.) throughout.
2.3. Procedure The flow system consists of three streams, a carrier stream of water (C.S.), a stream of starch solution (R.S.1) and a stream of [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer solution (R.S.2), as shown in Fig. 1(A). A sample solution containing amylase is injected into the carrier stream and merges with the stream of starch solution at the confluence point (C1). The starch is catalytically hydrolyzed by amylase to generate reducing sugar in the reaction tube (R.C.1). This mixed stream is subsequently merged with the stream of potential buffer at the confluent point (C2) and the reducing sugar is oxidized by [Fe(CN)6]3 − the potential buffer in a reaction coil (R.C.2), which changes the composition of the potential buffer solution. The composition change is detected by the redox electrode detector located downstream. The potential change of the detector is measured using the potentiometer and a peak-shaped signal is recorded. Since the reaction time of the injected sample in R.C.1 is kept strictly constant by controlling the pumping rate of solutions, the amount of reducing sugar generated in R.C.1 is related to
568
H. Ohura et al. / Talanta 45 (1998) 565–573
Fig. 1. Schematic diagram of FIA for amylase and analytical peak. (A) Flow injection system: C.S., carrier stream (water); R.S.1, reagent stream 1(10 mM soluble-starch (pH 5.0); R.S.2, reagent stream 2 (1×10 − 4 M [Fe(CN)6]3 − -1 ×10 − 4 M, [Fe(CN)6]4 − , 0.6 M NaOH); P, peristaltic pump; R.C.1, reaction coil 1 (5 m × 0.5 mm i.d.); R.C.2, reaction coil 2 (12 m ×0.5 mm i.d.); C.C., cooling coil (2 m× 0.5 mm i.d.); D, redox electrode detector; C1 and C2, confluence points; length of tube between S and C1, 50 cm× 0.5 mm i.d. (B) Calibration peaks injected every 6 min. (C) Repeated peaks of 3.84 ×10 − 3 U ml − 1 amylase injected every 4 min.
amylase activity, and the peak height of the potential signal is proportional to amylase activity. Amylase activity is determined from the calibration curve based on peak heights. The standard operational parameters for the determination of amylase activity are given in Fig. 1(A), which were optimized by experiments. The flow rates of CS, RS1 and RS2 are set up at 0.25, 0.25 and 0.5 ml min − 1, respectively, for the sake of dispersion suppression of the sample zone and completion of the redox reaction between reducing sugar and [Fe(CN)6]3 − in R.C.2.
Usually the sample solution was injected every 6 min, except for examination of the effect of its time interval on peak heights. It took about 6 min from sample injection to detect a peak maximum. The average residence times of the sample in R.C.1 and R.C.2 were ca. 2 min and ca. 2.5 min, respectively. The temperatures of R.C.1 and R.C.2 were maintained at 37 and 85°C, respectively. The Michaelis parameters, Km and Vm were calculated from the Lineweaver–Burk plot using data obtained from batchwise experiments.
H. Ohura et al. / Talanta 45 (1998) 565–573
2.4. Analytical method Determination of amylase activity by the FIA method, involved the two following reactions, enzymatic reaction of amylase with starch as a substrate depicted in Eq. (1) and the oxidation reaction of reducing sugar by [Fe(CN)6]3 − expressed by Eq. (2) amylase
starch + H2O reducing sugar
(1)
− 3 − OH
reducing sugar +n[Fe(CN)6]
n[Fe(CN)6]4 −
+ oxidation products
(2)
where n is the number of moles of [Fe(CN)6]3 − required to oxidize a mole of reducing sugar [6]. When the reaction of reducing sugar with [Fe(CN)6]3 − in the potential buffer solution goes to completion according to Eq. (2), the potential change (DE) (peak height) of the redox electrode is expressed by the following equation, which is derived from the Nernst equation [7]: DE/mV= − 59 log{1 − n[reducing sugar]/[Fe(CN)6]3o − } /{1 +n[reducing sugar]/[Fe(CN)6]4o − } (3) where [reducing sugar] is the concentration of reducing sugar produced from starch by Eq. (1), and [Fe(CN)6]3o − and [Fe(CN)6]4o − are the initial concentrations of [Fe(CN)6]3 − and [Fe(CN)6]4 − in the potential buffer solution, respectively. The amylase activity is determined from peak heights and the calibration curve which is obtained from the standard solution of injected amylase. The linear relationship between DE and amylase activity of the standard solution, exists where DE is less than 25 mV, according to Eq. (3).
569
official method shown in Scheme 1. This method is a modified Somogyi method. The enzyme reaction with amylase was determined using an iodine/thiosulphate titration method. At first, the activity of the sample solution prepared by diluting the standard solution was measured under reaction conditions at different concentrations of starch. The results are shown in Fig. 2. The amylase activity proportionally increased with concentration of starch solution up to 2 mM and became constant at concentrations higher than 10 mM. This constant value was 0.184 U ml − 1. This identical activity irrespective of starch concentration at higher concentrations than 10 mM indicates that this enzyme reaction obeys a law of zero-order reaction. Fig. 3 shows the Lineweaver–Burk plot for hydrolysis of starch by amylase calculated from the data of Fig. 2. A good linear relationship with a correlation coefficient of 0.99 was observed between the inverse of amylase activity and the inverse of starch concentration. From Fig. 3, Km and Vm values of 2.2 mM and 0.224 U ml − 1 min − 1 were obtained, respectively. This Km value for amylase from Aspergillus oryzae was almost the same as that obtained for amylase from Pseudomonas stutzeri (Km = 1.6 mM) as determined by Sakano et al. [8].
3. Results and discussion
3.1. Determination of amylase acti6ity of standard solution using the official method The determination of amylase activity of the standard solution was performed according to the
Fig. 2. Relationship between amylase activity and concentration of starch solution. Conditions: starch solution; 10 ml of 0.25 – 10mM starch solution, amylase solution: 1 ml of 20 mg l − 1 amylase.
570
H. Ohura et al. / Talanta 45 (1998) 565–573
Fig. 3. Lineweaver–Burk plot for the hydrolysis reaction of starch by amylase. V, amylase activity; S, concentration of starch; Km, Michaelis constant obtained at the intercept of the extrapolated line with the abscissa; Vm, maximum activity obtained at the intercept of the extrapolated line with the ordinate.
3.2. Optimization of analytical procedure of flow system The concentration of sodium hydroxide in the potential buffer and temperature in the reaction coil (R.C.2) are expected to have an affect on the reaction of reducing sugar with [Fe(CN)6]3 − i.e. on the sensitivity of amylase activity determination. The effects of sodium hydroxide concentration in the potential buffer and reaction temperature on the peak height were examined— 1 × 10 − 4 M potential buffer solution was used in the flow system without the stream of R.S.1 in Fig. 1. Moreover, the effect of reaction time on the reactivity between [Fe(CN)6]3 − in the potential buffer and reducing sugar in the R.C.2 was examined at different residense times of the sample zone by changing flow rates of C.S., R.S.1 and R.S.2. Since maltose is the main product of starch hydrolysis by amylase in neutral media [9], maltose was used as reducing sugar sample for optimization of the flow system. The concentration of the maltose sample solution injected was 5× 10 − 6 M. No peak was detected for the potential buffer containing sodium hydroxide at concentrations below 0.05 M. A peak was detected at 0.1 M sodium hydroxide and the peak height increased with increasing concentration of sodium hydroxide. The peak height was almost constant in the
concentration range of 0.6 M–1 M. Above 1 M, however, the peak became broad. This may be due to the increase in viscosity of the mixed solution of R.S.1 and R.S.2 caused by generation of a gel of starch in the reaction coil (R.C.2). No peak was observed when the temperature of the R.C.2 was below 40°C. Above 50°C, the peak height gradually increased and approached a constant value of about 80°C. The baseline potential was observed to shift more or less to the negative potential side, when temperature was over 90°C. This may be due to [Fe(CN)6]3 − in the potential buffer which is partially reduced to [Fe(CN)6]4 − by heating. Thus, the potential buffer solution containing 0.6 M sodium hydroxide and the R.C.2 kept at 85°C were used in subsequent experiments. The peak height was also dependent on the reaction time (residence time of the sample zone in R.C.2) in the oxidation reaction of maltose by [Fe(CN)6]3 − . The peak height gradually increased with increasing reaction time until about 2.5 min and then reached an almost constant value, when 1× 10 − 3 M potential buffer solution containing 0.6 M NaOH was used. This indicates that the reaction between maltose and [Fe(CN)6]3 − finishes in about 2.5 min. The response potential (DE) at a reaction time of 2.5 min was 19.0 mV, and this value corresponded to the n value which was equal to 7.1 as calculated from Eq. (3). This means that 7 mol of [Fe(CN)6]3 − are consumed in the oxidation of 1 mol of maltose under the above reaction conditions. This n value was slightly larger than that reported by Gupta (4–6 mol) obtained under reaction conditions in an ammoniacal medium [10]. Therefore, the flow system shown in Fig. 1(A) was decided to be the reaction time in the R.C.2 of 2.5 min.
3.3. Determination of amylase in digesti6e medicines The standard amylase solutions at different activity, determined previously using the official method, were injected into the carrier stream of the flow system shown in Fig. 1(A) to obtain the calibration curve. In this case, the concentrations of starch in the stream of R.S.1 was varied to
H. Ohura et al. / Talanta 45 (1998) 565–573
Fig. 4. Relationship between peak heights and amylase activities. Starch concentration (mM): (A), 0.5; (B), 1.25; (C), 2.5; (D), 5.0; (E), 10.0; (F), 15.0; (G), 20.0.
examine the concentration dependence on sensitivity, i.e. the slope of the calibration curve. As shown in Fig. 4, a linear relationships exists between the peak height and the amylase activity in the range from 1× 10 − 3 U ml − 1 to 6× 10 − 3 U ml − 1 when the 1× 10 − 4 M potential buffer solution was used. The sensitivity, however, increased with increasing starch concentrations up to 10 mM and was almost constant above 10 mM. Above 30 mM, the sensitivity decreased slightly and the peak became broad. The dependence of
571
starch concentration on sensitivity was similar to that the amylase activity observed in Fig. 2. Table 1 lists the amylase activity obtained using the official method and the sensitivity obtained by the FIA system at different starch concentrations. The typical flow injection peaks with calibration (curve (E) in Fig. 4) are shown in Fig. 1(B). The shape of the peak was broad and the peak width is about twice that compared with the peak width obtained using the same FIA manifold using the stream of water instead of the starch solution in R.S.1. The drifts of baseline potentials were less than 3 mV h − 1. The baseline potential drift for the determination of reducing sugars using the similar FIA manifold without the stream of the starch solution was 0.8 mV h − 1 [5,7]. Therefore the larger baseline potential drift of the present case may be due to the use of starch solution. FIA peaks for repeated injection of a sample of 3.8× 10 − 3 U ml − 1 amylase (n=5) every 4 min are shown in Fig. 1(C) and the relative standard deviation is 0.9% for their peak heights. The repeated injection every 4 min causes 5% positive error due to the overlapping of adjacent peaks in comparison with repeated injections every 6 min. The determination of 15 samples h − 1 was possible with good reproducibility using the proposed FIA method. The sensitivity of the proposed FIA method defined as peak height per activity (slope of the calibration curve, mV U − 1 ml − 1) depended on the concentration of potential buffer, as estimated
Table 1 Dependence of amylase activity and sensitivity on starch concentration Starch Concentration (mM)
Activitya (U ml−1)
Relative activityb
Sensitivityc (×103 mV−1 U−1 ml−1)
Relative sensitivityd
0.5 1.25 2.5 5.0 10.0 15.0 20.0
0.04 0.08 0.128 0.168 0.184 0.184 0.184
0.22 0.44 0.70 0.91 1.0 1.0 1.0
1.15 1.99 3.47 4.69 6.70 7.48 7.14
0.16 0.28 0.49 0.66 0.94 1.05 1.0
a
Amylase activity was obtained by the official method (Fig. 2). Normalized by the acitivity above the starch concentration of 10.0 mM c Sensitivity was obtained using the FIA method (Fig. 4). d Normalized by the sensitivity at the starch concentration of 20.0 mM b
H. Ohura et al. / Talanta 45 (1998) 565–573
572
Table 2 Effect of potential buffer concentration on sensitivity and measurable concentration range of amylase Concentration of the buffer (M)
Sensitivity (mV−1 U−1 ml−1)
Measurable concentration range (U ml−1)
1×10−3 1×10−4 5×10−5 1×10−5
1.0×103 6.6×103 9.9×103 3.9×104
(0.5–2.5)×10−2 (0.8–3.8)×10−3 (0.5–2.5)×10−3 (1.2–6.0)×10−4
from maltose sensitivity to potential buffer concentration. As shown in Table 2, the sensitivity increased by two order of magnitude when the concentration decreased from 10 − 3 M to 10 − 5 M, although the oxidation rate of reducing sugar by [Fe(CN)6]3 − is slower in the case of dilute buffer solution. When a 1×10 − 5 M potential buffer was used, the lower detection limit of amylase activity was 6.0 ×10 − 5 U ml − 1, and this is ca. 200 fold higher in comparison with the sensitivity obtained using other flow methods [3]. The proposed FIA method was applied to the determination of amylase in digestive medicines from five pharmaceutical companies, A, B, C, D and E. The sample solutions of the medicines, whose activities were determined previously using the official method, were injected into the flow system in Fig. 1(A). The amylase activity in the digestive medicine was determined by a peak height observed using the calibration curve obtained using the standard amylase solution. The correlation of analytical results using the official and FIA methods is shown in Fig. 5. From Fig. 5, the regression line expressed by Y =1.04X – (1.17× 10 − 4) with a correlation factor of 0.99 was obtained for all digestive medicines. The amylase content (mg g-tablet − 1) in the digestive medicine calculated using the official method and the proposed method are listed in Table 3. The analytical results obtained by the proposed method is in agreement with those obtained using the official method within relative error of 4.5–8.7%. The proposed FIA system using [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer can
Fig. 5. Comparison between amylase activities obtained using the FIA method and the official method on various commerical digestive medicines. : A, : B, 5: C, %: D, -, E. The solid line is the regression line. The circles are the results of the digestive medicines.
be applied to the determination of amylase in digestive medicines. In conclusion, the proposed FIA method using a combination of [Fe(CN)6]3 − –[Fe(CN)6]4 − potential buffer and the redox electrode detector was more sensitive than other flow methods [3,4], and was successfully applied to the determination of amylase in real samples of digestive medicines. Amylase in a wide activity range between two orders of magnitude was measured by selecting concentrations of potential buffer Table 3 Determination of the amylase content in digestive medicines Amylase content, mg g-tablet−1 Pharmaceutical company
Official method
FIA method
A B C D E
19.3 12.0 44.7 60.1 10.3
18.0 12.7 41.6 57.4 9.4
H. Ohura et al. / Talanta 45 (1998) 565–573
from 10 − 3 M to 10 − 5 M. We did not examine the interference from redox species. Some redox species must exist, which cause some interfering response to the redox electrode detector. Serious interference was not observed in the digestive medicine that was examined in this paper. Therefore, at this stage, applicability of the proposed method is limited to types of samples used in this study.
573
References
The authors are grateful to Mr. Nomura of Amano Pharmaceutical for his kind donation of the standard amylase (Biodiastase 1000) sample used in this work.
[1] The Japanese Pharmacopoeia 13th Technical Information, Yakugyo Jiho, 1966, p. 102. [2] J. Ruzicka, E.H. Hansen, Flow-Injection Analysis, Wiley, New York, 1981. [3] P.W. Hansen, Anal. Chim. Acta 158 (1984) 375. [4] D.P. Nikolelis, H.A. Mottola, Anal. Chem. 50 (1978) 1665. [5] H. Ohura, T. Imato, Y. Asano, S. Yamasaki, N. Ishibashi, Bunseki Kagaku 35 (1986) 807. [6] J.D. Roberts, M.C. Casserio, Basic Principles of Organic Chemistry, Benjamin, New York, 1965, p. 618. [7] H. Ohura, T. Imato, S. Yamaski, N. Ishibashi, Anal. Sci. 3 (1987) 453. [8] Y. Sakano, Y. Kashiwagi, T. Kobayashi, Agric. Biol. Chem. 46 (1982) 639. [9] F. Scheller, F. Schobert, Biosensors, Elsevier, Amsterdam, 1992. [10] K.C. Gupta, A. Sharma, V.D. Misra, Tetrahedron 37 (1981) 2887.
.
.
Acknowledgements
Potentiometric flow injection determination of amylase activity by using hexacyanoferrate(III)-hexacyanoferrate(II) potential buffer Hiroki Ohura a, Toshihiko Imato b,*, Yasukazu Asano c, Sumio Yamasaki a a
Department of Industrial Chemistry, Faculty of Engineering, Kyushu Sangyo Uni6ersity, Matsugadai, Higashi-ku, Fukuoka 813, Japan b Department of Chemical Sciences and Technology, Faculty of Engineering, Kyushu Uni6ersity, Hakozaki, Higashi-ku, Fukuoka 812, Japan c DKK Corporation, Kitamachi, Kichijoji, Musashino, Tokyo 180, Japan Received 3 April 1997; received in revised form 15 August 1997; accepted 18 August 1997
Abstract A highly sensitive potentiometric flow injection determination of amylase activity was carried out, utilizing a redox reaction of hexacyanoferrate(III) in alkaline media with reducing sugar as product of the enzymatic hydrolysis reaction of starch with amylase. The analytical method is based on the potential change detection of a flow-through type redox electrode detector due to the composition change of a [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer solution, which is caused by the redox reaction with the product of the enzymatic reaction. A linear relationship exists between the potential change (peak height) and the activity of amylase. Amylase of a wide activity range from 2.5 ×10 − 2 to 1.2 × 10 − 4 U ml − 1 can be determined by the changing the concentrations of the [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer from 10 − 3 to 10 − 5 M. The lower detection limit of amylase activity is 6.0× 10 − 5 U ml − 1. The sampling rate and relative standard deviation are 15 h − 1 and 0.9% (n= 5) for 3.8 × 10 − 3 U ml − 1 of amylase. The present method was successfully applied to determine amylase activity in real samples (commercial digestive medicines) with an accuracy of 4% compared with analytical results obtained using the present method with those achieved using the conventional titration method. © 1998 Elsevier Science B.V. Keywords: Flow-injection; Amylase activity; [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer; Digestive medicine
1. Introduction An accurate and rapid method for determination of amylase activity in digestive medicines has * Corresponding author. Fax: +81 92 6424134; e-mail: [email protected]
been requested from process and quality control in the pharmaceutical industry. Some standard official testing methods [1] for amylase activity, regulated by the pharmacopoeia of Japan are based on an iodide–thiosulphate titration modified Somogyi method and a spectrophotometric method, where the amount of product of
0039-9140/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved. PII S 0 0 3 9 - 9 1 4 0 ( 9 7 ) 0 0 3 0 3 - 2
566
H. Ohura et al. / Talanta 45 (1998) 565–573
the enzyme reaction of starch with amylase in unit reaction time is determined. The enzyme reactions are carried out at 37°C, pH 5.0 for 10 min. One of these methods is conducted using the procedure illustrated in Scheme 1. The same procedure is required for blank solutions as well as the sample solution. The official methods provide good accuracy but are cumbersome as well as time-consuming requiring about 40 min for the determination of one sample. A promising approach for a more rapid and simpler operation is the application of flow-injection analysis (FIA) techniques [2]. Spectrophotometric determination of fungel a-amylase activity using flow injection analysis has been reported for Aspergillus oryzae fermentations [3]. The method is based on the spectrophotometric measurement of residual starch by formation of an iodine complex, after degradation of starch by the enzyme. Nikolelis et al. [4] has reported the flow injection determination method of amylase using an enzymatic catalyzed reaction yielding glucose as a product, and an electrochemical oxygen detector with a three-electrode amperometric system. This method has been applied to real samples such as blood serum — for amylase. The limit of detection for amylase activity is 0.357 U ml − 1. We have reported a highly sensitive potentiometric flow injection determination of reducing sugars using a [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer solution [5]. The analytical method is based on the detection of change in the composition of potential buffer solution due to a reaction of reducing sugar with [Fe(CN)6]3 − using a flowthrough type redox electrode. The advantages of the present potentiometric method are (1) the electrode potential is very stable since the electrode is immersed in a well-defined potential buffer, (2) samples with a wide concentration range are determinable by appropriately selecting the concentration of the buffer solution. For example, the detection limit was 1×10 − 7 M for glucose and the measurable concentration range of reducing sugars was wide from 1× 10 − 7 M to ca. 0.1 M by appropriately changing the concentration of the potential buffer solution and injection volume [7]. If our FIA method is adapted to determine the reducing sugar generated by the
enzyme reaction of amylase with starch, activity of amylase is expected to be determined rapidly and with high sensitivity. In this paper, we describe an application of our FIA method to determine amylase activity and demonstrate the usefulness of this method for real samples of digestive medicines.
2. Experimental
2.1. Reagents and preparation of solutions Soluble starch was purchased from Wako and was used after drying. Biodiastase 1000, which is the commercial name of a-amylase (EC 3.2.1.1) from Aspergillus oryzae, was donated from Amano Pharmaceutical, and was used as the standard amylase reagent without purification. Digestive medicines containing biodiastase produced by five pharmaceutical companies in Japan, A, B, C, D and E, were purchased from local drug stores. Other reagents were of analytical grade and used as received. A 1% (w/v%) of starch solution was substrate for amylase and prepared according to the procedure of the Japanese Pharmacopoeia Official Testing Method [1]: Dried soluble starch (1 g) was added to ca. 25 ml of 0.4 N sodium hydroxide solution until the solution became a paste-like solution. After heating the solution for 3 min and cooling to room temperature, the solution was neutralized with a 2 N hydrochloric acid solution and was then adjusted to pH 5.0 by acetate buffer solution. The solution was finally made up 100 ml in a volumetric flask with deionized water. Concentration (molarity) of the starch solution was defined as the weight of starch per litre divided by the molecular weight of the monomer unit (glucose). A standard solution of a-amylase was prepared by dissolving 10 mg of biodiastase 1000 with 500 ml of a 0.1 M acetate buffer (pH 5.0) solution containing 6 mM CaCl2 and 20 mM NaCl. The addition of CaCl2 and NaCl to the amylase solution stabilizes enzyme activity. Decrease in activity by self-decomposition was less than 3% in 3 days when stored in a refrigerator. Amylase activity was determined using the official
H. Ohura et al. / Talanta 45 (1998) 565–573
method shown in Scheme 1. One unit of amylase activity was defined as the amount of enzyme that yielded reducing sugar equivalent to 1 mg of reducing sugar from starch min − 1 at pH 5.0 and 37°C g − 1 of solid-amylase. Activity of standard a-amylase solution was 9.2× 103 U g − 1 (0.184 U ml − 1) and was used as a stock solution. The sample solutions were prepared by dissolving 0.473, 0.561, 0.350, 0.463 and 0.325 g, which are the weights of one tablet of digestive medicines obtained from A, B, C, D and E pharmaceutical, respectively, with 500 ml of the same buffer solution used in the above standard solution. Amylase activity, U g-tablet − 1 (U ml − 1), of the digestive medicines from A, B, C, D and E pharmaceutical, respectively, were determined using the official method as being 1.8× 102 (0.168), 1.1× 102 (0.248), 4.1×102 (0.288), 5.5×102 (0.256) and 0.98 ×102 (0.064), respectively. These solutions were stored in a refrigerator as stock solutions. In the FIA method, the stock solution was diluted with the above buffer solution so as to measure activity using the flow system. A stock potential buffer solution consists of 0.1 M [Fe(CN)6]3 − , 0.1 M [Fe(CN)6]4 − and NaOH was prepared according to the procedure previously described [5]. For the sake of simplicity, this buffer is abbreviated to 0.1 M [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer solu-
Scheme 1. Procedure of the official method (modified Somogyi method).
567
tion. Buffer solutions of different concentrations were prepared by serial dilution of the stock solution with deionized water.
2.2. Apparatus The flow-injection apparatus consisted of a peristaltic pump (Model Minipuls 2, Gilson), a six-way valve (HPV-6, GLC Science) with a loop for sample injection (sample volume 140 ml), a flow-through type redox electrode detector which has a platinum-plate electrode and a silver/silver chloride reference electrode (DKK) and a potentiometer (IOC 10, DKK) for measuring the potential of the electrode detector. The potential signal of the detector was fed to a recorder (Model EPR221E, Toa Electronics). The manifold was constructed with Teflon tubing (0.5 mm i.d.) throughout.
2.3. Procedure The flow system consists of three streams, a carrier stream of water (C.S.), a stream of starch solution (R.S.1) and a stream of [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer solution (R.S.2), as shown in Fig. 1(A). A sample solution containing amylase is injected into the carrier stream and merges with the stream of starch solution at the confluence point (C1). The starch is catalytically hydrolyzed by amylase to generate reducing sugar in the reaction tube (R.C.1). This mixed stream is subsequently merged with the stream of potential buffer at the confluent point (C2) and the reducing sugar is oxidized by [Fe(CN)6]3 − the potential buffer in a reaction coil (R.C.2), which changes the composition of the potential buffer solution. The composition change is detected by the redox electrode detector located downstream. The potential change of the detector is measured using the potentiometer and a peak-shaped signal is recorded. Since the reaction time of the injected sample in R.C.1 is kept strictly constant by controlling the pumping rate of solutions, the amount of reducing sugar generated in R.C.1 is related to
568
H. Ohura et al. / Talanta 45 (1998) 565–573
Fig. 1. Schematic diagram of FIA for amylase and analytical peak. (A) Flow injection system: C.S., carrier stream (water); R.S.1, reagent stream 1(10 mM soluble-starch (pH 5.0); R.S.2, reagent stream 2 (1×10 − 4 M [Fe(CN)6]3 − -1 ×10 − 4 M, [Fe(CN)6]4 − , 0.6 M NaOH); P, peristaltic pump; R.C.1, reaction coil 1 (5 m × 0.5 mm i.d.); R.C.2, reaction coil 2 (12 m ×0.5 mm i.d.); C.C., cooling coil (2 m× 0.5 mm i.d.); D, redox electrode detector; C1 and C2, confluence points; length of tube between S and C1, 50 cm× 0.5 mm i.d. (B) Calibration peaks injected every 6 min. (C) Repeated peaks of 3.84 ×10 − 3 U ml − 1 amylase injected every 4 min.
amylase activity, and the peak height of the potential signal is proportional to amylase activity. Amylase activity is determined from the calibration curve based on peak heights. The standard operational parameters for the determination of amylase activity are given in Fig. 1(A), which were optimized by experiments. The flow rates of CS, RS1 and RS2 are set up at 0.25, 0.25 and 0.5 ml min − 1, respectively, for the sake of dispersion suppression of the sample zone and completion of the redox reaction between reducing sugar and [Fe(CN)6]3 − in R.C.2.
Usually the sample solution was injected every 6 min, except for examination of the effect of its time interval on peak heights. It took about 6 min from sample injection to detect a peak maximum. The average residence times of the sample in R.C.1 and R.C.2 were ca. 2 min and ca. 2.5 min, respectively. The temperatures of R.C.1 and R.C.2 were maintained at 37 and 85°C, respectively. The Michaelis parameters, Km and Vm were calculated from the Lineweaver–Burk plot using data obtained from batchwise experiments.
H. Ohura et al. / Talanta 45 (1998) 565–573
2.4. Analytical method Determination of amylase activity by the FIA method, involved the two following reactions, enzymatic reaction of amylase with starch as a substrate depicted in Eq. (1) and the oxidation reaction of reducing sugar by [Fe(CN)6]3 − expressed by Eq. (2) amylase
starch + H2O reducing sugar
(1)
− 3 − OH
reducing sugar +n[Fe(CN)6]
n[Fe(CN)6]4 −
+ oxidation products
(2)
where n is the number of moles of [Fe(CN)6]3 − required to oxidize a mole of reducing sugar [6]. When the reaction of reducing sugar with [Fe(CN)6]3 − in the potential buffer solution goes to completion according to Eq. (2), the potential change (DE) (peak height) of the redox electrode is expressed by the following equation, which is derived from the Nernst equation [7]: DE/mV= − 59 log{1 − n[reducing sugar]/[Fe(CN)6]3o − } /{1 +n[reducing sugar]/[Fe(CN)6]4o − } (3) where [reducing sugar] is the concentration of reducing sugar produced from starch by Eq. (1), and [Fe(CN)6]3o − and [Fe(CN)6]4o − are the initial concentrations of [Fe(CN)6]3 − and [Fe(CN)6]4 − in the potential buffer solution, respectively. The amylase activity is determined from peak heights and the calibration curve which is obtained from the standard solution of injected amylase. The linear relationship between DE and amylase activity of the standard solution, exists where DE is less than 25 mV, according to Eq. (3).
569
official method shown in Scheme 1. This method is a modified Somogyi method. The enzyme reaction with amylase was determined using an iodine/thiosulphate titration method. At first, the activity of the sample solution prepared by diluting the standard solution was measured under reaction conditions at different concentrations of starch. The results are shown in Fig. 2. The amylase activity proportionally increased with concentration of starch solution up to 2 mM and became constant at concentrations higher than 10 mM. This constant value was 0.184 U ml − 1. This identical activity irrespective of starch concentration at higher concentrations than 10 mM indicates that this enzyme reaction obeys a law of zero-order reaction. Fig. 3 shows the Lineweaver–Burk plot for hydrolysis of starch by amylase calculated from the data of Fig. 2. A good linear relationship with a correlation coefficient of 0.99 was observed between the inverse of amylase activity and the inverse of starch concentration. From Fig. 3, Km and Vm values of 2.2 mM and 0.224 U ml − 1 min − 1 were obtained, respectively. This Km value for amylase from Aspergillus oryzae was almost the same as that obtained for amylase from Pseudomonas stutzeri (Km = 1.6 mM) as determined by Sakano et al. [8].
3. Results and discussion
3.1. Determination of amylase acti6ity of standard solution using the official method The determination of amylase activity of the standard solution was performed according to the
Fig. 2. Relationship between amylase activity and concentration of starch solution. Conditions: starch solution; 10 ml of 0.25 – 10mM starch solution, amylase solution: 1 ml of 20 mg l − 1 amylase.
570
H. Ohura et al. / Talanta 45 (1998) 565–573
Fig. 3. Lineweaver–Burk plot for the hydrolysis reaction of starch by amylase. V, amylase activity; S, concentration of starch; Km, Michaelis constant obtained at the intercept of the extrapolated line with the abscissa; Vm, maximum activity obtained at the intercept of the extrapolated line with the ordinate.
3.2. Optimization of analytical procedure of flow system The concentration of sodium hydroxide in the potential buffer and temperature in the reaction coil (R.C.2) are expected to have an affect on the reaction of reducing sugar with [Fe(CN)6]3 − i.e. on the sensitivity of amylase activity determination. The effects of sodium hydroxide concentration in the potential buffer and reaction temperature on the peak height were examined— 1 × 10 − 4 M potential buffer solution was used in the flow system without the stream of R.S.1 in Fig. 1. Moreover, the effect of reaction time on the reactivity between [Fe(CN)6]3 − in the potential buffer and reducing sugar in the R.C.2 was examined at different residense times of the sample zone by changing flow rates of C.S., R.S.1 and R.S.2. Since maltose is the main product of starch hydrolysis by amylase in neutral media [9], maltose was used as reducing sugar sample for optimization of the flow system. The concentration of the maltose sample solution injected was 5× 10 − 6 M. No peak was detected for the potential buffer containing sodium hydroxide at concentrations below 0.05 M. A peak was detected at 0.1 M sodium hydroxide and the peak height increased with increasing concentration of sodium hydroxide. The peak height was almost constant in the
concentration range of 0.6 M–1 M. Above 1 M, however, the peak became broad. This may be due to the increase in viscosity of the mixed solution of R.S.1 and R.S.2 caused by generation of a gel of starch in the reaction coil (R.C.2). No peak was observed when the temperature of the R.C.2 was below 40°C. Above 50°C, the peak height gradually increased and approached a constant value of about 80°C. The baseline potential was observed to shift more or less to the negative potential side, when temperature was over 90°C. This may be due to [Fe(CN)6]3 − in the potential buffer which is partially reduced to [Fe(CN)6]4 − by heating. Thus, the potential buffer solution containing 0.6 M sodium hydroxide and the R.C.2 kept at 85°C were used in subsequent experiments. The peak height was also dependent on the reaction time (residence time of the sample zone in R.C.2) in the oxidation reaction of maltose by [Fe(CN)6]3 − . The peak height gradually increased with increasing reaction time until about 2.5 min and then reached an almost constant value, when 1× 10 − 3 M potential buffer solution containing 0.6 M NaOH was used. This indicates that the reaction between maltose and [Fe(CN)6]3 − finishes in about 2.5 min. The response potential (DE) at a reaction time of 2.5 min was 19.0 mV, and this value corresponded to the n value which was equal to 7.1 as calculated from Eq. (3). This means that 7 mol of [Fe(CN)6]3 − are consumed in the oxidation of 1 mol of maltose under the above reaction conditions. This n value was slightly larger than that reported by Gupta (4–6 mol) obtained under reaction conditions in an ammoniacal medium [10]. Therefore, the flow system shown in Fig. 1(A) was decided to be the reaction time in the R.C.2 of 2.5 min.
3.3. Determination of amylase in digesti6e medicines The standard amylase solutions at different activity, determined previously using the official method, were injected into the carrier stream of the flow system shown in Fig. 1(A) to obtain the calibration curve. In this case, the concentrations of starch in the stream of R.S.1 was varied to
H. Ohura et al. / Talanta 45 (1998) 565–573
Fig. 4. Relationship between peak heights and amylase activities. Starch concentration (mM): (A), 0.5; (B), 1.25; (C), 2.5; (D), 5.0; (E), 10.0; (F), 15.0; (G), 20.0.
examine the concentration dependence on sensitivity, i.e. the slope of the calibration curve. As shown in Fig. 4, a linear relationships exists between the peak height and the amylase activity in the range from 1× 10 − 3 U ml − 1 to 6× 10 − 3 U ml − 1 when the 1× 10 − 4 M potential buffer solution was used. The sensitivity, however, increased with increasing starch concentrations up to 10 mM and was almost constant above 10 mM. Above 30 mM, the sensitivity decreased slightly and the peak became broad. The dependence of
571
starch concentration on sensitivity was similar to that the amylase activity observed in Fig. 2. Table 1 lists the amylase activity obtained using the official method and the sensitivity obtained by the FIA system at different starch concentrations. The typical flow injection peaks with calibration (curve (E) in Fig. 4) are shown in Fig. 1(B). The shape of the peak was broad and the peak width is about twice that compared with the peak width obtained using the same FIA manifold using the stream of water instead of the starch solution in R.S.1. The drifts of baseline potentials were less than 3 mV h − 1. The baseline potential drift for the determination of reducing sugars using the similar FIA manifold without the stream of the starch solution was 0.8 mV h − 1 [5,7]. Therefore the larger baseline potential drift of the present case may be due to the use of starch solution. FIA peaks for repeated injection of a sample of 3.8× 10 − 3 U ml − 1 amylase (n=5) every 4 min are shown in Fig. 1(C) and the relative standard deviation is 0.9% for their peak heights. The repeated injection every 4 min causes 5% positive error due to the overlapping of adjacent peaks in comparison with repeated injections every 6 min. The determination of 15 samples h − 1 was possible with good reproducibility using the proposed FIA method. The sensitivity of the proposed FIA method defined as peak height per activity (slope of the calibration curve, mV U − 1 ml − 1) depended on the concentration of potential buffer, as estimated
Table 1 Dependence of amylase activity and sensitivity on starch concentration Starch Concentration (mM)
Activitya (U ml−1)
Relative activityb
Sensitivityc (×103 mV−1 U−1 ml−1)
Relative sensitivityd
0.5 1.25 2.5 5.0 10.0 15.0 20.0
0.04 0.08 0.128 0.168 0.184 0.184 0.184
0.22 0.44 0.70 0.91 1.0 1.0 1.0
1.15 1.99 3.47 4.69 6.70 7.48 7.14
0.16 0.28 0.49 0.66 0.94 1.05 1.0
a
Amylase activity was obtained by the official method (Fig. 2). Normalized by the acitivity above the starch concentration of 10.0 mM c Sensitivity was obtained using the FIA method (Fig. 4). d Normalized by the sensitivity at the starch concentration of 20.0 mM b
H. Ohura et al. / Talanta 45 (1998) 565–573
572
Table 2 Effect of potential buffer concentration on sensitivity and measurable concentration range of amylase Concentration of the buffer (M)
Sensitivity (mV−1 U−1 ml−1)
Measurable concentration range (U ml−1)
1×10−3 1×10−4 5×10−5 1×10−5
1.0×103 6.6×103 9.9×103 3.9×104
(0.5–2.5)×10−2 (0.8–3.8)×10−3 (0.5–2.5)×10−3 (1.2–6.0)×10−4
from maltose sensitivity to potential buffer concentration. As shown in Table 2, the sensitivity increased by two order of magnitude when the concentration decreased from 10 − 3 M to 10 − 5 M, although the oxidation rate of reducing sugar by [Fe(CN)6]3 − is slower in the case of dilute buffer solution. When a 1×10 − 5 M potential buffer was used, the lower detection limit of amylase activity was 6.0 ×10 − 5 U ml − 1, and this is ca. 200 fold higher in comparison with the sensitivity obtained using other flow methods [3]. The proposed FIA method was applied to the determination of amylase in digestive medicines from five pharmaceutical companies, A, B, C, D and E. The sample solutions of the medicines, whose activities were determined previously using the official method, were injected into the flow system in Fig. 1(A). The amylase activity in the digestive medicine was determined by a peak height observed using the calibration curve obtained using the standard amylase solution. The correlation of analytical results using the official and FIA methods is shown in Fig. 5. From Fig. 5, the regression line expressed by Y =1.04X – (1.17× 10 − 4) with a correlation factor of 0.99 was obtained for all digestive medicines. The amylase content (mg g-tablet − 1) in the digestive medicine calculated using the official method and the proposed method are listed in Table 3. The analytical results obtained by the proposed method is in agreement with those obtained using the official method within relative error of 4.5–8.7%. The proposed FIA system using [Fe(CN)6]3 − – [Fe(CN)6]4 − potential buffer can
Fig. 5. Comparison between amylase activities obtained using the FIA method and the official method on various commerical digestive medicines. : A, : B, 5: C, %: D, -, E. The solid line is the regression line. The circles are the results of the digestive medicines.
be applied to the determination of amylase in digestive medicines. In conclusion, the proposed FIA method using a combination of [Fe(CN)6]3 − –[Fe(CN)6]4 − potential buffer and the redox electrode detector was more sensitive than other flow methods [3,4], and was successfully applied to the determination of amylase in real samples of digestive medicines. Amylase in a wide activity range between two orders of magnitude was measured by selecting concentrations of potential buffer Table 3 Determination of the amylase content in digestive medicines Amylase content, mg g-tablet−1 Pharmaceutical company
Official method
FIA method
A B C D E
19.3 12.0 44.7 60.1 10.3
18.0 12.7 41.6 57.4 9.4
H. Ohura et al. / Talanta 45 (1998) 565–573
from 10 − 3 M to 10 − 5 M. We did not examine the interference from redox species. Some redox species must exist, which cause some interfering response to the redox electrode detector. Serious interference was not observed in the digestive medicine that was examined in this paper. Therefore, at this stage, applicability of the proposed method is limited to types of samples used in this study.
573
References
The authors are grateful to Mr. Nomura of Amano Pharmaceutical for his kind donation of the standard amylase (Biodiastase 1000) sample used in this work.
[1] The Japanese Pharmacopoeia 13th Technical Information, Yakugyo Jiho, 1966, p. 102. [2] J. Ruzicka, E.H. Hansen, Flow-Injection Analysis, Wiley, New York, 1981. [3] P.W. Hansen, Anal. Chim. Acta 158 (1984) 375. [4] D.P. Nikolelis, H.A. Mottola, Anal. Chem. 50 (1978) 1665. [5] H. Ohura, T. Imato, Y. Asano, S. Yamasaki, N. Ishibashi, Bunseki Kagaku 35 (1986) 807. [6] J.D. Roberts, M.C. Casserio, Basic Principles of Organic Chemistry, Benjamin, New York, 1965, p. 618. [7] H. Ohura, T. Imato, S. Yamaski, N. Ishibashi, Anal. Sci. 3 (1987) 453. [8] Y. Sakano, Y. Kashiwagi, T. Kobayashi, Agric. Biol. Chem. 46 (1982) 639. [9] F. Scheller, F. Schobert, Biosensors, Elsevier, Amsterdam, 1992. [10] K.C. Gupta, A. Sharma, V.D. Misra, Tetrahedron 37 (1981) 2887.
.
.
Acknowledgements