Development of a bienzymic graphite–Teflon composite electrode for the determination of hypoxanthine in fish
G. Cayuela, N. Pe ˜na, A. J. Reviejo and J. M. Pingarr´on*
Department of Analytical Chemistry, Faculty of Chemistry, Complutense University of Madrid, 28040-Madrid, Spain
A bienzymic amperometric graphite–Teflon composite biosensor, in which xanthine oxidase and peroxidase, together with the mediator ferrocene, are incorporated into the electrode matrix, was developed for the determination of hypoxanthine in fish samples. These biosensors were fabricated by simple physical inclusion of the enzymes and the mediator in the bulk of the
graphite–Teflon matrix. A Teflon content of 60%, a working potential of 0.00 V, a temperature of 30 ± 1 °C and a pH of 7.4 were selected as the optimum working conditions. The composite bioelectrode operated over long periods owing to the renewability of its surface by polishing. Reproducible amperometric responses were achieved with different electrodes fabricated from different composite matrices, and no significant loss of the enzyme activity was observed after more than 6 months of storage at 4 °C. A detection limit for hypoxanthine of 9.0 31028mol l21was obtained by amperometry in stirred solutions. An interference study of different substances which may be present in samples together with
hypoxanthine demonstrated very good selectivity for the determination of this analyte. The bienzymic composite biosensor was applied to the determination of
hypoxanthine in sardine muscle tissue, and the method was validated by comparing the results with those obtained by applying a recommended reference method.
Keywords: Graphite–Teflon composite enzyme electrodes;
hypoxanthine; fish
The major contribution to the well known rapid alteration of the organoleptic and health qualities of fish after death is that due to non-proteic nitrogen components.1 Among them, nucleotides such as ATP are among those most affected by the degradation of fresh fish. In this degradation process, purine bases such as hypoxanthine (Hx) are produced,1 and therefore its detection and quantification can be used as an indication of the fish freshness. In this context, different indexes, in which the concentration of Hx is one of the parameters included, have been proposed for the control of fish freshness.2,3Hence the determination of Hx is usually carried out in the course of the analysis of fish and fish products.4Most of the methods used for the determination of Hx are based on the enzymic reaction of the Hx oxidation catalysed by xanthine oxidase (XOD), and different biosensors have been developed for this applica- tion5–13 and for the determination of the freshness index in fish.14–19
We have demonstrated recently that graphite–Teflon com- posite enzyme electrodes, constructed by simple physical inclusion of the enzyme in the bulk of a graphite–Teflon pellet with no need for covalent attachments, constitute robust and practicable amperometric biosensors under both batch and flow-through conditions.20Furthermore, these composite bio- electrodes are suitable for working in predominantly non- aqueous media such as reversed micelles and acetonitrile–water mixtures.21 They exhibit fairly good renewability of the
electrode surface by polishing and long-term operation and stability, and offer the possibility of easy co-immobilization of other components (mediators, other enzymes, etc.) into the electrode matrix. These advantageous properties were exploited in this work for the development of a bienzymic amperometric composite biosensor, in which XOD and horseradish peroxidase (HRP), together with the mediator ferrocene, were incorporated into the graphite–Teflon matrix, useful for the determination of hypoxanthine in fish samples. Different examples of peroxidase electrode designs and their application can be found in a review by Ruzgas et al.22
Experimental
Apparatus, electrodes and electrochemical cells
Experiments were performed on a Metrohm (Herisau, Switzer- land) 641 VA potentiostat connected to a Linseis (Selb, Germany) L6512B recorder. The electrochemical cell was a BAS (W. Lafayette, IN, USA) Model VC-2 cell with a BAS RE- 1 Ag/AgCl/KCl (3 m) reference electrode and a platinum wire auxiliary electrode. A Meditronic (Barcelona, Spain) P-Selecta centrifuge, a P-Selecta Ultrasons ultrasonic bath and a P-Selecta thermostatic bath were also used.
Graphite–Teflon–XOD–HRP–ferrocene composite electrode Composite bienzymic electrodes were fabricated in the form of cylindrical pellets as follows. Graphite (ultra F purity; Carbon of America, Bay City, MI, USA), 0.25 g, XOD (EC 1.1.3.22, activity 0.067 units per mg of solid; Sigma, St. Louis, MO, USA), 0.1776 g, and HRP (EC 1.11.1.7, type II, activity 180 units per mg of solid, Sigma), 0.0089 g, were accurately weighed and thoroughly mixed by mechanical stirring for 2 h in a 0.30 ml suspension of a 0.05 mol l21phosphate buffer solution of pH 7.4 at 4 °C. Water was then evaporated by passing an argon stream through the mixture and 0.0096 g of ferrocene (Fluka, Buchs, Switzerland), dissolved in 0.3 ml of ethyl acetate (Aldrich, Milwaukee, WI, USA), was added. The resulting mixture was shaken mechanically for 15 min and then ethyl acetate was evaporated by passing an argon stream. Next, the appropriate amount of Teflon powder (Aldrich) to obtain the desired final Teflon content (0.36 g for a Teflon content of 60%) was added and mixed thoroughly by hand. The mixture was pressed into pellets by using a Carver pellet press (supplied by Perkin-Elmer, Norwalk, CT, USA) at 10 000 kg cm22for 10 min. These pellets were of 1.3 cm diameter and approximately 0.4 cm thick. Several 3.0 mm diameter cylindrical portions of the pellet were bored, and each portion was press-fitted into a Teflon holder. Electrical contact was made through a stainless- steel flat-tipped screw.
Reagents and solutions
Reagents used, apart from those mentioned above, were hypoxanthine (Fluka), d-(+)-galactose (Sigma Ultra), l-methio- nine (Sigma Ultra), l-alanine (Merck, Darmstadt, Germany), Analyst, February 1998, Vol. 123 (371–377) 371
ascorbic acid (Merck) and oxalic acid (Probus, Barcelona, Spain). All chemicals were of analytical-reagent grade and the water used was obtained from a Milli-Q purification system (Millipore, Bedford, MA, USA).
A stock standard solution of hypoxanthine (0.010 mol l21) was prepared in 0.01 mol l21NaOH with the help of ultrasonic stirring. Working standard solutions were prepared by suitable dilution with 0.05 mol l21phosphate buffer solution of pH 7.4.
A stock standard solution of XOD (0.064 U ml21) was also prepared by dissolving 0.0038 g of the enzyme in 4.0 ml of the 0.05 mol l21phosphate buffer solution.
Procedures
The biocomposite electrode was immersed in the electro- chemical cell and amperometric measurements in stirred solutions were performed by applying the desired potential and allowing the steady-state current to be reached. Solutions were maintained at the desired temperature by means of a thermo- static bath. When the response obtained with the composite bioelectrode was significantly lower than the original response (after having been used for a period of time), regeneration of the electrode surface was performed by polishing for 5 s on 150 grit SiC paper. After use, the composite bioelectrode was stored at 4 °C in a refrigerator.
Determination of hypoxanthine in sardine muscle tissue The procedure employed is a modification of the recommended method for the analysis of hypoxanthine by the Fish Products Sub-committee of the Analytical Methods Committee,4which has been used as the reference method for validation of the methodology proposed in this work.
About 5 g of sample were accurately weighed, then the sample was macerated with 50 ml of 0.6 mol l21perchloric acid for 1 h. After centrifugation at 4000 rpm for 1 h, an aliquot of 30 ml of the supernatant extract was transferred into an electrochemical cell containing 5.0 ml of a 0.05 mol l21 phosphate buffer solution of pH 7.4. Amperometry in stirred solutions was carried out at 0.00 V. Determination of hypox- anthine was performed by applying the standard additions method, which involved the addition of successive 3 ml aliquots of a 5.0 31023mol l21hypoxanthine standard solution.
Results and discussion
The biocatalytic scheme used in this work involved the oxidation of hypoxanthine catalysed by XOD. Uric acid and hydrogen peroxide are formed as products in this reaction, the reduction of H2O2 catalysed by peroxidase being used for monitoring the oxidation of hypoxanthine. Ferrocene was selected as a mediator for this latter enzyme reaction because of its poor solubility in the working medium used. Consequently, the amperometric signal employed for monitoring the process corresponded to the electrochemical reduction of ferricinium.
Optimization of the percentage of Teflon in the electrode matrix
As a previous step to the development of the bienzymic electrode, the content of the insulator material, Teflon, in the electrode matrix was optimized by constructing composite electrodes containing only XOD. In this way, only the main enzyme reaction, i.e., the oxidation of hypoxanthine to uric acid catalysed by XOD, was monitored. This monitoring was accomplished by means of the electrochemical oxidation of the uric acid generated as a product in the enzymic reaction, the choice of the applied potential for detecting this oxidation being made from the cyclic voltammogram of a 1.0 31023mol l21
uric acid solution in phosphate buffer of pH 7.4 (Fig. 1). As can be seen, an oxidation peak was obtained at 0.50 V. The contribution of H2O2, which is also formed as a product in the enzymic reaction, to this response can be neglected, as can be deduced from voltammograms (b) and (c) in Fig. 1 obtained from solutions containing 1021 and 1023 mol l21 H2O2, respectively. Therefore, a potential of 0.60 V was selected for the amperometric monitoring of the reaction.
Regarding the amount of XOD used, preliminary studies were made by increasing the enzyme loading into the electrode.
As mentioned in the Introduction, these electrodes were fabricated by simple physical inclusion of the enzyme in the bulk of the electrode with no need for covalent attachments, which makes the fabrication procedure easier, faster and cheaper, and avoids the possible loss of sensitivity due to the covalent linkages. The amount of XOD selected (see Experi- mental) corresponded to the enzyme loading at which suffi- ciently sensitive amperometric signals in stirred solutions were obtained for hypoxanthine. Amounts of enzyme larger than that selected did not yield noticeable increases in the substrate electroanalytical response.
The amperometric steady-state current for 4.0 3 1025 mol l21hypoxanthine solutions remained virtually constant for electrodes containing 40, 50 and 60% Teflon. However, this current decreased sharply when the composite electrode was fabricated with 70% Teflon. On the other hand, the RSD for sets of 10 current measurements decreased from 7.9% for a 40%
Teflon electrode to 4.0% for an electrode containing 60% of insulator material. Moreover, as expected, both the background current and the time necessary to reach a stable background current, decreased as the Teflon content increased. This behaviour, together with the higher mechanical strength of the 60% Teflon electrodes, led us to choose this insulator percentage in the bioelectrode matrix for all subsequent work.
Choice of the enzymic system and effect of the applied potential
Obviously, the amperometric detection of hypoxanthine at +0.60 V, at which the uric acid formed in the enzyme reaction is oxidized, has very poor selectivity. At this potential, many other organic substances (e.g., ascorbic acid) that might be present in samples for hypoxanthine determination are electro- chemically oxidizable. Hydrogen peroxide is also formed as a product in the enzymatic reaction, and the reduction of H2O2, catalysed by peroxidase and using ferrocene as a mediator, can be carried out at a potential of 0.00 V at a graphite–Teflon–
Fig. 1 Cyclic voltammograms at a graphite–40% Teflon composite electrode from (a) 1.0 31023mol l21uric acid in 0.05 mol l21phosphate buffer solution (pH 7.4), (b) and (c) 1021and 1023mol l21H2O2in the same supporting electrolyte, respectively and (d) background voltammo- gram.
HRP–ferrocene composite electrode.20A bienzymic ampero- metric composite electrode, in which XOD and HRP, together with the mediator ferrocene, are incorporated in the graphite–
Teflon matrix, was therefore considered for the selective detection of hypoxanthine.
The immobilization of XOD, HRP and ferrocene into the bulk of the composite electrode can give rise to different electrode reactions produced by different electroactive sub- stances generated in both the enzyme reactions involved. Hence the influence of the applied potential on the steady-state current for a 1.0 31024mol l21hypoxanthine solution in 0.05 mol l21 phosphate buffer (pH 7.4) was investigated at various compos- ite electrodes constructed with different components: (I) graphite–Teflon–XOD; (II) graphite–Teflon–XOD–ferrocene;
(III) graphite–Teflon–XOD–HRP; and (IV) graphite–Teflon–
XOD–HRP–ferrocene. Schemes showing the enzyme and electrode reactions for the four types of composite electrodes are summarized in Fig. 2. A fixed amount of mediator and HRP (see Experimental), sufficient to ensure an excess with respect to the substrate,20was employed in the electrodes containing these compounds. As can be easily deduced, the amperometric signals obtained for each electrode corresponded (I) to the electrochemical oxidation of uric acid, (II) to the electro- chemical oxidation of ferrocene, (III) to the reduction of oxidized HRP to its reduced form23 and (IV) to the electro-
chemical reduction of ferricinium. It should be mentioned in the case of electrode (II) that, although the mediator for the enzymic reaction is ferricinium, the species immobilized into the composite electrode is ferrocene. This is possible because the amperometric signal appears at such a positive potential for all the ferrocene at the electrode surface to have been oxidized to ferricinium, which is again converted into ferrocene by the enzyme reaction. Consequently, the monitored current corre- sponded to the oxidation of ferrocene.7
Results obtained in the study of the influence of the applied potential on the hypoxanthine steady-state current at the four composite electrodes are displayed in Fig. 3. Regarding electrode I, the oxidation current does not seem to reach a plateau in the potential range tested. Moreover, a small steady- state current was obtained at this electrode. On the other hand, electrode II showed a considerably higher steady-state oxida- tion current, as a consequence of the presence of the mediator, but the current plateau was obtained at relatively high potentials. Concerning the electrode formed with XOD and HRP (III), where the electrode process is the reduction of oxidized HRP, the highest intensity at the steady state was reached at potentials lower than 0.15 V. However, when ferrocene was also present in the electrode matrix (electrode IV), the amperometric signal increased significantly at a potential of 0.00 V at which the reduction of ferricinium is
Fig. 2 Schematic diagram displaying the enzyme and electrode reactions involved in the hypoxanthine detection at (I) graphite–Teflon–XOD, (II) graphite–
Teflon–XOD–ferrocene, (III) graphite–Teflon–XOD–HRP and (IV) graphite–Teflon–XOD–HRP–ferrocene electrodes.
produced.20With this electrode, another signal at more positive potentials was also observed, corresponding to the same oxidation process that occurred at the XOD–ferrocene elec- trode. This ferrocene oxidation signal was higher than that of the ferricinium reduction, but the potential where the plateau was obtained was so high that substances capable of being oxidized at these potentials could interfere.
Taking into account the above results, the enzymic system selected for the development of the hypoxanthine amperometric biosensor consisted of both enzymes, XOD and HRP, and ferrocene as a mediator. A potential of 0.00 V was selected to be applied to the electrode. At this value, possible interferences can be minimized. Furthermore, as will be discussed below, if any substance was still oxidized at potentials near 0.00 V, it would be possible to apply more negative potentials to the electrode, as can be observed in Fig. 3.
The composite bienzymic electrode responded very quickly to the changes in the hypoxanthine concentration, the steady- state current being reached in 5 s. This fast response is due both to the absence of a membrane barrier on the electrode surface, which is usually needed for working in aqueous solutions to keep the enzyme adhering to the electrode, and to the absence of covalent linkages between the enzymes and graphite, which could decrease the rate of the enzymic reactions.
Effect of temperature and pH
The steady-state current measured at 0.00 V for a 4.0 31025 mol l21 hypoxanthine solution increased regularly with increasing temperature over the range 20–45 °C. However, it has been established that the XOD biosensors exhibit a shorter term operation as the temperature is raised.3As a compromise between sensitivity and stability, a working temperature of 30 ± 1 °C was chosen.
When the pH of the phosphate buffer solution was varied within the range 6.0–9.9, the electrode yielded the highest amperometric current between pH 7.0 and 8.0. A working pH of 7.4 was selected for subsequent work.
Stability of the bienzymic composite electrode
Enzyme composite electrodes constitute three-dimensional biocomponent reservoirs whose surface can be easily re- generated. Consequently, one of the most important features of these biosensors from a practical point of view is how long they can operate whilst giving reproducible responses. Furthermore, immobilization of the enzymes by simple physical entrapment in the electrode matrix could suggest possible enzyme sol- ubilization in the working medium used. Hence different aspects regarding the stability of the bienzymic XOD–HRP
composite electrode were considered: (a) repeatability of the amperometric signal without regeneration of the electrode surface; (b) repeatability of the amperometric signal when the electrode surface was regenerated by polishing; (c) reproduci- bility of different electrodes fabricated from the same pellet and from different pellets (see preparation of the electrode in Experimental); (d) stability of one composite bienzymic electrode with time and (e) effect of the storage time of the pellet from which electrodes are constructed.
First, a set of 10 successive amperometric measurements at 0.00 V for 4.0 31025mol l21hypoxanthine with no electrode surface regeneration yielded an RSD for the steady-state current of 4.3%, indicating good repeatability of the measurements. On the other hand, one of the most advantageous properties of the use of rigid composite enzyme electrodes is the possibility of obtaining a ‘new’ electrode surface and, consequently, a ‘new’
biosensor, by simple polishing of the surface. Therefore, the study of the reproducibility of the amperometric signal after regeneration of the electrode surface by polishing (for approx- imately 5 s on a 150 grit SiC paper) is one of the fundamental aspects of the evaluation of the performance of the bienzymic composite electrode developed. Ten sets of six successive measurements were carried out for a hypoxanthine concentra- tion of 4.0 31025mol l21. The electrode was polished after each set. Fig. 4 shows current–time recordings obtained in this way for three series of measurements. An RSD of 7.0% was obtained for the 10 steady-state current mean values of each series, indicating that the bienzymic composite electrode yielded reproducible amperometric signals after being sub- jected to the regeneration procedure. This suggested that both enzymes are uniformly distributed in the bulk of the electrode matrix. Moreover, the RSD values for the six measurements of each series ranged between 2.3 and 6.0%, thus confirming the good repeatability of successive measurements mentioned above.
Similarly to an HRP–GOD–ferrocene bienzymic composite electrode,20the same XOD–HRP–ferrocene electrode could be used for 4–5 days with no need to regenerate the electrode surface. When the amperometric response obtained for a fixed concentration of hypoxanthine decreased noticeably, the initial
Fig. 3 Influence of applied potential (E) on the steady-state current (i) for 1.0 31024mol l21hypoxanthine in 0.05 mol l21phosphate buffer solution (pH 7.4) at (A) electrode I (graphite–Teflon–XOD), (B) electrode II (graphite–Teflon–XOD–ferrocene), (C) electrode III (graphite–Teflon–
XOD–HRP) and (D) electrode IV (graphite–Teflon–XOD–HRP–ferro- cene).
Fig. 4 Current–time recordings at a graphite–60% Teflon–XOD–HRP–
ferrocene composite electrode. Six successive measurements followed by polishing (5 s) and further similar sets of six measurements. Ten sets in all, of which three are illustrated. 4.0 31025mol l21hypoxanthine in 0.05 mol l21phosphate buffer solution (pH 7.4); Eapp = 0.00 V.
signal could then be restored by polishing. Furthermore, electrode storage periods of 10 d between successive sets of measurements did not give rise to any apparent decrease in the electrode activity. As explained in a previous paper,21 the number of polishings that can be performed with the same electrode depended on the thickness of the pellet. If repetitive regeneration of the electrode surface by polishing results in too thin a composite pellet, then the initial amperometric response can no longer be restored. Consequently, the electrode becomes useless and the biocomposite pellet should be changed.
Therefore, the reproducibility of the analytical signals obtained from different electrodes fabricated from the same pellet and from different pellets is an essential aspect to be evaluated in order to assess the practical usefulness of the amperometric bienzymic electrode developed.
Table 1 summarizes the results obtained for five different electrodes, three of them fabricated from the same pellet and the other two from each of two different pellets. Amperometric measurements for five different 4.0 3 1025 mol l21 hypox- anthine solutions were carried out with each electrode. The RSD for all the measurements made with the five electrodes was 5.4%, whereas, for example, the RSD for the three electrodes constructed from the same pellet was 4.2%. These results indicated that the fabrication procedure of the composite bienzymic electrodes was reliable, whether they are constructed from the same pellet or from different pellets, thus allowing reproducible electroanalytical responses to be achieved with different electrodes constructed in the same manner.
Finally, the measurements made with electrodes 1.1, 1.2 and 1.3, fabricated from the same composite pellet, were carried out at intervals of 3 months after storage dry at 4° C. Therefore, the time elapsed between the measurements performed with the first and third electrode was 6 months. No significant decrease in the enzyme activity occurred for at least 6 months under the above-mentioned storage conditions, demonstrating the suit- ability of the method for enzyme immobilization in the electrode matrix used.
All these results illustrate the stability and robustness of this type of design of composite bienzymic electrodes.
Analytical characteristics of the calibration graph for hypoxanthine
Under the optimized working conditions described above, a calibration curve was constructed for hypoxanthine. A calibra- tion curve typical of enzymic systems was obtained for the steady-state current versus the hypoxanthine concentration with a range of linearity, by fitting data points by least-squares regression, of 5 31027–1 31025mol l21(r = 0.9989), and a loss of linearity for higher hypoxanthine concentrations. The slope and intercept of the linear range were (1.1 ± 0.1) 3104mA l mol21and 0.00 ± 0.01 mA, respectively. Fig. 5 shows this calibration graph (triangles), together with that obtained by
immobilizing in the electrode matrix half of the amount of XOD usually used (circles). Neither the slope nor intercept varied significantly with change of enzyme loading, and only a slightly wider range of linearity was obtained with the larger amount of XOD.
The limits of determination and detection were calculated according to the criteria of 10 times the standard deviation24and 3sb/m25 respectively, where m is the slope of the linear calibration graph and sbwas taken as the standard deviation (n
= 10) of the signals from 5.0 31027mol l21hypoxanthine.
The values obtained were 3.1 31027and 9.0 31028mol l21, respectively; they are similar to or even lower than those reported previously. The RSD calculated from the signals of 10 different solutions of 5.0 31027 mol l21 hypoxanthine was 6.2%.
Interferences
The effect of the presence of potential interferents on the hypoxanthine steady-state amperometric response was checked under the experimental conditions specified above for hypox- anthine. The substances tested, which may be present in samples for hypoxanthine determination,26were amino acids (such as alanine and methionine), carbohydrates (such as galactose, glucose and ribose), ascorbic acid and oxalic acid.
None of them, except ascorbic acid, affected the response of 4.0 31026mol l21hypoxanthine even for an analyte-to-interferent ratio of 1 : 100. However, ascorbic acid, whose electrochemical oxidation at the graphite–Teflon electrode takes place near 0.00 V, did affect the amperometric response of hypoxanthine, yielding relative errors in the measurement of the steady-state current of 270% and 2120% for hypoxanthine-to-ascorbic acid concentration ratios of 1 : 5 and 1 : 10, respectively. No appreciable relative error was observed when the hypoxanthine- to-ascorbic acid concentration ratio was 1 : 1 (Fig. 6). With the aim of minimizing the interference from ascorbic acid, and taking into account the dependence of the hypoxanthine biosensor response on the applied potential shown in Fig. 3, the effect of the presence of ascorbic acid on the amperometric signal of hypoxanthine was tested at different values of this applied potential (see Fig. 6). As expected, the interference from ascorbic acid decreased as the potential became more negative, and for a potential of 20.05 V, the relative error for an analyte- to-ascorbic acid concentration ratio of 1 : 5 was now 240%.
These results indicated that, depending on the interferent substances which may be present in a given sample, avoiding or minimization of their interfering effects can be achieved by a judicious choice of the potential to be applied to the bienzymic hypoxanthine electrode. Obviously, although it has not been studied in this work, if any interferent species was electro-
Table 1 Reproducibility of different graphite–60% Teflon–XOD–HRP–
ferrocene electrodes. Five different electrodes, three of them fabricated from the same pellet and the other two from each of two different pellets;
five 4.0 31025mol l21hypoxanthine steady-state current (i) measurements were carried out with each electrode. Confidence intervals for a significance level of 0.05; 0.05 mol l21phosphate buffer (pH 7.4); Eapp. = 0.00 V
Pellet Electrode i/mA RSD (%) ¯ı/mA RSD (%)
1 1.1 0.25 ± 0.01 4.8
1.2 0.23 ± 0.01 4.8
1.3 0.241 ± 0.009 3.8
0.24 ± 0.01 5.4
2 2.1 0.26 ± 0.01 4.5
3 3.1 0.23 ± 0.01 4.5
Fig. 5 Dependence of the steady-state current on the concentration of hypoxanthine at a graphite–Teflon–XOD–HRP–ferrocene composite elec- trode: (A) 0.1776 and (B) 0.0888 g of XOD per pellet. Other conditions as in Fig. 4.
chemically reduced at 0.00 V, it would be possible to apply more positive potentials to the bienzymic electrode (see Fig. 3) in order to minimize such interference. All these results demonstrated the good selectivity of the composite electrode developed for the determination of hypoxanthine.
Determination of hypoxanthine in sardine muscle tissue The performance of the bienzymic XOD–HRP composite electrode for the analysis of real fish samples was tested by determining the hypoxanthine content in sardine muscle tissue, following the procedure specified under Experimental. Further, the method was validated by comparing these results with those obtained by applying the spectrophotometric method recom- mended for the determination of hypoxanthine in fish by the Analytical Methods Committee.4
Significant differences were observed when comparing the slope of the calibration graph obtained by applying the standard additions method [(0.5 ± 0.1) 3 104 mA l mol21] with that previously mentioned obtained with standard hypoxanthine solutions [(1.1 ± 0.1) 3 104 mA l mol21], and therefore the standard additions method was used to minimize the matrix effect.
Sardine samples were stored at 4 °C for different periods, and then the content of hypoxanthine was determined. The results obtained from five replicates of each sample are summarized in Table 2, the confidence interval being calculated for a significance level of 0.05. Following the literature on this subject, the hypoxanthine concentration level in the samples is expressed as mmol g21 of sample. As can be observed, the amount of hypoxanthine increased with the time of sample storage, although from approximately the tenth day this rate was considerably smaller, which is consistent with the observations made by Karube et al.3
In order to evaluate the accuracy of the proposed method, these results were compared with those of the recommended reference method by applying Student’s t-test. Furthermore,
comparison was also made for another different sample of sardine (sample 2). The results are summarized in Table 3. The standard deviation for five replicates showed that the differ- ences were not significant at the 0.05 significance level. The experimental t values (texp) were always below those tabulated at the same significance level, which indicated the absence of determinate errors at a probability level of 95%.
Conclusions
The results demonstrate that bienzymic graphite–Teflon–XOD–
HRP–ferrocene composite electrodes constitute robust am- perometric biosensors for the determination of hypoxanthine in fish samples. These composite bioelectrodes are very easily and reproducibly fabricated by simple physical inclusion of both enzymes and the mediator in the bulk of the electrode with no need for covalent binding. They allow the regeneration of the electrode surface by polishing and exhibit long-term operation and stability. Furthermore, both the sensitivity and the selectiv- ity of the developed method for hypoxanthine are similar to or even better than those of other more complicated methodologies available, and its accuracy has been demonstrated and validated by comparing the results obtained for real fish samples with those obtained by a recommended reference method.
Financial support from the Subdirecci´on General de Formaci´on y Promoci´on del Conocimiento (Project 9B96-0640) and from the Complutense University of Madrid (Project PR181/
96-6804) is gratefully acknowledged.
References
1 Ehira, S., and Uchiyama, M., Bull. Jpn. Soc. Sci. Fish, 1969, 35, 1080.
2 Saito, T., Arai, A., and Matsuyoshi, M., Bull. Jpn. Soc. Sci. Fish, 1959, 24, 749.
3 Karube, I., Matsuoka, H., Suzuki, S., Watanabe, E., and Toyama, K., J. Agric. Food Chem., 1984, 32, 314.
4 Analytical Methods Committee, Analyst, 1979, 104, 434.
Fig. 6 Current–time plots at a graphite–60% Teflon–XOD–HRP–ferrocene composite electrode at different applied potentials for 4.0 31026mol l21 hypoxanthine, followed by sequential additions of a 1.0 31023mol l21ascorbic acid standard solution. Ratios correspond to the final hypoxanthine-to- ascorbic acid concentration ratio in each case.
Table 2 Determination of hypoxanthine in sardine muscle tissue after storage at 4 °C at a bienzymic graphite–60% Teflon–XOD–HRP–ferrocene electrode
Hypoxanthine/
Time of storage/d mmol g21
0 1.6 ± 0.3
6 2.8 ± 0.3
8 3.1 ± 0.2
13 4.1 ± 0.2
23 5.4 ± 0.5
Table 3 Comparison of the results obtained for the determination of hypoxanthine in sardine muscle tissue by using the bienzymic graphite–
Teflon–XOD–HRP–ferrocene electrode and a spectrophotometric reference method4
Hypoxanthine/mmol g21 Time of
Sample storage/d Reference method Proposed method texp
1 8 3.2 ± 0.2 3.1 ± 0.2 0.86
23 5.1 ± 0.4 5.4 ± 0.5 1.34
2 21 6.0 ± 0.3 6.3 ± 0.3 1.73
5 Tamiya, E., Seki, A., and Karube, I., Anal. Chim. Acta, 1988, 215, 301.
6 Sanghera, G. S., Anal. Proc., 1988, 25, 108.
7 Okuma, H., Takahashi, H., and Sekimukai, S., Anal. Chim. Acta, 1991, 244, 161.
8 Gonz´alez, E., Pariente, F., Lorenzo, E., and Hern´andez, L., Anal. Chim.
Acta, 1991, 242, 267.
9 Korell, U., and Spichiger, U. E., Electroanalysis, 1994, 6, 305.
10 Zhao, S., and Luong, J. H. T., Electroanalysis, 1994, 6, 830.
11 Albery, J. W., Driscoll, B. J., and Kalia, Y. N., J. Electroanal. Chem., 1995, 399, 13.
12 Hasebe, Y., Gokan, A., and Uchiyama, S., Anal. Chim. Acta, 1995, 302, 21.
13 Arai, G., Takahashi, S., and Yasumori, I., J. Electroanal. Chem., 1996, 410, 173.
14 Watanabe, E., Toyama, K., Karube, I., Matsuoka, H., and Suzuki, S., Appl. Microbiol. Biotechnol., 1984, 19, 18.
15 Moody, G. J., Sanghera, G. S., and Thomas, J. D. R., Analyst, 1987, 112, 65.
16 Luong, J. H. T., Male, K. B., and Nguyen, A. L., Ann. Biotech. Lab., 1988, 6, 38.
17 Malle, P., and Le Pezennec, I., Sci. Aliments, 1992, 12, 257.
18 Okuma, H., Takahashi, H., Yazama, S., and Sekimukai, S., Anal.
Chim. Acta, 1992, 260, 93.
19 Volpe, G., and Mascini, M., Talanta, 1996, 43, 283.
20 Del Cerro, M. A., Cayuela, G., Reviejo, A. J., Pingarr´on, J. M., and Wang, J., Electroanalysis, 1997, 9, 1113.
21 Ortiz, G., Gonz´alez, M. C., Reviejo, A. J., and Pingarr´on, J. M., Anal.
Chem., 1997, 69, 3521.
22 Ruzgas, T., Csoregi, E., Emneus, J., Gorton, L., and Marko-Varga, G., Anal. Chim. Acta, 1996, 330, 123.
23 J¨onsson, G., and Gorton, L., Electroanalysis, 1989, 1, 465.
24 ACS Committee on Environmental Improvement, Anal. Chem., 1983, 55, 2210.
25 Hasebe, K., and Osteryoung, J., Anal. Chem., 1975, 47, 2412.
26 Househam, B. C., Van der Berg, C. M. C., and Riley, J. P., Anal. Chim.
Acta, 1987, 200, 291.
Paper 7/07190F Received October 6, 1997 Accepted November 17, 1997
