Talanta 51 (2000) 547 557 www.elsevier.

com/locate/talanta

Determination of salicylate in blood serum using an amperometric biosensor based on salicylate hydroxylase immobilized in a polypyrroleglutaraldehyde matrix
Lae rcio Rover Ju nior a, Graciliano de Oliveira Neto b,*, Joa o Roberto Fernandes c, Lauro Tatsuo Kubota a
a b

Instituto de Qu mica -UNICAMP, P.O. Box 6154, 13083 -970, Campinas, SP, Brazil Faculdade de Cie ncias Farmace uticas -USF, 12900 -000, Braganc a Paulista, SP, Brazil c Departamento de Qu mica -FC /UNESP, 17033 -360, Bauru, SP, Brazil

Received 29 July 1999; received in revised form 14 October 1999; accepted 15 October 1999

Abstract The use of an amperometric biosensor for the salicylate determination in blood serum is described. The biosensor is based on salicylate hydroxylase (EC 1.14.13.1) electropolymerized onto a glassy carbon-working electrode with polypyrrole and glutaraldehyde, to improve the biosensor lifetime. The hexacyanoferrate (II) was also incorporated to work as a redox mediator to minimize possible interferences. The salicylate is enzymatically converted to catechol, which is monitored amperometrically by its electrooxidation at + 0.170 V versus SCE (saturated calomel electrode). Salicylate determination was carried out maintaining the ratio between b-NADH and salicylate at 4:1 (30C). The amperometric response of the biosensor was linearly proportional to the salicylate concentration between 2.3 10 6 and 1.4 10 5 mol l 1, in 0.1 mol l 1 phosphate buffer (pH 7.8), containing 0.1 mol l 1 KCl and 5.0 10 4 mol l 1 Na2H2EDTA, as supporting electrolyte. The recovery studies, in the presence of several interfering compounds, showed recoveries between 96.4 and 104.8%. The useful lifetime of the biosensor in the concentration range evaluated was at least 40 days, in continuous use. Blood serum samples analyzed by this biosensor showed a good correlation compared to the spectrophotometric method (Trinder) used as reference, presenting relative deviations lower than 7.0%. 2000 Elsevier Science B.V. All rights reserved.
Keywords: Salicylate determination; Serum samples; Polypyrrole; Salicylate hydroxylase

1. Introduction Salicylate, acetylsalicylic acid (aspirin) and their derivatives have been used as fungicidal and antimicrobial agents in pharmaceuticals preparations (external use) as well as in the treatment of

* Corresponding author. Fax: + 55-11-7844-8044. E -mail address: gon@usf.com.br (G. de Oliveira Neto)

0039-9140/00/$ - see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 9 9 ) 0 0 3 1 1 - 2

548

L. Ro6er Ju nior et al. / Talanta 51 (2000) 547557

inammatory processes as antipyretic and analgesic drugs (internal use). Salicylate has been used in beverages and foods for preservation, but it has been forbidden since the sixties in several countries due to its toxicity [1 3]. After the aspirin ingestion, this compound hydrolyses to salicylic acid and circulates in the blood in an ionized form as salicylate. When the salicylate concentration in the blood is higher than 2.2 10 3 mol l 1 (300 mg l 1), it becomes toxic, requiring control and monitoring of the salicylate level in the serum. The effective therapeutic range is between 1.1 and 2.2 10 3 mol l 1 (150300 mg l 1), which is very close to the toxicity stage. Salicylate concentration values higher than 4.3 10 3 mol l 1 (600 mg l 1) are regarded as lethal [2,4].

method for salicylate determination is necessary, as they often require extraction and pre-concentration steps. Voltammetric methods [2123] have been developed in the recent years due to the high sensitivity that these techniques offer, principally when various redox mediators [24] are employed, decreasing the interference levels in the determinations of compounds present in complex matrices such as blood serum. An alternative approach is the use of enzymatic methods based on salicylate hydroxylase enzyme [25] (salicylate-1-monooxygenase, EC 1.14.13.1) immobilized onto an electrode surface or incolumn reactors [2633]. This enzyme catalyses the irreversible hydroxylation of salicylate conversion to catechol in the presence of b-NADH and molecular oxygen:

Salicylate is the main aspirin metabolite in the body, reaching its maximal level in the blood serum 2 h after aspirin ingestion. Others aspirin metabolites analogous to salicylate, such as gentisic acid (2,5-dihydroxybenzoic acid) and salicyluric acid are present in the blood, but at minor levels. Paracetamol (acetaminophen) is another compound with analgesic effects similar to the aspirin very employed on the pain relief and can be present in plasma at relatively high concentrations [4,5]. Several methods have been described in the literature for the salicylate determination analysis. In clinical diagnosis, the most commonly used is the Trinder test [6,7], based on the formation of a purple complex between salicylate and Fe (III) ions that is monitored spectrophotometrically at 540 nm. Although this test is inexpensive and easy to use, it can undergo interference from aliphatic enolic (e.g. acetoacetate) and phenolic (e.g. L-tyrosine) substances [5]. Other methods include gas (GC) and liquid (HPLC) chromatographies [8 13], spectrouorimetry [14,15] and potentiometry [1620] (ion-selective electrodes). These are less susceptible to interference problems, but are not suitable in many cases where a fast and accurate

A recent paper [34] describes the evaluation of an amperometric biosensor for determination of salicylic acid in urine samples. This biosensor employs a second enzyme, tyrosinase, that further oxidises catechol giving o -quinone, which is electrochemically reduced yielding catechol, resulting in a signal amplication. Spectrophotometric assays monitor the consumption of b-NADH via decreasing in the absorbance at 340 nm [3537] or the formation of a catechol complex by reaction with 4-aminophenol [38] or 4-aminophenazone [39] to yield a blue color compound. Electrochemical detection of oxygen uptake [5] can also be used or even a potentiometric sensor to quantify carbon dioxide formation [40]. This paper describes the construction of an electrochemical biosensor for salicylate determination in blood serum samples, based on catechol detection, generated in the enzymatic reaction using hexacyanoferrate (II) as redox mediator to minimize interferences observed at high potentials. The electrochemical immobilization of the enzyme together the pyrrole and glutaraldehyde [4143] may increase the useful lifetime of the biosensor by formation of cross-linkages between the bifunctional reagent and amino groups from enzyme into the conducting polymer.

L. Ro6er Ju nior et al. / Talanta 51 (2000) 547557

549

2. Experimental

2.1. Apparatus
Voltammetric data were obtained with a potentiostat from AUTOLAB model PGSTAT20 (Eco Chemie), connected to a PC microcomputer for potential control and data acquisition, utilizing a three-electrode cell of 5.0 ml capacity. The components used in the voltammetric measurements were a glassy carbon (GC) electrode (Metrohm) as working electrode (A = 0.19 cm2), a platinum wire as auxiliary electrode and a saturated calomel electrode (SCE) as reference. Spectrophotometric measurements were performed with a Pharmacia Biotech model Ultrospec 2000 spectrophotometer using a cell with a 1.0 cm optical path. The solution pH values were measured by an OP-271 pH/ion analyzer ( 9 0.1 mV) using an OP-808P glass electrode and an OP-003P temperature sensor, all from Radelkis (Hungary). A Colora type KT10K ( 9 0.1C) thermostatic bath was also used.

of gentisic and o -acetylsalicylic acids, paracetamol, b-D( + )glucose and urea (Sigma), 3-hydroxy- and 4-hydroxy-benzoic acids and L-tyrosine (Aldrich), L-ascorbic, citric and o amino-benzoic acids (Merck), besides benzoate, tartrate and succinate (Fisher). All chemicals were of analytical grade reagents.

2.3. Treatment of the working electrode

The cleanness of the glassy carbon electrode surface was made with hand polishing on an appropriate support using a 0.5 and a 0.1 mm a-Al2O3 suspension. After washing ultrasonically in water for 15 s, the electrode was cycled repeatedly in a 1 mol l 1 KCl solution within the potential range 0.2 to 0.4 V versus SCE at a scan rate of 20 mV s 1, to verify the electrode condition. The cyclic voltammograms (under argon atmosphere) did not present redox peaks indicating good cleaning of the working electrode. This procedure avoids the adsorption of particles on the electrode surface.

2.2. Chemicals and materials

For the preparation of polymeric lms, 0.1 mol l 1 pyrrole, PY (Merck) freshly distilled at low pressure, 0.1 mol l 1 sodium hexacyanoferrate (II), Na4Fe(CN)6 (Fluka), 10 3 mol l 1 sodium perchlorate (Merck), 0.05 mg ml 1 sodium dodecyl sulfate, DS (Aldrich), all dissolved in 0.1 mol l 1 PIPES (piperazine-N-N-bis[2-ethanesulfonic acid], from Sigma) buffer (pH 6.8) were utilized. The argon was bubbled into the solutions for 10 min to eliminate oxygen from the cell. In the enzymatic incorporation, glutaraldehyde, GA (Aldrich) 0.5% (v/v) and 10 mg (1.2 U mg 1) of salicylate hydroxylase (Sigma) were employed. The measurements were obtained with sodium salicylate (Merck) solutions in a range of 2.0 to 20.0 mmol l 1 and b-NADH (nicotinamide adenine dinucleotide, reduced form from Sigma) at concentrations 4 times that of the substrate salicylate, both prepared in 0.1 mol l 1 phosphate buffer (pH 7.8), containing 0.1 mol l 1 KCl and 5.0 10 4 mol l 1 Na2H2EDTA (Aldrich). The interference studies were also made with solutions

2.4. Electropolymerization of pyrrole with mediator on an electrode surface

The polypyrrole (PPY) lm was prepared using the galvanostatic mode applying a 1.0 mA current for 50 s (Q = 0.255 C cm 2, with estimated thickness [44] of 0.41 mm), in 0.1 mol l 1 PIPES buffer, with fresh 0.1 mol l 1 PY, 0.1 mol l 1 3 Fe(CN)4 mol l 1 perchlorate and 0.05 mg 6 , 10 1 ml DS solutions in the cell, to improve the lm characteristics, such as homogeneity and adhesion. The thickness was estimated from the charge consumed, which includes the contribution from Fe(CN)4 oxidation during the electrochemical 6 process. After the electrochemical polymerization has nished, the sensor electrode was rinsed in distilled water. The mediator incorporation into the PPY was conrmed by means cyclic voltammetry in a 1 mol l 1 KCl solution in the potential range between 0.2 and 0.4 V versus SCE, at a scan rate of 20 mV s 1. A scan rate study for the sensor was achieved with the purpose to evaluate the kinetics of redox mediator into the PPY matrix, in the range of 0.05 to 0.1 V s 1.

550

L. Ro6er Ju nior et al. / Talanta 51 (2000) 547557

2.5. Enzyme immobilization

The biosensor was prepared by a new galvanostatic electropolymerization on the sensor containing mediator and PPY, using the same solution as before, also prepared by dissolving 10 mg of lyophilized enzyme and 0.5% (v/v) glutaraldehyde solution. A 0.5 mA current was applied for another 50 s (Q = 0.128 C cm 2, with nal estimated thickness [44] of 0.62 mm). After 1 h drying, the biosensor was conditioned in 0.1 mol l 1 phosphate buffer for 24 h at 4C. The inuence of glutaraldehyde on the biosensor response and its lifetime was evaluated in the concentration range between 0.1 and 2.0% (v/v), to optimize the amperometric response and lifetime. Analytical parameters for this biosensor such as linear response range, solution pH, temperature, amount of b-NADH coenzyme and stability were evaluated. These parameters were studied by amperometry (E = 0.17 V vs. SCE), in 0.1 mol l 1 phosphate buffer. The biosensor was applied in real sample analysis, after optimization of the best experimental conditions for salicylate determination.

Ions of the supporting electrolyte are incorporated into the PPY lm on the electrode simultaneously, since the electropolymerization reaction proceeds in the aqueous solution. Most of the anions, which are trapped into the PPY lm, can /4 present electroactivity, such as Fe(CN)3 re6 dox couple and have been used as redox mediators in many enzymatic processes [48,49]. A PPY lm may exist in one of three different states [50]: neutral, not conducting with highly dense packing, oxidized, conducting and more porous than the neutral state and overoxidized, which is not conducting but has a stable structure. The formation and oxidation of a PPY lm can be described by the following reactions [50]:

2.6. Salicylate determination in blood serum samples

Blood serum samples were supplied by the Clinical Hospital of State University of Campinas by a single centrifugation of the whole blood. All samples containing sodium salicylate were analyzed by the proposed (biosensor for salicylate) and reference (Trinder test) methods. The samples did not undergo any pre-treatment before the analysis, only dilution in 0.1 mol l 1 phosphate buffer (pH 7.8) containing 0.1 mol l 1 KCl and 5.0 10 4 mol l 1 Na2H2EDTA in the same studied range of salicylate concentrations.

3. Results and discussion

3.1. Characterization of the sensor

The electropolymerization of PY has been a subject of great attention in recent years [45 47].

The black color of the PPY lm prepared is also indicative of its oxidized state, useful in the electrochemical measurements due to conductive characteristics. Also, the electrolyte type and the concentration are fundamental for the nal properties of the conducting polymer lm as well as the potential or current applied and the electropolymerization time. It has been reported [51] that the zwitterionic buffers, such as derivative sulfonic acids like PIPES, are suitable for PPY lms preparation as conductive polymer matrices for biological sensors as these buffers display a convenient pKa. For the PIPES buffer (pKa = 6.8), the results described have shown PPY lms with optimum electrochemical behavior exhibiting good stability after several hundred of electrochemical cycles [51]. In this work, PPY lm was obtained by the galvanostatic technique, applying a charge density of 0.255 C cm 2, because good conductive and adherence properties were obtained for the lm. The use of dodecyl sulfate and perchlorate improved the electrochemical growth of the lm by

L. Ro6er Ju nior et al. / Talanta 51 (2000) 547557

551

the homogenization facility of the pyrrole in aqueous solution. The experimental conditions were optimized to get lms with appropriate thickness, minimizing the capacitive current of the system. The PPY lm doped with the redox mediator, exhibited one pair of peaks at 0.170 V (anodic) and at 0.135 V (cathodic) attributed to /4 the redox couple Fe(CN)3 (Fig. 1 (C)). 6 The dependence of the peak current, ip, on the potential scan rate for the sensor was also studied, showing a linear dependence between 0.05 and 0.1 V s 1, suggesting that the electroactive species are strongly bound in the PPY. This behavior is observed even at higher scan rates, indicating a good electron transfer between the mediator and the PPY lm, as described before, presenting a work potential of 0.17 V versus SCE. These re/4 sults obtained for the sensor GC/PPY/Fe(CN)3 6 are in agreement to the similar systems reported in the literature [52,53].

3.2. Electrochemical characteristics and performance of the biosensor

After redox mediator incorporation into the PPY matrix, conrmed by cyclic voltammetry, another PPY layer containing the enzyme was obtained on the previously prepared sensor. It was experimentally veried that applying a charge density of 0.128 C cm 2, one more layer of PPY lm with enzyme was obtained, keeping its catalytic activity. It is possible to observe (Fig. 1 (D)) an increase in the resistance of the system in different electropolymerization steps, showing the incorporation of mediator and enzyme into the PPY matrix. The enzymatic activity of the salicylate hydroxylase was determined in solution before and after the electropolymerization [25,54]. Then, the amount of enzyme immobilized in the polymeric lm was evaluated by the difference, obtaining a value of ca. 8.2 units, indicating that the most of enzyme was immobilized in the PPY lm. The enzymatic incorporation was also developed in glutaraldehyde media. This kind of immobilization results in a greater physical and chemical stability of the catalytic material due to the cross-linking formed with the glutaraldehyde and enzyme. In this case, the active sites of the enzyme could be more accessible for the enzymatic reaction. Besides, an increase in the useful lifetime of the biosensor was veried when it was kept under refrigeration (4C) in phosphate buffer, after it has been used in the salicylate determinations. The cyclic voltammograms (Fig. 2) obtained with the biosensor illustrate the amperometric response relative to salicylate, maintaining the coenzyme amount 4 times more than the salicylate concentration, where an increase in anodic peak current is veried. In this sense, it was conrmed that the b-NADH does not display any signicant response for the biosensor when added alone to the system. When salicylate was added to the medium an electrocatalytic current was observed as shown in Fig. 2. This current was assigned to the electrooxidation of catechol produced in the enzymatic reaction.

Fig. 1. Cyclic voltammograms obtained with the glassy carbon (GC) electrode at 20 mV s 1, in 0.1 mol l 1 KCl/0.1 mol l 1 phosphate buffer (pH 7.8): (A) bare GC electrode; (B) GC /4 with PPY lm;(C) GC containing PPY and Fe(CN)3 6 (sensor); (D) salicylate hydroxylase immobilized on the sensor (biosensor).

552

L. Ro6er Ju nior et al. / Talanta 51 (2000) 547557

Fig. 2. Cyclic voltammograms obtained with the biosensor at 20 mV s 1, in phosphate buffer at pH 7.8: (A) without b-NADH and salicylate; (B) with 0.60 mmol l 1 b-NADH and without salicylate; (C) 0.60 mmol l 1 b-NADH and 0.15 mmol l 1 salicylate solutions.

The potentials in which catechol electrooxidation occurs using phosphate buffer (pH 7.8), directly on bare GC and a PPY lm without redox mediator were of ca. 0.25 and 0.24 V versus SCE, respectively (Fig. 3). Thus, a potential of 0.17 V (PPY lm doped with mediator) nevertheless is lower than those observed to the other systems and can be chosen to minimize interferences from other redox species observed at higher potentials. In a biosensor for salicylate it has been shown that the b-NADH may begin to oxidize at a potential of ca. 0.50 V versus SCE (pH 7, 25 C) in phosphate buffer [55], salicylate at more positive potentials and catechol at potentials of about 0.25 V [28,56]. Even so, other compounds present in several complex samples can have redox behavior interfering with the measurements. In fact, the hexacyanoferrate (II) decreases the potential necessary to electrooxidize catechol by the better interaction between catechol and the doped lm, facilitating electron transfer to the electrode. Additionally, the other components involved in the electrochemical polymerization of the enzyme may act like a diffusion barrier excluding some interfering compounds that could affect the biosensor response. Fig. 4 shows the biosensor dependence with different applied potentials, suggesting that the potential affects the sensitivity of the biosensor response signicantly. It is possible to note a great difference in the applied working potential in some electrochemical biosensors employed over the last years in several complex matrices for salicylate determination [27,30,31], denoting the advantage in the use of redox mediators to reduce the interference level, such as urate and cystein.

3.3. Effects of the pH, temperature and i -NADH amount on the biosensor response
The optimum pH and temperature values for the salicylate hydroxylase have been reported to be 7.6 and 30C for the soluble enzyme [25,54]. However, with the enzyme immobilization an alteration of these parameters can occur, thus accurate evaluation must be carried out. Fig. 5 shows the optimal conditions obtained for the enzyme electropolymerized into a PPY matrix.

Fig. 3. Cyclic voltammograms obtained in the 0.1 mmol l 1 catechol electrooxidation employing 0.1 mol l 1 KCl/0.1 mol l 1 phosphate buffer (pH 7.8) and at 20 mV s 1: (A) bare GC electrode; (B) catechol over GC; (C) GC with PPY lm; (D) catechol over PPY.

L. Ro6er Ju nior et al. / Talanta 51 (2000) 547557

553

Therefore, a pH of 7.8 and temperature of 30C was selected for all assays. Finally, the coenzyme amount that presented the maximum amperometric response was found to be 4 times the salicylate concentration, in accordance with the other biosensors for salicylate cited in the literature [28,29,32,33]. The experimental data showed that for pH values lower than 7.6, the response decreases and this fact can be attributed to the lower activity of the enzyme. However, pH values higher than 8.0, also generated a decreasing in the signal, presumably due to the protein denaturation at alkaline media. Temperatures higher than 30C exhibited a decreasing in the biosensor response, suggesting a thermal denaturation of the enzyme.

3.4. Inuence of glutaraldehyde on the response and stability of the biosensor

The inuence of the glutaraldehyde (GA) quantity on the biosensor response was evaluated, showing a signicant increase in the signal when 0.5% (v/v) of glutaraldehyde was employed in the

Fig. 5. Performance of the biosensor for salicylate relative to: (A = ) solution pH prole (T = 25C); (B =
) b-NADH effect on the reaction; (C = ) inuence of temperature in the system (pH 7.8). General conditions: 0.1 mol l 1 KCl/0.1 mol l 1 phosphate buffer. [salicylate] = 4.0 mmol l 1. E = 0.17 V versus SCE.

Fig. 4. The dependence on the electrode potential for the response of the salicylate biosensor, obtained in 0.1 mol l 1 KCl/0.1 mol l 1 phosphate buffer (pH 7.8). [salicylate] = 4.0 mmol l 1. [b-NADH]/[salicylate] = 4; T = 30C.

biosensor preparation (Fig. 6). The use of higher amounts of this chemical did not improve the response (current density values) for the salicylate biosensor, which exhibited a decrease in the signal probably due to the increase in the PPY lm thickness, making the diffusion of the electroactive species trough the electrode more difcult. The biosensors useful lifetime was at least 40 days when glutaraldehyde concentrations from 0.5 to 2.0% (v/v) were employed (Fig. 6) without signicant alterations. For all studies a glutaraldehyde concentration of 0.5% (v/v) was chosen for enzyme immobilization, where the biosensor showed the highest response and a good lifetime in relation to the other salicylate biosensors which presented lifetimes between 10 and 20 days [5,3133,40]. The calibration curves for salicylate obtained periodically with the biosensor exhibited an in-

554

L. Ro6er Ju nior et al. / Talanta 51 (2000) 547557 Table 1 Recovery percentages obtained with the biosensor for salicylate solutions containing interfering compounds at 50%, in 0.1 mol l1 phosphate buffer (pH 7.8) Interfering compound [Salicylate], mmol l1 Recoveries (%)d

crease in the response after 5 days of use, and then stabilized its activity for more than 1 month in continuous use. The small increase in the response can be assigned to the catalytic material that remains weakly adsorbed in the PPY matrix, leaching out to the solution with time. Afterwards, there is stabilization in the system producing a stable signal proportional to salicylate concentration. After forty days of use, the relative response of the biosensor begins to decrease, indicating a loss of the catalytic activity.

Added (Salicylate) 2,5-Dihydroxy-benzoic acida 3-Hydroxybenzoic acid 4-Hydroxybenzoic acid Succinate Benzoate b-D(+) Glucose o -Acetyl-salicylic acidb L-tyrosine Urea N -acetyl-pamino-phenol c Citrate Tartrate o -Amino-benzoic acid L-ascorbic acid L-ascorbic acid*
a

Found 4.05 9 0.05 4.34 9 0.09 4.19 9 0.02 4.17 9 0.05 4.16 9 0.06 4.12 9 0.02 4.11 9 0.11 4.08 9 0.07 4.07 9 0.05 4.00 9 0.04 3.98 9 0.05 3.96 9 0.08 3.89 9 0.05 3.86 9 0.03 3.75 9 0.09 4.07 9 0.05 100.1 9 1.2 108.4 9 1.1 104.8 9 0.8 104.2 9 1.6 103.9 9 1.2 103.1 9 1.1 102.8 9 1.0 101.9 9 0.6 101.7 9 0.5 100.0 9 2.5 99.4 9 1.3 99.0 9 0.7 97.2 9 1.2 96.4 9 1.3 93.7 9 0.7 101.7 9 0.4

4.00 4.00

3.5. Interference study

The inuence on the assay of a series of salicylate metabolites and structurally related substances and a number of common blood components was examined. Solutions of these compounds were prepared in the same conditions of salicylate (phosphate buffer) at 50% of the salicylate concentration. The biosensor response was monitored and compared with the signal obtained with 4.0 mmol

4.00

4.00

4.00 4.00 4.00 4.00 4.00 4.00 4.00

4.00 4.00 4.00 4.00 4.00

Gentisic acid. ASA. c Paracetamol. d Average values for 3 determinations. * Recovery value obtained after the sample treatment with ascorbate oxidase enzyme.
b

Fig. 6. Graphic showing: (A) amperometric response and (B) stability for the salicylate biosensor as a function of glutaraldehyde concentration (a 4.0 mmol l 1 salicylate solution was used in the period). Conditions: 0.1 mol l 1 KCl/0.1 mol l 1 phosphate buffer (pH 7.8); E = 0.17 V versus SCE; [bNADH]/[salicylate] = 4; T = 30C.

l 1 salicylate without interference (Table 1). The results show signicant interference in the both cases of gentisic and L-ascorbic acid. It must be noted that gentisic acid concentration in real serum is much lower than that of main aspirin metabolites (salicylate) and is capacitylimited in blood [3,4]. Although, after aspirin ingestion, this analgesic is almost totally con-

L. Ro6er Ju nior et al. / Talanta 51 (2000) 547557

555

verted to salicylate in blood serum at a proportion of ca. 95% and only 5% is converted to gentisic acid [4]. In this manner, despite the higher interference due to gentisic acid on the biosensor response, the amount of this compound in clinical samples is low, thus not affecting the biosensor response in the evaluated salicylate concentration range. On the contrary, L-ascorbic acid exhibited a decrease in the biosensor response, working like an inhibitor in the system. This behavior can be assigned due to the high reactivity of this molecule. The enzymatic determination involving salicylate requires molecular oxygen to generate the reaction products, and the L-ascorbic acid present in the media can be oxidized by dissolved oxygen, decreasing the salicylate enzymatic kinetics. Nevertheless, the interference of the L-ascorbic acid was eliminated by adding 50 ml of a crude extract of cucumber (Cucumis sati6us L) containing about 100 U ml 1 of ascorbate oxidase [57].

3.6. Calibration cur6e and salicylate determination in serum samples

A typical calibration curve for salicylate with the developed biosensor was carried out by amperometry at + 0.17 V versus SCE, in 0.1 mol l 1 phosphate buffer (pH 7.8) containing 0.1 mol l 1 KCl and 5.0 10 4 mol l 1 Na2H2EDTA. A linear range between 2.3 and 14.1 mmol l 1 salicylate solutions was observed, with good correlation coefcient (r = 0.9988), t to the equation: DJ = 15.2( 9 1.5) + 6.6( 9 0.2) [Salicylate] where DJ is the current density in mA cm 2 and [salicylate] is the salicylate concentration in mmol l 1. As expected, amperometric-based enzyme electrodes, which consume the reaction product, display a more expanded linear response range than their electrochemical counterparts [58,59], presenting a dynamic range for salicylate determination of the same order in comparison with the others cited [26,27,30,33,60]. Fig. 7 exhibits the proportional change of the signal of the biosensor with the addition of salicylate solutions in different concentrations. The biosensor response time was estimated in 16 s, time necessary to reach 95% of the maximum signal. The analyses of ten samples of blood serum was carried out by the proposed method (electrochemical biosensor) and by a reference method (Trinder test). The results obtained (Table 2) illustrate the accuracy of the biosensor with relative deviations lower than 7.0%, an acceptable value in the case of biological interest samples. Some of the samples show a small difference between Trinder and biosensor method, although lower than 7.0%. These differences can be assigned to the interference in the Trinder method caused by the sample coloration and/or opacity due to the high protein content in the blood serum samples. In conclusion, the great advantage of this salicylate biosensor compared to others is the kind of electrochemical immobilization used, using glutaraldehyde additionally to extend its stability. The developed biosensor durability was about 40 days, which is greater than the useful lifetime of

Fig. 7. Analytical curve and amperometric responses for the salicylate biosensor after successive injections of a 4.0 10 4 mol l 1 salicylate solution in the cell: (A) 30 ml; (B) 60 ml; (C) 90 ml; (D) 120 ml; (E) 150 ml; (F) 180 ml. Conditions: 0.1 mol l 1 KCl/0.1 mol l 1 phosphate buffer (pH 7.8); E = 0.17 V versus SCE; [b-NADH]/[salicylate] = 4; T = 30C.

556

L. Ro6er Ju nior et al. / Talanta 51 (2000) 547557

Table 2 Concentration values obtained from the proposed and standard methods for salicylate analyses in serum samples, with estimated standard deviation for 3 determinations Serum samplesa [Salicylate] (mmol l1) Amperometric methodb 01 02 03 04 05 06 07 08 09 10
a

Relative deviationd (%) Spectrophotometric methodc 4.5 9 0.6 5.5 9 0.9 3.6 9 0.9 4.5 9 0.5 2.8 9 0.4 5.5 9 0.8 5.8 9 0.7 5.8 9 0.8 5.1 9 0.4 5.9 9 1.0 2.2 3.6 2.7 6.4 3.5 3.6 6.9 5.2 5.9 1.7

4.4 9 0.3 5.3 9 0.5 3.5 9 0.4 4.2 9 0.3 2.9 9 0.5 5.7 9 0.6 6.2 9 0.4 6.1 9 0.2 4.8 9 0.3 6.0 9 0.4

Blood serum samples containing salicylate for analysis supplied by Hospital das Cl nicas of UNICAMP. Developed biosensor with salicylate hydroxylase immobilized in PPY-GA matrix on GC electrode. c Reference method using Trinder test (u = 540 nm). d Relative deviations between amperometric and spectrophotometric methods.
b

several salicylate biosensors. This shows that it is possible improve the performance of these probes as the use of enzymes as analytical reagents continues to be explored. This feature permits a more frequent use of the biosensor without loss of the enzymatic activity. The simplicity of the electrochemical immobilization of salicylate hydroxylase involving one single step demonstrates a considerable economic advantage. The biosensor sensitivity allows salicylate determinations in which the serum sample did not suffer any treatment, only simple dilution. The blood serum samples analyzed by the biosensor for salicylate presented good agreement with the reference method (Trinder), showing the potential of this electrochemical biosensor.

References
[1] K.D. Rainsford, Aspirin and the Salicylates, Butterworths, London, 1984, p. 245. [2] J.J. Thiessen, Acetylsalicylic Acid: New Uses for an Old Drug, In: H.J.M. Barnett, J. Hirsh, J.F. Mustard, Raven Press, New York, 1992, p. 49. [3] T.M. Brown, A.T. Dronseld, P.M. Ellis, J.S. Parker, Educ. Chem. 35 (1998) 47. [4] P. Nietsch, Therapeutic Applications of ASPIRIN, wbnVerlag, in: Bingen/Rhein, Bayer AG (Pub.), Germany, 1989, p. 7. [5] M.A.N. Rahni, G.G. Guilbault, G.O. Neto, Anal. Chim. Acta 181 (1986) 219. [6] P. Trinder, Biochem. J. 57 (1954) 301. [7] E.S. Kang, T.A. Todd, M.T. Capaci, K. Schwenzer, J.T. Jabbour, Clin. Chem. 29 (1983) 1012. [8] S.L. Ali, J. Chromatrogr. 126 (1976) 651. [9] J.N. Buskin, R.A. Upton, R.L. Williams, Clin. Chem. 28 (1982) 1200. [10] I.M. Jalal, S.I. Sasa, Talanta 31 (1984) 1015. [11] B.S. Yu, L.H. Nie, S.Z. Yao, HRC J. High Resol. Chromatogr. 20 (1997) 227. [12] R. Pirola, S.R. Bareggi, G. DeBenedittis, J. Chromatogr. B 705 (1998) 309. [13] G.P. McMahon, M.T. Kelly, Anal. Chem. 70 (1998) 409. [14] A.M. Pen a, F. Salinas, I.D. Me ` ras, Anal. Chem. 60 (1988) 2493. [15] R.N. Gupta, M. Zamkanei, Clin. Chem. 36 (1990) 1690.

Acknowledgements The authors thank FAPESP, CNPq and The European Economic Community for nancial support received and Professor Carol H. Collins for English revision of the manuscript. LRJ is indebted to CAPES for a fellowship.

L. Ro6er Ju nior et al. / Talanta 51 (2000) 547557 [16] N.A. Chaniotakis, S.B. Park, M.E. Meyerhoff, Anal. Chem. 61 (1989) 566. [17] T. Katsu, Y. Mori, Talanta 43 (1996) 755. [18] D. Liu, W.C. Chen, G.L. Shen, R.Q. Yu, Analyst 121 (1996) 1495. [19] R.S. Hutchins, P. Bansal, P. Molina, M. Alajar n, A. Vidal, L.G. Bachas, Anal. Chem. 69 (1997) 1273. [20] L. Rover, C.A.B. Garcia, G.O. Neto, L.T. Kubota, F. Galembeck, Anal. Chim. Acta 366 (1998) 103. [21] A.G. Fogg, M.A. Ali, M.A. Abdalla, Analyst 108 (1983) 840. [22] K. You, Clin. Chem. 149 (1985) 281. [23] Y.S. Fung, S.F. Luk, Analyst 114 (1989) 943. [24] M.J. Lobo, A.J. Miranda, P. Tun o n, Electroanalysis 9 (1997) 191. [25] H. Kamin, R.H. White-Stevens, R.P. Presswood, Biomembranes, in: S. Fleischer, H. Packer (Eds.), Methods in Enzymology Vol. LIII, 52, Academic Press, New York, 1978, p. 527. [26] J.E. Frew, S.W. Bayliff, P.N.B. Gibbs, M.J. Green, Anal. Chim. Acta 224 (1989) 39. [27] P. Bertocchi, D. Dottavio, M.E. Evangelisti, M. Mascini, G. Palleschi, Clin. Chim. Acta 207 (1992) 205. [28] M. Neumayr, O. Friedrich, G. Sontag, F. Pittner, Anal. Chim. Acta 273 (1993) 469. [29] M. Neumayr, G. Sontag, F. Pittner, Anal. Chim. Acta 305 (1995) 26. [30] T.J. Moore, M.J. Joseph, B.W. Allen, L.A. Coury Jr, Anal. Chem. 67 (1995) 1896. [31] M. Ehrendorfer, G. Sontag, F. Pittner, J. Fresenius, Anal. Chem. 356 (1996) 75. [32] L.T. Kubota, B.G. Milagres, F. Gouvea, G.O. Neto, Anal. Lett. 29 (1996) 893. [33] B.G. Milagres, G.O. Neto, L.T. Kubota, H. Yamanaka, Anal. Chim. Acta 347 (1997) 35. [34] C. Martin, E. Dominguez, J. Pharmac. Biom. Anal. 19 (1999) 107. [35] R.W. Longenecker, J.E. Trafton, R.B. Edwards, Clin. Chem. 30 (1984) 1369. [36] K. You, J.A. Bittikofer, Clin. Chem. 30 (1984) 1549. [37] P. Bouvrette, J.H.T. Luong, Anal. Chim. Acta 335 (1996) 169. [38] S.A.P. Chubb, R.S. Campbell, J.R. Ramsay, P.M. Ham-

557

[39]

[40] [41]

[42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55]

[56] [57] [58] [59] [60]

mond, T. Atkinson, C.P. Price, Clin. Chim. Acta 155 (1986) 209. H.C. Morris, P.D. Overton, J.R. Ramsay, R.S. Campbell, P.M. Hammond, T. Atkinson, C.P. Price, Clin. Chem. 36 (1990) 131. T. Fonong, G.A. Rechnitz, Anal. Chim. Acta 158 (1984) 357. R.W. Murray, Techniques of Chemistry-Molecular Design of Electrodes Surfaces, In: Saunders W.H. (Ed.), Vol. 22, Wiley, New York, 1992, p. 403 J.L. Anderson, E.F. Bowden, P.G. Pickup, Anal. Chem. 68 (1996) 379R. W.H. Scouten, Solid Phase Biochemistry, vol. 66, Wiley, New York, 1983, p. 260. M. Zagorska, H. Wycistik, J. Przyluski, Synth. Met. 20 (1987) 259. E. Yurtsever, B. Erman, Polymer 34 (1993) 3887. D.J. Ferm n, H. Teruel, B.R. Scharifker, J. Electroanal. Chem. 401 (1996) 207. W. Schuhmann, Mikrochim. Acta 121 (1995) 1. G. Lian, S. Dong, J. Electroanal. Chem. 260 (1989) 127. I. Katakis, E. Dom nguez, Mikrochim. Acta 126 (1997) 11. S. Zhao, J.H.T. Luong, Electroanalysis 7 (1995) 633. E.M. Genie ` s, M. Marchesiello, G. Bidan, Electrochim. Acta 37 (1992) 1015. A. Amine, J.M. Kauffmann, G.G. Guilbault, S. Bacha, Anal. Lett. 26 (1993) 1281. M. Kawakami, N. Uriuda, H. Koya, S. Gondo, Anal. Lett. 28 (1995) 1555. L.H. Wang, S.C. Tu, J. Biol. Chem. 259 (1984) 10682. L. Gorton, B. Persson, P.D. Hale, L.I. Boguslavsky, H.I. Karan, H.S. Lee, T.A. Skotheim, H.L. Lan, Y. Okamoto, Biosensors and Chemical Sensors, ACS Symposium Series 487, Washington, 1992, p. 56. J. Kulys, G. Gleixner, W. Schumann, H.L. Schmidt, Electroanalysis 5 (1993) 201. S. Maritano, T. Kohzuma, S. Suzuki, A. Marchesini, Phytochemistry 41 (1996) 349. L.D. Mell, J.T. Maloy, Anal. Chem. 47 (1975) 299. C.T. Minh, G. Broun, Anal. Chem. 47 (1975) 1359. D.M. Zhou, P. Nigam, J. Jones, R. Marchant, J. Chem. Tech. Biotechnol. 64 (1995) 331.