Characterization of the composition of bovine urine and its effect on the electrochemical analysis of the model mediator, p-aminophenol

Transcription

1 Analytica Chimica Acta 554 (2005) Characterization of the composition of bovine urine and its effect on the electrochemical analysis of the model mediator, p-aminophenol Mamun Jamal a, Mark A. Crowe b, Edmond Magner a, a Materials and Surface Science Institute, Chemical and Environmental Science Department, University of Limerick, Ireland b Faculty of Veterinary Medicine and Conway Institute, University College Dublin, Ireland Received 7 June 2005; received in revised form 27 July 2005; accepted 18 August 2005 Available online 5 October 2005 Abstract The reliable identification of compounds such as illegal growth promoters in cattle is generally based on expensive gas chromatography mass spectrophotometric analysis in urine, a method that does not allow on a large-scale screening. The use of simple, semi-quantitative electrochemical biosensors may provide a means of screening for the presence of compounds such as illegal growth promoters. Before such sensors can be utilised, it is necessary to understand which factors influence the response of an electrochemical sensor in bovine urine. The concentration range of protein ( %), uric acid ( mm), xanthine ( mm) and ascorbic acid ( mm) in 26 individual urine samples were determined. Using p-aminophenol (p-ap) as a model system, the electrochemical response increased by 5% in the presence of 6.0 mm uric acid, by 10% on the addition of 0.2 mm xanthine and by 22% in the presence of 1.0 mm ascorbic acid. Exposing urine to air and light for 75 min eliminated interference from ascorbic acid. Addition of Cu 2+ (10 M) reduced the time required to 34 min. Binding of species such as growth promoters to proteins may be disrupted by the addition of 8-anilino-1-naphthalene sulphonic acid (ANS) to the urine samples. Addition of 10 M ANS did not affect the limit of detection of p-ap. The ph of fresh bovine urine samples was monitored over the period 7 to 192 h after collection and ranged from 8.00 to The ph of lyophilised urine samples ranged from 8.24 to Amperometry was the most sensitive method among a range of electrochemical techniques in the detection of p-ap with a limit of detection (LOD) in urine of 1.0 gml 1 (10 M) on a glassy carbon electrode Elsevier B.V. All rights reserved. Keywords: Electrochemical biosensors; Urine matrix effects; p-aminophenol; Interferents 1. Introduction The use of hormonal substances as animal growth promoters is prohibited for food safety reasons (EU Directive 88/146/EEC). Growth promoters are screened in a number of matrices such as kidney, fat, urine, and meats using different chemical and immunochemical [1] screening techniques followed by confirmatory techniques such as LC [2], GC MS [3], LC MS MS [4], LC ESI MS [5]. All these techniques rely on random sampling, primarily at the point of slaughter, with specific tissue (muscle, liver, fat) or fluid (bile, urine) samples being taken to a centralised laboratory for analysis. The procedures involved in analysis can be relatively complex, and do not lend themselves to rapid measurement. These testing methods are not in widespread use due to the inherent cost involved, the ex situ Corresponding author. address: edmond.magner@ul.ie (E. Magner). nature of the analysis technique employed, the time lag involved (typically h) and the relatively complex techniques used which require skilled laboratory personnel. Due to increasing enforcement of regulations governing the use of growth promoting agents, producers are rapidly changing the compounds used and the amounts administered, making detection much more difficult. Control of the use of growth promoters requires the development of a thorough screening programme [6]. The techniques used should be rapid, continuous and allow on the spot decisions to be made regarding the androgen residue status of an animal to assure the safety and integrity of the food supply. The development of techniques which can reliably provide such assurances is difficult from a technical point of view, and probably, more importantly, in view of cost. A test kit which could provide semi-quantitative information, even if only on a restricted number of analytes, would be beneficial. Biosensors with an electrochemical transducer [7] may enable the implementation of such a screening program. As a sample matrix, blood is difficult to work with due to the large variable amount /$ see front matter 2005 Elsevier B.V. All rights reserved. doi: /j.aca

2 80 M. Jamal et al. / Analytica Chimica Acta 554 (2005) of red blood cells [8], and its high protein content. Sample processing of meat, faeces, and fats is also complicated. Urine contains less interference and is readily available at abattoirs and farms. As growth promoters are not generally electrochemically active, a recognition element such as an antibody coupled to an electrochemical transducer is required. The sensitivity of such an electrochemical ELISA technique depends on the ability of removing the effect of any species which interfere with electrochemical detection. It is necessary to determine the effect of each interfering species on the electrochemical response. In biological fluids, ascorbic acid, uric acid and xanthine are common electrochemical interferences. The composition of human urine [9] is well established. To our knowledge, the corresponding information on the composition of bovine urine is unknown. In this work, the concentrations of electrochemical interferences were determined and their effect on the electrochemical response of a model mediator, p-ap, was examined. We also describe methods of eliminating these interferences from biological samples and evaluated the best electrochemical technique for the detection of p-ap in bovine urine. The results of this work will then enable the development of an immunobiosensor for the detection of growth promoters in bovine urine. 2. Experimental 2.1. Reagents and chemicals Lyophilised urine from 20 individual animals was obtained from the National Institute for Public Health and the Environment (RIVM) (Zeist, The Netherlands). 4-Aminophenol, urea, uric acid, 8-anilino-1-naphthalene sulphonic acid (ANS), bovine serum albumin (BSA), EDTA, xanthine, urea, uric acid and l-ascorbic acid were purchased from Sigma. Cu(NO 3 ) 2, KH 2 PO 4, K 2 HPO 4, uricase and Nafion perfluorinated ion exchange resin were purchased from Aldrich. HCl and H 2 SO 4 were obtained from Merck, uric acid test kits (catalogue no. 292) from Sigma. All aqueous solutions were prepared using purified water (18.2 M ) from an ElgaStat SPECTRUM system Apparatus Electrochemical experiments were conducted using a threeelectrode cell in which glassy carbon (GC, surface area 0.69 cm 2 ), platinum wire and Ag AgCl were used as the working, counter and reference electrodes (CH Instruments), respectively. Electrochemical experiments were conducted using CHI 600 (CH Instruments) and PGSTAT 10 (Ecochemie) potentiostats. UV vis spectra were obtained using a Shimadzu 1601 spectrophotometer. An Orion 420A ph meter was used to monitor the ph. Urine samples were lyophilised on an FTS Systems freeze-dryer Urine preparation Urine specimens were collected from nine animals at the Lyons Research Farm (University College, Dublin) between 4 April 2003 and 7 May The study was performed in compliance with protocols approved by the Ethics Committee, University College Dublin, the Cruelty to Animals Act (Ireland, 1876), and the European Union Directive, 86/609/EC. All specimens were collected in 0.5 L containers (brown translucent material, light protected, with screw cap) on the farm and stored immediately at 20 Cin70mL containers. In addition, aliquouts (5 ml) of three fresh samples were lyophilised. The lyophilised urine samples obtained from RIVM were reconstituted by the addition of 5 ml of water Characterization of urine samples Electrodes were polished successively with 1.0, 0.3 and 0.05 m Al 2 O 3 slurry on micro-cloth pads (Buehler), rinsed with distilled water and briefly sonicated. No additional electrolyte was added to the samples. Potentials are reported with respect to Ag/AgCl. Electrodes were modified by placing 10 L of Nafion solution on the electrode surface, followed by drying in air for 30 min. The concentrations of ascorbic acid and xanthine were determined using HPLC. The chromatographic system was comprised of a Waters 484LC system utilising a reversed phase column (C18, 5 m, 4.6 mm 150 mm) (WAT ) and a UV detector. All experiments were performed at room temperature. In order to obtain good separation and reproducible results, the column was equilibrated for 4 days prior to use [10]. Levels of ascorbic acid were determined using a modification of the method described by Ross [11], with detection at 261 nm. The mobile phase was a mixture of K 2 HPO 4 (5 mm) and H 2 SO 4 (0.5 mm) ph 3 containing EDTA (200 mg/l), delivered at 0.35 ml/min. Urine samples were diluted 10 fold using 1% trichloroacetic acid to prevent sample decomposition and to adjust the ph to ca Stock solutions of ascorbic acid (1 mm) were prepared in 1% trichloroacetic acid. All solutions used were freshly prepared before each set of experiments. The same column was used to determine the concentration of xanthine. Urine was diluted by a factor of 10 using 0.01 M phosphate buffer to ph 5.6. A mixture of K 2 HPO 4 (0.01 M) and H 2 SO 4 (0.5 mm) at ph 5.6 was used as the mobile phase and xanthine was detected at a wavelength of 267 nm. Standard xanthine solutions were prepared using the same procedure as with urine. Peaks due to ascorbic acid and xanthine were identified on the basis of the retention times of standards injected separately and by the addition of standards to urine. Protein concentrations in urine were determined using the Bradford assay [12]. Solutions containing % (g/g) of BSA were used as standards. Uric acid test kits were used to quantify the concentration of uric acid in urine. Aliquots of urine (0.2 ml), glycine buffer (1.0 ml) and water (6.0 ml) were mixed and 3 ml of this mixture was placed in two different test tubes ml of uricase enzyme ( U ml 1 ) was added to one test tube and 0.05 ml water to the control solution. After 15 min, the absorbance was measured at 292 nm. The ph was measured in urine collected from the same animal at different time intervals from 7 to 192 h. During this time period, the animal had access to the same food area.

3 M. Jamal et al. / Analytica Chimica Acta 554 (2005) To measure the effect of interferences, ascorbic acid, uric acid and xanthine were added to the urine samples in the presence of 1 mm p-ap. All experiments were conducted within 40 min of addition of p-ap. The electrochemical signal was recorded using chronoamperometry in the presence of increasing amounts of uric acid (1.0 mm), ascorbic acid (0.2 mm), xanthine (0.04 mm). In order to select the best electrochemical technique to detect p-ap, four techniques were examined differential pulse voltammetry (DPV), square wave voltammetry (SWV), linear sweep voltammetry (LSV) and chronoamperometry (CA). A range of parameters for each of these techniques was optimised using a GC electrode in 1 mm p-ap in 0.1 M phosphate buffer at ph 8. For CA, these parameters were: conditioning potential at open circuit potential (OCP) for 20 s, equilibration time of 5 s, applied potential of 0.08 V for 10 s, current sampled at 10 Hz. For LSV, the initial and final potentials were 0.2 and 0.3 V, respectively at a scan rate of 20 mv s 1. For DPV, the initial and final potentials were 0.2 and 0.3 V, respectively, increment potential V, pulse period 0.2 s, pulse amplitude 0.1 V and pulse width 0.05 s. For SWV, the initial and final potential were 0.2 and 0.3 V, respectively, increment potential V, frequency 15 Hz and pulse amplitude V. 3. Results and discussion Voltammograms of three representative bovine urine samples are shown in Fig. 1. The large currents observed at potentials above 0.1 V can be ascribed to ascorbic acid (0.15 V versus Fig. 1. Cyclic voltammograms of three different blank urine samples on a glassy carbon electrode; scan rate of 100 mv s 1. Ag/AgCl) [13], uric acid (0.4 V) [14] and xanthine (0.67 V) [15]. The substantial variations in the responses of individual samples indicate that the concentrations of each electroactive species varied significantly. The concentration range of ascorbic acid in 26 different urine samples was determined using HPLC and found to vary between 0.1 and 0.95 mm with an average of 0.5 mm (Table 1). For xanthine and uric acid, concentrations ranged from 0.04 to 0.13 mm (mean 0.05 mm) and from 0.24 to 0.75 mm (mean 0.4 mm), respectively. Table 1 ph and amounts of ascorbic acid, uric acid and protein in bovine urine samples Urine sample Ascorbic acid (mm) Xanthine (mm) Uric acid (mm) Protein (%) ph 1 n n n a a a a a a n n: no response. a Fresh urine.

4 82 M. Jamal et al. / Analytica Chimica Acta 554 (2005) Fig. 2. Plot of % change in electrochemical response of 1 mm p-ap in three urine samples on addition of ascorbic acid (A), uric acid (B) and xanthine (C). Conditions: integration time of 0 10 s, applied potential of 0.08 V. p-ap was chosen as a model mediator. The effects of ascorbic acid, uric acid and xanthine on the electrochemical response of 1 mm p-ap were examined (Fig. 2). As shown in Table 1, bovine urine can contain up to 1 mm ascorbic acid. Sequential addition of a constant amount of ascorbic acid was used to assess the effect of ascorbate on the electrochemical response of p-ap (Fig. 2A). After the addition of 0.2 mm ascorbic acid, when the initial concentration of ascorbate was mm, there was a large increase in the current. On increasing the concentration further, the response became more complex and varied from sample to sample. Such a variation would introduce unacceptable changes in the response of a sensor. Since Nafion repels negatively charged species such as ascorbate [16], Nafion modified electrodes were prepared. An electrode coated with a solution containing 0.05% Nafion blocked the response from 1 mm ascorbate (Fig. 3A). However, the response to p-ap decreased by 40% (Fig. 3B), excluding the possibility of using Nafion to repel ascorbate. An alternative Fig. 3. Cyclic voltammograms of (A) 1 mm ascorbic acid in buffer on a bare (a) and Nafion (0.05%) coated (b) glassy carbon electrode; (B) of 1 mm p-ap on a bare (a) and Nafion (0.05%) coated (b) glassy carbon electrode; (C) 1 mm ascorbic acid in urine on glassy carbon electrode (a) on exposure to air and light (b) and on addition of 10 M Cu 2+ (c). Scan rate of 100 mv s 1.

5 M. Jamal et al. / Analytica Chimica Acta 554 (2005) Fig. 4. Charge passed on oxidation of p-ap after exposing urine to air and light ( ) and on addition of 10 M Cu 2+ ( ). Integration time 0 10 s. procedure, whereby ascorbic acid is oxidized to the electrochemically inactive dehydroascorbic acid on exposure of the samples to air and light was examined [17]. The presence of Cu 2+ enhances the rate of oxidation of ascorbic acid [18,19] to dehydroascorbic acid, reducing the time required to remove the ascorbate (Fig. 3C). After exposure of urine for 75 min to air and light, the LOD of p-ap was 10 M. On addition of 10 MCu 2+ to the urine sample, the time required to remove 1 mm ascorbic acid (added externally) was 34 min. In the presence of this concentration of Cu 2+, the LOD increased to 30 M(Fig. 4). Unlike in human urine, due to the presence of uricase in bovines, most of the uric acid is oxidized to allantoin [20], which is electrochemically inactive. Addition of allantoin did not affect the LOD of p-ap. However, urine can contain some residual levels of uric acid which could interfere with the electrochemical detection of the analyte. The addition of up to 6 mm (in 1 mm increments) uric acid to three representative samples altered the current response to p-ap by an average of less than 5% (Fig. 2B) indicating that uric acid does not interfere in the detection of p- AP. A slightly higher effect (10%) was obtained with xanthine (Fig. 2C). Ascorbate is then the major electrochemical interference in the detection of p-ap. While its effect can be eliminated as outlined above, the time required is long and may not be acceptable, particularly outside of laboratory conditions. To our knowledge, there are no published data available detailing the concentration of proteins in bovine urine. The amount of protein in a series of urine samples from different animals was determined using the Bradford assay. The amount of protein ranged from 0.01 to 0.04% (by weight) with an average of 0.015% (Table 1). Usually species such as steroid growth promoters are found in the free form and bound to protein. Detection of the latter form generally requires disruption of the mode of binding between the hormone and the protein. The binding constants for a hormone such as testosterone range from M 1 (BSA) to M 1 (sex hormone binding globulin (SHBG)) [21]. The addition of ANS to clinical samples is used to disrupt the binding of steroids to proteins [21 22]. ANS binds to BSA with a binding constant of M 1. The binding constant of testosterone to SHBG is unknown. If it is considered similar to that of testosterone (10 9 M 1 ), addition of ANS to a concentration of 0.1 M would be required to liberate 1 M (100 ng ml 1 )of testosterone. If the binding constant were higher, higher concentrations of ANS would be required. Concentrations up to 10 M ANS had no significant effect on the LOD of p-ap in urine. 4. Screening electrochemical techniques The limits of detection for p-ap in urine were ascertained by DPV, SWV, LSV and CA (Fig. 5). For the 26 different urine Fig. 5. Plot of response obtained by (A) chronoamperometry, (B) DPV, and (C) LSV in the detection of p-ap in bovine urine. Conditions are as described in experimental section. All the results are background subtracted.

6 84 M. Jamal et al. / Analytica Chimica Acta 554 (2005) Table 2 ph ranges and amounts of acid required to reduce ph of bovine urine to 7.0 Urine type Number of samples ph range Titre (0.01 M HCl) (ml) Lyophilised Fresh Freeze-dried (fresh) samples, a LOD of 10 M was obtained using CA on GC electrode with a linear range of M. With DPV, the LOD varied over the range M and LSV M. The background signal obtained with SWV was very high; making it very difficult to reliably measure the current due to the interfering peaks. The background signal for CA was lower than for each of the other techniques, rendering CA the optimal method of detection for p-ap. The use of a low applied potential in this method will minimise the effect of electrochemical interferences. 5. ph adjustment The volume of 0.01 M HCl required to adjust the ph of urine to 7.0 from its original ph (Table 2) and the peak currents due to interfering species (in case of blank urine) were determined (Table 3). However, when aliquots of the urine samples were evaporated to dryness and re-dissolved in a volume of water equal to the initial volume of urine, the final ph was less than that of the fresh urine samples. Urine contains bicarbonate, phosphate and ammonium ions, the relative ratio of which usually determines the ph of the urine samples. In this case, it appears that when the sample was evaporated to dryness, there was a loss of ammonia, leading to a shift to lower ph values [23]. The lyophilised urine samples from RIVM had higher ph, in contrast to the results obtained with the fresh urine samples (Table 2). These differences may arise from the lyophilisation procedure utilised or from differences in the feed, management or breed of the two sets of animals. The range of ph values of 26 different urine samples is listed in Table 1. For the type of biosensor outlined here (and indeed for any biosensor) to perform optimally, it is necessary that the ph of the measuring solution be controlled. This can be achieved by addition of buffer (dried) to the sensor or by dilution of the urine sample with a buffer solution at the appropriate ph. Fig. 6. Plot of (A) ph as a function of time of collection from three different animals, and (B) peak current at 650 mv obtained in fresh urine as a function of the time of collection from a single animal. Table 3 Comparison of ph and volume of acid required to neutralize a urine sample with time Time of collection (h) [AA] (mm) [UA] (mm) [XA] (mm) [Protein] % (g g 1 ) i p (650 mv) ( A) ph Volume of acid (ml)

7 M. Jamal et al. / Analytica Chimica Acta 554 (2005) The ph of urine samples from three different animals varied from 8.15 to 9.0 over a period of 100 h (Fig. 6A). Such changes do not affect the detection of p-ap, but obviously could affect the response by altering the activity of an enzyme label if measurements were made in this ph range. In terms of electrochemical interferences, changes in the peak currents at 650 mv of more than 10 A were observed (Fig. 6B). The amounts of ascorbic acid, uric acid, xanthine, protein and the ph of urine samples from a single animal were determined at different times (Table 3) and showed substantial changes over time. 6. Conclusions The amounts of uric acid, ascorbic acid, xanthine and protein were quantified in bovine urine. On examination of the effect of these interferences on the detection of p-ap, ascorbic acid was shown to be the major electrochemical interference. While exposure of the sample to air and light can eliminate the effect of ascorbate, the time required is long, ca. 30 min in the presence of 10 M Cu 2+. A simple washing step can remove all electrochemical interferences in the detection of p-ap at 0.1 V. The other major interference arises from the screening of analytes by proteins. Addition of ANS at concentrations up to 10 M can be used to eliminate this screening while not affecting the limit of detection of p-ap. Chronoamperometry was found to be the optimal electrochemical method of detection of p-ap with a limit of detection of 0.5 gml 1 in urine. The ph of a series of urine samples ranged from 8.22 to Work is now in progress to develop an immunosensor for steroids in bovine urine. Acknowledgements This research was funded by the EC Quality of Life Programme (QLRT ). References [1] S.A. Hewitt, M. Kearney, J.W. Currie, P.B. Young, D.G. Kennedy, Anal. Chim. Acta 473 (2002) 99. [2] Y. Martin, Analyst 125 (2000) [3] E. Sangiorgi, M. Curatolo, W. Assini, E. Bozzoni, Anal. Chim. Acta 483 (2003) 259. [4] R. Draisci, L. Palleschi, C. Marchiafava, E. Ferretti, F. Delli Quadri, J. Chromatogr. A 926 (2001) 69. [5] M. Horie, H. Nakazawa, J. Chromatogr. A 882 (2000) 53. [6] A. Boenke, Anal. Chim. Acta 473 (2002) 83. [7] E. Magner, Analyst 123 (1998) [8] E. Magner, Analyst 126 (2001) 861. [9] D. Pharon, Introduction to Biochemistry, hurine3.htm. [10] V.R. Meyer, Practical High Performance Liquid Chromatorgraphy, John Wiley and Sons, 1988, p [11] M.A. Ross, J. Chromatogr. B 657 (1994) 197. [12] M.M. Bradford, Anal. Biochem. 72 (1976) 248. [13] T. Selvaraju, R. Ramaraj, Electrochem. Commun. 5 (2003) 667. [14] J. Chen, L. Gorton, B. Akesson, Anal. Chim. Acta 474 (2002) 137. [15] E.T.G. Cavalheiro, K.A. Nour, A.B. Toth, J. Braz. Chem. Soc. 11 (2000) 512. [16] P. Hulthe, B. Hulthe, K. Johannessen, J. Engel, Anal. Chim. Acta 198 (1987) 197. [17] L. Henning, M.M. Baizer, Organic Electrochemistry, An Introduction and Guide, Marcel Dekker, New York, 1991, p [18] G.F. Zhang, H.Y. Chan, Jpn. Soc. Anal. Chem. (2000) 16. [19] T. Khan, A.E. Martell, J. Am. Chem. Soc. 89 (1967) [20] C.A. Burtis, E.R. Ashwood, Tietz Fundamentals of Clinical Chemistry, W.B. Saunders Company, Pennsylvania, 1996, p [21] M.E. Georgiou, C. Georgiou, M.A. Koupparis, Anal. Chem. 71 (1999) [22] C.A. Burtis, E.R. Ashwood, Tietz Fundamentals of Clinical Chemistry, W.B. Saunders Company, Pennsylvania, 1996, p [23] A. Fura, T.W. Harper, H. Zhang, L. Fung, W.C. Shyu, J. Pharm. Biomed. Anal. 32 (2003) 513.