Analyst, May 1994, Vol. 119 949

Selective Membranes for the Construction and Optimization of an Amperometric Oxalate Enzyme Electrode

Subrayal M. Reddy, S h u s P. J. Higson, Ian M. Christie and Pankaj M. Vadgama* Department of Medicine, (Section of Clinical Biochemistry), University of Manchester, Hope Hospital, Eccles Old Road, Salford, UK M6 8HD

Oxalate oxidase has been immobilized within a membrane laminate. Integration of the laminate within an amperometric electrode afforded a sensor for determining oxalate via electrochemical oxidation of H202 and subsequent current generation. A study of novel poly(viny1 chloride) (PVC) and cellulose acetate (CA) membranes with various outerhner membrane configurations made it possible to eliminate direct electrochemical interference from species present in urine, typically ascorbate (60-500 pmol PI), homovanillic acid (404% pmol l-1) and oxalate itself (50-500 p o l 1-1). Of the membranes studied, only CA and plasticized PVC were found to screen out the direct oxalate electrochemical signal effectively. Cellulose acetate exhibited the best selectivity ratio for H202 over urine interference. An outer 0.05 pm pore radius polycarbonate membrane imparted sensor linearity in the narrow range 2- 200 pmoll-1 oxalate, which fell below the normal range of oxalate in urine. The application of unplasticized PVC as an outer membrane allowed the sensor to exhibit linear characteristics which extended beyond the clinically relevant range (>700 pmol l-1). The MichaebMenten constant of the immobilized enzyme was measured and found to be 10-3 mol 1-1. Keywords: Oxalate oxidase; enzyme electrode; Membrane; poly(viny1 chloride); cellulose acetate

Introduction Measurement of oxalate in urine is important in the investiga- tion and correct treatment of urolithiasis, chronic renal failure and primary hyperoxaluria.1 Earlier methods for measuring oxalate2 involved liquid-liquid extraction, precipitation and isotope dilution. These were found to be cumbersome, expensive and inaccurate. Later methods focused on enzyme- based determinations, e.g., utilizing the enzyme oxalate decarboxylase (ODC) (EC 4.1.1.2) :

( O W

Here, oxalate was estimated by measuring the C02 evolution directly either by a pH change or by a colour change of a suitable indicator. This method, however, led to underestima- tion owing to loss of C02. Alternatively, the formate product was quantified by reaction with a second enzyme, formate dehydrogenase, which required the co-factor nicotinamide adenine dinucleotide (NAD+). This method was plagued by the contamination of commercial NAD+ with formic acid.3 The enzyme oxalate oxidase4 (OOD) (EC 1.2.3.4) has proved to be more reliable for oxalate determinations:

@OD) (COOH);! + 02- 2C02 + H202 The H202 product is determined enzymically with horseradish peroxidase, which effects a chromogen coupling reaction

(1) (COOH)2 - C02 + HCOOH

(2)

To whom correspondence should be addressed.

resulting in an indamine dye which can be measured spectrophotometrically.5 This method, although currently adopted by most routine laboratories, requires time-consum- ing pre-treatment procedures including multiple step pH adjustments, activated charcoal treatment (notably to remove ascorbate interference) and centrifugation of the samples.

More recently, oxalate determination has been exploited by enzyme electrode technology utilizing oxalate decarboxylase and monitoring the C02 product potentiometrically6 [eqn. (l)]. An alternative, simplified method for oxalate quantifica- tion is presented here in the form of an amperometric enzyme electrode which uses oxalate oxidase. The peroxide product of the enzyme reaction is oxidized and subsequently detected amperometrically at an oxidizing potential, as reported by Bradley and Rechnitz7 and Amini and co-~orkers:8~9

H202 + O2 + 2H+ + 2e- (3) This paper highlights the use of novel selective poly(viny1 chloride) (PVC) and cellulose acetate (CA) membranes to achieve the necessary screening from electrochemical interfer- ents.10 Poly(viny1 chloride) membranes have been widely used in potentiometric electrodes,ll providing a matrix for the selective ligand and facilitating selectivity by ionic exclusion through the incorporation of plasticizer and/or ion exchangers in the polymer matrix.l2 Cellulose acetate membranes also possess charge rejection properties notably to anions, but such membranes can also reject neutral molecules on the basis of size.13 In urine, the interferents are primarily ascorbate, urate and aromatic hydroxy compounds, typically homovanillic acid and DL-vanillomandelic acid.

The appropriate choice of inner and outer covering membranes for the oxalate enzyme electrode would enable enzyme kinetics and consequently the electrode signal output to be optimized for measurement in urine.

Experimental

An oxygen electrode (Rank Brothers, Bottisham, Cambridge- shire, UK) was used for electrochemical detection in the oxalate oxidase enzyme electrode. The working electrode was polarized at +650 mV (versus Ag-AgC1) for oxidation of H202. The cell consisted of a central 2 mm diameter platinum disc anode, with an outer pre-chloridizedl4 silver ring (12 mm 0.d. and 1 mm i.d.) acting as a combined reference and counter electrode. The polarizing source was an 'in-house' built potentiostat (Chemistry Workshops, University of New- castle) equipped with an ammeter. An x-t chart recorder was used to monitor electrode responses.

Oxalate oxidase from barley seedlings (0.94 U mg-1 solid) (1 U = 16.67 nkat) was obtained from Boehringer Mannheim (Lewes, East Sussex, UK). The polycarbonate membranes used (0.05 and 5 pm pore radii) were purchased from the Poretics Corporation (Livermore, CA, USA). Cuprophan dialysis membranes were obtained from Gambro (Lund, Sweden). High relative molecular mass (M, ) PVC ( M , =

Publ

ishe

d on

01

Janu

ary

1994

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 28

/10/

2014

17:

39:1

1.

View Article Online / Journal Homepage / Table of Contents for this issue

950 Analyst, May 1994, Vol. 11 9

200000) was supplied by BDH (now Merck) (Poole, Dorset, UK); the tetrasodium salt of ethylenediaminetetraacetic acid (EDTA), CA (39.8% acetyl content), glutaraldehyde (50% v/ v) and 4-hydroxy-3-methoxyphenylacetic acid (homovanillic acid; HVA) were from Aldrich (Gillingham, Dorset, UK); succinic acid, sodium oxalate and bovine serum albumin (BSA) were from Sigma (Poole, Dorset, UK); isopropyl myristate (IPM) was from Fluka (Glossop, Derbyshire, UK); the solvents tetrahydrofuran (THF) and acetone were obtained from Fisons (Loughborough, Leicestershire, UK). Fresh urine samples were kindly provided by Hope Hospital (Clinical Biochemistry Laboratories); these had been pre- acidified to pH 2.85 using HCl (5 mol 1-l) and stored at 4 "C.

The following polymer solutions were prepared: 0.06 g of PVC in THF (5 ml); 0.06 g of PVC and 150 pl of IPM in THF (5 ml); and 0.15 g of CA in acetone (10 ml). Membranes were cast in the following manner: each polymer solution was poured on a glass Petri dish (9 cm i.d.) and then rotated manually on a horizontal, flat surface; this gave an even distribution of polymer over the glass. Use of a covering glass lid allowed for slow, controlled evaporation of solvent (2 d) and the creation of membranes of even thickness. Membrane thickness was measured using a micrometre screw-gauge except for the IPM plasticized PVC membrane; owing to the fragile nature of this membrane, thickness was assessed by scanning electron microscopy.

A succinate buffer consisting of 0.05 mol 1-l succinic acid and 1 x mol 1-l EDTA (pH 2.85) in distilled water was prepared. For the selectivity studies a 1 cm2 portion of the membrane under study and/or dialysis membrane was placed on the working electrode with the dialysis membrane facing the electrode surface. This ensured electrolyte communciation between the working and Ag-AgC1 reference electrodes for each of the selective-membranes studied. For enzyme elec- trode preparation a solution mixture of oxalate oxidase (2.69 U ml-l) and BSA (0.05 g ml-l) was prepared in doubly distilled water. A 6 pl aliquot of the enzyme solution and 3 pl of glutaraldehyde (2.5% v/v in doubly distilled water) were then mixed rapidly and placed on a 1 cm2 portion of designated inner membrane (Fig. 1). A 1 cm2 portion of an outer membrane was then placed over the enzyme layer, and two glass plates were used to compress the enzyme and membrane laminate under mild finger pressure (approxi- mately 5 min). The resulting laminate (comprising the immobilized enzyme) was washed free of excess glutaral- dehyde using a jet of buffer solution and then placed over the working electrode where final electrode assembly was estab- lished by fixation with a rubber O-ring.

Results and Discussion The fabricated membranes were transparent and colourless in

Outer protective membrane

C ross-lin ked enzyme matrix

Inner permselective membrane

H,O, + 2C0,

* 0,+2H++2e-

Fig. 1 Schematic diagram of the oxalate enzyme electrode laminate system

appearance. The plasticized PVC (pPVC) membrane adhered to the dish surface and could only be removed in portions. Firstly, an area of the membrane was brought into contact with a portion of wet dialysis (support); the membrane was then lifted off the glass surface, and the required area cut out. In contrast the dry unplasticized PVC (uPVC) and CA membranes did not adhere to the glass surface and were freely removed from the Petri dish.

In the development of a clinical amperometric sensor, one of the foremost considerations is to remove electrochemical interference. A variety of organic compounds present in urine may interfere in this instance,l5 including interference from the direct electrochemical oxidation of oxalate itself. 16 Of the organic compounds present in urine, low relative molecular mass species such as ascorbate and aromatic hydroxy com- pounds, typically HVA,*7 have been identified as being the major potential contributors to interference in amperometric assays. Hydrodynamic voltammograms obtained for key interferents were compared with that of an equivalent concentration of H202 (Fig. 2). The anions ascorbate and oxalate exhibited classical electrochemical wave characteris- tics and reached plateau regions. It was evident that at potentials below 550 mV (versus Ag-AgCl), responses due to both oxalate and HVA were negligible compared with the corresponding peroxide response. Therefore, lowering the polarizing potential for peroxide detection provided a possible rationale for removing unwanted signals. The corresponding ascorbate signal, however, was still significant and comparable to that of the peroxide even down to +350 mV (versus Ag- AgCl); this suggested that with a typical mediator system18 to reduce the required overpotential, the interferent signal could not be completely eliminated.

A range of membranes varying in permeability (namely porous polycarbonate, Cuprophan dialysis, uPVC, pPVC and CA membranes) were studied in an attempt to identify one that exhibited good selectivity for H202 and which could then be used as the inner membrane for the oxalate enzyme electrode. A study of standard membranes (Fig. 3) showed that neither the porous polycarbonate nor dialysis membranes were sufficiently selective for H202 and that urine contained important interferents.

Recently, Christie et al. 19 reported the application of IPM plasticized PVC membranes to amperometric sensors. The membrane conferred high selectivity, notably exhibiting a 7- fold greater selectivity for H202 over the anionic interferent ascorbate. A study of the pPVC membrane was consistent with this observation [Fig. 4(a)] with additional quenching of direct oxalate interference. The excellent screening towards these anionic species was most probably a consequence of the inability of such charged species to penetrate the highly

80 I 1

60 1 40

20

300

fl

500 700 900 Bias voltagdmv

Fig. 2 Comparative hydrodynamic voltammograms for 0.1 mmol l-l solutions of A, H202; B, ascorbate; C, HVA; and D, oxalate through a non-selective Cuprophan dialysis membrane

Publ

ishe

d on

01

Janu

ary

1994

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 28

/10/

2014

17:

39:1

1.

View Article Online

Analyst, May 1994, Vol. 119 951

hydrophobic barrier presented by the pPVC membrane. The corresponding responses to urine and HVA, however, sur- passed that for peroxide. This latter problem was probably due to the lipophilic nature of this membrane maintaining partitioning of neutral and uncharged organic species,20 the effect being augmented for urine which comprised a pool for such species.

The uPVC membrane was devoid of lipid and was able to reduce relative responses to urine and HVA [Fig. 4(b)]. Notably the membrane was also permeable to oxalate and ascorbate alike, presumably owing to the more polar nature of the polymer matrix in the absence of plasticizer. However, the general permeability observed for uPVC precluded its use for urine measurement. The response time to the interferents, and particularly to oxalate, were retarded in comprison with the other oxalate permeable membranes (Table 1). It was probable that the uPVC membrane was exhibiting a reduced oxalate permeability coefficient as compared with either the polycarbonate or dialysis membranes.

Selectivity studies for the CA membrane [Fig. 4(c)] revealed an optimum for the peroxide signal in conjunction with significantly reduced interference from urine. A direct comparison of the response ratios (H202 : urine) together with membrane thickness data (Table 1) best exemplified the obvious attraction of the CA membrane (to be employed as an inner membrane in the enzyme electrode) over the other homogeneous and porous membranes studied. Cellulose acetate membranes13-21 have previously been shown to exhibit enhanced selectivity for H202 over ascorbate22 and have consequently found application for both lactate” and urate24 enzyme electrodes.

A calibration of the oxalate enzyme electrode consisting of an outerhner configuration of 5 pm/CA compared with those of 5 pm/pPVC and 5 pm/uPVC configurations revealed an

100

n Ascorbate HVA Urine

Oxalate Peroxide Fig. 3 Permeability characteristics of (a) a bare electrode, (b) a 0.05 pm polycarbonate membrane and (c) a Cuprophan dialysis membrane to key interferents at 0.1 mmol 1-l concentrations (except undiluted urine)

inner membrane dependence on the signal size and possibly the linear range of the sensor (Fig. 5). Molecular oxygen is the natural co-substrate for the oxidase-based enzymes and is a requisite for the enzyme reaction [eqn. (l)]. A restriction on PO:! in oxidase-based enzyme electrodes has been shown to have a detrimental effect on the linear range of the sensors.25 In an enzyme electrode, substrate levels are usually in excess of oxygen content. The oxygen consumed in the enzyme reaction essentially renders the enzyme layer deficient in the co-substrate. This may be replenished by a further supply of oxygen (as is dissolved in the bulk solution) through an oxygen-selective outer membrane or by permeation of elec- trochemically generated oxygen [eqn. (3)] back through the inner membrane. For the latter, oxygen supply will be influenced by membrane permeability. Literature values for the oxygen permeability coefficients of pPVC and CA membranes (Table 2) suggested that for Fig. 5, linearity may well be enhanced by a more oxygen-permeable inner mem- brane. An analogous effect was reported for silane-treated oxygen-permeable polycarbonate membranes used as inner membranes at a glucose enzyme electrode.28

20

10

0 0.50

a 0.25 g 0 Q

g o CL:

30

20

10

n 1

Ascorbate HVA Urine Oxalate Peroxide

Fig. 4 Permeability characteristics of (a) PVC-IPM, (b) uPVC and (c) CA membranes to key interferents at 0.1 mmol l-l concentrations (except undiluted urine)

Table 1 Relationship between membrane selectivity, thickness and amperometric response time to oxalate

Response ratio, H202 : urine Membrane

Membrane (H202 selectivity thickness*/pm 0.05 pm

Dialysis 0.27 14 UPVC 1.77 10 pPVC 0.49 14 CA 16.8 10 * f2 pm.

polycarbonate 0.50 7

Response time

oxalate/min (t99) to

2-5 2-5 8-10 No response No response

Publ

ishe

d on

01

Janu

ary

1994

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 28

/10/

2014

17:

39:1

1.

View Article Online

952 Analyst, May 1994, Vol. 119

The established diffusion-limiting properties of the uPVC membrane towards oxalate (Table 1) were exploited in an attempt to improve the linear characteristics of the enzyme electrode (with a CA inner membrane). Comparison of uPVC and 0.05 pm polycarbonate as outer membranes confirmed the expected extension in the operability range of the sensor when using the former (Fig. 6); this surpassed the upper limit of detection quoted by Bradley and Rechnitz7 (500 pm) for their amperometric oxalate sensor which consisted of a dialysidlO- 15 pm Gore-Tex outerlinner membrane configuration. Evi- dently, by using a uPVC outer membrane, the low permeabi- lity of substrate to the enzyme layer placed no apparent limitation through excessive co-substrate consumption. The Michaelis-Menten constant (K,) for the immobilized enzyme system was determined by constructing a Lineweaver-Burk plot from the calibration graph obtained when using uPVC as the outer membrane of the oxalate enzyme electrode. The value obtained was mol l-l, which was an order of magnitude higher than for the enzyme in solution4 ( K , = mol 1-l), indicating further that the functioning of the sensor was independent of enzyme kinetics. Additionally, with an outer uPVC membrane, responses were found to be depen- dent only on transport through the membrane and, therefore, independent of bulk solution turbulence. Stirring of the bulk solution, therefore, became unnecessary.

Conclusions The CA membrane exhibited superior selectivity properties against the common interferents present in urine with an associated enhanced permeability towards H202. The pPVC membrane was shown to be unsuitable for oxalate sensor application owing to its ability to facilitate the transport of

40 I

0 40 80 120 160 200 [C,o,~pmol I-’

Fig. 5 Extension in linear range when using A, PVC-IPM and B, uPVC inner membranes compared with a CA inner membrane (C) (all with underlying dialysis)

Table 2 Oxygen permeability coefficients for homogeneous mem- branes at standard temperature and pressure (STP)

Membrane p 0 2 cm3-(STP) c d s cm2 cmHg Ref. Homogeneous dense CA 1.22 x 10- lo 26 Homogeneous pPVC 40.3 X 10-lO 27

0 200 400 600 800 [C20,2Ypmol r1

Fig. 6 Extension in linear range when using a uPVC outer membrane (A) compared with a 0.05 pm polycarbonate outer membrane (B) (both with a CA inner membrane)

~~

organic electrochemical interferents likely to be present in urine, although it did exhibit complete screening af oxalate and ascorbate. The uPVC membrane proved to be non- selective, although its general diffusion-limiting properties helped to extend linearity. It is possible that a uPVC outer membrane may eventually be used to modify the linear range of enzyme electrodes for anionic substrates, viz., lactate and urate. The application of uPVC and pPVC as inner mem- branes also imparted stable linear characteristics, but at the expense of signal size attenuation, owing to restricted permeation of the signal species, Hz02. The best option for practical urine measurement therefore remains with the exploitation of an inner cellulosic membrane.

S. M. R. thanks the SERC for financial support in conducting this work.

1

2

3

4 5

6

7 8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26 27

28

References Hodgkinson, A., Oxalic Acid in Biology and Medicine, Aca- demic Press, New York, 1977. Hodgkinson, A., Clin. Chem. (Winston-Salem, N.C. ) , 1970,16, 547. Costello, J., Hatch, M., and Bourke, E., J. Lab. Clin. Med., 1976,87,903. Chiriboga, J., Biochem. Biophys. Res. Commun., 1963,11,277. Laker, M. F., Hofmann, A. F., and Meeuse, B. J. D., Clin. Chem. (Winston-Salem, N . C . ) , 1980, 26, 827. Vadgama, P., Sheldon, W., Guy, J. M., Covington, A. K., and Laker, M. F., Clin. Chim. Acta, 1984, 142, 193. Bradley, C. R., and Rechnitz, G. A., Anal. Lett., 1986,19,151. Amini, M. A. S., Vallon, J. J., and Bichon, C., Anal. Lett., 1989,22,43. Amini, M. A. S., and Vallon, J. J., Anal. Chim. Acta, 1991,245, 129. Palleschi, G., Rhani, M. A. N., Lubrano, G. J., Ngwainbi, J. N., and Guilbault, G. G., Anal. Biochem., 1986, 159, 114. Moody, G. J., Oke, R. B., and Thomas, J. D. R., Analyst, 1970, 95,910. Morf, W. E., and Simon, W., in Ion Selective Electrodes in Analytical Chemistry, ed. Freiser, H., Plenum Press, New York, 1978, vol. 1, p. 236. Koochaki, Z., Christie, I., and Vadgama, P., J . Membr. Sci., 1991,57, 83. Higson, S. P. J., and Vadgama, P., Electroanalysis, 1993, in the press. Tietz, N. W., Fundamentals of Clinical Biochemistry, Saunders, London, 2nd edn., 1976. Beden, B., Cetin, I., Kahyaoglu, A., Takky, D., and Lamy, C., J. Catal., 1987, 104, 37. Richards, D. A., and Titheradge, A. C., Biomed. Chromatogr., 1987, 2 , 115. Turner, A. P. F., Karube, I., and Wilson, G. S., Biosensors: Fundamentals and Applications, Oxford University Press, Oxford, 1987. Christie, I. M., Treloar, P. H., and Vadgama, P., Anal. Chim. Acta, 1992, 269,65. Baker, N., and Hadgraft, J., Faraday Discuss. Chem. SOC., 1984,77,97. Tsuchida, T., and Yoda, K., Enzyme. Microb. Technol, 1981,3, 326. Tone, S., Demiya, M., Shinohara, K., and Otake, T., J . Membr. Sci., 1984, 17,275. Mullen, W. H., Churchouse, S. J., Keedy, F. H., and Vadgama, P. M., Clin. Chim. Acta, 1986, 157, 191. Keedy, F. H., and Vadgama, P., Biosens. Bioelectron., 1991,6, 41. Tang, L. X., and Vadgama, P., Med. Biol. Eng. Comput., 1990, 28, B18. Puleo, A. C., and Paul, D. R., J. Membr. Sci., 1989,47, 301. Pal, S. N., Ramani, A. V., and Subramanian, N., J . Appl. Polym. Sci., 1992,46,981. Higson, S . P. J., Koochaki, Z., and Vadgama, P. M., Anal. Chim. Acta, 1993,276, pp. 335-340.

Paper 3104533A Received July 29, 1993

Accepted November 10, 1993

Publ

ishe

d on

01

Janu

ary

1994

. Dow

nloa

ded

by U

nive

rsity

of

Illin

ois

at C

hica

go o

n 28

/10/

2014

17:

39:1

1.

View Article Online