- Email: [email protected]
A disposable biosensor for urea determination in blood based on an ammonium-sensitive transducer
Biosensors & Bioelectronics 14 (1999) 33–41
A disposable biosensor for urea determination in blood based on an ammonium-sensitive transducer Claudia Eggenstein a,b, Michael Borchardt a, Christoph Diekmann a, Bernd Gru¨ndig a, Christa Dumschat a, Karl Cammann a, Meinhard Knoll a, Friedrich Spener a,b,* b
a Institut fu¨r Chemo- und Biosensorik Mu¨nster, Mendelstr. 7, D-48149 Mu¨nster, Germany Institut fu¨r Biochemie, Universita¨t Mu¨nster, Wilhelm-Klemm-Str. 2, D-48149 Mu¨nster, Germany
Received 17 November 1997; received in revised form 10 August 1998; accepted 13 August 1998
Abstract A potentiometric urea-sensitive biosensor using a NH+4-sensitive disposable electrode in double matrix membrane (DMM) technology as transducer is described. The ion-sensitive polymer matrix membrane was formed in the presence of an additional electrochemical inert filter paper matrix to improve the reproducibility in sensor production. The electrodes were prepared from one-side silver-coated filter paper, which is encapsulated for insulation by a heat-sealing film. A defined volume of the NH+4-sensitive polymer matrix membrane cocktail was deposited on this filter paper. To obtain the urea-biosensor a layer of urease was cast onto the ionsensitive membrane. Poly (carbamoylsulfonate) hydrogel, produced from a hydrophilic polyurethane prepolymer blocked with bisulfite, served as immobilisation material. The disposable urea-sensitive electrode was combined with a disposable Ag/AgCl reference electrode to obtain the disposable urea biosensor. The sensor responded rapidly and in a stable manner to changes in urea concentrations between 7.2 ⫻ 10−5 and 2.1 ⫻ 10−2 mol/l. The detection limit was 2 ⫻ 10−5 mol/l urea and the slope in the linear range 52 mV/decade. By taking into consideration the influence of the interfering K+- and Na+- ions the sensor can be used for the determination of urea in human blood and serum samples (diluted or undiluted). A good correlation was found with the data obtained by the spectrophotometric routine method. 1999 Elsevier Science S.A. All rights reserved. Keywords: Disposable biosensor; Disposable reference electrode; Potentiometric biosensor; Urea biosensor; Blood urea
1. Introduction The basic prerequisite for the availability of disposable potentiometric biosensors is a reproducible low cost production technology for ion-sensitive electrodes. To meet this requirement, coated-film electrodes were developed by the Eastman Kodak Company (Battaglia et al., 1980). However, controlled manufacturing of such electrodes requires sophisticated, elaborate and rather expensive processes. Our solution to this problem is the so-called double matrix membrane (DMM) technology in which an ion-selective polymeric membrane is formed in the volume of an electrochemical inert micro fiber matrix (simple filter paper) by a casting technique
* Corresponding author. Tel.: 00-49-141-83-33100; fax: 00-49-25183-32132; e-mail: [email protected]
(Knoll, 1991; Dumschat et al., 1994). There the polymer matrix homogeneously distributes by capillary forces and, after the evaporation of the solvent, the double matrix membranes are obtained which have better mechanical stability than traditional polyvinylchloride (PVC) membranes. As described for K+-sensitive membranes (Knoll, 1991) and for pH-sensitive membranes (Dumschat et al., 1994) the additional filter paper has no influence on the electrochemical response behaviour of the membranes. The DMM technology was used to develop disposable Na+- and K+-sensitive electrodes (Borchardt et al., 1994, 1995). An important parameter in clinical analysis is urea where an increased concentration in blood and a reduced level in urine is a strong indication for renal dysfunction. The determination of urea in body fluids is one of the most frequent analyses in routine clinical laboratories. The normal range in human serum is between 1.7 and 8.3 mmol/l (Rick, 1990) and levels increase up to
0956-5663/98/$ - see front matter 1999 Elsevier Science S.A. All rights reserved. PII: S 0 9 5 6 - 5 6 6 3 ( 9 8 ) 0 0 1 0 3 - 1
34
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
100 mmol/l under pathophysiological conditions (Rick, 1990). Urea is hydrolysed by urease according to the reaction (NH2)2CO ⫹ 2 H2O ⫹ H+→2 NH+4 ⫹ HCO−3. In conventional urea sensors, pH electrodes (Ruzicka et al., 1979;Vadgama et al., 1982; Miyahara et al., 1985; Karube et al., 1986; Hamann et al., 1986) and NH+4selective electrodes (Guilbault and Montalvo, 1969, 1970; Campanella et al., 1990) have been used to detect the respective hydrogen ions or ammonium ions produced in the enzymatic reaction. The major problem for pH-sensitive electrodes is that the sensor response is strongly dependent on the buffer capacity of the sample because the pH change produced in the course of the enzyme-catalysed reaction is suppressed by the buffer used, which leads to a narrow dynamic range and a loss in sensor sensitivity (Ruzicka et al., 1979; Eddowes, 1987; Koncki et al., 1992). The general problem of NH+4-sensitive electrodes is interference by Na+- and K+ions, which are present in serum (4.5 mmol/l K+ and 140 mmol/l Na+) and urine. In the present study the development of a low cost disposable urea-sensitive biosensor in the form of a test strip is described. The sensor consists of a disposable NH+4-sensitive electrode in DMM technology as transducer, which is covered with a layer of immobilized urease, and a disposable Ag/AgCl reference electrode. The fundamental characteristics of the sensor configuration are presented. The influence of ionic interferences on the response behaviour of the urea electrode is shown and furthermore results of urea measurements in blood and serum samples are presented.
2. Experimental section 2.1. Reagents Tetrahydrofuran (THF) was purchased from Aldrich (99.5 ⫹ %, spectrophotometric grade, inhibitor-free), cyclohexanone (extra pure) from Merck. PVC of high molecular weight and all other chemicals used for the preparation of the NH+4-sensitive membranes (selectophore quality) were obtained from Fluka and applied as received. Urease (260 U/mg) was purchased from Serva. Tris (hydroxymethyl)-aminomethane (Tris) and ethylenediaminetetraacetic acid (disodium salt) (EDTA) were obtained from Sigma (analytical grade). KCl, NaCl, NH4Cl and urea were purchased from Merck. Urease (5 U/mg) used for the spectrophotometric Berthelot method was from Merck. Phenol and sodium hypochloride were purchased from Sigma and sodium nitroprusside dihydrate was obtained from Fluka. As micro fiber matrix (MFM) material filter paper from (Schleicher & Schuell (5893 Blue ribbon ashless, Ref. No. 300211) was used. To support and insulate the filter
paper a 150-m thick polyester/polyethylene heat-sealing film (Team Codor, DUBO-Schweitzer GmbH) was used. The poly(carbamoylsulfonate) (PCS)-prepolymer used for the entrapment of urease is a research product and was a kind gift by Professor Vorlop (Forschungsanstalt fu¨r Landwirtschaft, Braunschweig). Polyethyleneimine (PEI) was obtained from Sigma. 2.2. Solutions All calibration and buffer solutions were prepared with double-distilled deionized water. The sample buffer (pH 7.5) containing 20 mmol/l Tris and 1.0 mmol/l EDTA was adjusted with HCl. Stock solutions of urea, 1.0 and 0.1 mmol/l, and those of NH+4, 1.0 and 0.1 mmol/l, were prepared daily in Tris buffer. The sample buffer solution containing 140 mmol/l NaCl and 4.5 mmol/l KCl was also prepared in Tris buffer. From the latter solutions containing other concentrations of NaCl and KCl were obtained by serial dilution with Tris buffer. 2.3. Sensor preparation Fig. 1 shows the final configuration of the disposable urea biosensor consisting of the urea-sensitive electrode and the reference electrode. The urea electrode is based on a disposable NH+4-sensitive DMM electrode as transducer. The transducer electrodes were produced in a batch process as described in detail by Borchardt et al. (1994) and Eggenstein et al. (1995). To prepare the silver contact and the conducting track, silver was evaporated on one side of the filter paper (vacuum coater, Univex 300, Leybold). For insulation the filter paper was encapsulated with a preperforated heat-sealing film. By this procedure the filter paper was completely covered with heat sealing film except for a hole of 3 mm diameter on one side for membrane deposition and a 5-mm long contact pad for the electrical contact on the other side. The following mixture was used for the preparation of the NH+4-sensitive DMM (wt%): 1.0% Nonactine, 66.8% bis (2-ethylsebacate) (DOS), 32.2% PVC. 400 mg of the membrane components were dissolved in 1.5 ml THF and 0.5 ml cyclohexanone. The components were allowed to dissolve overnight and shaken vigorously until a homogeneous solution was obtained. The deposition of the PVC-membrane onto the filter paper was carried out automatically with an Asymtek Automove 402 fluid dispenser to apply defined volumes into each hole. To prevent an overflow of substances three dispensing steps of 10 l for each hole were done at intervals of 5 min. After solvent evaporation overnight the NH+4sensitive electrodes were ready on the one hand for NH+4 measurements and on the other hand for the deposition of the enzyme layer on the surface of the ionsensitive PVC-membrane. The enzyme membrane was
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
Fig. 1.
35
Schematic view of the disposable urea biosensor consisting of disposable measuring and reference electrode.
prepared following the procedure of Vorlop et al. (1992) and as modified by Kotte et al. (1995). A defined amount of the aqueous PCS prepolymer solution (polymer content: 38% w/w) was diluted with water to obtain a 30% (w/w) solution. Polyethyleneimine (PEI) (2.5% w/v in water) was added dropwise to adjust the pH between 6 and 6.5. A 50-l aliquot of the hydrogel matrix was mixed with 50 l of enzyme solution (1 mg urease (260 U) in 50 l of water). A volume of 3 l of this aqueous hydrogel-enzyme suspension was deposited with a micro syringe onto the surface of the ion-sensitive membrane of the electrodes and allowed to form a gel at 4°C in a horizontal position. After 2 to 3 h, the enzyme electrodes were immersed in a dilute aqueous PEI solution (0.5% w/v) for 30 s and rinsed with water. PEI was used as coupling reagent to prevent the enzyme from bleeding out. The electrodes were stored under dry conditions at 4°C. The disposable Ag/AgCl reference electrodes were prepared in the same way as described by Diekmann et al. (1995) for a redox reference electrode, based on the same filter paper described above. An Ag/AgCl paste from Acheson (AV458) was used to prepare the con-
ducting track and 3 mmol/l KCl provided the internal electrolyte solution. For storage, the electrodes were immersed in 3 mmol/l KCl to avoid drying up. To obtain the final disposable urea biosensor the ureasensitive electrode and the reference electrode were glued onto a separating and insulating plastic film right before the measurements. This combination could be stored at 4°C in Tris buffer (pH 7.5) containing 1 mmol/l EDTA. 2.4. Measurements The performance of the electrodes was evaluated by their limits of detection, slopes, selectivities and by the reproducibility of the sensors produced. The calibration curves for identifying the detection limit, slope and linear range were obtained by adding known amounts of the relevant substrate and respective ion to the measuring solution from low to high concentrations. The selectivity coefficients were evaluated by the fixed interference method according to the IUPAC (1976) recommendations. Measurements were carried out either with a conventional Ag/AgCl-reference electrode from Orion
36
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
(Model 90-02) with double junction (inner filling: 3 mmol/l KCl; outer filling: sample buffer, pH 7.5) or with the disposable reference electrode developed here. For the determination of reproducibility of the sensors, eight NH+4-sensitive and five urea-sensitive electrodes from one batch, respectively, were compared with regard to slope and potential at equal concentrations, in this case the conventional Ag/AgCl electrode served as reference electrode. All measurements were made at room temperature with the aid of a pH-meter (microprocessor pH-meter pH 3000/multiplex 3000, WTW Weilheim, Germany) to which the sensors were plugged in. Before measurements the NH+4- and the urea-sensitive electrodes were equilibrated in the sample buffer until a stable potential was obtained (0.5–1 h) and then immersed with the reference electrode in 10 or 40 ml stirred sample buffer (pH 7.5) and the calibration graphs were recorded as described above. To gauge the influence of interfering ions on the response behaviour of the urea biosensor the electrodes were dipped in 10 ml of the sample buffer containing various combinations of Na+ and K+ concentrations and the calibration was attained by stepwise addition of aliquots of standard urea solutions (1.0 and 0.1 mmol/l) to the stirred test solution.
3. Results 3.1. Characterisation of the measuring electrode At first NH+4-sensitive disposable electrodes in DMM technology were characterised to evaluate whether they could serve as transducer in sensing urea. The electrochemical response behaviour of the NH+4-sensitive electrode versus a commercial Ag/AgCl-reference electrode and the corresponding calibration graph are shown in Fig. 2. The electrode responded rapidly and in a stable manner to changes in the NH+4-concentration with a response time in the range of a few seconds. As determined from the calibration graph the detection limit was 5.0 ⫻ 10−6 mol/l, the lowest limit of linear range 3.2 ⫻ 10−5 mol/l and the slope in the linear range amounted to 55 mV/decade. The calibration curves of eight NH+4sensitive electrodes from one lot revealed an average slope in the linear range of 54.8 ⫾ 0.3 mV/decade. It can be seen from Fig. 3 that the slopes were reproducible indeed. Taking these data we concluded that the developed NH+4-sensitive electrode developed was a suit-
2.5. Procedure for the determination of urea in serum and blood
Blood and serum samples were collected from healthy persons as well as serum samples from haemodialysis patients from a renal laboratory (Centre of Home Dialysis and Kidney Transplant, Mu¨nster). Sodium-citrate was added to blood samples to prevent clotting. Measurements were carried out with the final disposal urea biosensor as developed here. Blood and serum samples were measured undiluted or diluted 1:1 or 1:5 with Tris/HCl buffer. The urea-sensitive biosensor was immersed in the samples (1 ml sample volume) and the potential was recorded after 2–3 min when a constant signal was reached. Between trials the electrode was rinsed with double-distilled water and immersed in the sample buffer. Testing of the blood and serum samples was started again only after the electrode potential returned to its original (urea-free) value. The amount of urea was determined from standard calibration curves recorded in the sample buffer with constant interfering Na+- and K+-ion levels corresponding to the average levels in blood samples (diluted or undiluted). The Berthelot reaction, a standard method for spectrophotometric urea determination, was used as a reference (Fawcett and Scott, 1960; Bernt and Bergmeyer, 1970); the absorbance of the indophenol produced was measured at 550 nm.
Fig. 2. Response curve (A) and calibration graph (B) of the disposable NH+4-sensitive electrode in combination with the conventional Ag/AgCl-reference electrode in 0.02 mmol/l Tris/HCl buffer, pH 7.5 (in the calibration graph the potential in buffer is set to zero).
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
37
Fig. 3. Reproducibility of the batch production as evidenced by the calibration curves of eight disposable NH+4-sensitive electrodes (the potential in buffer is set to zero).
able transducer basis for the potentiometric urea sensor. For the application of the sensor to in real samples, however, it was necessary to consider that the ionophore used was sensitive not only to NH4+. To gauge the crosswere reactivities the selectivity coefficients log K pot ij determined to be Ca2 ⫹ ⫽ ⫺ 4.20, Li+ ⫽ ⫺ 3.60, Na+ ⫽ ⫺ 2.95, K+ ⫽ ⫺ 0.80. The next step was the deposition of the enzyme membrane onto the NH+4-sensitive DMM. Different polymer materials like PVC, polyvinylpyrrolidone, polyestersulfonic acid and photo-crosslinkable prepolymers were tested as gel matrix for entrapping the urease. Best results concerning adhesion capability, limit of detection, slope, linear range and lifetime were obtained with PCS cross-linked with PEI as gel material. This material not only adhered excellently to the PVC-membrane, but showed minimal diffusion resistance for various substrates, as already described by Kotte et al. (1995). The electrochemical response behaviour of such a urea-sensitive electrode versus a commercial Ag/AgCl electrode is shown in Fig. 4. As demonstrated, the electrode responded rapidly and in a stable manner to changes in the urea concentration with a detection limit as low as 1.0 ⫻ 10−5 mol/l. Fig. 5 shows the calibration curves of five urea-sensitive electrodes from one lot; the average slope in the linear range was 55.5 ⫾ 0.7 mV/decade again indicating high reproducibility. With the aim to obtain the complete disposable biosensor the urea electrode required a compatible disposable reference electrode to replace the conventional Ag/AgCl-electrode in the sensor configuration. A simple reference electrode was developed for this purpose and glued to the urea electrode, as described in the Experimental section. Fig. 1 depicts the resulting test strip for the disposable urea biosensor. Then the performance of the urea biosensor was evaluated; the respective response curve and the calibration graph derived are presented in Fig. 6. The limit of detection was 2.0 ⫻ 10−5 mol/l (in 0.02 mmol/l Tris buffer without interfering
Fig. 4. Response curve (A) and calibration graph (B) of the disposable urea-sensitive electrode in combination with the conventional Ag/AgCl-reference electrode in 0.02 mmol/l Tris/HCl buffer, pH 7.5 (in the calibration graph the potential in buffer is set to zero).
Fig. 5. Reproducibility of the batch production process as evidenced by the calibration curves of five disposable urea-sensitive electrodes (the potential in buffer is set to zero).
ions) and the slope in the linear range was about 52 mV/decade. The sensor responded within 1 and 2 min in a stable manner to changes in the analyte concentration. The urea-sensitive test strip could be stored over-
38
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
respectively, with intermittant immersion in sample buffer during nights. During these 4 days a slight decrease in electrode activity and selectivity was observed; thereafter no activity remained. Enzyme membranes stored dry at 4°C showed no deterioration in sensor performance during the 4 days and up to 3– 4 months a slight activity loss only (data not shown). The data obtained with the final configuration of the urea biosensor were compared to those obtained with the urea-sensitive electrode measured against the conventional Ag/AgCl-reference electrode (Table 1). The results revealed that the electrochemical response characteristics of both sensor configurations were comparable, only a slight decline in sensitivity and detection limit was observed for the former. 3.2. Application of the biosensors in biological samples
Fig. 6. Response curve (A) and calibration graph (B) of the disposable urea biosensor, i.e., the combination of the disposable urea-sensitive electrode and the disposable Ag/AgCl-reference electrode, in 0.02 mmol/l Tris/HCl buffer, pH 7.5 (in the calibration graph the potential in buffer is set to zero).
night in a refrigerator continuously immersed in Tris buffer pH 7.5 with 1 mmol/l EDTA, and then used again each day for several measurements. Under these conditions the life time of the disposable test strip was 4 days. Fig. 7 shows calibration graphs from measurements with the one urea-sensitive sensor test strip, first immediately after the glueing process and immersion into the sample buffer, and then after 1, 2 and 3 days,
Fig. 7. Lifetime of the disposable urea biosensor (continuous immersion in 0.02 mmol/l Tris/HCl buffer, pH 7.5).
For the determination of urea in blood it was necessary to take into account interferences by Na+- and K+ions abundantly present in such samples. To investigate the effect of increasing concentrations of these ions on the electrode response performance and to evaluate whether the biosensor developed is applicable in blood samples, response curves were recorded in background buffer solutions at different levels of the interfering ions. The calibration graphs obtained for the disposable biosensor are shown in Fig. 8 and the data derived therefrom are summarised in Table 2. Concentrations up to a level of 140 mmol/l for Na+ and 4.5 mmol/l for K+ions, corresponding to the normal level of these ions in human blood and serum, were tested. Without Na+- and K+-ions the slope in the linear range between 7.2 ⫻ 10−5 and 2.1 ⫻ 10−2 mol/l was 52.0 mV/decade. As the concentration of these ions increased, the lower detection limit of urea moved towards higher concentrations and the dynamic range of the biosensor was narrowed and, as a result, the apparent slope of the response of the sensor become smaller. The limit of detection increased from 2.0 ⫻ 10−5 up to 1.2 ⫻ 10−4 mol/l and the slope of the calibration curve decreased from 52.0 down to 31.7 mV/decade. This effect could be expected from the determination of the selectivity coefficients of the NH+4sensitive DMM for the interfering Na+- and K+-ions as descibed before. With a biosensor reference value for a linear range from 0.5 to 10 mmol/l urea (Fig. 8 and Table 2), which covers the reference interval of 2–8 mmol/l for urea in blood of a healthy person (Rick, 1990), the applicability of the urea biosensor was then checked in undiluted and diluted blood samples. For the measurement of undiluted samples (up to 10 mmol/l urea) the biosensor was calibrated in background buffer containing 140 mmol/l Na+ and 4.5 mmol/l K+, for measurements of samples diluted 1:1 (with buffer free of interfering ions) the test strip
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
39
Table 1 Electrochemical response characteristics of the disposable urea-sensitive electrode in combination with a conventional and the disposable reference electrode, respectively Urea-sensitive electrode versus
Slope (mV/decade)
Lowest limit of linear range (mol/l)
Limit of detection (mol/l)
Conventional reference electrode Disposable reference electrode
55.5 52.0
4.5 ⫻ 10−5 7.2 ⫻ 10−5
1.0 ⫻ 10−5 2.0 ⫻ 10−5
Fig. 8. Influence of increasing Na+- and K+-concentrations on the response of the urea biosensor in 0.02 mmol/l Tris/HCl buffer, pH 7.5. Table 2 Response characteristics of the urea biosensor in dependence on the concentration of interfering Na+- and K+- ions Concentration of interfering ions (mmol/l) Na+
K+
– 14 28 70 140
– 0.5 1.0 2.5 5.0
Slope
Lowest limit Limit of of linear range detection (mV/decade) (mol/l) (mol/l)
52.0 46.6 40.5 35.2 31.7
7.2 1.5 1.0 5.0 5.0
⫻ ⫻ ⫻ ⫻ ⫻
10−5 10−4 10−4 10−4 10−4
2.0 5.0 5.0 1.5 1.2
⫻ ⫻ ⫻ ⫻ ⫻
Fig. 9. Blood urea determined with the urea biosensor with reference to the spectrophotometric Berthelot method (samples from healthy persons).
10−5 10−5 10−5 10−4 10−4
was calibrated in background buffer containing 70 mmol/l Na+ and 2.25 mmol/l K+, and so on. Whereas the slope of the calibration curve remained constant during the course of a day, the entire standard curve shifted several millivolts. Therefore, the urea biosensor was calibrated after every 8 to 10 measurements. The results obtained with the urea electrode (y) were compared with values (x) obtained with the standard spectrophotometric Berthelot method in Figs. 9 and 10. The potentiometric values were mean values of double determinations with a standard deviation of 4%. The spectrophotometric values were also mean values of double determinations with a standard deviation of 5%. Fig. 9 shows the results of blood measurements of 13 healthy persons. The difference between the results obtained with the urea electrode
Fig. 10. Serum urea determined with the urea biosensor with reference to the spectrophotometric Berthelot method (samples from patients; inset: samples from healthy persons).
40
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
and the standard method was about 0.5–7.0%, the correlation plot was y ⫽ 1.086x ⫺ 0.341 (r ⫽ 0.991). Fig. 10 represents the results of serum measurements. Of the 18 samples compared, 7 were from healthy persons (inset of Fig. 10) and 11 from a laboratory attached to a renal unit. The difference between the results obtained with the potentiometric and the spectrophotometric method in this case was about 0.8–6.5% for the samples of healthy persons and 1.5–8.0% for the samples from the renal laboratory except 2 outliners (11%). The overall correlation plot was y ⫽ 0.927x ⫹ 0.824 (r ⫽ 0.994), with y ⫽ 1.154x ⫺ 0.666 (r ⫽ 0.997) for samples from the routine laboratory and y ⫽ 0.904x ⫹ 0.828 (r ⫽ 0.993) for samples from the renal laboratory.
4. Discussion The data reported here demonstrated that the DMM technology enabled the production of disposable ureasensitive electrodes, and in combination with the disposable reference electrode the production of a disposable urea biosensor with a low cost potential. The evolution of this DMM concept can be described by the following: at first the conventional ion-sensitive PVC matrix membrane was supplemented with the micro fiber matrix. For the resulting DMM no significant differences in detection limit, slope and selectivity were observed, as the additional filter paper did not influence the electrochemical responses of the cation-sensitive membranes; yet the mechanical stability was considerably improved (Knoll, 1991; Dumschat et al., 1994). Moreover, the development of the ion-sensitive electrode from an electrode with liquid internal junction to an electrode with solid contact in the form of a test strip had no detrimental effect on the response behaviour as already described for the Na+- and K+-sensitive electrodes (Borchardt et al., 1994, 1995) and demonstrated here for the NH+4-sensitive electrode. In comparison to the performance of previously published ion-selective electrodes based on a single PVC matrix, the present NH+4-sensitive DMM revealed a similar Nernstian behaviour, even to those based on a gaseous diffusion electrode (Campanella et al., 1990; Miyahara et al., 1991; Schindler et al., 1994). The main advantage of the new biosensor configuration is the technology of electrode production. Only cheap mass products were used and high productivity can be achieved by simple batch technology. Subsequent transformation of the NH+4-sensitive transducer to the ureasensitive electrode created a biosensor with an excellent response behaviour. The biosensor works in an analytical useful range, with a short response time. The slope of about 55.5 mV/decade in the linear range affords a sensitivity which is expected for NH+4-sensitive potentiometric transducers (Guilbault and Nagy, 1973; Miyahara et al., 1991).
The final goal was attained by integrating the Ag/AgCl-reference electrode into the disposable urea biosensor. The materials needed for the reference electrode, produced in a batch process, were cheap mass products as well. Investigations on the reproducibility of manufacturing and on the glueing process are under way. The slight decline seen today in the detection limit and the linear range compared to the combination of the urea-sensitive electrode with a conventional reference electrode may be caused by the contamination of the test solution with K+-ions from the internal electrolyte diffusing out of the disposable reference electrode. The good storage stability up to 4 months and the good operational stability over 4 days already are excellent criteria for a disposable biosensor. The measurements in background buffer with increasing concentrations of Na+- and K+-ions were a mandatory prelude before application of the biosensor in clinically relevant samples. From the selectivity coefficients of the NH+4-sensitive membrane for Na+- and K+-ions it was reasonable to expect that the slope and the dynamic range of the urea biosensor in the presence of the latter ions would be poorer than those measured in the absence of interfering ions. Urea could be measured directly in blood samples, if the electrode was calibrated in solutions with constant levels of interfering Na+- and K+-ions, as borne out by the close agreement with the spectrophotometric reference data. Yet for some of the 11 samples from the renal laboratory differences up to 11% were observed between the potentiometric and the spectrophotometric method. In these cases calibrations might have been carried out under a wrong assumption, as patients with renal dysfunction often not only suffer under increased urea levels in blood, but also from a disturbance in their electrolyte balance (Girndt, 1990; Pindur and Pindur, 1991). Consequently the method developed up to this stage furnishes data with sufficient accuracy for routine samples. Urea determination of samples coming out from a renal laboratory thus must be complemented by determinations of Na+- and K+-ions. This will become possible with a sensor configuration in DMM technology measuring Na+-, K+-ions and urea simultaneously or consecutively. This technology enabled single analyte determination in each case as shown earlier (Borchardt et al., 1994, 1995) and in the present investigation. An alternative possibility for urea determination, also based on DMM technology, could employ the pH-based system (Eggenstein et al., 1995).
Acknowledgements We are grateful to the Ministry of Science and Research of Northrhine-Westphalia (Germany) for financial support.
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
References Battaglia, C.J., Chang, J.C., Daniel, D.S., 1980. US Patent No. 4 214 968. Bernt, E., Bergmeyer, U., 1970. Harnstoff. In: Bergmeyer, U. (Ed.), Methoden der enzymatischen Analyse, vol. 1. Verlag Chemie GmbH, Weinheim, pp. 1738–1748. Borchardt, M., Diekmann, C., Dumschat, C., Cammann, K., Knoll, M., 1994. Disposable sodium electrodes. Talanta 41, 1025–1028. Borchardt, M., Dumschat, C., Cammann, K., Knoll, M., 1995. Disposable ion selective electrodes. Sensors and Actuators B 24/25, 721–723. Campanella, L., Mazzei, F., Sammartino, M.P., Tomassetti, M., 1990. New enzyme sensors for urea and creatinine analysis. Bioelectrochem. Bioenergetics 23, 195–202. Diekmann, C., Dumschat, C., Cammann, K., Knoll, M., 1995. Disposable reference electrode. Sensors and Actuators B 24/25, 276–278. Dumschat, C., Borchardt, M., Diekmann, C., Hepke, J., Knoll, M., 1994. Double matrix membranes for potentiometric cation sensitive sensors. Fresenius J. Anal. Chem. 348, 553–555. Eddowes, M.J., 1987. Response of an enzyme-modified pH-sensitive ion-selective device: analytical solution for the response in the presence of pH buffer. Sensors and Actuators 11, 265–274. Eggenstein, C., Borchardt, M., Dumschat, C., Gru¨ndig, B., Cammann, K., Spener, F., Knoll, M., 1995. Potentiometric biosensor in double matrix membrane technology. Biosensors and Bioelectronics 10, 595–600. Fawcett, F.K., Scott, J.E., 1960. A rapid and precise method for the determination of urea. J. Clin. Pathol. 13, 156–160. Girndt, J., 1990. Nieren- und Hochdruckkrankheiten. Schattauer Verlags GmbH, Stuttgart, New York, pp. 57–236. Guilbault, G.G., Montalvo, J.G., 1969. Urea-specific enzyme electrode. J. Am. Chem. Soc. 91, 2164–2165. Guilbault, G.G., Montalvo, J.G., 1970. An enzyme electrode for the substrate urea. J. Am. Chem. Soc. 92, 2533–2538. Guilbault, G.G., Nagy, G., 1973. Improved urea electrode. Anal. Chem. 45, 417–419. Hamann, H., Ku¨hn, M., Bo¨ttcher, N., Scheller, F., 1986. Enzyme
41
sensor system for potentiometric urea determination in serum and milk. J. Electroanal. Chem. 209, 69–76. 1976. Recommendations for nomenclature of ion-selective electrodes. Pure Appl. Chem. 48, 129–132. Karube, I., Tamiya, E., Dicks, J.M., 1986. A microsensor for urea based on an ion-selective field effect transistor. Anal. Chim. Acta 185, 195–200. Knoll, M., 1991. German Patent No. P41 37 261.1. Koncki, R., Leszcynski, P., Hulanicki, A., Glab, S., 1992. Urea sensor based on glass pH electrodes with physically immobilized urease. Anal. Chim. Acta 257, 67–72. Kotte, H., Gru¨ndig, B., Vorlop, K.D., Strehlitz, B., Stottmeister, U., 1995. Methylphenazonium-modified enzyme sensor based on polymer thick films for subnanomolar detection of phenols. Anal. Chem. 67, 65–70. Miyahara, Y., Moriizumi, T., Ichimura, K., 1985. Integrated enzyme FETs for simultaneous determination of urea and glucose. Sensors and Actuators 7, 1–10. Miyahara, Y., Tsukada, K., Miyagi, H., 1991. Urea sensor based on an ammonium-ion-sensitive field-effect transistor. Sensors and Actuators B 3, 287–293. Pindur, G., Pindur, U., 1991. Klinische Chemie und serologische Laboratoriumsdiagnostik. Wissenschaftliche Verlags GmbH, Stuttgart, pp. 244–309 Rick, W., 1990. Klinische Chemie und Mikroskopie. Springer-Verlag, Berlin, pp. 245–247. Ruzicka, J., Hansen, E.H., Ghose, A.K., Mottola, H.A., 1979. Enzymatic determination of urea in serum based on pH measurement with the flow injection method. Anal. Chem. 51, 199–203. Schindler, J.G., Schindler, M.M., Herna, K., Reisinger, E., Burk, B., Kuhlmann, U., Knaack, J., Schmidt, B., Langer, H., 1994. Harnstoff-Durchflußsensor zur ammoniumselektiv-enzymatischen Prozeßkontrolle der ku¨nstlichen Niere. Eur. J. Clin. Chem. Clin. Biochem. 32, 145–152. Vadgama, P.M., Alberti, K.G.M.M., Covington, A.K., 1982. Determination of urea in blood plasma by enzyme-pH electrode. Anal. Chim. Acta 136, 403–406. Vorlop, K.D., Muscat, A., Beyersdorf, J., 1992. Entrapment of microbial cells within polyurethane hydrogel beads with the advantage of low toxicity. J. Biotechnol. Technol. 6, 483–488.
A disposable biosensor for urea determination in blood based on an ammonium-sensitive transducer Claudia Eggenstein a,b, Michael Borchardt a, Christoph Diekmann a, Bernd Gru¨ndig a, Christa Dumschat a, Karl Cammann a, Meinhard Knoll a, Friedrich Spener a,b,* b
a Institut fu¨r Chemo- und Biosensorik Mu¨nster, Mendelstr. 7, D-48149 Mu¨nster, Germany Institut fu¨r Biochemie, Universita¨t Mu¨nster, Wilhelm-Klemm-Str. 2, D-48149 Mu¨nster, Germany
Received 17 November 1997; received in revised form 10 August 1998; accepted 13 August 1998
Abstract A potentiometric urea-sensitive biosensor using a NH+4-sensitive disposable electrode in double matrix membrane (DMM) technology as transducer is described. The ion-sensitive polymer matrix membrane was formed in the presence of an additional electrochemical inert filter paper matrix to improve the reproducibility in sensor production. The electrodes were prepared from one-side silver-coated filter paper, which is encapsulated for insulation by a heat-sealing film. A defined volume of the NH+4-sensitive polymer matrix membrane cocktail was deposited on this filter paper. To obtain the urea-biosensor a layer of urease was cast onto the ionsensitive membrane. Poly (carbamoylsulfonate) hydrogel, produced from a hydrophilic polyurethane prepolymer blocked with bisulfite, served as immobilisation material. The disposable urea-sensitive electrode was combined with a disposable Ag/AgCl reference electrode to obtain the disposable urea biosensor. The sensor responded rapidly and in a stable manner to changes in urea concentrations between 7.2 ⫻ 10−5 and 2.1 ⫻ 10−2 mol/l. The detection limit was 2 ⫻ 10−5 mol/l urea and the slope in the linear range 52 mV/decade. By taking into consideration the influence of the interfering K+- and Na+- ions the sensor can be used for the determination of urea in human blood and serum samples (diluted or undiluted). A good correlation was found with the data obtained by the spectrophotometric routine method. 1999 Elsevier Science S.A. All rights reserved. Keywords: Disposable biosensor; Disposable reference electrode; Potentiometric biosensor; Urea biosensor; Blood urea
1. Introduction The basic prerequisite for the availability of disposable potentiometric biosensors is a reproducible low cost production technology for ion-sensitive electrodes. To meet this requirement, coated-film electrodes were developed by the Eastman Kodak Company (Battaglia et al., 1980). However, controlled manufacturing of such electrodes requires sophisticated, elaborate and rather expensive processes. Our solution to this problem is the so-called double matrix membrane (DMM) technology in which an ion-selective polymeric membrane is formed in the volume of an electrochemical inert micro fiber matrix (simple filter paper) by a casting technique
* Corresponding author. Tel.: 00-49-141-83-33100; fax: 00-49-25183-32132; e-mail: [email protected]
(Knoll, 1991; Dumschat et al., 1994). There the polymer matrix homogeneously distributes by capillary forces and, after the evaporation of the solvent, the double matrix membranes are obtained which have better mechanical stability than traditional polyvinylchloride (PVC) membranes. As described for K+-sensitive membranes (Knoll, 1991) and for pH-sensitive membranes (Dumschat et al., 1994) the additional filter paper has no influence on the electrochemical response behaviour of the membranes. The DMM technology was used to develop disposable Na+- and K+-sensitive electrodes (Borchardt et al., 1994, 1995). An important parameter in clinical analysis is urea where an increased concentration in blood and a reduced level in urine is a strong indication for renal dysfunction. The determination of urea in body fluids is one of the most frequent analyses in routine clinical laboratories. The normal range in human serum is between 1.7 and 8.3 mmol/l (Rick, 1990) and levels increase up to
0956-5663/98/$ - see front matter 1999 Elsevier Science S.A. All rights reserved. PII: S 0 9 5 6 - 5 6 6 3 ( 9 8 ) 0 0 1 0 3 - 1
34
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
100 mmol/l under pathophysiological conditions (Rick, 1990). Urea is hydrolysed by urease according to the reaction (NH2)2CO ⫹ 2 H2O ⫹ H+→2 NH+4 ⫹ HCO−3. In conventional urea sensors, pH electrodes (Ruzicka et al., 1979;Vadgama et al., 1982; Miyahara et al., 1985; Karube et al., 1986; Hamann et al., 1986) and NH+4selective electrodes (Guilbault and Montalvo, 1969, 1970; Campanella et al., 1990) have been used to detect the respective hydrogen ions or ammonium ions produced in the enzymatic reaction. The major problem for pH-sensitive electrodes is that the sensor response is strongly dependent on the buffer capacity of the sample because the pH change produced in the course of the enzyme-catalysed reaction is suppressed by the buffer used, which leads to a narrow dynamic range and a loss in sensor sensitivity (Ruzicka et al., 1979; Eddowes, 1987; Koncki et al., 1992). The general problem of NH+4-sensitive electrodes is interference by Na+- and K+ions, which are present in serum (4.5 mmol/l K+ and 140 mmol/l Na+) and urine. In the present study the development of a low cost disposable urea-sensitive biosensor in the form of a test strip is described. The sensor consists of a disposable NH+4-sensitive electrode in DMM technology as transducer, which is covered with a layer of immobilized urease, and a disposable Ag/AgCl reference electrode. The fundamental characteristics of the sensor configuration are presented. The influence of ionic interferences on the response behaviour of the urea electrode is shown and furthermore results of urea measurements in blood and serum samples are presented.
2. Experimental section 2.1. Reagents Tetrahydrofuran (THF) was purchased from Aldrich (99.5 ⫹ %, spectrophotometric grade, inhibitor-free), cyclohexanone (extra pure) from Merck. PVC of high molecular weight and all other chemicals used for the preparation of the NH+4-sensitive membranes (selectophore quality) were obtained from Fluka and applied as received. Urease (260 U/mg) was purchased from Serva. Tris (hydroxymethyl)-aminomethane (Tris) and ethylenediaminetetraacetic acid (disodium salt) (EDTA) were obtained from Sigma (analytical grade). KCl, NaCl, NH4Cl and urea were purchased from Merck. Urease (5 U/mg) used for the spectrophotometric Berthelot method was from Merck. Phenol and sodium hypochloride were purchased from Sigma and sodium nitroprusside dihydrate was obtained from Fluka. As micro fiber matrix (MFM) material filter paper from (Schleicher & Schuell (5893 Blue ribbon ashless, Ref. No. 300211) was used. To support and insulate the filter
paper a 150-m thick polyester/polyethylene heat-sealing film (Team Codor, DUBO-Schweitzer GmbH) was used. The poly(carbamoylsulfonate) (PCS)-prepolymer used for the entrapment of urease is a research product and was a kind gift by Professor Vorlop (Forschungsanstalt fu¨r Landwirtschaft, Braunschweig). Polyethyleneimine (PEI) was obtained from Sigma. 2.2. Solutions All calibration and buffer solutions were prepared with double-distilled deionized water. The sample buffer (pH 7.5) containing 20 mmol/l Tris and 1.0 mmol/l EDTA was adjusted with HCl. Stock solutions of urea, 1.0 and 0.1 mmol/l, and those of NH+4, 1.0 and 0.1 mmol/l, were prepared daily in Tris buffer. The sample buffer solution containing 140 mmol/l NaCl and 4.5 mmol/l KCl was also prepared in Tris buffer. From the latter solutions containing other concentrations of NaCl and KCl were obtained by serial dilution with Tris buffer. 2.3. Sensor preparation Fig. 1 shows the final configuration of the disposable urea biosensor consisting of the urea-sensitive electrode and the reference electrode. The urea electrode is based on a disposable NH+4-sensitive DMM electrode as transducer. The transducer electrodes were produced in a batch process as described in detail by Borchardt et al. (1994) and Eggenstein et al. (1995). To prepare the silver contact and the conducting track, silver was evaporated on one side of the filter paper (vacuum coater, Univex 300, Leybold). For insulation the filter paper was encapsulated with a preperforated heat-sealing film. By this procedure the filter paper was completely covered with heat sealing film except for a hole of 3 mm diameter on one side for membrane deposition and a 5-mm long contact pad for the electrical contact on the other side. The following mixture was used for the preparation of the NH+4-sensitive DMM (wt%): 1.0% Nonactine, 66.8% bis (2-ethylsebacate) (DOS), 32.2% PVC. 400 mg of the membrane components were dissolved in 1.5 ml THF and 0.5 ml cyclohexanone. The components were allowed to dissolve overnight and shaken vigorously until a homogeneous solution was obtained. The deposition of the PVC-membrane onto the filter paper was carried out automatically with an Asymtek Automove 402 fluid dispenser to apply defined volumes into each hole. To prevent an overflow of substances three dispensing steps of 10 l for each hole were done at intervals of 5 min. After solvent evaporation overnight the NH+4sensitive electrodes were ready on the one hand for NH+4 measurements and on the other hand for the deposition of the enzyme layer on the surface of the ionsensitive PVC-membrane. The enzyme membrane was
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
Fig. 1.
35
Schematic view of the disposable urea biosensor consisting of disposable measuring and reference electrode.
prepared following the procedure of Vorlop et al. (1992) and as modified by Kotte et al. (1995). A defined amount of the aqueous PCS prepolymer solution (polymer content: 38% w/w) was diluted with water to obtain a 30% (w/w) solution. Polyethyleneimine (PEI) (2.5% w/v in water) was added dropwise to adjust the pH between 6 and 6.5. A 50-l aliquot of the hydrogel matrix was mixed with 50 l of enzyme solution (1 mg urease (260 U) in 50 l of water). A volume of 3 l of this aqueous hydrogel-enzyme suspension was deposited with a micro syringe onto the surface of the ion-sensitive membrane of the electrodes and allowed to form a gel at 4°C in a horizontal position. After 2 to 3 h, the enzyme electrodes were immersed in a dilute aqueous PEI solution (0.5% w/v) for 30 s and rinsed with water. PEI was used as coupling reagent to prevent the enzyme from bleeding out. The electrodes were stored under dry conditions at 4°C. The disposable Ag/AgCl reference electrodes were prepared in the same way as described by Diekmann et al. (1995) for a redox reference electrode, based on the same filter paper described above. An Ag/AgCl paste from Acheson (AV458) was used to prepare the con-
ducting track and 3 mmol/l KCl provided the internal electrolyte solution. For storage, the electrodes were immersed in 3 mmol/l KCl to avoid drying up. To obtain the final disposable urea biosensor the ureasensitive electrode and the reference electrode were glued onto a separating and insulating plastic film right before the measurements. This combination could be stored at 4°C in Tris buffer (pH 7.5) containing 1 mmol/l EDTA. 2.4. Measurements The performance of the electrodes was evaluated by their limits of detection, slopes, selectivities and by the reproducibility of the sensors produced. The calibration curves for identifying the detection limit, slope and linear range were obtained by adding known amounts of the relevant substrate and respective ion to the measuring solution from low to high concentrations. The selectivity coefficients were evaluated by the fixed interference method according to the IUPAC (1976) recommendations. Measurements were carried out either with a conventional Ag/AgCl-reference electrode from Orion
36
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
(Model 90-02) with double junction (inner filling: 3 mmol/l KCl; outer filling: sample buffer, pH 7.5) or with the disposable reference electrode developed here. For the determination of reproducibility of the sensors, eight NH+4-sensitive and five urea-sensitive electrodes from one batch, respectively, were compared with regard to slope and potential at equal concentrations, in this case the conventional Ag/AgCl electrode served as reference electrode. All measurements were made at room temperature with the aid of a pH-meter (microprocessor pH-meter pH 3000/multiplex 3000, WTW Weilheim, Germany) to which the sensors were plugged in. Before measurements the NH+4- and the urea-sensitive electrodes were equilibrated in the sample buffer until a stable potential was obtained (0.5–1 h) and then immersed with the reference electrode in 10 or 40 ml stirred sample buffer (pH 7.5) and the calibration graphs were recorded as described above. To gauge the influence of interfering ions on the response behaviour of the urea biosensor the electrodes were dipped in 10 ml of the sample buffer containing various combinations of Na+ and K+ concentrations and the calibration was attained by stepwise addition of aliquots of standard urea solutions (1.0 and 0.1 mmol/l) to the stirred test solution.
3. Results 3.1. Characterisation of the measuring electrode At first NH+4-sensitive disposable electrodes in DMM technology were characterised to evaluate whether they could serve as transducer in sensing urea. The electrochemical response behaviour of the NH+4-sensitive electrode versus a commercial Ag/AgCl-reference electrode and the corresponding calibration graph are shown in Fig. 2. The electrode responded rapidly and in a stable manner to changes in the NH+4-concentration with a response time in the range of a few seconds. As determined from the calibration graph the detection limit was 5.0 ⫻ 10−6 mol/l, the lowest limit of linear range 3.2 ⫻ 10−5 mol/l and the slope in the linear range amounted to 55 mV/decade. The calibration curves of eight NH+4sensitive electrodes from one lot revealed an average slope in the linear range of 54.8 ⫾ 0.3 mV/decade. It can be seen from Fig. 3 that the slopes were reproducible indeed. Taking these data we concluded that the developed NH+4-sensitive electrode developed was a suit-
2.5. Procedure for the determination of urea in serum and blood
Blood and serum samples were collected from healthy persons as well as serum samples from haemodialysis patients from a renal laboratory (Centre of Home Dialysis and Kidney Transplant, Mu¨nster). Sodium-citrate was added to blood samples to prevent clotting. Measurements were carried out with the final disposal urea biosensor as developed here. Blood and serum samples were measured undiluted or diluted 1:1 or 1:5 with Tris/HCl buffer. The urea-sensitive biosensor was immersed in the samples (1 ml sample volume) and the potential was recorded after 2–3 min when a constant signal was reached. Between trials the electrode was rinsed with double-distilled water and immersed in the sample buffer. Testing of the blood and serum samples was started again only after the electrode potential returned to its original (urea-free) value. The amount of urea was determined from standard calibration curves recorded in the sample buffer with constant interfering Na+- and K+-ion levels corresponding to the average levels in blood samples (diluted or undiluted). The Berthelot reaction, a standard method for spectrophotometric urea determination, was used as a reference (Fawcett and Scott, 1960; Bernt and Bergmeyer, 1970); the absorbance of the indophenol produced was measured at 550 nm.
Fig. 2. Response curve (A) and calibration graph (B) of the disposable NH+4-sensitive electrode in combination with the conventional Ag/AgCl-reference electrode in 0.02 mmol/l Tris/HCl buffer, pH 7.5 (in the calibration graph the potential in buffer is set to zero).
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
37
Fig. 3. Reproducibility of the batch production as evidenced by the calibration curves of eight disposable NH+4-sensitive electrodes (the potential in buffer is set to zero).
able transducer basis for the potentiometric urea sensor. For the application of the sensor to in real samples, however, it was necessary to consider that the ionophore used was sensitive not only to NH4+. To gauge the crosswere reactivities the selectivity coefficients log K pot ij determined to be Ca2 ⫹ ⫽ ⫺ 4.20, Li+ ⫽ ⫺ 3.60, Na+ ⫽ ⫺ 2.95, K+ ⫽ ⫺ 0.80. The next step was the deposition of the enzyme membrane onto the NH+4-sensitive DMM. Different polymer materials like PVC, polyvinylpyrrolidone, polyestersulfonic acid and photo-crosslinkable prepolymers were tested as gel matrix for entrapping the urease. Best results concerning adhesion capability, limit of detection, slope, linear range and lifetime were obtained with PCS cross-linked with PEI as gel material. This material not only adhered excellently to the PVC-membrane, but showed minimal diffusion resistance for various substrates, as already described by Kotte et al. (1995). The electrochemical response behaviour of such a urea-sensitive electrode versus a commercial Ag/AgCl electrode is shown in Fig. 4. As demonstrated, the electrode responded rapidly and in a stable manner to changes in the urea concentration with a detection limit as low as 1.0 ⫻ 10−5 mol/l. Fig. 5 shows the calibration curves of five urea-sensitive electrodes from one lot; the average slope in the linear range was 55.5 ⫾ 0.7 mV/decade again indicating high reproducibility. With the aim to obtain the complete disposable biosensor the urea electrode required a compatible disposable reference electrode to replace the conventional Ag/AgCl-electrode in the sensor configuration. A simple reference electrode was developed for this purpose and glued to the urea electrode, as described in the Experimental section. Fig. 1 depicts the resulting test strip for the disposable urea biosensor. Then the performance of the urea biosensor was evaluated; the respective response curve and the calibration graph derived are presented in Fig. 6. The limit of detection was 2.0 ⫻ 10−5 mol/l (in 0.02 mmol/l Tris buffer without interfering
Fig. 4. Response curve (A) and calibration graph (B) of the disposable urea-sensitive electrode in combination with the conventional Ag/AgCl-reference electrode in 0.02 mmol/l Tris/HCl buffer, pH 7.5 (in the calibration graph the potential in buffer is set to zero).
Fig. 5. Reproducibility of the batch production process as evidenced by the calibration curves of five disposable urea-sensitive electrodes (the potential in buffer is set to zero).
ions) and the slope in the linear range was about 52 mV/decade. The sensor responded within 1 and 2 min in a stable manner to changes in the analyte concentration. The urea-sensitive test strip could be stored over-
38
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
respectively, with intermittant immersion in sample buffer during nights. During these 4 days a slight decrease in electrode activity and selectivity was observed; thereafter no activity remained. Enzyme membranes stored dry at 4°C showed no deterioration in sensor performance during the 4 days and up to 3– 4 months a slight activity loss only (data not shown). The data obtained with the final configuration of the urea biosensor were compared to those obtained with the urea-sensitive electrode measured against the conventional Ag/AgCl-reference electrode (Table 1). The results revealed that the electrochemical response characteristics of both sensor configurations were comparable, only a slight decline in sensitivity and detection limit was observed for the former. 3.2. Application of the biosensors in biological samples
Fig. 6. Response curve (A) and calibration graph (B) of the disposable urea biosensor, i.e., the combination of the disposable urea-sensitive electrode and the disposable Ag/AgCl-reference electrode, in 0.02 mmol/l Tris/HCl buffer, pH 7.5 (in the calibration graph the potential in buffer is set to zero).
night in a refrigerator continuously immersed in Tris buffer pH 7.5 with 1 mmol/l EDTA, and then used again each day for several measurements. Under these conditions the life time of the disposable test strip was 4 days. Fig. 7 shows calibration graphs from measurements with the one urea-sensitive sensor test strip, first immediately after the glueing process and immersion into the sample buffer, and then after 1, 2 and 3 days,
Fig. 7. Lifetime of the disposable urea biosensor (continuous immersion in 0.02 mmol/l Tris/HCl buffer, pH 7.5).
For the determination of urea in blood it was necessary to take into account interferences by Na+- and K+ions abundantly present in such samples. To investigate the effect of increasing concentrations of these ions on the electrode response performance and to evaluate whether the biosensor developed is applicable in blood samples, response curves were recorded in background buffer solutions at different levels of the interfering ions. The calibration graphs obtained for the disposable biosensor are shown in Fig. 8 and the data derived therefrom are summarised in Table 2. Concentrations up to a level of 140 mmol/l for Na+ and 4.5 mmol/l for K+ions, corresponding to the normal level of these ions in human blood and serum, were tested. Without Na+- and K+-ions the slope in the linear range between 7.2 ⫻ 10−5 and 2.1 ⫻ 10−2 mol/l was 52.0 mV/decade. As the concentration of these ions increased, the lower detection limit of urea moved towards higher concentrations and the dynamic range of the biosensor was narrowed and, as a result, the apparent slope of the response of the sensor become smaller. The limit of detection increased from 2.0 ⫻ 10−5 up to 1.2 ⫻ 10−4 mol/l and the slope of the calibration curve decreased from 52.0 down to 31.7 mV/decade. This effect could be expected from the determination of the selectivity coefficients of the NH+4sensitive DMM for the interfering Na+- and K+-ions as descibed before. With a biosensor reference value for a linear range from 0.5 to 10 mmol/l urea (Fig. 8 and Table 2), which covers the reference interval of 2–8 mmol/l for urea in blood of a healthy person (Rick, 1990), the applicability of the urea biosensor was then checked in undiluted and diluted blood samples. For the measurement of undiluted samples (up to 10 mmol/l urea) the biosensor was calibrated in background buffer containing 140 mmol/l Na+ and 4.5 mmol/l K+, for measurements of samples diluted 1:1 (with buffer free of interfering ions) the test strip
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
39
Table 1 Electrochemical response characteristics of the disposable urea-sensitive electrode in combination with a conventional and the disposable reference electrode, respectively Urea-sensitive electrode versus
Slope (mV/decade)
Lowest limit of linear range (mol/l)
Limit of detection (mol/l)
Conventional reference electrode Disposable reference electrode
55.5 52.0
4.5 ⫻ 10−5 7.2 ⫻ 10−5
1.0 ⫻ 10−5 2.0 ⫻ 10−5
Fig. 8. Influence of increasing Na+- and K+-concentrations on the response of the urea biosensor in 0.02 mmol/l Tris/HCl buffer, pH 7.5. Table 2 Response characteristics of the urea biosensor in dependence on the concentration of interfering Na+- and K+- ions Concentration of interfering ions (mmol/l) Na+
K+
– 14 28 70 140
– 0.5 1.0 2.5 5.0
Slope
Lowest limit Limit of of linear range detection (mV/decade) (mol/l) (mol/l)
52.0 46.6 40.5 35.2 31.7
7.2 1.5 1.0 5.0 5.0
⫻ ⫻ ⫻ ⫻ ⫻
10−5 10−4 10−4 10−4 10−4
2.0 5.0 5.0 1.5 1.2
⫻ ⫻ ⫻ ⫻ ⫻
Fig. 9. Blood urea determined with the urea biosensor with reference to the spectrophotometric Berthelot method (samples from healthy persons).
10−5 10−5 10−5 10−4 10−4
was calibrated in background buffer containing 70 mmol/l Na+ and 2.25 mmol/l K+, and so on. Whereas the slope of the calibration curve remained constant during the course of a day, the entire standard curve shifted several millivolts. Therefore, the urea biosensor was calibrated after every 8 to 10 measurements. The results obtained with the urea electrode (y) were compared with values (x) obtained with the standard spectrophotometric Berthelot method in Figs. 9 and 10. The potentiometric values were mean values of double determinations with a standard deviation of 4%. The spectrophotometric values were also mean values of double determinations with a standard deviation of 5%. Fig. 9 shows the results of blood measurements of 13 healthy persons. The difference between the results obtained with the urea electrode
Fig. 10. Serum urea determined with the urea biosensor with reference to the spectrophotometric Berthelot method (samples from patients; inset: samples from healthy persons).
40
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
and the standard method was about 0.5–7.0%, the correlation plot was y ⫽ 1.086x ⫺ 0.341 (r ⫽ 0.991). Fig. 10 represents the results of serum measurements. Of the 18 samples compared, 7 were from healthy persons (inset of Fig. 10) and 11 from a laboratory attached to a renal unit. The difference between the results obtained with the potentiometric and the spectrophotometric method in this case was about 0.8–6.5% for the samples of healthy persons and 1.5–8.0% for the samples from the renal laboratory except 2 outliners (11%). The overall correlation plot was y ⫽ 0.927x ⫹ 0.824 (r ⫽ 0.994), with y ⫽ 1.154x ⫺ 0.666 (r ⫽ 0.997) for samples from the routine laboratory and y ⫽ 0.904x ⫹ 0.828 (r ⫽ 0.993) for samples from the renal laboratory.
4. Discussion The data reported here demonstrated that the DMM technology enabled the production of disposable ureasensitive electrodes, and in combination with the disposable reference electrode the production of a disposable urea biosensor with a low cost potential. The evolution of this DMM concept can be described by the following: at first the conventional ion-sensitive PVC matrix membrane was supplemented with the micro fiber matrix. For the resulting DMM no significant differences in detection limit, slope and selectivity were observed, as the additional filter paper did not influence the electrochemical responses of the cation-sensitive membranes; yet the mechanical stability was considerably improved (Knoll, 1991; Dumschat et al., 1994). Moreover, the development of the ion-sensitive electrode from an electrode with liquid internal junction to an electrode with solid contact in the form of a test strip had no detrimental effect on the response behaviour as already described for the Na+- and K+-sensitive electrodes (Borchardt et al., 1994, 1995) and demonstrated here for the NH+4-sensitive electrode. In comparison to the performance of previously published ion-selective electrodes based on a single PVC matrix, the present NH+4-sensitive DMM revealed a similar Nernstian behaviour, even to those based on a gaseous diffusion electrode (Campanella et al., 1990; Miyahara et al., 1991; Schindler et al., 1994). The main advantage of the new biosensor configuration is the technology of electrode production. Only cheap mass products were used and high productivity can be achieved by simple batch technology. Subsequent transformation of the NH+4-sensitive transducer to the ureasensitive electrode created a biosensor with an excellent response behaviour. The biosensor works in an analytical useful range, with a short response time. The slope of about 55.5 mV/decade in the linear range affords a sensitivity which is expected for NH+4-sensitive potentiometric transducers (Guilbault and Nagy, 1973; Miyahara et al., 1991).
The final goal was attained by integrating the Ag/AgCl-reference electrode into the disposable urea biosensor. The materials needed for the reference electrode, produced in a batch process, were cheap mass products as well. Investigations on the reproducibility of manufacturing and on the glueing process are under way. The slight decline seen today in the detection limit and the linear range compared to the combination of the urea-sensitive electrode with a conventional reference electrode may be caused by the contamination of the test solution with K+-ions from the internal electrolyte diffusing out of the disposable reference electrode. The good storage stability up to 4 months and the good operational stability over 4 days already are excellent criteria for a disposable biosensor. The measurements in background buffer with increasing concentrations of Na+- and K+-ions were a mandatory prelude before application of the biosensor in clinically relevant samples. From the selectivity coefficients of the NH+4-sensitive membrane for Na+- and K+-ions it was reasonable to expect that the slope and the dynamic range of the urea biosensor in the presence of the latter ions would be poorer than those measured in the absence of interfering ions. Urea could be measured directly in blood samples, if the electrode was calibrated in solutions with constant levels of interfering Na+- and K+-ions, as borne out by the close agreement with the spectrophotometric reference data. Yet for some of the 11 samples from the renal laboratory differences up to 11% were observed between the potentiometric and the spectrophotometric method. In these cases calibrations might have been carried out under a wrong assumption, as patients with renal dysfunction often not only suffer under increased urea levels in blood, but also from a disturbance in their electrolyte balance (Girndt, 1990; Pindur and Pindur, 1991). Consequently the method developed up to this stage furnishes data with sufficient accuracy for routine samples. Urea determination of samples coming out from a renal laboratory thus must be complemented by determinations of Na+- and K+-ions. This will become possible with a sensor configuration in DMM technology measuring Na+-, K+-ions and urea simultaneously or consecutively. This technology enabled single analyte determination in each case as shown earlier (Borchardt et al., 1994, 1995) and in the present investigation. An alternative possibility for urea determination, also based on DMM technology, could employ the pH-based system (Eggenstein et al., 1995).
Acknowledgements We are grateful to the Ministry of Science and Research of Northrhine-Westphalia (Germany) for financial support.
C. Eggenstein et al. / Biosensors & Bioelectronics 14 (1999) 33–41
References Battaglia, C.J., Chang, J.C., Daniel, D.S., 1980. US Patent No. 4 214 968. Bernt, E., Bergmeyer, U., 1970. Harnstoff. In: Bergmeyer, U. (Ed.), Methoden der enzymatischen Analyse, vol. 1. Verlag Chemie GmbH, Weinheim, pp. 1738–1748. Borchardt, M., Diekmann, C., Dumschat, C., Cammann, K., Knoll, M., 1994. Disposable sodium electrodes. Talanta 41, 1025–1028. Borchardt, M., Dumschat, C., Cammann, K., Knoll, M., 1995. Disposable ion selective electrodes. Sensors and Actuators B 24/25, 721–723. Campanella, L., Mazzei, F., Sammartino, M.P., Tomassetti, M., 1990. New enzyme sensors for urea and creatinine analysis. Bioelectrochem. Bioenergetics 23, 195–202. Diekmann, C., Dumschat, C., Cammann, K., Knoll, M., 1995. Disposable reference electrode. Sensors and Actuators B 24/25, 276–278. Dumschat, C., Borchardt, M., Diekmann, C., Hepke, J., Knoll, M., 1994. Double matrix membranes for potentiometric cation sensitive sensors. Fresenius J. Anal. Chem. 348, 553–555. Eddowes, M.J., 1987. Response of an enzyme-modified pH-sensitive ion-selective device: analytical solution for the response in the presence of pH buffer. Sensors and Actuators 11, 265–274. Eggenstein, C., Borchardt, M., Dumschat, C., Gru¨ndig, B., Cammann, K., Spener, F., Knoll, M., 1995. Potentiometric biosensor in double matrix membrane technology. Biosensors and Bioelectronics 10, 595–600. Fawcett, F.K., Scott, J.E., 1960. A rapid and precise method for the determination of urea. J. Clin. Pathol. 13, 156–160. Girndt, J., 1990. Nieren- und Hochdruckkrankheiten. Schattauer Verlags GmbH, Stuttgart, New York, pp. 57–236. Guilbault, G.G., Montalvo, J.G., 1969. Urea-specific enzyme electrode. J. Am. Chem. Soc. 91, 2164–2165. Guilbault, G.G., Montalvo, J.G., 1970. An enzyme electrode for the substrate urea. J. Am. Chem. Soc. 92, 2533–2538. Guilbault, G.G., Nagy, G., 1973. Improved urea electrode. Anal. Chem. 45, 417–419. Hamann, H., Ku¨hn, M., Bo¨ttcher, N., Scheller, F., 1986. Enzyme
41
sensor system for potentiometric urea determination in serum and milk. J. Electroanal. Chem. 209, 69–76. 1976. Recommendations for nomenclature of ion-selective electrodes. Pure Appl. Chem. 48, 129–132. Karube, I., Tamiya, E., Dicks, J.M., 1986. A microsensor for urea based on an ion-selective field effect transistor. Anal. Chim. Acta 185, 195–200. Knoll, M., 1991. German Patent No. P41 37 261.1. Koncki, R., Leszcynski, P., Hulanicki, A., Glab, S., 1992. Urea sensor based on glass pH electrodes with physically immobilized urease. Anal. Chim. Acta 257, 67–72. Kotte, H., Gru¨ndig, B., Vorlop, K.D., Strehlitz, B., Stottmeister, U., 1995. Methylphenazonium-modified enzyme sensor based on polymer thick films for subnanomolar detection of phenols. Anal. Chem. 67, 65–70. Miyahara, Y., Moriizumi, T., Ichimura, K., 1985. Integrated enzyme FETs for simultaneous determination of urea and glucose. Sensors and Actuators 7, 1–10. Miyahara, Y., Tsukada, K., Miyagi, H., 1991. Urea sensor based on an ammonium-ion-sensitive field-effect transistor. Sensors and Actuators B 3, 287–293. Pindur, G., Pindur, U., 1991. Klinische Chemie und serologische Laboratoriumsdiagnostik. Wissenschaftliche Verlags GmbH, Stuttgart, pp. 244–309 Rick, W., 1990. Klinische Chemie und Mikroskopie. Springer-Verlag, Berlin, pp. 245–247. Ruzicka, J., Hansen, E.H., Ghose, A.K., Mottola, H.A., 1979. Enzymatic determination of urea in serum based on pH measurement with the flow injection method. Anal. Chem. 51, 199–203. Schindler, J.G., Schindler, M.M., Herna, K., Reisinger, E., Burk, B., Kuhlmann, U., Knaack, J., Schmidt, B., Langer, H., 1994. Harnstoff-Durchflußsensor zur ammoniumselektiv-enzymatischen Prozeßkontrolle der ku¨nstlichen Niere. Eur. J. Clin. Chem. Clin. Biochem. 32, 145–152. Vadgama, P.M., Alberti, K.G.M.M., Covington, A.K., 1982. Determination of urea in blood plasma by enzyme-pH electrode. Anal. Chim. Acta 136, 403–406. Vorlop, K.D., Muscat, A., Beyersdorf, J., 1992. Entrapment of microbial cells within polyurethane hydrogel beads with the advantage of low toxicity. J. Biotechnol. Technol. 6, 483–488.