Comparative study of controlled pore glass, silica gel and Poraver® for the immobilization of urease to determine urea in a flow injection conductimetric biosensor system

Biosensors and Bioelectronics 19 (2004) 813–821

Comparative study of controlled pore glass, silica gel and Poraver® for the immobilization of urease to determine urea in a flow injection conductimetric biosensor system Warakorn Limbut a,b , Panote Thavarungkul a,c,∗ , Proespichaya Kanatharana a,b , Punnee Asawatreratanakul a,d , Chusak Limsakul a,e , Booncharoen Wongkittisuksa a,e a

e

Biophysics Research Unit of Biosensors and Biocurrents, Prince of Songkla University, Hat Yai, Songkhla 90112, Thailand b Department of Chemistry, Faculty of Science, Prince of Songkla University, Hat Yai, Songkhla 90112, Thailand c Department of Physics, Faculty of Science, Prince of Songkla University, Hat Yai, Songkhla 90112, Thailand d Department of Biochemistry, Faculty of Science, Prince of Songkla University, Hat Yai, Songkhla 90112, Thailand Department of Electrical Engineering, Faculty of Engineering, Prince of Songkla University, Hat Yai, Songkhla 90112, Thailand Received 10 February 2002; received in revised form 17 August 2003; accepted 19 August 2003

Abstract This study compared the responses of three enzyme reactors containing urease immobilized on three types of solid support, controlled pore glass (CPG), silica gel and Poraver® . The evaluation of each enzyme reactor column was done in a flow injection conductimetric system. When urea in the sample solution passed though the enzyme reactor, urease catalysed the hydrolysis of urea into charged products. A lab-built conductivity meter was used to measure the increase in conductivity of the solution. The responses of the enzyme reactor column with urease immobilized on CPG and silica gel were similar and were much higher than that of Poraver® . Both CPG and silica gel reactor columns gave the same limit of detection, 0.5 mM, and the response was still linear up to 150 mM. The analysis time was 4–5 min per sample. The enzyme reactor column with urease immobilized on CPG gave a slightly better sensitivity, 4% higher than the reactor with silica gel. The life time of the immobilized urease on CPG and silica gel were more than 310 h operation time (used intermittently over 7 months). Good agreement was obtained when urea concentrations of human serum samples determined by the flow injection conductimetric biosensor system was compared to the conventional methods (Fearon and Berthelot reactions). These were statistically shown using the regression line and Wilcoxon signed rank tests. The results showed that the reactor with urease immobilized on silica gel had the same efficiency as the reactor with urease immobilized on CPG. © 2003 Elsevier B.V. All rights reserved. Keywords: Biosensor; Conductivity; Enzyme reactor; Immobilized urease

1. Introduction Enzyme-based biosensors for urea use various techniques to apply the enzyme urease to transducers. In the enzyme electrode system, a thin layer of immobilized enzyme is coupled to an electrochemical sensor. Thus, it is possible to determine urea using a variety of transducers, such as a carbon dioxide electrode (Guilbault and Shu, 1972), an ammonia electrode (Guilault and Tarp, 1974; Mascini and Guilbault, 1977), a pH electrode (Ruzicka et al., 1979;

∗

Corresponding author. Tel.: +66-74-288753: fax: +66-74-212817. E-mail address: [email protected] (P. Thavarungkul).

0956-5663/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2003.08.007

Vadgama, 1986; Walcerz et al., 1995, 1998) and an ammonium electrode (Guilbault and Montalvo, 1969, 1970; Guilbault and Nagy, 1973; Yasuda et al., 1984). However, the stability of an enzyme electrode is difficult to define because enzymes can lose activity, resulting in a shift of the calibration curve downward (Guilbault, 1984). To overcome the above disadvantage the enzyme urease has been covalently immobilized on controlled-pore glass (CPG), packed into a small column and placed prior to the transducer. This system offers the advantages of large area for reaction with little mass transport limitation and giving a fast response (Chandler et al., 1989, 1990). This technique has been applied successfully in several systems (Adams and Carr, 1978; Bowers et al., 1976; Chandler et al., 1989,

814

W. Limbut et al. / Biosensors and Bioelectronics 19 (2004) 813–821

1990; Gorton and Ogren, 1981; Jurkiewicz et al., 1998a,b; Thavarungkul et al., 1991, 1999; Thavarungkul and Kanatharana, 1994; Xie et al., 1994). Since CPG is rather expensive, cheaper support material, that is, silica gel and Poraver® will be studied. Silica gel and Poraver® were chosen because they are insoluble, have large surface areas due to their porous nature, and their surfaces (the SiO2 groups) can be modified to covalently coupled to the enzymes. They are also convenient to handle, have compressive strength, and can be stored easily. If silica gel and Poraver® are found to have similar efficiency as CPG the enzyme reactor biosensor system can be applied more widely due to its lower cost, especially in a disposable or an enzyme inhibition system. This paper reports the comparative study of the efficiency of immobilized urease on three different support materials for the determination of urea. The evaluation of each enzyme reactor columns was done using a flow injection conductimetric biosensor system. The change in conductivity of the solution was generated by the hydrolysis reaction of urea. System using different reactor columns were compared while working under optimal conditions. The comparison was done by observing several analytical parameters, such as sensitivity, linear range, limit of detection and dynamic characteristic of the response. The system performance of enzyme reactor columns has also been tested for the determination of urea level in human serum compared with Fearon reaction and Berthelot reaction.

2. Materials and methods 2.1. Materials Urease (amidohydrolase EC 3.5.1.5 Type IX: from jack beans, 62,100 units/g solid) and reagents for the colorimetric determination of urea (blood urea nitrogen, BUN, No. 535 test kit Fearon reaction) were obtained from Sigma (St. Louis, Missouri, USA). Silica gel (diameter 40–63 ␮m, mean pore diameter 60 Å, specific surface area 500 m2 /g) was obtained from Fluka. Controlled pore glass (mean diameter 41 ␮m, mean pore diameter 200 Å, specific surface area 300 m2 /g) was from EKA Nobel AB (Surte, Sweden). Poraver® (diameter 1–2 mm crushed to 390–630 ␮m, mean pore diameter, measured from stereomicroscope, was 7–45 ␮m) from Dennert Poraver GmbH (Germany) was a gift from Prof. Bo Mattiasson, Department of Biotechnology, Center of Chemistry and Chemical Engineering, University of Lund, Sweden. All other chemicals used were of analytical grade.

step is to prepare the support materials by placing them in an oven at 110 ◦ C for 3 h to get rid of any organic adsorbed to the support. It was then heated in 5% (w/v) nitric acid solution at 75 ◦ C for 45 min to remove any metal residues. The support was rinsed thoroughly with distilled water and dried at 90 ◦ C overnight. The next step is the derivatization of the support surface with organosilane. This was done by adding 1 g of clean support material into 18 ml of distilled water and 2 ml of 10% (v/v) 3-aminopropy-triethoxysilane. The pH of the mixture is adjusted to be between 3 and 4 with 6 M HCl. It was then heated at 75 ◦ C in a water bath for 2 h. The support was filtered on a Buchner funnel, and washed with 20 ml of distilled water and dried at 90 ◦ C overnight. This produces an alkylamine derivative on the surface of the support material (Fig. 1A). The final step is the coupling of enzyme to the support. The alkylamine support materials was first activated by glutaraldehyde to yield aldehyde support materials. This was done by adding 2.0 ml (sedimented volume) of alkylamine support materials into 25 ml of 5.0% (v/v) glutaraldehyde in 0.05 M sodium phosphate buffer pH 7.0 (Fig. 1B). The mixture was tumbled end over end for 60–90 min, during this time, the colour of the carrier changed to orange-red. It was washed with 500 ml of distilled water and then with buffer repeatedly on Buchner funnel, until it has no odour of glutaraldehyde.To immobilize the enzyme, 10 mg (620 units) of urease was dissolved in 5 ml of 0.05 M sodium phosphate buffer pH 7.0 and added to 2.0 ml (sedimented volume) of activated support materials (Fig. 1C). The mixture was tumbled at room temperature (around 23 ◦ C). After 4–5 h, 50 mg of sodium cyanoborohydride was added to reduce the Schiff’s bond between aldehyde and enzyme, thus stabilizing the coupling. The mixture was tumbled again for another 15 h and was then washed with 500 ml of buffer. After this 25 ml of 0.1 M ethanolamine pH 8.0 was added and the reaction was allowed for another 2 h. This step was to occupy all the aldehyde group which did not couple to the enzyme. The preparation was then washed with 500 ml of buffer and was packed into a small column (4 mm inner diameter, 30 mm long) to be used in the analytical process. When not used, the column reactor was stored in 0.05 M sodium phosphate buffer pH 7.0 +0.02% sodium azide at 4 ◦ C. 2.3. Determination of the amount of protein bound to the supports The amount of protein was determined according to Lowry et al. (1951). The quantity of protein bound to the three supports were determined by the difference between the concentration of the protein in the solution before and after immobilization.

2.2. Immobilization of urease

2.4. Instrumentation

The immobilization of urease on inorganic support material surface consists of a three-step procedure. The initial

Fig. 2 shows the basic principle of the flow injection conductimetric biosensor system. The sample is injected into

W. Limbut et al. / Biosensors and Bioelectronics 19 (2004) 813–821 O

O

O Si OH Support materials

O

+ (CH3CH2O)3Si(CH2)3NH2

Support materials

O

O

O Si O Si (CH2)3NH2

O

O

O Si O Si(CH2)3N = CH(CH2)3CHO O

O

(A)

+

O

O

CHO

O

(CH2)3

O Si O Si(CH2)3N = CH(CH2)3CHO (B)

O

O Si O Si (CH2)3NH2

O Si OH

Support materials

815

CHO

O

O

+ H2N-Protein

O Support materials

O

O Si O Si(CH2)3N = CH(CH2)3CH = N-Protein O O O Si O Si(CH2)3N = CH(CH2)3CH = N-Protein O

O (C)

Fig. 1. Immobilizing enzyme by covalent binding; (A) Silanization of the surface of support materials, (B) Primary amino acid groups (R–NH2 ) of support are activated by glutaraldehyde (C5 H8 O2 ) to give a carbonyl derivative by formation of a Schiff’s base, (C) The formation of a Schiff’s base linkage between carbonyl groups of the activated support and free amino groups on the protein. (Adapted from Cabral and Kennedy, 1991; Weetall, 1976; Weetall and Lee, 1989).

the sample carrier buffer that is pumped through a dialyser before being sent to waste. The dialyser has a cellulose ester dialysis membranes with a MWCO 6000–8000 (Spectra/Por 1, Spectrum Medical Industries Inc., Los Angeles, USA). This allows small molecules, including urea, to pass through the membrane and to be collected in the buffer on the other side of the membrane. When the solution containing urea passed through the enzyme column, the ions was generated by the hydrolysis reaction of urea by urease according to the reaction urease

(H2 N)2 CO + 3 H2 O −−→ 2 NH4 + + HCO3 − + OH−

The increase of charged products was measures as the change in the conductivity of the solution by the conductivity cell at the outlet of the enzyme reactor column. The conductivity cell consisted of two 10 mm long stainless steel tubes (approximately 0.9 mm o.d.) glued to the ends of a glass tube (17 mm length, 1.0 mm i.d.). The ends of the electrodes inside the glass tube are approximately 4 mm apart. The conductivity change within the conductivity cell is measured by a lab-built conductivity meter. In this system, the background conductivity signal of the solution can be adjusted to zero allowing only the changes to be detected and amplified. The response, a voltage signal, is recorded

Fig. 2. Schematic diagram showing the flow injection conductimetric biosensor system.

816

W. Limbut et al. / Biosensors and Bioelectronics 19 (2004) 813–821

on a chart recorder. This voltage signal is linearly related to the solution conductance.

ered by balancing between the sensitivity and the analysis time.

2.5. Responses of urease immobilized on CPG, silica gel and Poraver®

2.7. Determination of urea in serum samples

The main aim of this work was to compare the efficiency of immobilized urease on CPG, silica gel and Poraver® in a conductimetric system for the determining of urea. Preliminary study was done by testing the responses of urease immobilized on all three supports in a simple system without a dialyser. Solutions of urea were prepared in 0.05 M glycine–NaOH buffer pH 8.80. The sample solutions, 500 ␮l, were introduced as pulses in the continuous flow of buffer, which flowed at a flow rate of 0.50 ml/min. The responses were the measured peak heights. 2.6. Optimization The operating conditions of the conductimetric biosensors in a flow injection analysis system, with a dialyser (Fig. 2), using immobilized urease were optimized. The variables were flow rate, type, pH and concentration of buffer, and sample volume. These are summarized in Table 1. The buffer used throughout the experiment, except when the effects of buffer were tested, was 0.05 M glycine–NaOH buffer pH 8.80. The background solution the of sample line was the same buffer plus 0.9 % (w/v) NaCl. The salt was added so that the solution used will be isotonic to serum which would be later tested. All samples were prepared using this solution. This is also to provide the same conductivity baseline for all analysis, i.e. standard urea and serum samples. The sample volume used throughout the experiment was 500 ␮l except when the effect of sample volume was tested. The optimization of each parameters were performed by changing a single parameter and kept other parameters constant. The optimum conditions in these systems were consid-

To demonstrate the use of the conductimetric biosensor (CPG and silica gel), the system was tested using the serum samples obtained from Songklanagarind Hospital, Prince of Songkla University, Hat Yai, Thailand. The same samples were analysed by UV–VIS spectrophotometry using the commercially available BUN test kit and the Bethelot reaction (the results obtained by Songklanagarind Hospital). The measurement of urea was carried out under optimum condition; 0.80 ml/min flow rate of buffer line, 0.3 ml/min of sample lines and 500 ␮l sample volume. The serum samples were diluted using 0.05 M glycine–NaOH buffer pH 8.80 (contained 0.9% (w/v) NaCl) at a serum: buffer ratio of 1:9. To calibrate the system, standard urea solutions (0.5, 1.0, 2.0, 3.0, 4.0, 5.0, 7.0 and 10.0 mM prepared in 0.05 M glycine–NaOH buffer pH 8.80 +0.9% (w/v) NaCl) were injected into the system. The calibration curve was prepared by plotting the conductivity change versus corresponding urea concentration (mM). The sample solutions were then injected into the system. The change in the conductivity of each sample was used to calculate the urea concentration from the calibration done prior to the test. For the Fearon reaction (BUN test kit; Sigma, procedure no. 535), it uses the direct interaction of urea with diacetyl monoxime. The absorbance of the solution was recorded at 525 nm. The plot of the absorbance versus urea concentration (mM) was used as the calibration curve. From the value of the absorbances the concentrations of urea in serum samples were calculated from the calibration curve. The conductimetric biosensor systems using CPG and silica gel were validated by comparing the results to those of the Fearon and the Berthelot reactions. In making such a comparison, the principle interest will be whether the proposed method gives results that are significantly higher or

Table 1 Assayed and optimized values of the conditions under study for the flow injection conductimetric biosensor system Parameters

Concentration of urea (mM)

Optimum CPG

Silica gel

0.30, 0.50, 0.70, 1.00, 1.20, 1.50 0.20, 0.30, 0.40, 0.50, 0.75, 1.00 0.60, 0.70, 0.80, 1.00, 1.20

0.1, 0.5, 1, 3, 5 5, 10, 15, 20, 30 5, 10, 15, 20, 30

1.00 0.30 0.80

1.00 0.30 0.80

5, 10, 15, 20, 30

Glycine–NaOH

Glycine–NaOH

Concentration

0.05 M Glycine–NaOH pH 8.80, 0.05 M Tris–HCl pH 7.40, 0.05 M imidazole-HCl pH 7.50 0.050, 0.010, 0.005, 0.001 M Glycine–NaOH pH 8.80

0.050

0.050

pH

0.05 M Glycine–NaOH pH 8.60, 8.80, 9.00, 9.50, 10.00

0.1, 0.5, 1, 3, 5, 10, 15, 20 1, 3, 5, 10, 15

8.80

8.80

200, 300, 400, 500

3, 5, 7, 10, 15

500

500

Flow rate (ml/min) (a) Without dialyser (b) Sample line with dialyser (c) Buffer line with dialyser Buffer solution Type

Sample volume (␮l) CPG: controlled pore glass.

Values

W. Limbut et al. / Biosensors and Bioelectronics 19 (2004) 813–821

lower than the established methods. So, the analysis using the regression line (Miller and Miller, 1993) and the Wilcoxon signed rank test (Triola, 1998) were used in this work.

3. Results and discussion 3.1. Responses of urease immobilized on CPG, silica gel and Poraver® The responses of the three enzyme reactor columns with urease immobilized on CPG, silica gel and Poraver® , in a flow system without a dialyser, are shown in Fig. 3. They are linear in the concentration range between 5 to 45 mM. The sensitivity (slope of the linear range) of the enzyme reactor column using Poraver® was 9.8 times less than CPG and 9.2 times less than silica gel. The responses of the enzyme reactor columns with urease immobilized on CPG and silica gel were similar and were much higher than that of Poraver® . This may due to two reasons, one being the percentage of SiO2 group in Poraver® (67%) which was 31% less than CPG (>97%) and 32% less than silica gel (>99%) (quoted by the manufacturer). SiO2 is important for the linkage between support and enzyme. Less SiO2 means less immobilized enzyme and, thus, smaller response. The other reason, the surface area, both the particle size and pore size of Poraver® are much larger than those of CPG and silica gel. Therefore, for the same reactor volume the reactor with Poraver® has much less surface area where the enzyme can be attached. In view of this, one would expect a higher response from enzyme immobilized on silica gel since its surface area is higher than CPG. However, as shown in Fig. 3 the response from silica gel reactor is slightly lower than CPG. According to one manufacture (CPG Inc.-Online, 2002), a mean pore size of 75 Å excludes globular proteins with molecular weights higher than 30,000. Since the molecular weight of urease is about

817

480,000 (Reithel, 1971) this would make the pore size of silica gel, 60 Å, too small for the enzyme to enter but its external surface is large enough for the immobilized enzyme to give similar response to the reactor using CPG. For the flow injection analysis of urea in human serum samples, a dialyser is used to filter off large molecules to prevent them from clogging up the enzyme reactor. In effect only some urea molecules pass through the dialysis membrane and are analysed. Therefore, a high sensitivity system is required to cope with the dilution of urea in this analysis system. Since Poraver® had a much lower sensitivity compared to CPG and silica gel, the latter two supports were chosen for further studies. 3.2. The amount of protein bound to the supports The amounts of protein bound to the supports were taken as the differences between the amounts of protein in the solution before and after immobilization. The concentrations of the enzyme urease on the supports (U/cm3 of support) were found to be, CPG 220 U/cm3 , silica gel 207 U/cm3 , and Poraver® 33 U/cm3 . The results were of the same trend as the responses in Fig. 3. CPG with the highest enzyme concentration provided the best responses, silica gel with a slightly lower enzyme concentration gave a slightly lower responses, and lastly Poraver® , with the least amount of enzyme that was much lower than the other two supports, gave much lower responses than CPG and silica gel. 3.3. Optimization 3.3.1. Flow injection system 3.3.1.1. Flow rate. In a flow system, the flow rate of the solution passing through the reactor and the detector is the main factor affecting the dispersion of the analyte particles, yield of the reaction and response of the detector. The dispersions that occur will affect the result of flow system. An

500

Responses (mV)

450 400 350 300

CPG silica gel Poraver

250 200 150 100 50 0 0

50

100

150

200

250

Concentration (mM) Fig. 3. Comparison of the responses obtained from enzyme reactor columns with urease immobilized on CPG, silica gel and Poraver® .

W. Limbut et al. / Biosensors and Bioelectronics 19 (2004) 813–821

increasing flow rate can decrease the dispersion effects. On the other hand, the yield of the reaction and the response of the detector depend on the retention time of the sample in the reactor column. So optimization of flow rate is necessary. In this proposed system, a dialyser was used and this consisted of two flow lines. A sample line where sample was introduced and a buffer line which carried the analyte that passed through the dialysis membrane from the sample to the reactor. The flow rate of these two lines would also affect the diffusion of the sample molecules through the membrane and hence the dispersion of the analyte particles. Therefore, both must be optimized. This was done in three steps: (a) optimized the flow rate of the system without a dialyser, (b) incorporated the dialyser into the system and optimized the flow rate of the sample line while the flow rate of the buffer line was fixed at the optimum as found in (a), and (c) fixed the flow rate of the sample line as found in (b) and re-optimized the flow rate of the buffer line. The peak heights and sensitivity of the responses decreased as the flow rate increased. For slower flow rate, the calibration curve gave higher sensitivity but also required longer analysis time. Therefore, the balance between peak height, sensitivity and analysis time were considered. The optimum condition flow rate of the sample and the buffer line for both system by enzyme reactor columns with urease immobilized on CPG and silica gel are shown in Table 1. The optimum flow rate of the sample and buffer line found in these experiments were used in the remainin experiments. 3.3.1.2. Buffer solution. Type. In the conductimetric system, the sensitivity is controlled by the ratio G/G, where G is the conductance of the medium (background level) and G is the conductance change that results from the enzymatic process (Mikkelsen and Rechnitz, 1989). In the case of high conductance buffer solution, the background noise level would be high and this would lead to a low sensitivity. Thus, it is necessary to find the suitable type of buffer solution and the results are shown in Table 1. The peak heights and sensitivities differed significantly between three types of buffer. The highest sensitivity was obtain from glycine–NaOH, and was about 40–70% higher than the other two buffers. This is because glycine–NaOH had the lowest conductance, that is, the background noise level would be low and this would lead to a high sensitivity. Concentration. The effect of buffer concentration (viscosity) has to be examined because viscosity of buffer solution would influence the hydrodynamic flow of fluid permeation rate through the dialysis membrane. The response of the low buffer concentration (low viscosity) was more than that of the high buffer concentration (high viscosity), because the fluid had high permeation rate through the dialysis membrane, but its handicap was the low buffer

capacity. In both reactors at 0.001 M, the responses gave the best sensitivity and the lowest limit of detection (0.1 mM). This might be attributed to local pH changes within the enzyme column due to the low capacity of buffer (Mikkelsen and Rechnitz, 1989). From these results, it seems that the lower the concentration the better response. However, another matter to be considered is whether the buffer capacity would be enough when uses with serum sample. From our test with real samples, the lowest buffer concentration that provided enough buffering capacity for serum was 0.050 M. Therefore, 0.050 M was chosen (Table 1). pH. Enzyme activity is known to be dependent upon the pH. The activity of urease as quoted by the manufacturer (Sigma, USA) was maximal at pH 7.0. However, when the enzyme is immobilized, the optimal pH may shift, depending on the nature of the support material (Cabral and Kennedy, 1991; Guilbault, 1984). Thus, it is necessary to find the optimum pH of buffer solution. In the conductimetric system, pH 7.00 is not obtainable using glycine–NaOH buffer since its pH range is between 8.60 and 10.60. In this case, pH 8.60 should be the best value since it is nearest to pH 7.00. However, the responses were maximal at both pH 8.60 and 8.80, and pH 8.80 was chosen because it has more buffering capacity than pH 8.60. 3.3.1.3. Sample volume. One way of improving the response of the system is to increase analyte by increasing the sample volume. However, in an enzymatic analysis the reaction yield also depends on the amount of enzyme. So, too much of the analyte for the same amount of enzyme can not increase the response. Moreover, large sample volume may increase the particle dispersion. Therefore, a suitable sample volume should be found. The peak heights and sensitivities increased as the sample volume increased and they differed significantly (P = 0.05) between each sample volume. In our systems, the sample Sensitivity of response (mV/ mM)

818

60 silica gel CPG

50 40 30 20 10 0 0

50

100 150 200 250 Operation time (h)

300

350

Fig. 4. Comparison of the sensitivity of responses of the enzyme reactor columns with urease immobilized on CPG and silica gel at different operation times.

W. Limbut et al. / Biosensors and Bioelectronics 19 (2004) 813–821 Table 2 Performances of CPG and silica gel enzyme reactor columns Parameters

to the denaturation of the immobilized urease. Although the sensitivities were decreased, they are still at sufficient values to be used in the systems.

Performances

Limit of detection (mM) Linear range (mM) Sensitivity (mV/mM) R2 Stability of enzyme reactor is more than (h) Sensitivity decreasing rate of the enzyme reactor column (mV/mM h) Analysis time (min)

819

CPG

Silica gel

0.5 Still linear up to 150 53 0.9997 310

0.5 Still linear up to 150 51 0.9995 310

−0.0902

−0.0675

4–5

4–5

3.5. Performances of CPG and silica gel enzyme reactors in the flow injection conductimetric biosensor systems Table 2 summarized the performances of the conductimetric system under optimum conditions. Both columns gave approximately the same results. The sensitivity of the enzyme reactor column with urease immobilized on CPG was only slightly better, that is, 4% higher than the column with silica gel. 3.6. Comparison between the results obtained from the biosensor system and other analytical methods

volume is limited to 500 ␮l due to the volume of the sample loop in the sample injector and this was chosen because it gave the highest sensitivity (Table 1).

The analysis of urea using the conductimetric biosensor system (CPG and silica gel), the Fearon, and the Berthelot reactions were done on the same serum samples. Comparisons were done for four pairs of different analysis techniques using the regression line method. The results are shown in Table 3. In all cases, the slope and the intercept did not differ significantly from the ideal value of 1 and 0, respectively, thus, there is no evidence for systematic differences between the methods. To verify the reliability, the Wilcoxon signed rank test was also used. In this test, the null hypothesis (there is no difference between the two methods) is rejected at a significance level P = 0.05 if the experimental value is less than or equal to the critical values (Triola, 1998). For all cases (Table 4), the null hypothesis is retained, that is, there is no evidence for a systematic difference between the results

3.4. Stability of immobilized enzyme Generally, after prolonged use of the enzyme, denaturation or inhibition of the enzyme may effect the response (Lehninger, 1987). The long-term performance of both enzyme reactor columns (urease immobilized on CPG and silica gel) was evaluated intermittently over a period of 7 months by monitoring its response to urea standard solution. Fig. 4 shows the sensitivities of both enzyme reactor columns at different operation times. After 310 h operation time the enzyme reactor columns using CPG and silica gel retained about 52 and 54% of their original sensitivities, respectively. The decay of sensitivity should be contributed

Table 3 Regression line statistics obtained from the comparison of the analytical methods Comparison of two analytical methods Y-axis

R2

Regression line X-axis

Conductimetric Conductimetric Conductimetric Conductimetric

biosensor biosensor biosensor biosensor

(CPG) (CPG) (silica gel) (silica gel)

Berthelot reaction Fearon reaction Berthelot reaction Fearon reaction

Y Y Y Y

= 1.01 = 0.99 = 0.99 = 0.97

± ± ± ±

0.03X − 0.44 ± 0.91 0.03X + 0.23 ± 0.95 0.03X + 0.29 ± 0.97 0.05X + 0.79 ± 1.56

0.996 0.995 0.995 0.987

Table 4 The Wilcoxon sign rank test for the comparison of two analytical methods Comparison of two analytical methods

Biosensor Biosensor Biosensor Biosensor

(CPG) and (CPG) and (silica gel) (silica gel)

Berthelot reaction Fearon reaction and Berthelot reaction and Fearon reaction

Test statistic T is the sum of

Critical Value at P = 0.05

Positive ranks

Negative ranks

n

Value

208 130 195 143

117 147 130 157

25 23 25 24

90 73 90 81

Test results

× × × ×

The null hypothesis (there is no difference between the two methods) is rejected if the test statistic T (the lower of the sum of positive rank or negative rank-shown as italic) is less than or equal to the critical value. The null hypothesis can not be rejected if the test statistic T is greater than the critical value. ×: fail to reject the null hypothesis.

820

W. Limbut et al. / Biosensors and Bioelectronics 19 (2004) 813–821

of conductimetric biosensor technique (CPG, silica gel), Berthelot and Fearon reactions. Therefore, the concentrations determined by biosensor conductimetric technique (CPG, silica gel) are in good agreement to the other two methods. It should be noted here that, in this work the system was applied to analyze urea in serum. The background solution for the sample line in the flow system contained 0.9% (w/v) NaCl. This was added so that the solution used will be isotonic to serum. That is, the baseline conductivity was adjusted to be the same as the samples that were analyzed. However, in a more general case where the salt concentration, i.e. the conductivity, of the sample is not known it is better to determine the differential signal of the solution conductivity before and after passing through the enzyme reactor. Therefore, further development for a differential signal measurement can increase the versatility of the conductimetric biosensor system.

4. Conclusions The enzyme reactor with urease immobilized on silica gel had the same efficiency as the enzyme reactor column with urease immobilized on CPG. Since silica gel is much cheaper than CPG the enzyme reactor technique, using silica gel, can then be used more widely especially in the disposable on inhibition system where material cost plays an important part. However, in an analysis system which does not require very high sensitivity the enzyme reactor with immobilized on Poraver® is also a good alternative since it is much cheaper than CPG.

Acknowledgements This project was supported by the National Electronics and Computer Technology Center (NECTEC); Higher Education Development Project: Postgraduate Education and Research Program in Chemistry (PERCH), funded by The Royal Thai Government and the Graduate School, Prince of Songkla University, Hat Yai, Thailand. The scholarship for Mr. Warakorn Limbut from PERCH was also gratefully acknowledged.

References Adams, R.E., Carr, P.W., 1978. Coulometric flow analyzer for urea with immobilized enzyme reactors. Anal. Chem. 50, 944–950. Bowers, L.D., Canning, L.M., Sayers, C.N., Carr, P.W., 1976. Rapid-flow enthalpimetric determination of urea in serum, with use of an immobilized urease reactor. Clin. Chem. 22, 1314–1318. Cabral, J.M.S., Kennedy, J.F., 1991. Covalent and coordination immobilization of proteins. In: Taylor, R.F. (Ed.), Protein Immobilization;

Fundamentals and Applications. Marcel Dekker, New York, pp. 73– 138. Chandler, G.K., Dodgson, J.R., Eddowes, M.J., 1989. ISFET-based enzyme sensors for urea: enzyme-modified ISFETs and columnimmobilised enzyme flow injection analysis. Anal. Proc. 26, 154– 156. Chandler, G.K., Dodgson, J.R., Eddowes, M.J., 1990. An ISFET-based flow injection analysis system for determination of urea: experiment and theory. Sens. Actuat. B1, 433–437. CPG Inc.-Online, 2002. Controlled-pore glass bio-supports application guide. http://www.cpg-biotech.com/cpgp/appguide1.html. Gorton, L., Ogren, L., 1981. Flow injection analysis for glucose and urea with enzyme reactors and on-line dialysis. Anal. Chem. Acta 130, 45–53. Guilbault, G.G., 1984. Analytical Uses of Immobilized Enzymes. Marcel Dekker, New York, p. 449. Guilbault, G.G., Montalvo, J.G., 1969. A specific urea 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. Guilbault, G.G., Shu, F.R., 1972. Enzyme electrode based on the use of a carbon dioxide sensor. Anal. Chem. 44, 2161–2165. Guilault, G.G., Tarp, M., 1974. A specific enzyme electrode for urea. Anal. Chem. Acta 73, 355–365. Jurkiewicz, M., Alegret, S., Almirall, J., Garcia, M., Fabregas, E., 1998a. Development of a biparametric bioanalyser for creatinine and urea. Validation of the determination of biochemical parameters associated with hemodialysis. Analyst 123, 1321–1327. Jurkiewicz, M., Alegret, S., Fabregas, E., 1998b. Comparison of flow injection analytical biosystems based on open-tube and packed-bed enzyme reactors. Anal. Chem. Acta 370, 47–58. Lehninger, A.L., 1987. The urea clycle. In: Principle of Biochemistry, fifth edition. Worth Publishers, New York, p. 531. Lowry, O.H., Rosebrough, N.J., Farr, A.L., Randall, R.J., 1951. Protein measurement with the folin phenol reagent. J. Biol. Chem. 193, 265– 275. Mascini, M., Guilbault, G.G., 1977. Urease coupled ammonia electrode for urea determination in blood serum. Anal. Chem. 49, 795– 798. Mikkelsen, S.R., Rechnitz, G.A., 1989. Conductometric Transducers for enzyme-based biosensors. Anal. Chem. 61, 1737–1742. Miller, J.C., Miller, J.N., 1993. Statistics for Analytical Chemistry, third edition. Simon & Schuster International Group, West Sussex. Reithel, F.J., 1971. Ureases. In: Boyer, P.D. (Ed.), The Enzymes, vol. IV, Hydrolysis. Academic Press, New York, pp. 1–21. Ruzicka, J., Hansen, E.H., Ghose, K., 1979. Enzymatic determination of urea in serum based on pH measurement with the flow injection method. Anal. Chem. 51, 199–206. Thavarungkul, P., Asawatreratanakul, P., Kanatharana, P., Duenjumroon, J., Chaibundit, C., 1999. Flow-through enzyme reactor system for urea determination in blood serum using conductimetric measurement. ScienceAsia 25, 157–163. Thavarungkul, P., Håkanson, H., Holst, O., Mattiasson, B., 1991. Continuous monitoring of urea in blood during dialysis. Biosens. Bioelectron. 6, 101–107. Thavarungkul, P., Kanatharana, P., 1994. An enzyme sensor for urea based on conductivity measurement. J. Sci. Soc. Thailand 20, 23– 30. Triola, M.F., 1998. Wilcoxon signed-rank test for two dependents samples. In: Triola, M.F. (Ed.), Elementary Statistics. Addison-Wesley, USA, pp. 655–668. Vadgama, P., 1986. Urea pH electrods: charaterisation and optimisation for plasma measurements. Analyst 111, 875–878. Walcerz, I., Koncki, R., Leszczynska, E., Glab, S., 1995. Enzyme biosensors for urea determination based on an ionophore

W. Limbut et al. / Biosensors and Bioelectronics 19 (2004) 813–821 free pH membrane electrode. Anal. Chem. Acta 315, 289– 296. Walcerz, I., Glab, S., Koncki, R., 1998. Potentiometric enzyme electrode in a flow injection system for the determination of urea in human serum samples. Anal. Chem. Acta 369, 129–137. Weetall, H.H., 1976. Covalent coupling methods for inorganic support materials. Methods Enzymol. 44, 134–148.

821

Weetall, H.H., Lee, M.J., 1989. Antibodies Immobilized on Inorganic Supports. Appl. Biochem. Biotechnol. 22, 311–330. Xie, U.H., Mecklenburg, M., Danielsson, B., 1994. Miniature thermal biosensor determines urea and lactate. Clin. Chem. 40, 2282–2287. Yasuda, K., Miyagi, H., Hamada, Y., Takata, Y., 1984. Determination of urea in whole blood using a urea electrode with an immobilised urease membrane. Analyst 109, 61–64.