Anal. Chem. 2001, 73, 5965-5971
A Planar Amperometric Creatinine Biosensor Employing an Insoluble Oxidizing Agent for Removing Redox-Active Interferences Jae Ho Shin, Yong Suk Choi, Han Jin Lee, Sung Hyuk Choi, Jeonghan Ha, In Jun Yoon, Hakhyun Nam, and Geun Sig Cha*
Chemical Sensor Research Group, Department of Chemistry, Kwangwoon University, Seoul 139-701, Korea
Creatinine, a cyclic anhydride of creatine, is an end product of muscle metabolism that is synthesized in the body at a fairly constant rate from creatine.1 Since the concentration of creatinine in serum and urinary excretion relative to that of urea is less affected by factors such as sepsis, trauma, fever, or dietary changes, creatinine levels give a more sensitive and specific index for evaluating glomerular filtration rate and, in general, for assessing renal, thyroid, and muscular functions.2 Therefore, the accurate and rapid determination of creatinine in biological fluids is of utmost relevance in the diagnosis and treatment of muscular and kidney disorders. According to a recent review article,3 creatinine is the most frequently measured analyte in clinical laboratories.
The normal physiological range of blood creatinine for adults is between 0.6 and 1.2 mg/dL (53-106 µM).1,4 However, in infants and children or in persons with small stature, decreased muscle mass, or inadequate dietary protein, creatinine is present in much lower concentrations (27-53 µM).4,5 On the other hand, its pathological values observed during impaired renal function, chronic nephritis, or urinary tract obstruction can increase to 1000 µM.6 Thus, a method for the determination of creatinine should have a lower detection limit in at least micromolar levels and remain linear up to millimolar concentrations as well. For routine assays of creatinine, most commercially available analyzers have employed the spectrophotometric procedure based on the Jaffe´ reaction, a nonenzymatic method, which involves the formation of the colored product with picrate in alkaline solution.7 The applicability of the Jaffe´ and its modified systems, however, has often been limited for two reasons. First, severe interferences from numerous metabolites and drugs cause poor selectivity. Second, the processes are costly, time-consuming, complicated, or difficult to automate.8 For such drawbacks, these methods have failed to keep up with the new trends in modern health care settings that demand point-of-care or decentralized testing devices capable of rapidly and reliably monitoring creatinine levels at or near the patient’s bedside. Of all the feasible approaches to fulfill such requirements, biosensors have received attention as the most attractive solution. A biosensor is a complex device comprising biochemical recognition elements and an electrochemical transducer, e.g., potentiometric or amperometric probe. Such a device not only is more specific for creatinine but also provides rapid measurement, and is inexpensive and simple to microfabricate and integrate in sensing cartridges. Rechnitz and co-workers9 first reported on creatinine electrodes using a potentiometric gas-sensing electrode to measure the ammonia liberated from the hydrolysis of creatinine by creatinine iminohydrolase. Substantial improvements in creatinine biosensor sensitivity and fabrication have been accomplished by employing
* Corresponding author: (fax) +822-911-8584; (e-mail) gscha@daisy. kwangwoon.ac.kr. (1) Whelton, A.; Watson, A. J.; Rock, R. C. Renal Function and Nitrogen Metabolites. In Tietz Textbook of Clinical Chemistry, 3rd ed.; Burtis, C. A., Ashwood, E. R., Eds.; W. B. Saunders Co.: Philadelphia, PA, 1999; Chapter 35. (2) Spencer, K. Ann. Clin. Biochem. 1986, 23, 1-25. (3) Bakker, E.; Diamond, D.; Lewenstam, A.; Pretsch, E. Anal. Chim. Acta 1999, 393, 11-18.
(4) Kee, J. L. Laboratory and Diagnostic Tests with Nursing Implications, 2nd ed.; Appleton & Lange: Norwalk, CT, 1987; pp 132-133. (5) Mendelssohn, D. C.; Barrett, B. J.; Brownscombe, L. M.; Ethier, J.; Greenberg, D. E.; Kanani, S. D.; Levin, A.; Toffelmire, E. B. Can. Med. Assoc. J. 1999, 161, 413-417. (6) Sena, S. F.; Syed, D.; McComb, R. B. Clin. Chem. 1988, 34, 2144-2148. (7) Jaffe´, M. Z. Physiol. Chem. 1886, 10, 391-400. (8) Weber, J. A.; van Zanten, A. P. Clin. Chem. 1991, 37, 695-700. (9) (a) Thompson, H.; Rechnitz, G. A. Anal. Chem. 1974, 46, 246-249. (b) Meyerhoff, M. E.; Rechnitz, G. A. Anal. Chim. Acta 1976, 85, 277-285.
A planar microchip-based creatinine biosensor employing an oxidizing layer (e.g., a PbO2 film), where interfering redox-active substances are broken (i.e., oxidized) to redox-inactive products, was developed to facilitate the microfabrication of the sensor and to provide improved, reliable determination of creatinine in physiological samples. The feasibility of using hydrophilic polyurethanes in permselective barrier membranes for creatinine biosensors and the effect of adding a silanizing agent (adhesion promoter) on the sensor performance (e.g., sensitivity, stability, and lifetime) are described. The proposed creatinine microsensor with a three-layer configuration, i.e., enzyme, protecting, and oxidizing layers, exhibits good electrochemical performance in terms of response time (t95% ) 98 s at 100 f 200 µM creatinine change), linearity (1-1000 µM, r ) 0.9997), detection limit (0.8 µM), and lifetime (∼35 days). The creatinine biosensor devised in a differential sensing arrangement that compensates the erroneous results from creatine is considered to be suitable for assay of serum specimens.
10.1021/ac010497a CCC: $20.00 Published on Web 11/15/2001
© 2001 American Chemical Society
Analytical Chemistry, Vol. 73, No. 24, December 15, 2001 5965
Figure 1. Interference-removing principle of the creatinine biosensor employing the oxidizing layer.
ammonium ion-selective membrane electrodes10 or amperometric ammonia sensors11 as signal-transducing elements. Nevertheless, the use of such a detection scheme still has inherent problems in accuracy, reliability, and construction due to the interfering effect of endogenous ammonia present in natural specimens.10a,11b,12 Particular consideration also has been given to the enzyme catalytic sequence proposed by Tsuchida and Yoda,13 which is based on the three-enzyme system consisting of creatinine amidohydrolase (creatininase, CA), creatine amidinohydrolase (creatinase, CI), and sarcosine oxidase (SO): CA
creatinine + H2O y\z creatine CI
creatine + H2O y\z sarcosine + urea
(1) (2)
SO
sarcosine + H2O + O2 y\z formaldehyde + glycine + H2O2 (3)
Most creatinine biosensors incorporating such biocatalytic reactions have been coupled with the detection of the hydrogen peroxide ultimately generated.14-17 However, the assay of creatinine still remains one of the most challenging analytical problems, owing to the possible interferences and the presence of creatine in serum at nearly the same concentrations as those of creatinine. Because the direct electrooxidation of H2O2 requires a relatively high working potential (∼+0.7 V vs Ag/AgCl), its amperometric detection often accompanies serious interferences arising from readily oxidizable metabolites such as ascorbic acid, uric acid, and acetaminophen. To address this problem with interferences, several attempts have been explored previously, including the use of electron-transfer mediators (e.g., ferrocene and its derivatives,18 hydroquinone,19 and cobalt phthalocyanine-modified electrode20), conducting polymers15,21 doped with a polyanion to lower the working potential, and the use of permselective membranes (e.g., cellulose acetate,16a,22 Nafion,23 and electropolymerized films16b,24) to discriminate H2O2 from interfering species by size exclusion or electrostatic repulsion. In yet another approach, electrocatalytic reductive methods modified with a peroxidase enzyme have been extensively studied for the deter5966
Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
mination of H2O2.17,25 Recently, it was found that the introduction of insoluble, strong oxidizing agents, e.g., lead dioxide, at the sample pretreatment zone offer a promising strategy to remove electroactive substances, resulting in practical elimination of interferences.26,27 (10) (a) Matuszewski, W.; Trojanowicz, M.; Meyerhoff, M. E.; Moszczynska, A.; Lange-Moroz, E. Electroanalysis 1993, 5, 113-120. (b) Jurkiewicz, M.; Alegret, S.; Almirall, J.; Garcia, M.; Fabregas, E. Analyst 1998, 123, 13211327. (11) (a) Trojanowicz, M.; Lewenstam, A.; Krawczynski vel Krawczyk, T.; Lahdesmaki, I.; Szczepek, W. Electroanalysis 1996, 8, 233-243. (b) Shih, Y.-T.; Huang, H.-J. Anal. Chim. Acta 1999, 392, 143-150. (12) (a) Collison, M. E.; Meyerhoff, M. E. Anal. Chim. Acta 1987, 200, 61-72. (b) Rui, C.-S.; Sonomoto, K.; Ogawa, H.-I.; Kato, Y. Anal. Biochem. 1993, 210, 163-171. (13) Tsuchida, T.; Yoda, K. Clin. Chem. 1983, 29, 51-55. (14) (a) Sakslund, H.; Hammerich, O. Anal. Chim. Acta 1992, 268, 331-345. (b) Kim, E. J.; Haruyama, T.; Yanagida, Y.; Kobatake, E.; Aizawa, M. Anal. Chim. Acta 1999, 394, 225-231. (15) (a) Yamato, H.; Ohwa, M.; Wernet, W. Anal. Chem. 1995, 67, 2776-2780. (b) Schneider, J.; Grundig, B.; Renneberg, R.; Cammann, K.; Madaras, M. B.; Buck, R. P.; Vorlop, K.-D. Anal. Chim. Acta 1996, 325, 161-167. (c) Khan, G. F.; Wernet, W. Anal. Chim. Acta 1997, 351, 151-158. (16) (a) Madaras, M. B.; Popescu, I. C.; Ufer, S.; Buck, R. P. Anal. Chim. Acta 1996, 319, 335-345. (b) Madaras, M. B.; Buck, R. P. Anal. Chem. 1996, 68, 3832-3839. (17) Kinoshita, H.; Torimura, M.; Kano, K.; Ikeda, T. Electroanalysis 1997, 9, 1234-1238. (18) (a) Liaudet, E.; Battaglini, F.; Calvo, E. J. J. Electroanal. Chem. 1990, 293, 55-68. (b) Gilmartin, M. A. T.; Hart, J. P. Analyst 1995, 120, 1029-1045. (19) Kajiya, Y.; Sugai, H.; Iwakura, C.; Yoneyama, H. Anal. Chem. 1991, 63, 49-54. (20) (a) Hart, J. P.; Wring, S. A. Trends Anal. Chem. 1997, 16, 89-103. (b) Sergeyeva, T. A.; Lavrik, N. V.; Rachkov, A. E.; Kazantseva, Z. I.; Piletsky, S. A.; El’skaya, A. V. Anal. Chim. Acta 1999, 391, 289-297. (21) (a) Yabuki, S.; Shinohara, H.; Aizawa, M. J. Chem. Soc., Chem. Commun. 1989, 14, 945-946. (b) Cooper, J. M.; Bloor, D. Electroanalysis 1993, 5, 883-886. (22) (a) Gilmartin, M. A. T.; Hart, J. P. Analyst 1994, 119, 2331-2336. (b) Vaidya, R.; Wilkins, E. Electroanalysis 1994, 6, 677-682. (23) (a) Harrison, D. J.; Turner, R. F. B.; Baltes, H. P. Anal. Chem. 1988, 60, 2002-2007. (b) Gilmartin, M. A. T.; Hart, J. P.; Birch, B. Analyst 1992, 117, 1299-1303. (24) Moser, I.; Jobst, G.; Aschauer, E.; Svasek, P.; Varahram, M.; Urban, G.; Zanin, U. A.; Tjoutrina, G. Y.; Zharikova, A. V.; Berezov, T. T. Biosens. Bioelectron. 1995, 10, 527-532. (25) (a) Mulchandani, A.; Wang, C.-L.; Weetall, H. H. Anal. Chem. 1995, 67, 94-100. (b) Bartlett, P. N.; Birkin, P. R.; Wang, J. H.; Palmisano, F.; De Benedetto, G. Anal. Chem. 1998, 70, 3685-3694. (c) Xiao, Y.; Ju, H.-X.; Chen, H.-Y. Anal. Chim. Acta 1999, 391, 299-306. (26) Johnston, J. B.; Daubney, S. D.; Palmer, J. L. International Patent, WO90/ 12,113, 1990.
In this work, a planar microchip-based creatinine biosensor was designed with an oxidizing layer to facilitate microfabrication. The biosensor incorporates biocatalytic components and a lead oxide film, both integrated on an electrode-sensing area, to provide a device capable of improved, reliable determination of creatinine in physiological samples (Figure 1). Interfering redox-active substances are converted (i.e., oxidized) to redox-inactive products during their diffusion through the oxidizing PbO2 layer before they reach the platinum working electrode. In addition, erroneous results from creatine are minimized by subtracting the signal from a creatine electrode from the creatinine signal in a differential sensing arrangement. Herein we describe the optimal compositions of three polymeric composites, i.e., enzyme, protecting, and oxidizing layers, and the effects of a silanizing agent (adhesion promoter) on the sensor performance (e.g., sensitivity, stability, and lifetime). The feasibility of using hydrophilic polyurethanes (HPUs) as outer membranes of creatinine sensors was also investigated. Finally, the practical applicability of the proposed PbO2-based differential creatinine sensor is demonstrated by determining creatinine content in control human serum specimens. EXPERIMENTAL SECTION Reagents. The sources of reagents used were as follows: creatinine amidohydrolase (creatininase; from recombinant Escherica coli., 600 units/mg), creatine amidinohydrolase (creatinase; from recombinant E. coli., 14.2 units/mg), and sarcosine oxidase (from recombinant E. coli., 31.6 units/mg) from Kikkoman (Tokyo, Japan); creatinine anhydrous, creatine anhydrous, ascorbic acid, uric acid, and acetaminophen from Sigma (St. Louis, MO); poly(vinyl alcohol) (PVA; MW 9000-10 000), lead dioxide, ceric dioxide, manganese dioxide, poly(ethylene glycol) (PEG; MW 1500), poly(propylene glycol) (PPG; MW 1000), ethylene glycol (EG), and dibutyltin di-n-laurate (DBTDL) from Aldrich (Milwaukee, WI); methylene bis(4-cyclohexyl isocyanate) (Desmodur W) from Miles (Pittsburgh, PA); (3-aminopropyl)triethoxysilane (Z-6011) from Dow Corning (Midland, MI); platinum, silver, and glass-filled dielectric pastes from Du Pont (Research Triangle Park, NC); and control human serum samples from Nissui Pharmaceutical (Tokyo, Japan). HPUs were prepared as described in our earlier work28 by reacting Desmodur W with a mixture of PEG, PPG, EG, and DBTDL. Three types of HPUs were synthesized by varying PEG/PPG ratios, i.e., HPU-A, 0.005/0.015; HPUB, 0.01/0.01; and HPU-C, 0.015/0.005. These hydrophilic polymers exhibit different water uptake properties (the weight ratio between water absorbed and dry HPU) and molecular weights: HPU-A, 40%/MW 574 000; HPU-B, 100%/MW 280 000; and HPU-C, 200%/ MW 87 000, respectively. All other solvents and chemicals used were analytical-reagent grade. Standard solutions and buffers were prepared with deionized water. Preparation of Microchip-Based Biosensors. The planartype amperometric sensor chips with a two-electrode configuration (Figure 2) were fabricated by sequentially screen-printing Pt, Ag, (27) Cui, G.; Kim, S. J.; Choi, S. H.; Nam, H.; Cha, G. S.; Paeng, K.-J. Anal. Chem. 2000, 72, 1925-1929. (28) (a) Shin, J. H.; Yoon, S. Y.; Yoon, I. J.; Choi, S. H.; Lee, S. D.; Nam, H.; Cha, G. S. Sens. Actuators, B 1998, 50, 19-26. (b) Lee, J. S.; Lee, S. D.; Cui, G.; Lee, H. J.; Shin, J. H.; Cha, G. S.; Nam, H. Electroanalysis 1999, 11, 260267.
and dielectric pastes onto an alumina plate, thermally treating the plate at 1100 °C for Pt, 850 °C for Ag, and 850 °C for dielectric film, respectively.28b To prepare the pseudoreference Ag/AgCl electrode, the printed Ag surface was treated with 0.1 M FeCl3 for 10 min. Planar creatinine sensors were made by sequentially forming the following three layers: enzyme, protecting, and oxidizing films on the base working and reference electrodes. The enzyme hydrogel layer (30 µm in thickness) was deposited by dispensing 3 µL of an aqueous mixture consisting of appropriate amounts of enzymes (i.e., 100 units of CA, 100 units of CI, and 100 units of SO for a creatinine-sensing site; and 100 units of CI and 100 units of SO for a creatine-sensing site, respectively) and 12 mg of PVA dissolved in 300 µL of phosphate buffer (0.053 M Na2HPO4, 0.015 M NaH2PO4, and 0.05 M NaCl; pH 7.6) and then drying the film for 20 min. The protecting membrane (20 µm in thickness) was formed by applying 6 µL of stock solutions of three different HPUs (i.e., HPU-A, -B, and -C) dissolved in cyclohexanone (500 µL for 20 mg of HPU) and evaporating the solvent for 20 min. To complete the preparation of biosensors, the oxidizing layer (30 µm in thickness) was subsequently deposited onto the protecting film. The cocktails for the oxidizing layer were prepared by additionally incorporating several different amounts (i.e., 140, 200, and 260 mg/mL) of PbO2 into the composite solution used for depositing the HPU-based protecting membrane. Furthermore, various amounts (i.e., 0, 12, 24, and 48 µL/mL) of Z-6011 were mixed with deposition solutions prepared for both the protecting and the oxidizing layers. The 8-µL aliquot of HPU/PbO2 casting solutions was dispensed on the HPU layer, and the membrane was cured overnight at room temperature. All polymeric layers were deposited with a pneumatic dispenser (EFD model 1000XL, Providence, RI). Evaluation of Sensor Performance. Amperometric measurements were performed using an eight-channel potentiostat (model cDAQ-0804, Dailinfo Co.; Seoul, Korea) controlled by a personal computer with a RS232 serial port for I/O communication and the data acquisition software written in LabVIEW programming development language (National Instruments; Austin, TX). For data analysis and presentation, IGOR Pro (WaveMetrics; Lake Oswego, OR) was utilized. Response and calibration curves were obtained through the additions of standard solutions (creatinine or creatine solution) to 25 mL of a background electrolyte (0.053 M Na2HPO4, 0.015 M NaH2PO4, and 0.05 M NaCl; pH 7.6) at room temperature with stirring. The creatinine and creatine solutions were freshly prepared every 3-4 days. Current was recorded every 1 s at an applied potential of +0.8 V versus Ag/AgCl. When not in use, biosensors were stored at 4 °C in phosphate buffer. To evaluate the analytical utility of the PbO2-based creatinine sensor system, the creatinine measurements in various control solutions containing typical interfering substances (i.e., ascorbic acid, uric acid, acetaminophen, and creatine) and human serum samples were undertaken in a differential setup. In this experiment, the creatinine and the creatine electrodes were constructed on the same chip. The differences between two signals (Idiff ) Icreatinine - Icreatine) were obtained using the aforementioned multichannel electrochemical system. RESULTS AND DISCUSSION The analytical range of most existing amperometric creatinine sensors based on the multiple-enzymatic reaction scheme is far Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
5967
Figure 2. Schematic drawings of the planar-type amperometric microelectrode and the differential creatinine sensor configuration.
from the physiological levels required in clinical applications. This is due to a loss in system sensitivity caused by the coupling of three enzymes and very low biological concentrations of creatinine.15c,29 Several attempts have been made to devise highly sensitive creatinine biosensors. They include the application of a poly(carbomoyl)sulfonate hydrogel as an enzyme-entrapment matrix15b and the use of a platinized shapable electroconductive polymer (polyanion-doped polypyrrole)-modified electrode.15c Warsinke et al. reported that an amperometric antibody-based creatinine sensor using the specific immunochemical interaction between creatinine and anti-creatinine antibodies exhibited a high sensitivity and a very low detection limit in nanomolar concentrations for creatinine measurements.30 However, these methods either require complex fabrication or measuring procedures or have a narrow detection range. Thus, our initial effort to develop a creatinine biosensor was devoted to improving its sensitivity and to achieving an extended analytical range necessary for pathological as well as normal specimens. Enzyme Layer. The realization of improvement in the sensor’s analytical range relies greatly on the method used for enzyme immobilization, the amount and ratio of enzymes employed, and the nature of the diffusion barrier (i.e., the outer protecting membrane). Conventional techniques for enzyme immobilization include physical entrapment into hydrogels15b,31 and electropoly(29) Killard, A. J.; Smyth, M. R. Trends Biotechnol. 2000, 18, 433-437. (30) Benkert, A.; Scheller, F.; Schossler, W.; Hentschel, C.; Micheel, B.; Behrsing, O.; Scharte, G.; Stocklein, W.; Warsinke, A. Anal. Chem. 2000, 72, 916921.
5968
Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
merized films,15a,16b,32 chemical cross-linking,15c,16a and surfacetopological processing.33 Among these methods, the use of a hydrogel system in which enzymes are entrapped into the aqueous environment provides distinctive advantages in terms of stability, reproducibility, dynamic range, and electrode fabrication without significant loss of enzyme activity. To construct an enzyme layer (3.3 mm2 in area; 30 µm in thickness) with the two- or threeenzyme system, appropriate amounts of enzymes, i.e., 0.3 unit/ mm2 CA, 0.3 unit/mm2 CI, and 0.3 unit/mm2 SO for a creatinine sensor; and 0.3 unit/mm2 CI and 0.3 unit/mm2 SO for a creatine sensor, respectively, were coimmobilized in a hydrogel network based on PVA. The above-specified units and ratios of enzymes were optimized through extensive experimentation. Protecting Membrane. Biosensors employing such a watersoluble PVA hydrogel as the enzyme-immobilizing matrix require an outer membrane to minimize enzyme leaching. A benefit of using an outer membrane is that it often serves as a diffusion barrier that can control analyte permeability, thereby extending the analytical range of sensing devices. Besides these essential utilities, the outer layer has another important function in the present work: it prevents direct contact between the enzyme layer (31) (a) Doretti, L.; Ferrara, D.; Gattolin, P.; Lora, S. Talanta 1997, 44, 859-866. (b) Sirkar, K.; Pishko, M. V. Anal. Chem. 1998, 70, 2888-2894. (32) (a) Scouten, W. H.; Luong, J. H.; Brown, R. S. Trends Biotechnol. 1995, 13, 178-185. (b) Iwuha, E. I.; de Villaverde, D. S.; Garcia, N. P.; Smyth, M. R.; Pingarron, J. M. Biosens. Bioelectron. 1997, 12, 749-761. (33) (a) Gooding, J. J.; Praig, V. G.; Hall, E. A. H. Anal. Chem. 1998, 70, 23962402. (b) Muguruma, H.; Hiratsuka, A.; Karube, I. Anal. Chem. 2000, 72, 2671-2675.
and an oxidizing film. Contact with H2O2 generated through biocatalytic reactions and oxidizing agents, i.e., PbO2, causes undesirable side reactions that could alter the working electrode current. Segmented polyurethanes, a family of block copolymers consisting of alternating hard- and soft-segment units, have attracted particular interest in many different fields, particularly in biomedical applications.34 HPUs are composed of soft blocks, e.g., PEG or PPG.34,35 The hydrophilic nature of HPUs is strongly dependent on the properties and compositions of the soft oligomers used. HPUs designed as medical-grade biomaterials have been exploited for formulating permselective films in carbonateand chloride-selective electrodes,28a,36 gas-permeable membranes in pCO2 and pO2 sensors,37 and enzyme-supporting layers in potentiometric biosensors.28b,38 These HPUs offer strong adhesion to various solid substrates, excellent biocompatibility, and short preconditioning time before use. Although several reports have demonstrated the usefulness of HPU-based membranes, there has been little study for their application to amperometric biosensing devices. Thus, we investigated the feasibility of using aliphatic polyether-based HPUs as a permselective barrier membrane for creatinine biosensors. In this work, three types of HPUs with different water uptake properties were prepared: i.e., 40% (HPUA), 100% (HPU-B), and 200% (HPU-C). The permeability to substrate and O2 may be varied depending on the variations in the composition of these HPUs (i.e., the PEG/PPG ratio). Figure 3 compares calibration curves for creatinine (A) and creatine (B) for creatinine sensors using different HPUs. As can be seen, the type of HPU used as an outer protecting membrane greatly influences the sensor slopes. In general, HPUs with less water uptake yield sensors with lower amperometric responses, possibly due to their decreased permeability. The use of HPU-A (curve a) as the protecting membrane resulted in a sensor with largely diminished creatinine response. For the sensor based on HPU-C (curve c), the calibration plot was reflected at high creatinine concentrations, i.e., at above 700 µM. In addition, detachment of the HPU-C film was observed frequently, owing to its poor adhesion and high internal osmotic pressure stemming from high water absorption. On the other hand, the creatinine sensor using HPU-B (curve b) showed an excellent amperometric response toward creatinine and provided improved stability and reproducibility. Interestingly, these HPU-based creatinine sensors exhibited significantly reduced responses toward creatine, when compared with responses toward creatinine: the creatinine/ creatine selectivities (∆Icreatinine/∆Icreatine) evaluated at 1000 µM concentration of each substance are 2.07 for HPU-A, 1.93 for HPUB, and 1.56 for HPU-C, respectively. This is unexpected considering that the generation of H2O2 from creatine requires fewer enzymatic steps than creatinine. The molecular weights of crea(34) (a) Ratner, B. D.; Hoffman, A. S.; Schoen, F. J.; Lemons, J. E. Biomaterials Science: An Introduction to Materials in Medicine; Academic Press: San Diego, CA, 1996; Chapter 7. (b) Lamba, N. M. K.; Woodhouse, K. A.; Cooper, S. L. Polyurethanes in Biomedical Applications; CRC Press: Boca Raton, FL, 1997. (35) Yilgor, I.; Yilgor, E. Polymer 1999, 40, 5575-5581. (36) Sakong, D. S.; Cha, M. J.; Shin, J. H.; Cha, G. S.; Ryu, M. S.; Hower, R. W.; Brown, R. B. Sens. Actuators, B 1996, 32, 161-166. (37) Choi, S. H.; Ha, J.; Shin, J. H.; Choi, Y. S.; Han, S. H.; Nam, H.; Cha, G. S. Anal. Chim. Acta 2001, 431, 261-267. (38) Cho, Y. A.; Lee, H. S.; Cha, G. S.; Lee, Y. T. Biosens. Bioelectron. 1999, 14, 435-438.
Figure 3. Calibration curves for creatinine (A) and creatine (B) for creatinine biosensors using three different HPUs as a protecting membrane: (a) HPU-A, (b) HPU-B, and (c) HPU-C.
tinine and creatine are 113.1 and 131.1, respectively; therefore, size is not a significant factor in the enhanced selectivity. A more likely explanation is found in the molecules’ ionic nature: creatinine is a positively charged site while creatine is a zwitterion near neutral pH. It is the authors’ experience30b that the permeation of anionic species through HPU films is more restricted than that of neutral species. Oxidizing Membrane. In an attempt to overcome the problem of interfering electroactive substances, we investigated the use of an additional oxidizing layer on the protecting membrane. Of the insoluble oxidizing agents testedsincluding PbO2, CeO2, and MnO2sPbO2 was found to exhibit the best performance in terms of the capability of eliminating interfering substances.39 To optimize the PbO2 layer formulation, creatinine sensors were prepared by using different amounts of PbO2 (i.e., 140, 200, and 260 mg/mL) in the HPU membrane (i.e., HPU-B). In this experiment, we found that 200 mg/mL PbO2 is the optimum amount for the proposed creatinine sensor system. A lower amount (i.e., 140 mg/mL) of PbO2 is not sufficient for eliminating the interfering responses, whereas a larger amount (i.e., 260 mg/ mL) of PbO2 reduces adhesion. Figure 4 shows dynamic responses toward creatinine and various interfering species (ascorbic acid, uric acid, acetaminophen) observed with two-layer (without PbO2 (39) Choi, S. H.; Lee, S. D.; Shin, J. H.; Ha, J.; Nam, H.; Cha, G. S., unpublished results.
Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
5969
Figure 4. Dynamic responses toward creatinine and various interfering species observed with two-layer (without PbO2 film; A) and three-layer (with PbO2 film; B) creatinine sensors: (a-e) 100, 200, 400, 700, and 1000 µM creatinine; (f) 228 µM ascobic acid; (g) 264 µM acetaminophen; and (h) 916 µM uric acid.
film; A) and three-layer (with PbO2 film; B) creatinine sensors. The concentrations of added interfering substances were approximately twice that of the highest levels normally present in physiological samples: (f) 228 µM ascorbic acid, (g) 264 µM acetaminophen, and (h) 916 µM uric acid. As can be seen, the two-layer sensor displayed a huge interfering response toward all oxidizable species. On the other hand, the three-layer creatinine sensor with the PbO2 film did not exhibit observable signal changes toward the tested concentrations. It should be noted, however, that the creatinine sensitivity of the three-layer sensor is somewhat lower than that of the two-layer sensor, owing to the presence of an additional membrane. The response behaviors of the optimized creatinine microsensor employing the three-layer system (i.e., enzyme-PVA/HPU/PbO2-HPU) in low-level creatinine concentrations are presented in Figure 5. The response time is relatively fast, i.e., t95% ) 118 s at 40 f 50 µM creatinine change, and t95% ) 98 s at 100 f 200 µM creatinine change. The linear range covers up to 1-1000 µM (correlation coefficient, r ) 0.9997), with a detection limit of 0.8 µM creatinine. From these observations, the electrochemical properties of the proposed creatinine sensor are considered to be adequate for use in physiological samples. Effect of Silanizing Reagent. It is well known that polymeric membranes incorporating a silanizing agent usually display much better adhesion and, thus, yield sensors with enhanced stability and lifetime.40 To investigate the effect of an adhesion promoter on sensor sensitivity, stability, and lifetime, we varied the amount of Z-6011 incorporated in both the protecting and oxidizing films of the planar creatinine sensor. Figure 6 compares the variation in the response slopes (A) and interfering signals (B) measured in the 1-1000 µM creatinine range and normal biological concentrations of interferences (i.e., 120 µM ascorbic acid, 140 (40) Shin, J. H.; Lee, J. S.; Choi, S. H.; Lee, D. K.; Nam, H.; Cha, G. S. Anal. Chem. 2000, 72, 4468-4473.
5970 Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
Figure 5. Dynamic response and calibration curves of the planar creatinine sensor employing the three-layer configuration in low-level creatinine concentrations. Initial and final concentrations are 1 and 150 µM, and the added concentration interval is 10 µM.
Figure 6. Variations in the response slopes (A) and the interfering signals (B) for planar creatinine sensors with varying amounts of Z-6011: (a) 0, (b) 12, (c) 24, and (d) 48 µL/mL.
µM acetaminophen, and 460 µM uric acid) for planar creatinine sensors with varying amounts of Z-6011: i.e., (a) 0, (b) 12, (c) 24, and (d) 48 µL/mL. It is evident that when 24 µL/mL Z-6011 was used, the creatinine sensor exhibited the best sensitivity and stability. In addition, the sensor showed no significant deterioration in its interference-removing capability. A gradual increase in
Table 1. Determination of Creatinine Levels in Various Control and Serum Samples with the Differential Creatinine Sensor Arrangementa,b
Figure 7. Creatinine measurements in various sample solutions with the PbO2-based differential creatinine microsensor system.
sensor slope over a period of 34 days may be attributed to the swelling of PVA hydrogel and to the increased pore size of HPU and PbO2-HPU membranes. The sensor with a lower amount of Z-6011 (i.e., 12 µL/mL) yielded unsatisfactory implementation owing to insufficient membrane adhesion. The sensor with a higher amount of Z-6011 (i.e., 48 µL/mL) quickly lost its sensitivity, even though the initial slope of this sensor was relatively high. This is believed to be due to the increased rigidity and porosity of HPU films by the increased amount of Z-6011 in HPU membranes: the pore size of different oxidizing membranes, 7.8 ( 1.2 for 0 µL/mL , 14.2 ( 4.3 for 12 µL/mL , 21.7 ( 4.8 for 24 µL/mL , and 40.1 ( 20.3 µm for 48 µL/mL Z-6011. Similarly, the greater interfering response of this sensor (see curve d in Figure 6B) may be explained by increased pore size that could decrease the effective surface area of the oxidizing layer. Differential Sensor. Although the HPU protecting layer reduces interference from endogeneous creatine, a differential creatinine sensor system was investigated to eliminate completely the errors caused by creatine present in physiological samples. Such a sensor arrangement consists of creatinine- and creatinesensing sites as shown in Figure 2. In this sensor design, the creatine sensor does not detect creatinine, while both sensors respond to creatine. Therefore, creatine signals cancel out in the differential measurement setup. The characteristics of this differential system were studied first by evaluating response behaviors of both creatinine and creatine sensors toward creatine. In the results, the creatinine and creatine sensor sites displayed a similar level of response signal toward added creatine, and this initial behavior was not changed significantly even after 8 days of use: signal deviations (∆I) at 200 µM creatine, 0.01 (first day), 0.14 (third day), 0.87 (eighth day), 0.23 (ninth day), and 2.01 nA (10th day); and at 500 µM creatine, 1.04 (first day), 1.29 (third day), 1.94 (eighth day), 2.54 (ninth day), and 8.27 nA (10th day). The planar differential creatinine-sensing device was fabricated on a single chip, and its analytical performance was examined. For this, various sample solutions were prepared by adding several interfering substances (i.e., ascorbic acid, uric acid, acetami-
sample type
composition of sample (or manufacturer’s specification)
creatinine value determined
control sample 1c control sample 2d normal serume abnormal serume
80 40 82.2 ( 8.8 495.1 ( 39.8
78.9 ( 3.0 39.1 ( 2.0 78.2 ( 5.0 486.1 ( 10.0
a In micromolar. b Number of samples, n ) 3. c Concentration levels of interfering species: ascorbic acid, 200 µM; acetaminophen, 200 µM; uric acid, 500 µM; and creatine, 80 µM. d Concentration levels of interfering species: ascorbic acid, 100 µM; acetaminophen, 100 µM; uric acid, 200 µM; and creatine, 40 µM. e Control human serum samples (Model Suitrol N or A) from Nissui Pharmaceutical (Tokyo, Japan).
nophen, and creatine) at upper clinical concentrations to creatinine standard solutions and by reconstituting control human serum. Figure 7 illustrates the results of the experiment performed with the proposed differential creatinine sensor system employing four standard solutions and two control sample solutions. As can be seen, the differential creatinine microsensor yields very fast response and recovery properties with stable signal output. The analytical results obtained for creatinine measurements with the proposed system are in excellent agreement with those of manufacturer’s specification (see Table 1). CONCLUSION In this report, we devised a clinically relevant planar creatinine biosensor by combining the use of an interference-removing oxidizing membrane with a differential amperometric sensing arrangement. The use of an HPU-based permselective barrier membrane yielded a sensor that exhibited an excellent creatinine response, while showing a significantly reduced interfering response toward creatine, with improved stability and reproducibility. A silanizing agent was utilized as an adhesion promoter to further enhance the stability and lifetime of the creatinine sensor. The creatinine biosensor devised in a differential sensing format that compensates the erroneous results from creatine is considered to be suitable for assay of serum specimens. ACKNOWLEDGMENT The authors gratefully acknowledge the financial support from the Korea Science and Engineering Foundation through the Center for Integrated Molecular Systems (POSTECH). This research was also supported by Dade Behring Inc. (Newark, DE), and in part by a grant from Kwangwoon University in 2001. Received for review May 1, 2001. Accepted September 26, 2001. AC010497A
Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
5971
