Comprehensive Study of Bioanalytical Platforms ... - ACS Publications

Nov 30, 2005 - For the past decade, highly specific analytical devices have been developed by coupling immobilized redox enzymes with electrochemical ...
0 downloads 13 Views 450KB Size
Anal. Chem. 2006, 78, 530-537

Comprehensive Study of Bioanalytical Platforms: Xanthine Oxidase E. Casero,† A. Martinez G. de Quesada,† J. Jin,‡ M. C. Quintana,† F. Pariente,† H. D. Abrun˜a,‡ L. Va´zquez,§ and E. Lorenzo*,†

Departamento de Quı´mica Analı´tica y Ana´ lisis Instrumental, Universidad Auto´ noma de Madrid, Campus de Cantoblanco, 28049 Madrid, Spain, Instituto de Ciencia de Materiales de Madrid (CSIC), C/Sor Juana Ine´ s de la Cruz, no. 3. 28049 Madrid, Spain, and Department of Chemistry and Chemical Biology, Baker Laboratory, Cornell University, Ithaca, New York 14853-1301

A comprehensive study of a general bioanalytical platform for biosensor applications is presented using xanthine oxidase (XnOx) as a case study within the framework of developing approaches of broad applicability. In this context, emphasis is placed on amperometric biosensors based on XnOx, which has been immobilized by covalent binding to gold electrodes modified with dithiobis-Nsuccinimidyl propionate. The immobilized XnOx layers have been characterized using atomic force microscopy under liquid conditions and quartz crystal microbalance techniques. In addition, spatially resolved mapping of enzymatic activity has been carried out using scanning electrochemical microscopy. Redox dyes of phenothiazine derivatives, specifically, thionine and methylene blue, have been found to work well as electron acceptors for reduced XnOx. The kinetic parameters and equilibrium constants of the mediated enzymatic oxidation of xanthine in the presence of the above-mentioned redox dyes have been calculated. The response of the enzymatic electrode to varying xanthine concentrations has been obtained in the presence of thionine or methylene blue as redox mediator in solution. Under these conditions, xanthine could be determined amperometrically at +0.2 V versus SSCE. For the past decade, highly specific analytical devices have been developed by coupling immobilized redox enzymes with electrochemical sensors.1 In comparison with other analytical techniques, such sensors possess simplicity of operation, high sensitivity, and selectivity. In addition, the immobilization of an enzyme has the advantage that a small amount of enzyme can be used for a large number of analytical determinations. Therefore, there is a great deal of continued interest in the development of enzyme immobilization methods that give rise to robust and successful biosensors. A crucial step in these studies is to know both the amount of immobilized enzyme and its conformation/ distribution on the surface. Thus, the characterization of the * To whom correspondence should be addressed. E-mail: encarnacion.lorenzo@ uam.es. † Universidad Auto´noma de Madrid. ‡ Cornell University. § Instituto de Ciencia de Materiales de Madrid. (1) Wang, J. Anal. Chem. 1995, 67, 487R-492R.

530 Analytical Chemistry, Vol. 78, No. 2, January 15, 2006

immobilization process, using techniques that allow quantification as well as visualization of the immobilized enzymes, would represent a great advantage. Among the techniques that have been employed to determine the mass deposited on a surface, the quartz crystal microbalance (QCM) has been shown to be most useful. In addition, this technique can be used not only for quantification but also to obtain kinetic information. Atomic force microscopy (AFM) has achieved particular relevance as a surface characterization technique, which provides morphological and mechanical information at the nanometer level. Since the development of tapping operation mode under aqueous environment,2 AFM has been able to provide morphological data of protein deposits without damaging the sample surface.3 In addition, AFM provides very valuable information on protein-protein and protein-surface interactions, through force spectroscopy measurements4 and on surface-induced conformational changes of proteins.5 The assessment of the enzymatic activity is also very important for the optimization of enzyme immobilization and functional parameters of sensors.6 Imaging of local activity of redox enzymes often requires the availability of electrochemically detectable compounds indicative of the biocatalytic reaction. In addition, the use of redox mediators with fast interfacial kinetics can enhance the sensitivity and reproducibility of the determinations.7 In this sense, scanning electrochemical microscopy (SECM) has provided a very sensitive and versatile tool for determining the local activity of immobilized enzymes.8 In the case of oxidases, mapping of enzymatic activity can be achieved either by detecting the enzymatic reaction product (hydrogen peroxide) or by following changes in the steady-state current of a mediator as a consequence of the enzymatic reaction. Among the main advantages of the SECM detection, one can highlight its suitability to detect molecules in very small volumes9 and that it does not require an (2) Putman, C. A.; Van der Werf, K. O.; De Grooth, B. G.; Van Hulst, N. F.; Greve, J. Appl. Phys. Lett. 1994, 64, 2454-2456. (3) Forbes, J. G.; Jin, A. J.; Wang, K. Langmuir 2001, 17, 3067-3075. (4) Feldman, K.; Ha¨hner, G.; Spencer, N. D.; Harder, P.; Grunze, M. J. Am. Chem. Soc. 1999, 121, 10134-10141. (5) Moulin, A. M.; O’Shea, S. J.; Badley, R. A.; Doyle, P.; Welland, M. E. Langmuir 1999, 15, 8776-8779. (6) Csoka, B.; Kova´cs, B.; Nagy, G. Electroanalysis 2003, 15, 1335-1342. (7) Pierce, D. T.; Unwin, P. R.; Bard, A. J. Anal. Chem. 1992, 64, 1795-1804. (8) Wittstock, G. Fresenius J. Anal. Chem. 2001, 370, 303-315. (9) Fan, F. R.; Bard, A. J. Science 1995, 267, 871-874. 10.1021/ac051676l CCC: $33.50

© 2006 American Chemical Society Published on Web 11/30/2005

incubation time period since the products of the enzymatic reaction are detected directly in proximity to the target.10 In addition, the biorecognition reactions can be localized in very small areas by micropattering techniques,11,12 which not only decrease reagent consumption but also provide the basis for multianalyte assays. The characterization studies described above are preliminary and essential steps in the design of biosensors based on (redox) enzymes. In this sense, the goal of this work was to carry out a comprehensive characterization of a general bioanalytical platform for biosensor applications using xanthine oxidase (xanthine oxygen oxidoreductase; XnOx) as a case of study with the purpose of developing approaches of broad applicability. XnOx is a complex metalloprotein involved in purine catabolism, oxidizing hypoxanthine to xanthine and xanthine to uric acid, with the concomitant reduction of oxygen.13-15 In addition, the enzyme catalyzes the hydroxylation of a large number of nitrogencontaining heterocyclic compounds and the oxidation of many aldehydes. XnOx-modified electrodes have been frequently used for the determination of biological purines, in particular, hypoxanthine,16-18 a major metabolite in the degradation of adenine nucleotide, which is found to accumulate in fish and beef. As a result, the level of hypoxanthine is generally used in the food industry as an index of freshness. Most of these reported sensors are generally based on the determination of either consumed oxygen or enzymatic reaction products, i.e., peroxide19,20 or uric acid.21 Electrical communication/connection between redox proteins and the electrode surfaces provides a general means of enhancing the activity of the redox-active biocatalyst.22 Direct contact between the protein’s redox center and the electrode interface is, however, generally ineffective due to the insulation of the active site by the protein matrix. Various methodologies have been employed to enhance the interaction between redox proteins and electrodes. One of the more common approaches involves the use of redox mediators in solution. In addition, mediated oxidase-based biosensors frequently operate at much lower potentials than those based on the determination of an enzymatic reaction product, i.e., peroxide. In solution, electron-transfer mediators operate by a diffusional route facilitating electrical contact between the enzyme’s redox center and the electrode surface. In the case of oxidoreductases, it has been shown that both, one- and twoelectron, mediators can be employed. Typical of the first would be ferrocenyl derivatives and hexacyanoferrates. In the latter case, (10) Shiku, H.; Matsue, T.; Uchida, I. Anal. Chem. 1996, 68, 1276-1278. (11) Turyan, I.; Matsue, T.; Mandler, Anal. Chem. 2000, 72, 3431-3435. (12) Wilhelm, T.; Wittstock, G. Langmuir 2002, 18, 9485-9493. (13) Harris, C. M.; Massey, V. J. Biol. Chem. 1997, 272, 28335-28341. (14) Krenitsky, T. A.; Neil, S. M.; Elion, G. B.; Hitchings, G. H. Arch. Biochem. Biophys. 1972, 150, 585-599. (15) Rosemeyer, H.; Seela, F. Eur. J. Biochem. 1983, 134, 513-515. (16) Lanqun, M.; Fang, X.; Qi, X.; Litong, J. Anal. Biochem. 2001, 292, 94-101. (17) Cayuela, G.; Pen ˜a, N.; Reviejo, A. J.; Pingarro´n, J. M. Analyst 1998, 123, 371-377. (18) McKenna, K.; Brajter-Toth, A. Anal. Chem. 1987, 59, 954-958. (19) Hu, S.; Xu, C.; Luo, J.; Cui, D. Anal. Chim. Acta 2000, 412, 55-61. (20) Rehak, M.; Snejdarkova, M.; Otto, M. Biosens. Bioelectron. 1994, 9, 337341. (21) Gonzalez, E.; Pariente, F.; Lorenzo, E.; Herna´ndez, L. Anal. Chim. Acta 1991, 242, 267-273. (22) Barlett, P. N.; Tebbut, P.; Whitaker, R. G. Prog. React. Kinet. 1991, 16, 55-60.

phenazine and phenothiazine derivatives, quinones/hydroquinones and related materials are often used. Recently, some redoxmediated XnOx-based sensors have been developed.23-25 In these previous studies, ferrocene and its derivatives as well as prussian blue have been employed as redox mediators either in solution23,24 or incorporated in the carbon paste electrodes.25 Detection limits in the micromolar regime were reported for both theophylline and hypoxanthine employing well-known immobilization schemes, such as enzyme retention behind a dialysis membrane, enzyme cross-linked with glutaraldhehyde on a porous Nylon membrane, or incorporation into electropolymerized pyrrole films. However, the drawbacks of the use of such enzyme immobilization schemes are well known and documented. In the present work, we have focused our attention on XnOxbased biosensor applications using artificial redox mediators. Specifically, phenothiazine dyes thionin and methylene blue as well as hydroxymethylferrocene have been employed as electrontransfer mediators in solution. As we have previously mentioned, the aim of this work is to describe bioanalytical sensor platforms based on oxidoreductase enzymes. In this sense, the work involves (1) the study of different artificial redox mediators that can be used as electron acceptors for XnOx and evaluation of their equilibrium constants and kinetics parameters, (2) the immobilization of XnOx by covalent binding to gold electrodes modified with dithiobis-N-succinimidyl propionate (DTSP) as a general reagent for a covalent enzyme immobilization, (3) the characterization of the resulting enzyme layer by QCM and AFM, (4) the assessment of the enzymatic activity of immobilized XnOx by SECM, and (5) the study of the enzyme electrode response to xanthine using thionin and methylene blue as solution redox mediators. EXPERIMENTAL SECTION Materials. XnOx (EC1.1.3.22) from buttermilk was commercially available from Sigma Chemical Co. (St. Louis, MO) and used without further purification. Stock solutions were prepared by dilution and stored at 4 °C. Under these conditions, the enzymatic activity remains stable for several weeks. Xanthine, thionin, methylene blue, hydroxymethylferrocene, 3,3′-dithiodipropionic acid di(N-succinimidyl ester) (DTSP), and dimethyl sulfoxide were obtained from Aldrich Chemical Co. (Milwaukee, WI) and were used as received. All other chemicals were of at least reagent grade quality and were used as received. Sodium phosphate (Merck) was employed for the preparation of buffer solutions (0.1 M, pH 6.5). Water was purified with a Millipore Milli-Q-System. All solutions were prepared just prior to use. Apparatus and Procedures. Spectrophotometric Measurements. Changes in the absorbance were measured for 0.1 mM solutions of the oxidized form of thionin or methylene blue after addition of xanthine to a 0.1 M pH 6.5 phosphate buffer solution containing 0.2 unit of XnOx. Absorbance measurements were carried out at 25 °C using a Shimadzu UV-1700 spectrophotometer and a quartz cell having a light path length of 1 cm. Before adding the xanthine solution, the phosphate buffer solution was deoxygenated by bubbling nitrogen for 20 min. (23) Liu, Y.; Nie, L.; Tao, W.; Yao, S. Electroanalysis 2004, 16, 1271-1277. (24) Luong, J. H. T.; Thatipamala, R. Anal. Chim. Acta 1996, 319, 325-333. (25) Stredansky, M.; Pizzariello, A.; Miertus, S.; Svorc, J. Anal. Biochem. 2000, 285, 225-229.

Analytical Chemistry, Vol. 78, No. 2, January 15, 2006

531

Quartz Crystal Microbalance Measurements. AT-cut quartz crystals (5.0 MHz) of 25-mm diameter with Au electrodes deposited over a Ti adhesion layer (Maxtek Inc., Santa Fe Springs, CA) were used for QCM measurements. An asymmetric keyhole electrode arrangement was used, in which the circular electrode geometrical areas were 1.370 (front side) and 0.317 cm2 (backside). The electrode surfaces were overtone polished. Prior to use, the quartz crystals were cleaned by immersion in piranha solution, H2SO4/H2O2 (3:1 v/v). Caution: Piranha solution is extremely reactive. They were subsequently rinsed with water and dried in air. The quartz crystal resonator was set in a probe (TPS-550, Maxtek) made of Teflon in which the oscillator circuit was included, and the quartz crystal was held vertically. The probe was connected to a cell by a homemade Teflon joint, which was immersed in water-jacketed beaker thermostated at the assay temperature (25.0 ( 0.1 °C) with a thermostatic bath (Digital Temperature Controller Haake F6). The frequency was measured with a plating monitor (PM-740, Maxtek Inc.) and simultaneously recorded by a personal computer. Atomic Force Microscopy Measurements. Supports used for protein attachment consisted of glass substrates (1.1 × 1.1 cm) covered with a chromium adhesion layer (1-4 nm thick) on to which a gold layer (200-300 nm thick) was deposited (Arrandee, Germany). Prior to use, the gold surfaces were annealed for 2 min in a gas flame in order to obtain Au (111) flat terraces. The atomic force microscope was a Nanoscope IIIa (Veeco, Santa Barbara, CA) used with a D scanner (maximal scan range of ∼14 µm). Measurements were carried out, under liquid environment, with Si3N4 cantilevers (nominal radius of 20 nm and spring constant of 0.38 N/m) from Veeco. The samples were imaged with different cantilevers in order to ensure that the imaged structures were not due to tip artifacts. The substrate was first imaged in buffer solution to ensure that the surface was flat and clean before carrying out the protein immobilization step. Both sample and cantilever were located within a Plexiglas fluid cell into which ∼50 µL of buffer was added. For morphological imaging of the different protein deposits and substrates, we employed tapping mode AFM microscopy. In this mode, the cantilever is oscillated at a frequency close to its resonant frequency, f0, in the ∼8.5-10.5 kHz range with an amplitude A0 (∼15 nm) above the surface. As the cantileversample distance decreases, the probe interaction with the sample surface increases, leading to a decrease of the cantilever amplitude, A (A < A0). Thus, the lower the A/A0 ratio, the higher the average applied force on the sample.26 Accordingly, one can measure, at a fixed spot, the change of A with the relative tipsample distance. These curves, when compared with those obtained with the same tip on a hard substrate, can provide an estimation of the indentation induced by the tip on the imaged material.27 Enzyme Electrode Response. The enzyme electrode was immersed in a stirred pH 6.5, 0.1 M phosphate buffer solution in a three-compartment electrochemical cell with standard taper joints so that all compartments could be hermetically sealed with Teflon adapters. A large-area coiled platinum wire was employed as a counter electrode. All potentials are reported against a sodium(26) Garcı´a, R.; Pe´rez, R. Surf. Sci. Rep. 2002, 47, 197-301. (27) Kopp-Marsaudon, S.; Lecle`re, Ph.; Dubourg, F.; Lazzaroni, R.; Aime´, J. P. Langmuir 2000, 16, 8432-8437.

532

Analytical Chemistry, Vol. 78, No. 2, January 15, 2006

saturated calomel electrode (SSCE) without taking into account the liquid junction. All solutions were deaerated with nitrogen gas before use, and the gas flow was kept over the solution during experiments. Cronoamperometric measurements were carried out by poising the enzyme electrode at a potential of +0.2 V, where oxidation of the mediator is assured. After a steady-state background current baseline was obtained, aliquots (typically 100 µL) of a xanthine stock solution (typically 5 mM) were added. After the mixture was stirred for 30 s and allowed 2 min for equilibration, the steady-state current in the unstirred solution was recorded. SECM Enzyme Assay Procedure. SECM measurements were carried out using CH Instrument model 900B equipment. The setup is formed by an electrochemical cell located on an XYZ positioning stage. The distance between the tip and the sample was controlled using a stepper motor (coarse approach) combined with a piezoelectric micropositioning system (fine approach). The three- or four-electrode cell was formed by a Ag/AgCl reference electrode, a platinum wire auxiliary electrode, a substrate (highly ordered pyrolytic graphite), and a Pt ultramicroelectrode with a 10-µm radius, usually termed the tip or probe electrode. Substrates were previously modified with the enzyme and subsequently mounted in the cell. Measurements were performed in the feedback mode (probe potential E ) + 0.4 V, sample at open circuit, lateral scan speed vt ) 15 µm s-1) in pH 6.5, 0.1 M phosphate buffer + 0.1 M KCl containing 1 mM hydroxymethylferrocene and 3 mM xanthine. Previously, the assay solution was thoroughly deoxygenated in the SECM cell. RESULTS AND DISCUSSION Spectrophotometric Study of Redox Mediators. Evaluation of Equilibrium Constants and Kinetic Parameters. Based on previous studies on artificial electron acceptors for oxidase enzymes, which, like XnOx have an FAD active redox center,28 we have examined the ability of several phenothiazine dyes, with emphasis on thionin and methylene blue, to accept electrons from the reduced XnOx. To carry out these studies, the absorption spectra of 0.1 mM solutions of the dye in pH 6.5, 0.1 M phosphate buffer containing 0.2 unit of XnOx were recorded (data not shown). Upon addition of xanthine (0.1 mM), a dramatic decrease in absorbance with reaction time was observed. In the case of thionin, the absorption peak at 600 nm decreased rapidly, and after 30 s, the spectrum was that of the known reduced form of thionin. However, a qualitatively similar behavior was observed for methylene blue except the final spectrum was reached after ∼60 s. These observations were consistent with the well-known ping-pong mechanism:

XnOx (ox) + S T XnOx-S T XnOx (red) + P

(1)

XnOx (red) + Medox T XnOx-Med T XnOx (ox) + Medred (2)

From such experiments we have obtained the time course of changes in the concentration of the oxidized form of the dye after addition of increasing concentrations of xanthine ranging from (28) Nakaminami, T.; Kuwabata, S.; Yoneyama, H. Anal. Chem. 1997, 69, 23672372.

Figure 1. (A) Time course of changes in the concentration of electron acceptors (thionin and methylene blue) in oxidized forms after addition of 0.4 µM xanthine in pH 6.5, 0.1 M phosphate buffer solution containing 0.2 unit of XnOx and 0.6 µM thionin or 0.5 µM methylene blue in initial concentration. (B) Lineweaver-Burk plots obtained from data presented in (A).

Table 1. Equilibrium Constants and Kinetic Parameters of XnOx

thionin methylene Blue

νmax (µmol min-1)

KM (µM)

Keq

τ (s)

125 55

0.325 0.077

6800 454

18 25

0.1 to 1.2 µM and a concentration of thionin and methylene blue of 0.6 and 0.5 µM, respectively. Figure 1A shows representative plots for both mediators at a xanthine concentration of 0.4 µM. As can be seen, the dye concentrations decrease with time. Given that the product of the enzymatic reaction is uric acid, the following equilibrium applies: XnOx

xanthine + Medox 79 8 uric acid + Medred K eq

(3)

where Keq is the equilibrium constant. The values of Keq were calculated for both thionin and methylene blue by determining, from the absorbance measurements, the concentrations of Medox and Medred once equilibrium has been reached. The values obtained are presented in Table 1. This table also includes kinetic parameters of the maximum rate of xanthine oxidation (Vmax) and the Michaelis constant for xanthine (KM) evaluated from the corresponding Lineweaver-Burk plots (see Figure 1B) of v-1 versus Cs-1, where Cs is the concentration of xanthine.29,30 The time employed to achieve equilibrium (τ) is also given in Table 1. This parameter is defined as the reaction time required to consume (1/n)∆CMedox, where n is a natural number and ∆CMedox is the difference between the initial concentration of the oxidized form of the mediator (Medox) and the corresponding value at equilibrium.31 As can be observed in Table 1, both Vmax and Keq, which reflect the ability of the two redox mediators to accept electrons from (29) Segel, I. H. Enzyme Kinetics; Wiley-Interscience: New York, 1993; Chapter 9-I. (30) Dixon, M.; Webb, E. C. Enzymes; Longman: London, 1979; Chapter 4. (31) Atkins, P. W. Physical Chemistry; Oxford University Press: New York, 1990; Chapter 26.

XnOx, are smaller for methylene blue than for thionin. Since thionin and methylene blue have very similar E0′ values (-0.21 and -0.22 V in 0.02 M pH 7.0 phosphate buffer solution, respectively) one would not, a priori, anticipate any differences in the ability of the mediators to accept electrons from XnOx. Thus, these results suggest that other effects, such as size or hydrophobicity of the redox mediators, may also play an important role in determining their reactivity. Although largely speculative in our part, the differences might arise from the enhanced ability of thionin to engage in hydrogen bonding. A similar effect has been previously reported by Nakaminami et al.28 in the case of cholesterol oxidase and several phenazines and phenothiazine derivatives, including thionin and methylene blue. Development of Xanthine Bioanalytical Platforms. One of the objectives of this investigation was the development of bioanalytical platforms based on XnOx. For this purpose, in a first step, we attempted the immobilization of XnOx on modified gold electrode surfaces. As we have previously described,32 gold electrodes can be modified with a monolayer of DTSP through the disulfide group, so that terminal succinimidyl groups allow further covalent immobilization of enzymes or other materials including XnOx. To characterize the XnOx immobilization process on DTSP-modified gold substrates, we have employed several techniques as described below. To determine the amount of XnOx immobilized on the DTSPmodified gold electrode, QCM studies were carried out. QCM measurements allow simultaneous acquisition of kinetic information about the binding process, as well the surface coverage of the resulting XnOx layer. The immobilization process was carried out, as described above, in a homemade cell coupled to the QCM probe containing the DTSP-modified Au quartz resonator in contact with 10 mL of buffer solution. After the temperature and the frequency were stabilized, an aliquot of XnOx stock solution was added to the cell in order to obtain a final concentration of protein of 1.8 µM. As can be seen in Figure 2A, upon addition of (32) Darder, M.; Takada, K.; Pariente, F.; Lorenzo, E.; Abrun ˜a, H. D. Anal. Chem. 2000, 72, 3784-3792.

Analytical Chemistry, Vol. 78, No. 2, January 15, 2006

533

Figure 2. Time dependence of the frequency changes of a DTSPgold substrate resonator (A) and a bare gold substrate (B) in pH 6.5, 0.1 M phosphate buffer solution upon addition of XnOx to a final concentration of 1.8 µM.

XnOx, the frequency decreased slowly during 40 min until a steady state was reached. Assuming that the frequency decrease (∆F) is only due to the change in mass arising from the adsorption of the enzyme, one can calculate the amount and surface coverage, Γ, of the immobilized XnOx by using the Sauerbrey equation33

∆m ) -Cf∆F

(4)

where ∆m is the mass change (ng cm-2) and Cf (17.7 ng Hz-1 cm-2) is a proportionality constant for the 5.0-MHz crystals used in this study. Assuming a molecular mass of 290.000 Da for XnOx,34 a surface coverage value of ∼1.4 × 10-12 mol cm-2 was obtained. It should be noted that, due to the asymmetric shape of the protein (see Figure 3), the theoretical surface coverage estimated from a monolayer of XnOx according to its molecular size would be in the range of 1.2 × 10-12-2.7 × 10-12 mol cm-2 depending on orientation and packing in the monolayer. As a comparison to the above measurements, QCM experiments were also performed to study the direct adsorption of XnOx onto a bare gold quartz crystal resonator under the same conditions as described above for DTSP-modified resonators. As (33) Sauerbrey, G. Z. Phys. 1959, 155, 206-222. (34) Enroth, C.; Eger, B. T.; Okamoto, K.; Nishino, T.; Nishino, T.; Pai, E. F. Proc. Natl. Acad. Sci. USA 2000, 97, 10723-10728.

534 Analytical Chemistry, Vol. 78, No. 2, January 15, 2006

Figure 3. Scheme of XnOx. The most exposed lysine groups are marked in pink. Table 2. First-Order Rate Constant (k), Frequency Change (∆Fmax), Mass of Xanthine Oxidase (m), and Surface Coverage (Γ) for the Immobilization Process of Xanthine Oxidase on a Gold Surface and on a DTSP-Gold Substrate

XnOx-DTSP-Au XnOx-Au

k (min-1)

∆Fmax (Hz)

m (ng cm-2)

Γ × 1012 (mol cm-2)

0.1 1.0

-23.90 -39.01

423 690

1.4 2.3

can be seen in Figure 2B, upon addition of XnOx, a rapid decrease in frequency was observed during the first 2 min and afterward a steady state was reached. These observations indicate that the enzyme adsorbs to bare gold surfaces. The shapes of the frequency time plots for bare and DTSP-modified surfaces are clearly different as can be ascertained from Figure 2 with the decay being faster for the bare gold surface. In addition, the surface coverages for bare gold surfaces were consistently larger than for DTSP-modified surfaces (Table 2). This could reflect differences in packing arising from the interaction of XnOx with the bare and DTSP-modified surfaces. As mentioned above, the shape of the frequency-time profile can be employed to study the kinetics of adsorption. The process can be controlled by either transport (diffusion) or by kinetics (activation controlled), which predict time dependencies of t1/2 and exp(t), respectively. Fits to both transport and kinetically controlled models were carried out, and the latter gave consistently better results. Assuming that the immobilization process is kinetically controlled, the data were fit (Figure 2A and B solid

line) to a first-order kinetics equation: ∆F ) ∆Fmax (1 - e-kt), where ∆F is the frequency change (in Hz), ∆Fmax is the frequency change between the initial and the final steady-state values, and k is the first-order rate constant (expressed in min-1). The values of ∆Fmax and k obtained from the fits are summarized in Table 2. It is evident from the different values obtained for the constant (k) that the frequency decreases more slowly when XnOx is immobilized via covalent binding than through direct adsorption. The difference in the kinetics of adsorption and in the final surface coverage may be explained based on potentially different mechanisms involved in the immobilization of XnOx on DTSP-modified and bare gold surfaces. On bare gold, the adsorption of the enzyme ostensibly takes place with a random distribution of protein adsorption geometries. However, when the immobilization of XnOx takes place through DTSP, the process involves a nucleophilic attack of the primary amino groups present in the enzyme to the terminal N-succinimidyl esters exposed to the solution in the monolayer. Therefore, in this case, the enzyme immobilization will occur only for particular orientations (see Figure 3) in which the amino groups are exposed to the substrate. Thus, one would anticipate slower kinetics than in the case of direct enzyme adsorption, which occurs in a nonspecific fashion. To further characterize the nature of the resulting XnOx layers obtained either by direct adsorption or covalently bonded through DTSP, and to establish the differences between them, we have carried out AFM experiments that provide morphological information that complement the results obtained by QCM. The advantage of AFM is that it allows characterization of the enzyme monolayer at the local (i.e., nanoscale) level. Panels A and B in Figures 4 show 800 nm × 800 nm images of a DTSP-modified gold substrate and a bare gold substrate after 10 min of adsorption of XnOx. From Figure 4, it is evident that the surface coverage obtained for the DTSP-modified surface is smaller than that obtained for the bare gold surface. In addition, closer inspection of Figure 4B reveals brighter spots, which correspond to XnOx molecules adsorbed on top of an apparently homogeneous layer in which globular structures are also observed. These results are qualitatively consistent with the surface coverage calculated from QCM plots (see Figure 2). Note that the typical diameter of the globular structures obtained by AFM is larger than the reported dimensions of XnOx.34 This is likely due to tip convolution effects since the tip radius is similar to the enzyme size. These effects become more evident for DTSPmodified surfaces because the enzyme molecules appear isolated on top of the flat substrate. In contrast, for bare gold, due to the monolayer enzyme packing, the measured enzyme size is smaller. In this case, the tip starts to image the next enzyme molecule before reaching the underlying substrate because the proteins are very close to each other. This relatively high enzyme packing precludes a more detailed analysis of the height of the enzyme structures. However, such analysis can be carried out for DTSPmodified surfaces since the isolated adsorption of the enzymes on the flat substrate allows us to determine their height. Note that the highlighted protein in Figure 4B, which lies clearly isolated on top of the protein layer, shows sizes similar to those observed in Figure 4A. The corresponding height histogram obtained from Figure 4A is shown in Figure 5. A relatively wide height distribution is observed, ranging from 5 to 24 nm. The

Figure 4. The 800 nm × 800 nm tapping mode AFM image taken under buffer conditions of XnOx layer adsorbed on a DTSP-modified gold substrate (A) and a bare gold substrate (B).

Figure 5. Height histogram obtained from Figure 4A.

largest values (>17 nm), which correspond to 25% of the observations, are likely related to aggregated structures since the protein dimensions are about 17, 7, and 9 nm. The rest of the features, with heights in the 5-17-nm range, represent 75% of the observations, displaying a maximum for heights in the 1415-nm range. These values, taking into account the likely deformation of the enzyme when imaged by AFM, are in agreement with the protein dimensions. In fact, the maximum at 14-15 nm suggests a predominant geometry of adsorption with the longest protein axis perpendicular to the substrate. The DTSP-modified surfaces allow further analysis of the distribution of the enzyme on the flat substrate and the protein deformation induced by the tip during AFM imaging. First, from the AFM image, one can obtain its autocovariance (Figure 6A). This figure shows an outer diffuse bright circle (relative maximum) containing both a black ring (minimum) and a bright central Analytical Chemistry, Vol. 78, No. 2, January 15, 2006

535

Figure 7. Approach AFM amplitude versus distance curves obtained on XnOx-Au and XnOx-DTSP-Au monolayers and Au surfaces in a liquid environment with the same cantilever. The dotted vertical line indicates the tip-sample surface first contact point.

Figure 6. (A) Autocovariance of the AFM image shown in Figure 4A. (B) Profile measured along the dotted line.

spot (global maximum). In Figure 6B is shown a profile of the autocovariance in which a central absolute maximum flanked by two smaller maximums is observed. The distance between the central maximum and the smaller maximums corresponds to the average distance between next protein neighbors. In our case, this value is in the 70-95-nm range, which is larger than that corresponding to that defined by two adjacent enzyme molecules (15-35 nm). These results support that during the adsorption process the XnOx molecules do not tend to adsorb close each other, but rather leave a relatively large gap between them. As we have mentioned previously, both isolated protein structures and flat substrate regions are observed in Figure 4A. Thus, we can obtain amplitude versus relative tip-surface distance curves (Figure 7) on both regions with the same tip. From Figure 7, it is clear that the slope of the curve obtained on XnOx structures is smaller than that corresponding to the DTSP-Au substrate. Also, the slope of the curve obtained on DTSP-Au substrates is smaller than that of the bare Au substrates. This result implies that the protein is deformed by the tip action to a larger extent than the DTSP layer. Furthermore, this deformation can be quantified by measuring the horizontal shift between both curves for a given A/A0 ratio. Thus, for A/A0 ) 0.75, the deformation obtained is close to 5 nm, whereas for A/A0 ) 0.92, it is close to 2-3 nm. It should be noted that for standard AFM topographical imaging we employ A/A0 ratios in the 0.9-0.93 range. Mapping of Local Enzyme Activity by Scanning Electrochemical Microscopy. To assess the reactivity of immobilized XnOx and to map its reactive sites, SECM experiments, using hydroxymethylferrocene as redox mediator, were carried out. Previous SECM studies based on oxidoreductases, especially glucose oxidase,7 have shown that in the feedback mode, the redox mediator, whose active form is generated at the SECM tip, 536 Analytical Chemistry, Vol. 78, No. 2, January 15, 2006

can be regenerated by the immobilized enzyme so that the feedback current provides a direct measurement of enzymatic activity. For systems with high catalytic rates, the SECM current will clearly show a positive feedback response, whereas for a slow process, a negative feedback response will be observed. We have carried out analogous studies with immobilized XnOx using hydroxymethylferrocene as redox mediator. Although thionin and methylene blue are two-electron mediators, it is well known that XnOx can also be catalyzed by one-electron mediator such as hydroxymethylferrocene. We have employed hydroxymethylferrocene as mediator rather than thionin or methylene blue because of the propensity of the latter two to strongly adsorb. Such adsorption could affect the measured response in an unpredictable way, and thus, we used hydroxymethylferrocene as mediator even though we recognize that it is not the most efficient. Figure 8B shows the activity of an XnOx layer over an area 50 × 50 µm2 with hydroxymethylferrocene acting as mediator after addition of xanthine to the solution. As can be observed, the tip current measured by the SECM exhibits clear positive feedback behavior. In contrast, no positive feedback was observed in the absence of xanthine. From the difference between both images, the spatially resolved enzymatic activity can be assessed. The variations in feedback current, which would correspond to variations in enzymatic activity, can be understood in terms of the spatial distribution of the enzyme and its activity. As mentioned in the Experimental Section, we employed direct adsorption for enzyme immobilization. This method often gives rise to a nonuniform spatial distribution of enzymes on surfaces, especially in the monolayer regime, as was the case in this study. In addition, it is often found that immobilized enzyme layers have nonuniform activity, reflecting variations in the specifics of adsorption (and conformation) for particular locations. Thus, such spatially inhomogeneous enzymatic activity would, in fact, be anticipated, as we indeed observed. Enzyme Electrode Response. In our configuration, the immobilized XnOx oxidizes xanthine to uric acid in the presence of thionin/methylene blue, which act as acceptors of electrons generated in the enzymatic reaction and are transformed to their reduced form. The mediator, in turn, diffuses to the electrode, where it is reoxidized back to its oxidized form. The electrode

Figure 8. SECM surface plot image (50 × 50 µm2) of immobilized XnOx. Image was taken with a platinum microelectrode tip (10 µm, rate 15 µm s-1). Positive feedback with hydroxymethylferrocene mediator (1 mM) at ET ) +0.4 V vs Ag/AgCl in pH 6.5, 0.1 M phosphate buffer solution in the absence (A) and in the presence of 3 mM xanthine (B).

acts as a secondary acceptor of electrons able to regenerate the redox mediator used in the enzymatic reaction. This sequence of redox events represents a catalytic process. Thus, the magnitude of this catalytic current can be employed as the analytical signal in the determination of the substrate (xanthine) concentration. The steady-state current response obtained at +0.2 V for either thionin or methylene blue was plotted as a function of the bulk concentration of xanthine in solution. The response was linear to xanthine concentration over the range of 5.0 × 10-5-8.0 × 10-4 M for both redox mediators. In the case of thionin, the response (y ) 0.002 + 130x; r ) 0.99) exhibited a good sensitivity (130 nA/mM). We estimate a limit of detection of ∼30 µM. Methylene blue (y ) 0.001 + 28x) exhibited a low sensitivity (28 nA/mM). The high sensitivity exhibited by thionin compared to that obtained with methylene blue agrees well with the results obtained from the spectrophotometric studies reported above. Concerning the stability of immobilized XnOx layer, it was found that the biosensor response decreases ∼30% of its initial value after one assay and then it remains constant for more than one month.

value of 1.4 × 10-12 mol cm-2. This value represents ∼50% of a compact monolayer, assuming an orientation with the longest enzyme axis perpendicular to the substrate. From QCM studies, we have also been able to determine that the kinetics of absorption is activation controlled with a value of 0.1 min-1 for the rate constant. AFM measurements have provided morphological characterizations that complement the results obtained by QCM. Moreover, the AFM studies provided data concerning the protein distribution on the substrate and the tip-induced deformation during AFM imaging. AFM measurements showed that on DTSPmodified surfaces XnOx molecules appear to adsorb in clusters or domain separated by 70-95-nm gaps. Thionin and methylene blue have been used as electron acceptors for reduced XnOx, and the kinetic parameters and equilibrium constants of the mediated enzymatic oxidation of xanthine have been calculated. Electrochemical and scanning electrochemical microscopy assays indicate that the immobilized enzyme retains its enzymatic activity after immobilization. Finally, a xanthine biosensor was developed employing thionin.

CONCLUSIONS A comprehensive study of a general bioanalytical platform for biosensor applications has been described. In particular, emphasis has been placed on amperometric biosensors based on XnOx, which was immobilized by covalent binding to gold electrodes modified with DTSP. This material provides a general way of covalent enzyme immobilization. The amount of immobilized enzyme has been estimated from QCM measurements giving a

ACKNOWLEDGMENT This work was supported by the Ministerio de Ciencia y Tecnologı´a of Spain through BQU2001-0163 and by the Universidad Auto´noma de Madrid through Proyecto Emergente 2005. Received for review September 20, 2005. Accepted November 1, 2005. AC051676L

Analytical Chemistry, Vol. 78, No. 2, January 15, 2006

537