Covalent Immobilization of β-Galactosidase onto a Gold-Coated

Covalent Immobilization of β-Galactosidase onto a Gold-Coated Magnetoelastic Transducer via a Self-Assembled Monolayer: Toward a Magnetoelastic ...
0 downloads 0 Views 100KB Size
Anal. Chem. 2003, 75, 6932-6937

Covalent Immobilization of β-Galactosidase onto a Gold-Coated Magnetoelastic Transducer via a Self-Assembled Monolayer: Toward a Magnetoelastic Biosensor J. Christopher Ball, Libby G. Puckett, and Leonidas G. Bachas*

Department of Chemistry, University of Kentucky, Lexington, Kentucky 40506-0055

The enzyme β-galactosidase has been covalently immobilized onto a gold-coated magnetoelastic film via a selfassembled monolayer (SAM) of ω-carboxylic acid alkylthiol. Use of magnetoelastic transduction allows for the wireless monitoring of enzymatic activity through the associated change in the frequency and amplitude of magnetic fields. The formations of SAMs of 3-mercaptopropanoic acid and thioctic acid were monitored by magnetoelastic transduction. After coupling of β-galactosidase to the SAMs, the enzyme activity was monitored by using a substrate that forms an insoluble product upon action of the enzyme. Specifically, an indolyl galactopyranoside substrate was employed in conjunction with an azo dye as the precipitating system. The immobilized enzyme was evaluated and found to have an apparent Michaelis-Menten constant (KM) of 1.2 mM for the indolyl galactopyranoside. Calibration plots for both substrates and inhibitors were generated to establish the versatility of this sensing system. Kinetic parameters for nonprecipitating substrates were determined in conjunction with a precipitating enzymatic substrate by way of a competitive inhibition study using β-galactosidase attached to magnetoelastic strips. The methods developed within this work allow for the fabrication of wireless enzyme sensing systems, which can also be used as another means of screening for enzyme inhibitors. Magnetoelastic alloy films offer the ability to fabricate wireless sensing devices. A typical use of such alloy films is as anti-theft devices.1 Recently, they have also been employed as a transduction platform for the production of sensing devices. Examples of sensing applications of these materials include the monitoring of physical parameters, such as temperature, humidity, pressure, density, and viscosity, and the sensing of chemical species including H+ ion (i.e., pH), ammonia, carbon dioxide, and glucose (ref 2 and references therein). Further, the use of magnetoelastic transduction to monitor the electrodeposition of polypyrrole has been demonstrated.3 Recently, a miniaturized sensing system * To whom correspondence should be addressed. E-mail: [email protected]. Phone: (859) 257-6350. Fax: (859) 323-1069. (1) Ryan, J., Jr. Sci. Am. 1997, 277, 120. (2) Grimes, C. A.; Mungle, C. S.; Zeng, K.; Jain, M. K.; Dreschel, W. R.; Paulose, M.; Ong, K. G. Sensors 2002, 2, 294-313.

6932 Analytical Chemistry, Vol. 75, No. 24, December 15, 2003

based on magnetoelastic films was applied to the long-term in situ monitoring of aqueous environments.4 Magnetoelastic transduction occurs with certain metal film alloys. Typically, to make use of such materials as sensing platforms, a dc-generated magnetic field is employed to offset the alloy’s magnetic anisotropy. A time-varying magnetic field, often generated by a sinusoidal ac, is then used as an excitation signal. When an alternating magnetic field sweep in the frequency domain is applied, the magnetoelastic film’s conversion of magnetic energy to elastic energy causes a mechanical resonance. This mechanical resonance gives rise to a time-varying magnetic flux that is at a maximum when the frequency of the excitation magnetic field matches the fundamental mechanical frequency of the alloy strip. This magnetic flux can be detected and converted into an electrical signal with a pickup coil. Observed changes in the peak resonance frequency of a magnetoelastic film and the amplitude of the film’s signal at that resonance frequency are directly related to any mass accumulated on the surface of the film. This principle can form a basis for using magnetoelastic films as sensing platforms. Depositing a layer that incorporates a biological receptor onto a magnetoelastic film could give rise to wireless biosensors. Many methods of immobilizing biological recognition elements on surfaces have been developed. Among these methods is the use of self-assembled monolayers (SAMs) on surfaces with subsequent covalent attachment of biomolecules. The formation of alkylthiol SAMs on a noble metal, such as gold, has become a welldeveloped technique for the design of biosensors.5-7 A variety of SAM-enzyme systems have been developed; see refs 8-12 for (3) Erso ¨z, A.; Ball, J. C.; Grimes, C. A.; Bachas, L. G. Anal. Chem. 2002, 74, 4050-4053. (4) Yang, X.; Ong, K. G.; Dreschel, W. R.; Zeng, K.; Mungle, C. S.; Grimes, C. A. Sensors 2002, 2, 455-472. (5) Wink, T.; van Zuilen, S. J.; Bult, A.; van Bennekon, W. P. Analyst 1997, 122, 43R-50R. (6) Ferretti, S.; Paynter, S.; Russell, D. A.; Sapsford, K. E.; Richardson, D. J. Trends Anal. Chem. 2000, 19, 530-540. (7) Gooding, J. J.; Hibbert, D. B. Trends Anal. Chem. 1999, 18, 525-533. (8) Dubrovsky, T. B.; Hou, Z.; Stroeve, P.; Abbott, N. L. Anal. Chem. 1999, 71, 327-332. (9) Mizutani, F.; Sato, Y.; Yabuki, S.; Sawaguchi, T.; Iijima, S. Electrochim. Acta 1999, 44, 3833-3838. (10) Gooding, J. J.; Pugliano, L.; Hibbert, D. B.; Erokhim, P. Electrochem. Commun. 2000, 2, 217-221. (11) Gaspar, S.; Zimmermann, H.; Gazaryn, I.; Cso ¨regi, E.; Schuhmann, W. Electroanalysis 2001, 13, 284-288. 10.1021/ac0347866 CCC: $25.00

© 2003 American Chemical Society Published on Web 11/06/2003

Figure 1. Representation of the system used to obtain magnetoelastic resonance frequency measurements. The solid lines depict the signal flow through the system.

examples. One method, in particular, involves the use of alkylthiol chains that terminate in ω-carboxylic acid. The latter can be activated by a carbodiimide and subsequently reacted with an N-hydroxysuccinimide moiety to provide a reactive site for covalent attachment of proteins (through the amino groups of lysine residues and the N-terminus) to the formed SAM.6 This approach has been employed in this work to attach an enzyme, β-galactosidase, to a mass-sensitive, gold-coated magnetoelastic alloy. It is well-known in the field of histochemistry that certain β-galactosidase substrates, after galactose cleavage, undergo a coupling with azo dyes to produce a precipitant.13 One such substrate is 5-bromo-6-chloro-3-indolyl-β-D-galactopyranoside (5-6-X-Gal). This precipitation phenomenon was employed herein to demonstrate the immobilization of a functional enzyme on the magnetoelastic alloy surface via ω-carboxylic acid terminated alkylthiol SAMs. A diazo dye, 5-chloro-4-benzamido-2-methylbenzenediazonium chloride hemi(zinc chloride) salt (also known as Fast Red Violet LB), was coupled with the 5-6-X-Gal enzymatic reaction product to increase the efficiency of the precipitation reaction.14 Additionally, this magnetoelastic biosensor could be used to determine the concentrations of compounds that are either substrates or inhibitors of β-galactosidase. This is accomplished through competitive inhibition of the β-galactosidase activity. This work demonstrates the feasibility of developing a class of magnetoelastic-transduced biosensors that could potentially be used for the monitoring or screening of any biological process involving the precipitation of analyte species. EXPERIMENTAL SECTION Chemicals. Metglas 2605SC was graciously provided by Honeywell (Conway, SC). This magnetoelastic alloy has the (12) Losic, D.; Shapter, J. G.; Gooding, J. J. Langmuir 2002, 18, 5422-5428. (13) Ashford, A. E. Protoplasma 1970, 71, 281-293. (14) Mohler, W. A.; Blau, H. M. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 1242312427.

composition Fe74B15C7Si4. High-purity (g99.9%) gold and titanium wires were obtained from Alfa Aesar (Ward Hill, MA). Absolute ethanol was purchased from Aaper Alcohol and Chemical (Shelbyville, KY). 3-Mercaptopropanoic acid (3-MPA) was purchased from Acros Organics (Springfield, NH). D-Galactose was from Aldrich (Milwaukee, WI). A number of other reagents were procured from Sigma (St. Louis, MO), including the following: D,L-6,8-thioctic acid (TA), 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDAC), N-hydroxysuccinimide (NHS), β-galactosidase from bovine liver, 5-6-X-Gal, and Fast Red Violet LB. All other chemicals and solvents used were of reagent grade. Solutions, unless otherwise stated, were prepared using deionized (Milli-Q, Millipore, Bedford, MA) RO water. Measurement of the Peak Resonance Frequency of the Magnetoelastic Films. The resonance frequency measurements were made using a setup previously described;15 a sketch of the system is shown in Figure 1. A brief overview of the operation of the system is as follows. Both a 3.0-A dc signal and a 50-mA sinusoidal ac signal were run through wire coils in the Helmholtz configuration to generate magnetic fields. The dc magnetic field offset the magnetic anisotropy of the alloy. The ac field was swept in the frequency mode for the interrogation of the film. Magnetic energy was then converted to elastic energy by the magnetoelastic film. The elastic energy caused a deformation of the strip, which, in turn, gave rise to a magnetic flux. A pickup coil was used to monitor the change in magnetic flux generated by the resonance frequency changes in the magnetoelastic films. The pickup coil signal was fed into a low-noise preamplifier and a lock-in amplifier before data acquisition by a computer. The computer also controlled the application of the dc and ac magnetic fields. Metal Coating of the Magnetoelastic Strips. The Metglas alloy was cut into strips with dimensions 30 mm × 6 mm. A coating of titanium was evaporated onto each side of the strips (15) Grimes, C. A.; Ong, K. G.; Loiselle, K.; Stoyanov, P. G.; Kouzoudis, D.; Liu, Y.; Tong, C.; Tefiku, F. Smart Mater. Struct. 1999, 8, 639-646.

Analytical Chemistry, Vol. 75, No. 24, December 15, 2003

6933

via a Balzer Union (Balzers, Lichtenstein) model 010 benchtop thermal evaporator; this was followed by a layer of gold. The intermediate titanium layer provided better adhesion of the gold to the alloy strips. The thicknesses of the layers were estimated by measuring the peak resonance frequency of particular strips before and after metal coating. The frequency shift can be used to calculate the mass deposited onto the surface of the strip and subsequently the metal coating thickness. This calculation is based on the modified Sauerbray expression presented in ref 3, which relates peak resonance frequency shifts to small mass loads placed on magnetoelastic alloy strip surfaces. Typically, the titanium coatings were ∼230 nm thick, and the gold coatings ∼100 nm thick. Formation of SAM Layers and Immobilization of β-Galactosidase. To allow self-assembled monolayers to form, strips of the magnetoelastic film were immersed in 25 mM solutions of either 3-MPA or TA in absolute ethanol overnight. These strips were then rinsed with absolute ethanol and dried in a stream of nitrogen. Then, unless otherwise stated, the strips were placed into a solution of 15 mM EDAC and 35 mM NHS prepared in dry acetonitrile, where they remained for 1.5 h. When the strips were removed from the EDAC/NHS solution, they were rinsed with dry acetonitrile and dried in a stream of nitrogen. To immobilize the β-galactosidase, magnetoelastic strips were placed in a 10 mM sodium phosphate buffer, pH 8.5, containing 200-500 µg/mL of the enzyme. The strips remained in this solution overnight at 4 °C. The film was then rinsed with buffer and placed in a glass tube with buffer, where it was kept refrigerated until it was used. All substrate solutions for β-galactosidase were prepared in 10 mM, pH 7.0 sodium phosphate buffer and contained 5 mM MgCl2. Magnetoelastic Measurements and Monitoring of the Precipitation of the Enzymatic Product. The magnetoelastic resonance measurements were made while the strips where in aqueous buffered solutions in cap-sealed glass tubes. The tubes were placed within the pickup coils that were within a set of Helmholtz-configured interrogation coils (see Figure 1). Resonance frequency scans were automatically obtained at timed intervals. The resulting peak resonance frequencies and their associated amplitudes (as voltages) were used to plot the data shown in the Results and Discussion section. The peak resonance frequencies and peak amplitudes were normalized by subtracting the initial value of each experiment from the rest of the data for that particular experiment. RESULTS AND DISCUSSION Magnetoelastic alloy films can be adapted for use as transducers for a wide variety of wireless sensing systems. The most typical use of magnetoelastic alloys is as antitheft markers in the retail industry. A signal versus frequency response obtained with the experimental setup described above from an unmodified antitheft strip from a retail product is compared in Figure 2 with the response of an unmodified alloy strip that was obtained directly from the manufacturer and used throughout this work. The response of a magnetoelastic alloy is dependent on the size of the film used, as well as the particular elemental composition of the alloy. The antitheft alloy is optimized as a positional sensor with a relatively frequency-stable signal response, whereas the Metglas 2605SC material has a more versatile range of uses. 6934 Analytical Chemistry, Vol. 75, No. 24, December 15, 2003

Figure 2. Magnetoelastic frequency scans of an unmodified commercial antitheft strip from a retail product (solid line) and of an unmodified alloy strip used to perform the experimentation described in this work. Table 1. Change in Peak Resonance Frequency during Self-Assembly of Different Concentrations of 3-MPA on Gold-Coated Magnetoelastic Films change in peak frequency (Hz) time (h)

5 mM

10 mM

25 mM

0.5 1.0 1.5 2.0 10.0 20.0

200.4 -199.3 -199.3 -199.3 -199.3 -199.3

200.8 0.1 -199.9 -199.9 -199.9 -199.9

400.7 0.4 -199.5 -199.5 -199.5 -199.5

Magnetoelastic transduction can be employed to investigate the formation of SAMs on the film surface. To that end, goldcoated magnetoelastic strips were placed into ethanol solutions of 3-MPA, and resonance frequency measurements were periodically taken while the formation of the monolayers took place. The resulting shifts in peak resonance frequency for three different concentrations of 3-MPA in ethanol are shown in Table 1. The blank sample, a strip immersed in thiol-free ethanol, showed no significant peak frequency shift over the period monitored. All solutions of 3-MPA demonstrated a total stable peak resonance shift of approximately 200 Hz toward lower frequencies after 90 min. This would indicate that complete monolayer formation was achieved in this time frame with the concentrations of 3-MPA tested. When a strip was left in a solution of 3-MPA for longer times, up to 20 h, the overall peak frequency shift did not significantly change for any concentration of 3-MPA. It was observed that, during 3-MPA SAM formation, there was an initial increase in the earliest peak resonance frequency measurements (at 30 min). Although the exact reason for this effect is not yet completely understood, it could be related to rearrangement of the initially adsorbed thiol-containing molecules. Gold-coated magnetoelastic strips exposed to two different types of SAM-forming compounds, 3-MPA and TA, were compared. The structures of 3-MPA and TA when attached to a goldcoated magnetoelastic strip surface are shown in Figure 3. When examining the peak frequency shifts that arise from the formation of the 3-MPA and TA monolayers, it is noticeable that, in the short term, the increase in peak resonance frequency seen with the 3-MPA monolayer does not appear with TA. Within a 90-min monitoring period, the two compounds exhibit very comparable

Figure 3. Molecular structures of 3-MPA and TA when attached to a gold surface.

overall peak shifts using solutions of the same concentration. This indicates that, over the 90-min period of monolayer formation, approximately the same amount of mass was immobilized from each compound. From this comparison and the fact that the molecular weight of TA is roughly twice that of 3-MPA (206.3 versus 106.1 g/mol, respectively), it would be expected that twice as many molecules of 3-MPA as TA had been attached to the gold surface. This conclusion seems to be logical given the molecular footprint of TA is a bidentate attachment, whereas 3-MPA has only a single attachment point (refer to Figure 3). The ability to immobilize biomolecules on the magnetoelastic alloy strip presents an opportunity for the development of wireless monitoring of biological reactions. Modification of the gold-coated magnetoelastic films with a self-assembled monolayer of a ω-carboxylic acid compound, 3-MPA or TA, generates functional groups onto which immobilization of β-galactosidase can proceed. The carboxylates of the SAM molecules were reacted with an acetonitrile solution containing carbodiimide and N-hydroxysuccinimide reagents to form a stable, linking intermediate to which β-galactosidase was covalently attached. The activity of the enzyme on the biofunctionalized magnetoelastic strips was monitored by resonance frequency scans using 5-6-X-Gal as the substrate. This substrate forms a product that is insoluble and, therefore, changes the mass load on the strip. The change over time of the voltage value at the peak resonance frequency of enzyme-coated strips during the reaction of β-galactosidase with 5-6-X-Gal is shown in Figure 4. Comparatively lower enzyme activity was found when β-galactosidase was physically adsorbed directly onto the goldcoated alloy film without first forming a SAM. Because the 3-MPAbased magnetoelastic enzyme system showed greater sensitivity than the TA-based system, all subsequent experiments were performed using 3-MPA SAMs only. It should also be noted that the overall peak amplitude change of even the 3-MPA system is in the low millivolt range when 5-6X-Gal is used as the precipitating reagent by itself (Figure 4). It is known that substrates such as 5-6-X-Gal generate enzymatic products that can be further reacted with azo dyes to form a more substantial amount of colored solid.13 In an effort to enhance the response of the magnetoelastic-enzyme sensing system, further experiments were performed using substrate solutions that contained both 5-6-X-Gal and the azo dye Fast Red Violet LB as detailed below. It should be noted that most diazo compounds are light-sensitive in solution. Magnetoelastic strips were immersed in a buffer solution containing 90 µM Fast Red Violet LB, and the change in peak amplitude was monitored over time while

Figure 4. Change in peak resonance frequency amplitudes of goldcoated magnetoelastic films with β-galactosidase immobilized via surface adsorption (×), TA SAM (b), and 3-MPA SAM (9), when placed in a sodium phosphate buffer (10 mM, pH 7.0) containing 100 µM 5-6-X-Gal. The signal change on the y axis refers to the change in amplitude of the peak resonance frequency.

Figure 5. Change in peak resonance frequency amplitudes of goldcoated magnetoelastic alloy strips placed in sodium phosphate buffer (10 mM, pH 7.0) containing 90 µM Fast Red Violet LB when exposed to light (0) and when kept in an opaque tube (b) during monitoring. The signal change on the y axis refers to the change in amplitude of the peak resonance frequency.

one solution was exposed to light and another was kept in an opaque tube. As can be seen in Figure 5, in the absence of any enzyme, the light-exposed azo dye forms measurable solid precipitate on the strip after approximately 15 min. When the light exposure was kept to a minimum, there was very little change in the magnetoelastic film response. All further work was performed while keeping solutions protected from light as much as possible. A calibration plot was constructed using substrate solutions containing 90 µM Fast Red Violet LB and various concentrations of 5-6-X-Gal up to 5 mM. The initial reaction rates of β-galactosidase immobilized on the magnetoelastic strip (within the first 120 s of the reaction) were calculated from the change in peak resonance frequency amplitudes for each concentration of 5-6-XGal used. From the resulting calibration plot (Figure 6), the apparent Michaelis-Menten constant (KM) of the immobilized enzyme on the magnetoelastic biosensor was determined. The apparent KM of the immobilized β-galactosidase for 5-6-X-Gal was found to be 1.2 mM by fitting the data to the Michaelis-Menten Analytical Chemistry, Vol. 75, No. 24, December 15, 2003

6935

Figure 6. Plot of initial enzymatic rates (as determined by changes in the amplitude of the peak resonance frequency) versus substrate concentration. The line represents the result of nonlinear regression analysis of the data points.

equation using nonlinear regression analysis. This value is slightly higher than the KM of β-galactosidase found in solution (0.9 mM). Published work on the competitive inhibition of signal-generating substrates for determinations of kinetic parameters of nonsignal-generating compounds point to the possibility of using the described system as a biosensor.16-18 To demonstrate this possibility, we employed D-galactose as an inhibitor of β-galactosidase immobilized on magnetoelastic strips; because D-galactose is also a substrate of β-galactosidase, it competes with 5-6-X-Gal for the active site of the enzyme and thus acts as a competitive inhibitor. Because it is not possible for D-galactose and 5-6-X-Gal to simultaneously occupy the active site of β-galactosidase, the presence of D-galactose results in a reduction of the precipitate formed when both substrates are in solution with β-galactosidase. When an enzymatic substrate competition was set up with both the 5-6-X-Gal/azo dye combination and a 14-fold greater concentration of the nonprecipitating substrate, D-galactose, the change in peak voltage still indicated reaction of the 5-6-X-Gal, but it was at a rate over 2.5 times lower than that observed with the precipitating reagents alone (initial rates of -0.069 vs -0.179 ∆mV/min, respectively). When a solution of D-galactose without any 5-6-X-Gal was used as a control, the corresponding signal change of the β-galactosidase-immobilized magnetoelastic strip was essentially zero for the time period monitored. A more complete study of the D-galactose competitive inhibition of β-galactosidase activity on 5-6-X-Gal was also performed. Solutions with D-galactose concentrations from 0 to 75 mM were prepared using a stock solution containing 5-6-X-Gal and MgCl2 in buffer. The initial enzymatic rates measuring the conversion of 5-6-X-Gal to precipitating product were determined using β-galactosidase immobilized on magnetoelastic strips. A plot of these rates versus the D-galactose concentration is shown in Figure 7. These data were utilized to construct a Dixon plot19 (not shown) (16) Blake, R. C., II; Vassall, R. F.; Blake, D. A. Arch. Biochem. Biophys. 1989, 272, 52-68. (17) Portaccio, M.; Stellato, S.; Rossi, S.; Bencivenga, U.; Mohy Eldin, M. S.; Gaeta, F. S.; Mita, D. G. Enzyme Microb. Technol. 1998, 23, 101-106. (18) Freeman, M. K.; Bachas, L. G. Biosens. Bioelectron. 1992, 7, 49-55. (19) Segel, I. H. In Biochemical Calculations: How to Solve Mathematical Problems in General Biochemistry, 2nd ed.; Wiley: New York, 1976; pp 208-323.

6936 Analytical Chemistry, Vol. 75, No. 24, December 15, 2003

Figure 7. Plot of the initial rate (as determined by changes in the peak resonance frequency) of enzymatic conversion of 0.1 mM 5-6X-Gal to precipitating product in sodium phosphate buffer (10 mM, pH 7.0) solutions also containing 5 mM MgCl2 and various concentrations of D-galactose.

from which the inhibition constant (Ki) of D-galactose for the system could be calculated. The Ki value for D-galactose was found to be 0.94 mM. The plot shown in Figure 7 demonstrates that β-galactosidase immobilized on magnetoelastic alloy films can be employed in determining D-galactose concentrations. The working range is dependent on the concentration of the precipitating substrate used in the experiment. One could envision that a number of magnetoelastic-based biosensing systems could be fabricated by using precipitating substrates for enzymes that have been immobilized on SAM-coated magnetoelastic alloys. Commercially available chromogenic insoluble substrates include benzidine- and naphthol-based compounds for use with peroxidases; phenazine methosulfate for use with glucose oxidase; indoyl-containing compounds for use with phosphatases or galactosidase; and tetrazolium salts, also for use with phosphatases.20 From the above results, it is reasonable to conclude that enzymes that are immobilized on the surface of magnetoelastic alloy film transducers could provide a platform for detecting possible enzyme inhibitors as encountered in high-throughput screening applications. CONCLUSIONS A method of covalently immobilizing an enzyme, β-galactosidase, to a magnetoelastic alloy strip has been devised. The system utilizes the formation of a self-assembled alkylthiol monolayer containing a terminal carboxylate on the gold-coated surface of the magnetoelastic film. The enzyme is then attached to the monolayer via carbodiimide/N-hydroxysuccinimide-mediated coupling. A time-varying magnetic field can be used to probe the signal of the magnetoelastic alloy. The addition of mass to the surface of the alloy strip causes a related change in the value of the peak frequency and amplitude. Enzyme activity can be monitored using an enzymatic precipitating reaction. It is also possible to monitor competing substrates and/or inhibitors through their effect on the precipitating substrate reaction with (20) Applications Handbook and Catalog; Pierce Biotechnology, Inc.: Rockford, IL, 2003; pp 255-276.

the enzyme. Because transduction is performed through the use of magnetic fields without the need for electrical connections or line of sight, this work demonstrates that it is possible to develop enzymatic sensors that can be placed and operated in sealed, nontransparent containers for in situ detection of substrates or inhibitors. Further, this method could also be used for screening enzyme inhibitors or determining concentrations of enzyme substrates or inbitors based on wireless, magnetoelastic transduction.

ACKNOWLEDGMENT The authors acknowledge funding through the NSF-IGERT program. Prof. Craig Grimes and his research group (Department of Electrical Engineering, Pennsylvania State University, University Park, PA) are thanked for the setup of the instrumentation. Received for review July 11, 2003. Accepted September 26, 2003. AC0347866

Analytical Chemistry, Vol. 75, No. 24, December 15, 2003

6937