Electrochemical Surface Plasmon Resonance Measurement in a

We developed an electrochemical surface plasmon resonance flow cell for the simultaneous measurement of the binding affinity and catalytic activity of...
0 downloads 0 Views 185KB Size
Download PDF
Anal. Chem. 2007, 79, 9572-9576

Electrochemical Surface Plasmon Resonance Measurement in a Microliter Volume Flow Cell for Evaluating the Affinity and Catalytic Activity of Biomolecules Ryoji Kurita,*,† Yoshimi Yokota,† Akio Ueda,‡ and Osamu Niwa†,‡

National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba Central 6, 1-1-1 Higashi, Tsukuba, Ibaraki, Japan, and Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama, Japan

We developed an electrochemical surface plasmon resonance flow cell for the simultaneous measurement of the binding affinity and catalytic activity of bifunctional biomolecules. These measurements will be useful for evaluating the performance of such biomolecules as ribozyme and abzyme. The simultaneous measurements were performed on a gold surface modified with a multilayer consisting of poly-L-lysine and poly(styrene sulfonate) assembled with the layer-by-layer method using an enzymelabeled monoclonal antibody as a model compound. We obtained the amount of immunocomplex formation from the surface plasmon resonance angle shift value by injecting the compound into the flow cell containing the multilayer modified with tumor necrosis factor-r. Then we compared this surface plasmon resonance result with that for the in situ electrochemical oxidation of p-aminophenol formed by the catalytic reaction of labeled enzyme on the same gold film. We were able to obtain a high correlation coefficient of 0.999 between the two responses. This is because the compound could be captured with high stability with a less than 3% coulometric response decrease in the catalyzed product in the multilayer whose thickness was easily controllable. In addition, we were able to measure the catalytic activity by coulometry and thus avoid the effect of peak broadening. We also report that the dephosphorylation activity of a bound compound could be estimated from the measurement results and an equation. Bifunctional biomolecules such ribozyme or abzyme have been extensively studied in recent decades. Since these molecules have both specific affinity and catalytic activity as regards their own cleavage or other biomolecules,1-4 they may be useful for * Corresponding author. Phone: +81-29-861-6158. Fax: +81-29-861-6177. E-mail: [email protected]. † National Institute of Advanced Industrial Science and Technology. ‡ Tokyo Institute of Technology. (1) Piccirilli, J. A.; Vyle, J. S.; Caruthers, M. H.; Cech, T. R. Nature 1993, 361, 85-88. (2) Brogan, A. P.; Eubanks, L. M.; Koob, G. F.; Dickerson, T. J.; Janda, K. D. J. Am. Chem. Soc. 2007, 129, 3698-3702. (3) Sagi, A.; Rishpon, J.; Shabat, D. Anal. Chem. 2006, 78, 1459-1461. (4) Westhof, E.; Massire, C. Science 2004, 306, 62-63.

9572 Analytical Chemistry, Vol. 79, No. 24, December 15, 2007

therapeutic applications, for example, as anticancer drugs. The most commonly used assay for these molecules has been based on polyacrylamide gel electrophoresis (PAGE).5,6 PAGE is useful for studying known or unknown samples by separating them in terms of molecular size in parallel on a slab gel. Unfortunately, this method is time-consuming, costly, and often involves labeling for fluorescence detection. In some instances, high-performance liquid chromatography or capillary electrophoresis has been used to quantify amounts of analytes and catalyzed product,7,8 and separation times are greatly improved by modern advances in microfluidics. However, these are not continuous measurement techniques, and so they are inappropriate as regards kinetics analysis for evaluating the affinity of biomolecules. We focus on a combination of the surface plasmon resonance (SPR) technique and electrochemical detection for the highthroughput measurement of bifunctional biomolecules.9,10 SPR is a surface-sensitive optical technique that has been used to study thin layers on metal surfaces.11-13 Recently, many researchers have reported the high-throughput screening of biomolecules on the basis of SPR angle shifts caused by interaction with a sensing surface modified with a reagent. The SPR-based affinity measurement is a simple, safe, and real-time method without any isotopes or fluorescence labels;14-18 however, it is difficult to measure (5) Paxon, T. L.; Brown, T. S.; Lin, H. Y. N.; Brancato, S. J.; Roddy, E. S.; Bevilacqua, P. C.; Ewing, A. G. Anal. Chem. 2004, 76, 6921-6927. (6) Forconi, M.; Herschlag, D. J. Am. Chem. Soc. 2005, 127, 6160-6161. (7) Thayer, J. R.; Rao, S.; Puri, N.; Burnett, C. A.; Young, M. Anal. Biochem. 2007, 361, 132-139. (8) Saevels, J.; Van Schepdael, A.; Hoogmartens, J. Anal. Biochem. 1999, 266, 93-101. (9) Kolb, D. M.; Ko ¨tz, R. Surf. Sci. 1977, 64, 96-108. (10) Abele`s, F.; Lopez-Rios, T.; Tadjeddine, A. Solid State Commun. 1975, 16, 843-847. (11) Niu, L.; Knoll, W. Anal. Chem. 2007, 79, 2695-2702. (12) Li, Y.; Lee, H. J.; Corn, R. M. Anal. Chem. 2007, 79, 1082-1088. (13) Lioubashevski, O.; Chegel, V. I.; Patolsky, F.; Katz, E.; Willner, I. J. Am. Chem. Soc. 2004, 126, 7133-7143. (14) Kanda, V.; Kariuki, J. K.; Harrison, D. J.; McDermott, M. T. Anal. Chem. 2004, 76, 7257-7262. (15) Fitzpatrick, B.; O’Kennedy, R. J. Immunol. Methods 2004, 291, 11-25. (16) Akimoto, T.; Ikebukuro, K.; Karube, I. Biosens. Bioelectron. 2003, 18, 14471453. (17) Kurita, R.; Yokota, Y.; Sato, Y.; Niwa, O. Anal. Chem. 2006, 78, 55255531. (18) Choi, J. W.; Park, K. W.; Lee, D. B.; Lee, W.; Lee, W. H. Biosens. Bioelectron. 2005, 20, 2300-2305. 10.1021/ac071412u CCC: $37.00

© 2007 American Chemical Society Published on Web 10/31/2007

catalytic activity at the same time with an identical surface. Thus, catalytic activity is measured after collecting the analytes that show an affinity with the modified reagents. On the other hand, electrochemical detection is advantageous for monitoring catalytic activity because the method is simple, inexpensive, and the electrodes are easy to miniaturize for smallvolume samples while maintaining relatively high sensitivity.19 However, when the electrochemical flow cell is connected in tandem downstream of the SPR flow cell, it is difficult to estimate the catalytic activity correctly. This is because we often need to change the flow rate considerably to prevent the SPR sensing surface from being in a diffusion-limited condition when measuring the association and dissociation kinetics. Therefore, it is more appropriate that coulometric catalytic activity measurements are undertaken on an identical sensing surface even if the response is distorted by the polymeric film because the products generated by the catalytic reaction simply diffuse across the modified layer and reach the electrode, which is less influenced by the flow stream. However, when immobilization techniques conventionally used in SPR measurements such gold-thiol binding are employed for electrochemical detection, we need to be careful with respect to the density because closely packed layers suppress the activity of the electrode surface.20,21 In contrast, low-density immobilization while maintaining high diffusion of the analyte reduces the amount of immobilized biomolecules and thus reduces the sensing capability. In addition, there is the problem of nonspecific adsorption on a part of a bare electrode surface which also reduces the stability of the electrochemical responses.22,23 Various selfassembled monolayer and polymer films have been modified to suppress nonspecific adsorption in bioassays.24 The film density on the electrode surface should be considered for electrochemical applications because surface modification strongly affects the electrochemical performance. For a simple, correct measurement with high collection efficiency, the detection surface should be modified with an ideal film that offers a high diffusion coefficient for catalytic products, a less nonspecific adsorption, and a stable immobilized reagent that retains its specific binding activity. In this paper, we propose a small-volume electrochemical SPR flow cell for simultaneously measuring affinity and catalytic activity on an identical gold sensing surface. We formed a thin multilayer of polylysine and poly(styrene sulfonate) as a scaffold for both measurements by injecting the two polyelectrolyte solutions onto the gold surface installed in our original electrochemical SPR flow cell with portable SPR equipment. After immobilizing tumor necrosis factor-R (TNF-R) on the multilayer, we compared two different responses from electrochemical and SPR signals by introducing anti TNF-R antibody labeled with alkaline phosphatase as a model compound that exhibits both specific affinity and catalytic activity. (19) Wang, J. Biosens. Bioelectron. 2006, 21, 1887-1892. (20) Nuzzo, R. G.; Fusco, F. A.; Allara, D. L. J. Am. Chem. Soc. 1987, 109, 2358-2368. (21) Fakunle, E. S.; Aguilar, Z. P.; Shultz, J. L.; Toland, A. D.; Fritsch, I. Langmuir 2006, 22, 10844-10853. (22) Kurita, R.; Tabei, H.; Iwasaki, Y.; Hayashi, K.; Sunagawa, K.; Niwa, O. Biosens. Bioelectron. 2004, 20, 518-523. (23) Kurita, R.; Yabumoto, N.; Niwa, O. Biosens. Bioelectron. 2006, 21, 16491653. (24) Prime, K. L.; Whitesides, G. M. Science 1991, 252, 1164-1167.

Figure 1. (a) Photograph of portable SPR equipment with an electrochemical SPR flow cell. (b) Schematic representation of the electrochemical SPR flow cell.

EXPERIMENTAL SECTION Reagents. Poly-L-lysine (average MW 80 000), TNF-R, and glycine were purchased from Sigma (St. Louis, MO). Poly(styrene sulfonate) (average MW 70 000) was purchased from Aldrich. N-Hydroxysulfosuccinimide (NHS) and N-ethyl-N′-(3-dimethyl aminopropyl) carbodiimide hydrochloride (EDC) were purchased from Pierce (Rockford, IL). Anti TNF-R antibody labeled with alkaline phosphate and p-aminophenolphosphate (PAPP) were purchased from R&D systems (Minneapolis, MN). p-Aminophenol (PAP) was purchased form Kanto-kagaku (Tokyo, Japan). We used three buffer solutions, acetate (pH 5.0), phosphate (pH7.0), and tris buffer (pH 8.0). The acetate buffer contained 10 mM acetic acid and 7 mM NaOH. The phosphate buffer contained 5.3 g/L KH2PO4 and 8.67 g/L Na2HPO4. The tris buffer contained 10 mM tris (hydroxymethyl) aminomethane, 0.15 M NaCl, and 5 mM MgCl2. Apparatus. Figure 1, parts a and b, shows a photograph and a schematic representation of an electrochemical SPR measurement system with a thin-layer flow cell. The flow cell was our original design made of acrylic resin, and it has a Ag-AgCl reference electrode and a stainless outlet pipe as a counter electrode. We installed the flow cell in the SPR equipment with an 80 µm poly(vinyl chloride) spacer and a glass wafer with a Analytical Chemistry, Vol. 79, No. 24, December 15, 2007

9573

thin gold film. The spacer has a 3 mm × 12 mm oval hole in it to provide a thin-layer flow channel. The thin gold film was deposited on a glass wafer (n ) 1.525 ( 0.0015, Matsunami Glass Industry, Osaka, Japan) as follows. We sealed the glass wafer using an adhesive sheet that had a 3 mm diameter round hole and an electrical wire pattern for connection between the round pattern and an electrochemical analyzer. We then deposited a thin titanium layer on the glass wafer with the adhesive sheet by using rf sputtering equipment (Seed Lab., Tokyo, Japan) and formed a gold film without breaking the vacuum.25 The total thickness of the gold and titanium film was 50 ( 2 nm. After peeling off the adhesive sheet, we used the round gold pattern as both a working electrode and an SPR measurement surface. We connected the three electrodes of the flow cell to an electrochemical analyzer (CHI instrument, model 802) and controlled the potential of the working electrode. We used portable SPR equipment (NTT Advanced Technology, Tokyo, Japan), which was developed in collaboration with one of the present authors. The portable SPR system is 16 cm wide, 6 cm deep, and 9.5 cm high, and weighs about 770 g. Index matching oil (n ) 1.510, Cargille Laboratories, Cedar Grove, NJ) was used to obtain optical contact between the glass wafer and the BK7 prism on the SPR equipment. A CMA102 dual syringe pump (CMA, Stockholm, Sweden) was used to introduce the sample solutions into the flow cell in a suction mode at a flow rate of 20 µL/min. Preparation of Layer-by-Layer Film for Electrochemical and SPR Measurements. We obtained layer-by-layer (LBL) film on the gold film by injecting poly(styrene sulfonate) and poly-Llysine solutions alternately into the flow cell while monitoring the SPR angle in real time. We injected 0.1 mg/mL poly(styrene sulfonate) into the flow cell followed by 0.1 mg/mL poly-L-lysine. We obtained LBL films with various thicknesses by repeating the alternating injection process. When we employed the LBL film for electrochemical and SPR measurements, we modified the LBL film with TNF-R by employing the common carbodiimide coupling reaction provided by the EDC/NHS system as follows. We injected acetate buffer containing both 0.4 mg/mL EDC and 1.1 mg/mL NHS for 15 min. We then injected 2 µg/mL TNF-R solution until the SPR angle increased 0.05°. This took approximately 2 min. Finally, we rinsed the gold surface by introducing a phosphate buffer solution for 1 h. Simultaneous Measurements of Affinity and Catalytic Activity. We injected anti TNF-R antibody labeled with alkaline phosphate as a model compound into the flow cell at a concentration of 50 to ∼900 ng/mL for 15 min. We labeled the antibody with alkaline phosphatase because many researchers have reported ribozymes and abzymes that exhibit catalytic activity with respect to phosphorylation or dephosphorylation.26-28 After rinsing the result by injecting phosphate buffer for 5 min, we observed that an amount of immunocomplex had formed on the LBL film by monitoring the SPR angle shift value. We then compared this SPR result with that obtained by employing electrochemical (25) Kurita, R.; Tabei, H.; Liu, Z.; Horiuchi, T.; Niwa, O. Sens. Actuators, B 2000, 71, 82-89. (26) Doudna, J. A.; Cech, T. R. Nature 2002, 418, 222-228. (27) Saran, D.; Held, D. M.; Burke, D. H. Nucleic Acids Res. 2006, 34, 32013208. (28) Schultz, P. G.; Yin, J.; Lerner, R. A. Angew. Chem., Int. Ed. 2002, 41, 44274437.

9574

Analytical Chemistry, Vol. 79, No. 24, December 15, 2007

detection with the alkaline phosphatase/PAPP system.29,30 We injected 2 mM PAPP in tris buffer into the flow cell. We then stopped the syringe pump and incubated the sample for 5 min to induce the enzyme reaction. During this incubation time, the alkaline phosphatase on the LBL film converted PAPP to PAP. After the incubation, we measured the oxidation current and coulomb value of PAP to p-quinoneimine by cyclic voltammograms in the -0.2 to +0.35 V (vs Ag/AgCl) range at a scan rate of 5 mV/s. PAPP is electroinactive in this potential range; thus, the current resulting from the oxidation of PAP to p-quinoneimine indicates the amount of captured compound on the LBL film. All measurements were carried out at room temperature. RESULTS AND DISCUSSION Electrochemical Response of PAP on a Layer-by-Layer Film Modified Electrode. In this work, it is important to form a thin film layer on a gold surface that can be used for immobilizing biomolecules with high stability while simultaneously minimizing any reduction in the diffusion coefficient of the catalyzed product on an electrode surface. In addition, a thin film must also be formed for SPR detection because the intensity of the evanescent field decreases as the distance from the gold surface increases. Therefore, we initially studied thin-film layer formation as a scaffold for reagent immobilization by performing SPR and electrochemical measurements of PAP, which is a catalyzed product in this study. Figure 2a shows cyclic voltammograms (CVs) of 100 µM PAP when the gold surface was unmodified and when it was modified with LBL film. We modified the gold surface with LBL film by injecting poly-L-lysine and poly(styrene sulfonate) solutions alternately from 0 to 8 times. We observed reversible voltammograms of PAP on a bare gold electrode at 80 and 31 mV. However, the shapes of the CVs distorted as we increased the number of injections we used to prepare the LBL film. The peak separation increased as the number of injections increased; however, the peak separation increase was only 27 mV when the LBL film was formed in two cycles. This value is very small compared with those in previous reports31,32 due to the small film thickness and porosity. From atomic force microscope observations, the average thicknesses of the films formed in 2 and 8 cycles are 11.9 and 60.8 nm, respectively (Supporting Information 1). We also observed a clear stepwise increase in the SPR angle by repeating the alternating injection process, and the film thickness increased around 7 nm per cycle injection. Although thiol-gold binding also provides a thinner monolayer, it is known that the shape of the voltammogram is often greatly distorted on a closely packed thiol electrode. Therefore, thiol density needs to be controlled. With our film, we were able to obtain a suitable structure for electrochemical detection with a simple assembly using two polyelectrolytes. Mizutani et al. reported that poly-L-lysine and poly(styrene sulfonate) complex film has a size exclusion property with a (29) Bauer, C. G.; Eremenko, A. V.; EhrentreichForster, E.; Bier, F. F.; Makower, A.; Halsall, H. B.; Heineman, W. R.; Scheller, F. W. Anal. Chem. 1996, 68, 2453-2458. (30) Das, J.; Jo, K.; Lee, J. W.; Yang, H. Anal. Chem. 2007, 79, 2790-2796. (31) Ionescu, R. E.; Gondran, C.; Gheber, L. A.; Cosnier, S.; Marks, R. S. Anal. Chem. 2004, 76, 6808-6813. (32) Darain, F.; Park, D. S.; Park, J. S.; Chang, S. C.; Shim, Y. B. Biosens. Bioelectron. 2005, 20, 1780-1787.

Figure 3. Variations in (a) SPR angle shift value and (b) electrochemical responses when the compound concentration was changed. The inset in (a) shows a Scatchard plot obtained from the SPR angel shift. The circles and squares in (b) indicate the coulomb value and the peak current, respectively.

Figure 2. (a) CVs for 100 µM PAP in 10 mM tris buffer containing 0.15 M NaCl on the LBL film modified thin gold electrode. The film was formed by the alternating injection of two polyelectrolyte solutions from 0 to 8 cycles. The scan rate was 5 mV/s. (b) Variations in coulomb value and peak current when the number of layers in the LBL film was changed. The circles and squares, respectively, indicate the coulomb value and the current density estimated from (a).

molecular weight cutoff a little over 100 Da.33 Thus, thin LBL film containing these two polyelectrolytes also had many relatively uniform nanoscopic pores, and the electron transfer of a small analyte (PAP, MW ) 109) was largely unimpeded. As a result, we were able to obtain a thin film without any degradation in electrochemical performance simply by employing polyelectrolyte injections. Figure 2b shows variations in the coulomb value and peak current estimated from Figure 2a when we changed the number of injections used to prepare the LBL film. The peak current decreases more than the coulomb value. This is mainly caused by the distortion of the voltammograms owing to the diffusion coefficient reduction as the LBL film thickness increases. In contrast, the coulomb value is almost maintained when the LBL film has few layers. This is because the PAP was oxidized with high efficiency on the electrode since the electrode was installed in a thin-layer (80 µm) flow channel. Since the electroactive volume of the electrode is 565 nL, the estimated coulomb value when all the PAP molecules are oxidized on the electrode is 152.8 µC/ cm2, which is identical to that on a bare gold electrode as shown (33) Mizutani, F.; Yabuki, S.; Hirata, Y. Anal. Chim. Acta 1995, 314, 233-239.

in Figure 2. We were able to obtain a stable response and a high collection efficiency on the LBL-modified electrode by combining it with our flow cell. This clearly indicates that the LBL film combined with our small-volume flow cell is advantageous as regards the correct measurement of catalytic activity by electrochemical detection without any interference from immobilized layers. In addition, our film is suitable for biomolecule immobilization because we can use well-known coupling techniques such the carbodiimide reaction in a mild aqueous condition because the film has an amino and carboxyl group. Comparison of SPR and Electrochemical Responses. Figure 3 shows the variations in the SPR and electrochemical responses when we injected various concentrations of antibodylabeled alkaline phosphatase onto a gold sensing surface. Before the measurements, the gold surface was modified with the two electrolytes by two injection cycles since this had almost no effect on the diffusion of the catalyzed product. In the SPR measurements, the SPR angle increased as the compound concentration increased. The detection limit of 50 ng/ mL (approximately 0.2 nM; the molecular weight of the compound is 250 kDa) was almost the same as that of previously reported SPR-based immunosensors.34-36 We also obtained a Scatchard plot as shown in the inset in Figure 3a. The binding constant (KA) of the immunoreaction was estimated to be 1.3 × 108 (M-1), which is a satisfactory level for the affinity of a monoclonal antibody. (34) Townsend, S.; Finlay, W. J. J.; Hearty, S.; O’Kennedy, R. Biosens. Bioelectron. 2006, 22, 268-274. (35) Bae, Y. M.; Oh, B. K.; Lee, W.; Lee, W. H.; Choi, J. W. Biosens. Bioelectron. 2005, 21, 103-110 (36) Gomes, P.; Andreu, D. J. Immunol. Methods 2002, 259, 217-230.

Analytical Chemistry, Vol. 79, No. 24, December 15, 2007

9575

slope )

C ∆SPR

(1)

Here, C (coulomb/cm2) is the charge density and ∆SPR (deg) is the SPR angle shift value. An SPR angle shift of 1° corresponds to 10-6 g/cm2, the Faraday constant is 96 500 (coulomb/mol), and the electron number for PAP oxidation is 2. Therefore eq 1 can be rewritten as follows,

slope )

Figure 4. Plots of the coulomb value and peak current vs the SPR angle shift value. Each value was estimated from Figure 3.

The circles and squares in Figure 3b indicate the electrochemical responses of the coulomb value and current density estimated from CVs after a 2 mM PAPP injection. We observed a stable increase in the coulomb value as the compound concentration increased. Furthermore the curve in the coulomb value is identical to that in the SPR responses. However, we observed saturation and some fluctuation in the current density increase. We previously reported a polyion complex film for SPR assay that we realized by placing and drying a mixture.37 The film showed both the suppression of nonspecific adsorption and the high specific affinity of an immobilized antibody. However, it is difficult to apply the film to electrochemical enzyme immunoassay because it is difficult to control the film thickness. LBL film consisting of poly(styrene sulfonate) and poly-L-lysine is beneficial not only for SPR but also for electrochemical detection because the film has good properties as a scaffold for the stable immobilization of biomolecules and its thickness can be controlled without losing the diffusion of the dephosphorylated product on the electrode. Figure 4 shows the correlation between the SPR and electrochemical responses in the flow cell estimated from Figure 3. We obtained a high correlation coefficient of 0.999 from the coulomb value measurements. In contrast, the correlation coefficient between the SPR results and the current density was 0.891. This indicates that a much closer relationship can be obtained between SPR and coulometric responses than between SPR and amperometric responses. The slope of the coulomb value versus the SPR response in Figure 4 is expressed as (37) Kurita, R.; Hirata, Y.; Yabuki, S.; Kato, D.; Sato, Y.; Mizutani, F.; Niwa, O. Electrochemistry 2006, 74, 121-124.

9576

Analytical Chemistry, Vol. 79, No. 24, December 15, 2007

N k W

(2)

Here, N (mol) is the catalyzed product number, W (g) is the weight of compound trapped on the sensing surface, and k is a constant (96 500 × 2 × 10-6). We can estimate the dephosphorylated product number per compound weight, i.e., the catalytic activity, from the slope and eq 2. The activity of the compound was about 7.7 units/mg from the slope. We achieved the simultaneous measurement of affinity and catalytic activity simply and rapidly by employing an in situ comparison of the coulomb value and SPR responses on an identical surface. This would also be useful, for example, for the efficiency screening of catalytic RNA (ribozyme) or catalytic antibody (abzyme). CONCLUSION We have developed an electrochemical SPR flow system with multiple layers of poly-L-lysine and poly(styrene sulfonate) assembled by the LBL method for the simultaneous determination of two sensor responses. We succeeded in obtaining the two responses with a high correlation coefficient between the SPR and electrochemical responses by controlling the multilayer thickness on an identical sensing surface located in a thin-layer flow channel. We only demonstrated the excellent correlation between the amounts of binding molecules and catalyzed products by an enzymatic reaction using alkaline phosphatase labeled antibody as a model compound. However, the combination of a bifunctional gold film detector (SPR and electrochemical) in a thin-layer flow cell and an LBL multilayer as a scaffold would be beneficial for the high-throughput screening of bifunctional biomolecules including ribozyme and abzyme, which exhibit specific affinity and catalytic activity. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review July 3, 2007. Accepted September 27, 2007. AC071412U