Wiring of Enzymes to Electrodes by Ultrathin Conductive Polyion

Jul 23, 2003 - Wiring of Enzymes to Electrodes by Ultrathin. Conductive Polyion Underlayers: Enhanced. Catalytic Response to Hydrogen Peroxide. Xin Yu...
0 downloads 0 Views 119KB Size
Download PDF
Anal. Chem. 2003, 75, 4565-4571

Wiring of Enzymes to Electrodes by Ultrathin Conductive Polyion Underlayers: Enhanced Catalytic Response to Hydrogen Peroxide Xin Yu,† Gregory A. Sotzing,†,‡ Fotios Papadimitrakopoulos,†,‡ and James F. Rusling*,†,‡,§

Department of Chemistry, Box U-60, University of Connecticut, Storrs, Connecticut 06269-3060, Institute of Materials Science, University of Connecticut, Storrs, Connecticut 06269-3136, and Department of Pharmacology, University of Connecticut Health Center, Farmington, Connecticut 06032

Stable electroactive films were grown layer by layer on rough pyrolytic graphite electrodes featuring 4-nm underlayers of sulfonated polyaniline (SPAN) covered with a film containing myoglobin or horseradish peroxidase grown in alternating layers with poly(styrenesulfonate). The self-doped polyanionic SPAN layer, grown on a 2-nm polycation layer, was conductive between about 0.1 and -0.4 V vs SCE at pH 4.5. The enzyme films had the architecture PDDA/SPAN/(enzyme/PSS)3, where PDDA is poly(diallyldimethylammonium) ion. Comparisons of voltammetric measurements of electroactive protein with quartz crystal microbalance measurements of total protein showed that 90% or more of the protein was coupled to the electrode when the SPAN underlayer was present, as opposed to ∼40% protein electroactivity when SPAN was absent. As a consequence of the highly efficient coupling between enzymes and electrode, the PDDA/SPAN/ (enzyme/PSS)3 films exhibited a higher sensitivity for the electrochemical catalytic reduction of hydrogen peroxide. Amperometry at a rotating disk electrode at 0 V gave sensitivity for hydrogen peroxide up to 14 µA µM-1 cm-2 in the submicromolar concentration range and a detection limit of ∼3 nM. Results suggest the future utility of ultrathin layers of conductive self-doping polyions in improving sensitivity of enzyme biosensors. Ultrathin films containing enzymes can be used to achieve direct electron exchange between redox cofactor sites and electrodes,1,2 avoiding the complications of mediators in devices such as biosensors and bioreactors. Peroxidases are iron heme enzymes that are useful tags in electrochemical biosensors,3,4 with * Corresponding author. E-mail: [email protected]. † Department of Chemistry. ‡ Institute of Materials Science. § Department of Pharmacology. (1) Armstrong, F. A.; Wilson, G. S. Electrochim. Acta 2000, 45, 2623-2645. (2) (a) Rusling, J. F.; Zhang, Z. In Handbook Of Surfaces And Interfaces Of Materials. Vol. 5. Biomolecules, Biointerfaces, and Applications; Nalwa, R. W., Ed.; Academic Press: San Diego, 2001; pp 33-71. (b) Rusling, J. F., Zhang, Z. In Electroanalytical Methods for Biological Materials; Chambers, J. Q., Bratjer-Toth, A., Eds.; Marcel Dekker: New York, 2002; pp 233-254. (3) Gorton, L.; Bremle, G.; Csoregi, E.; Jonsson-Pettersson, G.; Persson, B. Anal. Chim. Acta 1991, 249, 43-54. (4) Ruzgas, T.; Csoregi, E.; Emneus, J.; Gorton, L.; Marko-Varga, G. Anal. Chim. Acta 1996, 330, 123-138. 10.1021/ac034188r CCC: $25.00 Published on Web 07/23/2003

© 2003 American Chemical Society

applications that include DNA hybridization5-7 and immunosensing.8,9 Peroxidases on electrodes react with hydrogen peroxide to give a catalytic electrochemical reduction signal.3,4,10,11 Further, a peroxidase-tagged moiety bound to an antibody or oligonucleotide on an electrode gives a signal upon addition of hydrogen peroxide, providing the basis for a wide variety of biosensors. The reaction of horseradish peroxidase (HRP) with H2O2 to form an active oxidant is well documented.12 In this reaction, HRP forms compound I featuring an oxyferryl heme iron and a cation radical located on the porphyrin ring.13,14 In peroxidase catalysis, compound I can accept an electron from an organic substrate to form compound II (nonradical PFeIVdO, where P symbolizes protein), resulting in oxidation of the substrate. Compound II accepts a second electron to regenerate the ferric enzyme. Like peroxidases, the catalytic cycles of myoglobin (Mb), hemoglobin (Hb), and cytochrome P450s (cyt P450s) can be activated by H2O2, which is thought to generate reactive oxyferryl radical species similar to the formal oxidation state of compound I.15,16 We showed that Mb, HRP, and other peroxidases in lipid films give catalytic electrochemical reduction signals involving reactions with H2O2 by complex but similar pathways.17 (5) Palacek, E.; Fojto, M. Anal. Chem. 2001, 73, 75A-83A. (6) Azek, F.; Grossiord, C.; Joannes, M.; Limoges, B.; Brossier, P. Anal. Biochem. 2000, 284, 107-113. (7) Caruana, D. J.; Heller, A. J. Am. Chem. Soc. 1999, 121, 769-774. (8) Warsinke, A.; Benkert, A.; Scheller, F. W. Fresenius J. Anal. Chem. 2000, 366, 622-634. (9) Killard, A. J.; Micheli, L.; Grennan, K.; Franek, M.; Kolar, V.; Moscone, D.; Palchetti, I.; Smyth, M. R. Anal. Chim. Acta 2001, 427, 173-180. (10) Ruzgas, T., Lindgren, A., Gorton, L., Hecht, H.-J., Reichelt, J., Bilitewski, U. InElectroanalytical Methods for Biological Materials; Chambers, J. Q., BratjerToth, A., Eds.; Marcel Dekker: New York, 2002; pp 233-254. (11) Guo, Y.; Guadalupe, A. R. Chem. Commun. 1997, 1437-1438. (12) Everse, J., Everse, K. E., Grisham, M. B., Eds. Peroxidases in Chemistry and Biology; CRC Press: Boca Raton, Fl, 1991; Vol. II. (13) Schulz, C. E.; Rutter, R.; Sage, J. T.; Debrunner, P. G.; Hager, L. P. Biochemistry 1984, 23, 4743-4754. (14) Roberts, J. E.; Hoffman, B. M.; Rutter, R.; Hager L. P. J. Biol. Chem. 1981, 256, 2118-2121. (15) (a) Ortiz de Montellano, P. R, Ed. Cytochrome P450; Plenum Press: New York, 1995. (b) Schenkman, J. B., Greim H., Eds. Cytochrome P450; SpringerVerlag: Berlin, 1993. (16) (a) Ortiz de Montellano, P. R.; Catalano, C. E. J. Biol. Chem. 1985, 260, 9265-9271. (b) Rao, S. I.; Wilks, A.; Ortiz de Montellano, P. R. J. Biol. Chem. 1993, 268, 803-809. (c) Ortiz de Montellano, P. R.; Rao, S. I.; Wilks, A. Life Chem. Rep. 1994, 12, 29-32. (d) Choe, Y. S.; Rao, S. I.; Ortiz de Montellano, P. R. Arch. Biochem. Biophys, 1994, 314, 126-131.

Analytical Chemistry, Vol. 75, No. 17, September 1, 2003 4565

Entrapping enzymes in conductive polymer matrixes was reported by Foulds and Lowe in 1988.18 Films combining conductive polymers with enzymes may provide biosensors with enhanced sensitivity. Vectorial electron transport between electrode, enzyme, and substrate might become more efficient using a conducting polymer than in its absence by enhancing electron transport between electrode and enzyme, thereby improving signal-to-noise ratios in measured catalytic currents for substrates.19 Various transducer configurations employing such films for sensitive measurements of amperometry, resistance, or impedance are possible.20 Recent indicators of the viability of this approach involve films in which HRP was coated on electropolymerized polyaniline,20-22 and composite films of HRP, polycations, and poly(2-methoxyaniline-5-sulfonic acid).23 Redox metallopolymer hydrogels have also been used extensively to make films that “wire” enzymes to electrodes.24 Electrostatic layer-by-layer self-assembly of proteins and oppositely charged polyions was developed in the 1990s25-27 and has been used to make functional ultrathin films of more than 20 enzymes and proteins. In general, enzymes retain their native structures and activities in these films and can be used for catalysis. Film architecture can be tailored to predetermined specifications with this method. We first reported direct, reversible electron exchange between metalloproteins and electrodes in such films in 1998.28,29 We employed layered films of Mb and cyt P450cam with polyions to catalyze epoxidation of styrene and cismethylstyrene driven electrochemically or by adding hydrogen peroxide.2,28-31 Rubner et al. made conductive films using layerby-layer methods, employing both electrostatic attraction and hydrogen bonding to assemble films of conductive polyaniline with nonconductive polymers.32,33 Ultrathin conductive polyion layers have the potential to increase response speed and sensitivity in biosensor applications.20 In this paper, we explore this concept using proteins with (17) Zhang, Z.; Chouchane, S.; Magliozzo, R. S.; Rusling, J. F. Anal. Chem. 2002, 74, 163-170. (18) Foulds, N. C.; Lowe, C. R. Anal. Chem. 1988, 60, 2473-2478. (19) (a) Adeloju, S. B.; Wallace, G. G. Analyst 1996, 121, 699-703 and references therein. (b) Schuhmann, W. In Coordination Chemistry; Kauffman, G. B., Ed.; ACS Symposium Series 556; American Chemical Society: Washington, DC, 1994; pp 110-123. (20) Bartlett, P. N.; Astier, Y. Chem. Commun. 2000, 105-112. (21) Yang, Y.; Mu, S. J. Eletroanal. Chem. 1997, 432, 71-78. (22) Killard, A. J.; Zhang, S.; Zhao, H.; John, R.; Iwuoha, E. I.; Smyth, M. R. Anal. Chim. Acta 1999, 400, 109-119. (23) Tatsuma, T.; Ogawa, T.; Sato, R.; Oyama, N. J. Eletroanal. Chem. 2001, 501, 180-185. (24) Heller, A. Acc. Chem. Res. 1990, 23, 128-135. (25) Lvov, Y. M.; Decher, G. Crystallogr. Rep. 1994, 39, 628-647. (26) Lvov, Y. In Protein Architecture: Interfacing Molecular Assemblies and Immobilization Biotechnology; Lvov, Y., Mo ¨hwald, H., Eds.; Marcel Dekker: New York, 2000; pp 125-167. (27) Lvov, Y. In Handbook Of Surfaces And Interfaces Of Materials. Vol. 3. Nanostructured Materials, Micelles and Colloids; Nalwa, R. W., Ed.; Academic Press: San Diego, 2001; pp 170-189. (28) Lvov, Y. M.; Lu, Z.; Schenkman, J. B.; Zu, X.; Rusling, J. F. J Am. Chem. Soc. 1998, 120, 4073-4080. (29) Ma, H.; Hu, N.; Rusling, J. F. Langmuir 2000, 16, 4969-4975. (30) Zu, X.; Lu, Z.; Zhang, Z.; Schenkman, J. B.; Rusling, J. F. Langmuir 1999, 15, 7372-7377. (31) Rusling, J. F. In Protein Architecture: Interfacing Molecular Assemblies and Immobilization Biotechnology; Lvov, Y., Mo ¨hwald, H., Eds.; Marcel Dekker: New York, 2000; pp 337-354. (32) Cheung, J. H.; Stockton, W. B.; Rubner, M. F. Macromolecules 1997, 30, 2712-2716. (33) Stockton, W. B.; Rubner, M. F. Macromolecules 1997, 30, 2717-2725.

4566

Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

peroxidase activity, specifically Mb and HRP. We report herein that underlayers of sulfonated polyaniline (SPAN) of several nanometers nominal thickness greatly increased the sensitivity of films of polyanions and these proteins to hydrogen peroxide in voltammetry and amperometry. We used SPAN with 50% sulfonate per phenyl group since it is water soluble and amenable to layerby-layer electrostatic film assembly. This material is self-doped and is highly conductive in the emeraldine oxidation state over the pH range 0-7,34 enabling its use with native enzymes having optimal catalytic activities in the medium pH range. EXPERIMENTAL SECTION Chemicals. Horse heart Mb (MW ) 17 000) and HRP (MW ) 47 000) were from Sigma. Poly(sodium styrenesulfonate) (PSS, average MW 70 000) and poly(diallydimethylammonium chloride) (PDDA) were from Aldrich. SPAN (∼50% sulfonate groups by elemental analysis) was made by a published procedure.34a,b Water was purified with a Hydro Nanopure system to specific resistance of >16 mΩ‚cm. All other chemicals were reagent grade. Electrochemical Measurements. A CHI 430 electrochemical workstation was used for cyclic voltammetry and amperometry. A three-electrode cell was used employing a saturated calomel reference electrode (SCE), a platinum wire counter electrode, and a film-coated working electrode disk of ordinary basal plane pyrolytic graphite (PG, Advanced Ceramics, A ) 0.2 cm2). Voltammetry and amperometry were done at ambient temperature (22 ( 2 °C) in buffers of 0.05 M acetate, 0.05 M NaCl, pH 4.5, for films with Mb, or 0.02 M potassium phosphate, 0.05 M NaCl, pH 6.0, for films with HRP, unless otherwise noted. Buffers were purged with purified nitrogen for 20 min prior to a series of experiments. A nitrogen environment was maintained in the cell. Ohmic drop was compensated >95% by the CHI 430. Amperometry was done with a PG disk electrode rotated at 1000 rpm. Film Assembly. Films for electrochemistry were constructed on basal plane PG disk electrodes that were first abraded on 400grit SiC paper and then ultrasonicated in water for 1 min. Films were grown by alternate adsorption from aqueous solutions of polyions and proteins, using conditions established by previous studies.28,30 PG electrodes were first immersed in aqueous PDDA (2 mg mL-1) for 10 min. After being rinsed with water, these electrodes were then placed in pH 4.5 acetate buffer containing 3 mg mL-1 SPAN for 20 min. After rinsing again, the electrodes were then placed in 3 mg mL-1 pH 5.5 acetate buffer containing protein (HRP or Mb, where these proteins have a positive charge)28 for 20 min after which electrodes were rinsed again. These electrodes were then placed in 3 mg mL-1 PSS in 0.5 M NaCl for 20 min and then washed. The last two steps were repeated to make films with architecture denoted PDDA/SPAN/ (protein/PSS)3. Quartz Crystal Microbalance (QCM). A QCM (USI Japan) was used to monitor film assembly. QCM resonators (9 MHz, AT-cut, International Crystal Mfg. Co.) were covered by gold electrodes (A ) 0.16 cm2). To mimic a PG surface, Au-quartz (34) (a) Yue, J.; Epstein, A. J. J. Am. Chem. Soc. 1990, 112, 2800-2801. (b) Yue, J.; Wang, Z. H.; Cromack, K. R.; MacDiarmid, A. G. J. Am. Chem. Soc. 1991, 113, 2665-2671. (c) Wei, X.; Epstein, A. J. Synth. Met. 1995, 74, 123-125. (d) Wei, X.; Wang, Y. Z.; Long, S. M.; Bobeczko, C.; Epstein, A. J. J. Am. Chem. Soc. 1996, 118, 2545-2555.

Table 1. Average Characteristics of Protein Films from QCMa and CV

a

film achitecture

film d, nm

SPAN d, nm

total protein, nmol cm-2

Γb, nmol cm-2 (protein)

electroactive %

PDDA/SPAN/(Mb/PSS)3 PDDA/SPAN/(HRP/PSS)3 PDDA/(Mb/PSS)3 PDDA/(HRP/PSS)3

29 29 25 25

3.8 ( 0.7 3.8 ( 0.8

0.23 ( 0.02 0.088 ( 0.011 0.29 ( 0.02 0.098 ( 0.0014

0.196 ( 0.009 0.082 ( 0.004 0.106 ( 0.002 0.045 ( 0.002

85 93 37 46

As average of eight trials, d is the nominal thickness from eq 2. b Γ, amount of electroactive protein.

resonators were immersed into 0.7 mM 3-mercapto-1-propanol and 0.3 mM 3-mercaptopropionic acid in ethanol for 24 h and then washed.35 Films were then prepared as on PG. Resonators were immersed in adsorbate solutions, rinsed, and then dried in a stream of nitrogen, and the frequency change was measured. Similar procedures were used to make films for electrochemical quartz crystal microbalance (EQCM). The relationship between frequency change ∆F (Hz) for the 9-MHz resonators and the adsorbed mass per unit area M/A (g/cm2) is36

M/A ) -∆F/(1.83 × 108)

(1)

and utilizing the density of the materials nominal film thickness was estimated from36

d (nm) ≈ -(0.016 ( 0.002)∆F (Hz)

(2)

EQCM studies were done with the CHI model 430, using a CHI EQCM cell with a 8-MHz quartz crystal fitted into the bottom. RESULTS Film Characterization by QCM. Figure 1 shows a negative linear progression of frequency change with layer number for films made with initial layers of PDDA and SPAN and then alternate layers of protein and PSS. These results demonstrate reproducible formation of multiple layers during film construction. From eq 2 and the QCM data, we estimate nominal thicknesses of 2 nm for PDDA and ∼4 nm for layers of SPAN, Mb, HRP, and PSS. Table 1 summarizes the thicknesses and amounts of protein in PDDA/ SPAN/(protein/PSS)3 films and in equivalent films without SPAN. Roughly the same mass of each protein is adsorbed per adsorption step, but since Mb has a much smaller MW than HRP, this translated into larger molar amounts of Mb in the films than HRP. Voltammetry of PDDA/SPAN. Preliminary results with SPAN adsorbed directly onto PG electrodes showed that the films were not very stable. Also, when protein/PSS layers were grown on top of a single SPAN layer adsorbed on PG, the protein voltammetry was not chemically reversible, with reduction peaks much larger than oxidation peaks. Much more stable films and better reversibility were obtained when a layer of positively charged PDDA was first adsorbed, followed by the negative SPAN. Here the positive PDDA layer acts as an “electrostatic glue” to anchor a stable SPAN layer to the electrode surface. (35) Zhou, L.; Rusling, J. F. Anal. Chem. 2001, 73, 4780-4786. (36) Lvov, Y.; Ariga, K.; Ichinose, I.; Kunitake, T. J. Am. Chem. Soc. 1995, 117, 6117-6123.

Figure 1. QCM frequency shifts for cycles of alternate adsorption during growth of PDDA/SPAN/(protein/PSS)3 films on gold resonators first coated with mixed monolayer of mercaptoproionic acid and mercaptopropanol, with (a) Mb as the protein and (b) HRP as the protein. Error bars represent standard deviations for eight films.

Cyclic voltammetry of PG electrodes coated with a layer of SPAN (4 nm) on top of a layer of PDDA (2 nm) in pH 4.5 buffer (Figure 2a) showed two pairs of well-defined, reproducible, chemically reversible peaks. One peak pair was centered around 0.14 V versus SCE, and a second one was centered at -0.4 V versus SCE. These two sets of peaks give a pattern similar to thicker films of SPAN (50%) on Pt electrodes34d but are at considerably more negative potentials. CV comparisons of PDDA/ SPAN and SPAN films on PG electrodes showed voltammograms for adsorbed SPAN nearly identical to those reported for thicker SPAN layers on Pt electrodes34d at pH 1 but, for PDDA/SPAN films shifts of about -50 mV for the more positive peaks and -100 mV for the more negative peak, were found between pH 1 and 6. Peak potentials of PDDA/SPAN and SPAN films varied linearly with pH with slopes of ∼-70 mV pH-1. By analogy with the previously reported work,34 the oxidation peak at 0.14 V can be assigned to conversion of the conductive emeraldine form of SPAN to the insulating permigranline. The -0.45-V reduction peak can be assigned to the reduction of the emeraldine form to the insulating leucoemeraldine form. Thus, Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

4567

Figure 2. (a) Cyclic voltammograms of films on PG electrodes in pH 4.5 buffer at 0.3 V s-1. Solid line is PDDA/SPAN and dashed line is PDDA. (b) EQCM trace of the same SPAN/PDDA film in pH 4.5 buffer.

the conductive region of PDDA/SPAN films is between 0.15 and -0.45 V versus SCE. PDDA films showed no features in cyclic voltammetry. EQCM showed a large decrease of mass at the 0.14 reduction peak, suggesting that loss of bound cations from the film occurs (Figure 2b). This is consistent with the reported 2e-, 4H+ reduction process,34a,b The peak at -0.45 V involves no net change in charge or net mass, since it is a 2e-, 2H+ process, and no mass change is expected or observed in the negative potential range in EQCM. Cyclic voltammetry of PDDA/SPAN films showed negligible changes upon addition of up to 10 µM H2O2 (See Supporting Information.). Similarly, amperometry of SPAN films at 0 V versus SCE gave no response when 10 µM hydrogen peroxide was repeatedly injected into the electrochemical cell. These results are consistent with the documented insensitivity of polyaniline to hydrogen peroxide.20 Voltammetry of Protein Films. Cyclic voltammetry of PDDA/ SPAN/(Mb/PSS)3 revealed a pair of quasi-reversible peaks centered about -0.25 V versus SCE (Figure 3a). This potential range is characteristic of the Mb FeIII/FeII redox couple in PSS films.28 The Mb peaks seem to be superimposed on the SPAN peaks, and peaks for the conductive polymer can be seen at 0.15 V. The oxidation-reduction peak separation at 0.3 V s-1 was 100 mV, and the peak width at half-height was 240 mV. While considerable chemical reversibility is evident, the ratio of reduction/oxidation peaks is 1.5, possibly suggesting some unspecified interaction of SPAN and oxidized Mb. Reduction peak current for Mb was linearly proportional to scan rate up to 1 V s-1. These results are consistent with nonideal thin-layer voltammetry typical of thin protein films.2 4568 Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

Figure 3. Cyclic voltammograms in pH 4.5 buffer at 0.3 V s-1: (a) PDDA/SPAN/(Mb/PSS)3 film; (b) PDDA/(Mb/PSS)3 film.

Comparison with the CV of PDDA/(Mb/PSS)3 films (without SPAN) shows a qualitatively similar voltammogram with smaller peak currents than when SPAN is present (Figure 3b). The midpoint potential was -0.24 V versus SCE at 0.3 V s-1, oxidation-reduction peak separation was 90 mV, and the peak width at half-height was 200 mV. The ratio of reduction/oxidation peaks is 1.3, suggesting slightly better chemical reversibility than in the films with SPAN underlayers. Peak current for Mb was linearly proportional to scan rate up to 2 V s-1. The peak current of the film without SPAN in Figure 3 is smaller, and the amount of electroactive protein (Γ) averaged from integrating CVs in the low scan rate range below 0.1 V s-1 is ∼2.5-fold larger with SPAN than without SPAN (Table 1). Comparison of Γ values with total amount of protein from QCM shows that films with SPAN had ∼85% of the total Mb electroactive. As reported previously,28 repetitive scans of Mb/PSS films were nearly superimposable after the initial scan. On the other hand, PDDA/SPAN/(Mb/PSS)3 films showed small but significant decreases in peak height upon repetitive scanning. However, within several minutes in the buffer at open circuit, peak current recovered to its initial height or, after longer times, to a value slightly larger (See Supporting Information.). This slight hysteresis in CV did not greatly influence the responses to hydrogen peroxide described below. For films containing HRP, CVs of PDDA/SPAN/(HRP/PSS)3 gave 6-fold larger peak currents at 0.3 V s-1 than PDDA(HRP/ PSS)3 (Figure 4). The characteristic heme FeIII/FeII reductionoxidation peaks were found at -0.20 V versus SCE in films with SPAN. The peak shape was unsymmetric at 0.3 V s-1, suggesting the influence of charge diffusion.2 However, the peaks are still well defined and chemically reversible. HRP films without SPAN had peaks that were much broader and smaller, centered at ∼-0.23 V. The amount of electroactive HRP estimated by integrating CVs in a lower scan rate range, where the peak size

Figure 4. Cyclic voltammograms of films on PG electrodes in pH 6 buffer at 0.3 V s-1. Solid line is PDDA/SPAN/(HRP/PSS)3; dashed line is PDDA/(HRP/PSS)3.

difference with and without SPAN is not so large, was ∼2-fold larger with SPAN than without SPAN (Table 1). Films with SPAN have ∼93% of the total HRP electroactive. PDDA/SPAN/(protein/PSS)3 films were stored in buffer and in air, and CVs were measured periodically. In both cases, the Mb and HRP films showed good stability with less than 10% loss of CV peak height over 3 weeks. Catalytic Reduction of Hydrogen Peroxide. H2O2 converts the iron heme cofactors of Mb and HRP to oxyferryl radicals that can be reduced back to the FeIII form. When Mb or HRP is coupled to an electrode, a complex catalytic cycle for the reduction of H2O2 is set up that can be detected via a catalytic reduction current.3,4,17 Addition of nanomolar amounts of H2O2 to buffer solutions bathing PDDA/SPAN/(protein/PSS)3 films gave large increases in CV reduction current at the potential of the FeIII peak, accompanied by a disappearance of the FeII oxidation peak. This behavior is illustrated for PDDA/SPAN/(Mb/PSS)3 in Figure 5a at pH 4.5, where Mb is known to be in its native conformation in PSS films. It is characteristic of enzyme-catalyzed electrochemical reduction2 and is qualitatively similar to recently reported results for Mb and HRP in thin lipid films.17 In films without SPAN, 10fold larger amounts of H2O2 were need to get the same catalytic current (Figure 5b). Plots of catalytic peak current versus [H2O2] were linear up to ∼4 µM peroxide, and the sensitivity as judged by the slopes was larger, 3-fold for HRP and 10-fold for Mb, when SPAN was in the film compared to when it was absent (Figure 6). The slope for HRP films at pH 6, where HRP has near-optimal catalytic activity, was slightly larger than for Mb films at pH 4.5 Amperometric currents at 0 V versus SCE at rotating disk electrodes gave a more sensitive response to H2O2. A current increase of several hundred nanoamperes was observed within a few seconds after addition of H2O2 at 50 nM, illustrated for PDDA/ SPAN/(HRP/PSS)3 in Figure 7. Detection limits for H2O2 were ∼3 nM (signal at 3 times the noise level) in these pure solutions for a PDDA/SPAN/(protein/PSS)3 and were about 4-7-fold larger in equivalent films without SPAN. Plots of amperometric current after peroxide addition were linear up to ∼0.2 µM H2O2 for PDDA/SPAN/(HRP/PSS)3 and up to 0.5 µM H2O2 for PDDA/SPAN/(Mb/PSS)3 films (Figure 8). As for CV peak currents, the slope of amperometric current versus [H2O2] for HRP films with SPAN at 14.5 µA µM-1 cm-2 was slightly larger than for Mb films with SPAN at 11.0 µA µM-1 cm-2. These values were 5-6-fold larger than the slopes for

Figure 5. Cyclic voltammograms in pH 4.5 buffer at 0.3 V s-1 showing catalytic peaks for reduction of hydrogen peroxide: (a) PDDA/SPAN/(Mb/PSS)3 film; (b) PDDA/(Mb/PSS)3.

Figure 6. Influence of concentration of H2O2 on the catalytic CV peak current at 0.3 V s-1 for PDDA/SPAN/(protein/PSS)3 and PDDA/ (protein/PSS)3 films with Mb as the protein at pH 4.5 and HRP as the protein at pH 6.0.

corresponding films without SPAN, which were 1.9 µA µM-1 cm-2 for Mb and 3.0 µA µM-1 cm-2 for HRP. The higher sensitivity for HRP films is consistent with the 7-fold higher amperometric peroxidase activity than Mb found in lipid films.17 However, the amperometric sensitivity to peroxide depends not only on enzyme activity but also on the molar amount of protein in the films, which is smaller for the larger HRP molecules than for Mb (Table 1). Thus, the sensitivities per nanomole of protein for films with SPAN present are 165 µA µM-1 for HRP and 48 µA µM-1 for Mb. The stability of Mb and HRP films was tested in buffers containing 0.5, 5, and 50 µM hydrogen peroxide. After 30 min of repetitive CV scanning at 0.1 V s-1 through the region 0.4 to -0.8 V versus SCE, washing and returning to peroxide-free buffer gave Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

4569

Scheme 1. Possible Electrochemical Reduction Pathways Involving Peroxidases

Figure 7. Amperometry in pH 6 buffer at 0 V vs SCE with PDDA/ SPAN/(HRP/PSS)3 and PDDA/(HRP/PSS)3 films on rotating disk electrodes at 1000 rpm illustrating the response to hydrogen peroxide. Each step in current corresponds to the addition of 50 nM H2O2.

Figure 8. Influence of concentration of hydrogen peroxide on amperometric response at 0 V vs SCE for PDDA/SPAN/(protein/ PSS)3 and PDDA/(protein/PSS)3 on rotating disk electrodes at 1000 rpm for Mb films at pH 4.5 and HRP films at pH 6.0.

Figure 9. Amperometry with PDDA/SPAN/(HRP/PSS)3 films on rotating disk electrodes at 1000 rpm illustrating the response to hydrogen peroxide at various applied potentials. Each step in current corresponds to 50 nM addition of H2O2.

CV peak currents within ∼(5% of those obtained before peroxide exposure, with slightly larger decreases (∼10%) at the highest peroxide concentration. Furthermore, amperometric responses of those electrodes to 0.5, 5, and 50 µM hydrogen peroxide were within ∼5% of those obtained before peroxide exposure. Rotating disk electrode amperometry was also done at various applied potentials. Figure 9 and the resulting plots of current versus [H2O2] illustrate that the sensitivity to H2O2 was ∼3-fold 4570

Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

larger at 0 V than at -0.6 V or at 0.4 V. Similar results were found with both Mb and HRP films containing SPAN. That is, amperometric sensitivity was much larger at 0 V than at -0.6 V or at 0.4 V. At -0.6 V, SPAN is much less conductive than at 0 V. However, in Mb or HRP films without SPAN, sensitivity at 0 and -0.6 V was comparable, and sensitivity at 0.4 V was ∼25% smaller and comparable to that of films with SPAN at 0.4 V. DISCUSSION Results presented herein demonstrate the feasibility of constructing enzyme/polyion films on top of ultrathin conductive polymer layers nominally 4 nm thick (Table 1 and Figure 1) by using layer-by-layer alternate electrostatic assembly. SPAN provides a water-soluble material compatible with this film assembly method and has conductivity over a wide pH range compatible with optimal enzyme activity. Comparison of films with and without the conductive polymer layer clearly shows that SPAN provides efficient coupling of protein redox centers to the electrode (Figures 3 and 4, Table 1) such that roughly 90% of the protein exchanges electrons with the electrode in films with SPAN, while only ∼40% is electroactive in the absence of SPAN. Efficient wiring of HRP and Mb by SPAN in the films seems to be responsible for the high sensitivity of electrochemical catalytic reduction signals to hydrogen peroxide (Figures 5-8). As mentioned, addition of hydrogen peroxide to these iron heme proteins converts them to oxyferryl radicals (•PFeIVdO) that can be reduced to the ferric enzymes and water. However, hydrogen peroxide can also react with the oxyferryl radical to give ferric enzyme and oxygen, which can also be catalytically reduced, yielding hydrogen peroxide (Scheme 1). Thus, several catalytic cycles can be operative at a given applied potential.17 Amperometric sensitivity in films with SPAN for HRP films was 14.5 µA µM-1 cm-2 and for Mb was 11.0 µA µM-1 cm-2, roughly 6-fold larger than sensitivities for corresponding films without SPAN. Detection limits for the Mb and HRP films with SPAN without attempts at optimization were on the order of 3 nM. To place the above results in context, a recent report describes a mediator-containing carbon paste electrode containing HRP with a detection limit of 1 nM hydrogen peroxide37 and sensitivity comparable to our SPAN/HRP and SPAN/Mb electrodes. To our knowledge, this seems to be the mediated HRP system with the best detection limit and sensitivity reported to date. Mediatorless HRP-based electrodes have been reported38 with sensitivities up to 0.6 µA µM-1 cm-2, much smaller than with our enzyme/SPAN (37) Razola, S. S.; Aktas, E.; Vire, J.-C.; Kauffmann, J.-M. Analyst 2000, 125, 79-85 and references therein. (38) Ho, W. O.; Athey, D.; McNeil, C. J.; Hager, H. J.; Evans, G. P.; Mullen, W. H. J. Electroanal. Chem. 1993, 351, 185-197.

films. Films prepared on electrodes with electropolymerized polyaniline and adsorbed HRP21 or a composite mixture of poly(2-methoxyaniline-5-sulfonic acid), polycation, and HRP23 had detection limits 3-30-fold larger and poorer sensitivities than our SPAN/enzyme films. Our results suggest that the enhanced electrochemical sensitivity to hydrogen peroxide is a consequence of the efficient wiring of the active enzymes with SPAN. Similar enhanced performance for light-emitting diodes was found when the active luminescent layer was interfaced with a high surface area polyaniline underlayer.39 We expect that the SPAN layer is overlapped with one or more of the protein layers. Extensive mixing of neighboring layers in films grown layer by layer was deduced from neutron reflectance analysis of films of PSS and polycations,40 and PSS and Mb.27 While the films are constructed one layer at a time, the final structure features considerable intermixing of layers that may facilitate charge transport through these films. Matching the electroactive voltage range of SPAN with enzyme redox potentials may also be important. We found that the electroactive potential range of SPAN/PDDA films (Figure 2) was about 0.1 to -0.4 V versus SCE (Figure 2), shifted 50-100 mV negative compared to pure SPAN films on electrodes. This is likely the consequence of partial undoping of SPAN by the cationic PDDA underlayer. Nevertheless, the resulting electroactive range brackets the peak potentials of Mb and HRP in the films. At potentials where SPAN is insulating, e.g., -0.6 and 0.4 V, sensitivity to hydrogen peroxide was much less than it is at 0 V, where the SPAN layer is conducting (Figure 9). On the other hand, control amperometry done on films containing enzyme but no SPAN gave similar sensitivities at 0 and -0.6 V and 25% less sensitivity at 0.4 V, which was similar to the sensitivity of SPAN(39) Yang, Y.; Westerweele, E.; Zhang, C.; Smith, P.; Heegar, A. J. J. Appl. Phys. 1995, 77, 694-698. (40) Decker, G. Science 1997, 227, 1232-1237.

containing films at this applied potential. Thus, at -0.6 and 0.4 V, the conductivity of SPAN is unimportant, and it behaves like any nonconducting polyanion. In summary, an ultrathin layer of the sulfonated conductive polyion SPAN underneath enzyme/PSS films grown by layer-bylayer electrostatic assembly greatly improved the efficiency of electron transfer between enzymes and electrodes such that nearly all the redox sites in the enzyme films were efficiently addressed. As a consequence, much higher sensitivity of the enzyme films to hydrogen peroxide was achieved in an unmediated sensor configuration. These results may have important consequences for the future development of more sensitive biosensors, which are being pursued in our laboratory. Further, transistor-based biosensors20 utilizing ultrathin conductive self-doped polyion underlayers may provide even better sensitivity. ACKNOWLEDGMENT This work was supported by U.S. PHS Grant ES03154 from the National Institute of Environmental Health Sciences (NIEHS), NIH, and by grants from the Army Research Office (ARO), and the North Atlantic Treaty Organization (NATO, Contract DAAD1902-1-0381). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of NIH, ARO, or NATO. SUPPORTING INFORMATION AVAILABLE Two additional figures documenting the lack of influence of peroxide on SPAN films and hysteresis of SPAN/Mb film CV peaks under continual scanning. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review February 25, 2003. Accepted June 4, 2003. AC034188R

Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

4571