Cytochrome c Mutants for Superoxide Biosensors - Analytical

Mar 18, 2009 - Cytochrome c Mutants for Superoxide Biosensors ... The 11 mutants of human cyt c expressed in the course of this research ... 2015,301-...
0 downloads 0 Views 702KB Size
Anal. Chem. 2009, 81, 2976–2984

Cytochrome c Mutants for Superoxide Biosensors Franziska Wegerich,†,‡ Paola Turano,§,| Marco Allegrozzi,§ Helmuth Mo ¨ hwald,† and Fred Lisdat*,⊥ Interfaces, Max Planck Institute of Colloids and Interfaces, 14476 Potsdam-Golm, Germany, Analytical Biochemistry, University of Potsdam, 14476 Potsdam-Golm, Germany, Magnetic Resonance Center, Department of Chemistry, University of Florence, 50019 Sesto Fiorentino (FI), Italy, and Biosystems Technology, Wildau University of Applied Sciences, 15745 Wildau, Germany The effect of introducing positive charges (lysines) in human cytochrome c (cyt c) on the redox properties and reaction rates of cyt c with superoxide radicals was studied. The mutated forms of this electron-transfer protein are used as sensorial recognition elements for the amperometric detection of the reactive oxygen radical. The proteins were prepared by site-directed mutagenesis focusing on amino acids near the heme edge. The 11 mutants of human cyt c expressed in the course of this research have been characterized by UV-vis spectroscopy, circular dichroism, and NMR spectroscopy to verify overall structure integrity as well as axial coordination of the heme iron. The mutants are investigated voltammetrically using promoter-modified gold electrodes with respect to redox activity and formal redox potential. The rate constants for the reaction with superoxide have been determined spectrophotometrically. Four mutants show a higher reaction rate with the radical as compared to the wild type. These mutants are used for the construction of superoxide sensors based on thiol-modified gold electrodes and covalently fixed proteins. We found that the E66K mutant-based electrode has a clearly higher sensitivity in comparison with the wild-type-based sensor while retaining the high selectivity and showing a good storage stability. The superoxide anion radical, a reactive oxygen species, is produced by one-electron transfer to oxygen and is present in several pathophysiological situations, including reperfusion and sepsis.1,2 On-line detection of this short-lived species is thus particularly relevant in the medical field and can be realized by electrochemical biosensors.3,4 Therefore, either the natural reac* To whom correspondence should be addressed. Fax: +49-331-977-5153. E-mail: [email protected]. † Max Planck Institute of Colloids and Interfaces. ‡ University of Potsdam. § Magnetic Resonance Center, University of Florence. | Department of Chemistry, University of Florence. ⊥ Wildau University of Applied Sciences. (1) Valko, M.; Leibfritz, D.; Moncol, J.; Cronin, M. T. D.; Mazur, M.; Telser, J. Int. J. Biochem. Cell Biol. 2007, 39, 44–84. (2) Valko, M.; Rhodes, C. J.; Moncol, J.; Izakovic, M.; Mazur, M. Chem.-Biol. Interact. 2006, 160, 1–40. (3) Lisdat, F. In Encyclopedia of Sensors; Grimes, C., Pishko, M., Eds.; American Scientific Publishers, 2005. (4) Prieto-Simon, B.; Cortina, M.; Campas, M.; Calas-Blanchard, C. Sens. Actuators, B 2008, 129, 459–466.

2976

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

tion partner, superoxide dismutase (SOD), or the redox protein cytochrome c (cyt c) is preferentially used as the recognition element. First-generation SOD-based biosensors detect superoxide indirectly by measuring hydrogen peroxide concentration. H2O2 is enzymatically generated during superoxide reaction within immobilized SOD and then further oxidized at the electrode surface. Several approaches have been developed to enhance the selectivity for these types of sensors: a H2O2-impermeable Teflon membrane,5 combination with horseradish peroxidase,6 the entrapment of SOD between two membranes,7 and the incorporation of SOD in a polypyrrol film.8,9 Another approach (second-generation SOD-based biosensors) used different mediators, such as ferrocene-carboxyaldehyde or methyl viologen, to enable electron transfer from SOD to the electrode.10,11 Since the direct electron transfer of SOD to electrodes, without the need of mediators, results in a simple sensor design with high sensitivity and selectivity, third-generation biosensors have gained considerable attraction. The electron transfer to SOD was shown to be promoted using thiol self-assembled monolayers (SAMs),12-15 an electroactive sol-gel film,16 or a ZnO nanodisc film.17 SOD was also directly immobilized on gold using SOD mutants with additional cysteine residues.18,19 Superoxide biosensors using SOD often lack reproducibility due to immobilization problems. In contrast, third-generation cyt (5) Song, M. I.; Bier, F. F.; Scheller, F. W. Bioelectrochem. Bioenerg. 1995, 38, 419–422. (6) Lvovich, V.; Scheeline, A. Anal. Chim. Acta 1997, 354, 315–323. (7) Campanella, L.; Bonanni, A.; Favero, G.; Tomassetti, M. Anal. Bioanal. Chem. 2003, 375, 1011–1016. (8) Descroix, S.; Bedioui, F. Electroanalysis 2001, 13, 524–528. (9) Mesaros, S.; Vankova, Z.; Grunfeld, S.; Mesarosova, A.; Malinski, T. Anal. Chim. Acta 1998, 358, 27–33. (10) Endo, K.; Miyasaka, T.; Mochizuki, S.; Aoyagi, S.; Himi, N.; Asahara, H.; Tsujioka, K.; Sakai, K. Sens. Actuators, B 2002, 83, 30–34. (11) Ohsaka, T.; Shintani, Y.; Matsumoto, F.; Okajima, T.; Tokuda, K. Bioelectrochem. Bioenerg. 1995, 37, 73–76. (12) Ge, B.; Scheller, F. W.; Lisdat, F. Biosens. Bioelectron. 2003, 18, 295–302. (13) Tian, Y.; Mao, L.; Okajima, T.; Ohsaka, T. Anal. Chem. 2002, 74, 2428– 2434. (14) Tian, Y.; Mao, L. Q.; Okajima, T.; Ohsaka, T. Anal. Chem. 2004, 76, 4162– 4168. (15) Tian, Y.; Mao, L. Q.; Okajima, T.; Ohsaka, T. Biosens. Bioelectron. 2005, 21, 557–564. (16) Di, J. W.; Bi, S. P.; Zhang, M. Biosens. Bioelectron. 2004, 19, 1479–1486. (17) Deng, Z. F.; Rui, Q.; Yin, X.; Liu, H. Q.; Tian, Y. Anal. Chem. 2008, 80, 5839–5846. (18) Beissenhirtz, M. K.; Scheller, F. W.; Viezzoli, M. S.; Lisdat, F. Anal. Chem. 2006, 78, 928–935. (19) Kapp, A.; Beissenhirtz, M. K.; Geyer, F.; Scheller, F.; Viezzoli, M. S.; Lisdat, F. Electroanalysis 2006, 18, 1909–1915. 10.1021/ac802571h CCC: $40.75  2009 American Chemical Society Published on Web 03/18/2009

c based superoxide sensors are more stable and have already been used for in vivo applications.20,21 Here the reduction of the heme protein by superoxide and subsequent reoxidation by an electrode is used. Short-chain thiol-modified gold electrodes used for the immobilization of cyt c show a highly efficient communication between cyt c and the electrode22-25 and have been applied for cyt c based superoxide sensors.26 However, they are not completely blocking the electrode surface. For this reason, sensor electrodes using long-chain thiols have been developed to diminish effectively electroactive interferences. These cyt c electrodes have been applied for in vitro measurements of the radical27-32 as well as for in vivo studies.20,21 The sensitivity of these protein electrodes is dependent on the amount of electroactive protein and the reaction rate of protein with the radical. Thus, different approaches have been used for sensitivity enhancement: optimization of the promotor layer,33,34 increasing the surface roughness,35 and the construction of protein multilayer electrodes.36 Here we report on the preparation of cyt c mutants with a higher reaction rate with O2-. The design of the cyt c variants is based on studies of SOD mutants, which demonstrated an enhanced reaction rate with superoxide when positively charged amino acids were introduced near the activesite channel.37 On the basis of the assumption that enhanced electrostatic guidance of the negatively charged radical can also increase the reaction rate of cyt c with superoxide and thus the sensitivity of a cyt c based biosensor, we replaced neutral and negatively charged amino acids with positively charged lysines at single-mutation sites. Here we will present the results of the first mutation group containing 11 selected single-mutation sites. The proteins will be investigated using UV-vis spectroscopy, circular dichroism, and NMR spectroscopy to obtain information on the protein fold and characterize heme iron environment. Further, the mutants are studied using cyclic voltammetry and amperometry to get infor(20) Buttemeyer, R.; Philipp, A. W.; Mall, J. W.; Ge, B. X.; Scheller, F. W.; Lisdat, F. Microsurgery 2002, 22, 108–113. (21) Scheller, W.; Jin, W.; Ehrentreich-Forster, E.; Ge, B.; Lisdat, F.; Buttemeier, R.; Wollenberger, U.; Scheller, F. W. Electroanalysis 1999, 11, 703–706. (22) Frew, J. E.; Hill, H. A. O. Eur. J. Biochem. 1988, 172, 261–269. (23) Hinnen, C.; Parsons, R.; Niki, K. J. Electroanal. Chem. 1983, 147, 329– 337. (24) Nahir, T. M.; Clark, R. A.; Bowden, E. F. Anal. Chem. 1994, 66, 2595– 2598. (25) Taniguchi, I.; Toyosawa, K.; Yamaguchi, H.; Yasukouchi, K. J. Electroanal. Chem. 1982, 140, 187–193. (26) Tammeveski, K.; Tenno, T. T.; Mashirin, A. A.; Hillhouse, E. W.; Manning, P.; McNeil, C. J. Free Radical Biol. Med. 1998, 25, 973–978. (27) Beissenhirtz, M.; Scheller, F.; Lisdat, F. Electroanalysis 2003, 15, 1425– 1435. (28) Ignatov, S.; Shishniashvili, D.; Ge, B.; Scheller, F. W.; Lisdat, F. Biosens. Bioelectron. 2002, 17, 191–199. (29) Krylov, A. V.; Sczech, R.; Lisdat, F. Analyst 2007, 132, 135–141. (30) Lisdat, F.; Ge, B.; Ehrentreich-Forster, E.; Reszka, R.; Scheller, F. W. Anal. Chem. 1999, 71, 1359–1365. (31) Lisdat, F.; Ge, B.; Reszka, R.; Kozniewska, E. Fresenius’ J. Anal. Chem. 1999, 365, 494–498. (32) Beissenhirtz, M. K.; Kwan, R. C. H.; Ko, K. M.; Renneberg, R.; Scheller, F. W.; Lisdat, F. Phytother. Res. 2004, 18, 149–153. (33) Ge, B.; Lisdat, F. Anal. Chim. Acta 2002, 454, 53–64. (34) Gobi, K. V.; Mizutani, F. J. Electroanal. Chem. 2000, 484, 172–181. (35) Krylov, A. V.; Beissenhirtz, M.; Adamzig, H.; Scheller, F. W.; Lisdat, F. Anal. Bioanal. Chem. 2004, 378, 1327–1330. (36) Beissenhirtz, M. K.; Scheller, F. W.; Lisdat, F. Anal. Chem. 2004, 76, 4665– 4671. (37) Getzoff, E. D.; Cabelli, D. E.; Fisher, C. L.; Parge, H. E.; Viezzoli, M. S.; Banci, L.; Hallewell, R. A. Nature 1992, 358, 347–351.

mation on the electrochemical properties and the sensing characteristics for the detection of superoxide. EXPERIMENTAL SECTION Materials. Xanthine oxidase (XOD) was purchased from Roche (Mannheim, Germany). Horse heart cytochrome c, bovine erythrocyte superoxide dismutase (SOD), lysozyme, hypoxanthine, H2O2 (30% solution in water), isopropyl β-D-1-thiogalactopyranoside (IPTG), dithiothreitol (DTT), 1-ethyl-3(3-dimethyl aminopropyl) carbodiimide (EDC), 11-mercapto-1-undecanoic acid (MUA), 11-mercapto-1-undecanol (MU), 3-mercapto-1propanol, deuterium oxide (D2O), and ferriycyanide were provided by Sigma-Aldrich (Taufkirchen, Germany). Gold wire electrodes with a diameter of 0.5 mm were provided by Goodfellow (Bad Nauheim, Germany). Buffers. For the preparation of sodium phosphate buffers (5 and 10 mM), Na2HPO4 or NaH2PO4 was used with the pH adjustment using sodium hydroxide or phosphoric acid. The 10 mM sodium phosphate buffer containing 1 M KCl was adjusted for pH after KCl addition. Mutation of Human Cyt c. The plasmid pET21-CCHLHuCytC provided by Jeng et al.38 carries the human cyt c gene and the CCHL genes of the heme lyase. It was used for introducing additional mutations and is further referred as wild type (WT). Primers for mutation were designed with the webbased software PrimerX (http://bioinformatics.org/primerx/). Mutations were introduced using the QuickChange site-directed mutagenesis kit from Stratagene (La Jolla, CA) following the provided protocol. Gene sequencing was done to confirm the mutations. Recombinant Expression and Purification of Wild-Type and Mutated Human Cyt c. Human cyt c was expressed and purified using adapted protocols from the literature.38,39 E. coli cells containing the mutated plasmid DNA and additionally a pEC86 plasmid with the ccm genes for cyt c maturation40,41 were cultivated in minimal media M9 supplemented with minerals, vitamins, and glycerol in a shaker at 37 °C and 180 rpm until OD600 reached 1.0. The expression started after addition of IPTG (1 mM final concentration) and FeSO4 (100 mg/L final concentration), and the culture was incubated for 72 h at 30 °C and 60 rpm. After harvesting by centrifugation, the cells were lysed using lysozyme and sonication. In order to purify cyt c, two chromatography steps were involved: first, the supernatant of the centrifuged lysate was loaded onto a 5 mL SP Sepharose cation-exchange column (GE Healthcare, Sweden) and eluted with a linear NaCl gradient (0-500 mM) in 50 mM sodium phosphate buffer, pH 6.8. Pertinent fractions were determined by SDS-PAGE, and those containing cyt c were concentrated using an Amicon ultracentrifugal filter device with a molecular weight cutoff of 5000 kDa (Millipore, U.S.A.). The sample was then loaded onto a 120 mL dextran size exclusion column and eluted with 50 mM sodium phosphate buffer, pH 6.8. Fractions (38) Jeng, W. Y.; Chen, C. Y.; Chang, H. C.; Chuang, W. J. J. Bioenerg. Biomembr. 2002, 34, 423–431. (39) Rivera, M.; Walker, F. A. Anal. Biochem. 1995, 230, 295–302. (40) Barker, P. D.; Bertini, I.; Del Conte, R.; Ferguson, S. J.; Hajieva, P.; Tomlinson, E.; Turano, P.; Viezzoli, M. S. Eur. J. Biochem. 2001, 268, 4468– 4476. (41) Schulz, H.; Hennecke, H.; Thony-Meyer, L. Science 1998, 281, 1197–1200.

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

2977

were monitored by UV-vis spectroscopy and pooled together to obtain a final cyt c sample. NMR Measurements. Unlabeled and 15N-labeled cyt c samples, in both diamagnetic ferrous and paramagnetic ferric forms, were used for the NMR characterization. The reduced and oxidized states of the proteins were obtained by addition of DTT and an excess of ferricyanide, respectively. The oxidizing agent was then removed by ultrafiltration before NMR data acquisition. Typical protein concentrations were in the 200 µM to 1 mM range (samples were concentrated to meet this range with a Centricon centrifugal filter device, containing a molecular weight cutoff of 5000 kDa), in 50 mM phosphate buffer, pH 6.8 with 10% D2O. NMR experiments were carried out at 300 K on a Bruker 800 MHz spectrometer equipped with a TXI cryoprobe. Unidimensional 1H NMR spectra was recorded over a spectral width of 20 ppm and a 1.2 s recycle delay. Water suppression was achieved using excitation sculpting.42 The ferric form of the protein required detection of hyperfine-shifted resonances, achieved using a larger spectral width (100 ppm), faster repetition rates (600 ms), and presaturation for efficient water suppression. Two-dimensional 1H-15N heteronuclear singlequantum correlation (HSQC) experiments43 were acquired with 1024 × 256 data points, over a spectral width of 14 ppm for 1H and 40 ppm for 15N. Spectra were processed and analyzed with the Bruker Topspin software package. Spectrophotometric Measurements. The secondary structure content of cyt c mutants was characterized at 298 K by circular dichroism (CD). The spectra were collected in the 190-260 nm range on a Jasco J-810 spectropolarimeter, equipped with a Peltier unit. Sample concentrations were in the 2-20 µM range in 50 mM phosphate buffer, pH 6.8. Spectrophotometric measurements were performed with an UV-250 1PC UV-vis spectrophotometer (Shimadzu Scientific Instruments, Kyoto, Japan). Spectra were obtained between 350 and 650 nm for a 5 µM cyt c solution. In order to fully oxidize or reduce the protein, DTT and excess ferricyanide were used, respectively. To guarantee ultrapure samples for the determination of the reaction rate of cyt c with superoxide, mutant and wild-type cyt c were dialyzed before the spectrophotometric kinetic measurements with 50 mM sodium phosphate buffer pH 7.5 using the Slide-A-Lyzer dialyses cassettes (Thermo Scientific, Rockford, IL) with a molecular weight cutoff of 3500 kDa. The amount of oxidized cyt c used for this measurement was calculated according the spectra of the dialyzed protein before each measurement and the total amount of cyt c in the sample which could be retrieved of the spectra of reduced cyt (treated with DTT). The increase of absorbance at 550 nm after addition of XOD, with a final concentration of 30 mU/mL, was recorded for 60 s in the presence of 200 µM hypoxanthine. This corresponds to a steady-state superoxide concentration of about 520 nM. The initial rate of reduction of cyt c was measured for seven different cyt c concentrations (0.5-5 µM cyt c) for each cyt c form. This was performed at least twice. The resulting slope determined the

reaction rate constant k1 for the reaction of cyt c and superoxide in solution. Electrochemical Measurements. Cyclic voltammetry was performed with the Autolab PGSTAT 20 (Metrohm, Germany). Amperometric measurements were conducted on a model 720A potentiostate from CHI Instruments (Austin, TX). For the electrochemical studies, a custom-made 1 mL measurement cell, a Ag/AgCl/1 M KCl reference electrode with a potential of +0.237 V versus NHE (Biometra, Go¨ttingen, Germany), and a platinum counter electrode were used. Gold wire electrodes were cleaned following an established protocol.33 Electrochemical studies of cyt c in solution were carried out with mercaptopropanol-modified electrodes (incubation 24 h in 20 mM mercaptopropanol). Cyclic voltammograms were recorded with 20 µM cyt c in 1 M KCl and 10 mM sodium phosphate buffer, pH 7.0. The formal potential Ef was calculated as the midpoint between the anodic and cathodic peak potentials at different scan rates (100-400 mV/s). Diffusion coefficients were calculated from the peak currents at different scan rates according the Randles-Sevcik equation.44 For the electrochemical investigations of adsorbed cyt on gold electrodes, the needle electrodes were modified with mercaptoundecanoic acid and mercaptoundecanol (incubation for 48 h in 5 mM solutions with a volume ratio of 1:3). Subsequently the electrodes were incubated for 2 h in 30 µM cyt c (in 5 mM potassium phosphate buffer pH 7). Cyclic voltammograms were recorded between -300 and +300 mV versus Ag/AgCl at a scan rate of 100 mV/s. The formal potential Ef was calculated as the midpoint between the anodic and cathodic peak potentials at different scan rates. The heterogeneous electron-transfer rate constant ks was determined by the variation of the scan rate between 100 mV/s and 15 V/s followed by an evaluation of the peak separation according to the Laviron method.45 The amount of redox active cyt c (surface coverage) was derived from the mean of the oxidation and reduction peak area (charge) using Faraday’s law. The charge and the half-peak width were determined with the Autolab software (GPES version 4.8). In order to construct sensor electrodes for the amperometric measurement of superoxide, cyt c was immobilized covalently by incubation of the electrodes in 2.5 mM EDC, 30 µM cyt c (5 mM potassium phosphate buffer, pH 7.0) for 30 min. After washing with 5 mM potassium phosphate buffer, they were stored in the same buffer until they were measured. For long-time storage, electrodes were kept dry in closed 1.5 mL reaction vials at 4 °C. The amperometric measurements were performed at +150 mV versus Ag/AgCl under constant stirring in 1 mL of 50 mM sodium phosphate buffer, pH 7.5. An amount of 10 µL of 50 mM hypoxanthine was added after reaching a stable baseline. Enzymatic superoxide generation was performed by adding 2.5, 5, 10, or 15 µL of a 1 U/mL XOD solution. The sensor signal was derived from the steady-state current after baseline subtraction. The sensor was calibrated using the fact that the superoxide concentration corresponds to the square root of the XOD concentration used.33 At least three electrodes were measured for each cyt c type. The effect of pseudoperoxidase activity of cyt c was investigated with

(42) Hwang, T. L.; Shaka, A. J. J. Magn. Reson., Ser. A 1995, 112, 275–279. (43) Sklenar, V.; Piotto, M.; Leppik, R.; Saudek, V. J. Magn. Reson., Ser. A 1993, 102, 241–245.

(44) Bard, A. J.; Faulkner, L. R. Electrochemical Methods, Fundamentals and Applications; John Wiley and Sons: New York, 1980. (45) Laviron, E. J. Electroanal. Chem. 1979, 101, 19–28.

2978

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

covalently immobilized cyt c on gold electrodes using cyclic voltammetry within the range of -150 up to +250 mV at a scan rate of 20 mV/s. The difference between the cathodic current at 0 V in the absence and presence of 5 mM H2O2 was evaluated. The amperometric pseudoperoxidase control measurements were conducted at +0.15 V in the presence of up to 200 mM H2O2. RESULTS AND DISCUSSION Selection of Mutation Sites and Expression and Purification of Cyt c Mutants. With the intention of creating an improved recognition element for the radical superoxide through sitedirected mutagenesis, human cyt c was chosen as template protein. First photometric experiments have indicated a slightly higher reaction rate for human cyt c with superoxide, compared to horse heart and yeast forms. Furthermore, it was reported that human cyt c is more stable compared to other species.46 Eleven single-mutation sites have been selected for introducing positively charged lysines to increase the positive charge near the heme group. On the basis of earlier studies of the mechanism of O2- conversion by SOD37 and the reaction of cyt c with the radical,47 we propose that an increased positive charge could enhance the guidance of superoxide to the heme iron by electrostatic interaction. This would potentially increase the reaction rate of cyt c with superoxide, which could impose a higher sensitivity of a cyt c based biosensor. Mutations need to be selected at a position near the surface to take advantage of electrostatic interaction but also in a pertinent location near the heme group allowing proper guidance of the superoxide to the iron center. One caveat is that the structural integrity and the electrochemical properties, e.g., the redox potential of cyt c, need to be maintained. Seven hydrophobic residues have been selected to be replaced with lysine: Tyr46Lys, Ala50Lys, Ala51Lys, Tyr74Lys, Gly77Lys, Ile81Lys, and Phe82Lys. Furthermore, four positions are chosen to introduce lysines instead of negatively charged glutamic acids and aspartic acids: Glu61Lys, Asp62Lys, Glu66Lys, and Glu69Lys. Figure 1 shows all selected single-mutation sites and their position within the human cyt c molecule. All 11 mutants could be successfully expressed in E. coli and purified. Spectroscopic Characterization. Figure 2 summarizes the results of mutant characterization by UV-vis spectroscopy and shows spectra of the reduced variant form, G77K, and wild-type cyt c, respectively. It is noted that almost all mutants show the typical absorption maxima in the reduced state (Soret, R, and β bands) at the same wavelength as the wild-type cyt c. The γ band in the oxidized state slightly varies among a few mutants. Some variants show hypsochromic shifts consistent with a slight change in the heme group environment toward the nonpolar direction. We use far-UV CD spectra to evaluate the secondary structure of the expressed variants. The content in the dominating R-helices is the same for all mutants compared to the wild type (data not shown). The design of our mutants is guided by the idea of introducing an extra positive charge on the distal site of the heme, i.e., on

the side where the Met80 axial ligand is located. We therefore use monodimensional 1H NMR as a fingerprint to check the integrity of the iron-Met80 bond in the entire series of expressed mutants. NMR spectra of ferrocytochrome c are characterized by the presence of well-resolved resonances for Met80 side chain in the upfield region of the 1H spectrum.38,48,49 A50K, A51K, D62K, E61K, E66K, and G77K variants have clearly detectable chemical shift values, for the Met 80 resonances, that do not differ from the wild-type chemical shifts more than 0.008 ppm, thus proving intact axial coordination. Slightly larger chemical shift changes have been observed for Y74K (with variations from -0.05 to +0.04 ppm). The three most affected mutants are I81K, F82K, and Y46K. The introduction of a lysine at position 81 induces chemical shift changes in all the Met80 side-chain resonances (from -0.02 to +0.04 ppm), whereas the introduction of a lysine at position 82 has even larger effects (from -0.17 to +0.45 ppm). Finally, mutating Y46 to lysine causes the chemical shift changes at Met80 from -0.51 to -0.05 ppm. However, the chemical shift variations for Met80 signals in all expressed mutants are still small enough to be sure this residue remains the distal axial ligand of heme iron in the reduced form of the protein and its chemical environment is affected very little if not at all (see Supporting Information Figure S1).

(46) Olteanu, A.; Patel, C. N.; Dedmon, M. M.; Kennedy, S.; Linhoff, M. W.; Minder, C. M.; Potts, P. R.; Deshmukh, M.; Pielak, G. J. Biochem. Biophys. Res. Commun. 2003, 312, 733–740. (47) Butler, J.; Koppenol, W. H.; Margoliash, E. J. Biol. Chem. 1982, 257, 10747– 10750.

(48) Baistrocchi, P.; Banci, L.; Bertini, I.; Turano, P.; Bren, K. L.; Gray, H. B. Biochemistry 1996, 35, 13788–13796. (49) Banci, L.; Bertini, I.; Huber, J. G.; Spyroulias, G. A.; Turano, P. J. Biol. Inorg. Chem. 1999, 4, 21–31.

Figure 1. NMR structure of human cyt c with mutation sites [green, higher reaction rate with superoxide (E66, Y74, F82, E69); orange, lower reaction rate (Y46, A50, A51); yellow, reaction rate not affected (E61, D62; G77, I81)]. The mutant E66K with the highest reaction rate with superoxide is marked with a green arrow. The image was built with the software Pymol (http://pymol.sourceforge.net) using the pdb file 1J3S.

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

2979

Figure 2. Absorption spectra of the reduced wild type (s) and the mutant G77K (- - -) [5 µM protein sample in 50 mM sodium phosphate buffer pH 7.5]. Table: spectrophotometric properties of wild type (WT) and variant forms of human cyt c.

Electrochemical Characterization of Cyt c Mutants in Solution. Information about the thermodynamic and diffusional properties of all 11 purified mutants have been obtained by electrochemical investigations in solution with mercaptopropanolmodified gold electrodes. Cyclic voltammetry is performed at different scan rates to calculate the redox potential and the diffusion coefficients of the different mutants. All results are collected in Table 1, and an example of a typical cyclic voltammogram is shown in Figure 3A. For 10 mutants no remarkable influence on the redox potential has been observed. The formal redox potential, Ef, of Y46K is -40 mV (±8 mV) versus Ag/ AgCl which is clearly below the range of the other proteins; hence, Y46K is more easily oxidized. The diffusion coefficients calculated from the voltammetric measurements are found to be slightly below literature values for horse heart cyt c (4.7 × 10-7 cm2 s-1 50) which might be due to the use of chronocoulometry by these authors. The data in Table 1 also reveal the diffusion coefficients vary slightly among the mutants. E61K shows a 1.7-fold increased value, whereas F82K has a 4-fold decreased diffusion coefficient with respect to the wild type; the experimental error is approximately 25%. These variations can be explained by small conformational changes (50) Terrettaz, S.; Cheng, J.; Miller, C. J. J. Am. Chem. Soc. 1996, 118, 7857– 7858.

2980

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

caused by the mutations and thus a change in the solventaccessible surface area.51 Determination of Rate Constants between Cyt c Mutants and Superoxide. For all mutants, horse heart cyt c, and wildtype human cyt c, the reaction rate constants with superoxide are determined in solution by recording the cyt c reduction rate at 550 nm in the presence of the superoxide radical. Horse heart cyt c has been included in the investigation for comparative reasons since all cyt c based O2- sensors developed to date use this cyt c form; the results are collected in Figure 4. As already indicated by our preliminary experiments, the human form of cyt c has a slightly higher reaction rate constant (8.3 × 10-4 M-1 s-1) compared to horse heart cyt c (6.5 × 10-4 M-1 s-1). The three mutants E66K, Y74K, and F82K show a higher rate constant in the reaction with superoxide compared to the human wild type (12.5 × 10-4, 10.7 × 10-4, and 10.7 × 10-4 M-1 s-1); Figure 1 depicts these amino acids in green for a better visualization. The E69K variant has a reaction rate constant of 8.8 × 10-4 M-1 s-1, i.e., just slightly higher than the wildtype protein. For several mutants the reaction rate is not influenced, whereas three mutants (Y46K, A50K, A51K) show a reduced reaction with superoxide. In Figure 1 these mutation sites are marked in yellow and orange, respectively. Figure 1 suggests two areas of mutation sites, one with positive effects and one with negative effects (higher and lower reaction rates, respectively). The area with positive effects may be related to a possible access pathway for superoxide, which can guide the radical to the heme group. The decrease in the reaction rate with the radical, for three mutants, may be due to addition of a positive charge far from the access channel slowing the entrance of superoxide by misguiding the radical. Y46K has other factors contributing to the decreased reaction rate. This will be discussed in more detail in the last section of this study. Electrochemical Characterization of Cyt c Mutants and Sensor Behavior of Mutant Electrodes. Horse heart, human wild type, and four mutant forms of cyt c, which reveal a higher radical reaction rate constant, have been chosen to be investigated as a recognition element in a biosensor for the detection of superoxide. First, the thermodynamic and kinetic properties of the cyt c mutants have been characterized electrochemically with the protein adsorbed on a promoter-modified gold electrode. All chosen mutants can be adsorbed on the negatively charged mercaptoundecanoic acid/mercaptoundecanol (MU/MUA) layer. Figure 3B demonstrates that the cyclic voltammograms of the mutants are very similar to those achieved with the wild type. The behavior is also studied kinetically, and the values for the redox potential and the heterogeneous electron-transfer rate constants (ks) are presented in Table 1. The rather similar redox properties and surface coverage of the proteins, in comparison with the values for the wild-type and horse heart cyt c, are promising for sensor application. A similar behavior for the mutants can also be observed for the ks values and the peak shape, despite mutants F82K and Y74K, which have a 2.2- and 1.8-fold increased electron-transfer rate and a slightly smaller half-peak width (