Electrochemical generation of a high-valent state of cytochrome P450

JOURNAL OF

Inorganic Biochemistry Journal of Inorganic Biochemistry 100 (2006) 519–523 www.elsevier.com/locate/jinorgbio

Short communication

Electrochemical generation of a high-valent state of cytochrome P450 Andrew K. Udit b

a,*

, Michael G. Hill

a,*

, Harry B. Gray

b

a Department of Chemistry, Occidental College, 1600 Campus Road, Los Angeles, CA 91030, United States Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, CA 91125, United States

Received 15 January 2006; accepted 16 January 2006 Available online 28 February 2006

Abstract Cyclic voltammetry performed at rapid scan rates on cytochrome P450 from Pseudomonas putida (P450CAM) in didodecyldimethylammonium bromide (DDAB) films on graphite electrodes revealed a couple (E) at 830 mV (vs Ag/AgCl). E was not significantly observed at scan rates less than 30 V/s at room temperature, suggesting that the oxidized species is unstable. The lifetime of E could be prolonged at 4 C, which allowed reversible access to E at scan rates as low as 1 V/s. E was found to be sensitive to imidazole in solution and to variations in pH, suggesting that the redox reaction is occurring at the metal center (i.e., FeIV/III). Electrolysis reactions with different P450 substrates revealed that the electrochemically generated high-valent species is able to convert thioanisole to methyl phenyl sulfoxide.  2006 Elsevier Inc. All rights reserved. Keywords: Cytochrome P450; Electrochemistry; High-valent; Oxidation

1. Introduction Heme-thiolate proteins (cytochrome P450, nitric oxide synthase, chloroperoxidase) perform essential functions in biology. The hallmark reaction catalyzed by these enzymes is substrate oxidation, converting R–H to R–OH. Thiol ligation to the iron is postulated to be the critical factor that modulates the heme redox properties to activate dioxygen for substrate oxidation [1–3]: two electrons are delivered by native reductase proteins to reduce dioxygen, forming water and a high-valent ferryl species capable of performing oxygen atom transfer. This ability for oxidative catalysis has led to great interest in capturing heme-thiolate oxygenase activity in vitro for both medical and industrial applications [4,5]. Central to the catalytic cycle of heme-thiolate proteins is electron transfer (ET), leading to several investigations characterizing ET reactions in these enzymes [6–9]. In particular, protein-surfactant film voltammetry has become a reliable method for achieving direct ET with heme proteins *

Corresponding authors. Tel.: +1 323 259 2766; fax: +1 323 341 4912. E-mail addresses: [email protected] (A.K. Udit), [email protected] (M.G. Hill). 0162-0134/$ - see front matter  2006 Elsevier Inc. All rights reserved. doi:10.1016/j.jinorgbio.2006.01.011

[10]. Using this technique, the FeIII/II redox couple in various heme-thiolate enzymes has been investigated [11–14]. More recently, an FeII/I couple (1100 mV vs Ag/AgCl) was observed, with reductive dehalogenation catalyzed by the FeI species [15,16]. The ability to access FeIII/II and FeII/I redox couples in heme-thiolate proteins suggested to us that high-valent species might also be generated electrochemically in surfactant films. Conceivably, one- or two-electron oxidation of the FeIII–OH2 heme (resting state) could result in rapid conversion to the ferryl species FeIV@O–H (compound II) or FeV@O (compound I), as previously observed for horseradish peroxidase [17] and chloroperoxidase [18]. In the presence of substrates, these ferryl species could be capable of performing oxidative catalysis in the absence of both dioxygen and NAD(P)H, both of which are necessary in the native system. Directly generating ferryl species through heme oxidation (essentially running the P450 catalytic cycle backwards) is an attractive alternative to cathodic P450 bioelectrochemical catalysis, which can lead to direct dioxygen reduction at the electrode, production of destructive species (superoxide, peroxide), and uncoupling with futile redox cycling (through P450-catalyzed one-, two-, and four-electron reduction of dioxygen).

520

A.K. Udit et al. / Journal of Inorganic Biochemistry 100 (2006) 519–523

It is well established that both oxidized and reduced Fe– porphyrins can be electrochemically accessed in nonaqueous media [19,20]. Depending on the system, oxidation states from FeI to FeV can be generated, as well as porphyrin cations and anions. For P450-type systems, it has been shown that FeIII–porphyrins with basic axial ligands (cf. thiolate in P450) are oxidized first at the metal (FeIV), and then at the porphyrin macrocycle [21,22]. Indeed, Groves and Gilbert have shown that in the presence of water, these hemes can be electrochemically oxidized to produce ferryl species capable of catalytic epoxidation [22].

the coated electrode in a solution of enzyme (10 lM in 30 mM KPi, pH 7.4) for 30 min. Voltammetry was conducted in a three-compartment cell with an Ag/AgCl reference electrode (BAS) and Pt wire counter electrode. The buffer was 50 mM KPi/20 mM KCl/ pH 8. Buffers for variable pH measurements were kept at constant ionic strength by adding different proportions of the monobasic and dibasic forms of potassium phosphate salts such that the final concentration was always 50 mM; the resulting pH was determined with a pH electrode. Voltammetry experiments at 4 C were performed in an ice-cooled water bath.

2. Experimental methods 2.3. Electrolysis and thioanisole oxidation 2.1. Protein expression and purification Cytochrome P450 from Pseudomonas putida (P450CAM) was expressed as previously described [23]. Buffers for protein purification were as follows: buffer A – 50 mM KPi, pH 7.2, 500 lM camphor, 2 mM DTT; buffer B – 50 mM KPi, 1 M KCl, pH 7.2, 500 lM camphor, 2 mM DTT; buffer C – 50 mM KPi, 30% saturated ammonium sulfate (176 g/L) pH 7.2, 500 lM camphor, 2 mM DTT; buffer D – 50 mM KPi, 50 mM KCl, pH 7.4, 2 mM DTT, 500 lM camphor. All buffers were filtered and degassed using a vacuum filter. Cell pellets (3 g) were mixed with 9 mL of buffer A containing 1 mM phenylmethylsulfonyl fluoride (PMSF, protease inhibitor). Cells were lysed by sonication. The lysate was centrifuged at 24,000 rpm for 30 min at 4 C. The resulting supernatant was filtered using a 0.45 lm filter and then injected onto a 20 mL DEAE Fast Flow column equilibrated with buffer A. The protein was eluted with a linear gradient of 0–500 mM KCl in buffer A. Fractions containing heme protein (visibly dark red) were pooled. The protein from the pooled fractions was precipitated with 60% ammonium sulfate overnight at 4 C. The mixture was then centrifuged at 12,500 rpm for 25 min at 4 C. The resulting pellet was resuspended in 2–5 mL of buffer C. The solution was loaded onto a phenyl sepharose column equilibrated with buffer C. Protein was eluted with a linear gradient of ammonium sulfate from 30%–0%. Fractions with Abs395/Abs280 > 1 were collected, pooled, and concentrated to 1 mL with a Millipore Centriprep concentrator. The protein solution was injected onto a Sephacryl S200 column equilibrated with buffer D. The protein was eluted with a constant flow of buffer D. Fractions with Abs395/Abs280 > 1.45 were pooled and concentrated using a Centriprep concentrator. 2.2. Preparation of protein films for voltammetry Didodecyldimethylammonium bromide (DDAB) films were formed on basal plane graphite electrodes (0.07 cm2) by depositing 5 lL of 10 mM aqueous solution of DDAB on the electrode surface, followed by slow drying overnight. P450CAM was incorporated into the film by soaking

Electrodes for electrolysis were made from pyrolytic graphite plates, 1 cm · 2 cm, 0.025 cm thick. The plates were sanded, sonicated, and dried in air. Twenty microliters of a 10 mM solution of DDAB in ddH2O were applied and dried on each side of the graphite plate. Protein was incorporated into the films by soaking the filmed plates in a 5 lM P450CAM solution for 30 min. Electrolyses were performed in a three-neck flask at room temperature. Working and counter (Pt gauze) electrodes were placed directly into the solution, while the reference electrode was housed in a Luggin capillary. Since preliminary experiments revealed that the presence of dioxygen in solution did not affect the electrolysis reaction, degassing the buffers was not necessary. Reactions were performed in 6 mL of 50 mM KPi/20 mM KCl/pH 8, with 100 lL thioanisole as substrate. Reactions were conducted at 0.9 V with stirring for 10 min. After the reaction, 5 lL of 1 mM 4-chloro-styrene was added (internal standard), followed by extraction of the solution with 0.5 mL CHCl3. The CHCl3 layer was removed and analyzed by GC/MS using an Innowax column. The product methyl phenyl sulfoxide was quantified relative to the internal standard using a calibration curve that was constructed by relating the ratio of the areas of the product to the internal standard. Control reactions with BSA instead of P450CAM were conducted by soaking the filmed graphite plate in a solution that contained 6 lM BSA. Control reactions with catalase were performed by adding 2 lL of a 47 mg/mL catalase stock solution to the reaction. The amount of P450CAM incorporated into the films from solution was estimated by taking the absorption spectrum of the solution before and after soaking the filmed electrode. The difference in Abs416 (e416 = 115,000 M1 cm1) was used to estimate the amount of P450CAM incorporated into the film. Typically, 2–2.5 nmol of P450CAM were incorporated into the films on the PG plates. 3. Results and discussion We performed cyclic voltammetry (CV) on cytochrome P450 from P. putida (P450CAM) in didodecyldimethylammonium bromide (DDAB) films on the surface of basal

A.K. Udit et al. / Journal of Inorganic Biochemistry 100 (2006) 519–523

Current (mA)

0.6 0.2 -0.2 -0.6 -1.0 1.5

1

0.5 0 -0.5 Potential (V vs. Ag/AgCl)

-1

Fig. 1. Cyclic voltammogram of P450CAM in DDAB on graphite (0.07 cm2) at 50 V/s in 50 mM KPi/50 mM KCl/pH 8 at ambient temperature.

Current (μA)

10 0 -10 -20

0.4

Current (mA)

plane graphite electrodes and looked for evidence of protein and/or heme oxidation. Preliminary CV experiments were conducted at 200 mV/s and pH 7 at room temperature by beginning at 0 mV (vs Ag/AgCl), scanning negative to 700 mV, reversing the scan direction to 1300 mV, and then cycling between 1300 and 700 mV. The FeIII/II redox couple was observed at E1/2 = 227 mV, upshifted approximately 300 mV from the value in solution as is typically seen in P450-surfactant film electrochemical systems. In addition to the FeIII/II redox couple, we also observed a large catalytic current corresponding to solvent oxidation at potentials >900 mV. After scanning beyond this solvent window, subsequent scanning in the region of the FeIII/II couple resulted in no observable signal, suggesting that the heme had been oxidatively damaged. The above CV experiment was repeated at 50 V/s and pH 8 at room temperature. The resulting voltammogram (Fig. 1) clearly reveals two couples: the FeIII/II redox couple at 227 mV, and a second couple (E) with E1/2 = 830 mV. Accessing E was facilitated by scanning faster (>30 V/s), and performing the experiment at higher pH (pH > 7). E was found to be reversible above 30 V/s, and multiple CVs could be recorded without any significant loss in signal intensity. Notably, it was possible to reversibly generate E at scan rates down to 1 V/s when performed at 4 C (Fig. 2). There are three likely possibilities for the origin of E: (1) a porphyrin-centered redox process; (2) an amino

521

-0.6

-1.6 1.5

1

0.5

0

-0.5

-1

Potential (V vs. Ag/AgCl) Fig. 3. Cyclic voltammogram of P450CAM in DDAB on graphite (0.07 cm2) at 50 V/s in 50 mM KPi/50 mM KCl/pH 8 at ambient temperature in the presence of 500 mM imidazole.

acid redox process; (3) or an Fe-centered redox process (i.e., FeIV/III). ET reactions involving the macrocycle of an Fe–porphyrin are relatively insensitive to changes in the iron axial ligands or solution composition [24]. In contrast, CVs of P450CAM–DDAB films recorded under different solution conditions indicated significant perturbations to E. First, voltammetry at constant ionic strength and variable pH (6–9) revealed that E varied with pH according to 47 mV/pH, suggesting a proton-coupled ET mechanism.1 Indeed, pH-dependent ET is commonly observed for redox reactions in P450, possibly deriving from the heme axial water ligand.2 Second, voltammetry in the presence of 500 mM imidazole (Fig. 3) shows that E is inhibited, in accord with a redox reaction involving a water-ligated FeIII heme. Replacing the axial water with another ligand (imidazole in this case) would inhibit ferryl formation, as previously reported [25]. This finding also rules against the alternative interpretation that E is a tyrosine or tryptophan redox couple. It is noteworthy that these electrochemical data are fully consistent with recent EXAFS measurements and DFT calculations carried out on compound II from chloroperoxidase [1]: a relatively long Fe–O bond distance ˚ ) coupled to a contracted Cys–S–FeIV interaction (1.82 A ˚ ) strongly suggest that the FeIV@O unit in com(2.39 A pound II is protonated at neutral pH, with a pKa (for the protonated form) greater than 8. These results lead us to envision the reaction pathway depicted in Scheme 1. Beginning with FeIII–OH2 (resting state), oxidation of this species would lead to FeIV@O–H, which can decay in several ways. First, rapid reduction would regenerate the resting state, FeIII–OH2. Second, if reduction does not occur fast enough, oxidation of the surrounding protein environment by the ferryl species would likely result in damage to the protein. This is supported by the observation that voltammetry at slower scan rates, or

-30 -40 1.5

1

0.5 0 -0.5 Potential (V vs. Ag/AgCl)

-1

Fig. 2. Cyclic voltammogram of P450CAM in DDAB on graphite (0.07 cm2) at 1 V/s in 50 mM KPi/50 mM KCl/pH 8 at 4 C.

1

Notably, the FeIII/II couple varied similarly with pH according to 41 mV/pH. 2 The precise origin of the proton in the ET reaction is not wellunderstood, particularly in light of our recent work indicating that redox cycling likely affects the heme hydration state (see [14]).

522

A.K. Udit et al. / Journal of Inorganic Biochemistry 100 (2006) 519–523

H

H

H

O

O

FeIII

-é, -H+

FeIV

é, H+

H

H O

Sub H2O

Sub - OH Protein and / or heme oxidative damage

H

H

FeIII

O

FeIII

Scheme 1. Table 1 Methyl phenyl sulfoxide is the product of bioelectrocatalytic oxidation of thioanisole using P450CAM in DDAB on graphite Reaction

nmols methyl phenyl sulfoxide

P450CAM in DDAB on graphite P450CAM in DDAB on graphite + catalase DDAB on graphite DDAB on graphite + catalase BSA in DDAB on graphite

416 ± 14 368 ± 20 130 ± 14 115 ± 20 92 ± 14

Electrolyses were conducted at 0.9 V vs Ag/AgCl for 10 min in 6 mL 50 mM KPi/20 mM KCl/pH 8 at ambient temperature.

scanning at room temperature (cf. 4 C), results in loss of the FeIII/II couple. Third, it is possible that if a ferryl species is indeed generated, it may be capable of oxidizing substrates and subsequently regenerating the FeIII–OH2 resting state. To test the putative FeIV species for catalytic activity, we performed electrolysis reactions with P450CAM–DDAB films on graphite electrodes at 0.9 V.3 After trying several substrates (e.g., alkanes, alkenes) we found that only thioanisole was oxidized by the system. The results listed in Table 1 reveal that reactions with P450CAM consistently produce more methyl phenyl sulfoxide than control reactions without P450CAM [26].4 Additional controls with catalase present,5 and films made with bovine serum albumin instead of P450CAM, further support a P450 FeIV-catalyzed oxidation reaction. The observation that only thioanisole is oxidized by our putative FeIV-P450CAM species indicates that the 3

Electrolysis at more positive potentials resulted in loss of the protein response, and significantly decreased product formation. There are several possibilities for this type of response, including film desorbtion and oxidative damage to the protein or heme. 4 Direct oxidation of thioanisole to the sulfoxide (E0 = 1.3 V vs Ag/ AgCl) at the electrode can also occur. 5 Peroxide formed from water oxidation can oxidize thioanisole.

FeIV@OH unit is not powerful enough to perform more challenging alkane or alkene oxidations in the DDAB film. This is hardly unexpected considering that the active heme center in the native system, compound I, lies one full oxidation state higher than our (electrochemically accessible) compound II. Mayer [27] has proposed that the key determinant for C–H bond activation by metal-oxo species is the strength of the resulting O–H bond in the (M(n1)+)O–H complex following hydrogen-atom abstraction: if the O– H bond (D(O–H)) in the (formally one-electron reduced) ‘‘rebound’’ complex is stronger than the C–H bond (D(C–H)) being broken, the reaction can proceed. D(O– H) is equal to DH for the reaction: Mnþ @O þ H Mðn1Þþ AOH and can be estimated using a simple thermodynamic cycle,6 drawn in Scheme 2 (in our system Mn+@O is FeIV@OH). Values for the individual reactions in this cycle are available either experimentally (e.g., the potential of the metal redox couple (FeIV/III) and pKa for FeIII–OH2) or have been calculated (e.g., the bond energy of H2 in aqueous solution) [27]. Bordwell [28] has proposed an empirical expression for D(O–H) based on this cycle:7 DðO–HÞðkcal=molÞ ¼ 23:06  ½E0 ðFeIV=III Þ þ 1:37  pK a ðFeIII –OH2 Þ þ 57 Green et al. [1] recently used this approach to estimate D(O–H) for compound II in chloroperoxidase at 98 kcal/ mol, making the corresponding compound I (i.e., (P+)FeIV@O) potent enough to oxidize most alkanes Assuming equal entropies for Mn+@O and M(n1)+AOH. Although the value for proton reduction (line 2 in Scheme 2) is poorly defined under physiological conditions, the equation has been validated experimentally for many different compounds where the bond-dissociation energies were measured independently (see [28]). 6 7

A.K. Udit et al. / Journal of Inorganic Biochemistry 100 (2006) 519–523

FeIV=O-H + e-

FeIII-O-H

1/2 H2

H+ + e-

FeIII-O-H + H+

FeIII-OH2

H•

1/2 H2

FeIV=O-H + H•

FeIII-OH2

Scheme 2.

(D(C–H)  95–99 kcal/mol). In contrast, D(O–H) for FeIII–OH2 in P450CAM-DDAB films (calculated at pH 6)8 is only 94 kcal/mol. This provides an explanation for the inability of the system to perform reactions more demanding than thioanisole oxidation. Indeed, were it not for the substantial upshift in potential that occurs upon immobilizing the enzyme on electrode surfaces, the electrochemically generated compound II would be even less active. Assuming that the FeIV/III couple is shifted by an amount comparable to the FeIII/II process (300 mV vs solution), surface effects account for an increase of nearly 7 kcal/mol in D(O–H) for FeIII–OH2. 4. Abbreviations CV cyclic voltammetry DDAB didodecyldimethylammonium bromide ET electron transfer P450CAM cytochrome P450 from Pseudomonas putida EXAFS extended X-ray absorption fine structure DFT density functional theory KPi potassium phosphate buffer PMSF phenylmethylsulfonyl fluoride ddH2O double distilled water BSA bovine serum albumin DEAE diethylaminoethane DTT dithiothreitol

Acknowledgements We thank Nick Halpern-Manners and Geethani Bandara (Caltech) for assistance with protein purification. Supported by NIH (HBG, MGH), NSERC (AKU) and HHMI (AKU).

8 Eo 0 (FeIV/III) = 1.14 V vs NHE in DDAB, and pKa for FeIII–OH2 was approximated to be 9.

523

References [1] M.T. Green, J.H. Dawson, H.B. Gray, Science 304 (2004) 1653– 1656. [2] R. Silaghi-Dumitrescu, J. Biol. Inorg. Chem. 9 (2004) 471–476. [3] J. Conradie, J.C. Swarts, A. Ghosh, J. Phys. Chem. B 108 (2004) 452– 456. [4] F.P. Guengerich, Molecular Interventions 3 (2003) 194–204. [5] A. Glieder, E.T. Farinas, F.H. Arnold, Nat. Biotechnol. 20 (2002) 1135–1139. [6] A.R. Dunn, I.J. Dmochowski, J.R. Winkler, H.B. Gray, J. Am. Chem. Soc. 125 (2003) 12450–12456. [7] A.K. Udit, M.G. Hill, V.G. Bittner, F.H. Arnold, H.B. Gray, J. Am. Chem. Soc. 126 (2004) 10218–10219. [8] I.F. Sevrioukova, C.E. Immoos, T.L. Poulos, P.J. Farmer, Isr. J. Chem. 40 (2000) 47–53. [9] B. Munge, C. Estavillo, J.B. Schenkman, J.F. Rusling, ChemBioChem 4 (2003) 82–89. [10] J.F. Rusling, Acc. Chem. Res. 31 (1998) 363–369. [11] B.D. Fleming, Y. Tian, S.G. Bell, L. Wong, V. Urlacher, H.A.O. Hill, Eur. J. Biochem. 270 (2003) 4082–4088. [12] M. Bayachou, J.A. Boutros, J. Am. Chem. Soc. 126 (2004) 12722– 12723. [13] Z. Zhang, A.-E. Nassar, Z. Lu, J.B. Schenkman, J.F. Rusling, J. Chem. Soc., Faraday Trans. 93 (1997) 1769–1774. [14] A.K. Udit, W. Belliston-Bittner, E.C. Glazer, Y.H.L. Nguyen, J.M. Gillan, M.G. Hill, M.A. Marletta, D.B. Goodin, H.B. Gray, J. Am. Chem. Soc. 127 (2005) 11212–11213. [15] C.E. Immoos, J. Chou, M. Bayachou, E. Blair, J. Greaves, P.J. Farmer, J. Am. Chem. Soc. 126 (2004) 4934–4942. [16] E. Blair, J. Greaves, P.J. Farmer, J. Am. Chem. Soc. 126 (2004) 8632– 8633. [17] T. Egawa, D.A. Proshlyakov, H. Miki, R. Makino, T. Ogura, T. Kitagawa, Y. Ishimura, J. Biol. Inorg. Chem. 6 (2001) 46–54. [18] J. Berglund, T. Pascher, J.R. Winkler, H.B. Gray, J. Am. Chem. Soc. 119 (1997) 2464–2469. [19] K.M. Kadish, E.V. Caemelbecke, F. D’Souza, C.J. Medford, K.M. Smith, A. Tabard, R. Guilard, Inorg. Chem. 34 (1995) 2984–2989. [20] F.M. Hawkridge, I. Taniguchi, in: K.M. Kadish, K.M. Smith, R. Guilard (Eds.), The Porphyrin Handbook. Electrochemistry of Heme Proteins, vol. 8, Academic Press, San Diego, 2000. [21] T.S. Calderwood, W.A. Lee, T.C. Bruice, J. Am. Chem. Soc. 107 (1985) 8272–8273. [22] J.T. Groves, J.A. Gilbert, Inorg. Chem. 25 (1986) 123–125. [23] I.C. Gunsalus, G.C. Wagner, Methods Enzymol. 52 (1978) 166– 188. [24] M.A. Phillippi, E.T. Shimomura, H.M. Goff, Inorg. Chem. 20 (1981) 1322–1325. [25] C.E. Immoos, A.J. Di Bilio, M.S. Cohen, W. Van der Veer, H.B. Gray, P.J. Farmer, Inorg. Chem. 43 (2004) 3593–3596. [26] E. Baciocchi, M.F. Gerini, O. Lanzalunga, S. Mancinelli, Tetrahedron 58 (2002) 8087–8093. [27] J.M. Mayer, Acc. Chem. Res. 31 (1998) 441–450. [28] F.G. Bordwell, J. Cheng, G. Ji, A.V. Satish, X. Zhang, J. Am. Chem. Soc. 113 (1991) 9790–9795.