Kinetic Consequences of Introducing a Proximal Selenocysteine

Oct 13, 2015 - *E-mail: [email protected]. ... of a selenocysteine-ligated P450 compound I reveals direct link between electron donation and re...
2 downloads 0 Views 1MB Size
Subscriber access provided by NEW YORK MED COLL

Article

Kinetic consequences of introducing a proximal selenocysteine ligand into cytochrome P450cam An Vandemeulebroucke, Caroline Aldag, Martin Tillmann Stiebritz, Markus Reiher, and Donald Hilvert Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.5b00939 • Publication Date (Web): 13 Oct 2015 Downloaded from http://pubs.acs.org on October 25, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biochemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Kinetic Consequences of Introducing a Proximal Selenocysteine Ligand into Cytochrome P450cam An Vandemeulebroucke,a Caroline Aldag,a Martin T. Stiebritz,b Markus Reiherb and Donald Hilverta,* a

Laboratory of Organic Chemistry, bLaboratory of Physical Chemistry, ETH Zurich, CH-8093 Zurich, Switzerland.

AUTHOR INFORMATION Corresponding Author * E-mail: [email protected] Author Contributions The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Funding Sources This work was generously supported by the ETH Zurich. Notes The authors declare no competing financial interest.

1

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ABBREVIATIONS DFT, density functional theory; DTT, dithiotreitol; IPTG, isopropyl β-D-1-thiogalactopyranoside; NA DH, nicotinamide adenine dinucleotide; Pdx, putidaredoxin; Pdxred, pre-reduced putidaredoxin; PdR, putidaredoxin reductase; wtP450cam, wild-type cytochrome P450cam; P450cam*, R365L/E366Q P450cam; SeP450cam*, C357U/R365L/E366Q P450cam; SECIS, selenocysteine insertion sequence.

2

ACS Paragon Plus Environment

Page 2 of 50

Page 3 of 50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

ABSTRACT: The structural, electronic, and catalytic properties of cytochrome P450cam are subtly altered when the cysteine that coordinates to the heme iron is replaced by a selenocysteine. To map the effects of the sulfur-to-selenium substitution on the individual steps of the catalytic cycle, we carried out a comparative kinetic analysis of the selenoenzyme and its cysteine counterpart. Our results show that the more electron-donating selenolate ligand has only negligible effects on substrate, product and oxygen binding, electron transfer, catalytic turnover, and coupling efficiency. Off-pathway reduction of oxygen to give superoxide is the only step significantly affected by the mutation. Incorporation of selenium accelerates this uncoupling reaction approximately 50 fold compared to sulfur, but because the second electron transfer step is much faster, the impact on overall catalytic turnover is minimal. Density functional theory calculations with pure and hybrid functionals suggest that superoxide formation is governed by a delicate interplay of spin distribution, spin state, and structural effects. In light of the remarkably similar electronic structures and energies calculated for the sulfur and selenium-containing enzymes, the ability of the heavier atom to enhance the rate of spin-crossover may account for the experimental observations. Because the selenoenzyme closely mimics the properties of wildtype P450cam, even at the level of individual steps in the reaction cycle, selenium represents a unique mechanistic probe for analyzing the role of the proximal ligand and spin crossovers in P450 chemistry.

3

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Cytochrome P450 enzymes are ubiquitous heme-containing monooxygenases that catalyze diverse oxidative transformations, including C-H bond hydroxylation, O- and N-dealkylation, Nhydroxylation, S-oxidation and epoxidation reactions.1 In general, these enzymes play both catabolic and anabolic roles in their hosts. In humans, for example, they are involved in the breakdown of xenobiotics, such as ingested therapeutics and environmental compounds, as well as in the synthesis of steroid hormones and prostaglandins.2,3 P450 enzymes also contribute to the etiology of various cancers and arterial and inflammatory disorders.4,5 In addition to their physiological significance, the ability of these enzymes to catalyze difficult biotransformations with high rates and selectivities makes them very attractive for exploitation in stereospecific synthesis of highly functionalized compounds6, as well as for bioremediation of polluted areas.7 All P450 enzymes utilize iron-protoporphyrin IX as a cofactor and share a common overall fold and topology. The heme iron is axially coordinated by the thiolate of an absolutely conserved cysteine residue. Despite numerous structure-function studies, however, there is still considerable controversy regarding the catalytic mechanism by which these enzymes activate oxygen and insert it into organic substrates. The most accepted hypothesis concerning the catalytic cycle (Figure 1)8 starts with binding of the substrate at the distal heme pocket of the enzyme in the ferric resting state ①. Displacement of the coordinating water, ②, facilitates oneelectron reduction of the heme by a redox partner ③. Next, oxygen binds to the reduced ferrous intermediate to give a ferrous-dioxygen or ferric superoxide complex ④, which is then further reduced by transfer of a second electron to give a heme-peroxo intermediate ⑤. Subsequent proton transfer affords a hydroperoxide adduct ⑥, which breaks down to give a high-valent ironoxo species ⑦, called compound I. In the rebound mechanism,9 compound I is thought to abstract a hydrogen atom from substrate, yielding a substrate radical and a protonated 4

ACS Paragon Plus Environment

Page 4 of 50

Page 5 of 50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

oxidoiron(IV) or a protonated oxidoiron(III)-protoporphyrin radical, compound II ⑧. Hydroxylation of the substrate radical by compound II ⑨, followed by dissociation of the product from the active site, completes the catalytic cycle. The axially coordinating cysteine thiolate is generally believed to dictate P450 reactivity. Substitution of this residue by different proteinogenic amino acids that are utilized as proximal ligands in other natural heme enzymes adversely affects both cofactor binding and monooxygenase activity.10–13 Although not found in naturally occurring heme enzymes, the rare proteinogenic amino acid selenocysteine is potentially a better suited heme ligand given its close structural similarity to cysteine. We used the selenocysteine insertion machinery of Escherichia coli to incorporate this amino acid co-translationally into the well-characterized cytochrome P450cam isozyme from the soil bacterium Pseudomonas putida in place of Cys357.14 The resulting sulfur-to-selenium substitution afforded a stable, active enzyme (C357U P450cam* or SeP450cam*) with only subtly modified structural, electronic, and catalytic properties. Ortiz de Montellano and co-workers independently incorporated selenocysteine into two other P450 isozymes by reassignment of cysteine codons to selenocysteine in an auxotrophic host.15–17 Although the latter approach did not yield complete replacement of the axial cysteine by selenocysteine, the resulting selenoproteins, like SeP450cam*, were found to retain the basic catalytic properties of their parent enzymes. Selenium and sulfur share many features, including size, electronegativity and major oxidation states. However, selenium shows greater nucleophilic character than sulfur and is more polarizable, making it a softer nucleophile. Most importantly, selenols differ from thiols in their pKa and reduction potentials. Because selenols are substantially more acidic (pKa ca. 5.2) than

5

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

thiols (pKa ca. 8.3), selenocysteine is ionized at physiological pH, making it more reactive than cysteine. The greater electron donating capacity of selenium versus sulfur also results in a lower reduction potential for selenocysteine compared to cysteine.18,19 These distinctive electronic properties are potentially informative regarding mechanism.20 For example, quantum mechanical calculations suggest that a selenolate ligand should speed up the formation of compound I and slow subsequent reactions with substrate.21 Consistent with this hypothesis, selenocysteine incorporation into CYP125 favored products formed by heterolytic as opposed to homolytic pathways.17 In SeP450cam*, distinct changes in the electronic and chemical properties of the enzyme, including diagnostic redshifts of the absorption maxima, perturbation of the spin equilibrium in favor of the low spin resting state, and a lowered Fe(III)/Fe(II) electrode potential were observed.14 Such changes provide a unique window for probing the role of the proximal ligand in P450 chemistry. Here we present a detailed comparative pre-steady state kinetic analysis of SeP450cam* and its cysteine-containing counterpart. Our findings shed light on the effects of the sulfur-toselenium substitution on early steps in the catalytic cycle. Quantum mechanical (QM) calculations were employed to provide additional theoretical insight into the differences observed.

EXPERIMENTAL PROCEDURES Materials. All chemicals were of analytical or reagent grade and were used without further purification. (1R)-(+)-Camphor, nicotinamide adenine dinucleotide (NADH), sodium salicylate, trichloroacetic acid and catalase from bovine liver were purchased from Sigma Aldrich. Hydrogen peroxide and glucose were purchased from Merck KGaA. Sodium selenite and 6

ACS Paragon Plus Environment

Page 6 of 50

Page 7 of 50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

cysteine were purchased from ABCR Chemicals GmbH. Glucose oxidase and dithiothreitol (DTT) were purchased from Fluka. δ-Aminolevulinic acid and n-decane were purchased from Acros Organics. Xylenol orange was purchased from TCI Deutschland GmbH and isopropyl βD-1-thiogalactopyranoside (IPTG) from Axon Lab AG. Protein expression and purification. The P450cam variants were produced and purified as reported previously14 with slight modifications. The selenoenzyme (SeP450cam* = C357U/R365L/E366Q P450cam) was expressed in an XL1 blue E. coli strain cotransformed with plasmid pSUABC22 as before, whereas P450cam* (R365L/E366Q P450cam) and wild-type P450cam (wtP450cam) were expressed in a BL21 gold E. coli strain, which enhanced the yields of the latter proteins but not SeP450cam*. For purification of all P450cam proteins, the reported repeated ion-exchange chromatography on Q-Sepharose (GE Healthcare) was replaced by a single ion-exchange chromatography step, followed by hydrophobic interaction chromatography on butyl-Sepharose (GE Healthcare). The proteins were loaded on the pre-equilibrated butylSepharose column in 50 mM potassium phosphate buffer (pH 7.4) containing 1.5 M ammonium sulfate. The heme-containing proteins were eluted in the same buffer containing 0.45 M ammonium sulfate and 1 mM camphor; the heme-free proteins eluted at lower ammonium sulfate concentrations. This procedure afforded P450cam variants with higher A392nm/A280nm ratios (≥ 1.5), indicating greater purity. Before storage at -20 ˚C, the proteins were dialyzed overnight against 50 mM potassium phosphate buffer (pH 7.4) containing 0.5 mM DTT, 1 mM camphor, and 10% glycerol. Putidaredoxin (Pdx) and putidaredoxin reductase (PdR) were expressed in the E. coli BL21 strain from a pMG211 plasmid as C-terminally His6-tagged proteins. After cell lysis, the soluble cell fraction was subjected to Ni-NTA affinity chromatography. The colored elution fractions 7

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

were pooled and loaded onto a Superdex 75 column (GE Healthcare) and eluted with 50 mM potassium phosphate buffer (pH 7.4) containing 1 mM DTT and 10% glycerol as a final polishing step and stored at -80 ˚C. Only protein samples with A454nm/A280nm ≥ 0.6 for PdR and A280nm/A325nm ≤ 1.6 for Pdx were used for further experiments. 5-exo-Hydroxycamphor production. 5-exo-Hydroxycamphor was synthesized enzymatically in 50 ml 50 mM potassium phosphate buffer (pH 7.4) containing 5 mM camphor, 8 mM NADH, 5 µM Pdx, 1 µM PdR and 1 µM P450cam at room temperature. The depletion of camphor was monitored by GC-FID and the reaction was stopped by the addition of chloroform when the concentration of 2,5-diketocamphane started to increase. The organic compounds were extracted with chloroform, and the 5-exo-hydroxycamphor was purified by flash column chromatography (hexane/ethylacetate 3:2). A 57% yield based on the initial camphor concentration was obtained after purification, and the powder was stored at 20 ˚C. 1H NMR (300 MHz, CDCl3): δ 4.01 (dt, J = 7.0, 3.5 Hz, 1H), 2.33 (dd, J = 18.3, 5.2 Hz, 1H), 2.16 (d, J = 5.2 Hz, 1H), 1.85 (m, 2H), 1.81 – 1.74 (m, 1H), 1.69 (d, J = 18.3 Hz, 1H), 1.25 (s, 3H), 0.93 (s, 3H), 0.85 (s, 3H). 13C NMR (101 MHz, CDCl3) δ 218.35, 74.62, 58.71, 50.83, 46.50, 40.41, 39.98, 20.99, 20.12, 8.95. Control experiments with wtP450cam. SeP450cam* and the control protein P450cam* differ from wt P450cam by two amino acid substitutions (R365L and E366Q) that were required for efficient co-translational incorporation of the selenocysteine residue. To ensure the reliability of our kinetic measurements with the variants, all experiments were first conducted with wtP450cam. The data obtained with the wild-type enzyme agreed well with published kinetic parameters for each step in the catalytic cycle,23–27 confirming the validity of the experimental setup.

8

ACS Paragon Plus Environment

Page 8 of 50

Page 9 of 50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Steady-state kinetics. Steady-state kinetic parameters for the P450cam enzymes were determined by initial rate experiments measuring the absorbance decrease at 340 nm upon NADH consumption (ε = 6.22 mM-1cm-1). All measurements were performed at 25 ˚C in airsaturated 50 mM potassium phosphate buffer (pH 7.4) containing 1 mM camphor and 200 µM NADH. The net enzymatic rate was determined by subtraction of the background rate prior to P450cam addition from the rate measured after P450cam addition. In order to ensure that all the Pdx was in its reduced form during turnover we first determined the amount of PdR necessary to saturate the cytochrome at low Pdx concentration (1 µM) and a fixed cytochrome concentration (0.5 µM) for the P450cam variants.23 Next, saturating PdR concentrations were used to determine the maximal turnover rate of the P450cam-catalyzed reaction by varying the Pdx concentration. The kinetic parameters were derived by fitting the initial rates as a function of the electron donor (PdR or Pdx) concentration to the Michaelis-Menten equation by non-linear regression (Origin). Coupling efficiency. In order to determine the coupling efficiency between reducing equivalents consumed and hydroxylated product formed, reaction mixtures containing 0.5 µM P450cam variant, 0.5 µM PdR, 5 µM Pdx, 2 mM NADH and 2 mM camphor in 50 mM potassium phosphate buffer (pH 7.4) were run, monitoring the depletion of NADH and formation of 5-exo-hydroxycamphor and hydrogen peroxide. NADH oxidation was assayed by following the absorption decrease at 387 nm (ε = 0.622 mM-1 cm-1). Formation of 5-exo-hydroxycamphor was monitored by the combination of gas chromatography and mass spectrometry (GC-MS) or flame ionization detector (GC-FID). Substrate and product were extracted at different time points by quenching 200 µl aliquots with 100 µl chloroform containing 1 mM n-decane as an external standard. Peaks were assigned by 9

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

GC-MS and quantified by peak integration of the GC-FID traces and comparison with standard curves of substrate and product. Formation of H2O2 was monitored as previously described.14 In brief, 0.5 mL reaction aliquots were quenched after 1–5 min with 1 mL of cold 3% trichloroacetic acid, and H2O2 was quantified colorimetrically at 540 nm as described by Atkins and Sligar.28 Standard H2O2 concentration curves were determined by the xylenol orange assay.29 Equilibrium dissociation constants. The equilibrium dissociation constants for camphor and 5-exo-hydroxycamphor binding to the ligand-free P450cam variants were determined as described for wtP450cam.27 The absorbance at 417 nm, the heme absorbance maximum of the ligand-free P450cam protein, was monitored upon titrating the ligand-free P450cam variants (0.5 μM) with increasing concentrations of ligand (camphor 0.1 to 30 μM; 5-exo-hydroxycamphor 0.1 to 150 μM) at 25 °C. Because the enzyme variants were isolated in the presence of camphor, it was necessary to remove the substrate from the active site before conducting the binding assay. The camphorcontaining storage buffer was first exchanged with 50 mM potassium phosphate (pH 7.4) by a PD-10 desalting column (GE Healthcare), followed by addition of 100 mM imidazole to displace camphor from the active site. Imidazole was then removed in 50 mM Tris-HCl buffer (pH 7.4) on another PD-10 desalting column, followed by a final buffer exchange to 50 mM potassium phosphate (pH 7.4). Efficient removal of imidazole was confirmed by UV/Vis spectrometry. Imidazole-bound enzyme has an absorption maximum at 425 nm. The absorbance data obtained upon titration were converted to fraction of bound protein, and the dependence on ligand concentration was fitted by nonlinear regression (Origin) using the quadratic binding equation (1), where F represents the fraction of enzyme bound to the ligand, E0

10

ACS Paragon Plus Environment

Page 10 of 50

Page 11 of 50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

the initial enzyme concentration, ST the total ligand concentration, and KD the dissociation constant for the enzyme-ligand complex.

F=

E0 + ST + 𝐾𝐷 −

(E0 + ST + 𝐾D )2 − 4. E0 . ST (1) 2E0

Because KD is in the range of the enzyme concentration employed, this equation was needed to account for the fact that the total ligand concentration was not always equal to the free ligand concentration. General stopped-flow procedures and data analysis. All pre-steady-state experiments were performed on an Applied Photophysics model SX18 stopped-flow spectrometer equipped with a xenon arc lamp, which has a 1 cm path length and a dead time of 3 ms. All experiments were performed at 5 ˚C in 50 mM potassium phosphate buffer (pH 7.4) unless stated otherwise. The reported reactant concentrations in the Results section are the final concentrations, i.e. after volumes of 120 µl were mixed from each syringe. For all pre-steady-state experiments, the concentration of ligand or electron donor was kept at least 5-times higher than the P450cam concentration to ensure pseudo-first order conditions. For anaerobic experiments, solutions containing P450cam or Pdx were deoxygenated over a period of 45 min by seven to nine cycles of evacuation under house vacuum followed by equilibration with nitrogen, after which reduction was initiated under vacuum. Protein solutions were kept on ice during deoxygenation and reduction. All anaerobic experiments were performed in the presence of an O2 scrubbing system: catalase (3000 units/ml) and glucose oxidase (0.2 mg/ml) were added to the initial solution prior to deoxygenation, and glucose (1.2 mM) was added prior to the reduction step.24

11

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Typically, after averaging the traces of 5 to 10 shots, the average trace was fitted by non-linear regression (Origin) with the appropriate exponential equation after discarding the first 3 ms of data (dead time of the instrument). Inspection of residuals provided an assessment of the quality of the fits. The fitting functions had the following general form:

where y(t) is the observed signal at time t, i the number of transients, Ai the amplitude of the ith transient, ki the observed rate constant for the ith transient, and C the offset. Observed rate constants (kobs) obtained by fitting the data to equation 2 were used in replots of kobs versus concentrations of substrate, product, oxygen or reduced Pdx (Pdxred). Linear concentration dependencies of kobs, indicating a single-step binding mechanism, were fitted with equation 3, where k1 is the association rate constant, L the concentration of ligand, and k-1 the dissociation rate constant.

Equation 4 was fitted to the hyperbolic concentration dependencies of kobs for the electron transfer reactions, where ketr is the rate constant of electron transfer, KD the dissociation constant for the Pdxred-P450cam complex. Hyperbolic concentration dependency indicates a fast association step followed by unimolecular electron transfer; here the fast association step occurred within the dead time of the instrument and was not observed.

12

ACS Paragon Plus Environment

Page 12 of 50

Page 13 of 50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Substrate binding to the ferric enzyme. Association and dissociation rate constants for substrate binding were measured by stopped-flow spectroscopy as described for wtP450cam.27 The decrease in absorbance at 417 nm was monitored as a function of time upon mixing a fixed concentration of the ligand-free P450cam variant (2 μM) with increasing concentrations of camphor (10-80 μM). Electron transfer to the ferric enzyme. The rates of the first electron transfer were determined at 25 °C as described in the literature,25,30 except that the concentration of the P450cam variant was reduced to 1 μM. Protein solutions containing the oxygen scrubbing system were first deoxygenated, and then purged with CO gas for another 30 min. The Pdx solutions were reduced by adding NADH (360 μM) and a catalytic amount of PdR (0.5 μM) for 30-45 min before samples were loaded into gas-tight syringes. The reactions were initiated by mixing the ferric P450cam (1 μM) and Pdxred (5-700 μM) together and monitored by detecting formation of the ferrous CO form of the enzyme samples at 446 nm. Oxygen binding to the ferrous enzyme. Oxygenation reactions were performed as described previously.24 Solutions of P450cam (8 μM), containing catalytic amounts of PdR (0.4 μM) and Pdx (0.2 μM), were deoxygenated and reduced by adding NADH (100 μM) in buffer containing the oxygen scrubbing system. After further incubation for ~45 min, the protein solution was transferred to a gas-tight syringe. The oxygen-containing solutions were obtained by mixing airsaturated buffer (260 μM oxygen in 50 mM potassium phosphate, pH 7.4, 25 °C) in desired ratios with nitrogen-saturated buffer in a gas-tight syringe. Oxygen binding was monitored by detecting formation of the oxygenated form of the enzymes at 403 nm, an isosbestic point between ferric and ferrous cytochrome P450cam in the presence of camphor.

13

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Electron transfer to the ferrous dioxygen complex. Reduction of the ferrous dioxygen P450cam complex was analyzed by a published method using the sequential double-mixing mode of the stopped-flow spectrophotometer.25,30–32 Solutions of ferrous P450cam (5 μM) and reduced Pdx (30-160 μM) were prepared anaerobically in the standard buffer containing the oxygen scrubbing system. Subequimolar amounts of Pdx and PdR (0.1 μM each) were added to P450cam and the reduction was initiated by 100 μM NADH. A high stock concentration of Pdx was reduced by 0.05 μM PdR and 400 μM NADH. After reduction, the solutions were loaded into gas-tight syringes at the desired concentration. Ferrous P450cam was first mixed with airsaturated buffer and, after a 500 ms delay (100 ms for the selenocysteine variant), with increasing concentrations of Pdxred. The rate of reduction of the dioxygen adduct of ferrous P450cam (oxy-P450cam) was determined by following conversion of oxy-P450cam into ferric enzyme and product by the absorbance increase at 390 nm. Autoxidation of the ferrous dioxygen complex. The rate at which the ferrous dioxygen P450cam complex decays was determined by absorbance stopped-flow spectroscopy as described previously.26 The P450cam enzyme samples (5 μM) were reduced with a minimal amount of sodium dithionite in a glove box. Formation of the ferrous enzyme and the absence of residual dithionite were confirmed by UV/Vis spectrometry. The ferrous enzyme has an absorbance maximum at 408 nm and dithionite at 315 nm. The reduced P450cam samples were mixed with oxygen-saturated buffer and disappearance of the resulting oxyferrous complex was monitored at 418 nm. Computational methodology. All-electron calculations were carried out with density functional theory (DFT) programs of the TURBOMOLE package33 on minimal model complexes. The minimal model complexes for the active sites of P450cam* and SeP450cam* comprise the 14

ACS Paragon Plus Environment

Page 14 of 50

Page 15 of 50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

heme cofactor and a coordinating axial thiolate or selenolate. Coordinates were taken from the crystal structure of Pseudomonas putida P450cam (PDB entry 1DZ8) and then structurally optimized with a coordinating O2− ligand as a monoanion. The proximal ligands were represented by thioethanolate and selenoethanolate, respectively. All model complexes were treated as openshell systems in the unrestricted Kohn-Sham framework, in which different spin multiplicities were generated by choosing a corresponding number of unpaired electrons so that the total spin quantum number S is determined by the total spin magnetic quantum number MS (S = MS). For structure optimizations, we chose the TPSS exchange-correlation functional34 and a splitvalence basis set with polarization functions on all atoms (SVP)35 in combination with the resolution-of-the-identity density-fitting approximation as implemented in TURBOMOLE. Molecular structure optimizations were considered converged when the norm of the gradient had reached 10-3 Hartree/bohr and the energetic difference for the last twenty structures in the optimization procedure was below 0.2 kcal/mol. Dissociation of the O2 species from heme can occur as a neutral triplet dioxygen molecule or as a superoxide monoanion. In order to enforce the dissociation of superoxide in the structurally relaxed complexes, we fixed the O-O internuclear distance to the value of a free superoxide anion starting at the second point after the energy minimum towards dissociation. If this constraint was not introduced, triplet oxygen would be formed in the quintet and septet channels as the negative charge cannot be stabilized on the small dioxygen moiety dissociating from the much larger metal fragment in vacuo. Since our structural model is a minimal representation that neglects all environmental effects, the charged superoxide species is not stabilized upon dissociation because the dielectric environment of the active site is not modeled. As a consequence, our calculations suggest that dissociation of superoxide is energetically less 15

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

favorable than dissociation of triplet dioxygen (compare Figure 6 with Appendix Figures 7 D and 8 D where the latter process is high-lighted with a gray box), even though it is the chemically relevant process. If the entire model complex is embedded in a dielectric continuum capable of stabilizing the negative charge on the dioxygen species, then superoxide dissociation is stabilized by about 4 kcal/mol (TPSS/SVP with a conductor-like screening model (COSMO)36 as implemented in TURBOMOLE employing a relative permittivity of ϵ = 20) compared to the dissociation in vacuo; for comparison, neutral triplet oxygen is hardly affected by dielectric embedding. Since we are reporting gas-phase data, it is worth remembering that the true (electronic) dissociation energy at 0 K would likely be lower by about 4 kcal/mol if the environment were properly modeled (and even lower if the zero-point vibrational energy difference were included). For validation purposes, additional single-point calculations were performed with the hybrid functional TPSSh37 as it is known that the admixture of exact Hartree-Fock-type exchange energetically favors states with high multiplicity over those with low multiplicity;38–40 for our purposes it is therefore the most decisive parameter in the functional. For the derivation of the superoxide-dissociation curves, four different spin multiplicities of the dioxygen-bound species [Fe(III)-O2]- (i.e., singlet, triplet, quintet and septet) were considered in broken-symmetry calculations. The starting models were prepared by full structural optimization for each spin state of the O− 2 -coordinating model complexes. The dissociation curves were generated with an in-house computer program41 tailored to locate spin-state crossing points.

16

ACS Paragon Plus Environment

Page 16 of 50

Page 17 of 50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

RESULTS Protein production. The selenocysteine-containing SeP450cam* variant was produced recombinantly by bacterial expression of an engineered gene containing the requisite UGA stop codon for selenocysteine and a simplified selenocysteine insertion sequence (SECIS).14 The designed SECIS element, which is necessary for efficient stop codon suppression,42 introduced two additional surface mutations, R365L and E366Q, into the protein. To separate the effect of the sulfur-to-selenium substitution from that of the SECIS changes, we also produced the control protein P450cam*, which contains the R365L and E366Q mutations but retains the native cysteine as the axial ligand. It was analyzed in parallel with SeP450cam*. As reported previously,14 the SECIS mutations are well tolerated by the scaffold, causing only minor changes in specific activity and coupling efficiency. The latter effects can be largely ascribed to localized structural changes associated with the R365L mutation,14 which likely perturb binding of the redox partner to some extent and hence electron transfer into the active site.31,43 Although the spectroscopic properties of P450cam* and wtP450cam are essentially identical, the absorption maxima observed for SeP450cam* are slightly red shifted because of the more electron donating axial ligand. Nevertheless, these effects are small (3-5 nm, depending on the specific complex), so we analyzed all variants at the wavelengths used previously to investigate the individual steps in the catalytic cycle of the wild-type enzyme. Retention of overall activity and coupling efficiency. P450cam is the terminal monooxygenase in a three-component camphor-hydroxylating system. It obtains reducing equivalents from putidaredoxin (Pdx), which shuttles between the cytochrome and the NADHoxidizing putidaredoxin reductase (PdR). For wtP450cam, this process is tightly coupled to

17

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

camphor hydroxylation.28 As a result, the catalytic activity of P450cam can be determined by monitoring NADH depletion spectroscopically. It has been shown that at constant P450cam concentration, the concentration of PdR required to maintain Pdx in its reduced form during turnover increases with decreasing concentration of Pdx.23 We therefore first determined the saturating PdR concentration at low Pdx concentrations by varying the concentration of PdR at constant P450cam. Since the concentration of PdR required to obtain half maximal velocity was lower than 1 µM for all variants (Table 1), a 3 µM PdR concentration was assumed to be saturating for all variants at 0.5 µM P450cam and all concentrations of Pdx. Saturating PdR (3 µM), camphor (1 mM) and NADH (200 µM) concentrations were used to determine the specific activity of the different P450cam variants upon variation of the Pdx concentration. The steady-state kinetic analysis revealed that the selenocysteine substitution has essentially no effect on the specific catalytic activity or the binding interaction with the electron donor Pdx, as both kcat and KM,Pdx are very similar for SeP450cam* and P450cam* (Table 1). Direct comparison of 5-exo-hydroxycamphor formation determined by GC/FID analysis with NADH oxidation and hydrogen peroxide formation determined spectroscopically confirmed that both P450cam variants maintain high coupling efficiency between the electrons shuttled from NADH, via PdR and Pdx, to the cytochrome active site and hydroxylation of the substrate. Initial rates for NADH oxidation and product formation were the same within error for SeP450cam* and P450cam*. Moreover, the selenium substitution does not compromise coupling efficiency, judging from the similar amounts of hydrogen peroxide formed by the two enzymes (Table 1).

18

ACS Paragon Plus Environment

Page 18 of 50

Page 19 of 50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Unchanged binding affinity of substrate and product. Because cytochrome P450cam has a complex multistep kinetic mechanism, we embarked on a detailed pre-steady state kinetic analysis of SeP450cam* and P450cam* in order to determine the effect of the selenium substitution on the individual steps of the catalytic cycle. We started with binding of substrate (Figure 1, ① to ②) and product (Figure1, ① to ⑨) to the low-spin ligand-free enzymes. Titration of the ligand-free P450cam variants with increasing concentrations of substrate and product allowed determination of the respective equilibrium binding constants. These studies revealed that replacing cysteine with selenocysteine does not appreciably influence the affinity of either ligand (Table 2). In order to determine the rate constants defining the binding equilibrium, substrate binding transients were analyzed by stopped-flow absorbance spectroscopy. Upon mixing ligand-free P450cam with increasing amounts of substrate, single exponential binding transients were obtained. The linear dependence of the observed rate constant on substrate concentration afforded the association and dissociation rate constants, which are comparable for the control and selenoprotein (Table 2). The effect of selenium substitution on the intrinsic rate constants governing the substrate binding equilibrium is thus negligible. Electron transfer steps are unaffected. The P450 catalytic cycle includes two electron transfer steps from reduced Pdx to the cytochrome. The first electron transfer reduces the substrate-bound ferric form of the enzyme to the ferrous form (Figure 1, ② to ③) and has been shown to be the rate-determining step overall for wtP450cam.25,30 The second electron transfer results in reduction of the ferrous dioxygen bound intermediate (Figure 1, ④ to ⑤). Both steps were analyzed for SeP450cam* and P450cam* by absorbance stopped-flow spectroscopy.

19

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The first electron transfer event was monitored by mixing increasing concentrations of prereduced Pdx (Pdxred) with the respective substrate-bound ferric P450cam variant in CO-saturated buffer and following the increase in absorbance at 446 nm, which is diagnostic for the ferrousCO bound species.44 The transients were fitted with a single exponential for both variants and all Pdxred concentrations to give observed rate constants (kobs). In contrast to wtP450cam,25 the dependence of kobs on the Pdxred concentration (Figure 2A) did not show saturation behavior for either variant. Even concentrations up to 300 µM Pdxred failed to saturate the binding interaction between Pdxred and ferric P450cam*. Hence, only a second-order rate constant for electron transfer (ketr1/ KD1) can be accurately determined (equation 4, Table 2). The decrease in affinity of the reduced redox partner (Pdxred) for SeP450cam* and P450cam* compared to wtP450 probably reflects local structural changes caused by the SECIS mutations that perturb Pdx binding. Nevertheless, the apparent first-order rate constants measured at 100 µM Pdxred for SeP450cam* and 300 µM Pdxred P450cam* (Figure 2A, 7 s-1 and 10 s-1, respectively) are similar in magnitude to one another and comparable to the respective kcat values determined under steady-state conditions (Table 1). As for wtP450cam,25,30 the first electron transfer step is thus likely to be rate determining for both SeP450cam* and P450cam*. The second electron transfer step was analyzed by a previously published double-mixing experiment.25,32 Owing to the instability of the dioxygen-bound ferrous intermediate and the fast reduction reaction, measurements were carried out with a sequential mixing stopped-flow setup at low temperature (5 ˚C). In the first mixing step, the pre-reduced ferrous P450cam variant was mixed with oxygen-saturated buffer. After a short time delay of 100 ms for SeP450cam* and 500 ms for P450cam*, which allowed for formation of the oxygen complex with essentially no autoxidation, the solution was mixed with increasing concentrations of Pdxred. The resulting

20

ACS Paragon Plus Environment

Page 20 of 50

Page 21 of 50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

transients obtained at 390 nm were biphasic for both enzymes. Fitting the data obtained at all Pdxred concentrations to the sum of two exponentials gave observed rate constants of 5-25 s-1 and