Protonation state of the Cu4S2 CuZ site in nitrous oxide reductase: redox dependence and insight into reactivity

The edge ligand in the Cu4S2 CuZ form of nitrous oxide reductase is a μ2-thiolate in the 1-hole and a μ2-sulfide in the 2-hole redox state, leading to proton-coupled electron transfer reactivity.

and 40% Cu Z *) that had been incubated with 100 equivalents of reduced methyl viologen. The methyl viologen was removed using a PD-10 Sephadex G-25 medium (GE HealthCare) desalting column with 100 mM phosphate at pH 7.6 as the elution buffer. The protein-containing column fractions were concentrated by centrifugation using Amicon Ultra concentrators with an Ultracell regenerated cellulose membrane (Millipore). For pH dependence studies, during the concentration step samples were buffer exchanged to 100 mM MES pD 6.0, 100 mM phosphate pD 7.6, or 100 mM CAPS pD 10. The total spin intensity observed by EPR was not changed by buffer exchanging to different pHs. To determine the effect of deuteration, samples were prepared in parallel at pH/pD 7.6 and pH/pD 10. Samples of 2-hole Cu Z were prepared by reducing MhN 2 OR (60±10% Cu Z , 40±10% Cu Z *) with 10 equivalents of sodium ascorbate, which reduces the Cu A site rapidly and the 2-hole Cu Z site very slowly, and spectra were collected within 1 hour so that minimal reduction of 2-hole Cu Z was observed. In parallel, MhN 2 OR samples containing 90±10% Cu Z * were reduced with 10 equivalents of sodium ascorbate to obtain the spectral features of 1-hole Cu Z *. Corrected absorption spectra of 2-hole Cu Z were obtained by subtracting the spectral contribution of the appropriate concentration of 1-hole Cu Z *. Ascorbate reduced samples of MhN 2 OR containing Cu Z were buffer exchanged by centrifugation to 100 mM MES pD 6.0, 100 mM phosphate pD 7.6, or 100 mM CAPS pD 10 for pH dependence experiments. Typical MhN 2 OR concentrations used for spectroscopic samples were 0.1-0.3 mM for absorption, MCD and EPR, and up to 0.5 mM for resonance Raman. The concentration of the dimer MhN 2 OR was determined using the extinction coefficient of 7100 M -1 cm -1 at 640 nm for the dimer in the dithionite reduced spectrum in the purified protein and corrected according to the volume changes involved in spectroscopic sample preparation. 4 Absorption spectra were acquired in a Teflon-sealed 3 mm small volume quartz cell at room temperature using an Agilent 8453 UV-visible spectrophotometer with deuterium and tungsten sources. MCD samples were prepared by mixing protein samples in deuterated buffer 1:1 with deuterated glycerol. MCD spectra were collected on CD spectropolarimeters (Jasco J810 with an S20 PMT detector for the 300-900 nm region and a Jasco J730 with an InSb detector for the 600-2000 nm region) with sample compartments modified to insert magnetocryostats (Oxford Instruments SM4-7T). Low temperature absorption spectra were additionally obtained from the samples used for MCD using a double-beam Cary 500 spectrophotometer modified to accommodate a liquid helium cryostat (Janis Research Super Vari-Temp). Low temperature absorption spectra were corrected by subtracting the background spectrum from a cell containing a 50:50 mixture of buffer and deuterated glycerol and an additional scattering correction to account for the differences in glassing between the protein and background samples. EPR and resonance Raman samples were frozen in 4 mm diameter quartz sample tubes. EPR spectra were collected using a Bruker EMX spectrometer with an ER 041 XG microwave bridge, and an ER4102ST sample cavity for X-band and an ER 051 QR microwave bridge, an ER 5106QT resonator, and an Oxford continuous-flow CF935 cryostat for Qband. X-band samples were run at 77 K in a liquid nitrogen finger dewar. Q-band samples were run at 77 K using a cooling He gas flow. EPR spectra were baseline corrected using WinEPR (Bruker) and simulated using Simfonia (Bruker). Resonance Raman spectra were collected using a series of lines from a Kr + ion laser (Coherent 190CK), a Ti-sapphire laser (M-squared SolsTice, pumped by a 12 W Lighthouse Photonics Sprout diode pumped solid state laser), and a Dye laser (Rhodamine 6G, Coherent 699) with incident power of 20-30 mW arranged in a 130° backscattering configuration. The scattered light was dispersed through a triple monochromator (Spex 1877 CP, with 1200, 1800, and 2400 grooves mm -1 gratings) and detected with a backilluminated CCD camera (Andor iDus model). Resonance Raman samples were immersed in a liquid nitrogen finger dewar at 77 K. The spectrum of black carbon in an identical quartz EPR tube was subtracted to remove the spectral contribution from quartz scattering. The intensity of the ice peak at ~229 cm -1 was used to normalize the intensities of vibrations to obtain resonance Raman excitation profiles.

S1.3 Computational Details
A computational model of Cu Z was built from the atomic coordinates of the crystal structure of Pseudomonas stutzeri N 2 OR, the only known structure of the Cu 4 S 2 cluster (PDB ID 3SBP, resolution 1.7 Å). 5 The model included the Cu 4 S 2 core and 7 ligating His residues, where the α carbon and distal nitrogen were constrained at their crystallographic positions. A computational model for Cu Z * with a hydroxide bridging ligand and identical α carbon and distal nitrogen constraints was constructed from the crystal structure of Paracoccus denitrificans N 2 OR (PdN 2 OR, PDB ID 1FWX). 6 Two larger structural models were also optimized: (1) including a second sphere Lys-Glu salt bridge near the Cu I -Cu IV edge in both sites (Lys397 for PdN 2 OR and Lys454 for PsN 2 OR) 7 for the 1-hole redox state of Cu Z with an SHedge ligand and for 1-hole Cu Z * with an OHedge ligand and (2) including two second sphere carboxylate residues, Asp127 and Asp240 (which hydrogen bond to the His ligands of Cu I and Cu II ), in optimizations of 1-hole and 2-hole Cu Z with SHor S 2edge ligation. Including the second sphere residues did not significantly perturb the core Cu-S bond lengths (Table  S3), geometries or the spin distribution of the cluster, so the smaller computational model lacking second sphere residues was used for the analysis of the spectroscopic properties of Cu Z . Calculations were performed using Gaussian 09 (version d01). 8 Molecular structures and frequencies were visualized using Avogadro, an open source molecular builder and visualization tool (Version 1.1.1). 9 VMD 1.9.1 was used to visualize molecular orbitals, 10 and QMForge was used to obtain Mulliken spin populations of different orbitals and to analyze TD-DFT calculations. 11 Geometry optimizations were performed using the B3LYP functional, the TZVP basis set on all core atoms (Cu 4 S 2 ) and the ligating His nitrogens, and the SV basis set on all remaining atoms. The optimizations of the large model (2), including two second sphere Asp residues, were additionally performed using a larger basis set with TZVP on the Cu 4 S 2 core and all His ring heavy atoms. The resulting structures, spin distributions, and relative energies of the singlet and triplet ground states did not differ significantly from the smaller basis set optimizations (see Tables S4-S6), thus the structures optimized with TZVP only on the Cu 4 S 2 core and ligating nitrogens were used for frequency, TD DFT, and single point calculations. Optimizations, single points, frequencies, and TD-DFT calculations were performed with PCM values of 4.0 and 10.0 and no significant change in the spin distribution, frequencies, or TD-DFT was observed. Spin distributions, frequencies, and TD-DFT results reported are from calculations with a PCM of 4.0. TD-DFT calculations were additionally performed with the functional B98, which has previously shown to predict the experimental spectrum of a Cu 3 S 2 model complex. 12 As described in the Analysis, models with an edge S 2or SHwere optimized for both 1-hole and 2-hole redox states and the 2-hole models were optimized in both triplet and broken symmetry singlet spin states (the singlet states were always lower in energy). This was further tested with the functionals M06L, M06, and TPSSh, which all predict that singlet ground states are lower in energy by at least 5 kcal/mol for the 2-hole redox state; however, these functionals predict a restricted singlet ground state wavefunction for the 2-hole Cu Z site, while B3LYP predicts a more chemically reasonable spin polarized wavefunction.
To determine the relative energy of deprotonation (ΔΔE) of the edge SHin the 2hole versus 1-hole redox state, the larger structural models including two second sphere Asp residues were used. The directly deprotonated versions of the 2Asp model (with loss of H + to solvent) were considered and the energies of an internal proton transfer from the edge SHto Asp127 were also calculated for the 1-hole and 2-hole redox states. The PCM dependence of the ΔΔE of deprotonation was also evaluated by obtaining the singlet point energies with different PCM values for structures optimized with a PCM of 4.0. The ΔG for deprotonation of 1-hole Cu Z was also approximated from frequency calculations using structures with identical fixed atom constraints and the same number and magnitude for the imaginary frequencies associated with the fixed atom constraints. To minimize the error introduced into the vibrational energy correction by the fixed atom constraints, the masses of the fixed atoms were artificially increased until the calculated ΔG correction showed no further significant dependence on the fixed atom masses (~200 Da). The resulting ΔG was very similar to the calculated ΔE, indicating that the energy differences are electronic in origin.

S2 Supporting Figures, Tables, and Schemes
Figure S1: Methyl viologen reduction of three dithionite reduced samples of N 2 OR to quantify the % Cu Z from the % spin remaining after reduction. Black and red: N 2 OR prepared with two aerobic chromatographic steps ("Form I"), containing ~60% Cu Z and ~40% Cu Z *. Green: N 2 OR prepared with three aerobic chromatographic steps ("Form II"), containing ~90% Cu Z * and ~10% Cu Z . Figure S2: Second derivative (black) of the X band EPR spectrum of 1-hole Cu Z at 77 K, 9.6349 GHz, with simulation in red.  Table S1: Transition energies, C 0 /D 0 ratios, and assignments from simultaneous fitting of the low temperature absorption and MCD spectra of 1-hole Cu Z and 1-hole Cu Z * (values and assignments for Cu Z * reproduced from Ref. 7 ).

5
Scheme S1: Simplified acceptor and donor MOs for sulfide charge transfer transitions in resting Cu Z *, derived from DFT calculations of the Cu Z * cluster with a hydroxide bridged edge (B3LYP, tzvp on Cu 4 SON 7 , sv on remainder, PCM=4.0). 6 x y z Figure S3: Resonance Raman spectrum of resting Cu Z * at 77 K and 605 nm excitation; excitation profile of resting Cu Z *. Reproduced from Ref. 7 . Figure S4: Lack of pH dependence of the MCD, resonance Raman, and EPR spectra of 1-hole Cu Z at pD 6 (red), pD 7.6 (green) and pD 10 (blue). Note that the presence of weak S-H bends in the resonance Raman spectra at pD 6 and pD 7.6 is due to incomplete deuteration of the samples. Figure S5: Absorption spectrum of 2-hole Cu Z at 10 K, after ascorbate reduction and before (blue) or after (black) subtraction of the spectral contribution of 1-hole Cu Z *. Figure S6: Resonance Raman spectra of ascorbate reduced N 2 OR containing 2-hole Cu Z and 1hole Cu Z * at pD 6.0 (red), pD 7.6 (green), and pD 10 (blue), with energies of the vibrations of the 2-hole Cu Z site labeled.        Table S4.          Table S6.