Localized Electronic Structure of Nitrogenase FeMoco Revealed by Selenium K-Edge High Resolution X-ray Absorption Spectroscopy

The size and complexity of Mo-dependent nitrogenase, a multicomponent enzyme capable of reducing dinitrogen to ammonia, have made a detailed understanding of the FeMo cofactor (FeMoco) active site electronic structure an ongoing challenge. Selective substitution of sulfur by selenium in FeMoco affords a unique probe wherein local Fe–Se interactions can be directly interrogated via high-energy resolution fluorescence detected X-ray absorption spectroscopic (HERFD XAS) and extended X-ray absorption fine structure (EXAFS) studies. These studies reveal a significant asymmetry in the electronic distribution of the FeMoco, suggesting a more localized electronic structure picture than is typically assumed for iron–sulfur clusters. Supported by experimental small molecule model data in combination with time dependent density functional theory (TDDFT) calculations, the HERFD XAS data is consistent with an assignment of Fe2/Fe6 as an antiferromagnetically coupled diferric pair. HERFD XAS and EXAFS have also been applied to Se-substituted CO-inhibited MoFe protein, demonstrating the ability of these methods to reveal electronic and structural changes that occur upon substrate binding. These results emphasize the utility of Se HERFD XAS and EXAFS for selectively probing the local electronic and geometric structure of FeMoco.

mixtures were allowed to react at room temperature with stirring for 20 minutes. The mixtures were then were concentrated by ultrafiltration with 100,000 kDa MWCO filters and washed 3-5 times with cold size exclusion buffer (50 mM Tris/HCl, pH 7.5; 200 mM NaCl; 5 mM Na 2 S 2 O 4 ) in the anaerobic chamber. CO was regularly bubbled through the solution until flash-frozen for future use. After washing the samples, the protein was loaded onto a size exclusion column (S200, 26/60, GE Healthcare, 450 mL) which had been pre-equilibrated with degassed size exclusion buffer bubbled through with CO. Av1SeCO was constantly bubbled through with CO during the washing steps. For crystallography, Av1SeCO concentration was adjusted to 30 mg/mL. For HERFD, Av1SeCO concentration was adjusted to 200 mg/mL.

Reactivation of Se-labeled, CO-inhibited FeMo-cofactor -Av1Se reac
For Av1Se reac , the protein was labeled as usual with Se, inhibited with CO as described, above and then reactivated from CO. Reactivation from CO consisted of bubbling a fresh reaction mixture containing all activity components with Ar to replace excess CO. Importantly, a fresh reaction mixture consists of 20 mM creatine phosphate, 5 mM ATP, 5 mM MgCl 2 , 25 units/mL phosphocreatine kinase, and 25 mM Na 2 S 2 O 4 , and newly added Av2 at a ratio of Av2:Av1=1:1. Reactivation was assessed in a parallel assay vials for acetylene reduction activity.Then this reactivated Av1Se reac was concentrated via ultrafiltration as described, and then applied to a cold sizing column. The protein was then collected and the concentration was adjusted to approximately 20 mg/mL. Reactivation from CO of native FeMoco (no Se labeling) has previously been reported, 7 with the reincorporation of S at the 2B position confirmed crystallographically. Additionally, reactivated MoFe was found to exhibit similar N 2 reduction activity as native protein.

Crystallization of MoFe
All Av1-Se samples were crystallized using the sitting drop vapor diffusion method at 21° C in an anaerobic chamber (5% H 2 / 95% Ar), with protein concentration at 30 mg/mL. The reservoir solution consisted of: 24-28% PEG 6000 (v/v), 750-850 mM NaCl, 100 mM imidazole/malic acid (pH 8.0), and 5 mM Na 2 S 2 O 4 . 5, 7-8 For CO-inhibited samples, the reservoir solutions were bubbled through with CO before setting the drops. After the drops were set, clear Duck tape was used to seal the reservoirs. The tape was then punctured with a syringe and the wells were quickly filled with an additional 2 mL of CO, and resealed with a second layer of clear Duck tape. A seeding strategy using crushed Av1 crystals was employed to accelerate crystallization and to generate higher quality crystals. Cryo-protection of the crystals before freezing in liquid nitrogen consisted of dipping the crystals into a 5 uL drop of reservoir solution containing 8-12% MPD (v/v).

Preparation of protein samples for XAS
Protein samples used for XAS data collection were concentrated to a final volume of 100 -150 uL in an anaerobic chamber. The highly concentrated samples (200 mg/mL) required slow dispensation into XAS cuvettes, since capillary action with the highly concentrated sample made it difficult to add to the sample cell all at once. 10 uL of the highly concentrated sample were added to the cuvette at a time, and the sample cell was then centrifuged for 60 seconds at 1000 rpm. This process was repeated until the entire sample had been dispensed into the cuvette. Protein samples were then frozen inside the anaerobic chamber with liquid nitrogen, and kept frozen during every subsequent manipulation and sample measurement. The sample cells used were machined 5 x 5 x 20 mm Delrin cuvettes with a sample measurement window of 38um Kapton tape and a sample volume of approximately 120 uL. S6 At 12662 eV, the atoms of interest were determined to have experimental scattering factors , , and . The scattering factor at 12662 eV was found to ′′ = 1.5 ′′ = 4.42 ′′ = 0.24 ′′ vary by approximately 20% over the protein samples, and therefore the average value was used for all samples. This experimentally observed variation in the scattering factor is attributed to chemical effects at the Se absorption edge. While it is anticipated to affect the absolute Se occupancies, the observed variation in should have a negligible effect on the relative Se ′′ ratios. The data were indexed, integrated, and scaled using XDS and Aimless. [10][11][12] Molecular replacement was used to obtain phase information, with the 1.0 Å Av1 structure (PDB: 3U7Q) used as a model. 8 Structure refinement and modeling was performed using the CCP4-embedded programs REFMAC5 and Coot. [12][13][14] Se-anomalous density maps were calculated based on the data collected at 12662 eV, while the S-anomalous maps were calculated based on the data collected at 7110 eV. Refmac-refined electron density data files (mtz files) were merged with anomalous mtz files using the CCP4embedded program CAD. Maps were generated using fast-Fourier transform (FFT) of the merged mtz files and the Refmac-refined pdb coordinates. The CCP4 program Mapmask was used to orient the model based on the data resolution limits. Finally, the program Mapman performs map normalization and extracts the density values on an arbitrary scale at all atomic positions within a set sphere. 15 The structures were then rendered for publication in PYMOL ( Figure S1).
Samples Av1Se lo and Av1Se hi were found to exhibit similar Se occupancies at each of the three bridging sites-2B, 3A, and 5A-with roughly a 3:1 Se2B:Se3A/5A ratio (Table S1). Av1Se reac exhibits a smaller Se2B:3A/5A ratio of nearly 1:1, predominantly due to decreased Se occupancy at the 2B position, while Av1SeCO exhibits a ratio of nearly 1:9 Se2B:Se3A/5A. An important observation from the preparation of Av1Se reac is that the highest Se occupancy occurs at the 2B site. Consequently, a rearrangement occurs during reactivation of CO inhibited FeMoco such that species at the 3A/5A belt positions can migrate to the 2B position. This observation complements the finding that the 2B site migrates to the 3A/5A positions during reaction with CO, as described in the original characterization of the reactivities of Se and CO to the FeMoco. 6

Se Quantitation via Inductively-Coupled Plasma Mass Spectrometric (ICP-MS) Studies
Excess sample for HERFD experiments as well as protein crystals were used for ICP-MS. Both sets of samples were not subjected to X-ray radiation. The solution samples were directly dissolved in 3 mL of 2% nitric acid (HNO 3 ), while the crystal samples were resuspended in 50 uL of ddH 2 O before being mixed with 3 mL of 2% HNO 3 . All samples were analyzed using a quadrupole-based inductively coupled plasma-mass spectrometer (Agilent 8800 ICP-QQQ) at the Environmental Analysis Center at Caltech. Samples were analyzed in helium mode to avoid any argon-argon dimers that may form in the plasma. Standards were generated by comparing ion counts from prepared selenium, molybdenum, and iron standards, all prepared in 2% HNO 3 .
Absolute Se concentrations were taken relative to absolute Mo and Fe concentrations (assuming 1 Mo and 15 Fe per FeMoco/P cluster pair), and the average of the two measures is reported as Se:FeMoco in Table S1. For comparison, a separately prepared wild-type MoFe sample (Av1 wt ) was also quantified by ICP-MS and revealed 0.09 equivalents of Se per FeMoco. This value is interpreted as the minimum error for the ICP-MS experiments. S11 EPR Spectra of Se-substituted FeMoco samples EPR spectra were recorded at the Caltech Electron Paramagnetic Resonance (EPR) Facility with a Bruker EMX X-band CW-EPR spectrometer using an Oxford ESR 900 liquid helium/nitrogen flowthrough cryostat. The temperature was 10K, microwave power 2 mW, and microwave frequency 9.46GHz. The Se-substituted FeMoco sample Av1Se EPR ( Figure S2, red) reveals a broadened S = 3/2 signal centered around 1800 Gauss, qualitatively similar to the S = 3/2 native FeMoco signal ( Figure S2, blue). The S =1/2 signal in Av1Se EPR centered near 3400 Gauss corresponds to the reduced F-cluster (the sample was not purified via size-exclusion chromatography). The CO-inhibited, Se-substituted sample Av1SeCO EPR ( Figure S2, green) reveals the near-complete loss of the S = 3/2 signal and generation of a new S = 1/2 signal consistent with the lo-CO species previously reported in the literature. 16 Figure S2. Continuous wave X-band EPR of native Av1 (blue), Av1Se EPR (red) and Av1SeCO EPR (green). Principal g-values for S = 3/2 of native Av1 and S = 1/2 of Av1SeCO EPR are emphasized with gray vertical lines. Av1Se EPR data between 3000 and 4000 gauss scaled to 33% for clarity. S12 Figure S3 reveals linear correlations of pre-edge area and  d with the average formal Fe oxidation state in a synthetic model complex series supported by chelating dianionic bis(benzimidazolato) ligands (S XAS spectra and fits, Figure S4). This series is unique in that it is the only well characterized Fe 2 S 2 series spanning all three oxidation state levels (diferric, mixed-valent, and diferrous). Similar trends have been observed in other complexes for both terminal and bridging ligands at the S and Cl K-edges, as shown in Figure S5-S7. In these cases, the monomeric species exist only in two oxidation states, (Fe II and Fe III ), while dimeric species are only stable in one oxidation state (Fe III Fe III ). The experimental data (solid squares for the monomers, solid circles for the dimers) has been complemented by theoretical values determined for the hypothetical [Fe II Fe III ] and [Fe II Fe II ] species (open circles). The experimental data exhibits the general trend that pre-edge intensity increases and | d | increases as the average formal oxidation state of Fe increases, while the dimeric complexes (including the theoretical values) reproduces the linear correlations observed in Figure S3.

Fits of Experimental Se HERFD XAS and S XAS Spectra
The experimental data is shown in black. Individual edge Gaussians are shown as thin red lines and their sum is given as a thick red line. Individual pre-edge pseudo-Voigt functions are shown as thin dark blue lines and their sum is given as a thick dark blue line. The difference between the fit and the experimental data is shown as a thin light blue line.

Fits of Calculated TDDFT S and Se XAS Spectra
The TDDFT data is shown in black. Individual edge Gaussians are shown as thin red lines and their sum is given as a thick red line. Individual pre-edge pseudo-Voigt functions are shown as thin dark blue lines and their sum is given as a thick dark blue line. The difference between the fit and the TDDFT data is shown as a thin light blue line.