Environment and coordination of FeMo–co in the nitrogenase metallochaperone NafY

In nitrogenase biosynthesis, the iron-molybdenum cofactor (FeMo–co) is externally assembled at scaffold proteins and delivered to the NifDK nitrogenase component by the NafY metallochaperone. Here we have used nuclear magnetic resonance, molecular dynamics, and functional analysis to elucidate the environment and coordination of FeMo–co in NafY. H121 stands as the key FeMo–co ligand. Regions near FeMo–co diverge from H121 and include the η1, α1, α2 helical lobe and a narrow path between H121 and C196.


TABLE OF CONTENTS
Materials and methods S2 Table S1: Attenuation of spins resonances in core-NafY upon FeMo-co binding S6 Table S2: NafY wild type (WT) and site-directed variants generated for this work S7 Table S3: Primer sequences used for overlapping PCR reactions S7 Table S4: Primer combinations and PCR conditions for PCR 1 and PCR 2 S7 Table S5: Purified NafY variants used in this study S7 Table S6: Mean root mean square deviations values (RMSD) and the corresponding standard deviations () along MD simulations for the three MD trajectories of the core-NafY FeMo-co complex  15 N-core-NafY purification steps S9 Fig. S3: Binding site in core-NafY located in molecular modelling S10 Fig. S4: Poisson-Boltzmann electrostatic potentials S11 Fig. S5: Model structure of the complete 243-amino acids NafY protein S12 Fig. S6: Secondary structure of core-NafY identified with DSSP S13 Fig. S7: Neighborhood of 4 Å around any atom of homocitrate S13 Fig. S8: Change of C 125 .S -FeMo-co.Fe6 distance S14 Fig. S9: SDS-PAGE analysis of NafY variants purification processes S14 Fig. S10: Analysis of NafY variant interaction with apo-NifDK S15 Fig. S11: Comparison of output for the three MD trajectories of the core-NafY/FeMo-co complex starting at the following initial structures S15 Fig. S12: Activation of apo-NifDK by NafY/FeMo-co complexes S16 References S17

MATERIALS AND METHODS
Electronic Supplementary Material (ESI) for RSC Chemical Biology. This journal is © The Royal Society of Chemistry 2021 S2 All chemicals used were of analytical grade and were used as received from the chemical manufacturer, unless otherwise indicated. Ultrapure water used in all experiments was from a Millipore system.

Bacterial strains and growth conditions
Escherichia coli DH5 and BL21 (pREP-4) strains were grown in LB medium with shaking (220-250 rpm) at 37 o C. For growth on solid medium, 1.5% agar was added to the growth culture. Antibiotics were used at standard concentrations 1 . Overexpression of GST-core-NafY (Fig. S2) was induced by addition of 1 mM IPTG to the culture medium and incubation for 3 h. Isotopic labeling was performed according to Marley 2 .
E. coli BL21 (DE3) strains 2161, 2162, 2163 and 2164 were used as host to express NafY wild-type, H121L, C196A and H121L/C196A, respectively. Cultures of 100 ml were grown in LB media supplemented with ampicillin (in 500 ml flasks) at 37C and 200 rpm until the cultures reached an optical density (OD 600 ) of 0.8-1.0. Cultures were used to inoculate 1 L LB medium supplemented with ampicillin (in 4 L flasks) at an initial OD 600 of 0.02-0.05. Cells were grown at 37C and 200 rpm until the culture reached an OD 600 of 0.6-0.8, then induced by addition of 3 g/L lactose and cultured overnight at 30C and 105 rpm. The cells were collected by centrifugation at 4,500 rpm for 10 min at 4C, frozen in liquid N 2 and stored at -80C.

Site-directed mutagenesis and cloning of Twin-Strep-tagged NafY variants
NafY variants (Table S2) were expressed from a modified pET-16b where the 10xHis-Factor Xa site had been replaced by a Twin-Strep-TagII-TEV sequence (pN2LP29). Introduction of site-directed mutations in Azotobacter vinelandii nafY was carried out by overlapping PCR using primers listed in Table S3 as described 3 . Primer combinations and PCR conditions are listed in Table S4. Plasmid pRHB62 (with the A. vinelandii nafY wild-type gene inserted into a pGEX-4T-3 plasmid for expression of GST-tagged NafY 4 ) was used as the template to generate DNA fragments for the overlapping PCR reactions (PCR 1 and 2). The overlapping PCR was performed using primers 2660 (5'-TCTTTATTTTCAAGGTCATATGGTAACCCCCGTGAACATG-3´) and 2661 (5´-CGGGCTTTGTTAGCAGCCGGATCCTCATGCCCTGGCCGCC-3´) as external primers (Tm 65.5C), generating nafY gene variants with 19 bp and 24 bp extensions (5' and 3' respectively) complementary to the cloning vector. PCR reactions were carried out using Phusion™ High-Fidelity DNA Polymerase (Thermo Scientific) according to manufacturer recommendations. The nafY gene fragments were inserted into plasmid pN2LP29 digested with NdeI and BamHI by Exonuclease and Ligation-Independent Cloning (ELIC) 5 . DNA of digested vector and nafY gene variants were mixed in a 1:4 ratio and incubated at room temperature for 10 min, and then transformed into E. coli DH5α. All nafY variants were confirmed by DNA sequencing.

Purification of 15 N labeled core-NafY
The 15 N-GST-tagged core-NafY protein was purified in aerobic conditions from cell-free extracts of induced E. coli BL21 (pREP-4) cells transformed with plasmid pRHB263 (P tac -GST-core-NafY) 6 . Cell pellets were resuspended in buffer A (10 mM sodium phosphate, 1.8 mM potassium phosphate (pH 7.3), 140 mM NaCl, and 2.7 mM KCl) and disrupted at 12,000 psi using a French press. Cell lysates were clarified by two consecutive centrifugation steps at 25,000 x g for 40 min each. The cell-free extract was loaded onto a 25-ml GSH-Sepharose Fast Flow affinity column (GE Healthcare) and washed with 250 ml of buffer A supplemented with 1% Triton X-100. The GST-tagged protein was eluted from the column by applying 75 ml of buffer B (50 mM Tris-HCl (pH 8.0) and 10 mM GSH). The GST-tag was cleaved by the addition of 10 g of purified TEV protease/mg of GST-tagged protein, followed by 2 h incubation at 30 o C. The protease reaction mixture was then subjected to three sequential chromatographic steps: (i) Ni 2+ affinity (to remove TEV protease), (ii) gel filtration on Sephadex G-25 (to remove GSH), and (iii) GSH-Sepharose (to remove the free GST tag). A typical purification protocol yielded 0.2 mg of 15 N-core-NafY per gram of E. coli cell paste. Protein purity was estimated to be >95% based on SDS-PAGE analysis (Fig.  S2). Purified 15 N-core-NafY was concentrated by ultrafiltration through a YM10 membrane (Millipore) using an Amicon cell device to 0.6 mM for nuclear magnetic resonance analysis.

Anaerobic purifications of NafY variants for FeMo-co binding and apo-NifDK activation assays
Purification procedures aimed to assay NafY activities were carried out under anaerobic conditions. Buffers were previously made anaerobic by sparging with N 2 . Each NafY variant was isolated from about 25 g of E. coli cells. Cells were resuspended in 50 ml buffer A (100 mM Tris-HCl (pH 8.2), 200 mM NaCl, 10% glycerol, 2 mM dithionite (DTH)) supplemented with 1 mM phenylmethylsulfonyl fluoride, 1 μg/ml leupeptin and 5 μg/ml DNase I. Cells were lysed in an Emulsiflex-C5 homogenizer (Avestin Inc.) at 15,000 psi. Cell-free extracts (CFE) were prepared by centrifugation of the lysates for 1 h at 50,000 x g (4C).
Purifications were performed inside a glovebox (Coy Laboratory) (<5 ppm of O 2 ) using Strep-binding affinity chromatography. The CFE was loaded using a peristaltic pump (Masterflex C/L, Cole Parmer) at 2 ml/min into a 5-ml Strep-Tactin XT Superflow Cartridge (IBA Life Science) previously equilibrated with buffer A. The NafY proteins were eluted, following 5 consecutive washes of 15 ml buffer A, with 15 ml of buffer A supplemented with 50 mM biotin, concentrated using 10 kDa cut-off centrifugal membrane filters (Amicon Ultra-15, Millipore), and desalted using a PD-10 desalting column (GE Healthcare). NafY proteins were analyzed by SDS-gels stained with Coomassie blue and stored as protein beads in liquid N 2 .

Apo-NifDK activation assays
NafY/FeMo-co complexes were generated by incubating 10 µg of NafY and FeMo-co (2-fold molar excess) in 150 µl of anaerobic 22 mM Tris-HCl buffer (pH 7.4), 2 mM DTH for 5 min on ice followed by 20 min at room temperature. Samples were supplemented with anaerobic buffer to 500 µl total volume and concentration down to 100 µl using a 10-kDa cut-off centrifugal membrane filters. Samples were subjected to four additional cycles of concentration/dilution and were left at 100 µl final volume to decrease unbound FeMo-co 2,000-fold.

S4
Extracts of A. vinelandii UW146 (ΔnifB ΔnafY) were used to study the interaction between NafY variants and apo-NifDK (Fig. S10). Twenty-eight g of A. vinelandii UW146 cells were resuspended in 50 ml of the above-described buffer A. Soluble cell-free extracts were prepared as described above.
StrepTactin-XT resin (IBA Life Science) saturated with Twin-Strep tagged NafY variants (or without NafY as control) were incubated with 10 ml of UW146 CFE for 3 h (room temperature) with mild agitation. Following transfer to empty columns for gravity chromatography, resin was washed with 35 ml buffer A. Proteins were eluted using 3.5 ml of buffer A supplemented with 50 mM biotin. Eluted proteins were concentrated to 1.5 ml using 10 kDa cut-off spin filters (Microcon, Millipore). Protein interactions were analyzed by Coomassie blue staining of SDS-PAGE gels, and stored under anaerobic conditions at -80C.

NMR Spectroscopy
NMR samples of 15 N-core-NafY were obtained by concentrating purified protein in buffer A supplemented with 0.02% (w/v) NaN 3 in either 100% 2 H 2 O or 10% (v/v) 2 H 2 O to a 0.6 mM protein concentration. 15 N-core-NafY NMR data sets were collected on a 135-amino acid polypeptide containing 132 residues of NafY (Arg 100 -Phe 231 ) plus three N-terminal residues (Gly-His-Met) introduced during the cloning procedure ( Figure S1). NMR experiments were recorded at 298 K on a Bruker Avance II 900 equipped with a CPTXI cryoprobe. NMR data were processed using NMRPipe or rNMRtk, and spectra were analyzed with NMRDraw 9 and CARA 10 .

N-core-NafY/FeMo-co complex preparation for NMR determinations
FeMo-co was purified as previously described 11 . All glassware was rinsed with 4 N HCl overnight to remove traces of contaminant metals, and then rinsed thoroughly with deionized water. All steps of the preparation were performed inside an anaerobic (<5 ppm O 2 ) glove box (Coy) with a 95% N 2 /5% H 2 environment. 15 N-core-NafY was pre-treated with 1.5 mM DTH in anaerobic buffer C (10 mM sodium phosphate, 1.8 mM potassium phosphate (pH 7.3), 140 mM NaCl, and 2.7 mM KCl) for 30 min. FeMo-co was added to purified 15 N-core-NafY in a 1.5 mol of purified FeMo-co to 1 mol of protein ratio in 1 ml of anaerobic buffer A. The mixture was incubated at room temperature for 15 min. The unbound FeMo-co and the residual NMF solvent were separated from the 15 N-core-NafY/FeMo-co complex by applying at least five cycles (20-fold dilution per cycle) of concentration and dilution with anaerobic buffer C in an Amicon cell device equipped with a YM10 membrane (Millipore). NMF was estimated to be below 0.000003% at the end of the procedure.

Protein assays
S5 SDS-PAGE was performed according to Laemmli 12 . Protein concentration in the samples (Table S5) was determined by UV-Vis spectroscopy at 280 nm or by the bicinchoninic acid method (BCA reagent, Pierce) using bovine serum albumin (BSA) as a standard 13 . Samples were pre-treated with iodoacetamide before performing the BCA assay to eliminate the interfering effect of DTH 14 . UV-visible spectroscopy was carried out in a Shimadzu (Kyoto) UV1601V spectrophotometer. The structure of core-NafY (1P90) in Fig. 1b was represented with PyMOL 15 .

Computational modelling and molecular dynamics (MD) calculations
Possible binding sites were explored in protein-ligand docking calculations with AutoDock Vina 16 and Chimera 1.15 17 . The crystal structure 1P90 of core-NafY was used for the receptor and the geometry of FeMo-co (Fe 7 S 9 CMo-homocitrate) cluster in the crystal structure of NifDK from A. vinelandii (PDB 3U7Q) 18 was used for the ligand. After locating putative sites in the vicinity of H 121 , the best docking solutions were selected for optimizing the geometry of the corresponding NafY/FeMo-co complexes using the CHARMM 3.6 force field (FF) 19 with the improved version 3.6m 20 for the protein. FF parameters for FeMo-co were prepared following the procedure by R. Björnsson 21,22 (https://sites.google.com/site/ragnarbjornsson/mm-and-qm-mm-setup). The best energies corresponded to a FeMo-co binding site between H 121 and C 196 of NafY. This was the initial geometry for all-atom MD exploratory 10-ns simulations.
Three MD trajectories were obtained for the core-NafY/FeMo-co complex (i) in that initial geometry (trajectory 1), (ii) restraining only the presence of H 121 (trajectory 2), and (iii) restraining only the presence of C 196 (trajectory 3). All MD calculations were performed with the multicore CUDA version of NAMD 2.13 23 in the Tesla V100 GPU of the high-performance computing CBGP cluster. The systems composed of protein, FeMo-co and a periodic solvation box with 14 Å spacing and the TIP3P model of water 24 were prepared with VMD 1.9.3 25 . Sodium and chloride ions were added to counter the total charges of the systems setting a 0.150 M salt concentration. The particle-mesh Ewald summation method was employed for long-range electrostatics and a 10 Å cutoff was set for short-range nonbonded vdW interactions. Initial geometries in each MD run were first minimized at 5,000 conjugategradients steps, water was then equilibrated at 1 atm and 298 K for 100 ps at t = 2 fs, and 10-ns trajectories were then obtained also at t = 2 fs (5 million steps) in the NPT ensemble at 1 atm and 298 K. T and P control was treated by means of the Langevin dynamics (T) and Nosé-Hoover Langevin piston method (P). NAMD output was saved every 5,000 steps to render trajectories composed of 1,000 frames that were processed and analyzed with VMD 1.9.3 25 .
The complete sequence of NafY consisting of 243 amino acids was modelled with Robetta 26 using the TrRefineRosetta 27 method. This procedure predicted five "best" (highest scores) models in which the core segment 90-221 has the same architecture as in the crystal structure 1P90. Four of these five models agree in predicting nearly identical geometries for both the N-terminal domain (residues 1-98) and the C-terminal 222-243 segment composed of an -helix (residues 225-243) linked to the core domain through a short loop (residues 222-224).
Poisson-Boltzmann (PB) electrostatic potentials V(r) were computed using the nonlinear PB equation with the APBS 1.5 28 Table S1. Attenuation of spins resonances in core-NafY upon FeMo-co binding. "Ratio" means intensity ratios between FeMo-co-bound and FeMo-co-less measurements. Gaps denote that the FeMo-co-bound data is ambiguous or that the resonance was unassigned. Ratios equal to zero represent true complete attenuation of the signal.   Fig. S1 Schematic representation of NafY first, second, and core domains. The core-NafY polypeptide used in this work, contains 132 amino acids (Arg 100 -Phe 231 ) plus three N-terminal residues (Gly-His-Met) introduced during the cloning procedure.             NafY + FeMo-co NafY -FeMo-co