The crystal structure of methanogen McrD, a methyl‐coenzyme M reductase‐associated protein

Methyl‐coenzyme M reductase (MCR) is a multi‐subunit (α2β2γ2) enzyme responsible for methane formation via its unique F430 cofactor. The genes responsible for producing MCR (mcrA, mcrB and mcrG) are typically colocated with two other highly conserved genes mcrC and mcrD. We present here the high‐resolution crystal structure for McrD from a human gut methanogen Methanomassiliicoccus luminyensis strain B10. The structure reveals that McrD comprises a ferredoxin‐like domain assembled into an α + β barrel‐like dimer with conformational flexibility exhibited by a functional loop. The description of the M. luminyensis McrD crystal structure contributes to our understanding of this key conserved methanogen protein typically responsible for promoting MCR activity and the production of methane, a greenhouse gas.

Methyl-coenzyme M reductase (MCR) is a multi-subunit (a 2 b 2 c 2 ) enzyme responsible for methane formation via its unique F 430 cofactor.The genes responsible for producing MCR (mcrA, mcrB and mcrG) are typically colocated with two other highly conserved genes mcrC and mcrD.We present here the high-resolution crystal structure for McrD from a human gut methanogen Methanomassiliicoccus luminyensis strain B10.The structure reveals that McrD comprises a ferredoxin-like domain assembled into an a + b barrel-like dimer with conformational flexibility exhibited by a functional loop.The description of the M. luminyensis McrD crystal structure contributes to our understanding of this key conserved methanogen protein typically responsible for promoting MCR activity and the production of methane, a greenhouse gas.
For methanogen and anaerobic methanotrophic ANME-2 genomes, the MCR genes (mcrBGA) are almost always colocated with two other genes, mcrC and mcrD (mcrBDCGA) [16][17][18].The mcrC gene is located at a different locus in ANME-1 genomes while mcrD is not present in ANME-1 and anaerobic alkane oxidising archaeal ANKA genomes [19].McrC has been identified as a component of the A3a multienzyme complex in the F 430 nickel reduction pathway for MCR activation [9].McrD has been shown to interact with MCR in immunoprecipitation and co-expression experiments leading to the proposal that McrD acts as a chaperone in MCR/F 430 assembly, facilitating MCR post-translational modification and F 430 insertion [14,15,20,21].McrD was able to alleviate product inhibition in the final step in F 430 biosynthesis catalysed by enzyme CfbE [7].McrD is not present with MCR when MCR is purified on the basis of activity, it is not required for MCR in vitro methane production and its deletion in the Methanosarcina acetivorens genome had only a minor effect on cell doubling time [13,15,22,23 [15].
In order to obtain a greater understanding of McrD biology, we have determined the high-resolution crystal structure of McrD from the human gut methanogen Methanomassiliicoccus luminyensis.The McrD structure contains a ferredoxin-like fold assembled in a dimeric a + b barrel structure, with a key functional loop that can adopt alternate conformations.The McrD crystal structure adds insight into its role as a chaperone in the assembly of the MCR complex and its cofactors enhancing our understanding of the processes that ultimately lead to methane formation.

Cloning, expression and purification of McrD
The mcrD gene from M. luminyensis strain B10 (NCBI accession: WP_019176772.1)was cloned into the TOPO pET151D expression plasmid, essentially as described previously [24].Briefly, forward 5 0 -CACCATGAGTACTGA CAAATTTGAACCG and reverse primers 5 0 -TTACCT CTTGCCTTTCTTATAATCG were used in a PCR reaction to amplify the mcrD cDNA which was then inserted into pET151D (LifeTechnologies, Carlsbad, CA, USA) using topoisomerase-mediated cloning.Colony PCR was used to identify positives clones, and the resulting plasmid purified and sequenced to verify successful cloning.Escherichia coli BL21 (strain LOBSTR; Kerafast, Boston, MA, USA) were transformed with the pET151D-McrD plasmid and grown with 1 mM IPTG added to induce McrD expression.Nickel nitriloacetic acid affinity chromatography was performed on the cell lysate containing hexa-histidine tagged McrD from the transformed and cultured E. coli.The imidazole-eluted McrD-containing fractions were dialysed into a final storage buffer (20 mM MOPS buffer, pH 7, 50 mM NaCl, 2 mM b-mercaptoethanol, 2 mM TCEP) and concentrated to 22 mgÁmL À1 .SDS-gel protein electrophoresis confirmed a high level of purity for the protein preparation.

McrD crystal structure determination
Plate-like McrD crystals were obtained in 0.03 M MgCl 2 , 0.03 M CaCl 2 , 0.05 M MES, 0.05 M imidazole, pH 6.5, 20% glycerol, 10% PEG 4000, 0.02 M carboxylic acids mix (formate, acetate, citrate, tartrate and oxamate; solution G3 from the Morpheus I Crystal Screen (Molecular Dimensions, Rotherham, UK)) at 4 °C after 3-4 days.Crystals were flash frozen in the presence of additional glycerol to a final concentration of 25% (v/v) or ethylene glycol as cryoprotectant and diffraction data collected at the Australian Synchrotron MX2 beamline using an Eiger detection system [25].X-ray diffraction data were integrated with XDS [26] before being scaled and averaged with POINTLESS/ AIMLESS [27,28].The McrD structure was determined by single isomorphous replacement with anomalous scattering (SIRAS) using SHELX [29] as implemented in CRANK2 [30].SIRAS derivative data were collected on a crystal soaked for 20 s in mother liquor containing 25% (v/v) ethylene glycol and 0.5 M NaI.The resulting interpretable electron density map was autobuilt with BUCCANEER [31], followed by cycles of manual building and refinement with COOT [32] and REFMAC5 [33] respectively, with a final R and R free of 0.178 and 0.192 at 1.65 A resolution.X-ray diffraction and refined structure statistics are provided in Table 1.Figures were prepared with CCP4MG [34] and PYMOL [35].Protein oligomerisation analysis was performed with PISA [36].

Results
The high-resolution (1.65 A) McrD crystal structure has a ferredoxin-like fold with a signature core bab-bab topology fold corresponding to the b2a1b3-b6a2b7 secondary structure elements (Fig. 1).There is an extended glycine-rich loop region (loop 5, amino acids 49-69; 6 glycines) and a 2-stranded b-sheet (b-strands 4 and 5) between the two bab motifs.A short b-strand (b1) at the N terminus forms an anti-parallel interaction with the C-terminal b-strand (b7).There are 18 and 9 amino acids present at the N terminus of monomer A and B, respectively, from the pET151D affinity tag that have interpretable electron density and are included in the structure.
McrD crystallised in C222 1 space group with a unit cell containing two molecules in the asymmetric unit   1).The two monomers of the McrD dimer superimpose with a high degree of structural similarity within their core regions excluding loop 5 which adopts a different conformation in each monomer (Fig. 3), largely as a result of a rigid body shift (root mean square difference RMSD of 0.9 A for Ca 1-48 and 70-129).The McrD monomers superimpose on one another with RMSD 3.0 A (117 Ca) when loop 5 is included.NCBI Blast multiple sequence alignment [39] and ConSurf [40] analyses of homologous McrD sequences mapped to the M. luminyensis McrD structure did not reveal any obvious conserved surfaces that might be indicative of functional regions (Fig. 4).A section of helix a1 (amino acids [26][27][28][29][30][31] and residues within or nearby loop 5 are conserved (G47, P51, G57, P58, G61, V64 and H66) suggesting that the properties of these elements are important for McrD structure and/ or function.There are also a number of conserved residues at the dimer interface (R69, L81, G122, R123, Y124, K126 and P129 C terminus) and other conserved residues dispersed across the structure as a mixture of surface exposed (R20, Q42, K70, G86) and buried residues (E13, L22, I72 and E91).Helix a2 is the most sequence variable region.The positions of the MlMcrD C-terminal residues (10 for monomer A and 9 for monomer B) are not interpretable in the electron density maps and consequently are not modelled.The MlMcrD sequence is 139 amino acids in length; residues C-terminal to the highly conserved P129 (Fig. 4) are presumed to extend in a disordered manner away from the core a + b barrel structure into a solvent channel within the crystal lattice.

Discussion
The crystal structure of McrD shows that it has a core babbab ferredoxin-like fold with an extended loop 5 region and two b-strands inserted between the two bab motifs.The MCR a and b subunit N-terminal domains, and the c subunit, also contain a ferredoxin-like fold that are decorated with additional elements [5,41]   including methanogenesis [42,43].The ferredoxin-like domain is a common fold in archaea and is also widespread in eukaryotes and bacteria [44].Structural homology analysis using DALI [45], SALAMI [46] and FOLDSEEK [47]    The displaced MCR a subunit N-terminal domain forms a b-strand that inserts into the McrD b-sheet in a parallel manner alongside strand b4 [15].Superposition of the MaMcrD structure fitted to the MaMcrD/MCR cryo-EM image reconstruction to that of the MlMcrD monomers reveals that the core ferredoxin-like fold is conserved but reveals a conformational difference in the large loop 5 (Fig. 3).The MaMcrD structure superimposes with MlMcrD monomers A and B with RSMD values of 2.7 A (102 Ca) and 4.9 A (109 Ca), respectively.The structural diversity of loop 5 shown by MlMcrD monomers A and B, and MaMcrD at the interaction interface with MCR, suggests this structural heterogeneity is likely to be important in its MCR-binding role [15].
The significance of the dimer in the function of MlMcrD is unclear at this stage given that MaMcrD binds MCR as a single molecule.Superposition homology modelling analysis shows that dimeric MlMcrD would not able to bind the MCR complex as a dimer in an homologous way to MaMcrD owing to steric clashes from the second MlMcrD monomer, and even though the MCR a, b and c subunits all contain ferredoxin-like domains within their structures the MlMcrD dimer interface does not match the interaction interface between monomeric MaMcrD and MCR.MlMcrD has a shorter sequence than MaMcrD, and most other McrD proteins, with a C-terminal deletion of an approximately 30 amino acid domain.The functional significance of this McrD C-terminal domain is unclear considering its lack of absolute conservation and it not being observed in the MaMcrD/MCR complex cryo-electron microscopy image reconstruction [15].
Overall, the M. luminyensis McrD crystal structure adds additional structural insights into the biology of this key conserved methanogen protein responsible for promoting MCR activity and methane greenhouse gas production.

Fig. 1 .
Fig. 1.Orthogonal views of the M. luminyensis McrD dimer crystal structure in cartoon representation.Monomer A is sequence colour-ramped from the N terminus (dark blue) to the C terminus (red) with secondary structure elements labelled.Monomer B is grey scale-ramped from the N terminus (black) to the C terminus (light grey).N and C label the McrD N terminus (amino acid 1) and C terminus (amino acid 129/130), respectively.
placing McrA (Mcr a subunit), McrB (b subunit), McrG (c subunit) and McrD all in the same SCOPe Fold classification [37].The ferredoxin-like fold is considered to be an ancient domain central to the origin of metabolic pathways,

Fig. 2 .
Fig. 2. The McrD electron density across a region of the dimer interface.The McrD structure is shown in stick representation coloured by atom type; carbon atoms are coloured green or grey for monomer A and monomer B respectively, oxygen are red and nitrogen are blue.The weighted 2fo-fc electron density is contoured at 1 sigma and clipped to the atoms shown at a radius of 1.5 A.
identified numerous structurally related ferredoxin-like domain proteins as homologues of McrD, including some dimeric proteins, but these analysis methods did not identify any consensus homologues.This result likely reflects the widespread presence of ferredoxin-like domains within proteomes, the non-typical a + b barrel dimer McrD structure and also the specialised functional role of McrD.Comparison of the two molecules in the MlMcrD dimeric crystal structure reveals that loop 5 adopts a different conformation relative to the core ferredoxin-like domain for each of the two monomers.The M. acetivorans McrD-MCR complex structure determined by cryo-electron microscopy shows that McrD interacts asymmetrically as a single molecule with the MCR a 2 b 2 c 2 hexamer [15].MaMcrD binds to MCR in both the apo-apo (free of F 430 , CoM and CoB cofactors) or semi-apo where the cofactors are present in a single active site suggesting that McrD assists in the assembly of an intermediate MCR state [15].MaMcrD binds the MCR a 2 b 2 c 2 hexamer at the interface of the a, a 0 (a 0 is the second a subunit of the MCR a 2 b 2 c 2 hexamer) and c subunits with the McrD loop 5 located between the MCR a 0 and c subunits, occupying the position normally filled by the MCR a subunit N-terminal domain when McrD is not bound.

Fig. 3 .
Fig. 3. Superposition of the M. luminyensis (Ml ) and M. acetivorans (Ma) McrD structures in cartoon representation.MlMcrD monomer A (Ml-A) is colour-ramped N to C terminus from blue to red, MlMcrD monomer B (Ml-B) is grey scale-ramped N to C termini black to light grey MaMcrD (Ma) (PDB: 8GF6) [15] is shown in purple.Loop 5 and the termini are labelled.The MlMcrD amino acids within the plasmid-encoded affinity tag were not included in the superposition and are not shown.

Fig. 4 .
Fig. 4. McrD ConSurf [40] sequence conservation analysis mapped on to the M. luminyensis McrD crystal structure.The dimer structure is shown in cartoon representation with monomer A colour-ramped by sequence conservation as calculated from 150 sequences (dark magenta most conserved, white intermediate and cyan least conserved).Monomer B is coloured yellow.The N-terminal plasmid tag-encoded amino acids in the structure were removed for this analysis.

Table 1 .
Crystallographic X-ray diffraction data and refined structure statistics.