Structure of an Ancient Respiratory System.

KEY RESOURCES TABLE REAGENT or RESOURCE SOURCE IDENTIFIER

Bacterial and Virus Strains

Pyrococcus furiosus strain MW0414 ( McTernan et al., 2014 ) MW0414 Pyrococcus furiosus strain COM1 ( Lipscomb et al., 2011 ) COM1 Pyrococcus furiosus strain MW0574 This study ΔmbhABC Chemicals, Peptides, and Recombinant Proteins EPPS Sigma Cat#E9502 n-dodecyl-β-D-maltoside Inalco Cat#1758-1350 Critical Commercial Assays His-Trap crude FF Ni-NTA GE Healthcare Cat#17-5286-01 His-Trap HP Ni-NTA GE Healthcare Cat#17-5247-01 Superose 6, 10/300 GL GE Healthcare Cat#17517201 Deposited Data Coordinates of MBH complex This study PDB: 6CFW Cryo-EM map of MBH complex This study EMDB: EMD-7468 Coordinates of Thermus thermophilus Complex I ( Baradaran et al., 2013 ) PDB: 4hea Coordinates of D. gigas [NiFe] hydrogenase (Volbeda et al., 1996) PDB: 2frv Oligonucleotides taccccatacttccttacttgctcgtacattctttttgaaagctctgctc This study UFR F tgagggcctctagaatgttccgccaaacctccttaacatt This study UFR R gaacattctagaggccctcagtggg ( Lipscomb et al., 2011 ) Marker F gattgaaaatggagtgagctgagttaatgatgacc ( Lipscomb et al., 2011 ) Marker R tcattaactcagctcactccattttcaatcgtgagaaaaatgaatcttgacatga This study DFR F tttgagatggcatacataaccaaagcagtaacaaccccag This study DFR R gtcctcacctcctgccctaacttgg This study PCR Screen F cgtccggaaatctgtggagggctatg This study PCR Screen R Recombinant DNA pGL021 ( Lipscomb et al., 2011 ) pGL021 ΔmbhABC Knock-in cassette This study Knock-in cassette Software and Algorithms SerialEM ( Mastronarde, 2005 ) http://bio3d.colorado.edu/SerialEM FEI EPU FEI https://www.fei.com/software/epu/ MotionCor2 ( Zheng et al., 2017 ) http://msg.ucsf.edu/em/software/motioncor2.html CTFFIND4 ( Rohou and Grigorieff, 2015 ) http://grigoriefflab.janelia.org/ctffind4 RELION-2.0 ( Kimanius et al., 2016 ) http://www2.mrc-lmb.cam.ac.uk/relion EMAN2 ( Tang et al., 2007 ) http://blake.bcm.edu/emanwiki/EMAN2 ResMap ( Kucukelbir et al., 2014 ) http://resmap.sourceforge.net/ SWISS-MODEL server ( Arnold et al., 2006 ) https://swissmodel.expasy.org/ CHIMERA ( Pettersen et al., 2004 ) https://www.cgl.ucsf.edu/chimera Coot ( Emsley et al., 2010 ) https://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot PHENIX ( Adams et al., 2010 ) https://www.phenix-online.org/ MOLPROBITY ( Chen et al., 2010 ) http://molprobity.biochem.duke.edu/ PyMOL Schrödinger, LLC. http://www.pymol.org/2/ Other C-Flat Cu CF-1.2/1.3, 400-mesh grids ELECTRON MICROSCOPY SCIENCES Cat#CFT413-50 CONTACT FOR REAGENT AND RESOURCE SHARING Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Huilin Li ( huilin.li@vai.org ) EXPERIMENTAL MODEL AND SUBJECT DETAILS Pyrococcus furiosus strain MW0414, COM1 and MW0574 were grown in defined maltose medium consisting of 1x base salts, 1x trace minerals, 1x vitamin solution, 2x 19-amino-acid solution, 10 μM sodium tungstate, 0.25 mg/ml resazurin, and 0.5% (wt/vol) maltose, with added cysteine at 0.5 g/liter, sodium sulfide at 0.5 g/liter, and sodium bicarbonate at 1 g/liter; and adjusted to pH 6.8 (Detailed buffer composition was described in the METHODS DETAILS ). For protein purification, large-scale growth of P. furiosus strain MW0414 was carried out in a 20-liter fermenter at 90 °C with constant flushing of 20% (v/v) CO 2 and 80% (v/v) N 2 for 14 hours, while the pH was maintained at 6.8 by the addition of 10% (w/v) sodium bicarbonate. METHODS DETAILS Expression and purification of MBH The MBH holoenzyme (S-MBH) was solubilized and purified anaerobically from Pyrococcus furiosus strain MW0414, in which a His 9 -tag had been engineered at the N-terminus of the MbhJ subunit. The procedure was as previously described with some modifications ( McTernan et al., 2014 ). Frozen cells were lysed in 25 mM sodium phosphate, pH 7.5, containing 1 mM DTT and 50 μg/ml DNase I (5 ml per gram of frozen cells). After stirring for one hour, the cell-free extract was centrifuged at 100,000 × g for one hour. The supernatant was removed and the membranes were washed twice using 50 mM EPPS buffer, pH 8.0, containing 5 mM MgCl 2 , 50 mM NaCl, 10% (v/v) glycerol, 1 mM DTT and 0.1 mM PMSF. The membrane pellet was collected by ultracentrifugation at 100,000 × g for one hour after each wash step. The washed membranes were resuspended in 50 mM Tris-HCl, pH 8.0, containing 5 mM MgCl 2 , 50 mM NaCl, 5% (v/v) glycerol, 1 mM DTT, and 0.1 mM PMSF. MBH was solubilized by adding n-dodecyl-β-D-maltoside (DDM, Inalco) to 3% (w/v) followed by incubation at 4 °C for 16 hours. The solubilized membranes were centrifuged at 100,000 × g for 1 hour. The supernatant was applied to a 5-ml His-Trap crude FF Ni-NTA column (GE Healthcare) while diluting it 10-fold with buffer A (25 mM sodium phosphate, 300 mM NaCl, pH 7.5, containing 1 mM DTT and 0.03 % DDM). The column was washed with 10 column volumes of buffer A and the bound protein was eluted with a 20-column volume gradient from 0 to 100 % buffer B (buffer A containing 500 mM imidazole). The eluted protein was further purified by applying it to a 1-mL His-Trap HP Ni-NTA column (GE Healthcare) while diluting it 5-fold with buffer A. A 30-column volume gradient from 0 to 100 % buffer B was used to elute the bound protein. The MBH sample was concentrated and further purified using a Superose 6 10/300 GL column (GE Healthcare) equilibrated with 50 mM Tris-HCl, pH 8.2, containing 300 mM NaCl, 2 mM sodium dithionite, and 0.03 % DDM. Deletion of MbhABC The genetically tractable P. furiosus strain COM1 was used to delete the three genes PF1423-1425. 500 bp flanking regions were amplified from P. furiosus genomic DNA for the UFR and DFR, and the selection marker ( pyrF- P gdh ) was amplified by using pGL021 as the template ( Lipscomb et al., 2011 ). The knock-in cassette was assembled using overlapping PCR ( Bryksin and Matsumura, 2010 ). The genomic DNA was prepared using Zymobead Genomic DNA Kit (Zymo Research). P. furiosus transformants were grown in defined maltose media as previously described ( Lipscomb et al., 2011 ). The maltose medium was composed of 1x base salts, salts, 1x trace minerals, 1x vitamin solution, 2x 19-amino-acid solution, 0.5% (wt/vol) maltose, 10 μM sodium tungstate, and 0.25 mg/ml resazurin, with added cysteine at 0.5 g/liter, sodium sulfide at 0.5 g/liter, sodium bicarbonate at 1 g/liter, and 1 mM sodium phosphate buffer (pH 6.8). The 5x base salts stock solution contained (per liter): 140 g of NaCl, 17.5 g of MgSO 4 ·7H 2 O, 13.5 g of MgCl 2 ·6H 2 O, 1.65 g of KCl, 1.25 g of NH 4 Cl, and 0.70 g of CaCl 2 ·2H 2 O. The 1000x trace mineral stock solution contained (per liter) 1 ml of HCl (concentrated), 0.5 g of Na 4 EDTA, 2.0 g of FeCl 3 , 0.05 g of H 3 BO 3 , 0.05 g of ZnCl 2 , 0.03 g of CuCl 2 ·2H 2 O, 0.05 g of MnCl 2 ·4H 2 O 0.05 g of (NH 4 ) 2 MoO 4 , 0.05 g of AlK(SO 4 )·2H 2 O, 0.05 g of CoCl 2 ·6H 2 O, and 0.05 g of NiCl 2 ·6H 2 O. The 200x vitamin stock solution contained (per liter) 10 mg each of niacin, pantothenate, lipoic acid, p -aminobenzoic acid, thiamine (B 1 ), riboflavin (B 2 ), pyridoxine (B 6 ), and cobalamin (B 12 ) and 4 mg each of biotin and folic acid. The 25x 19-amino-acid solution contained (per liter) 3.125 g each of arginine and proline; 1.25 g each of aspartic acid, glutamine, and valine; 5.0 g each of glutamic acid and glycine; 2.5 g each of asparagine, histidine, isoleucine, leucine, lysine, and threonine; 1.875 g each of alanine, methionine, phenylalanine, serine, and tryptophan; and 0.3 g tyrosine. A solid medium was prepared by mixing an equal volume of liquid medium at a 2x concentration with 1% (wt/vol) Phytagel (Sigma) previously autoclaved to solubilize all chemicals, and both solutions were maintained at 95°C just prior to mixing in glass petri dishes. Aliquots of P. furiosus culture typically grown to mid-log phase (2 × 10 8 cells/ml) in defined liquid medium were mixed with DNA at a concentration of 2 to 10 ng DNA per μL of culture, spread in 30-μL aliquots onto defined solid medium. Plates were placed inverted in anaerobic jars and incubated at 90 °C for approximately 64 hours. Colonies were picked into 4 mL of defined medium in Hungate tubes and incubated anaerobically overnight at 90 °C. Genomic DNA, isolated using the Zymobead Genomic DNA Kit (Zymo Research), was used for PCR screening, which was carried out by using GXL polymerase (Takara, ClonTech). PCR screening was performed using a pair of primers outside the Mbh locus in order to confirm that the transformation cassette recombined into the correct locus.

Preparation of Cell Suspensions

COM1 and the ΔMbhABC strain were grown in the defined maltose medium in 1L culture bottles at 90 °C with shaking as described previously ( Lipscomb et al., 2011 ). Cells were harvested by centrifugation at 18,000 × g for 10 minutes in a Beckman-Coutler Avanti J-30i centrifuge. Cell suspensions were created by washing harvested cells with an anaerobic resuspension buffer containing 20 mM imidazole, 30 mM MgCl 2 ·6H 2 O, 0.5 M KCl, 2 mM cysteine-HCl, pH 6.5 and resuspending them in the same buffer at cell densities equivalent to OD 600 = 0.6. H 2 Production Assays H 2 production assays are modified from ( Lim et al., 2014 ). In brief, 2.0 mL of cell suspensions was added to rubber-sealed glass vials and the headspace was flushed with argon. Samples were incubated at 80 °C for 3 minutes and the reaction was initiated by the addition of the desired concentration of NaCl from an anaerobic 2 M stock solution. At various times, gas samples were removed by syringe and the amount of H 2 was determined using a 6850 Network Gas Chromatograph (Agilent Technologies). Protein concentrations were measured using the Bradford Protein Assay Dye.

Cryo-EM data acquisition

For cryo-EM analysis, 3 μl aliquots of the purified MBH complex at 2–3 mg/ml was applied to a glow-discharged holey carbon grids (C-Flat Cu CF-1.2/1.3, 400 mesh). The grids were blotted for 3–4 s at 10 °C with 95% humidity and flash-frozen in liquid ethane using an FEI Vitrobot IV. Cryo-EM data collection was performed on a 300 kV FEI Titan Krios electron microscopy with a K2 camera positioned post a GIF quantum energy filter. Automated data acquisition was performed with SerialEM ( Mastronarde, 2005 ) and FEI EPU package. Micrographs were recorded in super-resolution counting mode at a nominal magnification of 130,000×, resulting in a physical pixel size of 1.09 Å per pixel. Defocus values varied from 1.2 μm to 3 μm. The dose rate was 10.2 electron per pixel per second. Exposures of 6 s were dose-fractionated into 30 sub-frames, leading to a total accumulated dose of 51.7 electrons per Å 2 .

Image processing and 3D reconstruction

Two batches of data were collected. Dose fractionated movie frames were motion corrected (globally and locally), dose weighted and binned by 2 fold with MotionCor2 ( Grant and Grigorieff, 2015 ; Zheng et al., 2017 ), resulting in summed micrographs in a pixel size of 1.09 Å per pixel. Contrast transfer function (CTF) parameters for each micrograph were estimated by CTFFIND4 ( Rohou and Grigorieff, 2015 ). RELION-2.0 was used for further processing steps ( Kimanius et al., 2016 ). Bad micrographs revealed by manual inspection were excluded from further analysis, yielding 2804 and 2155 good micrographs for each dataset. For each dataset, a manually picked sets of particles were subject to 2D classification. This generated templates for reference-based particle picking, which yielded 674,607 and 574,955 automatically picked particles, respectively. Particle sorting and reference-free 2D classification was performed to remove contaminants and noisy particles, resulting in two datasets with 636,689 and 548,230 particles, respectively. Then 3D classification was performed using an ab initio map generated by EMAN2 ( Tang et al., 2007 ) as the initial reference model. For each dataset, one out of 4 classes with high-resolution features was obtained. The two identified good classes were combined as a new set of 301,300 particles, which was subjected to another round of 3D classification. The most populated 3D class (131,679 particles) was subsequently selected for the final 3D auto-refinement with a soft mask including the protein and detergent regions. This generated a map with an overall resolution of 3.7 Å. The resolution was estimated based on the gold-standard Fourier shell correlation 0.143 criterion ( Rosenthal and Henderson, 2003 ). The final map was corrected for the modulation transfer function (MTF) of the detector and sharpened by applying a negative B-factor, estimated by the post-processing procedure in RELION-2.0. Local resolution distribution was estimated using ResMap ( Kucukelbir et al., 2014 ). Atomic model building Most regions of the map, especially the transmembrane helices of membrane subunits, exhibit sufficient features for de novo model building of MBH, which started from the global assignment of its 14 subunits. Homology models for six subunits (hydrophilic MbhJ-MbhN and membrane MbhH and MbhM) were generated with the SWISS-MODEL server ( Arnold et al., 2006 ) using the structure of T. thermophilus Complex I as a template (PDB ID 4hea) ( Baradaran et al., 2013 ). They were fitted into the EM map as rigid bodies with CHIMERA ( Pettersen et al., 2004 ). The assignments of remaining 8 membrane subunits were assisted by their predicted secondary structural features and the excellent main chain connectivity of the map. MbhF, the only membrane subunit with 4 TMHs, was assigned first. Although they all contain 2 TMH, the specific structural features of MbhA, MbhE and MbhI helped to individually locate them to the map. MbhA TM2 is followed by a ferredoxin-like fold, which was predicted by I-TASSER ( Yang et al., 2015 ). For MbhI and MbhE, their TM1-TM2 linkers were predicted to be quite different: a long loop and a following long amphipathic helix for MbhI; a loop with a short helix in the middle for MbhE. Lastly, MbhB, MbhC, MbhD and MbhG all contain 3 transmembrane helices and they form 4 layers of three-helix bundle together. Among them, MbhC has the longest predicted helices and MbhG TM2-TM3 has the longest loop linker, which helped to locate them on the map. The positioning of last two subunits MbhB and MbhD was assisted by their sequence information. After the subunit assignment, for the six subunits MbhJ-MbhN, MbhH and MbhM, the fitted homology models were improved by manual adjustments and rebuilding using Coot ( Emsley et al., 2010 ). For each of the remaining 8 membrane subunits MbhA-MbhG and MbhI, a polyalanine model was first built with Coot and subsequent sequence assignment was mainly guided by bulky residues such as Arg, Tyr, Phe and Trp. In the final MBH model, 2470 of 2502 residues was assigned with side chains. MbhI loop Aa42-73 only allows the tracing of its main chain, which were built as polyalanine. The refinement of the MBH complex model against the cryo-EM map in real space was performed using the phenix.real_space_refine in PHENIX( Adams et al., 2010 ). The final model was assessed using MOLPROBITY ( Chen et al., 2010 ). All figures were prepared using PyMOL Schrödinger, LLC.) and CHIMERA. Statistics of the 3D reconstruction and model refinement were provided in Table S1 .

QUANTIFICATION AND STATISTICAL ANALYSIS

Resolution estimations of cryo-EM density maps are based on the 0.143 Fourier Shell Correlation (FSC) criterion ( Chen et al., 2013 ; Rosenthal and Henderson, 2003 ). DATA AND SOFTWARE AVAILABILITY Data Resources The accession number for the atomic coordinates reported in this paper is PDB: 6CFW. The accession number for the EM density maps reported in this paper is EMDB: EMD-7468.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Pyrococcus furiosus strain MW0414, COM1 and MW0574 were grown in defined maltose medium consisting of 1x base salts, 1x trace minerals, 1x vitamin solution, 2x 19-amino-acid solution, 10 μM sodium tungstate, 0.25 mg/ml resazurin, and 0.5% (wt/vol) maltose, with added cysteine at 0.5 g/liter, sodium sulfide at 0.5 g/liter, and sodium bicarbonate at 1 g/liter; and adjusted to pH 6.8 (Detailed buffer composition was described in the METHODS DETAILS ). For protein purification, large-scale growth of P. furiosus strain MW0414 was carried out in a 20-liter fermenter at 90 °C with constant flushing of 20% (v/v) CO 2 and 80% (v/v) N 2 for 14 hours, while the pH was maintained at 6.8 by the addition of 10% (w/v) sodium bicarbonate.

METHODS DETAILS Expression and purification of MBH The MBH holoenzyme (S-MBH) was solubilized and purified anaerobically from Pyrococcus furiosus strain MW0414, in which a His 9 -tag had been engineered at the N-terminus of the MbhJ subunit. The procedure was as previously described with some modifications ( McTernan et al., 2014 ). Frozen cells were lysed in 25 mM sodium phosphate, pH 7.5, containing 1 mM DTT and 50 μg/ml DNase I (5 ml per gram of frozen cells). After stirring for one hour, the cell-free extract was centrifuged at 100,000 × g for one hour. The supernatant was removed and the membranes were washed twice using 50 mM EPPS buffer, pH 8.0, containing 5 mM MgCl 2 , 50 mM NaCl, 10% (v/v) glycerol, 1 mM DTT and 0.1 mM PMSF. The membrane pellet was collected by ultracentrifugation at 100,000 × g for one hour after each wash step. The washed membranes were resuspended in 50 mM Tris-HCl, pH 8.0, containing 5 mM MgCl 2 , 50 mM NaCl, 5% (v/v) glycerol, 1 mM DTT, and 0.1 mM PMSF. MBH was solubilized by adding n-dodecyl-β-D-maltoside (DDM, Inalco) to 3% (w/v) followed by incubation at 4 °C for 16 hours. The solubilized membranes were centrifuged at 100,000 × g for 1 hour. The supernatant was applied to a 5-ml His-Trap crude FF Ni-NTA column (GE Healthcare) while diluting it 10-fold with buffer A (25 mM sodium phosphate, 300 mM NaCl, pH 7.5, containing 1 mM DTT and 0.03 % DDM). The column was washed with 10 column volumes of buffer A and the bound protein was eluted with a 20-column volume gradient from 0 to 100 % buffer B (buffer A containing 500 mM imidazole). The eluted protein was further purified by applying it to a 1-mL His-Trap HP Ni-NTA column (GE Healthcare) while diluting it 5-fold with buffer A. A 30-column volume gradient from 0 to 100 % buffer B was used to elute the bound protein. The MBH sample was concentrated and further purified using a Superose 6 10/300 GL column (GE Healthcare) equilibrated with 50 mM Tris-HCl, pH 8.2, containing 300 mM NaCl, 2 mM sodium dithionite, and 0.03 % DDM. Deletion of MbhABC The genetically tractable P. furiosus strain COM1 was used to delete the three genes PF1423-1425. 500 bp flanking regions were amplified from P. furiosus genomic DNA for the UFR and DFR, and the selection marker ( pyrF- P gdh ) was amplified by using pGL021 as the template ( Lipscomb et al., 2011 ). The knock-in cassette was assembled using overlapping PCR ( Bryksin and Matsumura, 2010 ). The genomic DNA was prepared using Zymobead Genomic DNA Kit (Zymo Research). P. furiosus transformants were grown in defined maltose media as previously described ( Lipscomb et al., 2011 ). The maltose medium was composed of 1x base salts, salts, 1x trace minerals, 1x vitamin solution, 2x 19-amino-acid solution, 0.5% (wt/vol) maltose, 10 μM sodium tungstate, and 0.25 mg/ml resazurin, with added cysteine at 0.5 g/liter, sodium sulfide at 0.5 g/liter, sodium bicarbonate at 1 g/liter, and 1 mM sodium phosphate buffer (pH 6.8). The 5x base salts stock solution contained (per liter): 140 g of NaCl, 17.5 g of MgSO 4 ·7H 2 O, 13.5 g of MgCl 2 ·6H 2 O, 1.65 g of KCl, 1.25 g of NH 4 Cl, and 0.70 g of CaCl 2 ·2H 2 O. The 1000x trace mineral stock solution contained (per liter) 1 ml of HCl (concentrated), 0.5 g of Na 4 EDTA, 2.0 g of FeCl 3 , 0.05 g of H 3 BO 3 , 0.05 g of ZnCl 2 , 0.03 g of CuCl 2 ·2H 2 O, 0.05 g of MnCl 2 ·4H 2 O 0.05 g of (NH 4 ) 2 MoO 4 , 0.05 g of AlK(SO 4 )·2H 2 O, 0.05 g of CoCl 2 ·6H 2 O, and 0.05 g of NiCl 2 ·6H 2 O. The 200x vitamin stock solution contained (per liter) 10 mg each of niacin, pantothenate, lipoic acid, p -aminobenzoic acid, thiamine (B 1 ), riboflavin (B 2 ), pyridoxine (B 6 ), and cobalamin (B 12 ) and 4 mg each of biotin and folic acid. The 25x 19-amino-acid solution contained (per liter) 3.125 g each of arginine and proline; 1.25 g each of aspartic acid, glutamine, and valine; 5.0 g each of glutamic acid and glycine; 2.5 g each of asparagine, histidine, isoleucine, leucine, lysine, and threonine; 1.875 g each of alanine, methionine, phenylalanine, serine, and tryptophan; and 0.3 g tyrosine. A solid medium was prepared by mixing an equal volume of liquid medium at a 2x concentration with 1% (wt/vol) Phytagel (Sigma) previously autoclaved to solubilize all chemicals, and both solutions were maintained at 95°C just prior to mixing in glass petri dishes. Aliquots of P. furiosus culture typically grown to mid-log phase (2 × 10 8 cells/ml) in defined liquid medium were mixed with DNA at a concentration of 2 to 10 ng DNA per μL of culture, spread in 30-μL aliquots onto defined solid medium. Plates were placed inverted in anaerobic jars and incubated at 90 °C for approximately 64 hours. Colonies were picked into 4 mL of defined medium in Hungate tubes and incubated anaerobically overnight at 90 °C. Genomic DNA, isolated using the Zymobead Genomic DNA Kit (Zymo Research), was used for PCR screening, which was carried out by using GXL polymerase (Takara, ClonTech). PCR screening was performed using a pair of primers outside the Mbh locus in order to confirm that the transformation cassette recombined into the correct locus.

Preparation of Cell Suspensions

COM1 and the ΔMbhABC strain were grown in the defined maltose medium in 1L culture bottles at 90 °C with shaking as described previously ( Lipscomb et al., 2011 ). Cells were harvested by centrifugation at 18,000 × g for 10 minutes in a Beckman-Coutler Avanti J-30i centrifuge. Cell suspensions were created by washing harvested cells with an anaerobic resuspension buffer containing 20 mM imidazole, 30 mM MgCl 2 ·6H 2 O, 0.5 M KCl, 2 mM cysteine-HCl, pH 6.5 and resuspending them in the same buffer at cell densities equivalent to OD 600 = 0.6. H 2 Production Assays H 2 production assays are modified from ( Lim et al., 2014 ). In brief, 2.0 mL of cell suspensions was added to rubber-sealed glass vials and the headspace was flushed with argon. Samples were incubated at 80 °C for 3 minutes and the reaction was initiated by the addition of the desired concentration of NaCl from an anaerobic 2 M stock solution. At various times, gas samples were removed by syringe and the amount of H 2 was determined using a 6850 Network Gas Chromatograph (Agilent Technologies). Protein concentrations were measured using the Bradford Protein Assay Dye.

Cryo-EM data acquisition

For cryo-EM analysis, 3 μl aliquots of the purified MBH complex at 2–3 mg/ml was applied to a glow-discharged holey carbon grids (C-Flat Cu CF-1.2/1.3, 400 mesh). The grids were blotted for 3–4 s at 10 °C with 95% humidity and flash-frozen in liquid ethane using an FEI Vitrobot IV. Cryo-EM data collection was performed on a 300 kV FEI Titan Krios electron microscopy with a K2 camera positioned post a GIF quantum energy filter. Automated data acquisition was performed with SerialEM ( Mastronarde, 2005 ) and FEI EPU package. Micrographs were recorded in super-resolution counting mode at a nominal magnification of 130,000×, resulting in a physical pixel size of 1.09 Å per pixel. Defocus values varied from 1.2 μm to 3 μm. The dose rate was 10.2 electron per pixel per second. Exposures of 6 s were dose-fractionated into 30 sub-frames, leading to a total accumulated dose of 51.7 electrons per Å 2 .

Image processing and 3D reconstruction

Two batches of data were collected. Dose fractionated movie frames were motion corrected (globally and locally), dose weighted and binned by 2 fold with MotionCor2 ( Grant and Grigorieff, 2015 ; Zheng et al., 2017 ), resulting in summed micrographs in a pixel size of 1.09 Å per pixel. Contrast transfer function (CTF) parameters for each micrograph were estimated by CTFFIND4 ( Rohou and Grigorieff, 2015 ). RELION-2.0 was used for further processing steps ( Kimanius et al., 2016 ). Bad micrographs revealed by manual inspection were excluded from further analysis, yielding 2804 and 2155 good micrographs for each dataset. For each dataset, a manually picked sets of particles were subject to 2D classification. This generated templates for reference-based particle picking, which yielded 674,607 and 574,955 automatically picked particles, respectively. Particle sorting and reference-free 2D classification was performed to remove contaminants and noisy particles, resulting in two datasets with 636,689 and 548,230 particles, respectively. Then 3D classification was performed using an ab initio map generated by EMAN2 ( Tang et al., 2007 ) as the initial reference model. For each dataset, one out of 4 classes with high-resolution features was obtained. The two identified good classes were combined as a new set of 301,300 particles, which was subjected to another round of 3D classification. The most populated 3D class (131,679 particles) was subsequently selected for the final 3D auto-refinement with a soft mask including the protein and detergent regions. This generated a map with an overall resolution of 3.7 Å. The resolution was estimated based on the gold-standard Fourier shell correlation 0.143 criterion ( Rosenthal and Henderson, 2003 ). The final map was corrected for the modulation transfer function (MTF) of the detector and sharpened by applying a negative B-factor, estimated by the post-processing procedure in RELION-2.0. Local resolution distribution was estimated using ResMap ( Kucukelbir et al., 2014 ). Atomic model building Most regions of the map, especially the transmembrane helices of membrane subunits, exhibit sufficient features for de novo model building of MBH, which started from the global assignment of its 14 subunits. Homology models for six subunits (hydrophilic MbhJ-MbhN and membrane MbhH and MbhM) were generated with the SWISS-MODEL server ( Arnold et al., 2006 ) using the structure of T. thermophilus Complex I as a template (PDB ID 4hea) ( Baradaran et al., 2013 ). They were fitted into the EM map as rigid bodies with CHIMERA ( Pettersen et al., 2004 ). The assignments of remaining 8 membrane subunits were assisted by their predicted secondary structural features and the excellent main chain connectivity of the map. MbhF, the only membrane subunit with 4 TMHs, was assigned first. Although they all contain 2 TMH, the specific structural features of MbhA, MbhE and MbhI helped to individually locate them to the map. MbhA TM2 is followed by a ferredoxin-like fold, which was predicted by I-TASSER ( Yang et al., 2015 ). For MbhI and MbhE, their TM1-TM2 linkers were predicted to be quite different: a long loop and a following long amphipathic helix for MbhI; a loop with a short helix in the middle for MbhE. Lastly, MbhB, MbhC, MbhD and MbhG all contain 3 transmembrane helices and they form 4 layers of three-helix bundle together. Among them, MbhC has the longest predicted helices and MbhG TM2-TM3 has the longest loop linker, which helped to locate them on the map. The positioning of last two subunits MbhB and MbhD was assisted by their sequence information. After the subunit assignment, for the six subunits MbhJ-MbhN, MbhH and MbhM, the fitted homology models were improved by manual adjustments and rebuilding using Coot ( Emsley et al., 2010 ). For each of the remaining 8 membrane subunits MbhA-MbhG and MbhI, a polyalanine model was first built with Coot and subsequent sequence assignment was mainly guided by bulky residues such as Arg, Tyr, Phe and Trp. In the final MBH model, 2470 of 2502 residues was assigned with side chains. MbhI loop Aa42-73 only allows the tracing of its main chain, which were built as polyalanine. The refinement of the MBH complex model against the cryo-EM map in real space was performed using the phenix.real_space_refine in PHENIX( Adams et al., 2010 ). The final model was assessed using MOLPROBITY ( Chen et al., 2010 ). All figures were prepared using PyMOL Schrödinger, LLC.) and CHIMERA. Statistics of the 3D reconstruction and model refinement were provided in Table S1 .

Supplementary Material 1 Movie S1. Surface-rendered cryo-EM 3D map of the MBH complex segmented according to the 14 individual subunits, and are colored as in Figure 1C . 10 2 Movie S2. Overall structure of 14-subunit MBH complex shown in cartoon. Individual subunits are colored as in Figure 1D . 3 Figure S1. Biochemical characterization of Pyrococcus furiosus MBH complex, Related to the STAR Methods and Figure 1 A–B, Na + -dependent H 2 production activity of P. furiosus. Effect of NaCl concentration (0, 50 or 150 mM) on the H 2 production activity of cell suspensions of the parent (COM1, panel A ) or Δ mbhABC ( B ) strains. Data are represented as mean ± SEM. C–E, Preparation of MBH for single-particle cryo-EM analysis. C , The predicted molecular weight (kDa) of each of the fourteen MBH subunits. D , A representative profile of size-exclusion chromatography (Superose 6 10/300 GL column) of the purified MBH complex solubilized in the detergent n-Dodecyl β-D-maltoside. E , SDS-PAGE of the pooled gel filtration peak fractions shows the presence of all fourteen MBH subunits as labeled to the right. Molecular weight markers in kDa are labeled to the left. 4 Figure S2.

Cryo-EM analysis of the archaeon Pyrococcus furiosus

MBH complex, Related to Figure 1 A, Representative cryo-EM micrograph of the MBH particles. B , Representative 2D class averages of the MBH particles. C , The workflow of cryo-EM data processing. Two datasets, each contains 2804 and 2155 micrographs, respectively, were collected and processed separately during 2D classification and the first round of 3D classification. The best 3D class from these two datasets were combined for the second round of 3D classification. Then the most populated 3D class (131,679 particles, ~44% of total particles) was subjected to the final refinement, which resulted in a 3D density map with an estimated resolution of 3.7 Å. See Methods for more details. D–F, Statistics of the cryo-EM 3D map. D, Angular distribution of all particles used in the final 3D reconstruction (top panel) is shown with the corresponding view of the 3D map (bottom panel). The number of particles in each orientation is indicated by bar length and color (blue, low; red, high). E, The gold-standard Fourier shell correlation (FSC) curve of the 3D map. F, Local resolution of the 3D density map, calculated with ResMap. The range of resolution is color-coded from the higher resolution blue (3 Å) to the lower resolution red (5 Å). 5 Figure S3. A gallery of EM density maps of ten membrane subunits (panels A–J) and four cytoplasmic subunits (panels K–N), superimposed with their corresponding atomic models in cartoon and stick views, Related to Figure 1 A, MbhA. B, MbhB. C, MbhC. D, MbhD. E, MbhE. F, MbhF. G, MbhG. H, MbhH. I, MbhI. Densities for Aa42-73 were weak and only allowed the tracing of their main-chain atoms. J , MbhM. K, MbhJ. L, MbhK. M, MbhL. N, MbhN. 6 Figure S4. Comparison of membrane arms between MBH and Complex I, Related to Figure 3 A–D , MBH complex ( A, C ) and Complex I ( B, D ; T. thermophilus ; PDB ID 4hea) were aligned by the membrane subunits immediately below their respective peripheral arms: Nqo8 of Complex I vs MbhM of MBH. Side views of this comparison are shown separately in panels A (MBH) and B (Complex I). The top views are shown separately in panels C (MBH) and D (Complex I), respectively. Note the large gap between subunits M and H in MBH ( A ). There are four elongated densities located to the lower region of the gap ( A inset; marked by blue dashed lines), which stack against several hydrophobic regions of subunits M and H. These densities are likely from two phospholipid molecules that may stabilize the structure and prevent ion leakage across the membrane bilayer. The dashed curves in panels C and D highlight the fact that the chain of hydrophilic residues found in Complex I is continuous ( D ), but is discontinuous in MBH ( C ). E, Superimposition of the MBH complex and T. thermophilus Complex I (PDB ID 4hea), aligned based solely on their respective antiporter-like subunit MbhH and Nqo14 (as Figure 3C ). By this alignment, the peripheral arm is docked to the right end of the membrane arm of Complex I and to the left end of the membrane arm of the MBH. F–I, Structural alignment shown in ( E ) revealed a module shared between the MBH complex and Complex I. The corresponding subunits are superimposed and shown in panels F–I: F , MbhH and Nqo14. G , MbhG and Nqo11. H , MbhD and MbhE together are equivalent to Nqo10. Notably, MbhD TMH3 contains a -bulge that is also present in Nqo10. I , Similarity between the MbhI C-terminal region (lateral helix HL and TMH2) with the Nqo12 C-terminal region (lateral helix HL and TMH16). 7 Figure S5. A comparison of the membrane-anchored hydrogenase module between the MBH complex and Complex I, Related to Figure 4 A, Overlay of MBH peripheral arm with the classic dimeric [NiFe] hydrogenase from D. gigas (PDB ID 2frv) by aligning MbhL with the large subunit of the two-subunit classic hydrogenase. Only MbhL and MbhJ are visible here. The membrane subunit MbhM of MBH is also shown, although it is absent in the dimeric [NiFe] hydrogenase. B, Overlay of MBH peripheral arm plus the membrane Mbh M and I with the corresponding T. thermophilus Complex I subunits – the Q-module and the membrane Nqo8, Nqo7. The alignment is based on MbhL and Nqo4. The two systems share a similar architecture except for the C-terminal regions of MbhI and Nqo7, suggesting that this sub-complex is well conserved between the MBH complex and Complex I. C, Like in Complex I, there is also a chamber at the interface between the peripheral arm and the membrane subunit MbhM in the MBH complex. The internal cavity is shown as a red surface. The entry to the chamber in the MBH complex, which is equivalent to the quinone entry site in Complex I, is closed due to the presence of several bulky residues there. Structural alignment was based on MbhM (MBH) and Nqo8 (Complex I). D–F, A comparison of coordinations of the [4Fe-4S] and [Ni-Fe] clusters in the peripheral arm of the MBH complex with those in D. gigas hydrogenase and Complex I. The structures are colored as in panels A and B . D, The coordination of the distal and medial [4Fe-4S] clusters in MbhN (MBH) is highly similar to that in Nqo9 (Complex I). E, The coordination of the proximal [4Fe-4S] cluster is similar between the MBH and D. gigas hydrogenase. In Complex I, coordination of the N2 cluster involves an unusual pair of tandem Cys residues (C45 and C46). F, The side chains coordinating the [NiFe] cluster in MBH are similar to those in the D. gigas hydrogenase. The structural elements for coordinating [Ni-Fe] cluster are not present in Complex I. 8 Figure S6. Structure and sequence comparison of the peripheral arms of the MBH complex with Complex I revealed several insertions evolved in Complex I for anchoring its N-module, Related to Figure 4 A, Overlay of MBH peripheral arm and the membrane MbhM with the T. thermophilus Complex I N-module, Q-module, and the membrane Nqo8. Alignment was based on MbhL and Nqo4. Individual subunits are coloured as in Figure S5B . The N-module of Complex I, shown in cartoon as well as in transparent surface view, is peripheral and evolved later. B–E, Structural comparisons (top panel) and sequence alignments (bottom panel) of the four N-module-interacting subunits of Complex I with their corresponding MBH subunits : B , Nqo4 and MbhL; C , Nqo9 and MbhN; D , Nqo6 and MbhJ; E , Nqo5 and MbhK. Protein sequence sources are Pfu, Pyrococcus furiosus; Tth, Thermus thermophilus; Ova, Ovis aries . Major structural differences are marked by gray ovals, and further highlighted by the gray square(s) in the sequence alignment shown below each panel. 9 Figure S7. Sequence alignments of individual MBH membrane subunits with their counterparts in the Mrp H + /Na + antiporter, Related to Figure 5 and 6 and Table 1 The TMH numbers shown above the primary sequences are based on the MBH structure. The predicted secondary structural elements of Bacillus subtilis Mrp subunits are shown below the sequences. Protein sequence sources are Pfu, Pyrococcus furiosus ; Bsu, Bacillus subtilis ; Bps, Bacillus pseudofirmus ; Sau, Staphylococcus aureus ; Rme, Rhizobium meliloti ; Pae, Pseudomonas aeruginosa ; Vch, Vibrio cholera . The filled black triangles mark residues shown in Fig. 3E , and the filled black squares mark residues shown in Fig. 5B–C .