Abstract
While early genetic and low resolution structural observations suggested that extracellular conductive filaments on metal reducing organisms such as Geobacter were composed of Type IV pili, it has now been established that bacterial c-type cytochromes can polymerize to form extracellular filaments capable of long-range electron transport. Atomic structures exist for two such cytochrome filaments, formed from the hexaheme cytochrome OmcS and the tetraheme cytochrome OmcE. Due to the highly conserved heme packing within the central OmcS and OmcE cores, and shared pattern of heme coordination between subunits, it has been suggested that these polymers have a common origin. We have now used cryo-EM to determine the structure of a third extracellular filament, formed from the Geobacter sulfurreducens octaheme cytochrome, OmcZ. In contrast to the linear heme chains in OmcS and OmcE, the packing of hemes, heme:heme angles, and between-subunit heme coordination is quite different in OmcZ. A branched heme arrangement within OmcZ leads to a highly surface exposed heme in every subunit, which may account for the formation of conductive biofilm networks, and explain the higher measured conductivity of OmcZ filaments. This new structural evidence suggests that conductive cytochrome polymers arose independently on more than one occasion from different ancestral multiheme proteins.
Introduction
While soluble electron acceptors easily diffuse to the cytoplasmic membrane to support most microbial respirations, bacteria must build conductive pathways out of the cell to respire using insoluble metals and conductive surfaces. This process, known as extracellular electron transfer, enables global Fe(III) and Mn(IV) reduction (Gralnick and Newman, 2007; Lovley et al., 2004), methane production in anaerobic digestors (Morita et al., 2011; Shrestha et al., 2013), methane oxidation in ocean seeps (Chadwick et al., 2022), corrosion (Tang et al., 2019), and electricity generation in microbial electrochemical devices (Wang and Ren, 2013).
When model organisms from the genus Geobacter utilize conductive surfaces as electron acceptors, they establish conductive multicellular biofilms (Chadwick et al., 2019; Yates et al., 2015). Many components, including pili and polysaccharides, are essential for formation of these biofilms (Reguera et al., 2005; Reguera et al., 2007; Rollefson et al., 2011), but the octaheme c-type cytochrome OmcZ is the only cytochrome out of over 80 multiheme proteins in the G. sulfurreducens genome necessary for long-distance conductivity (Nevin et al., 2009; Peng and Zhang, 2017; Richter et al., 2009). Deletion of omcZ affects anodic electron transfer to electrodes (Nevin et al., 2009; Peng and Zhang, 2017; Richter et al., 2009), especially at low redox potential, and also slows cathodic corrosion of Fe0 (Tang et al., 2019). In contrast, OmcZ is not needed for reduction of other extracellular metals, including Fe(III) and Mn(IV) oxides (Aklujkar et al., 2013).
The omcZ gene (GSU2076) is up-regulated 400-800% in conductive electrode biofilms (Franks et al., 2012; Nevin et al., 2009), producing a 50 kDa periplasmic preprotein that is processed by the OzpA protease (GSU2075) to a 30 kDa octoheme form (Inoue et al., 2010; Jimenez Otero et al., 2018; Kai et al., 2021). Similar operons containing homologus omcZ and ozpA are widespread throughout the Bacterial and Archaeal domains (Chadwick et al., 2022). In G. sulfurreducens, large heat-stable OmcZ molecules with a wide redox potential window (420 to −60 mV vs. the standard hydrogen electrode (SHE)) are abundant in the extracellular matrix between cells (Inoue et al., 2011; Qian et al., 2011), which under AFM imaging appear as ∼2.5 nm diameter conductive filaments (Yalcin et al., 2020).
OmcZ represents one of three multiheme cytochrome filaments identified to date, along with the proteins OmcS (Wang et al., 2019; Yalcin and Malvankar, 2020) and OmcE (Wang et al., 2022b). Cryo-EM structures of G. sulfurreducens OmcS and OmcE, which are involved in electron transfer to insoluble Fe(III) and Mn(IV), reveal a core of closely spaced c-type hemes, with the unique characteristic of histidine residues from the previous subunit coordinating a heme from the next. While the OmcS and OmcE cytochromes share no amino acid sequence or structural similarity, and show different patterns of surface glycosylation, the heme molecules in these two nanowires can be superimposed, suggesting a shared evolutionary path (Wang et al., 2022b). In conductive-probe AFM measurements of enriched filaments cast on gold surfaces, OmcZ can demonstrate up to 1000-fold higher conductivity than preparations primarily consisting of OmcS (Yalcin et al., 2020), indicating that internal packing or external accessibility of hemes may be uniquely suited for interfacing OmcZ with electrodes.
In this work, we describe the cryo-EM structure of G. sulfurreducens OmcZ, and show it forms a multiheme cytochrome nanowire different from OmcS and OmcE in both protein fold and heme arrangement. The core of closely spaced hemes shares no similarity with previously reported nanowires, two heme pairs are oriented at unique angles from one another, and one heme diverges from the central chain to create a solvent-exposed site along the wire. OmcZ also lacks the inter-subunit coordination shared by OmcS and OmcE, and has no evidence of glycosylation. These data show that OmcZ is a member of a new and widespread multiheme cytochrome nanowire family that arose independently from OmcS and OmcE.
Results
Cryo-EM of the OmcZ filament
To obtain OmcZ filaments, we grew a ΔomcS strain of G. sulfurreducens on graphite electrodes poised at +0.24 vs. SHE. Filaments were enriched via shearing, DNAse treatment, and salt precipitation, similar to protocols used for OmcS and OmcE, except that higher pH buffers were used. Using cryo-EM (Fig. 1A), we determined the structure of these filaments. Unlike OmcS filaments (Wang et al., 2019), OmcZ filaments did not show as strong a sinusoidal morphology (Fig. 1B-C). An averaged power spectrum from raw segments (Fig. S1A) showed a meridional layer-line at ∼1/(58 Å), corresponding to the rise per subunit in the filament, and another layer-line near the equator at ∼1/(132 Å) corresponding to the 132 Å pitch of a 1-start helix. After refining those parameters in the helical and subsequent non-uniform refinement, a ∼4.2 Å resolution reconstruction was obtained, judged by a map:map Fourier shell correlation (FSC, Fig. S1B). The model:map FSC gives a similar resolution estimation (Table 1). OmcZ filaments are ∼50 Å at their widest point and coordinate eight heme molecules per cytochrome subunit.
Building an atomic model de novo at such resolution is typically challenging, especially when protein secondary structure elements are sparse (Wang et al., 2022a). Fortunately, highly accurate protein structure prediction is now possible with AlphaFold2, even when the protein of interest has minimal similarity with any known protein structure (Jumper et al., 2021). The AlphaFold2 predicted full-length OmcZ model had a signal peptide, a N-terminal domain with eight pairs of histidines coordinating hemes that reasonably matched the cryo-EM map, and a C-terminal domain containing two β-sandwiches (Fig. S1C). After model building, residues P27 to S284 could be fit into the map, with the signal peptide (residues 1-26) and the C-terminal β-sandwich domain missing. It has been previously shown that OzpA, a subtilisin-like serine protease, cleaves the C-terminal part of OmcZ (Kai et al., 2021). While prior data detected amino acids consistent with residues 280-282 (FGN) at the C-terminus of purified OmcZ, the last visible C-terminal residues in the cryo-EM map extended to residue 284 (FGNSS) suggesting this could be the cleavage site. The next serine towards the C-terminus is S298, in the middle of a β-sandwich in the predicted model. The only other protein in G. sulfurreducens predicted to have a similar structure is GSU1334. However, OmcZ was the only 30 kDa cytochrome detected by mass spectrometry of SDS-PAGE separated filament preparations, consistent with transcriptional analysis showing OmcZ is induced to levels over 75-fold higher than GSU1334 during electrode growth (∼50 RPKM for GSU1334 vs. 4200 RPKM for omcZ) (Jimenez Otero et al., 2018). An OmcZ model generated from homology modeling has been previously reported (Yalcin et al., 2020). However, neither its protein fold nor heme arrangements bear any resemblance to our experimentally determined model.
For all eight hemes per OmcZ subunit, two histidines axially coordinate iron at the center of each heme, and the vinyl groups of each heme form covalent thioether bonds with cysteines. Most heme-heme arrangements in OmcZ are either T-shaped or anti-parallel (Fig. 1D). However, heme 6 (Fig. 1D), in each OmcZ subunit does not fit into the closely packed central linear heme chain. The edge-to-edge distances between adjacent porphyrin rings in the main heme chain are between 3.6 to 5.7 Å, comparable to those in OmcS and OmcE filaments. All hemes in OmcZ are thus close to two other hemes, with the exception of heme 6 which is only close to one other heme, heme 5, at an edge-to-edge distance of 4.6 Å.
OmcZ structurally differs from OmcS and OmcE
OmcZ filaments represent the third experimentally determined atomic structure of an extracellular cytochrome filament. Our first question was whether the protein fold or heme arrangement had been seen in other proteins. We have shown that OmcS and OmcE, while lacking sequence or structural similarity, share a conserved heme arrangement (Wang et al., 2022b). In contrast, OmcZ shares no similarity to OmcS or OmcE in sequence, protein fold, or heme arrangement (Fig. 2A-B). When we used the DALI (Holm, 2020) and Foldseek servers (van Kempen et al., 2022) to find structures with a similar fold to OmcZ, both servers returned no hits, suggesting the fold of OmcZ has not been previously observed. This may be due to the low percentage of secondary structure in OmcZ (18.2% helices and 4.7% β-strands), as the DALI and Foldseek servers also returned no hits for OmcS and OmcE, except for themselves and their homologs sharing extensive sequence similarity.
In the conserved heme packing of OmcS and OmcE (Fig. 2B), the heme-heme interactions are periodic, with pairs of anti-parallel and T-shape repeating units. In OmcZ, one of eight hemes per asymmetrical subunit is located outside of the chain, so the minimum periodic pattern requires a unique doublet of anti-parallel hemes (anti-parallel, T-shape, anti-parallel, anti-parallel, T-shape, anti-parallel, T-shape). We previously analyzed preferred heme-heme orientations by looking at all 800+ heme c (HEC) containing proteins available in the Protein Data Bank (Wang et al., 2022b). For comparison with OmcZ, we further improved the analysis by including other types of heme molecules (HEM, which is a non-heme c, in addition to HEC) and only including the atoms in the porphyrin ring for the analysis. Six out of eight heme pairs in OmcZ fall into the preferred clusters previously detected (Fig. 2C). Interestingly, two heme pairs, heme 4-heme 5 and heme 5-heme 6 (Fig. 1D), have a rare rotation angle of 56° and 82°, respectively. Heme pairs with a rotation angle between 50-90° normally have a 6 Å or larger edge-to-edge distances between porphyrin rings (Fig. 2C). This is not the case in OmcZ: both of their edge-to-edge distances are smaller than 5 Å, suggesting efficient electron hopping could occur.
The protein-protein interface in OmcZ filaments is also quite different from OmcS and OmcE. In the latter cases, a heme molecule is coordinated by a histidine coming from an adjacent subunit, providing extra strength to the interface. In OmcZ, all eight hemes are coordinated by histidines within the same subunit. The interfacial buried area in OmcZ is ∼1,200 Å2, comparable to the interface in OmcE (∼1,100 Å2) but much smaller than the interface in OmcS (∼1,900 Å2). Similar to OmcS and OmcE, an anti-parallel heme pair is observed at the interface (Fig. 2D). Also unique to OmcZ is the motif binding heme 1, which lies at N-terminal interface. Instead of the canonical CXXCH motif, additional amino acids form a 14-residue loop between C77 and C92 (rather than the expected pattern of C89, C92, and H93). This additional loop is involved in the subunit-subunit interface (Fig. 2D).
OmcZ filaments possess solvent accessible hemes
In OmcZ, heme 6 is largely solvent-exposed in every subunit in the filament (Fig. 3A). In contrast, the only solvent-exposed hemes in the OmcE and OmcS polymers are those at the two ends of a filament, similar to the chain inside the 10 heme MtrAB membrane-spanning complex (Edwards et al., 2020), where hemes are only solvent-exposed hemes at the ends of a linear “molecular wire” (Fig. 3B). The exposure of OmcZ was more analogous to the MtrC portion of the MtrABC complex, which has a branched heme chain introducing additional solvent-exposed hemes at the sides (Fig. 3B,C). For comparison, the surface exposed area of heme 6 in OmcZ is 326 Å2, the surface exposed area of the largely buried heme 5 is 12 Å2, while the most exposed heme in MtrABC (heme 901) has a surface exposed area of 292 Å2.
Discussion
The observation that the Geobacter sulfurreducens cytochromes OmcE and OmcS share the same highly conserved arrangement of hemes, even though the proteins themselves have no apparent sequence or structural similarity, was unexpected (Wang et al., 2019). These two proteins lack substantial amounts of secondary structure and mainly consist of loops and coils. The overall rmsd between OmcE and OmcS is 19 Å, which is what one might expect for two completely unrelated structures, but 28 atom pairs from these two proteins can be aligned with a rmsd of 1.1 Å. These 28 pairs represent the CxxC heme-binding motifs, proximal histidines, and distal histidines that coordinate the hemes in both proteins. The four hemes in OmcE can therefore be superimposed almost perfectly on the first four of the six hemes in OmcS. This suggests that there is strong selective pressure on residues coordinating hemes in both proteins, but almost no selective pressure on the intervening residues that form loops and coils, allowing these two proteins to diverge from a common ancestor until the sequences as well as overall folds retain no recognizable similarity.
The structure of OmcZ reveals a filament that does not share the heme packing found in OmcS and OmcE. Further, each of the eight hemes in an OmcZ subunit are coordinated axially by histidines from the same subunit, while in both OmcS and OmcE there is coordination of hemes in an adjacent subunit by a histidine in the neighboring subunit. This strongly suggests that polymers of multi-heme cytochromes have arisen independently at least twice in bacterial evolution. Given the many known multiheme packing motifs (Soares et al., 2022), it is likely that other conductive cytochrome polymers exist.
Electron transfer between adjacent hemes is slowed by increased distance and solvent exposure, and is more rapid between hemes in the T-shaped compared to parallel configuration (Blumberger, 2015; Jiang et al., 2017; Jiang et al., 2020; van Wonderen et al., 2019). In recent calculations, the solvent exposure of MtrC hemes combined with rate-limiting interfacial steps (up to 8 Å porphyrin-porphyrin spacing between hemes in MtrA and MtrC) was hypothesized to cause slower electron transfer through MtrCAB compared to OmcS. While the spacing of porphyrin rings in OmcZ (3.6 – 5.7 Å) and OmcS/OmcE (3.4 – 6.1 Å and 3.8 – 6.0 Å, respectively) are similar, the increased solvent exposure of OmcZ and additional hemes in parallel configuration predict OmcZ could have a conductivity lower than OmcS. However, enriched OmcZ filaments are reported to have 1,000 fold higher conductivity versus OmcS (Yalcin et al., 2020).
What could explain the higher apparent conductivity of OmcZ filaments (Yalcin et al., 2020)? First, measured conduction values are typically expressed as an intrinsic conductivity, which involves an estimate of the cross-sectional area. This likely distorts true values, as the heme chains provide the conductivity, while the surrounding protein acts largely as an insulator (Edwards et al., 2020). For example, a 1 mm diameter copper wire with 0.5 mm vs. 2 mm thick insulation would have similar measured conductivity, but if scaled by cross-sectional area, the less insulated wire will appear (5/2)2 or 6.25 times more conductive. Similar errors would occur when comparing the larger diameter MtrABC complex or OmcS filament to the much thinner OmcZ. A second and likely greater difference could be introduced by experimental measurements on conductive surfaces, where multiple exposed hemes of OmcZ could make very close contact with the electrode surface. In contrast, the exposed terminal heme in an OmcS or OmcE filaments may never be closer than ∼ 10 Å from the surface of an electrode or AFM probe, and due to post-translational modifications, OmcS and OmcE wires may be further insulated from conductivity measurements (Fig. 4).
The higher measured conductivity of OmcZ was proposed to be due to increased crystallinity and reduced disorder within OmcZ filaments compared to OmcS (Yalcin et al., 2020). However, the cryo-EM analysis reveals OmcZ filaments to be more flexible and less ordered than OmcS filaments. The earlier conclusion about crystallinity was based upon an increase in X-ray scattering from unoriented samples at ∼ 1/(3.6 Å), which was interpreted to arise from parallel stacking of hemes. Such conclusions appear formally similar to an argument made previously (Malvankar et al., 2015) that the same increase in diffraction at ∼ 1/(3.2 Å) in unoriented and dried samples was diagnostic of aromatic amino acid residue stacking in a hypothetical PilA-N structure.
Along with X-ray scattering, prior characterization of filaments assumed to be OmcZ (Yalcin et al., 2020) was based in large part upon scanning IR nanospectroscopy, which found that at pH 7 putative OmcZ filaments contained ∼ 41% of residues in β-sheets and ∼ 39% of residues in α-helices, while at pH 2 they contained ∼ 53% in β-sheets and ∼ 20% in α-helices and. In contrast, in the cryo-EM structure of filaments prepared at pH 10.5, OmcZ contained only ∼ 18% helices and 5% β-sheets. These data suggest that IR nanospectroscopy is not yet a reliable method for determining filament identity or composition, just as measures of diameter, which have dominated the field of extracellular nanowires, are also unreliable (Wang et al., 2022b).
With the availability of these new cytochrome filament structures, it is tempting to correlate their differences with phenotypes linked to each wire. The OmcE and OmcS extracellular wires of Geobacter are compatible with a model where filaments link a cell to a nearby metal oxide particle (Fig. 3D). Post-translational modifications, especially the extensive glycosylations of OmcE, would be expected to add insulation to limit electron escape to the filament tips. In contrast, the OmcZ solvent-exposed heme raises the possibility that networks of OmcZ filaments may form with a multiplicity of paths through which electrons can flow, supporting a more conductive biofilm (Fig. 3E). In fact, we observed by cryo-EM extensive meshes of OmcZ filaments (Fig. 1A) that were not seen for OmcS and OmcE. Such junctions, whether they be the crossing over of two filaments or one filament forming a branch with another, could be paths for electron flow, while exposed hemes and lack of glycosylation along the surface of the network could offer multiple sites for electron transfer to surfaces. Given the multitude of uncharacterized cytochromes in genomes of metal-reducing organisms, and evidence that conductive filaments have arisen multiple times, this binary model will likely become increasingly complex as new structures become available.
Materials and Methods
G. sulfurreducens growth and OmcZ filament preparation
G. sulfurreducens were grown in anoxic basal medium with acetate (20 mM) as the electron donor and fumarate (40 mM) as the electron acceptor with 0.38 g/L potassium chloride, 0.2 g/L ammonium chloride, 0.069 g/L monosodium phosphate, 0.04 g/L calcium chloride dihydrate, 0.2 g/L magnesium sulfate heptahydrate and 10 mL mineral mix (Chan et al., 2015). The pH of the medium was adjusted to 6.8, buffered with 2 g/L sodium bicarbonate, and sparged with N2:CO2 gas (80:20) passed over a heated copper column to remove trace oxygen. For all experiments, G. sulfurreducens strains were revived anaerobically from frozen DMSO stocks for single colony isolates on 1.2% agar plates. All cultures were grown at 30° C.
OmcZ filaments were isolated from electrode-grown G. sulfurreducens PCA ΔomcS strain (Wang et al., 2022b). Three-electrode bioreactors with a working volume of 100 ml with 40 mM acetate as the electron donor were assembled as previously described (Marsili et al., 2008). The potential of two polished graphite working electrode with a surface area of 9 cm2 each was maintained at 0.240 V vs. standard hydrogen electrode (SHE) using a VMP3 multichannel potentiostat (Biologic), a stainless-steel mesh to serve as a counter-electrode and calomel reference. Transcripts of omcZ were ∼6-10 times more abundant when electrodes were used as terminal electron acceptor compared to fumarate (Jimenez Otero et al., 2018). Reactors were inoculated 1:100 with fumarate-grown cells. Additional acetate (40 mM) was added to the reactor when current reached ∼0.5 mA/cm2. Bioreactors were maintained at 30°C under a constant stream of humidified N2:CO2 (80:20) scrubbed free of oxygen by passage over a heated copper furnace. In prior work, the oxygen concentration in the headspace of these reactors has been shown to be ∼1 ppm.
G. sulfurreducens biofilm were harvested when current density reached ∼ 1 mA/cm2. Biofilm from eight 9 cm2 electrodes was scraped into 80 mls of 150 mM ethanolamine buffer pH 10.5 with 0.25 U/ml benzonase. OmcZ filaments were sheared using a Waring Commercial Blender (Cat. No. 7011S) for 3 minutes on low setting, then 1 minute on high setting. Cell debris was removed by centrifugation at 8,000 x g for 20 minutes.
SDS-PAGE of filaments sample
The sheared filaments mixture was filtered and concentrated using 100 kDa and 300 kDa concentrator units (Sartorius Vivaspin) after multiple washing with buffer and subsequently ultrapure water. The concentrated filaments sample was diluted to 0.3 mg/mL with 50 mM ammonium bicarbonate (pH 7.8) buffer. The filaments sample was analyzed by sodium-dodecyl sulfate - polyacrylamide gel electrophoresis (SDS-PAGE) to assess the separation as previously described (Wang et al., 2019). Briefly, Acryl/Bis (37.5:1, 40% w/v) solution (VWR) was used to prepare 4-16% Tricine-SDS-PAGE (Schagger, 2006) gels. Diluted filament samples with a final concentration of 0.15 mg/mL from each 100 kDa and 300 kDa filtration step were mixed with sample buffer with 0.5% and 2% w/v final concentrations of SDS. Before loading the samples into the gel, samples were boiled for 20 min before cooling to room temperature, followed by spinning at 2000xg for 1 min. The samples were allowed to run with an initial voltage of 50 V for 10 min and next voltage step of 200V for 2.5 hours. A Spectra Multicolor Broad Range Protein Ladder (ThermoFisher) was used as mass standards for the protein bands. After washing the gels with ultrapure water three times, gels were stained with silver stain (Kavran and Leahy, 2014) or with 3,3′,5,5′-tetramethylbenzidine (TMB) (Thomas et al., 1976).
In-solution digestion by Trypsin/Chymotrypsin
In-solution digestion protocol was adapted from Promega protocols for Trypsin Gold and Chymotrypsin (Promega). A 0.3 mg/mL of filaments sample obtained after the 300 kDa filtration step was digested with a combination of Trypsin Gold and Chymotrypsin. For Trypsin Gold, a 50 µg was mixed with 50 mM acetic acid to reach 1µg/µL and diluted to 20 µg/mL in 50 mM ammonium bicarbonate (pH 7.8). For Chymotrypsin, a 25 µg was resuspended in 1 mM HCl for a final concentration of 1µg/µL and diluted to 20 µg/mL in 50 mM ammonium bicarbonate (pH 7.8). A 5 µL of 100 mM DL-dithiothreitol (Sigma-Aldrich) was used to reduce a 30 µL of the filaments sample followed by incubation at 37 ºC for 1 hour. The reduced sample was spiked with 10 µL of 100 mM iodoacetamide (Sigma-Aldrich) and incubated in the dark at room temperature for 45 min. Then, the mixture was diluted with 90 µL of 50 mM ammonium bicarbonate (pH 7.8) before adding 15µL of the combination of Chymotrypsin and Trypsin Gold. The filaments sample was incubated for 3 hours at 37 ºC.
Mass spectrometry of filament samples
The proteolyzed peptides mixture of filament sample was quenched after 3 hours incubation by adding 30 µL of 1% trifluoroacetic acid before MALDI-TOF/TOF (UltrafleXtreme, Bruker) characterization in positive ion mode. A saturated solution of α-cyano-4-hydroxycinnamic acid (Sigma-Aldrich) was used for the matrix solution by dissolving with a 2:1 solution of ultrapure water:acetonitrile (Ing et al., 2017). The sample spot was recrystallized on the MALDI plate with a 1:1 ratio of 0.1% of trifluoroacetic acid and acetonitrile. The data were analyzed using mMass© protein identification software (Niedermeyer and Strohalm, 2012; Strohalm et al., 2008; Strohalm et al., 2010).
Cryo-EM conditions and image processing
The cell appendage sample (ca. 3.5-4.0 μl) was applied to glow-discharged lacey carbon grids, and then plunge frozen using an EM GP Plunge Freezer (Leica). The cryo-EMs were collected on a 300 keV Titan Krios with a K3 camera (University of Virginia) at 1.08 Å/pixel and a total dose of ca. 48 e/Å2. Motion corrections and CTF estimations were done in cryoSPARC (Punjani et al., 2017; Rohou and Grigorieff, 2015; Zheng et al., 2017). Particles were auto-picked by “Filament Tracer” with a shift of 60 pixels between adjacent boxes. All auto-picked particles underwent multiple rounds of 2D classification, and all particles in bad 2D class averages were removed. After this, the OmcZ dataset had 92,170 particles remaining. The helical symmetry was originally determined in our previous study (Wang et al., 2019). After a helical refinement applying and optimizing the helical symmetry, a non-uniform refinement was performed to improve the resolution in the central area of the map. The resulting half maps were then sharpened by “highres” mode in DeepEMhancer (Sanchez-Garcia et al., 2021), and the cryo-EM parameters are listed in Table 1.
Model building of OmcE filaments
The hand of the OmcZ filaments was determined by the hand of an α-helix. The hand assignment also agreed with the AlphaFold (Jumper et al., 2021) predictions of OmcZ. The AlphaFold prediction was docked into the cryo-EM map, and the regions that did not fit well were manually adjusted in Coot (Emsley and Cowtan, 2004). To better refine heme-interacting areas at this resolution, bond/angle restraints for the heme molecule itself, His-Fe, and Cys-heme thioester bonds were restricted based on the geometries obtained in high resolution crystal structures such as NrfB20 (PDB 2P0B) and NrfHA21 (PDB 2J7A). Real-space refinement was performed with those restraints (Afonine et al., 2018). To clean up the protein geometry, OmcZ models were further rebuilt using Rosetta (Wang et al., 2015). MolProbity (Williams et al., 2018) was used to evaluate the quality of the filament model, and the refinement statistics are shown in Table 1.
Structural analysis of heme c pairs
All structural coordinates with heme ligand (HEC or/and HEM) were downloaded from the Protein Data Bank. All possible heme pairs were then filtered with a minimum distance less than or equal to 6 Å, with the “contact” command in UCSF-ChimeraX (Pettersen et al., 2021). For each qualified pair, the rotation matrix between two porphyrin rings were generated in ChimeraX using the “align” command. The rotation angle θ was then calculated from the rotation matrix with the following equation, where tr is the trace of the rotation matrix:
Data Availability
The atomic model and three-dimensional reconstruction of OmcZ have been deposited in the Protein Data Bank (8D9M) and Electron Microscopy Data Bank (EMD-27266), respectively
Acknowledgments
The cryo-EM imaging was done at the Molecular Electron Microscopy Core Facility at the University of Virginia, which is supported by the School of Medicine and built with NIH grant G20-RR31199. This work was supported by NIH Grant GM122510 (E.H.E.), K99GM138756 (F.W.), DOE grant DE-SC0020322 (A.I.H., D.R.B., E.H.E.), AFOSR grant FA9550-19-1-0380 (A.I.H), NSF grant 2030381 (D.S.), Office of Naval research grant N00014-18-1-2632 (C.H.C, K.J., and S. C.) and the SRCP Seed Grant at the University of Washington Bothell (D.S).