Methane-oxidizing microbes catalyze the oxidation of the greenhouse gas methane using the copper-dependent enzyme particulate methane monooxygenase (pMMO). Isolated pMMO exhibits lower activity than whole cells, however, suggesting that additional components may be required. A pMMO homolog, ammonia monooxygenase (AMO), converts ammonia to hydroxylamine in ammonia-oxidizing bacteria (AOB) which produce another potent greenhouse gas, nitrous oxide. Here we show that PmoD, a protein encoded within many pmo operons that is homologous to the AmoD proteins encoded within AOB amo operons, forms a copper center that exhibits the features of a well-defined CuA site using a previously unobserved ligand set derived from a cupredoxin homodimer. PmoD is critical for copper-dependent growth on methane, and genetic analyses strongly support a role directly related to pMMO and AMO. These findings identify a copper-binding protein that may represent a missing link in the function of enzymes critical to the global carbon and nitrogen cycles.
Methane-oxidizing microbes1 are nature’s primary sink for methane, which is the second most abundant greenhouse gas and has a global warming potential ~84 times that of CO2 over a 20-year period2. The first metabolic step carried out by methane-oxidizing bacteria (MOB), the oxidation of methane to methanol, is catalyzed by methane monooxygenase (MMO) enzymes3. Understanding how MMOs break a 105 kcal mol−1 4 methane C-H bond is central to efforts to develop biological gas-to-liquid technologies5. The primary MMO in nature is particulate MMO (pMMO), a copper-dependent, integral membrane enzyme composed of three subunits, PmoB, PmoA, and PmoC3,5. A pMMO homolog, ammonia monooxygenase (AMO), converts ammonia to hydroxylamine in ammonia-oxidizing bacteria (AOB)6 which produce another potent greenhouse gas, nitrous oxide7.
Studies of pMMO and AMO have been hindered for the past 30 years by low enzymatic activity upon isolation. pMMO activity decreases 10- to 100-fold upon membrane isolation and protein purification3, and a purified AMO sample with enzymatic activity has never been obtained8. The persistent difficulties in obtaining high-activity pMMO/AMO preparations may be due to the absence of unidentified protein components that facilitate loading, assembly, and stabilization of the active sites and/or delivery of electrons and protons. Additional genes encoded by the pmo and amo operons provide potential candidates for such missing links. In alpha-proteobacterial MOB (α-MOB), including the Methylosinus and Methylocystis genera, and in beta-proteobacterial AOB (β-AOB), including the Nitrosomonas, Nitrosospira, and Nitrosovibrio genera9, the gene denoted pmoD/amoD is adjacent to the genes encoding the three pMMO/AMO enzyme subunits (Fig. 1a), and is partially co-regulated with pmoCAB and amoCAB10,11,12. In beta-proteobacterial amo operons, an additional gene homologous to amoD, denoted amoE/orf4, precedes amoD (Fig. 1a)9. These extended operons also encode CopC, CopD, and sometimes PCuAC proteins, all of which have been implicated in copper homeostasis in a range of bacteria10. Gamma-proteobacterial pmo/amo operons contain fewer genes: in Nitrosococcus genera, only amoD follows amoCAB, and in related gamma-proteobacterial methanotrophs, including the Methylococcus genera, pmoD genes are present, but are not found in pmo operons11,13. Most MOB and AOB genomes contain at least two pmoD/amoD homologs, including those of MOB from the verrucomicrobial and NC-10 subgroups and AOB belonging to the Nitrospirae phylum, although not archaeal ammonia oxidizers. Here we show that PmoD forms a copper center exhibiting the features of a CuA site using a previously unobserved ligand set at the interface of a PmoD homodimer. We further show that PmoD is critical for the copper-dependent growth of Methylosinus trichosporium OB3b on methane. Finally, genetic analyses support a role for PmoD/AmoD related to the function of pMMO/AMO.
PmoD forms a CuA site as a homodimer
Sequence analysis indicates that PmoD/AmoD comprises an N-terminal signal peptide followed by a periplasmic domain containing two strictly conserved cysteine residues and a C-terminal transmembrane helix within the inner membrane (Supplementary Figure 1a). We heterologously expressed and purified the periplasmic domain of PmoD encoded by the Methylocystis species (sp.) strain Rockwell extended pmo operon (PmoDMet49242_1452; there are 10 other pmoD genes in the genome of this species) (Supplementary Figure 1b, d). Addition of CuSO4 to PmoDMet49242_1452 after reduction with dithiothreitol (DTT) gives rise to a species characterized by an intense purplish pink color. The absorption spectrum of this species exhibits peaks at 475, 530, and 770 nm, resembling spectra of the CuA-containing domains of cytochrome c oxidase (CcO), nitrous oxide reductase (N2OR), and engineered CuA proteins, including CuA azurin14 (Fig. 1b). The copper-loaded protein can be further resolved into two distinct species by size exclusion chromatography (Supplementary Figure 1c). The first of these retains the CuA-like spectral features and contains 1.4 ± 0.2 equivalents of Cu per protein molecule, while the second lacks the CuA-like spectral features and contains 3.1 ± 1.2 equivalents of Cu per protein molecule. Each was further analyzed using size exclusion chromatography with multi-angle light scattering (SEC-MALS): the CuA-like species has a molar mass of 30,540 Da (±2%), consistent with a PmoDMet49242_1452 dimer, while the second species has a molar mass of 16,250 Da (±3%), consistent with a PmoDMet49242_1452 monomer (Fig. 1c).
CuA centers feature a mixed valent Cu(+1.5)–Cu(+1.5) site in which two copper ions separated by ~2.5 Å are coordinated by two bridging cysteine thiolates, two in-plane histidine imidazoles, and a methionine thioether and backbone carbonyl as axial ligands. The canonical CuA amino acid ligands derive from a conserved Hx34Cx3Cx3Hx2M motif15 which is not present in PmoD sequences. In typical CuA centers, the 475 nm and 530 nm absorption maxima are assigned to sulfur (Cys) to Cu ligand-to-metal charge transfer transitions, and the longer wavelength near-infrared feature (~770 nm) arises from the Ψ → Ψ* Cu–Cu bonding to antibonding orbital transition16. The energy of this transition is inversely correlated to the length of the Cu–Cu bond14. In the PmoDMet49242_1452 spectrum, this feature is blue-shifted relative to natural CuA systems (775–808 nm for CcO and 780–800 nm and N2OR)14 and is more similar to that observed for the semisynthetic CuA azurin (765 nm)17, indicating that PmoDMet49242_1452 has a short Cu–Cu distance with strong Cu–Cu bonding. The presence of a CuA site with a Cu2S2 core is definitively established by the PmoDMet49242_1452 electron paramagnetic resonance (EPR) spectrum which exhibits the characteristic seven hyperfine line splitting along gz (Fig. 1d)18, defining a fully valence-delocalized dicopper center. However, the Az and gz values for the PmoDMet49242_1452 CuA site (Az = 141 MHz, gz = 2.13) are significantly smaller than those of CuA azurin (Az = 167 MHz, gz = 2.1719), while compared to CcO and N2OR (Az = 107-117 MHz, gz = 2.18, Az = 117 MHz, gz = 2.18, respectively)14, the Az is larger and the gz is smaller. The PmoD gz = 2.13 is the smallest observed for a CuA site to date14. By contrast, the monomeric species of PmoDMet49242_1452 exhibits an optical spectrum with no charge transfer bands (Fig. 1c, inset) and is EPR silent. The X-ray absorption near edge structure (XANES) spectra of PmoDMet49242_1452 exhibit features similar to those observed for CuA model compounds and the soluble CuA domain of Thermus thermophilus CcO20 (Fig. 1e, Supplementary Discussion). In addition, extended X-ray absorption fine structure (EXAFS) data can be fit with a 2.41 Å Cu–Cu interaction (Supplementary Discussion, Supplementary Figure 2a-d, Supplementary Table 1), consistent with the observed absorption spectrum.
PmoD uses a non-canonical ligand set to form the CuA site
Both the dimeric and monomeric species of PmoDMet49242_1452 were subjected to crystallization trials, but crystals were obtained from only the monomeric form. The 1.9 Å resolution structure solved by copper single wavelength anomalous dispersion (SAD) phasing (Supplementary Table 2, Supplementary Figure 3a) reveals a cupredoxin fold14 with a single copper ion located between the two PmoDMet49242_1452 monomers in the asymmetric unit. The copper ion is coordinated in tetrahedral geometry by Met42 and Met44 from each monomer, and the two invariant cysteine residues, Cys41 and Cys65, form an intramolecular disulfide bond (Fig. 2a and Supplementary Figure 3b, c). Given that the crystals were obtained from the monomeric PmoDMet49242_1452 species, it is likely that the observed copper-binding site is a crystallization artifact. Nevertheless, comparison of the overall structure to other CuA-containing domains shows that the CuA site in PmoDMet49242_1452 must be formed in a unique fashion.
The coordinates of PmoDMet49242_1452 can be superposed on those of the CuA domains of CcO21 (2CUA (https://doi.org/10.2210/pdb2cua/pdb]) and N2OR22 (1FWX (https://doi.org/10.2210/pdb1fwx/pdb)), as well as the engineered azurin CuA domain23 (1CC3 (https://doi.org/10.2210/pdb1cc3/pdb)) with root-mean-square deviation (r.m.s.d.) values of 3.06, 2.82, and 2.95 Å for 82, 80, and 82 Cα atoms, respectively (Supplementary Figure 3d, e). The PmoDMet49242_1452 structure differs from these canonical CuA domain structures in three distinct regions (Fig. 2b and Supplementary Figure 3b, c). First, the N-terminal histidine ligand in the Hx34Cx3Cx3Hx2M CuA motif typically derives from the loop between β3 and β4, but PmoDMet49242_1452 contains a small helix in place of this loop and lacks a histidine in the equivalent position. Second, the loop between β7 and β8 provides the other four CuA ligands, but PmoDMet49242_1452 lacks not only this sequence motif, but also any histidine, cysteine, or methionine residues in this area. Instead, this region in PmoDMet49242_1452 is an extended β hairpin (Fig. 2b). Finally, the N-terminal region of PmoDMet49242_1452 diverges in the orientation of β1 and β2, as well as the presence of an additional β strand prior to β1. These changes at the N terminus are likely due to the intramolecular disulfide bond (Supplementary Figure 3b).
Instead of the canonical CuA ligands, PmoDMet49242_1452 contains a cluster of surface-exposed cysteine, methionine, and histidine residues near its N terminus, including the universally conserved cysteine residues Cys41 and Cys65 (Fig. 2c). A variant of PmoDMet49242_1452 in which Cys65 is replaced with serine retains the CuA features (Fig. 1b,e, Supplementary Figure 2e-h), and exhibits nearly complete dimerization upon addition of copper (Supplementary Figure 1c) with 0.80 ± 0.02 equivalents of Cu bound per protein molecule. While these data indicate that Cys65 is not a CuA ligand, the reduced copper stoichiometry compared to wild type indicates that Cys65 may bind an additional copper ion. By contrast, replacement of Cys41 with serine abolished appearance of the CuA spectral features. Since PmoDMet49242_1452 only contains these two conserved cysteine residues and CuA sites require two bridging thiolates, the CuA site must be formed between two PmoDMet49242_1452 monomers, of which each contributes a Cys41 ligand. The periplasmic domain of PmoDMet49242_1452 contains four histidine residues, of which one, His51, is located near Cys41 (Fig. 2c). Replacement of His51 with alanine also eliminates CuA formation, as does replacement of Met49 with alanine (Fig. 1b), implicating these two residues as ligands. Thus, despite the overall similarity in fold to other CuA domains, the PmoD CuA site is formed by two protein molecules and a completely different part of the protein structure (Fig. 2d) in contrast to all previously characterized CuA centers which are housed within a single cupredoxin domain.
Disruption of pmoD in vivo results in a copper-dependent growth defect
To assess the role of PmoD in vivo, we disrupted the pmoD gene in the extended copy of the pmo operon (containing pmoCAB as well as the pmoD, copC, copD, and pCuAC genes) in Methylosinus trichosporium OB3b (PmoDMettrDRAFT_0381). Unlike Methylocystis sp. Rockwell, Methylosinus trichosporium OB3b also possesses genes for the soluble, iron-containing MMO (sMMO) and for proteins related to the biosynthesis, regulation, and transport of the siderophore-like copper chelator methanobactin (Mbn)24. Under copper starvation conditions, sMMO and Mbn are produced; both processes are associated with characteristic phenotypes. When copper is bioavailable, these genes are repressed as part of the so-called copper switch, and the pMMO genes are mildly up-regulated10. Growth of both the wild-type and ΔpmoD strains was monitored in either copper-free media or media supplemented with 10 μM CuSO4. Under copper starvation conditions, ΔpmoD Methylosinus trichosporium OB3b is phenotypically similar to the wild-type strain. The strains grow at a comparable rate (Fig. 3a), and both produce Mbn (Supplementary Figure 4c, d), exhibit sMMO activity (Supplementary Figure 4e), and produce sMMO (Supplementary Figure 4f). Under copper-replete conditions, however, ΔpmoD Methylosinus trichosporium OB3b exhibits a pronounced growth defect. Cultures grown in copper-supplemented media produce significantly less cell density after 1 week of growth (Fig. 3b), and the ΔpmoD strain also exhibits a growth defect on nitrate mineral salts (NMS) agar plates supplemented with CuSO4 (Supplementary Figure 4a, b). This phenotype is distinct from the wild-type strain which exhibits a slight increase in growth rate under copper-replete conditions25 (Fig. 3b). Despite the growth defect in the presence of copper, the ΔpmoD strain, like the wild-type strain, does not produce Mbn (Supplementary Figure 4g, h), and no longer exhibits sMMO activity (Supplementary Figure 4i). While the sMMO subunits are significantly less abundant in copper-replete cell lysates, the pMMO subunits are detected in both the wild-type and ΔpmoD Methylosinus trichosporium OB3b strains (Supplementary Figure 4j). Thus, PmoDMettrDRAFT_0381 does not appear to play a role under sMMO-utilizing conditions, but is important for growth under pMMO-utilizing conditions. It is not clear, however, whether the CuA site is linked to the observed growth defect.
PmoD homologs exhibit varying levels of copper-dependent regulation
Interestingly, this copper-dependent growth defect is observed despite the presence of seven additional PmoD homologs encoded in the Methylosinus trichosporium OB3b genome, suggesting that all PmoD proteins are not functionally equivalent. We previously showed that the components of the full Methylosinus trichosporium OB3b pmo operon (pmoCAB, pmoD, copC, copD, and pCuAC) are very mildly up-regulated by copper10, and other studies have shown that in AOB, amoD (and, if present, amoE) are co-up-regulated and co-transcribed with amoCAB upon exposure to ammonia12,26. The regulation of pmoD homologs beyond the pmo operon has not been investigated. Therefore, we conducted time-dependent, copper-responsive quantitative polymerase chain reaction (qPCR) analysis of all Methylosinus trichosporium OB3b non-operon pmoD genes. We observe varying patterns of copper-dependent regulation, with some pmoD genes mildly repressed, some up-regulated, and several not regulated by copper at all (Fig. 3c and Supplementary Figure 5), further supporting that not all PmoD proteins are functionally equivalent. A recent RNA-sequencing study on Methylosinus trichosporium OB3b that lacks the time-resolved component of these experiments is nevertheless consistent with the observed regulatory patterns for pmoDs with significant regulatory changes and/or correlated co-regulation with other copper-responsive genes27. Moreover, differential regulation of pmoD genes in response to copper is also observed for Methylococcus capsulatus (Bath)28 which has two pmoD genes of which neither is located in its pmo operon. These regulatory patterns combined with the ΔpmoD Methylosinus trichosporium OB3b phenotype suggest that PmoD homologs play distinct roles in vivo.
In support of nonredundant functionality, the production of many PmoD homologs is evident by proteomics at variable levels. We detected PmoD soluble domains and predicted C-terminal helices, but not the N-terminal signal peptides, in membrane fractions or whole cell lysate from five different methanotrophs (Supplementary Figure 6). For Methylocystis sp. Rockwell, 5 of the 11 PmoD homologs are observed in solubilized membrane fractions, including PmoDMet49242_1452 (Supplementary Figure 6a). In solubilized membrane fractions from Methylomicrobium buryatense 5GB1C (Supplementary Figure 6b) and Methylomicrobium alcaliphilum 20Z (Supplementary Figure 6c), whose genomes each contain two PmoD homologs, both homologs are detected. In the case of Methylococcus capsulatus (Bath) (Supplementary Figure 6d), which also has two PmoD homologs, one was observed. In whole cell lysate from Methylosinus trichosporium OB3b grown under copper-replete conditions, six of eight PmoD homologs are observed, including those encoded by genes adjacent to copCD, pCuAC, and pmoCAB (Supplementary Figure 6e).
Bioinformatics analyses suggest functional differences
To address potential functional differences among PmoD homologs, we performed a detailed bioinformatic analysis of all PmoD sequences identifiable in the Joint Genome Institute/Integrated Microbial Genomes (JGI/IMG) genome database. The vast majority of these are found in MOB and AOB, and all are found in species with genomes containing at least one operon encoding a copper membrane monooxygenase (CuMMO) related to pMMO/AMO (Fig. 4a). Notably, the only genomes that encode pMMO/AMO homologs, but not PmoD proteins, are species that oxidize neither methane nor ammonia such as strains of Mycobacteria and Nocardiodes29 (Fig. 4b). The number of PmoD homologs per genome ranges from one (e.g., Methylovulum and Methylogaea species, and some species outside of canonically methanotrophic genera) to 11 (Methylocystis sp. Rockwell), all of which differ in sequence and genomic context. Most canonical MOB and AOB genomes contain at least two PmoD homologs. An alignment of all putative PmoD amino acid sequences was used to construct a preliminary PmoD profile hidden Markov model (HMM) and to identify several core sequence motifs involving the two invariant cysteines and other putative metal binding residues (Supplementary Figure 7). Genomic neighborhood analysis provides additional insight into the link between the various PmoD proteins and pMMO/AMO (Supplementary Figure 8). PmoD proteins containing either a Cx7MxHxnC motif (like PmoDMet49242_1452) or a Cx8MHxnC motif comprise 23.6% and 13.4%, respectively, of all PmoDs (Fig. 4c, Supplementary Figure 7). These subsets are predominantly associated with pmo/amo operons (the latter annotated in AOB as amoD and the former as amoE) and are also found in the proximity of some lone pmoC genes (Fig. 4d). In α-MOB, the pmoD genes found in pmo operons encode PmoDs with the Cx7MxHxnC motif, as observed in the proteins encoded by amoE and not in amoD. The close association between genes encoding PmoD and pMMO/AMO is markedly different from the neighboring genes encoding CopCs, CopDs, and PCuACs which are quite common in other prokaryotes30. These in-operon PmoDs are likely to have functions specific to pMMO and AMO, consistent with the phenotype observed in the ΔpmoD Methylosinus trichosporium OB3b strain (Fig. 3a).
CuA site formation is limited to homologs with a Cx7MxHxnC motif
Since the ligand set identified for the CuA site in PmoDMet49242_1452 (Cys41, Met49, and His51), along with Cys65, which is not involved in the CuA site, corresponds to the Cx7MxHxnC motif (Fig. 4c, Supplementary Figure 7), we investigated whether other PmoD homologs containing this motif are also capable of forming a CuA center. We expressed and purified PmoDO5ODRAFT_02836, a homolog from Methylocystis parvus OBBP. According to metal quantitation and EPR and optical spectroscopic characterization, PmoDO5ODRAFT_02836, like PmoDMet49242_1452, binds copper (Supplementary Table 3) and forms a CuA species (Fig. 4e, Supplementary Table 4, Supplementary Figure 9), although a significant fraction (estimated at ~50%) of the PmoDMet49242_1452 EPR signal is attributable to mononuclear type 2 Cu(II) resonance31,32.
We also examined the genomic neighborhoods of genes encoding PmoDs with other putative metal binding motifs and conducted in vitro metal binding and spectroscopic experiments on representative members of these classes. The representative Cx8MxnC motif-containing (7.9% of all PmoD homologs, commonly near genes encoding multicopper oxidases, PCuACs, and ABC transporters) and Cx9HxnC motif-containing (22.2%, genes often found near copC) PmoDs Methylocystis parvus OBBP PmoDO5ODRAFT_02342 and Methylosinus trichosporiumOB3b PmoDMettrDRAFT_0370 bind copper (Supplementary Table 3), but do not exhibit optical or EPR spectroscopic features consistent with a CuA site (Fig. 4e, Supplementary Figure 9). In addition, a large number of PmoDs have no conserved metal binding residues beyond the CxnC motif (29.0%) and typically neighbor genes of unknown relevance or are adjacent to copC genes (Supplementary Figure 8). While one such PmoD, PmoDMet49242_0515, also binds copper, it too cannot form a CuA species (Supplementary Table 3, Fig. 4e, Supplementary Figure 9). Thus, CuA site formation is confined to Cx7MxHxnC-containing PmoDs encoded by genes within extended pmo/amo operons.
PmoD/AmoD represent a family of copper-binding proteins specific to MOB and AOB, the key players in global carbon and nitrogen cycling. The subset of PmoD homologs encoded in genomic proximity to and co-regulated with pMMO/AMO contains an unprecedented type of CuA site and is important for copper-dependent growth on methane. PmoD homologs are more closely correlated with CuMMOs than any other recently proposed members of the MOB/AOB copper homeostasis system, including the periplasmic copper sponge protein Csp1/2, the cytoplasmic copper sponge protein Csp3, the secreted copper-binding proteins CorA/MopE, and the copper-binding chalkophore Mbn (Fig. 4a). This observation strongly supports a unique role for PmoDs in the function of CuMMOs.
Potential functions of PmoD include providing reducing equivalents to drive pMMO reactivity, facilitating copper loading of pMMO, or even structurally stabilizing pMMO within the membrane. The specific biological role of PmoD, including whether its function in vivo is mediated via the homodimeric CuA species, remains to be determined. More comprehensive characterization of the unusual CuA site, combined with additional in vivo and in vitro studies of PmoD including its transmembrane helix, are needed to address these possibilities. Furthermore, functional differences among the entire PmoD/AmoD family remain to be elucidated. Given the prevalent genomic proximity of pmoD/amoD genes to those encoding a range of copper-containing proteins, it is likely that additional copper centers and copper-related functions will be uncovered.
Cloning and construct design for PmoD expression plasmids
The expression constructs for the N-terminally His6-tagged tobacco etch virus (TEV)-cleavable PmoD periplasmic domains were all prepared in the pSCG-His vector by Anthony Gizzi and Steven Almo (Albert Einstein College of Medicine) using the primers listed in Supplementary Table 5. The cysteine variants in PmoDMet49242_1452 were generated via a QuikChange Lightning kit as directed (Agilent), using primers listed in Supplementary Table 5. The Met49Ala and His51Ala variants were synthesized by Genscript along with a TEV cleavage site before the N-terminal residue and subcloned into the pET28a vector. All constructs include an additional vector-derived serine and methionine at the N terminus.
Expression and purification of the PmoD periplasmic domains
All PmoD periplasmic domain constructs were expressed and purified following the same protocol. Plasmid DNA was transformed into BL21*(DE3) cells (Novagen). Overnight cultures in LB media were grown at 37 °C with agitation at 200 rpm, followed by inoculation into 1 L autoinduction media33 and incubation at 37 °C with agitation at 200 rpm until reaching an optical density at 600 nm of 0.6. The temperature was then reduced to 22 °C, and the cells were harvested after ~18 h by centrifugation at 6000 × g for 15 min. The resulting cell pellet was resuspended in lysis buffer (20 mM imidazole, 20 mM Tris, pH 8.0, 500 mM NaCl) supplemented with 1 mM DTT, 1 mg mL−1 DNase I (Sigma), and 1 mM phenylmethylsulfonyl fluoride dissolved in ethanol (EtOH). Cells were lysed via sonication for 15 min with 1 s pulses. Cell debris was removed by centrifugation at 24,000 × g for 1 h.
The resulting supernatant was then applied to NiNTA resin (Qiagen) pre-equilibrated with chilled lysis buffer. The flow-through was discarded, and the resin was washed with 5 column volumes of lysis buffer. His6-TEV-PmoD was eluted with 2 column volumes of elution buffer consisting of lysis buffer with 500 mM imidazole. To remove the His6 tag, TEV protease was added to the eluate, and the sample was dialyzed in 10 kDa molecular weight cutoff (MWCO) SnakeSkin dialysis tubing (Thermo) against 20 mM Tris, pH 7.0, and 1 mM DTT overnight at 4 °C. Following cleavage, the dialyzed sample was then re-applied to NiNTA resin to remove the TEV protease and the cleaved tag. The flow-through containing purified PmoD was then concentrated in 10 kDa MWCO Amicon spin concentrators (Millipore) and either used immediately or flash frozen in liquid nitrogen and stored at −80 °C.
Metallation of PmoD
Purified PmoD samples were buffer exchanged by dilution into reductant-free buffer (20 mM Tris, pH 7.0) and concentrating in 10 kDa MWCO spin concentrators (Millipore) to 200–600 μM followed by the slow addition of 2 molar equivalents of CuSO4 with constant mixing on ice. Excess copper was then removed either using a disposable PD-10 desalting column (GE Healthcare Life Sciences) or by size exclusion chromatography using a HiLoad 16/600 Superdex 75 column (GE Healthcare Life Sciences). By the latter method, two distinct PmoDMet49242_1452 species can be resolved, the dimeric species that forms the CuA-like site and a monomeric copper-bound species. All protein concentrations were measured using the Bradford assay with a standard curve generated from known concentrations of bovine serum albumin. Optical spectra were collected using copper-loaded samples containing 200 μM protein in a quartz cuvette (Helma) at room temperature using an Agilent 8453 spectrophotometer. For metal analysis, all samples were first desalted on a PD-10 column into metal-free buffer (20 mM Tris, pH 7.5) and digested in 5% nitric acid in metal-free tubes (VWR). Standard curves were generated using a dilution series of a custom multi-element standard (Inorganic Ventures). The copper content of the samples was determined using a Thermo iCAP 7600 ICP-OES instrument in the Quantitative Bio-element Imaging Center (QBIC) core facility at Northwestern University.
Size exclusion chromatography with multi-angle light scattering
SEC-MALS was used to determine the molar masses of the two PmoD species resolved by SEC. Each peak was pooled and concentrated to 2–4 mg mL−1 in 10 kDa MWCO concentrators (Millipore). Samples were analyzed using an Agilent 1260 series high-performance liquid chromatography (HPLC) system equipped with diode-array detection absorbance in-line with a DAWN HELEOS II multi-angle static light scattering detector (Wyatt Technology), a QELS dynamic light scattering detector (Wyatt Technology), and a T-rEx differential refractive index detector (Wyatt Technology). Then, 300 μL protein was applied to a Superdex 75 Increase 10/300 GL column (GE Healthcare) that had been pre-equilibrated in a mobile phase consisting of 20 mM Tris, pH 7.0. The mobile phase and protein sample prior to injection were maintained at 4 °C and the column was kept at 8 °C. Each sample was run at 0.4 mL min−1 for 60 min. Optical spectra were also recorded at the apex of each peak. Data were processed and analyzed using Astra software version 5.3.4 (Wyatt Technology).
Electron paramagnetic resonance spectroscopy
EPR samples were prepared in 20 mM Tris, pH 7.0, at concentrations of 100–300 μM protein. Protein solution was loaded into Wilmad quartz X-band EPR tubes (Sigma Aldrich) and samples were then frozen in liquid nitrogen, where they were stored until analysis. Continuous wave (CW) X-band spectra were collected on a Bruker ESP-300 spectrometer with an Oxford Instruments ESR-900 helium flow cryostat.
X-ray absorption spectroscopy
The CuA-like dimeric species of wild-type PmoDMet49242_1452 isolated by size exclusion chromatography was concentrated to 1.8 mM Cu (1.26 mM protein) in 20 mM Tris, pH 7.0, 30% glycerol. The C65S sample was concentrated to 900 μM Cu (740 μM protein). Each sample was loaded into a 2 mm Lucite EXAFS cell wrapped in Kapton tape, flash frozen in liquid nitrogen, and stored in liquid nitrogen until exposure to X-ray radiation. X-ray absorption spectroscopy (XAS) data were collected at the Stanford Synchrotron Radiation Lightsource on beamline 9−3 which is equipped with a Si double-crystal monochromator with an upstream mirror for focusing and harmonic rejection. Fluorescence spectra were collected using a 100-element Ge solid-state detector (Canberra). During data collection, the continuous-flow liquid helium cryostat (Oxford Instruments) was stabilized at 10 K. Copper excitation data were collected using a 6 μm nickel filter and solar slits placed between the cryostat and detector to reduce scattering fluorescence. XAS spectra were measured using 5 eV steps in the pre-edge region (8750–8960 eV), 0.25 eV steps in the edge region (8986-9050 eV), and 0.05 Å−1 increments in the EXAFS region (to k = 13.3 Å−1, integrating from 1 to 25 s in a k3 weighted manner for a total scan length of approximately 40 min). A Cu foil spectrum was collected simultaneously with each protein spectrum for energy calibration, with an assigned first inflection point at 8980.3 eV. Spectra were closely monitored for any photodamage and slight photoreduction was observed in the second scan at each exposure position. To quantitate the extent and impact of photoreduction, individual spectra were collected at unique positions on the sample surface using a matrix positioning grid (ensuring a new face for radiation exposure), and only the initial exposure spectrum at each position was used during overall data analysis. Data represent the average of 6 scans.
XAS spectra were processed and analyzed using the EXAFSPAK program suite written for Macintosh OSX (EXAFSPAK; http://www-ssrl.slac.stanford.edu/~george/exafspak/exafs.htm, 2001), integrated with the Feff v8 software34 for theoretical model generation. EXAFS fitting analysis was performed on raw/unfiltered data. Single scattering models were calculated for oxygen, nitrogen, sulfur, copper, and carbon coordination to simulate possible copper ligand environments, with values for the scale factors (Sc) and E0 calibrated by previous fittings of characterized crystallographic copper model compounds35. Standard criteria for judging the best-fit EXAFS simulations included reasonable Debye–Waller factors (σ2 < 0.006 Å2)36, the spectral resolution calculated based on the energy range extent of usable data37, and the lowest mean square deviation between data and fit corrected for the number of degrees of freedom (F’)37. During the standard criteria simulations, only the bond length and Debye–Waller factor were allowed to vary for each ligand environment. A scale factor of 0.95 and E0 values of −12, −14, and –16 were used for Cu–C/N/O, Cu–S, and Cu–Cu theoretical model calibration; these values were obtained from simulating authentic Cu–Cu models35. Simulations applying a formal but static Cu–Cu interaction were also tested in an attempt to deconvolute potential Cu–Cu scattering in the Cu EXAFS for each sample set. Fits using the static Cu–Cu ligand environment followed the strategy outlined by Blackburn et al.20 which was applied to systems in which Cu–S and Cu–Cu interactions were nonresolvable, as is the case in these data. Static Cu–Cu fits held the Cu ligand parameters constant, manually stepping through the Cu–Cu bond length, but not allowing the bond length or Debye–Waller factor for this ligand environment to vary; however, the bond length and Debye–Waller factor in each additional ligand environment was allowed to vary during these static Cu–Cu simulations (Supplementary Table 1).
Crystallization and structure determination
Crystals of the copper-bound monomeric species were obtained using 3 mg mL−1 protein in the Classics Suite (Qiagen) condition H1 (0.1 M ammonium acetate, pH 4.6, 0.2 M ammonium sulfate, 25 % (w/v) polyethylene glycol (PEG) 4000). Crystallization conditions were further optimized, and the crystal used for data collection was obtained from sitting drop screens with 0.1 M sodium acetate, pH 4.6, 0.2 M ammonium sulfate, 18% PEG 4000 as the precipitant. Prior to data collection, the crystal was transferred to a cryoprotectant solution consisting of the precipitant supplemented with 10% ethylene glycol for 5 s and then submerged in liquid nitrogen. Data were collected at 100 K at a wavelength of 1.377 Å using a Pilatus detector at GM/CA beamline 23 ID-B at the Advanced Photon Source (APS) at Argonne National Laboratory. The crystals belong to space group P3221 with two PmoD monomers in the asymmetric unit. The data were processed using HKL2000 to 1.9 Å resolution (Supplementary Table 2). The structure was solved using the SAD method with phenix.autosol38, yielding a Bayes CC of 56.6 and a FOM (figure of merit) of 45.6, indicating that the solution was likely correct. Phenix.autobuild38 was used to obtain an initial model that included 245 residues in two protein chains and had an R/Rfree of 24.7%/27.0%. This model was improved through iterative rounds of model building using Coot39 and refinement using phenix.refine38 resulting in a final R/Rfree value of 17.2%/20.9% (Supplementary Table 2). The final model includes 253 residues (amino acids 38–163 in chain A and amino acids 37–163 in chain B), 1 Cu ion, 126 water molecules, and a single sulfate ion. The Ramachandran plot indicates that 99.2% of residues are in favored regions with the remaining 0.8% in allowed regions. The Molprobity40 score is 1.29 (99th percentile).
Construction of a Δ pmoD strain of Methylosinus trichosporium OB3b
The Methylosinus trichosporium OB3b ΔpmoD mutant was generated by chromosomal gene disruption into wild-type Methylosinus trichosporium OB3b (a gift from John Lipscomb) using methods described previously41. Briefly, a gentamicin resistance gene (genR) was inserted in the middle of PmoDMettrDRAFT_0381 via homologous recombination. Primers targeting regions 600 bp upstream and 600 bp downstream from the center of the pmoD gene were used to amplify the DNA flanking the inserted genR. The gentamicin resistance cassette was amplified from vector pFBOH-LIC using primers listed in Supplementary Table 5. The PCR products were assembled into pK18mobsacB_p15a41 using Gibson assembly with primers listed in Supplementary Table 5 and transformed into Escherichia coli S17-1 ATCC 47055 to produce plasmid pSYR16. This plasmid was then introduced into Methylosinus trichosporium OB3b via conjugation with E. coli S17-1 cells transformed with pSYR16 on NMS mating agar plates (0 μM CuSO4). After 2 days, mated cells were plated on NMS selection agar plates (0 μM CuSO4) containing kanamycin (25 μg mL−1) and nalidixic acid (10 μg mL−1). Single colonies from selection plates were picked and streaked onto NMS counter selection plates (0 μM CuSO4) containing 2.5% sucrose and gentamicin (10 μg mL−1). Single colonies from counter selection plates were then plated and maintained on NMS agar plates (0 μM CuSO4) containing gentamicin (10 μg mL−1). Genotyping and sequencing were performed to confirm generation of the mutant using the primers listed in Supplementary Table 542. Cells on NMS agar plates were incubated in GasPak plate incubation chambers (BD) at 30 °C and were subjected daily to vacuum and gas-exchange cycles with gas at a 1:1 methane-to-air ratio and 1 L min−1 for 3 min.
Phenotype analysis of Methylosinus trichosporium OB3b strains
Wild-type Methylosinus trichosporium OB3b and the ΔpmoD strain were cultivated as 50 ml cultures in 250 mL flasks in NMS medium (10 mM NaNO3, 1 mM K2SO4, 150 µM MgSO4, and 50 µM CaCl2, buffered with 3.5 mM phosphate buffer, pH 6.8) supplemented with trace metals solution (40 µM FeSO4, 1.6 µM ZnSO4, 1 µM H3BO3, 0.8 µM MnCl2, 0.5 µM KI, 0.2 µM Na2MoO4, and 0.2 µM CoCl2) at 30 °C shaking at 300 rpm. The cultures were sealed with gas-tight rubber septa and fed daily with a 1:3 air/methane mixture42. For copper-replete growths, the medium was supplemented with 10 μM CuSO4, and for ΔpmoD Methylosinus trichosporium OB3b, 5 μg mL−1 gentamicin was also included in the growth medium. Growth was monitored daily by withdrawing 1 mL aliquots and measuring the optical density of the cultures at 600 nm. After a week, phenotypes were analyzed using the naphthalene assay to assess sMMO activity. A saturating amount of naphthalene was added to 1 mL cell culture, followed by incubation with agitation at 30 °C for 1 h. sMMO activity was assessed qualitatively by the observation of a color change upon addition of 100 µL 1% solution of Fast Blue B salt (Sigma) to each reaction. The optical spectra of the spent media were analyzed for the presence of apo Mbn by the presence of distinct absorbance features at 345 and 390 nm. One milliliter of culture was centrifuged at 8000 × g for 10 min at room temperature and the ultraviolet (UV)–visible light spectrum of the supernatant was measured using an Agilent 8453 UV-visible spectrophotometer10. To account for growth variability between cultures, the samples for both the naphthalene assay and Mbn assay were adjusted to the same optical density in NMS medium. Cell lysates were prepared and analyzed using sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) as follows. A 50 mg cell pellet was resuspended in 85 μL buffer (250 mM NaCl, 25 mM PIPES, pH 7.0). Cells were lysed by the addition of 200 μL Bacterial Protein Extraction Reagent (B-PER) (Thermo Fisher Scientific) supplemented with 350 μg mL−1 lysozyme, 35 μg mL−1 DNase I, and 2 µL of a 25× stock solution of cOmplete EDTA-free Protease Inhibitor Cocktail (Roche). After nutation at 4 °C for 30 min, the solution was centrifuged for 25 min at 20,000 × g at 4 °C, 4× SDS loading dye was added to the clarified lysate (supernatant), and the sample was run on a 15% SDS-PAGE gel.
Time-dependent copper-responsive qPCR analysis of pmoD genes
All time-dependent copper-responsive qPCR experiments were performed in wild-type Methylosinus trichosporium OB3b. For these experiments, as in previous studies10,43, Methylosinus trichosporium OB3b cultures were grown in nitric acid-washed spinner flasks to minimize copper concentrations. Cells were grown in NMS medium buffered with 3.5 mM phosphate buffer, pH 6.8, and amended with trace metals at a final concentration of 50 µM CuSO4, 40 µM FeSO4, 1.6 µM ZnSO4, 1 µM H3BO3, 0.8 µM MnCl2, 0.5 µM KI, 0.2 µM Na2MoO4, and 0.2 µM CoCl2. Cultures were stirred at 200 rpm and sparged continuously with a 1:3 methane/air mixture. Once cultures achieved logarithmic growth (A600 = 0.6–1.0), cells were assayed for the production of Mbn and sMMO as proxies for copper starvation. The presence of Mbn was confirmed via absorbance spectroscopy of clarified spent medium and the production of sMMO was confirmed via the colorimetric naphthalene assay.
Once cells achieved logarithmic phase growth and copper starvation was confirmed, time-course experiments were initiated. Cells were harvested prior to the addition of CuSO4 to a final concentration of 12.5 µM, as well as at three subsequent timepoints (15 min, 5 h, and 24 h). All experiments were repeated with three distinct biological replicates. At each timepoint, harvested cells were re-analyzed for Mbn production and sMMO production as well as used for RNA isolation.
All experiments involving RNA used RNase-free reagents and supplies. At each timepoint, culture volumes containing approximately 3 × 109 cells were removed from the harvested cells and added to an ice-cold stop solution containing 5% phenol and 95% EtOH, for a final ratio of 10:1.25 cell culture to stop solution. The cells were then centrifuged at 6000 × g for 20 min at 4 °C, and the resulting pellets were resuspended in 0.6 mL TE buffer, supplemented with 20 mg lysozyme (Sigma) and 100 µL proteinase K (Qiagen.) The resuspended cells were incubated for 30 min at room temperature on a rocker. After this incubation step, 5 mL of Qiazol (Qiagen) heated to 65 °C was added to each reaction, and RNA purification was performed using an RNeasy Midi kit (Qiagen), following the protocol described in the RNeasy Lipid Handbook. This initial process included an on-column DNase treatment step. All RNA was eluted in RNase-free water and underwent preliminary quantification using a Nanodrop (Thermo). Any remaining genomic DNA was removed using a Turbo DNAfree kit (Life Technologies). The purified RNA was further processed using an isopropanol/ethanol precipitation step and was resuspended in 20 µL RNase-free H2O (Life Technologies). The final RNA concentrations were initially determined using a Nanodrop 1000 (Thermo), but these concentrations were validated using a BioAnalyzer (Agilent), which also yielded RNA integrity numbers for all samples. A SuperScript VILO complementary DNA (cDNA) synthesis kit using random hexamer primers (Life Technologies) was used to synthesize cDNA from 1 µg input RNA.
Some primers used in these experiments had been validated previously10, including those for the reference genes and for components of the main mmo and pmo operons; new primers were validated using the same process. In short, primers were designed using the IDT PrimerQuest tool using constraints including an amplicon length between 70 and 150 bp, no possible products in the genome with fewer than 5 mismatches and 1–2 gaps (as identified using the e-PCR tool (NCBI)), 60 °C ≤ Tm ≤ 64 °C, ∆Ghomodimer ≥ −7.5, and ∆Ghairpin ≥ −3. qPCR reactions were carried out in white 384-well PCR plates (Bio-Rad), and each reaction contained 1.2 µL SYBR GreenER Express Universal Master Mix (Life Technologies) and 1.2 µL containing varying amounts of cDNA or RNA (1 ng in standard reactions, fourfold dilutions from 2 ng to 488 fg in primer efficiency reactions), water, and 90 nL pooled forward and reverse primers (5 µM each). Each reaction was run with three technical replicates, and no-template controls and no-reverse-transcriptase controls were also included for every primer/sample. Once prepared, qPCR reactions were incubated at 4 °C for 1 h, centrifuged briefly, and then loaded on a CFX384 (Bio-Rad) thermocycler.
qPCR reactions required an initial 2 min UDG (uracil-DNA glycosylase) inactivation step at 50 °C; this step was followed by a 2 min heating step at 95 °C. These steps were followed by 40 cycles at 95 °C for 15 s and 60 °C for 1 m; plate reading occurred during the combined annealing and extension steps. After the completion of 40 cycles, samples were allowed to anneal at 50 °C for 1 min, and melt curves were then obtained by performing a 5 s plate-reading step at 0.5 °C intervals from 50 °C to 95 °C. Initial Cq values were determined in the Bio-Rad CFX Manager software using regression for each trace. Wells with no Cq value (where replicates had Cq values), a Cq value 10 greater or less than the other technical replicates, or with a melting curve not corresponding to the melting curve peak of the correct product were removed from consideration. The size of all qPCR products and the specificity of all qPCR reactions were confirmed by loading 5 µL reactions onto a 4% NuSieve 3:1 agarose gel (Lonza) in TBE (Supplementary Fig. 5a).
While initial data screening for primer efficiency and gene expression levels were both analyzed in CFX Manager (Bio-Rad), further analysis was performed in qBasePLUS44,45, including calculation of primer efficiency (Supplementary Data 1) and calculation of normalized relative quantity (NRQ) values for all genes (Supplementary Fig. 5), along with additional statistical analyses (Supplementary Data 1). Hierarchical clustering of log10-transformed NRQ data was performed using Cluster 3.046, and these data were used to generate dendrograms and heatmaps in JavaTreeView47.
Membrane-bound pMMO samples from Methylocystis sp. strain Rockwell (ATCC), Methylococcus capsulatus (Bath) (ATCC), Methylomicrobium alcaliphilum 20Z (a gift from Marina Kalyuzhnaya), and Methylomicrobium buryatense 5GB1C (a gift from Mary Lidstrom) were prepared from cells grown in copper-replete medium. For Methylocystis sp. strain Rockwell, as in previous studies48, the growth medium consisted of an adapted NMS-based salt solution consisting of 0.08% w/v KNO3, 0.02% w/v K2SO4, 0.003% w/v MgSO4∙7H2O, and 0.00007% w/v CaCl2∙2H2O buffered with 3.9 mM phosphate buffer, pH 7.0, and supplemented by a trace metal solution consisting of 50 µM CuSO4, 40 µM FeSO4, 2 µM ZnSO4, 2 µM H3BO3, 1.6 µM MnCl2, 1 µM KI, 0.4 µM Na2MoO4, and 0.4 µM CoCl2. Cells were grown in 12 L fermentations at 30 °C and 300 rpm with gas sparging at a 3:1 air-to-methane ratio. Once an A600 = 5.0–8.0 was reached, cells were harvested by centrifugation at 5–7000 × g for 10 min. Cells were washed 1–3 times with a buffer consisting of 25 mM PIPES, pH 6.8, before freezing in liquid nitrogen. For Methylococcus capsulatus (Bath), as in previous growths48, the growth medium consisted of NMS solution which contains 0.2% w/v KNO3, 0.1% w/v MgSO4∙7H2O, and 0.001% CaCl2∙7H2O, along with 3.9 mM phosphate buffer, pH 6.8. This medium was supplemented with trace metals, including 50 µM CuSO4, 80 µM NaFe(III)EDTA, 1 µM Na2MoO4, and a trace metal solution composed of 0.6 µM Na2•EDTA, 0.4 µM FeSO4•7H2O, 0.02 µM ZnSO4•7H2O, 0.008 µM MnCl2•4H2O, 0.24 µM H3BO3, 0.042 µM CoCl2•6H2O, 0.004 µM NiCl2•6H2O, and 0.006 µM Na2MoO4•2H2O. Cells were grown in 12 L fermentations at 300 rpm and 45 °C with gas sparging at a 4:1 air/methane ratio. Cells were harvested as described for Methylocystis sp. Rockwell. As in previous studies49, Methylomicrobium alcaliphilum 20Z was grown in modified NMS (NMSA) medium amended with 2.3 mM phosphate buffer at pH 7.2, 50 mM carbonate buffer pH 9.5, 500 mM NaCl, 40 µM CuSO4 and a trace metal solution containing 0.3 µM Na2•EDTA, 0.2 µM FeSO4•7H2O, 0.009 µM ZnSO4•7H2O, 0.004 µM MnCl2•4H2O, 0.1 µM H3BO3, 0.02 µM CoCl2•6H2O, 0.002 µM NiCl2•6H2O, and 0.003 µM Na2MoO4•2H2O. Cells were grown in 12 L fermentation volumes as described for Methylocystis sp. Rockwell. Cells were harvested at A600 of 8–10, centrifuged at 8000 × g for 30 min, and flash frozen in liquid nitrogen. As in previous studies50,51, Methylomicrobium buryatense 5GB1C was grown in modified NMSA medium amended with 2.3 mM phosphate buffer, pH 6.8, 50 mM carbonate buffer, pH 9.5, 130 mM NaCl, 40 µM CuSO4, and a trace metals solution including 5.4 µM Na2•EDTA, 14.4 µM FeSO4•7H2O, 5.5 µM ZnSO4•7H2O, 0.3 µM MnCl2•4H2O, 1 µM H3BO3, 1.7 µM CoCl2•6H2O, 0.2 µM NiCl2•6H2O, and 0.4 µM Na2MoO4•2H2O. Growth and harvest were performed as described for Methylomicrobium alcaliphilum 20Z.
To purify pMMO48,51,52, frozen cells were thawed in a lysis buffer consisting of 25 mM PIPES, pH 7.2 and 250 mM NaCl, except for Methylomicrobium alcaliphilum 20Z which required an alternate lysis buffer consisting of 25 mM PIPES, pH 7.2 and 500 mM NaCl. Cells were lysed via sonication, and cell debris was removed via centrifugation at 12,000 × g for 1 h at 4 °C. The supernatant was harvested, and membranes were pelleted at 100,000 × g for 1 h at 4 °C and flash frozen in liquid nitrogen. pMMO was solubilized from the membrane using 1.2 mg n-dodecyl-β-d-maltopyranoside (DDM) (Anatrace) per 1 mg of crude protein for 1 h at 4 °C. The membrane was pelleted via centrifugation at 100,000 × g for 30 min at 4 °C, and the supernatant containing the detergent-bound, solubilized pMMO was collected for further purification. Methylococcus capsulatus (Bath) pMMO was purified using a Superdex 200 size exclusion chromatography column (GE Healthcare) equilibrated with 25 mM PIPES, pH 7.3, 250 mM NaCl, and 0.02% DDM. pMMOs from Methylocystis sp. strain Rockwell, Methylomicrobium alcaliphilum 20Z, and Methylomicrobium buryatense 5GB1C were purified using a 15Q anion exchange chromatography column (GE Healthcare) with a 50–800 mM NaCl gradient in 25 mM PIPES, pH 7.3, 0.02% DDM. pMMO-containing fractions were concentrated and buffer exchanged into 25 mM PIPES, pH 7.3, 250 mM NaCl, 0.02% DDM using an Amicon centrifugal concentrator (Millipore) with a 100 kDa MWCO.
Proteins were separated via reverse-phase HPLC on an Agilent 1100 HPLC using a 214TP54 analytical C4 column (Grace Vydac: 5 μm, 250 × 4.6 mm i.d.). Approximately 100 µL pMMO (at a concentration of 100 µM) was loaded onto the column, and the protein was eluted using a gradient from 100% A (63.75% formic acid, 10% acetonitrile, 5% i-PrOH) to 100% B (70% formic acid, 30% i-PrOH) over 60 min at 0.5 mL min−1. Then, 100 μL of approximately 100 μM protein (concentration calculated per pMMO protomer) was injected with each run. Elution fractions were collected every 0.25 mL. Samples were then diluted 1:8 with water and digested with 2 μg pepsin at 37 °C overnight (Promega). Samples were heated at 95 °C for 10 min to inactivate pepsin.
Copper-replete Methylosinus trichosporium OB3b was grown as described above. Cells were harvested near the late stages of logarithmic growth (an optical density at 600 nm of ~1). Whole cell lysate from Methylosinus trichosporium OB3b cells was obtained using an SDS lysis buffer containing 4% SDS, 100 mM HEPES, pH 7.2, 10 mM DTT, and EDTA-free cOmplete protease inhibitors (Roche) at a final volume of 1 tablet per 50 mL lysis buffer. Proteins were precipitated with 8 volumes of cold acetone and one volume of trichloroacetic acid overnight at −20 °C. After washing the pellet with ice-cold acetone, resulting proteins were resuspended in 50 μL deionized 8 M urea in 100 mM ammonium bicarbonate, pH 7.8, reduced with 10 mM DTT at 50 °C for 30 min, and cysteines were alkylated with 35 mM iodoacetamide in the dark for 30 min. The solution was then diluted to <2 M urea (final concentration) and 1 μg porcine trypsin (Thermo) was added prior to overnight incubation at 37 °C with shaking.
For both samples, the resulting peptides were desalted using solid phase extraction on a Pierce C18 Spin column and eluted in 80 μL of 80% acetonitrile in 0.1% formic acid. After lyophilization, peptides were reconstituted with 5% acetonitrile in 0.1% formic acid and injected onto a trap column (150 μm i.d. × 3 cm) coupled with a nanobore analytical column (75 μm i.d. × 15 cm, both ReproSil C18aq, 3 µm). Samples were separated using a linear gradient of solvent A (95% water, 5% acetonitrile, 0.1% formic acid) and solvent B (5% water, 95% acetonitrile, 0.1% formic acid) over 60 min. Mass spectrometry (MS) data were obtained on a Velos Orbitrap Elite (Thermo) mass spectrometer. Data were searched using Mascot (Matrix Science) 2.5 against the SwissProt database and results were reported at 1% false discovery rate in Scaffold 4.5 (Proteome Software).
There are no pre-existing protein families to which PmoD sequences belong, and pmoD genes are commonly unannotated or are annotated solely as a hypothetical membrane protein. Thus, the PmoDMettrDRAFT_0381 sequence was used as an initial seed for a protein BLAST search against the JGI/IMG genome database. The most divergent result was used as the seed for a second BLAST round, and this was repeated until no additional PmoD sequences were obtained. The amino acid sequences of these genes and their associated metadata were downloaded on 10 October 2017.
All putative PmoD sequences were aligned via MAFFT (L-INS-I mode)53. After alignment, sequences were trimmed. Truncated sequences with less than 150 amino acids (aa) were removed from the dataset, as were sequences with more than 250 aa that did not contain the two characteristic cysteines along with an N-terminal signal peptide sequence and a C-terminal transmembrane sequence. The trimmed dataset included 860 putative PmoD sequences, of which 98.5% were found in species encoding an enzyme homologous to pMMO and AMO. The metadata file was updated to remove trimmed sequences (Supplementary Data 2), and the trimmed dataset (Supplementary Data 3) was realigned via MAFFT and used to establish a HMM (Supplementary Data 4) via HMMER 3.154. PmoD subgroups were aligned against the profile HMM using HMMalign (Supplementary Figure 7) to visualize differences in subgroup sequence conservation.
A sequence similarity network was constructed using the EFI-EST tool55, with an E-value cutoff of 21 and condensation of identical sequences into a single node. The resulting network (Supplementary Data 5) was visualized in Cytoscape 3.5.156. As in previous studies30, additional metadata was added for each gene for visualization within the context of the sequence similarity network. This included genus (from which subgroups of methane- and ammonia-oxidizing bacteria were identified) and conservation of specific residues of interest using the alignment generated via hmmalign. For genome neighborhood analysis, hierarchical clustering of pmoD genes against traits (presence of genes annotated with specific PFAM and TIGRFAM families within 5 genes up- and downstream from pmoD-specific metal binding motifs) was carried out as described previously. The resulting dataset (Supplementary Data 2) was examined to establish 20 main operon groups (Supplementary Figure 8).
To analyze the co-occurrence of specific copper-related genes with CuMMO-encoding operons, lists of genomes containing genes corresponding to specific protein families were generated by analyzing all genomes in the JGI/IMG database. The a/pmoB genes (as a proxy for amo/pmo operons) were identified as belonging to the PF04744 family; sMMO-encoding operons were identified as members of the TIGR04550 family which corresponds to mmoD genes, not generally found in non-sMMO soluble diiron monooxygenase operons; mbn operons were identified via BLAST and manually curated using an E-value cutoff of 2057 using the Methylosinus trichosporium OB3b mbnB gene (MettrDRAFT_3422, corrected to remove the unconserved internal stop codon in the published genome)10; MopE/CorA homologs were identified using BLAST (against Methylomicrobium album BG8 corA gene (MetalDRAFT_3218) and Methylococcus capsulatus Bath mopE gene (MCA2589), since the existing profile HMM, which defines family PF11617, matches only a small region of the MopE/CorA sequence and thus lacks specificity beyond the putative metal binding site); csp1/2 genes were identified as belonging to the TIGR04401 family; csp3 genes were identified as belonging to the PF03860 family or were identified as one of the top 500 BLAST results for Methylosinus trichosporium OB3b csp3 gene (MettrDRAFT_1424), since some csp3 genes more closely related to the Methylosinus trichosporium OB3b gene were not annotated as DUF326 genes (Supplementary Discussion); pmoD genes were identified as described above. Lists of genomes containing each class of gene were extracted from the JGI/IMG gene metadata, and the co-occurrence of genes of various classes in a given genome was quantified.
The coordinates and structure factors for the PmoDMet49242_1452 structure have been deposited in the Protein Databank with accession code 6CPD. Other data are available from the corresponding author upon reasonable request.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This work was funded by the Department of Energy grant DE-SC0016284 (A.C.R.), National Institutes of Health grants GM118035 (A.C.R.), GM111097 (B.M.H.), DK068139 (T.L.S.), R01AT009143 (N.L.K.) and F32GM119191 (O.S.F.), and American Heart Association grant 14PRE20460104 (G.E.K.). We thank the staff of the GM/CA and LS-CAT beamlines of the Advanced Photon Source, which is a DOE Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11356. Use of the Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, is supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under Contract No. DE-AC02-76SF00515. The SSRL Structural Molecular Biology Program is supported by the DOE Office of Biological and Environmental Research, and by the National Institutes of Health, National Institute of General Medical Sciences (including P41GM103393). The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of NIGMS or NIH. Steven Almo and Anthony Gizzi at Albert Einstein College of Medicine produced the wild-type His6-PmoDMet49242_1452 soluble domain construct. The Quantitative Bio-element Imaging Center at Northwestern is supported by NASA Ames Research Center NNA06CB93G. The Proteomics Center for Excellence at Northwestern is supported by National Resource for Translational and Developmental Proteomics under NIH grant P41 GM108569 and NCI CCSG P30 CA060553. The Keck Biophysics Facility at Northwestern is supported in part by NCI CCSG P30 CA060553. We thank Laura M. K. Dassama and Thomas J. Lawton for discussions.