Direct Methane Oxidation by Copper- and Iron-Dependent Methane Monooxygenases

Abstract

Methane is a potent greenhouse gas that contributes significantly to climate change and is primarily regulated in Nature by methanotrophic bacteria, which consume methane gas as their source of energy and carbon, first by oxidizing it to methanol. The direct oxidation of methane to methanol is a chemically difficult transformation, accomplished in methanotrophs by complex methane monooxygenase (MMO) enzyme systems. These enzymes use iron or copper metallocofactors and have been the subject of detailed investigation. While the structure, function, and active site architecture of the copper-dependent particulate methane monooxygenase (pMMO) have been investigated extensively, its putative quaternary interactions, regulation, requisite cofactors, and mechanism remain enigmatic. The iron-dependent soluble methane monooxygenase (sMMO) has been characterized biochemically, structurally, spectroscopically, and, for the most part, mechanistically. Here, we review the history of MMO research, focusing on recent developments and providing an outlook for future directions of the field. Engineered biological catalysis systems and bioinspired synthetic catalysts may continue to emerge along with a deeper understanding of the molecular mechanisms of biological methane oxidation. Harnessing the power of these enzymes will necessitate combined efforts in biochemistry, structural biology, inorganic chemistry, microbiology, computational biology, and engineering.

Graphical Abstract

1. INTRODUCTION

Methane is the second most abundant greenhouse gas next to carbon dioxide and has a global warming potential 84 times that of carbon dioxide over a 20-year period.¹ Atmospheric methane levels have increased rapidly in recent years with global methane emissions for 2008–2017 being 576 teragrams (Tg) yr⁻¹ (1 Tg = 1 million metric tons), exceeding those of the previous decade by 29 Tg yr⁻¹.² The largest yearly increase in atmospheric methane since recording began in 1983 was ~17 ppb in 2021.^3,4 Approximately 60% of these record methane emissions are anthropogenic, attributable to fossil fuel production and use, livestock, rice cultivation, landfills and wastewater, and biomass burning.^2,4 Of particularly high profile are frequent instances of methane leakage from oil and natural gas (composed primarily of methane) harvesting and handling systems. These increases put Earth on track for global temperature increases of >3 °C by the end of the century.² Methane is removed from the atmosphere primarily via reaction with hydroxyl radicals to form carbon dioxide and water. Because of its short perturbation lifetime (how long it takes to decay back to the original level after an emissions increase) of ~12 years, reducing methane emissions would have an immediate impact on climate change.^1,5,6

Conversion of methane to liquid fuels and chemicals would couple mitigating climate change with meeting rising energy demands, but gas-to-liquid conversion processes require steam reforming of methane to syngas (a mixture of carbon monoxide and hydrogen) followed by conversion to methanol or long chain hydrocarbons via Fischer-Tropsch synthesis. These indirect, technically demanding processes are carried out in large scale facilities and entail significant capital and operating expenses.^7,8 Direct conversion of methane to methanol is highly desirable, since methanol is used to generate the gasoline additive methyl tert-butyl ether, for substitution into the gasoline pool and as a feedstock for production of olefins, formaldehyde, and acetic acid.⁹ However, development of high yield homogeneous or heterogeneous catalysts for direct methane conversion is challenging for two reasons.¹⁰ First, methane has an unusually high C–H bond strength of 105 kcal/mol, rendering it less reactive than other alkanes.¹¹ Second, methanol is more reactive than methane and thus prone to further oxidation to CO₂. As such, direct methane oxidation has been referred to as one of the “Holy Grails” of catalysis.¹²

An alternative approach to homogeneous and heterogeneous catalysis is biological catalysis using microorganisms or their isolated enzymes to oxidize methane to methanol under ambient conditions. Microbial oxidation of methane occurs in both aerobic and anaerobic environments. Aerobic methane oxidation is performed by methanotrophs,¹³ bacteria that consume ~30 Tg yr⁻¹ of atmospheric methane.^2,14 Methanotrophs convert methane to methanol in the first step of their metabolic pathway using methane monooxygenase (MMO) enzymes, which react with methane and dioxygen to form methanol and water. Two evolutionarily distinct MMOs can catalyze this chemically difficult reaction: a soluble enzyme (sMMO) that uses a dinuclear iron catalytic site and a membrane-bound or particulate enzyme (pMMO) whose activity is dependent on copper.^15–17 In between aerobic and anaerobic methane oxidation is “intra-aerobic” methane oxidation carried out by the bacterial phylum NC10. These bacteria couple oxygen generation by nitrite reduction to methane oxidation by the pMMO system.^18,19 Finally, entirely anaerobic methane oxidation occurs in anaerobic methanotrophic archaea (ANME) via reverse methanogenesis with sulfate, nitrate, or metal ions as electron acceptors.²⁰ In contrast to aerobic methanotrophs, ANME and NC10 bacteria have not been isolated in pure culture, precluding biochemical studies. These anaerobic methane-oxidizing microbes also play a major role in offsetting methane emissions from soil and marine environments.^21,22

In this review, we focus on the enzymatic oxidation of methane by aerobic methanotrophs. Recent reviews have addressed methanotroph physiology, engineering, and applications^23–27 and the biochemistry, structure, and spectroscopy of MMOs.^{15–17,28,29} Here we address both the biology and chemistry of MMOs, spanning the history of MMO research, while highlighting recent developments and providing an outlook on unresolved questions. Progress toward understanding the molecular complexity of MMOs has required the use of a diverse scientific toolbox, involving methods in biochemistry, molecular biology, computational inorganic chemistry, spectroscopy, and structural biology. Structural approaches, in particular, have paved the way toward understanding biological methane oxidation, with recent applications of state-of-the-art methods, including cryoelectron microscopy (cryoEM) and X-ray free electron laser (XFEL) crystallography. We first address the ecology and biology of methanotrophs, focusing on the central role of copper in their physiology. We then review pMMO, addressing its molecular structure, metal centers, activity, active site, and proposed protein interaction partners. Finally, we discuss the structure, mechanism, and protein component interactions of sMMO. We also cover recent progress toward engineering both MMOs, which will be essential to their deployment in climate bioremediation and biological gas-to-liquid conversion processes.

2. BIOLOGY OF METHANOTROPHS

2.1. Taxonomy and Metabolism

Methanotrophs are gram-negative bacteria that live on methane gas as their source of carbon and energy. They were first reported in 1906,³⁰ but initial characterization did not happen until 50 years later with the isolation of Pseudomonas (Methylomonas) methanica,³¹ Methanomonas methanoooxidans,^32,33 and Methylococcus capsulatus,³⁴ which would become a workhorse strain for studies of methanotroph biochemistry. It was also established early on that the oxygen atom in methanol derives from dioxygen,^35,36 setting the stage for studies of MMO reaction chemistry. Methanotrophs were subsequently classified into types I, II, and X (a subset of type I) on the basis of their metabolic pathways, membrane lipid contents, cell morphologies, 16s rRNA sequences, and genomic characteristics,^37,38 with multiple revisions over the years.^39–41 All methanotrophs were long thought to be obligate, meaning that they can only live on one-carbon sources (primarily methane, but also possibly methanol, formate, formaldehyde and methylamines), but facultative methanotrophs that utilize multicarbon substrates such as acetate or ethanol have been isolated and characterized.^42,43

Another important early observation was the presence of prominent intracytoplasmic membranes (ICMs), which were also used for classification, with the type I methanotrophs exhibiting membranes shaped like vesicular discs (Figure 1a) and the type II methanotrophs exhibiting paired membranes around the cell perimeter^37,44 (Figure 1b). Thermoacidophilic methanotrophs, referred to as group III, were discovered much later^45–47 and have been the subject of much interest due to their growth requirement for rare earth elements.⁴⁸ The type I and type II methanotrophs are synonymous with the Gammaproteobacteria and Alphaproteobacteria classes of the Proteobacteria phylum, respectively, while the type III methanotrophs belong to the Methylacidiphilae class of the Verruocmicrobia phylum.^49,50 Methanotrophs are found in diverse environments, including soil, rice paddies, freshwater lakes, oceans, tundra wetlands, landfills, and volcanic mudpots.^13,51,52

Figure 1.

Micrographs of type I and type II methanotrophs. (a) Section of a type I Methylococcus strain magnified ×80,000. (b) Section of a type II Methylosinus strain magnified ×80,000. Adapted with permission from ref 37, copyright 1970, Society for General Microbiology.

The first step in methanotroph metabolism is the oxidation of methane to methanol by MMOs. Methanol is then oxidized to formaldehyde by methanol dehydrogenase (MDH). The next steps diverge depending on the type of methanotroph. In Gammaproteobacteria, carbon is assimilated at the stage of formaldehyde by the ribulose monophosphate pathway, whereas in Alphaproteobacteria, carbon is assimilated as formate via the serine pathway^39,53,54 (Figure 2). Verrucomicrobial methanotrophs fix CO₂ using the Calvin-Benson-Bassham cycle.^55,56 The proteobacterial assimilatory pathways have been targeted for metabolic engineering to produce a range of fuels and chemicals, including lactate and 2,3-butanediol (reviewed in refs 23 and 26). However, further advances will require increased rates of methane conversion to methanol,²⁸ which cannot be accomplished without a detailed understanding of MMO chemistry and regulation.

Figure 2.

Methanotroph metabolic pathways. The pMMO trimer is colored by protomer, showing its C3 symmetrical organization. sMMO is colored by subunits that comprise the complex, along with the MMO regulatory protein B (MMOB) bound (indigo) on the front and back (dashed line) sides of sMMO. MDH, methanol dehydrogenase; ICMs, intracytoplasmic membranes; PFLAs, phospholipid fatty acids; RuMP, ribulose monophosphate; CBB, Calvin-Benson-Bassham; H₄MPT, tetrahydromethanopterin pathway; H₄F, tetrahydrofolate pathway; FDH, formate dehydrogenase; SHMT, serine hydroxymethyltransferase; Hps, hexulose 6-phosphate synthase; RuBisCo, ribulose 1,5-bisphosphate carboxylase.

2.2. Copper Acquisition

2.2.1. Methanobactins.

As a required cofactor for pMMO activity^57–60 and an inducer of ICM formation,^61–63 copper is central to methanotroph physiology. Methanotrophs have several specialized copper acquisition systems. Some methanotrophs produce natural products called methanobactins (Mbns) under conditions of copper starvation.^64–66 Mbns are ribosomally synthesized, post-translationally modified peptide natural products that bind Cu(I) with particularly high affinity. The copper binding site consists of two nitrogen and two sulfur ligands provided by nitrogen-containing heterocycles and neighboring thioamide groups (Figure 3). Genes encoding the Mbn precursor peptide MbnA, biosynthetic enzymes, transporters, and other associated proteins are found in Mbn operons,⁶⁷ which are coregulated with the genes encoding sMMO.⁶⁸ All Mbn operons encode the MbnB/MbnC heterodimeric complex that uses a mixed valent Fe(II)Fe(III) site in MbnB to convert two cysteines in MbnA to oxazolone/thioamide groups.^69,70 Additional modifying enzymes present in some Mbn operons include the aminotransferase MbnN,⁷¹ a predicted flavin-dependent oxidoreductase, MbnF, a predicted sulfotransferase, MbnS, a predicted TauD-like nonheme iron enzyme, MbnD, and a protein related to MbnB called MbnX. Variations in MbnA sequences combined with the presence of different modifying enzymes lead to a diversity of Mbn structures (Figure 3).^72–76 Notably, Mbn operons are also found in a wide range of non-methanotrophic bacteria, suggesting a broader function in bacterial metal homeostasis.

Figure 3.

Structures of methanobactin from Methylosinus trichosporium OB3b and Methylosinus sp. LW3. The oxazolone moieties are highlighted in purple, and the pyrazinedione group is shown in green.

Under copper starvation conditions, methanotrophs secrete the apo (metal free) form of Mbn, which is then reinternalized in its copper-bound form (Figure 4).⁷⁷ Addition of copper-loaded Mbn to methanotrophs can promote methane oxidation activity and the copper switch between sMMO and pMMO (section 2.3).^78,79 Due to its high affinity for Cu(I), Mbn not only binds copper in solution but can also extract copper from mineral sources or glass.^78,80,81 The mechanism of Mbn secretion has not been established but is proposed to involve MbnM, a member of the multidrug and toxic compound extrusion family.^67,82 Uptake of copper-loaded Mbn is an active transport process mediated by the outer membrane TonB-dependent transporter MbnT, which is encoded both within Mbn operons and elsewhere in the methanotroph genomes.^82–85 After intact copper-loaded Mbn enters the cell,⁷⁷ it may interact with periplasmic proteins such as MbnE⁸² or MbnP, followed by import to the cytoplasm, perhaps by ABC transporters (Figure 4). MbnP was recently shown to bind Cu(I) using a kynurenine residue that is generated by the diheme enzyme MbnH.^86–88 The MbnP and MbnH genes are typically found adjacent to genes encoding MbnT. It is not known how copper is then delivered to pMMO or cytoplasmic cellular targets, including transcription factors.

Figure 4.

Model for copper homeostasis in the type II methanotroph Methylosinus trichosporium OB3b.

2.2.2. MopE and Csp Proteins.

Not all methanotrophs possess the ability to manufacture Mbn. There is evidence that methanotrophs can take up non-native Mbns, but so far, this Mbn piracy only involves other Mbn producers.^82,84,85 Some methanotrophs instead produce copper-binding proteins belonging the MopE/CorA family. The M. capsulatus (Bath) MopE protein is truncated and modified to contain a copper-binding kynurenine residue (MopE*) and then secreted. The surface-associated CorA from Methylomicrobium album BG8 also binds Cu(I) with kynurenine.^89–91 MopE* and CorA differ in overall structure and in the details of copper coordination from MbnP, which also has a kynurenine ligand.⁸⁷ In MopE* and CorA, the Cu(I) is ligated by two histidines, a kynurenine, and a water molecule, whereas the Cu(I) in MbnP is coordinated by one histidine, one methionine, a kynurenine, and a water molecule. Copper downregulates expression of MopE, CorA, and a Methylotuvimicrobium alcaliphilum comb. nov. 20Z (20Z) homolog, suggesting that these proteins function in copper acquisition.^91–93 How copper bound to these proteins is mobilized remains unclear.

Finally, members of the copper storage protein (Csp) family have been proposed to play a role in methanotroph copper handling.^94,95 The M. trichosporium OB3b Csp1 and Csp2 proteins are predicted to be secreted from the cytoplasm to the periplasm in a copper-bound form, which for Csp1 includes binding 13 Cu(I) ions using primarily cysteine residues housed in the interior of a four-helix bundle.⁹⁶ While the copper-binding properties of these proteins have been investigated in detail, their cellular localization in methanotrophs and evidence for a specific role in methane oxidation have not been reported. Disruption of the genes encoding both Csp1 and Csp2 led to a modest increase in sMMO iron-dependent activity, which could be consistent with a role in copper storage for pMMO.⁹⁶ Csp3, which does not have a signal sequence and thus should reside in the cytoplasm, binds 19 Cu(I) ions, also within a four-helix bundle, and is widespread in non-methanotrophic bacteria.^97,98

2.3. The Copper Switch

While the sMMO and pMMO genes were initially cloned in the late 1980s and early 1990s,^99,100 numerous genomes from all types of methanotrophs are now available.^23,101,102 The sMMO genes are encoded in the mmoXYBZDC operon, with mmoX, mmoY, and mmoZ encoding three subunits of the hydroxylase protein (MMOH), mmoB encoding the regulatory protein (MMOB), and mmoC encoding the reductase (MMOR) (Figure 5a).¹⁰³ The pMMO genes include pmoA, pmoB, and pmoC, encoding the PmoA, PmoB, and PmoC subunits of pMMO, respectively. Methanotroph genomes contain up to three copies of the pMMO genes, depending on the species,^{100,104–113} along with up to two additional copies of the pmoC gene sometimes referred to as PmoC singletons.^110,114 In alphaproteobacterial methanotrophs, the pmoD gene is found adjacent to the other genes (Figure 5b).¹¹⁵ Many methanotrophs, including the Verrucomicrobia, only contain the pMMO genes, while a few species from the Methylocella^42,116 and Methyloferula¹¹⁷ genera only possess the sMMO genes.⁴¹ Notably, the pmo operon is similar to that encoding ammonia monooxygenase (AMO),^118,119 the only enzyme besides pMMO and sMMO known to oxidize methane.^120,121 AMO converts ammonia to hydroxylamine in ammonia-oxidizing bacteria and ammonia-oxidizing archaea.^122–125 These nitrifying microbes also contribute to global warming by producing nitrous oxide, which is the third most important greenhouse gas next to carbon dioxide and methane.⁵

Figure 5.

MMO operons. The operons encoding sMMO in (a) M. capsulatus (Bath) and (b) M. trichosporium OB3b and pMMO in (c) M. capsulatus (Bath) and (d) M. trichosporium OB3b are shown.

A large subset of methanotrophs encodes both sMMO and pMMO in their genomes and can switch between them depending on copper-to-biomass ratios.⁴¹ This “copper switch” was discovered ~40 years ago with the observation that MMO activity was differentially associated with the membrane (particulate) or soluble fractions as a function of copper availability and that copper and particulate fraction activity are associated with ICM formation (Figure 1).^61–63 In these methanotroph strains, sMMO is prevalent at copper concentrations <1 μM, and pMMO becomes predominant at copper concentrations >5 μM. The copper-induced biogenesis of ICMs is not well understood despite their being imaged extensively by electron^{63,68,126,127} and fluorescence¹²⁸ micros-copies as well as cryoelectron tomography (cryoET).^129,130 These imaging studies indicate that the ICMs are continuous with the cytoplasmic inner membrane and form by invagination of this membrane.^128,130,131

In the well-studied M. capsulatus (Bath) and M. trichosporium OB3b strains, transcription of the sMMO genes is downregulated by copper.¹³² While the copper switch has been referred to as “reciprocal regulation”, pMMO is in fact expressed constitutively^133–135 and only mildly upregulated in the presence of copper.^132,136 Recent time-dependent qRT-PCR data showed less than an order of magnitude of upregulation of pMMO expression over 24 h of exposure to copper while sMMO expression is downregulated by 2–3 orders of magnitude within 24 h of exposure to copper.⁶⁸ Some studies have reported a more pronounced increase in pMMO expression,^80,137 but the consensus seems to be mild upregulation. This constitutive expression of pMMO raises the questions of whether it is actually present under low copper conditions and, if so, where it is localized and whether it contains copper.

It remains unclear how copper mediates the differential expression of the two MMOs. Several proteins encoded in the mmo operon, including the transcription factor MMOR and the GroEL homolog MmoG (Figure 5a), are essential for sMMO expression.^135,138 A two component system found in M. capsulatus (Bath) (Figure 5a), MmoQ/MmoS, may also play a role in sMMO regulation. Of these four proteins, only the soluble sensor domain of MmoS has been biochemically characterized, and it does not bind copper.¹³⁹ No regulatory factors for pMMO have been identified. The MMOD protein, which forms a complex with and inhibits the sMMO MMOH component (section 4.1),^140,141 has been proposed to bind copper and then to both repress pMMO expression and upregulate sMMO expression.^137,142 This model is based on characterization of an M. trichosporium OB3b mutant in which mmoXYBZD and the first three codons of mmoC are deleted (SMDM mutant).¹⁴³ For this mutant, pmoA expression decreases in the presence of copper as opposed to increasing in the wildtype strain. Since mmoD is the only disrupted gene in the SMDM mutant with an unclear function, it was suggested to mediate the copper switch.¹³⁷ However, mmoD is regulated with the rest of the sMMO genes, which is inconsistent with a regulatory role. Also incompatible with this model, MMOD has no DNA binding or metal binding motifs,¹⁴¹ does not bind copper, and does not bind to a heparin column,⁶⁸ often used as a diagnostic for DNA binding.

Several other strains of M. trichosporium OB3b have broken copper switches in that they constitutively express sMMO.¹²⁷ These mutants were generated by treatment with the mutagen dichloromethane and exhibit reduced intracellular copper levels.¹⁴⁴ Further studies of one of these mutants, the PP358 strain, showed that copper neither downregulates sMMO nor stimulates ICM formation. The PP358 genome has been sequenced, and of potential relevance to the copper switch, a frameshift deletion in the copD gene was detected.⁶⁸ The copD gene neighbors (Figure 5b) and is coregulated with the pmo genes in M. trichosporium OB3b. Since CopDs are putative copper importers,^145–148 a CopD disruption in Ms. trichosporium OB3b could prevent copper from reaching transcription factors in the cytoplasm (Figure 4). However, disruption of copD and the neighboring copC gene, which encodes a periplasmic copper binding protein,¹⁴⁹ does not affect the copper switch when tested at copper concentrations of 0 and 1 μM.¹⁵⁰ It is not known whether a partial deletion in copD, as found in the PP358 strain,⁶⁸ would have the same lack of phenotype. It is also possible that a phenotype would be observed using different conditions and copper concentrations.

3. PARTICULATE METHANE MONOOXYGENASE

3.1. Enzyme Structure

pMMO comprises three subunits, PmoB, PmoA, and PmoC, arranged in an α₃β₃γ₃ trimer (Figure 6a, b). All structurally characterized pMMOs, which include those from M. capsulatus (Bath),^60,151 M. trichosporium OB3b,¹⁵² Methylocystis species strain (sp.) M,¹⁵³ Methylocystis sp. Rockwell,^59,60 and M. alcaliphilum 20Z^60,154 (Table 1), form this trimer, and dissociation of the subunits or alternative oligomerization states have not been observed. One third of this trimer is typically referred to as the pMMO protomer. PmoB (42 kDa) consists of an N-terminal cupredoxin-like domain, two transmembrane helices, and a C-terminal cupredoxin-like domain. All PmoB subunits are predicted to have this architecture, although the related AmoB from the archaeal AMO system only contains one cupredoxin-like domain followed by a single transmembrane helix.^155–157 The cupredoxin-like domains face the periplasm and constitute the only soluble regions of pMMO. PmoA (24 kDa) comprises seven transmembrane helices, along with a small β hairpin that protrudes into the PmoB periplasmic domain. PmoA has a similar fold to the S components of bacterial energy-coupling factor (ECF) ABC transporters, which are responsible for uptake of vitamins such as riboflavin, thiamin, and biotin.^158,159 However, PmoA lacks a pocket equivalent to the ligand binding site in the S components.

Figure 6.

Trimeric structure of pMMO. (a) CryoEM structure of M. capsulatus (Bath) pMMO in native lipid nanodiscs (PDB ID: 7S4H). One protomer comprising PmoB (dark purple), PmoA (wheat), and PmoC (light purple) is highlighted. (b) CryoEM structure of M. sp. Rockwell pMMO in POPC nanodiscs (PDB ID: 7S4M). One protomer comprising PmoB (dark blue), PmoA (blue), PmoC (sky blue), and helix X (yellow) is highlighted.

Table 1.

pMMO Structures

	Resolution (Å)	PDB code
X-ray
M. capsulatus (Bath) pMMO	2.80	1YEW
M. capsulatus (Bath) pMMO	2.80	3RGBa
M. trichosporium OB3b pMMO	3.90	3CHX
M. sp. M pMMO	2.68	3RFR
M. sp. Rockwell pMMO	2.59	4PHZ
M. sp. Rockwell pMMO Cu(II) soaked	3.15	4PI0
M. sp. Rockwell pMMO Zn(II) soaked	3.33	4PI2
M. alcaliphilum 20Z pMMO	2.70	6CXH
CryoEM
M. capsulatus (Bath) pMMO in native lipid nanodisc	2.14	7S4H
M. capsulatus (Bath) pMMO in native lipid nanodisc	2.16	7S4J
M. capsulatus (Bath) pMMO in native lipid nanodisc	2.26	7S4I
M. capsulatus (Bath) pMMO in native lipid nanodisc	2.34	7S4K
M. capsulatus (Bath) pMMO in native lipid nanodisc + CN	3.65	7T4O
M. capsulatus (Bath) pMMO in native lipid nanodisc + CN and Cu	3.62	7T4P
M. capsulatus (Bath) pMMO in DDM	2.60	7EV9
M. capsulatus (Bath) pMMO in native lipid nanodisc + CN	3.21	8SR5
M. capsulatus (Bath) pMMO in native lipid nanodisc + CN and Cu	3.12	8SR4
M. capsulatus (Bath) pMMO in native lipid nanodisc + 20x TFE	2.19	8OYI
M. capsulatus (Bath) pMMO in native lipid nanodisc xlinked + 20x TFE	2.16	8SQW
M. capsulatus (Bath) pMMO in native lipid nanodisc + 20x TFB	2.36	8SR2
M. capsulatus (Bath) pMMO in native lipid nanodisc xlinked + 20x TFB	2.18	8SR1
M. sp. Rockwell pMMO in POPC nanodisc	2.42	7S4M
M. alcaliphilum 20Z pMMO in POPC nanodisc	2.46	7S4L
CryoET
M. capsulatus (Bath) pMMO	4.80	7YZY

^aPDB 3RGB is an improved version of structure 1YEW and should be used as the M. capsulatus (Bath) pMMO model; 1YEW is obsolete.

PmoC (22 kDa) consists of six transmembrane helices. Part of PmoC, spanning residues 225–253 in M. capsulatus (Bath) pMMO, is unmodeled in all the crystal structures due to a lack of electron density (Figure 7a). This region was finally resolved in the high resolution (up to 2.14 Å) cryoelectron microscopy (cryoEM) structures of pMMO reconstituted into nanodiscs (phospholipid bilayer discs surrounded by a membrane scaffold protein belt)¹⁶⁰ with native methanotroph lipids (Figure 7b).⁶⁰ These residues, which correspond to the most highly conserved part of the PmoC sequence, face the interior of the trimer and are stabilized by interactions with phospholipids. In the 2.6 Å resolution cryoEM structure of M. capsulatus (Bath) pMMO in n-dodecyl-β-d-maltoside (DDM) detergent, only PmoC residues 108–157 and 258–286 were modeled (Figure 7c),¹⁶¹ providing a less complete model than the crystal structures. The presence of lipids also stabilizes PmoA residues 192–212 (M. capsulatus (Bath) numbering).⁶⁰ These residues were not modeled in the M. capsulatus (Bath) crystal structure¹⁵¹ or a cryoEM structure of M. capsulatus (Bath) pMMO in DDM.¹⁶¹

Figure 7.

Structures of M. capsulatus (Bath) pMMO protomers showing PmoC (light purple), PmoA (wheat), PmoB (dark purple), copper ions (cyan), and zinc ions (gray) as modeled. (a) Crystal structure of pMMO showing PmoC and PmoA subunits with missing regions (PDB ID: 3RGB). (b) CryoEM structure of pMMO in native lipid nanodiscs showing the stabilized PmoC and PmoA architectures (PDB ID: 7S4H). (c) CryoEM structure of pMMO in detergent with perturbed PmoC and PmoA subunits (PDB ID: 7EV9).

In the crystal and cryoEM structures of pMMOs from the Alphaproteobacteria (M. trichosporium OB3b,¹⁵² M. sp. M,¹⁵³ M. sp. Rockwell),^59,60 strong density corresponding to an unidentified helix (helix X) is observed adjacent to a large groove in the surface of PmoC (Figure 6b). While helix X neighbors the PmoC N-terminus, ~15 residues of which are not modeled, its position and length are not consistent with it being connected to PmoC. Helix X could not be identified using mass spectrometry⁵⁹ and has been modeled as up to 25 alanine residues, extending from the periplasm (N-terminus) toward the cytoplasm (C-terminus). In M. sp. Rockwell pMMO, lipids located between helix X and PmoC interact with two conserved arginine residues in PmoC, Arg 102 and Arg 171.^59,60 Since all pMMO samples for structural characterization have been isolated directly from methanotrophs, it is likely that helix X represents a biologically relevant interaction partner. One possibility is that helix X corresponds to the transmembrane helix of PmoD (section 3.6), but attempts to model its side chain density with the PmoD sequence have not been successful. Regardless of helix X’s identity, the deep groove in the surface of PmoC is striking and is a likely binding site for either a protein partner or a large ligand. For example, an unusually shaped cryoEM density in this groove has been suggested to correspond to a quinone.⁶⁰

Recent serial cryo-focused ion beam (cryoFIB) milling/scanning electron microscope (SEM) volume imaging and cryoelectron tomography (cryoET) studies of pMMO in M. capsulatus (Bath) cells have revealed that the pMMO trimers assemble into higher order array structures.¹³⁰ The pMMO trimer in the intact cell was observed at 15 Å resolution in the subtomogram averaged map, and a 4.8 Å resolution map of a pMMO trimer surrounded by six lower resolution trimers was obtained by imaging isolated membranes.¹³⁰ The overall architecture agrees well with the crystal and cryoEM structures, and several intertrimer contacts involving the PmoB subunit were predicted from molecular dynamics simulations. Further studies, including simulations within a lipid bilayer, are needed to assess the molecular basis for array formation and may also shed light on the mechanisms of ICM biogenesis.

3.2. Metal Centers

While a 2007 study suggested that pMMO contains a catalytic diiron center similar to that in sMMO,¹⁶² iron detected in other preparations was attributed to heme from contaminating cytochromes, identifiable by optical, electron paramagnetic resonance (EPR), and X-ray absorption spectroscopies.¹⁶³ No further evidence for a diiron center has been obtained since the original report.¹⁶² Instead, pMMO is widely believed to contain copper active sites, consistent with observations that copper restores activity to pMMO samples that have been metal depleted by treatment with potassium cyanide.^57–60 The copper stoichiometries of purified pMMO from M. capsulatus (Bath) (the only pMMO studied by multiple research groups) over the past 20 years are in the range of either 2–3 copper ions or 13–15 copper ions per 100 kDa pMMO protomer.^41,164–166 As detailed below, 2–3 copper ions are consistent with the structural data obtained over the same time period, while the higher copper content, still favored by Chan and co-workers,^167,168 is not.

3.2.1. Metal Binding Sites in the PmoB Subunit.

The structures reveal two mononuclear copper centers in the PmoB subunit, in contrast to claims that PmoB is a “copper sponge” that can bind ~10 Cu(I) ions.^161,169,170 The first copper center, denoted the bis-His site, is coordinated by His48 and His72 (M. capsulatus (Bath) numbering) and is observed in the crystal¹⁵¹ and cryoEM^60,161 structures of M. capsulatus (Bath) pMMO (Figure 8a). This site is not present in the structures of pMMO from M. trichosporium OB3b,¹⁵² M. sp. M,¹⁵³ and M. sp. Rockwell^59,60 because His48 is replaced with asparagine in these alphaproteobacterial PmoB sequences (Figure 8b). Notably, His48 is conserved in M. alcaliphilum 20Z PmoB, but the site is devoid of metal.¹⁵⁴ Given that EPR spectra of pMMOs from alphaproteobacterial and gammaproteobacterial methanotrophs are virtually identical, this site in M. capsulatus (Bath) pMMO has been assigned as Cu(I).^171,172 Since this site is not conserved and not always occupied, it is unlikely to play a critical functional role.

Figure 8.

(a) The bis-His site in the PmoB subunit of M. capsulatus (Bath) (PDB ID: 7S4H) and (b) the corresponding residues in the PmoB subunit of M. sp. Rockwell pMMO (PDB ID: 7S4M).

The second site in the PmoB subunit, denoted Cu_B, has been the subject of much discussion in the literature. In the original M. capsulatus (Bath) pMMO crystal structure, this site was modeled with two copper ions, one coordinated by His137 and His139 and the other by the side chain of His33 as well as the amino terminal group of His33, which is the first residue in the PmoB subunit.¹⁵¹ The first 32 residues constitute the predicted signal sequence that is presumably removed upon export to the periplasm¹⁰⁰ and are not present in any pMMO preparation. The dicopper site model was influenced by extended X-ray absorption fine structure (EXAFS) data indicating the presence of a short (~2.5 Å) Cu–Cu interaction,^163,173 and a similar model was proposed for M. trichosporium OB3b pMMO¹⁵² and for one protomer of M. sp. M pMMO.¹⁵³ However, higher resolution, better quality crystallographic data obtained for pMMOs from M. sp. Rockwell⁵⁹ and M. alcaliphilum 20Z¹⁵⁴ (Table 1) were more consistent with a monocopper site in this location, as was quantum refinement¹⁷⁴ and high-energy-resolution fluorescence detected (HERFD) EXAFS analysis.¹⁷⁵ The crystal structure of the soluble portion of AmoB from the ammonia oxidizing archaeon Nitrosocaldus yellowstonii also revealed a single copper ion, although the amino terminal histidine was disordered in this structure.¹⁵⁶

The question of the Cu_B nuclearity was resolved through EPR studies of M. capsulatus (Bath) whole cells cultivated in the presence of ¹⁵N and ⁶³Cu.¹⁷¹ Consistent with prior whole cell EPR studies,^176,177 a single type 2 Cu(II) signal was observed with superhyperfine splitting indicative of four equatorial nitrogen ligands (Figure 9). Three of these four nitrogen ligands were assigned to histidine side chains on the basis of electron nuclear double resonance (ENDOR) spectroscopic analysis. The only location in the pMMO structure with three histidines positioned to coordinate copper is the Cu_B site, so this EPR signal is attributable to Cu_B, which must be a mononuclear Cu(II) site. The same results were obtained for M. sp. Rockwell pMMO.¹⁷² In addition, ¹⁷O and ¹H ENDOR data indicate the presence of an axially bound water molecule,^171,172 and ¹H ENDOR signals attributable to the bound amino group are observed.¹⁷² The EPR parameters of the Cu_B site are the same in whole cells, isolated membranes, and purified pMMO in detergent, bicelles, and nanodiscs prepared with both synthetic and native lipids.^{154,171,172,178}

Figure 9.

EPR spectra of pMMO from M. capsulatus (Bath) (adapted from refs 171 and 178). (a) pMMO in M. capsulatus (Bath) whole cells showing a Cu(II) EPR signature corresponding to the Cu_B(II) site. Asterisk indicates an organic radical signal. (b) pMMO in isolated membranes showing signals corresponding to the Cu_B(II) site and to the Cu_C/D(II) site, which exists as Cu(I) in vivo but is oxidized to Cu(II) upon membrane isolation. (c) pMMO solubilized in DDM and purified shows signals for the Cu_B(II) and Cu_C/D(II) sites. (d) pMMO in native lipid nanodiscs exhibits signals for the Cu_B(II) and Cu_D(II) sites as supported by cryoEM. (e) KCN-treated pMMO in native lipid nanodiscs shows an attenuated Cu(II) EPR spectrum with only a weak signal corresponding to partial loading of the Cu_B(II) site, consistent with metal depletion and supported by cryoEM. (f) KCN-treated, then copper-reloaded pMMO in native lipid nanodiscs shows recovered signals for the Cu_B(II) and Cu_D(II) sites, as supported by cryoEM.

Further support for a mononuclear Cu_B site is derived from native top-down mass spectrometry (nTDMS) of pMMO.¹⁷⁹ In these studies, M. alcaliphilum 20Z PmoB ejected from a detergent micelle exhibited a mass consistent with the presence of a single Cu(II) ion, as did M. sp. Rockwell PmoB ejected from micelles. In contrast to the typical metal analysis of pMMO by inductively coupled plasma mass spectrometry (ICP-MS) or optical emission spectroscopy (ICP-OES), nTDMS enables subunit-specific localization of bound metal ions. Finally, the significantly higher resolution cryoEM structures of pMMO from M. capsulatus (Bath) (Figure 10), M. sp. Rockwell, and M. alcaliphilum 20Z (Table 1) provided unequivocal evidence for a mononuclear Cu_B site.⁶⁰

Figure 10.

The Cu_B site in the cryoEM structure of M. capsulatus (Bath) pMMO in native lipid nanodiscs (PDB ID: 7S4H). The cryoEM density is shown as a transparent surface contoured at (a) 6σ and (b) 5σ.

3.2.2. Metal Binding Sites in the PmoC Subunit.

The pMMO crystal structures revealed one metal binding site in the PmoC subunit adjacent to the disordered region, denoted Cu_C and coordinated by Asp156, His160, and His173 (M. capsulatus (Bath) numbering) (Figure 11a). In the structures of pMMO from M. capsulatus (Bath)¹⁵¹ and M. sp. M,¹⁵³ this site is occupied by zinc, identified by analysis of anomalous diffraction data. Both of these pMMOs were crystallized in the presence of excess ZnSO₄, which not only occupies this site but also binds to the protein surface and mediates crystal lattice contacts. Zinc was not required for crystallization of M. trichosporium OB3b¹⁵² and M. sp. Rockwell⁵⁹ pMMOs, and the site is occupied by copper in these structures (Figure 11b). Soaking of M. sp. Rockwell crystals in CuSO₄ significantly increases the occupancy of the Cu_C site, while treatment with ZnSO₄ results in occupancy with zinc and ordering of 10 additional residues, including a glutamic acid coordinated to the zinc ion.⁵⁹

Figure 11.

Metal binding sites in the PmoC subunit. (a) Crystal structure of M. capsulatus (Bath) pMMO (PDB ID: 3RGB) showing a zinc ion in the Cu_C site. (b) Crystal structure of M. sp. Rockwell pMMO (PDB ID: 4PI0) showing a copper ion in the Cu_C site. (c) CryoEM structure of M. capsulatus (Bath) pMMO in native lipid nanodiscs (PDB ID: 7S4H) showing an empty Cu_C site and a copper ion in the Cu_D site.

Surprisingly, the Cu_C site is unoccupied in the cryoEM structures of M. capsulatus (Bath) pMMO in native lipid nanodiscs. Instead, another metal binding site is apparent ~5.7 Å from the Cu_C site location, with ligands Asn227, His231, and His245, all derived from the PmoC region that was not observed in the crystal structures (Figure 11c).⁶⁰ Unlike crystallography, there is no method to directly identify metal ions in cryoEM density maps, but structures of samples depleted of metals by potassium cyanide treatment⁵⁹ and then reloaded with CuSO₄ indicate that this site, denoted Cu_D, is indeed occupied by copper.^60,178 Instead of a metal ion in the Cu_C site, the M. capsulatus (Bath) pMMO cryoEM maps contain density assigned as a water molecule within hydrogen bonding distance of Cu_C ligands Asp156, His160, and His173 (Figure 11c). The Cu_C site is occupied in the cryoEM map of one sample of M. capsulatus (Bath) pMMO as well as in the cryoEM maps of pMMOs from M. alcaliphilum 20Z and M. sp. Rockwell. In these maps, the residues surrounding the Cu_D site are poorly ordered.⁶⁰ Thus, occupancy of Cu_C appears to correlate with disorder in the highly conserved region spanning residues 225–253 (M. capsulatus (Bath) numbering).

While whole cells exhibit a single Cu(II) EPR signal attributed to Cu_B (section 3.2.1), isolated membranes and purified pMMO exhibit a second Cu(II) EPR signal (Figure 9) that was initially assigned to Cu_C using Cu–Cu distances determined from double electron-electron resonance (DEER) spectroscopic analysis.^{171 15}N ENDOR experiments performed at fields where this signal does not overlap with that of Cu_B indicate that the Cu(II) ion is coordinated by two histidine ligands, consistent with assignment to Cu_C. An axial water molecule was also detected by ¹H ENDOR at a distance of ~3 Å from the Cu(II) ion. Similar results were obtained for pMMO reconstituted into 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) nanodiscs, although the axial water was not present.¹⁷²

The cryoEM structures⁶⁰ raised the question of whether this EPR signal might instead derive from the Cu_D site given that both Cu_C and Cu_D have two histidine nitrogen ligands and one oxygen ligand, with the only difference being the presence of asparagine instead of aspartic acid in Cu_D (Figure 11c). To address this question, parallel samples of M. capsulatus (Bath) pMMO in native lipid nanodiscs were interrogated by EPR and used for cryoEM structure determination. These enzymatically active samples exhibited the same two Cu(II) signals and showed occupancy only of Cu_D in the cryoEM structure.¹⁷⁸ Therefore, the second EPR signal is attributable to Cu_D in native nanodisc samples and perhaps in isolated membranes as well. It remains unclear whether this signal in detergent-solubilized pMMO arises from Cu_C, Cu_D, or some combination of the two sites, which are separated by ~5.7 Å. Regardless, since this EPR signal is not observed in whole cells, the corresponding site must be Cu(I) in vivo.

3.3. Enzymatic Activity

3.3.1. Delivery of Electrons.

The activity of pMMO is measured by monitoring either propylene epoxidation or methane oxidation. Propylene epoxidation, which may occur by a different mechanism than that of methane oxidation, is used for whole cell activity assays, as methanol is further metabolized by downstream enzymes and thus not detectable. Methane oxidation by isolated membranes and solubilized or purified pMMO are most accurately measured using ¹³C-labeled methane, which ensures that all detected methanol product derives from pMMO activity.¹⁵⁴ pMMO activity assays require a reductant, typically formate for whole cells, NADH or duroquinol for isolated membranes, and duroquinol for purified enzyme.^{173,180–182} Duroquinol is a synthetic analog of ubiquinol, and while ubiquinol is produced by methanotrophs,^183,184 duroquinol is not a native cofactor, despite being included in some computational studies.^185,186

Although these reductants are effective in vitro, the physiological source of electrons for pMMO remains unresolved, with several models under consideration. In the first model, NADH is proposed to reduce ubiquinol via a type 2 NADH:quinone oxidoreductase followed by the transfer of electrons from ubiquinol to pMMO.^136,187 This scenario, which is consistent with the use of NADH and duroquinol in vitro, is referred to as the “redox arm” model (Figure 12).¹⁸⁸ The second model, denoted “direct coupling”, involves transfer of electrons from MDH to pMMO via a cytochrome c electron shuttle (Figure 12).¹⁸⁹ A number of metabolic modeling studies have attempted to distinguish between these pathways by correlating growth parameters and other experimental data with flux balance analysis. Depending on the methanotroph species, these studies indicate that either pathway or a combination of the two, termed “uphill electron transfer”, could be operational.^{188,190–195}

Figure 12.

Proposed models for electron delivery to pMMO. The direct coupling model may also include the transfer of methanol from the pMMO active site to the MDH active site in a proposed supercomplex.

3.3.2. Activity of Isolated pMMO Preparations.

The specific activity of pMMO decreases upon isolation of the membranes and is significantly reduced or completely abrogated after detergent solubilization and purification (Table 2). The lack of activity upon purification for crystallization was suggested by Chan and co-workers to result from the loss of as many as 12 copper ions.^168,196 However, reconstitution of pMMO into bicelles (phospholipid bilayer discs surrounded by detergent)¹⁹⁷ recovered activity without altering the copper content or EPR spectroscopic signature, indicating that removal from the membrane, rather than massive copper loss, adversely affects pMMO activity.¹⁵⁴ Reconstitution into nanodiscs in the presence of copper also recovers activity.^60,179 The activity of M. capsulatus (Bath) pMMO nanodiscs was tested using several different lipids, including 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), POPC, and native lipids isolated directly from M. capsulatus (Bath) cells, of which the latter conferred the most activity (Table 2). The native lipids include a mixture of phosphatidylethanolamine (PE), phosphatidylcholine (PC), phosphatidylglycerol (PG), and cardiolipin (CL), as well as a significant fraction (~20%) of unidentified lipids.⁶⁰ It is not clear why the native lipids confer higher activity to pMMO in nanodiscs, as lipid densities in cryoEM structures of pMMO in native lipid nanodiscs resemble those in POPC nanodiscs in the same locations, suggesting that most of the lipids observed are PCs or native lipids that remain stably bound regardless of the peripheral lipid environment.⁶⁰

Table 2.

pMMO Activity Data

Sample	Methanotroph	Substrate	Reductant	Turnover number per protomer (s−1)	Specific activity (nmol mg total protein−1 min−1)	Refs
Cells producing pMMO	M. capsulatus (Bath)	propylene	formate	1.4−2.5a	167–300	136,162
	M. trichosporium OB3b	methane	NA	0.68–2.5b	82–300	28,357
Membrane-bound pMMO	M. capsulatus (Bath)	methane	NADH	0.083–0.146c	40–70	59,154
			duroquinol	0.025–0.042d	12−20	59,154
		propylene	NADH	0.044–0.246e	21–118	136,173
			duroquinol	0.033−0.179e	16−86	136,173
	M. sp. Rockwell	methane	NADH	0.017–0.024d,f	8–11.5	59,179
			duroquinol	0.004–0.006d,f	1.8–3	59,179
	M. alcaliphilum 20Z	methane	NADH	0.006c,f	3	154
			duroquinol	0	0	154
Purified pMMO in DDMg	M. capsulatus (Bath)	methane	NADH	0	0	154
			duroquinol	0.002h,i	1	60,154
	M. capsulatus (Bath)	propylene	NADH	0	0	136,173
			duroquinol	0.032–0.21i,j	18–126	136,173
	M. sp. Rockwell	methane	NADH	0	0	59
			duroquinol	0	0	59
	M. alcaliphilum 20Z	methane	NADH	0	0	154
			duroquinol	0	0	154
Purified pMMO in bicelles	M. capsulatus (Bath)	methane	NADH	0.009f,g	5.2	154
			duroquinol	0.006f,h	3.5	154
	M. alcaliphilum 20Z	methane	NADH	0	0	154
			duroquinol	0.007f,h	4.4	154
Purified pMMO in nanodiscsi
DMPCg	M. capsulatus (Bath)	methane	duroquinol	0.005h	3	60
POPCg	M. capsulatus (Bath)			0.009h	5.4	60
native lipids	M. capsulatus (Bath)			0.012h	7.2	60
POPCg	M. sp. Rockwell	methane	duroquinol	0.011g	6.6	179
POPCe	M. alcaliphilum 20Z	methane	duroquinol	0	0	60

^aCalculated from rate of propylene epoxidation monitored by gas chromatography (GC) and assuming pMMO is 20% of the total protein.

^bCalculated from rate of methane uptake and assuming that pMMO is 20% of the total protein.

^cCalculated from rate of conversion of 13CH4 to ¹³CH₃OH monitored by GC/mass spectrometry (GC/MS) and assuming that membrane-bound protein is 80% pMMO.

^dCalculated from rate of conversion of CH₄ to CH₃OH monitored by GC and assuming that membrane-bound protein is 80% pMMO. Values from ref 59 were converted to ¹³C values by applying a correction factor of 0.5.¹⁵⁴

^eCalculated from rate of propylene epoxidation monitored by GC and assuming that membrane-bound protein is 80% pMMO.

^fActivity assay was performed at 30 °C. All other activity assays were performed at 45 °C.

^gAbbreviations used: DDM, n-dodecyl-β-D-maltoside; DMPC, 1,2-dimyristoyl-sn-glycero-3-phosphocholine; POPC, 1-palmitoyl-2-oleoylphosphatidylcholine.

^hCalculated from rate of conversion of ¹³CH₄ to ¹³CH₃OH monitored by GC/MS.

ⁱThe samples used for the M. capsulatus (Bath) pMMO crystal structure determination were not assessed for methane oxidation activity and did not exhibit propylene epoxidation activity.¹⁵¹

^jCalculated from rate of propylene epoxidation monitored by GC.

The level of recovered activity in membrane mimetic systems approaches that of the isolated membranes but is still significantly less than that of whole cells (Table 2). In whole cells, pMMO is densely packed in the ICMs, forming hexagonal arrays (Figure 13).^130,198 These array structures can be recapitulated to some extent in nanodiscs by altering the reconstitution conditions, and these higher order pMMO nanodisc arrays exhibit increased activity compared to single particle nanodiscs.¹³⁰ Thus, pMMO activity may be enhanced in vivo by the properties of these ordered membrane structures and perhaps by protein-protein, protein-lipid, or protein-quinol interactions within these arrays. Overall, the issues with retaining activity have precluded using isolated pMMO for biotechnological applications, although one promising study demonstrated that a stable and reusable catalytic material could be generated by embedding pMMO-containing membranes in polymer hydrogels.¹⁹⁹

Figure 13.

Negative stain micrograph showing pMMO in isolated membranes. The inset shows a magnified view of the isolated membranes with a single pMMO trimer circled.

Besides methane and propylene, pMMO can oxidize C1–C5 n-alkanes and terminal alkenes to 2-alcohols and 1,2-epoxides, but it does not react with aromatic and cyclic hydrocarbons.^{180,181,200–202} pMMO can also oxidize ammonia, the substrate of the homologous enzyme AMO, to nitrite.²⁰³ Inhibitors of pMMO activity include metal chelating agents, alkynes, and excess copper.^57,203–206 Notably, inhibition by excess copper can be reversed by removal with potassium cyanide and reconstitution with stoichiometric amounts of copper.⁵⁷ Zinc is also an inhibitor, with excess zinc completely inhibiting activity²⁰⁷ and stoichiometric amounts leading to 40–60% inhibition.⁵⁹ Loading of apo pMMO with zinc almost completely abolishes activity in membranes. Zinc may occupy the copper active site and has also been proposed to interfere with proton transfer.⁵⁹

3.4. Assignment of the Active Site

3.4.1. Proposed Tricopper Site.

Models for the pMMO active site have evolved as new spectroscopic and structural data have been obtained. One model proposed prior to the first crystal structures and perpetuated in the literature involves a trinuclear copper center in the PmoA subunit.^165,196,208 However, no metal binding sites were observed in PmoA in any of the crystal structures. Three copper ions were modeled in PmoA in the cryoEM structure of M. capsulatus (Bath) pMMO in detergent,¹⁶¹ but the cryoEM structures of M. capsulatus (Bath) pMMO in native lipid nanodiscs clearly show that this region is occupied by a water molecule and a glutamate residue⁶⁰ (Figure 14). While the absence of the tricopper center in the crystal structures was ascribed to the loss of activity in the crystallized samples,²⁰⁸ its absence in the cryoEM structure of active pMMO in native nanodiscs⁶⁰ indicates that it is not a viable candidate for the active site.

Figure 14.

Proposed tricopper center in the PmoA subunit. (a) CryoEM structure of M. capsulatus (Bath) pMMO in native lipid nanodiscs showing the proposed tricopper center ligands and the corresponding density (PDB ID: 7S4H). (b) CryoEM structure of M. capsulatus (Bath) pMMO in DDM with copper ions and ligands shown as modeled with the corresponding density superimposed (PDB ID: 7EV9).

3.4.2. Cu_B Site.

The Cu_B site somewhat resembles the catalytic center of lytic polysaccharide monooxygenases (LPMOs), which hydroxylate and cleave glycosidic bonds of polysaccharides.^209,210 The LPMO active site consists of a Cu(I) ion coordinated by the side chain and amino group of an N-terminal histidine and the side chain of a second histidine, together called a histidine brace. By contrast, the Cu_B site binds Cu(II) with three, not two, histidines, and differs in some details of coordination. In particular, the LPMO non-amino terminal histidine coordinates copper with its ε nitrogen atom while one of the non-amino terminal histidine residues in Cu_B uses its δ nitrogen atom. Nevertheless, the ability of LPMOs to activate strong C–H bonds led to the suggestion that Cu_B could be the site of methane oxidation.²¹¹ In support of this model, a soluble fragment of PmoB comprising the two periplasmic domains connected by a flexible linker (spmoB)⁵⁷ or by monomers of apo ferritin¹⁸⁵ ostensibly exhibited methane oxidation activity. However, further investigation of spmoB and variants thereof indicated that the activity was not from the Cu_B site but instead was likely attributable to reactions of the reductant duroquinol with O₂.¹⁷¹ Consistent with this conclusion, the activity of the apo ferritin PmoB constructs was highly dependent on the presence of duroquinol.¹⁸⁵

A number of additional observations are incompatible with Cu_B being the active site. First, the three histidine ligands are not conserved in the PmoB sequences of verrucomicrobial pMMOs, which instead contain methionine, proline, and glycine at these positions.^113,212,213 Second, Cu_B is always present as Cu(II), even in whole cells,¹⁷¹ and is coordinatively saturated with four nitrogen ligands. Binding and activation of O₂ would require reduction and the presence of an open coordination site. Relatedly, there are members of the LPMO family that have saturated copper coordination and do not exhibit LPMO activity.²¹⁴ Third, Cu_B is exposed at the protein surface, and there is no obvious hydrophobic pocket for substrate binding. Finally, mutation of one of the Cu_B ligands in a related hydrocarbon monooxygenase from Mycobacterium strain NBB4 did not completely abolish activity.²¹⁵

3.4.3. Cu_C Site.

In contrast to Cu_B, several lines of evidence suggest Cu_C as the likely active site. First, all the ligands to the Cu_C site are strictly conserved, including in the verrucomicrobial PmoC sequences. Second, an increase in the methane oxidation activity of M. sp. Rockwell pMMO nanodiscs observed upon copper supplementation is correlated with increased copper in the PmoC subunit as measured by nTDMS. This experiment, while not specifically pinpointing Cu_C, demonstrated that copper bound to PmoC is critical for activity.¹⁷⁹ Third, the EPR signal attributed to Cu_C in purified pMMO is perturbed by the addition of ¹⁵N NO₂⁻, and ENDOR data are consistent with NO₂⁻ binding to Cu(II) via its oxygen atom(s).¹⁷¹ This finding is significant, as NO₂⁻ inhibits methane oxidation^216,217 and is therefore likely to bind at the active site. While there is no apparent substrate binding cavity near Cu_C in the crystal structures, a hydrophobic pocket adjacent to Cu_C and Cu_D is present in the cryoEM structures (section 3.4.4). Finally, mutation of any of the three residues corresponding to the Cu_C ligands in the M. strain NBB4 hydrocarbon monooxygenase completely abrogated activity.²¹⁵

3.4.4. Cu_D Site.

When the full architecture of the region adjacent to Cu_C and the presence of Cu_D were revealed by the cryoEM structures of active pMMO in native lipid nanodiscs (sections 3.1 and 3.2.2), the Cu_C active site model was revised. The cryoEM maps of active samples, including M. capsulatus (Bath) pMMO in both native and POPC nanodiscs, revealed an occupied Cu_D site whereas the maps of samples with no activity, including M. alcaliphilum 20Z and M. sp. Rockwell pMMOs in POPC nanodiscs, exhibit an occupied Cu_C site and disorder at the Cu_D site.⁶⁰ Thus, Cu_D occupancy appears to correlate with activity. Another key finding from the cryoEM structures in native nanodiscs is the presence of a hydrophobic cavity lined by residues from PmoA and PmoC, including three invariant phenylalanines from PmoC.⁶⁰ Prior to these structures, there was no sign of a potential substrate binding cavity in pMMO.

The possibility of a Cu_D active site was further investigated by parallel ENDOR and cryoEM studies of M. capsulatus (Bath) pMMO in native nanodiscs in the presence of the inhibitor 2,2,2-trifluorethanol (TFE).¹⁷⁸ Analysis of ¹⁹F ENDOR data revealed ¹⁹F couplings (Figure 15a) attributable to TFE interacting with the Cu_D site in an axial fashion with respect to the Cu_D ligand plane, with the fluorine-nuclei centroid ~5 Å away from the Cu(II) ion (Figure 15b). Modeling TFE bound with this Cu–F distance placed the TFE oxygen atom ~2 Å from Cu_D. CryoEM maps of the same samples showed new density connected to Cu_D, which was modeled well as TFE (Figure 15c). The average Cu–F distance is ~4.8 Å, consistent with the geometric information yielded by ENDOR analysis, and the TFE is situated in the aforementioned hydrophobic cavity, tilted axially out of plane with respect to the Cu_D-coordinating ligands. Similar experiments with 4,4,4-trifluorobutanol (TFB) showed ¹⁹F couplings to Cu_D via ENDOR with a larger density, modeled as TFB, connected to Cu_D in the cryoEM map. These combined data strongly support a model in which Cu_D and the surrounding cavity is the site of substrate binding and product formation. The possibility that Cu_C and Cu_D, separated by 5.7 Å, could be occupied by copper simultaneously in some form of pMMO remains open and is an important area for future investigation.

Figure 15.

Product analog binding at the Cu_D site. (a) Q-band ¹H/¹⁹F Mims pulsed ENDOR of M. capsulatus (Bath) pMMO in native lipid nanodiscs with (light purple) and without (dark purple) the addition of 20× TFE at g_|| = 2.14 (~11200 G). (b) Model for the binding of TFE to Cu(II) based on the ENDOR-derived Cu(II)–F distance of ~5 Å. (c) Model of TFE bound at the Cu_D site based on the 2.19 Å resolution cryoEM map of M. capsulatus (Bath) pMMO in native lipid nanodiscs with 20× TFE added (PDB ID: 8OYI).

3.5. Interaction with Methanol Dehydrogenase

MDHs are dimeric enzymes that use a pyrroloquinoline quinone (PQQ)/calcium ion cofactor to convert methanol to formaldehyde. The MxaFI MDHs consist of two subunits, MxaF (64 kDa), which houses the PQQ cofactor, and MxaI (8.5 kDa), of which the function is not known, arranged in an α₂β₂ dimer (Figure 16a).^218–220 A second type of MDH, XoxF, forms an α₂ dimer of a single subunit and utilizes lanthanide ions instead of calcium (Figure 16b).^221,222 In methanotrophs that possess both MDHs, expression is regulated by the presence of lanthanides, which repress transcription of MxaFI and activate transcription of XoxF.^223–226 The verrucomicrobial methanotroph Methylacidiphilum fumariolicum SolV only possesses XoxF, and the presence of lanthanides is essential for its growth.^45,48 Crystal structures of several methanotroph MDHs have been determined, including M. capsulatus (Bath) MxaFI,²²⁷ Methylotuvimicrobium buryatense 5GBC1 XoxF with lanthanum,²²⁸ and M. fumariolicum SolV XoxF in the presence of cerium,⁴⁸ europium,²²⁹ and neodymium.²³⁰

Figure 16.

General architecture of methanol dehydrogenases and their active sites. (a) The calcium-dependent M. capsulatus (Bath) MDH shown with one αβ protomer highlighted (PDB ID: 4TQO). The MxaF subunit is shown in purple, and the MxaI subunit is shown in wheat. The inset shows the calcium (green) and the PQQ cofactor (purple) binding site. (b) The lanthanide-dependent XoxF MDH from M. fumariolicum SolV shown with one subunit of the homodimer highlighted in green (PDB ID: 4MAE). The inset shows the cerium (silver) and PQQ cofactor (green) binding site.

Several lines of evidence suggest that pMMO interacts directly with MDH. First, MDH, despite being a periplasmic enzyme, is typically associated with the ICMs.^231–233 Second, transient interactions between M. capsulatus (Bath) pMMO and its cognate MxaFI as well as between M. buryatense 5GBC1 pMMO and its cognate XoxF have been detected by biolayer interferometry with K_D values of ~9 and ~50 μM, respectively.^227,228 An interaction between M. capsulatus (Bath) MxaFI and the spmoB protein was also detected, consistent with the location of MxaFI in the periplasm.²²⁷ Such interactions would facilitate channeling of the pMMO product methanol to the MDH active site and are consistent with the direct coupling model for electron transfer (section 3.3.1).

However, a stable pMMO-MDH complex has not been isolated by size exclusion chromatography or by reconstitution of purified proteins.^227,228 A putative supercomplex between M. capsulatus (Bath) pMMO and its cognate MxaFI was reported based on a 16 Å resolution 3D volume acquired by cryoEM, but three MxaFI monomers were fit to the density,²³⁴ inconsistent with the dimeric structure of M. capsulatus (Bath) MxaFI. More recently, attempts to reproduce this result or determine the high resolution structure of a pMMO-MDH complex using improved cryoEM technology have been unsuccessful. It may be that pMMO-MDH complexes can only assemble on the membrane in the context of the pMMO arrays present in cells.¹³⁰ On the basis of crystal packing interactions and the presence of multiple lysine residues, the small MxaI subunit was proposed to mediate interactions with negatively charged phospholipid headgroups in the membrane.²²⁷ This model is not generalizable to the XoxFs though, as these enzymes lack the second subunit. Nevertheless, support for the direct coupling electron transfer model, at least in type I methanotrophs,^{188,190,191,194} and longstanding evidence of MDH association with the membranes^231–233 underscore the importance of pMMO-MDH interactions as an area for future study.

3.6. The PmoD Protein

The PmoD and AmoD/AmoE proteins belong to a unique protein family found only in methanotrophs and ammonia-oxidizing bacteria, suggesting that they are functionally linked to pMMO and AMO.¹¹⁵ In type II methanotrophs, including the Methylosinus and Methylocystis genera, and in gammaproteobacterial ammonia oxidizers, the pmoD/amoD gene is located directly adjacent to pmoB (Figure 5a).^115,119,235 In betaproteobacterial ammonia oxidizers, amoE and amoD follow the pmo genes. Type I methanotrophs have genes encoding PmoD located elsewhere in the genome, typically adjacent to genes encoding multicopper oxidases or CopC proteins (Figure 5a).^115,235 Notably, the genomes of methanotrophs and ammonia oxidizers contain multiple copies (2–11) of pmoD/amoD genes. In support of a function related to pMMO, M. trichosporium OB3b pmoD expression is coregulated with that of the pMMO subunits.⁶⁸ Furthermore, its genetic disruption leads to a growth defect under pMMO-utilizing conditions, while growth under sMMO-utilizing (copper-starvation) conditions is not affected.¹¹⁵

PmoD proteins are predicted to comprise an N-terminal periplasmic domain followed by a transmembrane helix (Figure 4). The N-terminal domain of the PmoD protein encoded in the Mc. sp. Rockwell pmo operon has been biochemically and structurally characterized.^115,236 In the presence of copper, this domain forms a dimeric species with optical and EPR spectroscopic features (Figure 17a) characteristic of a Cu_A site.²³⁷ Mutagenesis data indicate that unlike typical Cu_A sites, the ligands derive from two monomers, resulting in a symmetric site with distinct electronic properties (Figure 17b). The crystal structure of the Cu_A-bridged dimer has not been determined, but that of a monomeric species¹¹⁵ reveals key differences in the regions that provide the ligands in typical Cu_A domains (Figure 17c).^238–240 The PmoD Cu_A site is also unusually unstable, decaying slowly to form two type 2 Cu(II) sites.²³⁶

Figure 17.

Structure of PmoD and a model for Cu_A site formation. (a) CW X-band (~9.5 GHz) EPR spectrum of the Cu_A of PmoD. Brackets indicate hyperfine splitting A_z (adapted from ref 115). (b) Crystal structure of the PmoD soluble domain from M. sp. Rockwell (PDB ID: 6CPD) showing potential Cu_A-forming residues. (c) Model of Cu_A site formation between two PmoD proteins.

Formation of the Cu_A site is associated with the presence of a Cx₇MxHx_nC motif, which is characteristic of PmoDs encoded within pmo operons. PmoD homologs encoded in different genomic neighborhoods contain a variety of other potential metal binding motifs and also bind copper but do not form Cu_A sites.¹¹⁵ Full-length PmoD, including the C-terminal transmembrane helix, has not been biochemically characterized, so it remains unclear whether the Cu_A site or any type of copper site forms when PmoD is embedded in the membrane. Further, it is not known whether in vitro copper binding or Cu_A formation is related to the growth phenotype upon disruption of the pmo operon copy of the PmoD gene in the M. trichosporium OB3b.¹¹⁵ While PmoD has been proposed to play a role in pMMO copper loading, catalytic activity, and/or stabilization,^115,236 further investigation is needed to elucidate its functional significance.

3.7. Mechanisms of Dioxygen and Methane C–H Bond Activation

Despite the continually evolving picture of the pMMO copper active site, computational chemists have attempted to elucidate its mechanism of O₂ activation, with a number of intermediates under consideration. Early studies by Yoshizawa and co-workers utilized the Cu_B site as a model, albeit with an oxygen atom from a nearby glutamic acid rather than the amino terminal group as a fourth ligand. Their calculated reaction pathways for M. capsulatus (Bath) pMMO involved conversion of a μ-η²:η²-peroxo-Cu(II)Cu(II) or a μ-η¹:η²-peroxo-Cu(I)-Cu(II) species to a reactive bis(μ-oxo)Cu(II)Cu(III) or (μ-oxo)(μ-hydroxo)Cu(II)Cu(III) species (Figure 18) capable of methane oxidation.^241–243 Formation of the latter species was proposed to occur via homolytic cleavage of the O–H bond in a nearby tyrosine residue, Tyr 374, followed by proton transfer to the μ-η²:η²-peroxo-Cu(II)Cu(II) core, yielding a μ-η¹:η² hydroperoxo-Cu(I)Cu(II) species that is converted to the (μ-oxo)(μ-hydroxo)-Cu(II)Cu(III) species.²⁴³ This mechanism was revisited more recently, this time suggesting that the nearby glutamic acid, Glu35, receives the proton from Tyr374, followed by transfer to the dicopper core.²⁴⁴ While Tyr374 is not strictly conserved, all the pMMO structures have a tyrosine near the Cu_B site. Since Cu_B is not dinuclear and no longer believed to be the site of methane oxidation (sections 3.2.1 and 3.4.2), these studies are likely not relevant. However, some of the proposed dicopper intermediates could be of interest in a scenario with Cu_C and Cu_D occupied simultaneously, assuming that a Cu–Cu distance significantly less than the 5.7 Å indicated by the cryoEM structures⁶⁰ could be achieved via conformational changes.

Figure 18.

Copper-oxygen species proposed on the basis of computational studies to mediate methane oxidation by pMMO.

The reactivity of O₂ with a monocopper center has also been investigated computationally. Using the bis-His site, which is not conserved and unlikely to bind O₂, as a model, Yoshizawa and co-workers suggested that a Cu(III)-O (Cu(II)-O^•) (Figure 18) species might be able to oxidize methane.²⁴¹ Later studies by Ryde and co-workers using the mononuclear Cu_B site also suggested that a Cu(III)-O species can activate the methane C–H bond,¹⁷⁴ though Cu(III) has yet to be detected in any biological system.^245–247 A mechanism that does not invoke Cu(III) was proposed recently by Wang and co-workers.¹⁸⁶ In this calculated mechanism, duroquinol binds at the Cu_C site, forming a Cu(II) duroquinol anion species. The duroquinol anion is then replaced by O₂, coupled with electron transfer from O₂ to the duroquinol anion to yield a Cu(II)-O₂^•− species and a duroquinol radical. A sequence of hydrogen atom abstraction and electron transfer steps involving a second duroquinol molecule then results in a Cu(II)-O reactive intermediate that reacts with methane. This mechanism is unlikely to be relevant to pMMO activity in vivo because duroquinol is not the physiological reductant of pMMO. Duroquinol is a synthetic analog of ubiquinol, which is too large to dock at the Cu_C site and thus could not participate in this mechanism. While the larger plastoquinol could be docked at the Cu_C site in this study, the structure used for docking lacked the crystallographically disordered region of PmoC and the Cu_D site, creating an artificial cavity.¹⁸⁶ A well-folded PmoC subunit would likely preclude the binding of quinols at the Cu_C or Cu_D sites. Another recent study simulated ubiquinol binding directly at the Cu_D site,²⁴⁸ which, while more biologically plausible than the binding of synthetic duroquinol at this site, might be precluded by amino acid side chains and lipids blocking access to this cavity.

Several studies have employed different substrates to experimentally address the mechanism of C–H activation by pMMO. The reaction of membrane-bound pMMO with chiral ethane gave an intramolecular kinetic isotope effect k_H/k_D of 5.2–5.5 and was completely stereoselective, eliminating a mechanism involving alkyl radicals or cations and instead suggesting a concerted mechanism with a pentacoordinate hydrocarbon intermediate.²⁴⁹ Enantioselective hydroxylation was also observed for other substrates such as n-pentane, n-butane, and alkenes, albeit with less stereoselectivity, especially for the alkenes, as compared to ethane.^{200,201,250–252} No ¹²C/¹³C carbon kinetic isotope effect was detected using propane as a substrate, consistent with a concerted mechanism.²⁵³ Such studies should be interpreted cautiously, since sMMO is known to react with different substrates via different mechanisms (section 4.5).

3.8. Overexpression and Engineering

Heterologous expression of pMMO has had limited success, with no laboratory ever obtaining expression of all three subunits in E. coli. Soluble proteins corresponding to one or both of the PmoB periplasmic domains have been expressed in E. coli, but these proteins require refolding or the presence of fusion proteins and do not exhibit methane oxidation activity (section 3.4.2).^{57,170,171,185} The periplasmic domain of AmoB from the ammonia-oxidizing archaeon N. yellowstonii was expressed solubly without fusion tags but did not exhibit methane oxidation activity.¹⁵⁶ There is one report of heterologous expression of M. trichosporium OB3b pMMO in Rhodococcus erythropolis LSSE8–1, but the whole cell activity is 2 orders of magnitude less than that of M. trichosporium OB3b, and the protein expression levels were not reported.^28,254 Initial steps toward expression of the pMMO genes in plants have been reported, but evidence for assembly of pMMO or activity was not obtained.²⁵⁵ Finally, the hydrocarbon monooxygenase from M. strain NBB4, which is related to pMMO and oxidizes C2–C4 alkanes, was expressed in Mycobacterium smegmatis, conferring ethane, propane, and butane monooxygenase activity²⁵⁶ and allowing interrogation of several site-specific mutants,²¹⁵ but further work with this system has not been reported. Thus, standard site-directed mutagenesis studies of pMMO have not been possible.

Another option for producing pMMO variants is genetic manipulation of native methanotrophs. Protocols for methanotroph gene disruption have been developed, providing insight into the functions of several proteins and facilitating metabolic engineering.^26,257 Efforts to alter pMMO specifically are complicated by the presence of multiple pmo operons in most methanotroph genomes. Genetic tools exist for M. buryatense 5GB1C, which contains a single pmo operon,^258,259 and site-directed mutagenesis should be possible in this strain. While growth under sMMO-utilizing conditions ought to be a viable strategy for obtaining pMMO variants, such efforts have not been successful. One possibility is that pMMO is still required for cell viability under copper-starvation conditions. In support of this idea, M. trichosporium OB3b pMMO is expressed constitutively and only mildly upregulated upon copper addition.⁶⁸ Another strategy for generating point mutants is CRISPR-Cas9 genome editing, which can be performed with ~10% efficiency in M. capsulatus (Bath).^257,260 The feasibility of this approach for site-directed mutagenesis of pMMO remains unproven, however.

Cell free protein synthesis (CFPS) represents a way to circumvent both difficulties with heterologous expression and the possibility that mutants with impaired pMMO viability will not grow under sMMO-utilizing conditions. In CFPS, the transcription and translation machinery is isolated from the cell,²⁶¹ obviating the need for a functional pMMO for methanotroph cell viability. In recent work, M. capsulatus (Bath) pMMO was expressed in an E. coli lysate system directly into POPC nanodiscs.²⁶² To generate the amino terminal histidine residue of the PmoB subunit, the native signal sequence was replaced by a SUMO fusion protein, and expression was conducted in the presence of SUMO protease. Remarkably, the pMMO trimer was assembled as demonstrated by negative stain EM and 2D class averaging. Activity assays on the cell-free reaction mixtures as well as on pMMO isolated from the mixture yielded no measurable methane oxidation, however. Nevertheless, this promising approach should be revisited as CFPS technology develops and more factors important for pMMO activity are elucidated.

4. SOLUBLE METHANE MONOOXYGENASE

4.1. Enzyme Structure

Three different proteins are required for methane oxidation by sMMO.^16,263,264 The diiron active site is located in the multisubunit hydroxylase protein (MMOH). A reductase, MMOR, transfers two electrons from NADH to the MMOH diiron site via its two cofactors, FAD and a [2Fe-2S] cluster. The third component, MMOB, binds to MMOH and significantly increases its activity, as evidenced by a 1000-fold increase in reaction rate with dioxygen and a 150-fold increase in turnover number.^265–267 A fourth protein encoded in the sMMO operon (Figure 5a), MMOD, inhibits sMMO activity.^140,268 As noted above (section 2.3), MMOD has also been proposed to function in the copper switch,^137,142 but its structure¹⁴¹ and biochemical properties⁶⁸ are not consistent with this role. Another proposed role is iron loading of MMOH, but MMOD instead prevents reconstitution of apo MMOH with iron¹⁴⁰ and reduces the rate of iron removal from MMOH.²⁶⁸

Components of sMMO from both M. capsulatus (Bath) and M. trichosporium OB3b have been structurally characterized (Table 3). The hydroxylase (MMOH), characterized first by crystallography^269–271 and visualized 30 years later by cryoEM,²⁷² comprises two copies each of the α, β, and γ subunits arranged in a 245 kDa α₂β₂γ₂ dimer (Figure 19a). The α and β subunits are primarily α-helical and form a dimeric heart-shaped structure, similar to that of the R2 subunit of ribonucleotide reductase.²⁷³ The α subunit houses the diiron center in a four-helix bundle formed by helices labeled B, C, E, and F. The N-terminus of the β subunit comprises a helix that docks on the α subunit followed by a loop region that connects to the rest of the subunit. The two γ subunits, also helical, are found on opposite sides of the dimeric structure. NMR structures of MMOB (Figure 19b)^274,275 and of the individual FAD/NADH binding and [2Fe-2S] cluster-containing domains of MMOR^276–278 (Figure 19c) have been determined as well. The N-terminal 35 residues of MMOB are disordered in the NMR structures but were shown through NMR²⁷⁴ and DEER spectroscopies²⁷⁹ to interact with MMOH.

Table 3.

sMMO Structures

M. capsulatus (Bath) MMOH (hydroxylase)	Resolution (Å)	PDB code
oxidized 4 °C	2.20	1MMO
oxidized	1.96	1FZ1
oxidized	1.70	1MTY
reduced in crystal	2.15	1FYZ
anaerobically grown reduced	2.40	1FZ5
mixed valent, reduced in crystal	2.15	1FZ2
anaerobically grown mixed valent	2.07	1FZ0
methanol soaked	2.05	1FZ6
ethanol soaked	1.96	1FZ7
Xe pressurized	3.30	1FZI
Xe pressurized	2.60	1FZH
dibromomethane grown	2.10	1FZ8
iodoethane grown	2.30	1FZ9
pH 8.5 soaked	2.38	1FZ4
pH 6.2 soaked	2.03	1FZ3
Mn(II) soaked	2.32	1XMF
apo (metal free)	2.10	1XMG
Co(II) reconstituted	2.32	1XMH
phenol soaked	1.96	1XU5
6-bromohexanol soaked	1.80	1XVB
8-bromooctanol soaked	2.00	1XVC
4-fluorophenol soaked	2.30	1XVD
3-bromo-3-butenol soaked	2.40	1XVE
chloropropanol soaked	2.00	1XVF
bromoethanol soaked	1.96	1XVG
bromophenol soaked	2.30	1XU3
cryoEM structure using graphene	2.40	7TC8
cryoEM structure using quantifoil	2.90	7TC9
M. trichosporium OB3b MMOH (hydroxylase)	Resolution (Å)	PDB code
oxidized	2.00	1MHY
oxidized	2.70	1MHZ
oxidized	1.52	6VK6
reduced in crystal	2.12	6VK7

MMOB, MMOR, and protein-protein complexes	Resolution (Å)	PDB code
M. capsulatus (Bath) MMOB NMR		1CKV
M. trichosporium OB3b MMOB NMR		2MOB
M. capsulatus (Bath) MMOR [2Fe-2S] domain NMR		1JQ4
M. capsulatus (Bath) MMOR FAD/NADH binding domain NMR		1TVC
M. sporium MMOR FAD/NADH binding domain	1.50	6L2U
M. capsulatus (Bath) MMOH-MMOB complex	2.90	4GAM
M. trichosporium OB3b MMOH-MMOB with benzoate	1.86	6VK5
M. trichosporium OB3b MMOH-MMOB with succinate	2.03	6VK8
M. trichosporium OB3b MMOH-MMOB with one site reduced	2.35	6VK4
M. trichosporium OB3b MMOH-MMOB 5FW	2.80	7M8Q
M. trichosporium OB3b MMOH-MMOB BTFA/K15C/5FW	2.20	7M8R
M. trichosporium OB3b MMOH-MMOB S109A/T111A form 1	1.96	7S6Q
M. trichosporium OB3b MMOH-MMOB S109A/T111A form 2	2.40	7S7H
M. trichosporium OB3b MMOH-MMOB H5A	1.89	7S6R
M. trichosporium OB3b MMOH-MMOB N107G/S110A	1.98	7S6S
M. trichosporium OB3b MMOH-MMOB H33A	1.82	7S6T
M. trichosporium OB3b diferric MMOH-MMOB XFEL	1.95	6YD0
M. trichosporium OB3b diferrous MMOH-MMOB XFEL	1.95	6YDI
M. trichosporium OB3b reoxidized MMOH-MMOB XFEL	1.95	6YDU
M. trichosporium OB3b diferrous MMOH-MMOB t = 0 XFELa	2.00	6YY3
M. sporium MMOH-MMOD	2.60	67DK

^aTreated the same as 6YDU but exposed to helium rather than O₂.

Figure 19.

Structures of the sMMO proteins. (a) Overall structure of MMOH highlighting one protomer of the α₂β₂γ₂ dimer (PDB ID: 1MTY). The α subunit is shown in light purple, the β subunit is shown in dark purple, and the γ subunit is shown in wheat. (b) Structure of MMOB (green, PDB ID: 4GAM). (c) Structure of the MMOR ferredoxin domain (light pink, PDB ID: 1JQ4) and the MMOR FAD domain (salmon, PDB ID: 1TVC).

Crystal structures of protein-protein complexes are also available (Table 3). The structures of both M. capsulatus (Bath)^280,281 and M. trichosporium OB3b²⁸² MMOH in complex with MMOB show that two molecules of MMOB bind symmetrically to the MMOH dimer, altering the conformations of α subunit helices E, F, and H (Figure 20a). The N-terminal 35 residues of MMOB order into a ring-like structure on the surface of MMOH, explaining why removal of the N-terminus obviates or significantly reduces sMMO activity^280,283,284 and why mutation of specific N-terminal residues affects steps in the catalytic cycle.²⁸⁵ The MMOB C-terminus also becomes more ordered upon complexation, consistent with its truncation decreasing the MMOH turnover number.²⁸⁶ While a structure is not available for the MMOH-MMOR complex, hydrogen-deuterium exchange coupled to mass spectrometry analysis²⁸⁷ and chemical cross-linking data²⁸⁸ indicate that MMOR binds to the same region of MMOH as MMOB, specifically with its [2Fe-2S] cluster-containing domain occupying the MMOB binding site. Finally, a structure of the MMOH-MMOD complex from Methylosinus sporium strain 5 shows that MMOD binds in the same site as MMOB (Figure 20b), rationalizing its inhibitory effect in vitro.¹⁴¹ MMOD consists of four antiparallel β strands and a C-terminal α helix followed by an unstructured region comprising ~35 residues. Notably, MMOD displaces the N-terminal helix of the MMOH β subunit, causing helices B and C in the α subunit to shift position, thereby altering the geometry of the diiron active site.¹⁴¹

Figure 20.

Structures of MMOH protein-protein complexes. (a) Structure of the M. capsulatus (Bath) MMOH-MMOB complex with MMOB shown in green (PDB ID: 4GAM). (b) Structure of the M. sporium MMOH-MMOD complex with MMOD shown in pink (PDB ID: 6D7K). The α subunits are shown in light purple, the β subunits are shown in dark purple, and the γ subunits are shown in wheat.

4.2. Active Site Structure

The diiron active site of sMMO, first identified by EPR, Mössbauer, and EXAFS spectroscopies, consists of two iron ions, Fe1 and Fe2, which are antiferromagnetically coupled in the Fe(III)Fe(III) state and weakly ferromagnetically coupled in the Fe(II)Fe(II) state.^289–293 A mixed valent Fe(II)Fe(III) state can be generated, but it is not part of the catalytic cycle. MMOH has been crystallographically characterized in all three of these oxidation states (Table 3). In the Fe(III)Fe(III) state, the two iron ions are separated by 3.1 Å. Fe1 is coordinated by Glu114, His147, and a solvent molecule, while Fe2 is coordinated by Glu209, Glu243, and His246 (M. capsulatus (Bath) MMOH numbering). The two iron ions are bridged by two hydroxides and Glu144 (Figure 21a).^{269–271,294} In the Fe(II)Fe(II) state, the Fe–Fe distance increases to 3.3 Å, and Glu243 shifts to bridge Fe1 and Fe2, displacing a bridging a hydroxide and adopting a bidentate coordination to Fe2 (Figure 21b).^282,295 In the Fe(II)Fe(III) state, the Fe–Fe distance increases to 3.3–3.4 Å and Glu144 no longer coordinates Fe2 (Figure 21c).²⁹⁴ MMOH has also been crystallized in the apo, Co(II)Co(II), and Mn(II)Mn(II) forms (Table 3).²⁶⁸ The latter two structures exhibit metal coordination geometries similar to that of reduced Fe(II)Fe(II) MMOH. In the MMOH-MMOB complex from M. capsulatus (Bath), Glu243 adopts coordination more similar to that of reduced MMOH.²⁸⁰ By contrast, the M. trichosporium OB3b MMOH-MMOB X-ray free electron laser (XFEL) structure determined at room temperature reveals a coordination similar to that of oxidized MMOH, suggesting that photoreduction occurred in the M. capsulatus (Bath) structure and that MMOB binding does not perturb Glu243.²⁸²

Figure 21.

Active site of sMMO from M. capsulatus (Bath) with helices B, C, E, and F labeled. (a) The diiron cluster in the oxidized Fe(III)Fe(III) state (PDB ID: 1MTY). (b) The diiron cluster in the reduced Fe(II)Fe(II) state (PDB ID: 1FYZ). (c) The diiron cluster in the mixed valent Fe(II)Fe(III) state (PDB ID: 1FZ0).

Multiple structures of MMOH with substrates, substrate analogs, products, and product analogs bound at the diiron site are available (Table 3). The substrates dibromomethane and iodomethane and the substrate mimic xenon, often used to probe for O₂ binding sites, bind in cavities extending from the diiron site to the surface (section 4.3),²⁹⁶ as do a range of halogenated product analogs.²⁹⁷ The products methanol, ethanol, 2-bromoethanol, 3-chloropropanol, 6-bromohexanol, and 3-bromo-3-butenol bind at the diiron site with the oxygen atom bridging the two iron ions.^297,298 These structures are consistent with EPR and ENDOR data showing the binding of methanol, ethanol, DMSO, and TFE to the diiron center.^{291,299–301}

4.3. Substrate Access to the Active Site

4.3.1. Chain of Cavities.

Possible pathways for substrate access to the MMOH diiron site have been investigated extensively. There are three hydrophobic cavities extending from the active site to the protein surface, denoted cavities 1, 2, and 3, as well as a pore connecting cavity 1 directly to the surface.^269,302 Binding of substrate and product molecules in cavities 2 and 3 as well as at the diiron site-housing cavity 1 suggested that these pockets provide a route for methane and O₂ entry.^296–298 In particular, residues Phe188 and Leu110 form a gate, which is closed in oxidized M. capsulatus (Bath) MMOH and was proposed to control access to the diiron site from cavities 2 and 3.^270,297 In support of this gating model, these two residues shift in the M. capsulatus (Bath) MMOH-MMOB complex, connecting the two cavities (Figure 22a).²⁸⁰

Figure 22.

Proposed channels to and from the sMMO active site (PDB ID: 6YDI). (a) Substrate delivery channel to the hydrophobic pocket. (b) W308 tunnel 1 shown with key gating residues.

Different results were obtained for M. trichosporium OB3b MMOH: the gate is open in both oxidized MMOH and the oxidized MMOH-MMOB complex.³⁰³ Further complicating the interpretation, the gate is closed in both the oxidized and the reduced M. trichosporium OB3b MMOH-MMOB XFEL structures determined at room temperature.²⁸² Thus, it seems that MMOB may serve to close, rather than open, the gate. Interestingly, the gate is open in structures of M. trichosporium OB3b MMOH-MMOB with bound benzoate and succinate, and further examination of the M. capsulatus (Bath) MMOH-MMOB complex electron density map suggests that an unmodeled substrate molecule might be present and perturb the gate in that structure.³⁰³ Another issue with this pathway is that the reaction kinetics are not consistent with methane accessing the diiron site from the 35–40 Å chain of cavities 1–3.^16,304,305 In particular, the linear decay rate of reactive intermediate Q (section 4.4) with substrate concentration^266,306 is inconsistent with the cavities filling with methane prior to reaction.

4.3.2. W308 Tunnels.

There are two possible alternatives to the cavity path. First, direct entry to the active site might be available through the pore region. However, binding of MMOB covers this region and blocks the diiron center.^280,303 Second, a narrow tunnel, denoted W308 tunnel 1 (Figure 22b), has been identified recently using a probe with a solvent radius of 1.1 Å as opposed to the typically used water solvent radius of 1.4 Å. This tunnel is gated by residues Pro215 and Trp208 and is lined with conserved hydrophobic residues. The tunnel is closed in the structure of reduced MMOH but open in the reduced MMOH-MMOB complex from M. trichosporium OB3b.^303,307 The binding of MMOB leads to organization of a dome of hydrophobic residues at the tunnel entrance, proposed to facilitate O₂ entry.³⁰³ The tunnel is also adjacent to a number of MMOB residues shown by mutagenesis to be important for catalysis.^284,304 Notably, replacement of MMOB residue Val41, located at the tunnel entry to the α subunit, with arginine and other bulky residues almost completely abrogated enzymatic activity.³⁰³ While both the chain of cavities (section 4.3.1) and W308 tunnel 1 have been proposed as access routes for both methane and O₂, recent work suggests that a different, related path exists for methane access. This path, denoted W308 tunnel 2, is widened in the complex between MMOH and a double mutant of MMOB, S109A/T111A,³⁰⁸ consistent with this MMOB variant exhibiting increased rates of reactivity with larger substrates.³⁰⁹

4.4. Mechanism of Dioxygen Activation

Activation of O₂ by sMMO has been studied extensively, with the first iron-oxygen intermediates reported 30 years ago.^266,310 Single-turnover kinetic and spectroscopic studies of reduced MMOH with O₂ in the presence of MMOB have established a detailed reaction cycle (Figure 23).^{15,16,285,311} The first intermediate, O, is an Fe(II)Fe(II) species that is proposed to have O₂ bound to the protein but not at the diiron center, as it exhibits the same optical and EPR spectroscopic features as reduced MMOH.^{266,267,312,313} Intermediate O forms irreversibly, and its existence explains why formation of the subsequent intermediates does not depend on the O₂ concentration.

Figure 23.

sMMO reaction cycle. All of the intermediates, with the exception of R, have been detected directly.

The binding of O₂ to the diiron center then yields intermediate P* followed by intermediate P. For M. capsulatus (Bath) MMOH, intermediate P* was proposed to be an Fe(III)Fe(III) species differing from the subsequent intermediate P in the protonation of a coordinating ligand or solvent molecule.³¹⁴ Intermediate P* in the M. trichosporium MMOH reaction cycle was also originally proposed to be an Fe(III)Fe(III) species, but Mössbauer data indicate that it is actually an Fe(II)Fe(II) species.^315,316 These studies of M. trichosporium MMOH P* were facilitated by using the MMOB His33Ala variant, which slows the decay of P*.^304,316 Intermediate P is an Fe(III)Fe(III) peroxo species, identified by its optical and Mössbauer spectra, which are consistent with a cis or trans μ−1,2 bridging coordination.^306,311,315 The formation and decay of P depends on pH, and kinetic solvent isotope effects are observed in D₂O, indicating that proton transfer, likely involving a bound solvent molecule or one of the carboxylate ligands, is involved.^314,315

In the next step, the O–O bond is cleaved to form Q, the intermediate that reacts directly with methane. Intermediate Q is an antiferromagnetically coupled Fe(IV)Fe(IV) species,^266,312,317 of which the exact structure has been the subject of ongoing debate (Figure 23). On the basis of Mössbauer parameters and EXAFS data fit with a short 2.46 Å Fe–Fe interaction, Q was proposed to have a diamond core structure,^317,318 an assignment later supported by time-resolved resonance Raman data.^319,320 However, difficulties reproducing the short Fe–Fe distance computationally³²¹ and reactivity comparisons of biomimetic diamond and open core model compounds³²² suggested that alternative structures might be plausible. Using HERFD XAS, it was possible to compare the pre-edge energy of Q with those of a range of Fe(IV)Fe(IV) model complexes. Combined with calculations, these data led to the conclusion that Q is better described as an open core structure.^323,324 A comparison of newly acquired HERFD EXAFS data with the prior partial fluorescent yield (PFY) EXAFS results further indicated that the 2.46 Å Fe–Fe distance could derive from background contamination, and gave a revised Fe–Fe distance of 3.30–3.34 Å, which is more consistent with an open core.³²⁵ The tide then turned back, with a systematic nuclear resonance vibrational spectroscopic (NRVS) study supporting only closed core models.³²⁶ In addition, DFT calculations predict that concerted motions of the two oxo bridges in the closed core structure provide the reactivity necessary to break the methane C–H bond.³²⁶ Once Q reacts with methane (section 4.5), an oxo-bridged Fe(III)Fe(III) product complex with an oxygen atom derived from O₂,³¹⁹ intermediate T, is formed. Finally, methanol is released, regenerating the resting Fe(III)Fe(III) state (Figure 23).

4.5. Mechanism of C–H Activation

The reaction of intermediate Q with methane has been studied by a range of experimental and computational approaches. In contrast to pMMO (section 3.7), reactions with chiral ethane and chiral butane yield some inversion of stereochemistry, consistent with hydrogen abstraction by Q producing a short-lived radical intermediate.^327–331 Radical clock substrates have also been employed as probes, including substituted cyclopropanes, methylcubane, and norcarane.^332–337 In these studies, rearrangement of the probe substrate upon reaction with Q can inform upon the existence and lifetime of transient intermediates. The overall results are consistent with the involvement of a short-lived radical but suggest that different substrates are oxidized by different mechanisms, rendering it difficult to draw conclusions regarding methane oxidation.^{16,321,338,339}

Kinetic isotope effect (KIE) measurements have also provided insight into the sMMO mechanism. A remarkably large KIE is obtained for the reaction of deuterated methane with intermediate Q, 50 for M. trichosporium OB3b sMMO and 28 for M. capsulatus (Bath) sMMO, while no KIE is measured using other substrates.^306,340,341 Analysis of the temperature dependence of the KIE for methane is consistent with significant quantum tunneling,³⁰⁹ which is facilitated by interactions with MMOB. Using the MMOB quad variant (N107G/S109A/S110A/T111A), which increases the decay rate of Q with larger substrates presumably by increasing the size of the entry pathway (pore or other),³⁰⁴ reduces the methane KIE to 6. This result indicates that conformational changes upon interaction with MMOB are not only relevant to substrate access and methane specificity but also important for tunneling.³⁴¹ Thus, MMOB facilitates selectivity for methane, despite it having the highest C–H bond strength, both by modulating substrate access and by enabling quantum tunneling.

4.6. Interplay between MMOR and MMOB Binding

The role of MMOR in providing electrons to MMOH is clear and the ways in which MMOB regulates substrate access and steps in the catalytic cycle have emerged over the years, but how the binding of these two proteins is orchestrated has been the subject of debate. MMOB is believed to prevent further reduction of intermediate Q by MMOR before it can react with methane. In support of this role, reduced MMOH in the presence of MMOR, but not MMOB, exhibits significantly less activity.³⁴² As summarized recently, several distinct models for regulation of electron transfer from MMOR to MMOH have been proposed.³⁴³ In one scenario, MMOR and MMOB bind to MMOH simultaneously using separate binding sites, consistent with cross-linking data suggesting formation of a ternary complex.³⁴⁴ Alternatively, only one component can interact with MMOH at a time. This model is supported by hydrogen-deuterium exchange coupled to mass spectrometry analysis showing that the binding sites overlap and fluorescence anisotropy measurements indicating that the MMOR [2Fe-2S] domain can displace MMOB from MMOH.²⁸⁷ In this scenario, MMOR reduces MMOH and is then replaced by MMOB, which might remain loosely associated via its N-terminal region or might dissociate completely.

Displacement of MMOR from reduced MMOH by MMOB is consistent with fluorescence anisotropy data showing that M. capsulatus (Bath) MMOB has a higher affinity for reduced MMOH than for oxidized MMOH.²⁷⁹ The increased affinity of MMOB for reduced MMOH contradicts early reports that MMOB decreases the MMOH redox potential,^345,346 which would mean it binds oxidized MMOH with higher affinity. Initial studies of M. trichosporium OB3b MMOB and MMOH using fluorescent probes did indicate a higher affinity for the oxidized form,³⁴⁷ but recent reinvestigation of the affinity of M. trichosporium OB3b MMOB for MMOH using ¹⁹F NMR gave different results.³⁴³ Instead of attaching large probes to cysteine residues in MMOB as done previously, two tryptophan residues in MMOB and one in MMOR were replaced with 5-fluorotryptophan. In addition, an MMOB variant labeled with 3-bromo-1,1,1-trifluoroacetone was generated. The combined use of these less disruptive labels and the sensitivity of ¹⁹F NMR provided new insight into the interactions between the components. In particular, the same binding constants were measured for the interactions of both oxidized and reduced M. trichosporium OB3b MMOH with MMOB. The observed effect of MMOB on the MMOH redox potential^345,346 is not consistent with this finding but may be due to experimental considerations in the redox titration.³⁴³

The ¹⁹F NMR study also showed that the affinity of MMOR for MMOH is similar to that of MMOB for MMOH, regardless of the MMOH oxidation state. These observations suggest that there is an equilibrium and that MMOR can completely displace MMOB, ruling out the simultaneous binding mechanism. Thus, a third model was proposed in which MMOB and MMOR compete for the binding site on MMOH regardless of oxidation state (Figure 24).³⁴³ In this dynamic equilibrium model, irreversible reaction steps, including reduction of MMOH while in complex with MMOR and subsequent reactions with O₂ and methane while in complex with MMOB, pull the reaction forward. The slow kinetics of MMOB dissociation from MMOH are proposed to protect intermediate Q from unproductive reduction by MMOR before it can react with methane.

Figure 24.

Model for regulation of electron transfer and substrate binding in sMMO adapted with permission from ref 343.

4.7. Overexpression and Engineering

Many studies of the MMOH mechanism and its interactions with MMOB and MMOR were facilitated by the ability to heterologously express MMOB in E. coli and produce site-specific and truncated variants. It has proven much more difficult to generate variants of MMOH. Expression of MMOH, MMOB, and MMOR in Pseudomonas strains has been reported, and Pseudomonas putida expressing sMMO degrades trichloroethylene (TCE) at 12.5% of the rate of TCE degradation by M. trichosporium OB3b.^28,348,349 TCE oxidation was also observed for strains of Agrobacterium tumefaciens and Rhizobium meliloti expressing sMMO,³⁵⁰ but none of these systems were reported to oxidize methane.

More recently, attempts to express sMMO have focused on coexpression with the E. coli chaperone proteins GroES and GroEL. Coexpression of the Methylomonas methanica MC09 sMMO operon with E. coli GroEL and GroES led to assembly of the MMOH dimer as detected by native PAGE.³⁵¹ This MMOH exhibited nitrobenzene oxidation activity at about half the level of M. trichosporium OB3b sMMO and an EPR signal consistent with the presence of a mixed valent Fe(II)Fe(III) center. While the sMMO operon encodes a GroEL homolog, MmoG (Figure 5a), it is unclear whether it interacts with a GroES homolog and MmoG alone is not sufficient to yield soluble MMOH. In a preprint report, screening and directed evolution yielded soluble M. capsulatus (Bath) MMOH upon coexpression with a GroEL/GroES pair from M. capsulatus (Bath). Methane conversion to methanol was observed in the E. coli cells expressing sMMO, and further metabolic engineering to produce acetone was successful.³⁵² In addition, mutations that enhance activity, likely by enhancing solubility, were identified using directed evolution. Overall, this approach is promising both for biochemical studies and for biotechnological applications.

Another strategy for generating MMOH variants is homologous expression in methanotrophs. Site-directed mutagenesis of M. trichosporium OB3b MMOH has been performed by introducing genes with mutations into a strain lacking part of the sMMO operon. Unlike pMMO (section 3.8), sMMO is not required for cell viability, and these strains can be cultivated under pMMO-utilizing conditions followed by expression of sMMO variants as copper levels are lowered.^{143,353–355} This system has been used to alter several residues near the diiron center, including Cys151 and Thr213,³⁵³ as well as a leucine, Leu110, separating the active site from cavity 2.¹⁴³ Two Cys151 variants, Cys151Glu and Cys151Tyr, could not be produced at high levels, while a Thr213Ser variant was purified and exhibited diminished propylene oxidation activity.³⁵³ Mutation of Leu110 to glycine, cysteine, arginine, and tyrosine resulted in differences in regioselectivity,¹⁴³ as did alterations to Phe192, which resides close to the diiron center and to Arg98, which is part of a hydrogen bonding network proposed to modulate access to the cavity pathway.³⁵⁶ Despite these promising results over the past 20 years, this approach has yet to be deployed on large scale, likely due to limitations in working with methanotrophs. Finally, as noted above (section 3.8), CRISPR-Cas9 gene editing can be performed with ~10% efficiency in M. capsulatus (Bath),^257,260 but its efficacy in producing point mutants has not yet been demonstrated.

5. CONCLUSIONS AND OUTLOOK

Within the time of preparing this article, the climate crisis has intensified, as manifested in dangerous air and water temperatures, wildfires and accompanying air pollution, and extreme weather. As an abundant yet short-lived greenhouse gas, methane is a prime target for immediate mitigation efforts, and methanotrophs and MMOs present a promising route forward. While significant progress has been made toward an atomic level understanding of both pMMO and sMMO function, important questions remain unanswered. The picture of pMMO has been revised multiple times, with the most recent studies indicating that the active site is located at the PmoC Cu_D site, with the possibility that Cu_D and Cu_C can be occupied simultaneously still on the table. Activity and structural studies of pMMO in membranes and native lipid nanodiscs have underscored the importance of studying the enzyme in its native environment, and future work should prioritize in situ characterization. Protein-protein interactions with candidates such as PmoD and MDH may only be detectable in situ and have the potential to shed light on the physiological reductant(s). The state of the field is more advanced for sMMO, with an established catalytic cycle and a detailed model of interactions between MMOH and its partner proteins MMOR and MMOB. Regulation of intermediate Q formation, methane selectivity, and active site access are also well understood. The nature of Q remains controversial, but the continual application of advanced techniques should resolve this debate in the near future. While computational studies have indicated that a pMMO monocopper site could oxidize methane, experimental evidence for reactive intermediates analogous to sMMO intermediate Q has not yet been obtained. It is likely that pMMO also has specific mechanisms for preventing inactivation of intermediates and the enzyme subunits themselves. Engineering of both MMOs is nascent, with progress in heterologous and homologous expression, cell free protein synthesis, and CRISPR-Cas9 gene editing over the past few years. Further efforts will be required to establish robust systems that will ultimately be scalable to a commercially viable level. Such systems will require a deep understanding of methanotroph physiology, particularly as pertains to the copper switch, metal acquisition, and ICM formation. In the future, biochemists and structural biologists will need to interface closely with microbiologists, synthetic biologists, and engineers to leverage the full potential of these remarkable bacteria and enzymes for the sake of our planet.

Biographies

Frank J. Tucci, from Hartford, CT, received his B.A. in Chemistry and in Neuroscience from Wesleyan University, where he conducted research in the laboratory of Professor Erika A. Taylor, studying an enzyme involved in lipopolysaccharide biosynthesis, Heptosyltransferase I. Frank is now a Ph.D. candidate in the laboratory of Professor Amy C. Rosenzweig at Northwestern University, where he is using cryoEM to study pMMO, with particular interests in its copper centers, its interactions with the native membrane, and electron microscopy methods. He is a former trainee of the NIH-funded Chemistry of Life Processes Training Program and is currently supported by a Ruth L. Kirschstein Predoctoral Individual National Research Service Award and the Northwestern University Rappaport Award for Research Excellence.

Amy C. Rosenzweig, originally from Pittsburgh, PA, is the Weinberg Family Distinguished Professor of Life Sciences in the Departments of Molecular Biosciences and Chemistry at Northwestern University. She received a B.A. in Chemistry from Amherst College and a Ph.D. in Inorganic Chemistry from Massachusetts Institute of Technology. The Rosenzweig laboratory uses structural, biochemical, and biophysical approaches to attack problems at the forefront of bioinorganic chemistry. Their work has been honored recently by the American Chemical Society Alfred Bader Award in Bioinorganic or Bioorganic Chemistry and the Protein Society Hans Neurath Award. Rosenzweig, a member of the National Academy of Sciences and a fellow of the American Academy of Arts and Sciences, is currently chair of the Department of Molecular Biosciences.