Methanobactin and the Link between Copper and Bacterial Methane Oxidation

SUMMARY

Methanobactins (mbs) are low-molecular-mass (<1,200 Da) copper-binding peptides, or chalkophores, produced by many methane-oxidizing bacteria (methanotrophs). These molecules exhibit similarities to certain iron-binding siderophores but are expressed and secreted in response to copper limitation. Structurally, mbs are characterized by a pair of heterocyclic rings with associated thioamide groups that form the copper coordination site. One of the rings is always an oxazolone and the second ring an oxazolone, an imidazolone, or a pyrazinedione moiety. The mb molecule originates from a peptide precursor that undergoes a series of posttranslational modifications, including (i) ring formation, (ii) cleavage of a leader peptide sequence, and (iii) in some cases, addition of a sulfate group. Functionally, mbs represent the extracellular component of a copper acquisition system. Consistent with this role in copper acquisition, mbs have a high affinity for copper ions. Following binding, mbs rapidly reduce Cu²⁺ to Cu¹⁺. In addition to binding copper, mbs will bind most transition metals and near-transition metals and protect the host methanotroph as well as other bacteria from toxic metals. Several other physiological functions have been assigned to mbs, based primarily on their redox and metal-binding properties. In this review, we examine the current state of knowledge of this novel type of metal-binding peptide. We also explore its potential applications, how mbs may alter the bioavailability of multiple metals, and the many roles mbs may play in the physiology of methanotrophs.

INTRODUCTION

Methanobactins (mbs) were first identified in aerobic methane-oxidizing bacteria (methanotrophs). This remarkable group of bacteria can grow using methane as their sole source of carbon and energy. They are ubiquitous in environments where oxygen and methane are available, and they play a major role in consuming much of the methane produced in the biosphere, thus mitigating its effects in global warming (1–4). Due to their growth on an inexpensive, readily available, and, if generated via methanogenesis (5), renewable carbon source, methanotrophs also have considerable potential for the production of bulk and fine chemicals and for bioremediation of pollutants in the environment (2, 6–8).

The first report of a bacterium growing on methane was by Söhngen, working in Beijerinck's laboratory in Delft in The Netherlands, who in 1906 reported the isolation of Bacillus methanicus from aquatic plants and pond water (9). It was not until 50 years later that this microbe was reisolated and renamed Pseudomonas methanica (10, 11). A second methanotroph, Methylococcus capsulatus (Texas strain), was isolated in 1966 (12). A landmark in methanotroph biology came in 1970, when Whittenbury and colleagues isolated, from a variety of terrestrial and freshwater environments, and described over 100 new aerobic methanotrophs growing on methane (13). They then devised a classification scheme, i.e., type I versus type II, based on the ability of these methanotrophs to grow on methane, pathways of carbon assimilation, formation of resting stages (cysts and spores), morphology, the possession of complex intracytoplasmic membrane arrangements, and the moles percent G+C content of their DNA. Subsequently, Bowman and colleagues isolated a similar number of methanotrophs from various environments and classified them according to the scheme of Whittenbury and colleagues and according to their 16S rRNA phylogeny (14, 15). It is remarkable that though there was no DNA sequencing at the time, the overall classification scheme of Whittenbury and colleagues still remains a robust and convenient way of grouping methanotrophs today.

Accordingly, there are currently 15 genera of methanotrophs within the family Methylococcaceae and 3 in the Methylothermaceae family of the Gammaproteobacteria class. Methylobacter, Methylocaldum, Methylococcus, Methylogaea, Methyloglobulus, Methylomagnum, Methylomarinum, Methylomicrobium, Methylomonas, Methyloparacoccus, Methyloprofundus, Methylosoma, Methylosphaera, Methylosarcina, and Methylovulum are the methanotrophs in the Methylococcaceae family, and Methylohalobius, Methylomarinovum, and Methylothermus are the methanotrophs in the Methylothermaceae family (16–21, 227). Within the class Alphaproteobacteria, the genera Methylosinus and Methylocystis are found in the family Methylocystaceae and the genera Methylocella, Methyloferula, and Methylocapsa in the family Beijerinkiaceae. In the last 15 years, there have been increasing reports of facultative methanotrophs within the Methylocella, Methylocapsa, and Methylocystis genera that can use multicompounds for growth in addition to methane (22–26). Also known today are filamentous methanotrophs from other genera, such as Crenothrix and Clonothrix, and nonproteobacterial (verrucomicrobial) methanotrophs of the genus Methylacidiphilum growing at high temperatures and low pH have also been discovered recently (27). Finally, it was shown that “Candidatus Methylomirabilis oxyferans,” a member of the NC10 phylum, generates dioxygen for the oxidation of methane despite being an obligate anaerobe (28, 29). Taken together, these data clearly illustrate the widespread nature of methanotrophic bacteria in most ecosystems of our planet.

Physiology and Biochemistry of Methanotrophs

Methanotrophs can use methane as an energy source and also to provide carbon for all of their cellular constituents (6, 30, 31). The initial oxidation of methane to methanol is catalyzed by the enzyme methane monooxygenase (MMO). There are two structurally and biochemically distinct forms of MMO, a membrane-associated or particulate MMO (pMMO) and a cytoplasmic or soluble MMO (sMMO), which represent evolutionarily independent solutions to the same molecular problem of methane oxidation (32–37). The sMMO is a three-component binuclear iron active-center monooxygenase that belongs to a large group of bacterial hydrocarbon oxygenases, known as soluble di-iron monooxygenases (SDIMOs) (38), which are also homologous to the R2 subunit of class I ribonucleotide reductase. Two very similar sMMO systems, from Methylococcus capsulatus (Bath) (39–43) and Methylosinus trichosporium OB3b (44–47), have been studied in detail. sMMO is encoded by a six-gene operon, mmoXYBZDC, and has three components: (i) a 250-kDa hydroxylase with an α₂β₂γ₂ structure in which the α subunits (MmoX) contain the binuclear iron active center where substrate oxygenation occurs, (ii) a 39-kDa NAD(P)H-dependent reductase (MmoC) with flavin adenine dinucleotide (FAD) and Fe₂S₂ prosthetic groups, and (iii) a 16-kDa component (MmoB) known as protein B or coupling/gating protein that does not contain prosthetic groups or metal ions (39, 48). There are X-ray crystal structures for the hydroxylase component (49–52), nuclear magnetic resonance (NMR)-derived structures for protein B (39, 53, 54), and an NMR-derived structure for the flavin domain of the reductase (55). The complex formed by the three components has been studied structurally via small-angle X-ray scattering analysis and biophysically by electron paramagnetic resonance spectroscopy, ultracentrifugation, and calorimetric analysis (56, 57). The catalytic cycle of sMMO has been extensively studied, and excellent progress has been made toward understanding the mechanism of oxygen and hydrocarbon activation at the binuclear iron center (45) (45, 58–62). Detailed reviews of the structure and catalytic mechanisms of sMMO are available (6, 37, 47, 63, 64).

The pMMO, in contrast, is a copper- and possibly iron-containing, membrane-associated enzyme that is associated with unusual intracytoplasmic membranes that take the form of vesicular disks in type I methanotrophs and paired peripheral layers in type II organisms (65–75). Intracytoplasmic membranes are enriched in pMMO and can be physically separated from the cytoplasmic membrane on the basis of sedimentation velocity in sucrose density gradients (76). An understanding of the structure and mechanism of pMMO has emerged more slowly than that for sMMO because of losses of activity when the enzyme is solubilized. pMMO consists of three polypeptides of approximately 49, 27, and 22 kDa encoded by the genes pmoCAB (77). There are often multiple copies of these pmo genes in methanotrophs (78, 79). Recent studies have shown that native pMMO forms a complex with methanol dehydrogenase (MeDH), which may supply electrons to the pMMO (80, 81), similar to what has been found for the hydroxylamine oxidoreductase and ammonia monooxygenase redox couples (82–85).

Some methanotrophs, such as M. capsulatus (Bath) and M. trichosporium OB3b, can produce either form of MMO. Most known methanotrophs possess only pMMO, e.g., Methylomonas methanica, Methylomicrobium album BG8, Methylocystis parvus OBBP, and the verrucomicrobial and NC10 methanotrophs. Only a few methanotrophs within the Beijerinckiaceae family, e.g., Methylocella silvestris and Methyloferula stellata, have sMMO but do not possess pMMO (21, 86).

The methanol produced by MMO is oxidized to formaldehyde by a calcium or rare-earth-dependent pyrroloquinoline quinone (PQQ)-containing MeDH (87–91). Formaldehyde is an important branch point in the metabolism of methanotrophic metabolism and represents the point at which one-carbon (C₁) intermediates can be either oxidized to CO₂ to derive energy or assimilated into biomass. Since formaldehyde is toxic, methanotrophs must protect themselves against accumulation of this metabolic intermediate. Multiple pathways for metabolism of formaldehyde are found in methanotrophs (2, 26, 92–96). For example, oxidative dissimilation of formaldehyde can occur by its conjugation to tetrahydromethanopterin (H₄MPT) (97, 98), via dye-linked membrane-associated (93) or via NAD⁺-dependent (95, 96, 99) formaldehyde dehydrogenases. Formate, resulting from the oxidation of formaldehyde by formaldehyde dehydrogenases, is further oxidized to carbon dioxide by an NAD⁺-dependent formate dehydrogenase which generates NADH, which is then available for oxidation of methane, biosynthetic reactions, and energy generation within the cell (100–102). Methanotrophs also possess two pathways for fixation of formaldehyde into biomass, the serine and ribulose monophosphate (RuMP) cycles, which are active in alphaproteobacterial and gammaproteobacterial methanotrophs, respectively. The pathways of carbon fixation in methanotrophs have been reviewed extensively (see, e.g., reference 6).

The “Copper Switch” in Methanotrophs

Early attempts to characterize methane oxidation were complicated by different reports on the cellular location of the MMO. MMOs were described either as soluble or as membrane associated, depending on the strain and, for some strains, on the reporting laboratory. Several groups initially reported activity in the particulate or membrane fraction (103, 104), whereas other groups detected activity in the soluble fraction (105, 106). Subsequent studies showed that the cellular location varied with cultivation conditions. Oxygen limitation was reported to induce methane oxidation in the soluble fraction in M. trichosporium OB3b (36, 107). However, it was subsequently shown that oxygen was not the regulatory factor and that the switch between the membrane-associated and soluble activities was related to biomass concentration (108). The defining moment in the discovery of this “switch” was when Dalton and colleagues attempted to grow M. parvus OBBP to high cell densities in chemostat culture. They observed that cultures of M. parvus OBBP, when supplied with methane, air, and a nitrate mineral salts (NMS) solution at relatively low cell densities, simply stopped growing. However, when additional trace element solution was added, the cultures immediately started growing again. The “secret ingredient” in the trace element solution was narrowed down to copper ions (108). It was subsequently realized that M. parvus OBBP contained only pMMO, resulting in the high requirements for copper ions, and did not contain an sMMO which would have allowed it to grow to the same high cell densities observed at the time with M. capsulatus Bath and M. trichosporium OB3b under copper limitation. The latter strains, when confronted with the same conditions, switched to expression of the sMMO and carried on growing (35, 109). Interestingly, copper had earlier been shown to enhance growth of Methanomonas margaritae, a methanotroph which does not contain sMMO, but these original observations were never investigated further (110).

Dalton and colleagues followed up their observations in detail and established the existence of this “copper switch,” i.e., the regulation of expression of the two different forms of MMO in methanotrophs in response to the copper-to-biomass ratio of cultures of methanotrophs which possess both sMMO and pMMO. This allowed them to explain many of the earlier observations on metal-dependent growth of methanotrophs. For instance, they showed that in M. capsulatus Bath, expression of the sMMO was observed only at high cell densities when copper ions in the medium were depleted, whereas additions of excess copper ions allowed this methanotroph to express active pMMO. Subsequently, Murrell and colleagues showed at the molecular level that under growth with low concentrations of copper ions, expression of sMMO was initiated at a σ⁵⁴ promoter upstream of the sMMO gene cluster (mmoXYBZDC). Conversely, under high-copper growth conditions, expression of sMMO was repressed, and high levels of expression of the genes encoding pMMO (pmoCAB) allowed both M. trichosporium OB3b and M. capsulatus Bath to grow using pMMO (34, 111–113).

Further research established that copper affects methanotrophic physiology and gene expression much more broadly. For example, it was found that the intracytoplasmic membrane content in methanotrophs increased with increasing copper in the growth medium (66, 109, 114). This was not totally unexpected, however: considering that the pMMO is localized in the intracytoplasmic membranes, then greater expression and activity of pMMO would logically require more of these membranes. More surprisingly, however, it was recently discovered that pMMO and the PQQ-linked MeDH encoded by the mxa operon form a supercomplex anchored in the intracytoplasmic membranes and that electron transfer from the PQQ-linked MeDH to pMMO in vivo may drive the oxidation of methane (80, 81). In support of the latter finding, it was recently found that not only does expression of pmo genes increase with increasing copper, but that of genes in the mxa operon does so as well (91).

It has also been shown by proteomics that additional steps in the oxidation of methane to carbon dioxide are overexpressed with increasing availability of copper, such as that of proteins involved in lipid, cell wall, and membrane synthesis (66, 115). Conversely, the ability of methanotrophs to direct carbon from methane to poly-3-hydroxybutyrate increases with decreasing copper availability (66, 116), suggesting that to some extent energy metabolism of methanotrophs is controlled by copper. Such a conclusion was actually already reached earlier by Dalton and colleagues, who showed that the biomass yield and carbon conversion efficiency in methanotrophs when grown on methane increased as copper increased, i.e., when methanotrophs switch from expressing sMMO to expressing pMMO (117). The “surfaceome,” or proteins on the outer surface of the outer membrane, is also controlled by copper availability in some methanotrophs. The expression of several multi-c-type cytochromes and proteins believed to be involved in copper uptake also varies, and in most cases decreases, as copper availability increases (118–123).

Additional insight into how methanotrophs handle copper has been provided by the recent discovery of the Csps, a new family of copper storage proteins, in M. trichosporium OB3b (124). This methanotroph possesses three Csps: Csp1 and Csp2, which have predicted twin arginine translocase-targeting signal peptides and are therefore thought to be exported after folding, as well as the cytosolic Csp3. Csp1 forms a tetramer of four helix bundles that can bind up to 52 Cu¹⁺ ions via Cys residues that point into the core of the bundle. Switchover to sMMO is accelerated in the Δcsp1 csp2 mutant compared to the wild type, suggesting that these proteins play a role in storing copper for the pMMO and provide an internal copper source when copper becomes limiting. Under such conditions, mb is produced, which can readily remove all Cu¹⁺ from Csp1 and therefore may play a role in helping to utilize Csp1-bound copper.

Evidence Suggesting a Copper-Specific Uptake System

The first evidence for a copper-specific uptake system and for the production of an extracellular copper-binding ligand came during the phenotypic characterization of constitutive sMMO mutants (sMMO^C) of M. trichosporium OB3b (125, 126). Phelps et al. (126) isolated five sMMO^C mutants by culturing M. trichosporium OB3b in the presence of dichloromethane, acting as a mutagen following its cometabolic transformation by methane monooxygenase to formyl chloride. In addition to the sMMO^C phenotype, sMMO^C mutants were defective in copper acquisition (125, 127, 128). A drastic increase in soluble versus insoluble copper in the culture medium was also observed and fostered speculation on the production of an extracellular Cu²⁺-complexing agent(s) analogous to Fe³⁺-complexing siderophores (127). Subsequent studies revealed the presence of a low-molecular-mass copper-binding ligand, although the identity of this compound was not determined (125, 127).

Initial Identification and Isolation of Methanobactin (a Copper-Binding Compound or “Chalkophore”)

Somewhat paradoxically, the “copper-binding ligand” or “copper-binding compound” was first isolated from M. capsulatus Bath during the purification of pMMO and thus under conditions of high copper concentration (73). Separation of this copper-binding compound from the pMMO revealed a low-molecular-mass yellow-fluorescent copper-containing molecule. The color and fluorescence properties of this molecule were similar to those of the water-soluble pigment observed in cells cultured in low-copper medium (73). Subsequent characterization of this water-soluble pigment revealed that it was identical to the copper-binding compound that copurified with the pMMO (73, 128). The production of water-soluble pigments by methanotrophs had been noted during the initial isolation of many type strains over 40 years ago but was associated with cells cultured in low-iron medium (13). The copper-binding compound was eventually termed methanobactin (mb), based on the molecule's antimicrobial activity toward Gram-positive bacteria (129, 130). Once identified, this copper-binding compound was isolated or identified in a number of different methanotrophs, including Methylomicrobium album BG8 (131), Methylocystis strain SB2 (132), Methylocystis rosea (strain SV97) (133), Methylocystis hirsuta CSC1 (133), and Methylocystis strain M (133). The crystal structures of the mbs from M. trichosporium OB3b (134, 135), M. hirsuta CSC1 (133), and Methylocystis strain M (133) have been determined. Furthermore, the chemical structures of the mbs from Methylocystis strain SB2 (132) and M. rosea (133) have been deduced. This review will focus on these mbs that have allowed putative mbs to be identified from methanotrophs with sequenced genomes.

Methanobactins as Chalkophores

Functionally, mbs are similar to siderophores. Like siderophores, they are all low-molecular-mass (<1,200-Da) compounds produced by bacteria under low-copper conditions (125, 126, 128, 136–138). Following the nomenclature of siderophores, which is from the Greek for iron bearing or iron carrying, mbs are chalkophores (copper bearing or copper carrying) (134, 139). mbs are currently the only known representatives of this group. Consistent with their role as the extracellular component of a copper acquisition system, mbs have some of the highest known binding affinities for copper ions (2, 131, 135, 138, 140, 141) (Table 1).

TABLE 1.

Metal binding affinities (K) of mbs from M. trichosporium OB3b, Methylocystis strain M, Methylocystis hirsuta CSC1, Methylocystis rosea, and Methylocystis strain SB2^a

mb	Method	pH	Reference(s)	K (M−1)
				Cu1+	Cu2+	Cu2+/Cu1+ b	Au3+	Ag1+	Hg2+	Pb2+	Cd2+	Co2+	Fe3+	Mn2+	Ni2+	Zn2+
mb4-OB3b	ITC	6.8	141, 148	—c	—	3.2 × 1034	1.0 × 1040	—	UTF	—	—	—	—	—	—	—
mb2-OB3b		6.8		—	—	2.6 × 108	1.0 × 105	2.6 × 107	9.9 × 109	—	—	—	—	—	—	—
mb-OB3b		6.8		—	—	6.5 × 106	1.8 × 105	4.7 × 104	9.0 × 105	6.8 × 105	—	—	—	—	—	—
mb4-OB3b	PT	6.8	172	—	—	1.0 × 1058	—	—	—	—	—	—	—	—	—	—
mb2-OB3b		6.8		—	—	1.0 × 1052	—	—	—	—	—	—	—	—	—	—
mb-OB3b		6.8		—	—	1.0 × 1025	—	—	—	—	—	—	—	—	—	—
mb-OB3b	Competition titrations with BCS	7.5	135	∼7 × 1020	—	—	—	—	—	—	—	—	—	—	—	—
mb-OB3b		7.5		—	∼3 × 1012d	—	—	—	—	—	—	—	—	—	—	—
mb-M	Competition against mb-OB3b	7.5	133	∼1 × 1021	—	—	—	—	—	—	—	—	—	—	—	—
mb-M		7.5		—	∼2 × 1011d	—	—	—	—	—	—	—	—	—	—	—
mb-CSC1	Competition titrations with BCS	7.5	133	∼8 × 1020	—	—	—	—	—	—	—	—	—	—	—	—
mb-CSC1		7.5		—	∼5 × 1014d	—	—	—	—	—	—	—	—	—	—	—
mb-roseae	Competition against mb-OB3b	7.5	133	∼5 × 1020	—	—	—	—	—	—	—	—	—	—	—	—
mb-rosea		7.5		—	∼7 × 1011d	—	—	—	—	—	—	—	—	—	—	—
mb4-SB2	DITC	6.9	154	—	—	7.6 × 1026	—	—	—	—	—	—	—	—	—	—
mb2-SB2	DITC	6.6		—	—	1.2 × 109	—	—	—	—	—	—	—	—	—	—
mb-SB2	DITC	6.9		—	—	1.0 × 106	—	—	—	—	—	—	—	—	—	—
mb-SB2	Competition titrations with TRIEN	6.8		—	—	1.9 × 1027	—	—	—	—	—	—	—	—	—	—
mb-SB2	ITC	6.8		—	—	—	—	1.2 × 108	—	—	—	—	—	—	—	—
mb4-OB3b	ITC	6.8	148	—	—	—	—	—	—	—	1.3 × 106	—	9.7 × 105	—	—	4.5 × 106
mb2-OB3b		6.8		—	—	—	—	—	—	—	1.1 × 107	1.1 × 106	1.7 × 105	7.7 × 105	4.9 × 105	1.8 × 104

^aAbbreviations, ITC, isothermal titration calorimetry; PT, potentiometric titration; DITC, displacement isothermal titration calorimetry; mb, methanobactin; mb₂, mb dimer; mb₄, mb tetramer; UTF, unable to fit; BCS, bathocuproine disulfonate; TRIEN, triethylenetetramine.

^bCopper was added as Cu²⁺, and therefore the oxidation state of the copper bound by mb is in question due to its ability to reduce Cu²⁺ to Cu¹⁺.

^c—, not determined.

^dCalculated using the reduction potential of the Cu²⁺-mb/Cu¹⁺-mb couple.

^eFor the form of mb-rosea missing the C-terminal Asn and Thr residues.

Although chalkophores and siderophores share a number of properties, these two groups of metal-binding compounds can be distinguished in a number of ways. With the exception of the phytosiderophore domoic acid (142), siderophores are expressed under iron limitation, whereas chalkophores are expressed under copper limitation. Many siderophores bind copper, and chalkophores can bind iron (142–148); however, different metal binding constants characterize the two groups, and colorimetric assays have been developed to distinguish between them based on this difference (138, 149, 150). Structurally, chalkophores differ from siderophores by their typical heterocyclic rings and associated thioamide groups (Fig. 1 and 2). Different ring systems have characteristic UV-visible absorbance, circular dichroism (CD), and fluorescent spectral properties, which can be used for identification and to characterize the metal binding properties of the molecule (131, 133–136, 138, 148, 151–153). Excitation energy transfer occurs between the two rings in mb (136, 154), as observed for the chromophores of light-harvesting complexes (155–157), resulting in the fluorescent properties of mbs. For example, the emission intensity of mbs increases with selective hydrolysis of the one of the rings, and emission intensity often increases following metal addition (132, 136–138, 141, 148, 154). Again, this property can be used in both identification and characterization of mbs.

FIG 1.

Chemical structures of full-length mb-OB3b (135, 151) (A), mb-M (133) (B), mb-CSC1 (133) (C), mb-rosea (133) (D), and mb-SB2 (132) (E). Amino acids marked with asterisks are observed in some but not all samples.

FIG 2.

Core features of mbs. AA, amino acid(s). R-groups can be Arg, Ile, Met, or Pro.

However, not all of the properties just described are sufficient to identify mbs. For instance, Clostridium cellulolyticum produces a copper-binding secondary metabolite, closthioamide, with a molecular mass similar to that of mbs (158–160), and like mbs, closthioamide has thioamide groups, will reduce Cu²⁺ to Cu¹⁺, and has a high Cu¹⁺ binding affinity (>10¹⁵ M⁻¹). Closthioamine will also test positive with the copper-chrome azural S (Cu-CAS) assay, a liquid or plate assay used to screen for mb production (138, 149, 161). However, closthioamide can be distinguished spectroscopically from mbs; e.g., closthioamide shows a single UV-visible absorption maximum at 270 nm arising from its two characteristic phenolic groups separated by six thioamide moieties. Furthermore, closthioamide is able to form a dinuclear Cu¹⁺ complex. Also in contrast to that of mbs, closthioamide synthesis is not regulated by copper, and unlike for mbs, the molecule is believed to be produced by a polyketide synthase (see below).

Similarly, under low-copper conditions, Paracoccus denitrificans also produces a low-molecular-mass 716.18-Da porphyrin, coproporphyrin III, that appears to be involved in copper acquisition (162). Unfortunately, the copper binding properties were not reported in the initial publication, and we are not aware of any follow-up studies. Coproporphyrin III has a typical heme UV-visible absorption spectrum and shows different γ, α, and β maxima depending on the metal coordinated by the heme group, allowing it to be distinguished from mbs based on this property.

STRUCTURAL PROPERTIES OF METHANOBACTINS

Structural Diversity and Core Features

The core structural features of mbs are shown in Fig. 1 and 2. As mentioned above, mbs are modified peptides characterized by the presence of one oxazolone ring and a second oxazolone, imidazolone, or pyrazinedione ring, which are separated by 2 to 4 amino acid residues (see Biosynthesis of Methanobactin below). Each ring has an adjacent thioamide group. Structurally, mbs can be divided into two types (Fig. 3 and 4). One type (group I) is represented by mb from M. trichosporium OB3b (mb-OB3b) (Fig. 1A, 3A, and 4). In addition to these core properties, mb-OB3b also contains Cys residues in the mature peptide, which are linked by an intramolecular disulfide bond (Fig. 1A and 3A). Cu¹⁺-mb-OB3b has a pyramid-like shape with the metal coordination site at the base (133–135). The disulfide bond is present in both apo- and Cu¹⁺-mb-OB3b. mb-OB3b is structurally the most complex mb characterized so far and is currently the only structurally characterized representative of this mb type. However, based on sequence similarity and alignments, the putative mbs from Methylosinus sp. strain LW3 (mb-LW3), Methylosinus sp. strain LW5 (mb-LW4), and Methylosinus sp. strain PW1 (mb-PW1) and one of the two mbs from M. parvus OBBP [mb-OBBP(1)] would also fall within this group (Fig. 4). In this group, the core peptide is predicted to contain 2 or more Cys residues in the mature peptide and/or to contain an additional ring(s). If the Cys residues are not incorporated into a ring, the second Cys and either the third or fourth Cys are predicted to form an intermolecular disulfide bond.

FIG 3.

Crystal structures of Cu¹⁺-mb-OB3b (A), Cu¹⁺-mb-M (B), and Cu¹⁺-mb-CSC1 (C), with a Thr residue missing at the C terminus in panels B and C compared to the largest form isolated. The copper ions are represented as gray spheres, and the oxazolone (oxa) and pyrazinedione (pyra) rings are labeled, as are the coordinating atoms. Hydrogen-bonding interactions are shown in panels B and C as dashed orange lines.

FIG 4.

mb precursor peptides. Sequences were detected in bacteria with known genome sequences, in methanotrophs with characterized mbs as well as from other methanotrophs, and from selected nonmethanotrophs, directly from the DNA sequence using Fuzztran in Mobyle (http://mobyle.pasteur.fr), with the optimized mb sequence motif (Prosite format) [ILMV]-[AIKST](1,3)-[IV]-[KNRT]-[IV]-X-[AKQ]-[KRT]-X-[ILM]-X-[IV]-X-[GV]-R-X(2)-[AL]-X-C(1,2)-[GA](0,1)-[ST](0,2)-X(0,2)-C(1,2) as a query. Only sequences also featuring downstream mb biosynthesis cassette genes mbnB and mbnC (see the text and Table 2) are shown. Known and predicted leader sequences and sequences not observed in the final product are shown in black, sequences detected in bacteria of known genome sequence from methanotrophs with mbs either whose crystal structure has been determined (mb-OB3b) or whose primary structures were predicted (mb-SB2 and mb-rosea) are shown in red, sequences detected in bacteria of known genome sequence from methanotrophs are shown in blue, and sequences detected in bacteria of known genome sequence from nonmethanotrophs are shown in green. Amino acids observed in some but not all samples are shown in tan. Amino acids with a gray background represent the amino acid pair shown or predicted to be posttranslationally modified into an oxazolone, imidazolone, or pyrazinedione group in structurally characterized mbs; underlined Cs represent Cys residues known or predicted to be present in the mature peptide. mbs were from M. trichosporium OB3b (mb-OB3b), Methylosinus sp. strains LW3 (mb-LW3), LW4 (mb-LW4), and PW1 (mb-PW1), M. parvus OBBP (mb-OBBP), M. rosea (mb-rosea), Methylocystis strains SB2 (mb-SB2), SC2 (mb-SC2), and LW5 (mb-LW5), Cupriavidus basiliensis B-8 (mb-B-8), Pseudomonas extremaustralis 14-3 (mb-14-3), Azospirillum sp. strain B510 (mb-B510), Tistrella mobilis KA081020-065 (mb-mobilis), Comamonas composti DSM 21721 (mb-21721), and Gluconoacetobacter sp. strain SXCC (mb-SXCC). Numbers in parentheses after a strain name are used for species with more than one mb gene cluster. Stop codons are indicated by an asterisk and gaps in the sequence alignment by a hyphen.

The second group (group II) is represented by the isolated mbs from M. rosea (mb-rosea) and Methylocystis strain SB2 (mb-SB2), which show high similarity to the structurally characterized (133) mbs from M. hirsuta CSC1 (mb-CSC1) and Methylocystis strain M (mb-M). In this group, mbs lack the Cys residues in the mature peptide and are smaller and probably less rigid due to the absence of the disulfide bond found in mb-OB3b. Heterocyclic rings are separated by two or three amino acids. In contrast to members of group II mbs from methanotrophic organisms, putative mbs from nonmethanotrophs (mb-B-8, mb-14-3, mb-B510, and mb-21721) contain four Cys residues in the apo-protein. However, based on the location of Cys residues, we predict that all 4 Cys residues are incorporated in the two heterocyclic rings. mbs from the structurally characterized members in this group contain a sulfate group, which appears to aid in the formation of a tight bend in the molecule (Fig. 3B) by making a hydrogen bond with the backbone amide of Ser2. The sulfate group also increases affinity for copper ions (133). Overall, mbs from the Methylocystis strains display a hairpin-like shape (Fig. 3B and C) (133). The conserved T/S adjacent to the putative C-terminal ring suggests that the other members of this group also contain a sulfate group (Fig. 4).

Truncated forms have been identified for all characterized mbs (133, 135, 138), which results from the loss of one or more C-terminal amino acid residues (Fig. 1 and 4). The different forms show very similar copper binding properties (133, 135) but distinct reduction potentials (133, 135) and spectral properties (138). The mechanism leading to the loss of C-terminal amino acids remains an open question but does not appear to involve N-protonation as observed for microcin B17 (163), since the C-terminal oxazolone ring of mb remains intact. In microcin B16, this reaction results in autoproteolysis of the ring and protein splicing. In Methylocystis strains, the sulfate group can also be lost, which does affect copper affinity (133).

mbs thus have a number of unusual structural features. Modifications of two residues to form oxazolone or imidazolone rings are rare but have been observed in several classes of ribosomally synthesized and posttranslationally modified peptide (RiPP) secondary metabolites (163–166). Pyrazinedione rings are even more uncommon and have been detected only in the non-amino-acid-containing molecules selerominol (167) and flutamide (168) from fungi. The presence of thioamide groups in natural products is also rare and has been identified only in closthioamide from Clostridium cellulolyticum (153, 160) and thioviridamide from Streptomyces olivoviridis (169, 170). Although rare, O-sulfonation of Ser or Thr has been reported in eukaryotic proteins (171), but to our knowledge, mb is the only bacterial peptide that contains this posttranslational modification and is ribosomally produced, as described below.

Copper Coordination Site

In the three mbs whose crystal structures have been determined, Cu¹⁺ is coordinated in a similar manner by an N₂S₂ ligand set with a distorted tetrahedral geometry (Fig. 3) (133–135). The C-terminal S and N ligands in the mb-CSC1 and mb-M are switched compared to their position in mb-OB3b (133); however, the resulting coordination geometry is very similar. The coordination sites are stabilized by the two 5-membered chelate rings formed upon ligation and a number of hydrogen bonding and π interactions (Fig. 3), for example, between the backbone amide of Cys3 and a coordinating sulfur in Cu¹⁺-mb-OB3b and a π-anion interaction between the sulfate and the pyrazinedione ring in mb-CSC1 and mb-M (133, 135).

METAL BINDING PROPERTIES

Binding and Reduction of the Primary Metal, Copper

mb binds both Cu²⁺ and Cu¹⁺, and the binding by mb appears to depend on pH (135, 172) and on the ratio of Cu²⁺/Cu¹⁺ to mb (136, 141, 172) (Table 1). As discussed below, mb is able to complex both soluble and insoluble forms of Cu¹⁺ and Cu²⁺. Thermodynamic, spectral, and kinetic studies have been carried out on the addition of Cu²⁺ to mb-OB3b and mb-SB2 (136, 141). These studies (136, 141, 172) are complicated by the fact that mb rapidly reduces Cu²⁺ to Cu¹⁺. Since the oxidation state of copper responsible for the high bonding constant for experiments in which Cu²⁺ has been added to apo-mbs is not known, we will indicate that a mixture of these oxidation states is present by using “Cu²⁺/Cu¹⁺” from this point on. At low ratios of Cu²⁺/Cu¹⁺ to mb-OB3b, mb-OB3b initially binds Cu²⁺/Cu¹⁺ as an oligomer/tetramer (136, 141). Pre-steady-state kinetic data also suggest that the initial binding of metal is on only one of the rings and its associated thioamide. In mb-OB3b, initial Cu²⁺/Cu¹⁺ coordination is to the oxazolone A, followed by a short (8- to 10-ms) lag period and then coordination to oxazolone B (141). At higher ratios of Cu²⁺/Cu¹⁺ to mb-OB3b, mb-OB3b coordinates Cu²⁺/Cu¹⁺ as a dimer, followed by a monomer at Cu²⁺/Cu¹⁺-to-mb-OB3b ratios above 0.5 Cu²⁺ per mb-OB3b. mb-SB2 appears to follow a similar tetramer-dimer-monomer binding sequence depending on the copper-to-mb ratio. For mb-SB2, the initial binding of Cu²⁺/Cu¹⁺ is to the imidazolone ring, followed by coordination to the oxazolone ring (154).

A number of different methods have been used to determine metal binding affinity constants for mbs (Table 1). Affinities for Cu¹⁺ can be determined from competition studies using a well-established approach (173–175) with chromophoric ligand such as bathocuproine disulfonate. The measurement of the reduction potential of the Cu-mb then allows the Cu²⁺ affinity to be calculated. The Cu¹⁺ affinity is ∼10²¹ M⁻¹ for all mbs analyzed using this approach (133, 135) and is one of the highest known for biological systems. Such a high affinity raises the issue of how copper is released from mb within a methanotroph. The reduction of the disulfide, in mb-OB3b (135), and removal of the sulfate, in mb-CSC1 (133), do decrease the Cu¹⁺ affinity by ∼2 orders of magnitude, but the values are still high. Given that the Cu²⁺ affinities are significantly weaker (Table 1), oxidation may be utilized for release of the metal, which may be more likely for those Cu-mbs with lower reduction potentials, such as Cu-mb-CSC1 (133).

Other copper-chelating compounds, primarily for Cu²⁺, have also been used to estimate affinities (91, 125, 128, 133, 135), as well as isothermal titration calorimetry (136, 141) and displacement isothermal titration calorimetry (136) (Table 1). mbs have been shown to solubilize and bind insoluble forms of Cu¹⁺ under anaerobic conditions (141) and to extract Cu from copper minerals (176), humic material (177), and glass (178). The ability of mbs to extract copper from copper-containing minerals as well as other organic molecules should be considered in modeling bioavailable sources of copper (179–183).

Binding of Other Metals

In addition to copper, mb-OB3b and mb-SB2 have been shown to bind a number of transition and near-transition metals (137, 148, 154, 184) (Table 1). In general, the spectral properties following binding are unique for each metal and can be used in initial analysis of metal binding and in metal competition studies (137, 154). The primary function of mbs appears to be copper acquisition, and the capacity to bind other metals appears to be an inadvertent consequence of the Cu¹⁺ coordination site. The peptide backbone of mbs appears to influence the binding of other metals. For example, the order of metal binding preference by mb-OB3b is Cu²⁺/Cu¹⁺ = Hg²⁺ = Au³⁺ > Zn²⁺ > Cd²⁺ > Co²⁺ > Fe³⁺ > Mn²⁺ > Ni²⁺, whereas the metal binding preference for mb-SB2 is Au³⁺ > Hg²⁺ > Cu²⁺/Cu¹⁺ > Zn²⁺ > Cd²⁺ = Co²⁺ = Fe³⁺ > Mn²⁺ > Ni²⁺ (137, 154).

Although strong binding of other metals such as mercury by mbs may seem counterproductive for methanotrophic growth, it may actually be of some benefit. Specifically, by strongly binding mercury, mbs may significantly alter its speciation and bioavailability, thereby reducing its toxicity. Indeed, growth studies showed that in the presence of mercury, added as mercuric chloride, methanotrophic growth was completely inhibited, but growth did occur when mb was also added (184). Interestingly, such a protective effect was seen in a variety of methanotrophs when exposed to mercury and mb-OB3b. These results suggest that mb secreted by any methanotroph could serve to protect the broader microbial community from mercury toxicity (184).

Further indirect evidence of the importance of mb binding to metals other than copper was afforded by the observation that in M. trichosporium OB3b sMMO is expressed and active in the presence of both gold and copper. This finding was unexpected since, as discussed above, copper strongly represses expression of sMMO. Also, the amount of copper associated with biomass significantly decreased in the presence of gold, compared to when gold was not added (91). This suggests that mb-OB3b is limited in its ability to bind copper in the presence of gold. Indeed, if mb was preloaded with copper and added exogenously to M. trichosporium OB3b, sMMO expression was not observed, and copper levels associated with biomass increased (91). These findings suggest that although methanotrophs synthesize mb for copper uptake, such uptake may be compromised if other metals such as gold are also present due to their ability to compete effectively for binding to mb. These observations suggest that in such situations, sMMO may be expressed and active even in the presence of copper, which may alter the methanotrophic community structure as well as activity.

Metal Reduction

As mentioned above, another property exhibited by all mbs examined so far is their ability to reduce Cu²⁺ to Cu¹⁺ (133, 135, 136, 141, 185). In addition, mb-OB3b and mb-SB2 have also been shown to reduce Au³⁺ and Ag⁺ to Au⁰ and Ag⁰, respectively (137, 148). In the case of gold, no other oxidation states were observed. In contrast, no change in the oxidation states of Zn²⁺, Cd²⁺, Co²⁺, Fe³⁺, Mn²⁺, and Ni²⁺ was observed following binding by mb-OB3b and mb-SB2 (137, 148). Surprisingly, in the absence of an external reductant, both mb-OB3b and mb-SB2 have been shown to reduce multiple Cu²⁺and Au³⁺ ions, which raises the question of the electron donor for this reaction.

BIOSYNTHESIS OF METHANOBACTIN

Identification of the Polypeptide Precursor of Methanobactin

Although a potential peptide sequence precursor of mb-OB3b was identified following hydrolysis of the oxazolone rings (132), the question whether this peptide was assembled by a nonribosomal peptide synthetase or encoded by DNA and synthesized on the ribosome remained. Determination of the genome sequence of M. trichosporium OB3b helped resolve this question. BLAST searches using the predicted mb peptide precursor revealed a short open reading frame (ORF) with a perfect match at a location within the M. trichosporium OB3b genome sequence where no protein product had been predicted (Fig. 4). The genome region of the putative mb precursor matching sequence in M. trichosporium OB3b had a number of distinctive and striking features (Fig. 5; Table 2). These included (i) a precursor peptide composed of leader and core peptide sequences followed by a stop codon, as expected for posttranslationally modified peptide natural products (164), (ii) a potential cleavage site between the leader and core peptide, suggestive of secretion, and (iii) genes upstream and downstream of the mb precursor gene encoding protein sequences compatible with possible roles in maturation of the mb precursor sequence, transport, and regulation of mb biosynthesis (Fig. 5, Table 2).

FIG 5.

mb gene clusters. Gene clusters of complete genomes of methanotrophs M. trichosporium OB3b (OB3b) and Methylocystis sp. strain SB2 and M. rosea SV97 (SB2/SV97), which produce mbs whose chemical structures have been determined/deduced (Fig. 1). Shown are the gene for the mb peptide precursor MbnA (black) and associated genes for the mb biosynthesis cassette protein MbnB (red), mb biosynthesis cassette protein MbnC (orange), MATE efflux pump MbnM (dark brown), aminotransferase MbnN (yellow), diheme cytochrome c peroxidase MbnH (dark green) and associated MbnP of unknown function (light green), FAD-dependent oxidoreductase MbnF (violet), sulfotransferase MbnS (dark blue), TonB-dependent receptor domain MbnT (light brown), FecI-like RNA polymerase sigma-70 domain MbnI (light blue), and FecR-like membrane sensor MbnR (light purple). Other genes of unknown function are shown in white, and genes flanking the clusters with predicted functions not thought to be associated with mb maturation or transport are shown in gray.

TABLE 2.

Genes in the methanobactin gene cluster

Gene	Proposed annotation	Characteristics/comments	InterProScan ID
mbnA	mb peptide precursor	Short ORF with N-terminal leader peptide	NA
mbnB	mb biosynthesis cassette protein MbnB	TIM barrel enzyme (aldolase/isomerase type)	IPR026432, IPR0078801
mbnC	mb biosynthesis cassette protein MbnC	Less well defined than MbnB	IPR023973
mbnE	FAD-dependent oxidoreductase	Often also found when mbnH and mbnP are missing	IPR003042
mbnH	Di-heme cytochrome c peroxidase		IPR023929, IPR004852
mbnI	RNA polymerase sigma 70 domain	Putative σ70 factor; in some cases (e.g., strains SB2, SC2, SV97), methanobactin precursor peptide is the C-terminal peptide of MbnI
mbnM	MATE efflux pump	Multiantimicrobial extrusion protein	IPR004839, IPR005814
mbnN	Aminotransferase	Rarely found (e.g., strains OB3b, LW4), pyridoxal-phosphate dependent	IPR023977
mbnP	Partner of mbnH	Conserved protein of unknown function
mbnR		FecR-like, putative sigma factor activator	IPR000863
mbnS	Sulfotransferase	Rarely found (e.g., strains SB2, SC2, SV97); sulfonation of Thr?	IPR012373
mbnT	TonB-dependent receptor	Plug domain, large membrane protein family	IPR000531

Elaboration on this initial search revealed a series of genomes containing gene clusters with characteristics matching those of the M. trichosporium OB3b mb gene cluster, e.g., in M. parvus OBBP (186, 187) and Methylosinus sp. strain LW3 (186), as well as nonmethanotrophs Azospirillum sp. strain B510 (132, 187), Azospirillum sp. strain B506 (186), Pseudomonas extremaustralis (187), Pseudomonas extremaustralis subsp. laumondii TT01 (186), Tistrella mobilis (187), Gluconacetobacter sp. strain SXCC (132, 186). Gluconacetobacter oboediens (186, 187) Methylobacterium sp. strain B34 (186), Cupriavidus basilensis B-8 (186), Photorhabdus luminescens (186), and Vibrio caribbenthicus BAA-2122 (186). Notably, this list includes not only methanotrophic strains known to produce mb for which a genome sequence was available but also genomes of other strains, suggesting that such gene clusters may encode peptide-derived products with more diverse functions than those associated with mbs.

Considering the mb precursor peptides identified so far, it is striking that the leader peptides are much more strongly conserved than core peptide sequences (Fig. 4). It is also interesting to note that sequence conservation does not follow taxonomic affiliation and that identical precursor sequences are found in gene clusters of strains from different genera (Fig. 4). This strongly suggests that horizontal gene transfer has contributed to dissemination of mb gene clusters in the environment, and indeed, mobile genetic elements have often been identified in the sequences immediately adjacent to mb gene clusters (186).

The lack of sequence conservation in mb core peptides also implies that designing sequence motifs to detect mb gene clusters is challenging and that database hits using sequence motifs based on known mb precursor sequences need to be substantiated with other evidence. In our view, for a sequence hit to be considered indicative of an mb precursor, additional evidence should include (i) the presence of the only conserved genes in the mb gene cluster (Fig. 5), i.e., the genes for mb biosynthesis cassette proteins MbnB and MbnC, in the immediate vicinity of the mb peptide precursor and (ii) the presence of 2 or more cysteines in the core mb peptide sequence (Fig. 4; Table 2). Regarding the latter, it is striking that mb peptide precursors identified so far contain either 2 or at least 4 cysteines, just as in the two types of mbs characterized so far. Nevertheless, the spacing between cysteine residues in the mb peptide precursor sequence appears to be quite variable, with 1 or 2 residues between the first and the second cysteines and, if present, 2 to 4 residues between the second and the third cysteines and 1 to 3 residues between the third and the fourth cysteines, respectively (Fig. 4). It is also striking to note that “doublets” of two cysteine residues immediately adjacent to each other have so far been found only in putative mb precursor peptides of presumably nonmethanotrophic strains (Fig. 4).

The overall structure of mb gene clusters and related gene clusters is beyond the scope of this review, but their analysis, using a combination of sequence-based, motif-based, and gene synteny-based searches, will undoubtedly yield many interesting findings in the future, as more mb gene clusters are identified in the increasing number of microbial genomes that are being sequenced. Summarizing the quite large diversity of mb gene clusters that have already been detected (see, e.g., reference 186), two major types of mb gene clusters seem to emerge with respect to the localization of the mbnA gene for the mb precursor peptide (Fig. 5). Whereas mbnA is always directly upstream of and in the same orientation as mbnB and mbnC, it may be found as a short ORF, as exemplified for strain OB3b (Fig. 5), or actually represent the 3′ end of regulatory gene mbnI, as in Methylocystis strain SB2 and M. rosea (strain SV97) (Fig. 5). The implications of the latter localization of the mbnA gene, notably in terms of regulation of mb expression in strains with this gene arrangement (also see Regulation of Gene Expression below), are completely unknown at present. Currently, the cytochrome c peroxidase MbnH (Fig. 5; Table 2), as well as the FAD-dependent oxidoreductase MbnF, present instead or sometimes in addition to MbnH in methanotroph gene clusters (186, 187), are likely candidates to be involved in the oxidation steps required for ring formation (see below) (Fig. 6). In addition, the aminotransferase MbnN, found in the mb-OB3b but not the mb-SB2/mb-rosea gene cluster (Fig. 5), may be involved in formation of the N-terminal keto-isopropyl group from Ile1 in the peptide precursor, and the sulfotransferase MbnS, found in the mb-SB2 and mb-rosea but not the mb-OB3b gene cluster, may catalyze sulfonation of the threonine (Fig. 1 and 5). Two other gene products, TonB receptor and multidrug and toxin extrusion (MATE) protein) (Fig. 5; Table 2) have been suggested (186, 187) to be involved in uptake and secretion of mature mbs, respectively. See a recent review by Kenney and Rosenzweig (186) for a detailed description and phylogenetic analysis of several mb gene clusters.

FIG 6.

Proposed reaction schemes for biosynthesis of the oxazoline rings with associated thioamide groups via a tandem two-step sequence of peroxidation and dehydration reactions. Cysteine thiols are likely protected against oxidation, possibly as disulfides involving one of the proteins of the mb gene cluster (blue circles). For imidazolone and pyrazinedione ring formation, oxazolone rings are modified via a transamination/deamination step followed by an aminolysis step to open the oxazolone ring followed by ring formation and dehydration.

Current Hypotheses on the Nature of the Methanobactin Biosynthesis Pathway

Despite the identification of the genetic determinants of mb production, the precise mechanism involved in maturation of mbs from peptide precursors still is an open question. The presence of heterocyclic rings would suggest a pathway similar to that of other postribosomal peptide synthesis (PRPS) proteins (163–166, 188). For instance, in the PRPS protein microcin B17, McbB (cyclodehydrase), McbC (flavin mononucleotide [FMN] dehydrogenase), and McbD convert Gly-Cys and Gly-Ser dipeptide sequences into thiazole and oxazole rings, respectively (163). In this reaction sequence, the cyclodehydrase catalyzes ring formation via the amine bond with the thiol or alcohol, followed by oxidation by an FMN dehydrogenase (163–166). In mb, ring formation is initiated from an X-Cys dipeptide sequence, resulting in formation of either an oxazolone, imidazolone, or pyrazinedione ring with a neighboring thioamide group. If the catalytic sequence in ring formation in mb followed the microcin B17 example, the thiol group on Cys from mb would have to be replaced by a hydroxyl group, possibly with an amine intermediate. The thioamide group then would have to be introduced via an amide-to-thioamide replacement, as proposed for two other natural products with thioamide groups (153), closthioamide from Clostridium cellulolyticum (160) and thioviridamide from Streptomyces olivoviridis (169, 170). As an alternative, the thioamide group of mbs was proposed to originate from the Cys thiol, as hydrolysis of the oxazolone rings resulted in an X-Gly-thioamide sequence (132). Based on the precursor peptide sequences that have now been identified, the latter reaction mechanism now appears highly likely.

The absence of a cyclodehydrase-like gene in mb gene clusters also suggest that ring formation differs from that of other natural products in other aspects as well. The simplest and shortest reaction sequence in oxazolone ring formation in mbs would involve an oxidation followed by a rearrangement that changes the connectivity of the peptide backbone (Fig. 6). In such a scheme, Cys thiols would likely need to be protected against oxidation (blue circle), as protein-linked thioesters, or alternatively as disulfides (as shown in Fig. 6), possibly via glutathionylation as observed with MetE in Escherichia coli (189). This reaction may occur during the initial reaction of peroxidase, which is one of the proteins of the mb gene cluster (Table 2). The reaction sequence could operate during formation of both oxazolone rings of mb-OB3b, from Leu1 and Cys2 and from Pro7 and Cys8 of the precursor peptide, respectively. Methylocystis strain SB2 and M. rosea share the same mb gene clusters and identical precursor peptides (Fig. 4 and 5), but the mature mb-SB2 (132) as characterized is suggested to contain an imidazolone ring, whereas mb-rosea is predicted to have a pyrazinedione ring that is also present in mb-CSC1 and mb-M (133). However, imidazolone and pyrazinedione rings are actually isomeric and can be interconverted by a simple sequence of hydration, rearrangement, and dehydration reactions (190). Nevertheless, the mechanism of formation of this ring is likely to be more complex than that of oxazolone ring formation (Fig. 6). One possible sequence would replace the transamination reaction, proposed for mb-OB3b, with an oxidative deamination of the N-terminal amine, which would produce ammonia that could be used for subsequent aminolysis of the oxazolone ring (Fig. 6). A condensation reaction would then lead to cyclization and the formation of the imidazolone or, with an intervening rearrangement step, to the pyrazinedione. It should be noted that the Methylocystis species lack the aminotransferase found in the mb-OB3b gene cluster. Alternate reaction schemes would involve two changes of connectivity in the peptide backbone and not just one as proposed for oxazolone ring formation (Fig. 6). Clearly, the formation of the different types of heterocyclic ring systems in mbs now needs to be addressed experimentally. The production and characterization of mutants will hopefully help in this respect, and such studies have now been initiated (187).

REGULATION OF GENE EXPRESSION

The discovery of the mb biosynthesis gene cluster raised two main questions. First, is mb biosynthesis regulated with respect to copper? Second, what is the role of mb in the “copper switch” controlling the expression of the two forms of MMO? The first question was easily answered through the use of reverse transcription-quantitative PCR (RT-qPCR) using primers specific to some gene(s) of the mbn operon. Initial studies found that in M. trichosporium OB3b, expression of mbnA, the gene encoding the polypeptide precursor of mb, did indeed vary with respect to copper, with expression dropping over three orders of magnitude when copper concentrations increased from zero (no amendment) to 1 μM copper, and expression was largely invariant at higher copper concentrations (140). Such a decrease in expression was reflected in the finding that mb in the spent medium was highest at copper concentrations less than 1 μM and dropped ∼5-fold when the copper concentration was increased to 5 μM (136, 138, 141). Given that the two forms of MMO are also regulated by copper, it was possible that mb was directly involved in this copper switch. Further genetic analyses, however, found that it is not (187). RT-qPCR of mmoX and pmoA from both wild-type M. trichosporium OB3b and an mbnA mutant where a gentamicin cassette was inserted, knocking out mbnA, showed that mmoX and pmoA expression followed similar patterns in both strains as the copper/biomass ratio varied but that the magnitudes of such changes were lower in the mbnA::Gm^r mutant than in the wild-type strain (187). To answer the second question above, a mutant in which the mmo operon was deleted was constructed. In this mutant, the “copper switch” was inverted; i.e., pmoA expression was greatest in the absence of copper and dropped ∼2 orders of magnitude when copper concentrations increased (187). Further, expression of mbnA was largely invariant in this mutant with respect to copper. Since all deleted genes in the mmo operon with the exception of mmoD are known to encode polypeptides of sMMO and since MmoD is not required for sMMO activity (32, 51, 63), these expression data suggest that MmoD is a key component of the copper switch, with MmoD serving to regulate expression of mb and mb then amplifying the magnitude of the response in the copper switch.

On the basis of these findings, a model for the regulation of expression of the mmo and pmo operons by mb and MmoD was proposed (187). During growth at low copper/biomass ratios, MmoD protein represses expression of the pmo operon and also upregulates expression of the mmo operon, including that of mmoR and mmoG, which were shown previously to play key roles in regulating mmo expression (113). Further, MmoD is postulated to enhance expression of mb, which serves to amplify expression of the mmo operon. During growth with high copper concentrations, mb binds copper and can no longer enhance the expression of the mmo operon. Under these conditions, it is presumed that MmoD also binds copper and no longer represses expression of the pmo operon or induces expression of sMMO or mb.

There are, however, a number of issues with the model as proposed in Fig. 7. For such a model to be accurate, mmoD would be required to be constitutively expressed with respect to copper in order for sMMO and mb expression to occur after copper is removed. This version of the model implies, however, that as the copper/biomass ratio increases, expression of the entire mmo operon no longer occurs. Recent qRT-PCR analysis, however, shows that mmoD, unlike mmoX, is constitutively expressed with respect to copper (J. D. Semrau, unpublished data). Nevertheless, expression of mmoD, like that of mmoX, increases with increasing addition of exogenous mb, suggesting that expression of mmoD is at least partly coregulated with that of the entire mmo operon.

FIG 7.

Model for the regulation of gene expression in the mmo, pmo, and mbn gene clusters by copper, mb, and MmoD. (A) Low copper/biomass ratio; (B) high copper/biomass ratio. (Reproduced from reference 187 with permission of the publisher.)

In addition, this model, although explaining the genetic basis for the copper switch in methanotrophs, does not explicitly consider the role of two genes known to play roles in the regulation of the mmo operon, i.e., mmoR and mmoG upstream of mmoX (113, 191). Based on available data, it appears that mmoR and mmoG do not play a significant role in the copper switch; i.e., they do not appear to have any control over expression of the pmo operon. It is possible, however, that regulation of the mmo operon is the result of a concerted interaction of mb and/or MmoD with these gene products.

This model also ignores the possibility that the presence of putative regulatory genes (e.g., mbnI and mbnR [Fig. 5]) in mb gene clusters may play a role in regulating gene expression, particularly the mb gene cluster and/or the mmo operon. Given this, a revised model of the copper switch in methanotrophs is now proposed in Fig. 8. Here MmoD is again postulated to be a key component of the copper switch, but MbnI is proposed to be responsible for inducing expression of the mb gene cluster, as well as mmoR and mmoG. Collectively, mb, MmoR, and MmoG interact to induce expression from the σ^N promoter upstream of mmoX, while MbnI binds to the σ⁷⁰ promoter upstream of mmoY. In this updated and revised current model, mmoD is constitutively expressed but associates with copper when it is present, preventing repression of the pmo operon or expression of the mb gene clusters. Further experiments will show whether this new model is correct, but it is already clear that the exact details of the copper switch mechanism are more complex than previously thought.

FIG 8.

Revised regulatory scheme for expression of the mmo, pmo, and mbn gene clusters as a function of copper, mb, and MbnI. (A) Low copper/biomass ratio; (B) high copper/biomass ratio.

PHYSIOLOGICAL FUNCTIONS

Copper Acquisition

Although mb has a high affinity for copper, its importance in copper acquisition by methanotrophs is still unclear. Nonnative Cu¹⁺-mbs have been shown to be taken up by methanotrophs and facilitate switchover to pMMO (133). Copper uptake and switchover are faster if the native mb is used but still take more than 24 h. However, the role of mb in meeting the overall copper requirements of the cell is still in question. It has been found that M. trichosporium OB3b possesses at least two mechanisms for copper uptake, one clearly based on mb and involving active transport of copper-mb complexes and another, nonspecific passive transport pathway (192). It was also found that the amount of copper associated with biomass was the same regardless of whether copper was added as CuSO₄ or as the Cu¹⁺-mb complex at the same concentration. Subsequently, it was found for M. trichosporium OB3b that when the gene for the precursor polypeptide of mb (mbnA) is knocked out, copper is still associated with biomass, and it increases with increasing amounts of copper in the growth medium, as does copper found associated with wild-type cultures. The role of mb in copper uptake thus remains unclear. It should be kept in mind that these studies were performed using a well-defined growth medium with limited diversity of copper speciation and little if any copper-containing precipitate. It is tempting to speculate that mb may have a more significant role in copper uptake in situ, where copper speciation and distribution will be much more complex and will include copper associated with a wide range of organic materials (e.g., humic and fulvic acids), as well as found either sorbed onto or part of various mineral phrases. Indeed, it has been shown that expression of pmoA is strongly dependent on the form of copper present, with copper associated with metal oxides found to induce smaller changes in pmoA expression, and that concomitant addition of mb increased the magnitude of the copper switch in M. trichosporium OB3b (176, 193).

Do Methanobactins Have a Role as Signaling Molecules?

Given the small size of mbs and their secretion into the environment when copper availability is low, a role for mbs as signaling molecules may be envisaged. Indeed, addition of mb-OB3b enhances expression of mmoX in in cultures of both wild-type M. trichosporium OB3b and its mbnA::Gm^r mutant (184, 187), suggesting that mb may control gene expression by acting as a signaling molecule. Further, given that all known forms of mb have significant structural similarity, mbs may allow for cross-species communication. Recent studies show that this is indeed the case. mb-SB2 increased both mmoX expression and sMMO activity in M. trichosporium OB3b in the presence of copper. However, it had no significant effect (P > 0.05) on expression of either pmoA or mbnA, nor did it increase the amount of cell-associated copper (228). The addition of mb-SB2 preloaded with copper, however, reduced both mmoX and mbnA expression when M. trichosporium OB3b was grown in the absence of any added copper, and no sMMO activity was detected. The latter results were likely due to the increased amount of cell-associated copper and may reflect the role of mb in copper uptake.

Membrane Development

Several studies have demonstrated a direct correlation between intracytoplasmic membrane development and expression of pMMO in methanotrophs (33, 35, 66, 109, 114, 194). As stated above, mbnA plays a secondary role in the regulation of both the membrane-bound and soluble MMO (see Regulation of Gene Expression above). However, deletion of mmoD along with the rest of the sMMO operon had little effect on intracytoplasmic membrane development in M. trichosporium OB3b (Fig. 9C and D). In contrast, a deletion mutation in the mbnA structural gene for mb appears to affect intracytoplasmic membrane development well beyond its effects on pMMO expression (187) (Fig. 9E and F). Currently, knowledge of the effect(s) of mb on intracytoplasmic membrane development is observational, and additional research is required to determine whether mb is directly involved in the development of these membranes in methanotrophs.

FIG 9.

Transmission electron micrographs of wild-type M. trichosporium OB3b (A and B), an M. trichosporium OB3b sMMO deletion mutant (C and D), and an M. trichosporium OB3b mb deletion mutant (E and F) cultured in nitrate mineral salts (NMS) medium (A, C, and E) or NMS medium amended with 5 μM CuSO₄ (B, D, and F).

Detoxification of Reactive Oxygen Species

Copper is one of the essential metals required by microorganisms for growth (179). Like iron, however, copper in oxygenated solutions can generate a variety of toxic reactive oxygen species via Fenton and Haber-Weiss reactions (195–197). The coordination of copper by mb is likely to reduce this toxicity, as observed with the iron- and copper-binding siderophore schizokinen (146). Nevertheless, both copper-bound and copper-free forms of mb-OB3b have been shown to reduce dioxygen to superoxide in the presence of a reductant (198). In general, this would represent a daunting problem for bacteria that accumulate high concentrations of Cu-mb. Perhaps not unexpectedly, the superoxide dismutase (SOD) activity of Cu-mb is approximately 30,000 times higher than the oxidase activity associated with either mb-OB3b or Cu-mb-OB3b. In addition, in the presence of a reductant, Cu-mb-OB3b also has a hydrogen peroxide reductase (HPR) activity which is approximately 50 times the oxidase activity (198). The overall result of these three activities is the reduction of O₂ to H₂O with concomitant loss of reductant. In addition to their protective roles in the context of mb-catalyzed oxidations, high SOD and HPR activities of mb are also likely to provide protection against oxygen radicals that are potentially formed during methane oxidation and respiration of dioxygen (196, 199, 200) and may thus explain the stabilizing effects of mb-OB3b on pMMO activity in cell extracts (66, 73, 198). Similar activities have also been observed for mb-SB2 (N. L. Bandow et al., unpublished results), and we predict that this is a property of all mbs.

MEDICAL, INDUSTRIAL, AND ENVIRONMENTAL IMPLICATIONS

Metal Mobilization or Immobilization for Remediation of Polluted Subsurface Environments

Based on the published yields of mb in laboratory cultures (136, 138), methanotrophs are predicted to export 3 to 50 copies of mb per cell per second depending on the copper concentration during growth. Given this high export rate and the fact that mb can bind most transition and near-transition metals, it is likely that mb is able to influence metal mobilization and thus bioavailability of different metals in soil systems and that this affects the structure and functioning of microbial communities. Recent studies have shown that mb will bind both inorganic and organic forms of mercury (154) and thereby protect the host bacterium as well as other bacteria from mercury toxicity (184). However, additional studies are required to determine which metals are affected in their mobility by mbs and the effects of mbs on the soil microflora.

Formation of Gold Nanoparticles

The potential application of gold nanoparticles as anti-inflammatory agents, in the targeting of cancer cells, in the detection of melamine, as biosensors, and in glucose oxidation has motivated the development of various production methods for such materials (201–206). Most approaches require either high (>100°C) or low (−18°C) temperatures in the presence of reducing and stabilizing agents. Alternatively, microbially mediated production of gold nanoparticles, both intracellularly and extracellularly, has also been extensively investigated (207). However, the size distribution of gold nanoparticles produced by microbial processes is usually quite broad, ranging from 10 to 6,000 nm (208). Further, most reported processes yield particles with a wide variety of shapes (spherical, octahedral, triangular, irregular, etc.), and rates of formation vary over a large range, from minutes to hours (209). Crucially, both mb-OB3b and mb-SB2 have been shown to efficiently reduce Au³⁺ ions to metallic gold (typically >5 ions/mb) and in so doing to form well-defined size distributions of spherical nanoparticles (137, 148). In a further study with mb-SB2, gold spherical nanoparticles of well-defined sizes (2.0 ± 0.7 nm) were formed, with mb-SB2 suggested to act both as a reductant and as a stabilizing agent, thereby demonstrating that the use of mb-SB2 represents a viable approach to produce well-defined gold nanoparticles in this size range (137).

Possible Therapeutic Treatment of Human Ailments

Copper has been implicated as a key factor in the development of a variety of human ailments, perhaps most notably Wilson's disease (210, 211), in which affected individuals have mutations that inactivate the Cu-transporting P-type ATPase ATP7B (212), impeding copper export to the bile and causing copper accumulation in the liver. Based on in vivo and in vitro studies, copper accumulation specifically affects liver mitochondria, via copper-induced modification of membrane protein thiols leading to multiprotein cross-linking (213). In the final stages of Wilson's disease, oxidative damage, primarily in the liver, is observed. Other symptoms of Wilson's disease include serious neurological issues that can lead to additional psychiatric disorders (210, 214, 215).

There is as yet no cure for Wilson's disease. Currently, treatment consists of one or more of the following options: (i) the use of chelating agents to remove copper via enhanced urine excretion, (ii) implementation of a low-copper diet, and (iii) zinc supplements added to the diet to stimulate production of the copper-binding protein metallothionein (210, 214). Of these options, the use of chelating agents is the most commonly prescribed treatment, with two drugs, penicillamine and trientine, currently approved by the U.S. Food and Drug Administration (FDA). Penicillamine, a breakdown product of penicillin, binds copper (216) and also induces the production of metallothionein. Although the clinical benefit of penicillamine is widely reported, it has serious side effects, including bone marrow suppression, degeneration of elastic tissue, and proteinuria. In a large fraction of cases (20 to 50%), neurological deterioration is observed (217). In such cases, the polyamine trientine, which has fewer reported side effects than penicillamine (217) and a comparable if not higher affinity for copper, is substituted. In both penicillamine and trientine treatments, copper is excreted through the urine and not the bile. Another chelating agent, tetrathiomolybdate, allows for copper to be excreted from liver primarily into the blood, but some evidence suggests that part of the copper may also be transferred to the bile (219) and that insoluble copper is found in both liver and kidneys at tetrathiomolybdate/copper ratios greater than two (219–221). The physiological impact of such an increased burden of copper in the blood, as well as the formation of insoluble copper, is unclear.

mbs may represent a superior alternative as a copper-chelating agent to treat Wilson's disease, as indicated by preliminary studies. In the animal model for Wilson's disease, the Long Evans Cinnamon (LEC) rat, intravenous application of mb-OB3b extracted copper from the liver into the bile as Cu-mb-OB3b, without measurable side effects (213, 222). It was also found that mb-OB3b reversed both the buildup of copper in liver mitochondria and protein cross-linking, allowing the liver to function normally (213), and also stripped copper from metallothionein (222). Insoluble forms of copper were not observed. Lastly, and in contrast to other chelators used for the treatment of Wilson's disease, mb-OB3b was shown to work in circumstances of advanced liver failure, further suggesting that mb may be a viable alternative in the treatment of Wilson's disease in the future. It is also tempting to speculate that the use of mbs may be of interest in other copper-associated pathologies as well, including certain forms of cancer (223–226). For example, it was recently reported that active oncogenesis of a wide variety of cancers is directly affected by the presence of copper through the role of BRAF kinase (226), since disrupting kinase copper binding through site-directed mutagenesis or preventing copper uptake through the addition of tetrathiomolybdate inhibited tumorigenesis. In our view, these observations make it clear that mbs, first discovered as an elusive factor associated with the exclusively bacterial process of methane utilization, may hold great promise as molecules useful to humankind as well.

CONCLUSIONS

The combination of a heterocyclic ring and an associated thioamide is a unique structural feature of mbs among all natural products and the primary cause of the capabilities observed for these molecules. Although rare, oxazolone and imidazolone rings have been identified in several ribosomally synthesized and posttranslationally modified natural products, and this feature appears to be responsible for the observed antimicrobial and metal binding properties of such molecules. Pyrazinediones, in contrast, have so far been observed only in two nonribosomally synthesized natural products, selerominol and flutamide, which show activity against several cancer cell lines and antiviral activity, respectively. Like for thioamide groups, they also confer antimicrobial and metal binding properties in the two natural products identified so far that contain this functional group. Crucially, combining both of these structural features in small peptide derivatives such as mbs appears to enhance various properties, including copper acquisition, metal detoxification, detoxification of reactive oxygen species, electron transfer, nanoparticle formation, and gene regulation. Production of high concentrations of mbs by a ubiquitous microbial group also raises a number of ecological questions such as its effects on metal bioavailability and mobilization, as well as on the structure and functioning of microbial communities in the environment. From an application viewpoint, mbs now have demonstrated potential for the production of nanoparticles as well as in the treatment of copper-related diseases, and we predict that a variety of other commercial and medical applications have yet to be identified.

Biographies

Alan A. DiSpirito received his B.S. in biology from Providence College and his M.S. and Ph.D. in microbiology from The Ohio State University with Olli H. Tuovinen and Patrick Dugan. His postdoctoral training was with Alan B. Hooper at The University of Minnesota and with Mary E. Lidstrom at The University of Washington—Seattle, Center for Great Lakes Studies, and California Institute of Technology. He was an Assistant Professor at The University of Texas at Arlington before moving to the Department of Microbiology at Iowa State University. He is currently a Professor in the Roy J. Carver Department of Biochemistry, Biophysics and Molecular Biology at Iowa State University. His basic research interest is in the basic metabolism, environmental importance, and commercial application of methanotrophic and chemoautotrophic bacteria. Current research projects include structural and functional characterization of methanobactins, the mechanism of methanobactin biosynthesis, and the influence of methanobactin on metal mobility in soil systems.

Jeremy D. Semrau, Arthur F. Thurnau Professor, has appointments in the College of Engineering and the College of Literature, Science and Arts at the University of Michigan. His research focuses primarily on methanotrophs, with a particular emphasis on the biochemistry, genetics, and regulation of methane monooxygenase and methanobactin synthesis. As part of his research, Professor Semrau has also isolated and characterized novel facultative methanotrophs, developed applications of methanotrophy for use in bioremediation and biofuel production, and developed novel applications of methanobactin for the production of gold nanoparticles and the treatment of Wilson's disease. He holds several patents in these areas and has published many papers, abstracts, and book chapters on fundamental and applied aspects of methanotrophy. Professor Semrau is currently on the editorial board of Applied and Environmental Microbiology.

J. Colin Murrell is Professor in Microbiology in the School of Environmental Sciences and Director of the Earth and Life Systems Alliance (ELSA) at the University of East Anglia, United Kingdom. He has wide-ranging research interests centered around the bacterial metabolism of methane and other one-carbon compounds, such as methanol, methylamines, methanesulfonate, dimethyl sulfide, methyl halides, and isoprene, in the terrestrial, aquatic, and marine environments. Other areas of research include the microbiology of the rhizosphere, the sea surface microlayer, caves, alkaline soda lakes, salt marshes, and cold-water corals and cultural heritage microbiology, regulation of gene expression by metals, microbial genomics, metagenomics, bioremediation, biocatalysis, and industrial biotechnology. His research over the past 35 years has resulted in around 280 publications and six edited books. Dr. Murrell is a member of the editorial boards of Environmental Microbiology, The ISME Journal, and FEMS Microbiology Letters and was the Chair of the 2015 Gordon Research Conference on Applied and Environmental Microbiology. Dr. Murrell is Vice President of The International Society for Microbial Ecology and was recently elected Member of the European Molecular Biology Organization and Member of the European Academy of Microbiology.

Warren H. Gallagher was born in Providence, RI, in 1952 and received his A.B. degree in chemistry and biology from Albion College, Albion, MI, in 1974. He went on to receive his Ph.D. in biophysics from the University of Pittsburgh in 1980 under the direction of Professor Max Lauffer. He then held postdoctoral positions with Professor Victor Bloomfield and Professor Clare Woodward at the University of Minnesota, Twin Cities, before joining the chemistry faculty at the University of Wisconsin—Eau Claire, where his current title is Professor and Chair of Chemistry. He has a long-term interest in using spectroscopic methods to determine the structures of peptide-derived molecules, and over the past several years the methanobactins have presented him and his students with some exciting challenges that demanded detective-like solutions.

Christopher Dennison received his Ph.D. in inorganic chemistry (1994) from Newcastle University with Professor A. G. Sykes and Professor W. McFarlane and then worked as a postdoctoral fellow with Professor G. W. Canters at Leiden University. He was appointed as a Lecturer at University College Dublin in 1997 and returned to Newcastle University in 1999 as a Lecturer in Inorganic Chemistry in the Department of Chemistry. In 2004 he became a Senior Lecturer in the Institute for Cell and Molecular Biosciences (Medical School) and in 2010 Professor of Biological Chemistry. He studies the role of metals, and particularly copper, in biological systems. This has included the analysis of electron transfer proteins and multicopper redox enzymes. More recently, his work has focused on copper homeostasis in a wide range of organisms, including understanding how methane-oxidizing bacteria acquire, handle, and store copper.

Stéphane Vuilleumier studied natural sciences at ETH Zurich, Switzerland, and obtained a Ph.D. in organic chemistry at the University of Basel, Switzerland. After postdoctoral work in protein engineering at the University of Cambridge, United Kingdom, he returned to ETH Zurich, where he became a group leader at the Institute of Microbiology and obtained his habilitation. He has been a Professor of Environmental Biology and Microbiology in Strasbourg, France, since 2002, heads a CNRS research team, and shares responsibility for the Chemistry and Biology Master at Université de Strasbourg. In 2005, his interest in bacterial metabolism of halogenated methanes led him to become involved in international collaborative genome sequencing projects on methylotrophic and methanotrophic bacteria at JGI (United States) and Genoscope (France) and in their subsequent analysis using the MicroScope platform at Genoscope. Dr. Vuilleumier was elected Chair of the 2016 Gordon Research Conference “Molecular Basis of Microbial One-Carbon Metabolism.”