Oxygen Generation via Water Splitting by a Novel Biogenic Metal Ion-Binding Compound

ABSTRACT

Methanobactins (MBs) are small (<1,300-Da) posttranslationally modified copper-binding peptides and represent the extracellular component of a copper acquisition system in some methanotrophs. Interestingly, MBs can bind a range of metal ions, with some being reduced after binding, e.g., Cu²⁺ reduced to Cu⁺. Other metal ions, however, are bound but not reduced, e.g., K⁺. The source of electrons for selective metal ion reduction has been speculated to be water but never empirically shown. Here, using H₂¹⁸O, we show that when MBs from Methylocystis sp. strain SB2 (MB-SB2) and Methylosinus trichosporium OB3b (MB-OB3) were incubated in the presence of either Au³⁺, Cu², or Ag⁺, ^18,18O₂ and free protons were released. No ^18,18O₂ production was observed in the presence of either MB-SB2 or MB-OB3b alone, gold alone, copper alone, or silver alone or when K⁺ or Mo²⁺ was incubated with MB-SB2. In contrast to MB-OB3b, MB-SB2 binds Fe³⁺ with an N₂S₂ coordination and will also reduce Fe³⁺ to Fe²⁺. Iron reduction was also found to be coupled to the oxidation of 2H₂O and the generation of O₂. MB-SB2 will also couple Hg²⁺, Ni²⁺, and Co²⁺ reduction to the oxidation of 2H₂O and the generation of O₂, but MB-OB3b will not, ostensibly as MB-OB3b binds but does not reduce these metal ions. To determine if the O₂ generated during metal ion reduction by MB could be coupled to methane oxidation, ¹³CH₄ oxidation by Methylosinus trichosporium OB3b was monitored under anoxic conditions. The results demonstrate that O₂ generation from metal ion reduction by MB-OB3b can support methane oxidation.

IMPORTANCE The discovery that MB will couple the oxidation of H₂O to metal ion reduction and the release of O₂ suggests that methanotrophs expressing MB may be able to maintain their activity under hypoxic/anoxic conditions through the “self-generation” of dioxygen required for the initial oxidation of methane to methanol. Such an ability may be an important factor in enabling methanotrophs to not only colonize the oxic-anoxic interface where methane concentrations are highest but also tolerate significant temporal fluctuations of this interface. Given that genomic surveys often show evidence of aerobic methanotrophs within anoxic zones, the ability to express MB (and thereby generate dioxygen) may be an important parameter in facilitating their ability to remove methane, a potent greenhouse gas, before it enters the atmosphere.

INTRODUCTION

Aerobic methane-oxidizing bacteria (methanotrophs) oxidize methane to carbon dioxide via a series of two-electron steps with methanol, formaldehyde, and formate as intermediates (1). The initial oxidation of methane to methanol is an oxygen- and energy-dependent reaction and is catalyzed by either a soluble cytoplasmic methane monooxygenase (sMMO) or a particulate or membrane-associated methane monooxygenase (pMMO) (1–8). The reductant for the initial oxidation of methane is supplied by NADH for the sMMO and by quinols for the pMMO (2, 3, 9–11). Methanol is oxidized to formaldehyde by a calcium- or rare-earth-dependent methanol dehydrogenase using a c-type cytochrome as an electron acceptor (12–17). Formaldehyde is either assimilated or oxidized using either NAD⁺ or quinone as the electron acceptor (10, 18–20). The final two-electron oxidation of formate to carbon dioxide is catalyzed by the NAD⁺-linked formate dehydrogenase (21–23). Electrons from NADH, quinol, or cytochrome c are either utilized in biosynthetic reactions or oxidized for energy using either dioxygen (11), nitrate (24), or ferric iron (25) as the terminal electron acceptor.

In methanotrophs capable of expressing both forms of the MMO, expression is regulated by copper (1, 9, 26–28). In addition to the MMOs, a number of genes are regulated by copper (1), and some methanotrophs of the Alphaproteobacteria have novel copper acquisition systems based on the extracellular copper-binding peptide methanobactin (MB) (29–31). MBs are low-molecular-mass (<1,300-Da), high-potential (E_m of 483 to 745 mV) ribosomally synthesized posttranslationally modified peptides (RiPPs) and were the first examples of a chalkophore, i.e., a compound excreted by bacteria for the purpose of scavenging copper from the surrounding environment (30, 32). Structurally, MBs are divided into two groups. Both group I and II MBs are characterized by an internal oxazolone group with an associated thioamide and a second N-terminal 5- or 6-membered ring, which in group I MBs is either an oxazolone or a pyrazinedione group with an associated thioamide, while group II MBs have either an imidazoline or a pyrazinedione group with an associated thioamide (30, 32–36). The ring and associated thioamide are derived from an X-Cys dipeptide via a series of partially characterized posttranslational modifications (29, 31, 37). Group I MBs are characterized by an internal disulfide bridge and the copper-bound form of a dicyclic structure (32–34). Group II MBs lack this disulfide bridge, and the copper-bound form has a hairpin-like structure and is characterized by a central sulfonated threonine (30, 36).

In addition to copper ions, MBs will bind many metal ions (38–41) and reduce some but not all metal ions that are bound (38, 42). In MB from Methylosinus trichosporium OB3b (MB-OB3b), metal ions such as copper, silver, and gold are coordinated via an N₂S₂ ligand set utilizing an N from each ring and the two thioamides, and these metals are reduced after binding (30, 34, 38, 43). Other metal ions such as iron, nickel, and cobalt are coordinated via an N₁S₁ ligand set using one ring and its associated thioamide and are not reduced (38). Based on coordination, metals were classified as either group A metals coordinated by an N₂S₂ ligand set or group B metals coordinated by an N₁S₁ ligand set. In contrast, all of the metals bound by MB from Methylocystis sp. strain SB2 (MB-SB2) are coordinated by an N₂S₂ ligand set (39, 40, 42, 44; this study).

Since metal ion reduction assays are often carried out in unbuffered reaction mixtures in the absence of an external reductant, water has been proposed, but not shown, to be the electron donor (36). Here, we examine the binding and reduction of oxidized forms of gold (as HAuCl₄), copper (as CuCl₂), silver (as AgF), iron (as FeCl₃), nickel (as NiCl₂), mercury (HgCl₂), cobalt (as CoCl₂), and potassium (as KCl) in the presence and absence of H₂¹⁸O by MB-SB2 as well as the binding and reduction of gold, copper, and silver in the presence of H₂¹⁸O by MB-OB3b.

RESULTS

Spectral and thermodynamic properties of AuCl₄⁻ binding by MB-SB2.

UV-visible absorption, fluorescence, and circular dichroism spectra (see Fig. S1 to S3 in the supplemental material) and thermodynamic measurements (Fig. S4 and Table S1) demonstrate that changes following the addition of HAuCl₄ to MB-SB2 were complex, with transitions being apparent at 0.25, 0.5, 0.75, 1.0, and 2.0 Au per MB-SB2. As MB-SB2 has only one identified metal-binding motif (i.e., an N₂S₂ ligand set), we therefore interpret these data to indicate changes in Au coordination, when MB-SB2 transitions from an oligomeric state(s) to a monomer.

The increased fluorescence emission intensity following HAuCl₄ addition may be due to disruption of internal quenching between the imidazolone and oxazolone groups and is consistent with the intramolecular exciton transfer previously demonstrated following hydrolysis of the oxazolone group (42) (Fig. S2). The decreased fluorescence at HAuCl₄/MB-SB2 ratios of >1.0 suggests direct metal quenching or intra-/interexciton transfer between the oxazolone and imidazolone groups or may be associated with nanoparticle formation, which occurs at Au/MB-SB2 ratios of >1:1.

As Au nanoparticle formation requires Au³⁺ reduction (45), nanoparticle formation (Fig. S5 and Table S2) by MB-SB2 indicated that MB-SB2 binds and reduces multiple Au³⁺ molecules to Au⁰. Such findings accentuated the need to determine the electron source for metal reduction by MBs. To extend these preliminary studies, we examined the reduction of HAuCl₄ via MB-SB2 when dissolved in either H₂¹⁶O or H₂¹⁸O.

X-ray photoelectric spectroscopy, kinetics, and chloride determination.

In reaction mixtures containing HAuCl₄ and MB-SB2 dissolved in H₂¹⁶O, MB-SB2 was observed to reduce AuCl₄⁻ to Au⁰ plus 4Cl⁻, as determined by X-ray photoelectric spectroscopy (XPS) and argentometric titrations, respectively (Fig. 1 and Table 1). MB-SB2 reduced up to 19 Au³⁺ to 19 Au⁰ with a time-dependent average Au³⁺-to-Au⁰ reduction rate of 0.3 ± 0.06 min⁻¹ for those assays where rates could be determined. This time-dependent reduction was the reason why samples were frozen in liquid nitrogen and lyophilized overnight to stop the reaction and dry the samples for analysis. As observed with MB-OB3b (38), Au³⁺ and Au⁰ were the only oxidation states detected, indicating a direct three-electron reduction of HAuCl₄ (Fig. 1).

FIG 1.

Gold X-ray photoelectric spectra of MS-SB2 at a gold/MB-SB2 molar ratio of 14 to 1 after a 30-min incubation (red circles) (A) and at a gold/MB-SB2 molar ratio of 19 to 1 after a 30-min incubation (B). Experimental results (circles) were fit with CASA XPS software to four Gaussian/Lorentzian curves, using two peaks for Au³⁺ (orange curves) and two peaks for Au⁰ (blue curves). Gold 4f core electrons are spin-orbit split as 4f_7/2 and 4f_5/2, with a splitting of 3.7 eV and area ratio of 4:3, so that only two peaks are independently fit: the 4f_7/2 peaks for Au³⁺ and Au⁰. The 4f_5/2 peaks’ positions and areas are determined by spin-orbit splitting; these parameters and the peak widths are fixed in the fitting program. The background used was of a Shirley type.

TABLE 1.

Distribution of Au as Au³⁺ and Au⁰ following incubation of MB-SB2 and HAuCl₄^a

HAuCl4/MB-SB2 ratio	Time (min)	% Au3+	% Au0	Reduction rate (Au3+ reduced min−1)
0.9	30	0	100
2.25	30	0	100
9	30	8	92	0.27
14	30	11	89	0.41
19	30	59	41	0.26
9	60	0	100
14	60	0	100
19	60	10	90	0.28
19	360	0	100

^aReduction rates were determined from samples where <100% reduction was observed.

Kinetics of AuCl₄⁻ binding and reduction.

The time course for the binding of Au³⁺ to the oxazolone and imidazolone rings in MB-SB2 was measured as the decrease in the absorbances at 341 and 389 nm, respectively, following stopped-flow mixing of MB-SB2 with Au³⁺ at 4°C (Fig. 2A). Unfortunately, even at 4°C, initial binding rates could be determined only for the oxazolone ring since binding to the imidazolone ring was complete before the mixing of the sample was complete (1.4 ms). In contrast, the rates of binding to the oxazolone ring were low, 12 to 57 s⁻¹, at Au³⁺/MB-SB2 ratios of <0.3 and increased at Au³⁺/MB-SB2 ratios of between 0.3 and 1.5 Au³⁺ per MB-SB2, up to a maximum rate of ∼1,600 s⁻¹, followed by a decline in the rate at molar ratios of >1.5 Au³⁺ per MB-SB2 (Fig. 2A).

FIG 2.

(A) Kinetics of Au binding by MB-SB2 at 4°C. (A) Rate of HAuCl₄ binding to the imidazolone (△) and oxazolone (○) rings of MB-SB2 at 4°C as measured from the absorbance changes at 386 nm and 341 nm, respectively. The rates for Au binding were >2,000 s⁻¹ and were set at 2,000 s⁻¹ in the figure. (B) Emission at 429 nm from SB2-MB following excitation at 341 nm after the addition of 0 (light blue), 0.25 (orange), 0.5 (gray), 0.75 (yellow), or 2.25 (dark blue) HAuCl₄ per MB-SB2. a.u., arbitrary units.

Au³⁺ reduction rates (0.3 ± 0.06 min⁻¹) (Table 1) were much lower than the initial binding rates (>2,000 s⁻¹) (Fig. 2A). The difference may be due to the different rates of binding between the imidazolone and oxazolone groups. Monitoring the fluorescence changes over time at HAuCl₄/MB-SB2 ratios of below 1:1 suggested that final Au coordination required several minutes to complete (Fig. 2B). At gold/MB-SB2 ratios of >1.0, an initial disruption of exciton coupling resulted in an increased fluorescence intensity followed by quenching (Fig. 2B). What is pertinent to this discussion is that Au initially binds primarily if not exclusively to the imidazolone group, followed by binding to the oxazolone group and a final reorientation. The time scales for these changes are in keeping with gold reduction rates.

Oxidation of H₂O coupled to Au³⁺ reduction by MB-SB2.

As four Cl⁻ molecules were generated in reaction mixtures for every HAuCl₄ molecule reduced to Au⁰, chlorine was ruled out as a potential electron donor. To determine if H₂O was the electron donor, H⁺ concentrations were monitored during HAuCl₄ titrations of MB-SB2. Unfortunately, pH changes associated with the addition of HAuCl₄ to unbuffered reaction mixtures made pH changes associated with the binding of AuCl₄⁻ difficult to determine (Fig. 3A). To examine if H₂O could serve as an electron source for Au³⁺ reduction, ^18,18O₂ production was monitored in reaction mixtures containing 97% H₂¹⁸O. No ^18,18O₂ production was observed in reaction mixtures containing either MB-SB2 alone (Fig. 4A) or HAuCl₄ alone (results not shown). However, following HAuCl₄ addition to MB-SB2, ^18,18O₂ was observed, demonstrating the coupling of water oxidation with metal reduction (Fig. 4B and Fig. 5).

FIG 3.

(A) pH changes following the addition of HAuCl₄ to aqueous solutions (gray triangles) or an aqueous solution of 40 μM MB-SB2 (yellow circles). (B) pH changes following the addition of CuCl₂ to aqueous solutions (gray triangles) or an aqueous solution of 40 μM MB-SB2 (blue circles). (C) pH changes following the addition of KCl to aqueous solutions (gray triangles) or an aqueous solution of 40 μM MB-SB2 (black circles).

FIG 4.

Mass spectra of the headspace gas of a reaction mixture containing 2 mM MB-SB2 in 97% H₂¹⁸O (A) and following the addition of 20 mM HAuCl₄ (B), 20 mM CuCl₂ (C), 10 mM CuCl₂ (D), 20 mM KCl (E), 20 mM AgF (F), 20 mM FeCl₃ (G), 20 mM HgCl₂ (H), 20 mM NiCl₂ (I), and 20 mM CoCl₂ (J).

FIG 5.

^18,18O₂ concentration in the headspace of a reaction mixture containing 2 mM MB-SB2 plus 20 mM HAuCl₄ (yellow triangles) or 20 mM CuCl₂ (light blue circles) in 97% H₂¹⁸O and following the addition of 7.3 mM catalase.

It should be noted that there is seemingly an electron imbalance, with three electrons being required for Au³⁺ to Au⁰, while four electrons are released for every two molecules of water oxidized. There are two possibilities to resolve this issue (46): (i) the reduction of four atoms of Au³⁺ is coupled to the oxidation of six molecules of water, and (ii) the reduction of one atom of Au³⁺ is coupled to the oxidation of two molecules of water, with the fourth electron being used to reduce dioxygen to superoxide. Assays show that reduced MB-SB2 will reduce dioxygen to superoxide (Table S3). Under the low-pH conditions following HAuCl₄ addition to MB-SB2 (Fig. 3), superoxide would be expected to undergo dismutation reactions, generating H₂O₂ (47). In addition, as observed for MB-OB3b (48), Au–MB-SB2 complexes show superoxide dismutase activity (Table S3). Thus, H₂O₂ should appear in reaction mixtures if the fourth electron was used to reduce dioxygen. The rate of ^18,18O₂ production increased by approximately 18% following the addition of catalase, suggesting the production of H₂O₂ (Fig. 5).

Oxidation of H₂O coupled to Cu²⁺ reduction by MB-SB2.

Although the oxidation of water to dioxygen coupled to Au³⁺ reduction is chemically interesting, it is not likely biologically relevant. To determine if the oxidation of H₂O was specific to Au³⁺ reduction or a more general property of metal ion reduction by MB-SB2, similar experiments in H₂¹⁸O were carried out with CuCl₂ as it is believed that the primary purpose of MB is the collection of copper critical for methanotrophic activity. Previous spectral and thermodynamic studies have shown that MB-SB2 will reduce multiple Cu²⁺ molecules to Cu⁺ in the absence of an external reductant, suggesting that water served as the reductant (42). ^18,18O₂ evolution was observed following the addition of CuCl₂ to an H₂¹⁸O solution of MB-SB2 (Fig. 4C). Furthermore, such evolution followed a trend similar to that for HAuCl₄, and a substantial pH drop was observed (Fig. 3B to Fig. 5). Perhaps of greater environmental relevance is the finding of substantial (>100 μM) evolution of dioxygen from water oxidation when MB-SB2 bound and reduced copper (Fig. 5).

The ratio of AuCl₄⁻ and Cu²⁺ to MB-SB2 in the experiments described above, as well as other metals showing ^18,18O₂ production (Fig. 4) described below, was 10:1. To determine the number of electrons needed to be extracted from MB-SB2 before water oxidation occurs, reaction mixtures containing 0.5, 1, 2, 3, 4, and 5 Cu²⁺ per MB-SB2 in 97% H₂¹⁸O were examined. No ^18,18O₂ was observed in samples containing 0.5, 1, 2, 3, or 4 Cu²⁺ per MB-SB2 (results not shown). ^18,18O₂ was observed in samples containing 5 Cu²⁺ per MB-SB2, indicating that for the initial water oxidation to occur, five electrons must be extracted from MB-SB2 (Fig. 4D).

K⁺ was also examined as a metal ion bound by MB-SB2 (Fig. S6A) but not reduced, as no evidence of the formation of metallic K⁰ was observed (49). No ^18,18O₂ was observed following the addition of KCl (Fig. 4E), and comparatively minor changes in pH (Fig. 3C) were observed, demonstrating that water oxidation by MB-SB2 after binding a metal ion is contingent upon that metal being reduced. MB-SB2 does not bind Mo²⁺ (Fig. S6B) and was used as a negative control. As expected, no ^18,18O₂ was observed in reaction mixtures containing NaMoO₄ and MB-SB2 (results not shown).

Oxidation of H₂O coupled to Ag⁺, Hg²⁺, Fe³, Ni²⁺, and Co²⁺ reduction by MB-SB2.

As described above, group A metal ions bound by MB-OB3b are reduced following binding. Ag⁺ and Hg²⁺ are group A metals; MB-SB2 bound both metals via an N₂S₂ coordination (Fig. S6C and D), and ^18,18O₂ was observed in reaction mixtures containing MB-SB2 and AgF (Fig. 4F) or HgCl₂ (Fig. 4G) at levels similar to those observed with gold and copper.

In contrast to MB-OB3b (38), MB-SB2 binds all metal ions tested via an N₂S₂ coordination (39, 40, 42) (Fig. S6). Also, in contrast to MB-OB3b, MB-SB2 will reduce Fe³⁺ to Fe²⁺ at a rate of 1.02 ± 0.09 min⁻¹ as measured via the ferrozine assay (50, 51) (Fig. 6A). In fact, MB-SB2 will dissolve insoluble Fe³⁺ hydroxides in the light (Fig. S6E, inset) or dark (Fig. 6C). The one-electron ferric iron reduction rate was approximately three times higher than the three-electron gold reduction rate. In reaction mixtures containing MB-SB2 and FeCl₃ (Fig. 4H), ^18,18O₂ was observed at concentrations 1.3- ± 0.1-fold higher than those observed with Au³⁺ and Cu²⁺. ^18,18O₂ was also observed in reaction mixtures containing NiCl₂ (Fig. 4I and Fig. S6F) or CoCl₂ (Fig. 4H and Fig. S6G) and MB-SB2, although the concentration of ^18,18O₂ was consistently low with CoCl₂.

FIG 6.

(A and B) Iron reductase activities of MB-SB2 (A) and MB-OB3b (B). The absorption change at 562 nm of reaction mixtures containing 1 mM ferrozine plus 10 mM FeCl₃, 1 mM ferrozine plus 23.4 μM MB-SB2, 1 mM ferrozine plus 10 mM FeCl₃, and either 5.8, 11.6, 17.4, or 23.4 μM MB-SB2 (A) or MB-OB3b (B) was measured. (C) Aqueous 4 M FeCl₃ solution and 4 M FeCl₃ solution plus 20 mM MB-SB2 4 h after the addition of MB-SB2.

Oxidation of H₂O coupled to Au³⁺, Cu²⁺, and Ag⁺ reduction by MB-OB3b.

To determine if water oxidation coupled to metal ion reduction was specific to MB-SB2, a group II MB, or a more general property of MBs, water oxidation was examined in the group I MB from M. trichosporium OB3b (MB-OB3b) (31). Previous studies have shown that MB-OB3b binds and reduces Au³⁺, Cu²⁺, and Ag⁺ to Au⁰, Cu⁺, and Ag⁰, respectively, and binds but does not reduce Fe³⁺ (Fig. 6B) (38, 43). Thus, ^18,18O₂ production was monitored in reaction mixtures containing HAuCl₄, CuCl₂, AgF, and FeCl₃ with or without MB-OB3b prepared in 97% H₂¹⁸O. Again, no ^18,18O₂ production was observed in reaction mixtures containing MB-OB3b either alone (Fig. 7A) or with HAuCl₄, CuCl₂, or AgF alone (results not shown). However, following HAuCl₄ (Fig. 7B), CuCl₂ (Fig. 7C), or AgF (Fig. 7D) addition to reaction mixtures containing MB-OB3b, ^18,18O₂ was observed, although the concentrations of ^18,18O₂ were <25% of the ^18,18O₂ concentrations produced in similar reactions with MB-SB2. No ^18,18O₂ was observed following FeCl₃ addition to a reaction mixture containing MB-OB3b.

FIG 7.

Mass spectra of the headspace gas of a reaction mixture containing 2 mM MB-OB3b in 97% H₂¹⁸O (A) and following the addition of 20 mM HAuCl₄ (B), 20 mM CuCl₂ (C), and 20 mM AgF (D).

Methane oxidation coupled to O₂ generated from Cu²⁺ reduction by MB-OB3b.

To determine if dioxygen generated during metal ion reduction could support methane oxidation by M. trichosporium OB3b, incubations with ¹³CH₄ in the presence and absence of MB-OB3b and Cu²⁺ were performed under anoxic conditions in an anaerobic glove box. In cell suspensions with no additional amendments of either copper or MB-OB3b, 0.72 ± 0.17 μmol ¹³CO₂ was observed after 3 days (assumed to be driven by the presence of residual dioxygen in the reaction mixtures [Fig. 8]). In cell suspensions amended with 25 μM Cu²⁺, 0.97 ± 0.03 μmol ¹³CO₂ was observed (a 34% increase, not significantly different from the amount of ¹³CO₂ measured with no amendment [P = 0.06]). If 5 μM MB-OB3b was added instead, 1.47 ± 0.08 μmol ¹³CO₂ was measured (an increase of ∼104%, significantly higher than with no amendment [P = 2.2 × 10⁻³], presumably due to MB-OB3b binding and reducing metals that are part of the growth medium). If both 25 μM Cu and 5 μM MB-OB3b were added, 2.5 ± 0.37 μmol ¹³CO₂ was observed, an increase of ∼250% from that with no amendment (again, significantly different [P = 1.5 × 10⁻³]), indicating that metal ion reduction by MB can support methane oxidation under anoxic conditions (Fig. 8).

FIG 8.

¹³CO₂ production by the M. trichosporium OB3b wild type (WT), the WT plus 25 μM CuCl₂, the WT plus 5 μM MB-OB3b, and the WT plus 25 μM CuCl₂ and 5 μM MB-OB3b incubated in an anaerobic glove box for 3 days.

DISCUSSION

Metal ion binding by MBs has focused on MB-OB3b, a group I MB (34, 38, 43). MB-OB3b bound group A metal ions such as Cu²⁺, Au³⁺, Ag⁺, and Hg²⁺ via an N₂S₂ ligand set. Other metal ions such as Fe³⁺, Co²⁺, Cd²⁺, Mn²⁺, Ni²⁺, and Zn²⁺ showed an N₁S₁ coordination and were placed in group B. Of the metal ions examined, group A metals were reduced following binding, whereas group B metal ions were not. In this and previous reports (39, 40, 42, 44), MB-SB2, a group II MB, coordinates all metals bound via N₂S₂ coordination and reduced metals previously placed in groups A and B. With the exception of K⁺, metal ions bound via an N₂S₂ coordination are reduced, and here, we show that H₂O can serve as an electron donor driving metal ion reduction.

The finding that MB, after binding specific metal ions, can split water to form dioxygen and that this reaction can drive methane oxidation under anoxic conditions suggests that this may be another strategy whereby aerobic methanotrophy can occur in an anoxic environment. That is, it has been shown that methane oxidation via aerobic methanotrophy occurs in anoxic zones of shallow lakes (i.e., at a depth of ∼10 m), with such activity being driven by oxygenic photosynthesis as sunlight could penetrate to this depth (52, 53). In these studies, it was found that methane oxidation rates increased in the light versus the dark, and such activity was abolished when a selective inhibitor of photosynthesis was added. Thus, it appears that methanotrophs can form very effective relationships with oxygenic photosynthetic microbes to scavenge trace amounts of dioxygen and, by so doing, enhance methane removal from these environments.

More germane to the findings here, however, is the discovery that aerobic methanotrophs were also active in deep lake water (∼160 m) where oxygenic photosynthesis is highly unlikely as sunlight cannot penetrate to this depth (54). Such activity, however, could be stimulated by the addition of dioxygen and oxidized metals. Here, it was concluded that methanotrophs may survive anoxic environments by utilizing alternative electron acceptors. Others have shown that aerobic methanotrophs of the Methylobacter genus can be stimulated in anoxic lake waters through the addition of either nitrate or sulfate (55). Indeed, it has been shown that some aerobic methanotrophs can respire nitrate (24) or ferric iron (25). Such a strategy could conserve trace amounts of dioxygen to enable methane oxidation by the MMOs. Alternatively, it has been shown that some methanotrophs will couple methane oxidation to fermentation to putatively conserve dioxygen (56), and such a strategy has been speculated to be responsible for methanotrophic activity in dioxygen-limited lakes (55). Finally, it has been speculated that alternatively or in conjunction, methanotrophs may form syntrophic partnerships with other microbes to facilitate methane oxidation (55) when dioxygen is limiting.

It should be noted, however, that in studies of methane oxidation in anoxic lake water samples, great care was taken to exclude any oxygen intrusion, and any trace amounts of oxygen present were quite small and could not explain the extent of methane oxidation observed. How these microbes then are able to oxidize methane in the absence of dioxygen is still unclear. That is, for the identified methanotrophs to oxidize methane, dioxygen is required for either form of MMO regardless of whether the alternative terminal electron acceptors can be used or an effective microbial partnership(s) can be formed. Thus, either unknown sources of dioxygen exist in these environments or these microbes possess some novel, as-yet-undescribed mechanism of anaerobic methane oxidation, i.e., novel forms of MMO that can utilize oxidized sulfur and nitrogen species in place of dioxygen.

Here, we present an alternative explanation for the presence and activity of aerobic methanotrophs in anoxic environments, particularly alphaproteobacterial methanotrophs. That is, genes for MB biosynthesis have been found only in the genomes of various Methylosinus and Methylocystis species of the Alphaproteobacteria (31). It has been repeatedly shown that these genera prefer high-methane/low-oxygen conditions found at the oxic-anoxic interface in situ (57, 58). Furthermore, they are the predominant methanotrophic genera present in completely anoxic zones of rice paddy soils (59). Thus, it is tempting to speculate that the ability to produce MB enables methanotrophs to colonize methane-rich environments by self-producing dioxygen to ensure that methane oxidation can continue even when ambient concentrations of dioxygen are quite low. Such a strategy is particularly important for methanotrophs that colonize the oxic-anoxic interface in soils because not only are these locations dark (thus excluding the possibility of methanotrophy/phototrophy synergy), this interface also shifts quickly in response to episodic precipitation and drying periods. As such, methanotrophs that colonize this interface must be prepared to tolerate periodic and possibly quite extended anoxic conditions. The ability to produce dioxygen from water would thus enable these microbes to continue to oxidize methane under anoxia, thereby generating ATP, as well as providing intermediates required for carbon assimilation (i.e., formaldehyde). Doing so would enable them to survive extended periods in the absence of oxygen, if not allowing for some continued growth under anoxic conditions.

It should be noted, however, that in the above-mentioned lake studies concluding that aerobic methanotrophy occurs in anoxic environments, gammaproteobacterial methanotrophs appeared to be predominantly responsible for methane oxidation, and to date, no representatives of this group have been shown to have the genes required for MB biosynthesis, although it is clear that at least some can and do secrete a copper-binding compound (60). It may be that these methanotrophs utilize dioxygen created by others via MB production (i.e., some sort of collaboration between gamma- and alphaproteobacteria as concluded between methanotrophs and oxygenic photosynthetic microbes) and/or can generate dioxygen via some unknown mechanism.

Finally, prior to the discovery of dioxygen production via the splitting of water by metal-MB complexes reported here, dioxygen production by biological systems has been observed in only four known pathways: oxygenic photosynthesis (61, 62), detoxification of oxygen radicals (63, 64), (per)chlorate respiration (65), and nitric oxide dismutation by “Candidatus Methylomirabilis oxyferans” of the NC10 phylum (66). The latter two mechanisms may provide some explanation for the significance of MB-mediated water oxidation. That is, it has been shown that dioxygen evolution from (per)chlorate respiration occurs when the intermediate chlorite is dismutated to chloride and dioxygen, and it is speculated that the dioxygen is then used for an antibiotic-producing monooxygenase in Haloferax volcanii (67). Furthermore, “Ca. Methylomirabilis oxyferans” is a methanotroph but respires nitrite rather than dioxygen. Interestingly, dioxygen is critical for its growth as this microbe utilizes the membrane-associated methane monooxygenase for methane oxidation to methanol (66). Stable-isotope studies showed that “Ca. Methylomirabilis oxyferans” dismutates nitric oxide to dinitrogen and dioxygen, the latter of which is used for methane oxidation (the mechanism[s] by which this occurs, however, is still unknown). It may be that MB-expressing aerobic methanotrophs perform a similar feat to ensure that there is adequate dioxygen for continued MMO activity under hypoxic/anoxic conditions.

In conclusion, the discovery of water oxidation by specific metal-methanobactin complexes not only is unusual but also implies a strategy whereby aerobic methanotrophs can survive, if not thrive, under anoxic conditions. As such, MB-driven dioxygen generation may be an important but hitherto unrecognized process whereby methane emissions are regulated.

MATERIALS AND METHODS

Materials.

Anhydrous CuCl₂ (Acros Organics, Geel, Belgium), HAuCl₄ (Acros Organics, Geel, Belgium), HgCl₂ (Acros Organics, Geel, Belgium), AgF (Acros Organics, Geel, Belgium), FeCl₃ (Acros Organics, Geel, Belgium), NiCl₂ (Acros Organics, Geel, Belgium), CoCl₂ (Acros Organics, Geel, Belgium), NaMoO₄ (Sigma-Aldrich), and KCl (Sigma-Aldrich) were stored in a desiccator under Ar₂. H₂¹⁸O was obtained from Cambridge Isotope Laboratories, Inc. (Andover, MA, USA), and ^18,18O₂ was obtained from Sigma-Aldrich. Ar₂, ^16,16O₂, and chemically pure (CP)-grade CH₄ were obtained from Airgas USA LLC. High-performance liquid chromatography (HPLC)-grade acetonitrile, methanol, and other reagents/chemicals were purchased from Fisher Scientific and used without additional purification. Dianion HP-20 was purchased from Sigma-Aldrich LLC.

Organisms, culture conditions, and isolation of methanobactin.

Methylocystis strain SB2 and M. trichosporium OB3b were cultured in nitrate mineral salts (NMS) medium (68) amended with either 0.2 or 1.0 μM CuSO₄ to optimize the production of their methanobactin (MB-SB2). MB-SB2 was purified from the spent medium as previously described (69), with the following exception. The freeze-dried sample from the Dianion HP20 column was resuspended in deionized H₂O and loaded onto a 250-mm by 20-mm Targa C₁₈ column (Higgins Analytical, Inc., Mountain View, CA, USA) on an Azura HPLC system (Knauer, Berlin, Germany). MB-SB2 eluted in the 12 to 25% methanol fraction in a methanol-H₂O gradient. The purified methanobactin was then freeze-dried as described above.

X-ray photoelectric spectroscopy.

X-ray photoelectric spectroscopy (XPS) was performed as previously described (38, 43), with the following modifications. Samples containing HAuCl₄ and HAuCl₄ plus MB-SB2 were dried onto highly oriented pyrolytic graphite by freeze-drying. The 1-cm² graphite substrates were immersed in a solution containing either HAuCl₄ or HAuCl₄ plus MB-SB2, frozen in liquid nitrogen, and lyophilized overnight. The graphite was then mounted onto an XPS puck and analyzed. Other drying methods were employed, such as drying in air under a stream of He gas with a drying time of 30 min or filtering through a porous alumina filter followed by a 2-min drying time. However, samples produced by these methods showed additional reduction.

As previously observed (38), XPS analysis of Au was complicated by X-ray-induced reduction during the measurement process. Au 4f peak areas were therefore measured as a function of X-ray exposure, the peak areas for a given X-ray dose were determined using the CASA XPS fitting program, and the areas were plotted as a function of time. An exponential fit to the data using the Igor Pro fitting program allowed the determination of the unirradiated sample’s Au³⁺ and Au⁰ peak areas.

XPS measurements were carried out on a custom-designed system that incorporated a Specs hemispherical analyzer (Specs Scientific Instruments, Sarasota, FL, USA), an Al X-ray source, and a load lock to allow rapid sample exchanges.

Kinetics of Au³⁺ binding.

The rates of Au³⁺ binding to MB-SB2 were determined by measuring the absorption changes at 338 nm and 387 nm using a four-syringe Biologic SFM/4000/S stopped-flow reactor coupled to a MOS-500 spectrophotometer (Biologic Science Instrument SA, Claix, France) at 4°C as previously described (39). In contrast to the absorbance maxima using a Cary 50 spectrometer, the absorbance maximum for the oxazolone was 338 nm and that for the imidazolone ring was 387 nm with this system. Stock solutions of HAuCl₄ were prepared in >18 MΩ · cm H₂O. The stock solutions for MB-SB2 were prepared by dissolving freeze-dried MB-SB2 in >18 MΩ · cm H₂O. Final concentrations of the stock solutions of MB-SB2 were determined after filtration by UV-visible absorption spectroscopy as previously described (39). The path length for the cuvette used in the Biologic SFM/4000/S stopped-flow reactor was 1.5 mm, and the dead time of the system was 1.4 ms. The system was cooled and maintained at 4°C. Reaction mixtures contained 400 μM MB-SB2 and either 40, 100, 200, 240, 280, 320, 360, 400, 600, 700, or 800 μM HAuCl₄. Rates obtained for each concentration were averages from either 5 or 7 traces. The rates were determined by fitting the traces to the exponential function in Biokine operational software (Biologic Science Instrument SA). Binding rates were calculated in moles of Au bound per second per mole of MB-SB2 and are reported as the binding number (per second).

Fluorescence changes over time were monitored at 429 nm on a Cary Eclipse instrument (Agilent Technologies, Inc., Santa Clara, CA, USA) following excitation at 341 nm.

Water oxidation.

Saturated solutions of anhydrous CuCl₂, HAuCl₄, HgCl₂, AgF, FeCl₃, NiCl₂, CoCl₂, NaMoO₄, and KCl were prepared in a Coy anaerobic chamber (atmosphere of 95% Ar and 5% H₂) (Coy Laboratory Products, Ann Arbor, MI, USA). The oxidation of 2H₂O to O₂ plus 4H⁺ in reaction mixtures containing a metal ion and either MB-SB2 or MB-OB3b was determined by monitoring the production of ^18,18O₂, H⁺, and, in the case of HAuCl₄, Cl⁻. In oxygen evolution experiments, freeze-dried MB-SB2, MB-OB3b, and catalase as well as an anhydrous metal stock solution were prepared in 97% H₂¹⁸O (Sigma-Aldrich, St. Louis, MO, USA) in 0.8-ml brown airtight vials (DWK Life Sciences, Millville, NJ, USA). Reaction mixtures contained 0 or 2 mM MB-SB2 or MB-OB3b and 0 to 20 mM metal ion in a final volume of 100 μl H₂¹⁸O. Reaction mixtures were prepared in 2-ml brown serum vials sealed with Teflon-lined silicone septa. Initial experiments were performed with aluminum foil-wrapped vials, but that practice was discontinued once it was clear that identical results were produced regardless of whether the vials were wrapped or not. The generation of ^18,18O₂ from H₂¹⁸O was monitored by the direct injection (1 μl or 2 μl) of headspace.

Gas samples were manually injected into an Agilent (Santa Clara, CA, USA) 7890B gas chromatography (GC) system with a 7250 accurate-mass quadrupole time of flight (Q-TOF) GC-mass spectrometry (MS) system and a DB5-ms column. Except for the ^18,18O₂ injections for standard curves, all injection volumes were 1 μl using gastight Hamilton syringes. Standard curves were generated with 1-μl, 1.5-μl, and 2-μl injections of 97% ^18,18O₂ (Sigma-Aldrich, St. Louis, MO, USA). The headspace in the vials was sampled before and after the addition of the metals, as was the outside air in the mass spectroscopy instrument, as controls. After the standards and controls were injected, the samples were mixed, and headspace samples were immediately collected, with subsequent samples being taken every 30 to 60 s. After several minutes, collection slowed to 1 sample every 2 to 3 min. The quantization of generated ^18,18O₂ came from an extracted-ion chromatogram set to 35.9978 Da. A small shift in the MS location of the ^18,18O₂ was observed on some dates. If a drift in the MS location of ^18,18O₂ was observed, the identity of the peak was verified with the 97% ^18,18O₂ standard.

Oxidase, superoxide dismutase, hydrogen peroxide reductase, and iron reductase activities and pH measurements.

Oxidase, superoxide dismutase, and hydrogen peroxide reductase activities were determined as previously described by Choi et al. (48). A ferrozine assay was used to determine iron reductase activity (50, 51).

pH changes during metal titrations were monitored on either a PHM 220 pH meter with a pH2005-7 combined pH electrode (Radiometer Analytical, Villeurbanne, France) or an Oakton Ion 700 pH meter (Cole-Parmer, Vernon Hills, IL, USA).

Free chloride produced from the binding and deduction of HAuCl₄ to Au⁰ was measured via argentometric titration (70). HAuCl₄–MB-SB2 solutions were prepared at a molar ratio of 9:1 and incubated for at least 72 h. Following the incubation period, the solution was titrated with a standardized AgNO₃ solution, delivered with a ramé-hart 2.0-ml microsyringe. The titration processes were monitored with a custom-made Ag wire working electrode and an Ag/AgCl reference electrode.

Methane oxidation coupled to O₂ generated from Cu²⁺ reduction by MB-OB3b. (i) Sample preparation.

M. trichosporium OB3b was grown on NMS medium (68) at 30°C in a 250-ml sidearm flask sealed with rubber stoppers. Cultures were shaken at 200 rpm under a methane-to-air ratio of 1:2 until the mid-exponential phase (optical density at 600 nm [OD₆₀₀] of ∼0.3) was reached. Two milliliters of the cell culture was then transferred to 8.5-ml serum vials containing a Teflon-coated magnetic stir bar. Four separate conditions were used: (i) M. trichosporium OB3b with no amendments, (ii) M. trichosporium OB3b plus 25 μM copper (5 μl added from a filter-sterilized [0.22-μm] 10 mM stock solution of CuCl₂), (iii) M. trichosporium OB3b plus 5 μM MB-OB3b (10 μl added from a filter-sterilized [0.22-μm] 1 mM stock solution of MB-OB3b), and (iv) M. trichosporium OB3b plus 25 μM copper and 5 μM MB-OB3b. Biological triplicate samples were prepared for all conditions. The vials were then crimp sealed and degassed using prepurified-grade filter-sterilized (0.22-μm) nitrogen gas (N₂ [99.998%]) for 20 min at a flow rate of 3.42 ml/s using 22- and 25-gauge needles for N₂ gas flow in and out. After degassing, the needles were removed, and the samples were immediately placed in an anaerobic chamber filled with an H₂-N₂ gas mixture (1:9 mixing ratio). Once in the anaerobic chamber, 1 ml of ¹³C-labeled methane (¹³CH₄ [99%]) (Sigma-Aldrich, St. Louis, MO) was added using a 10-ml gastight syringe (Hamilton Company, Reno, NV). Vacuum grease was then spread on the top of the sealed septa. The vials were finally covered with aluminum foil and incubated inverted (septum side down) on a magnetic stir plate in the anaerobic chamber for 3 days at 25°C.

(ii) Gas chromatography-mass spectrometry analysis.

GC-MS analyses were performed using an Agilent 7890B gas chromatograph system coupled with an Agilent 5977B single-quadrupole mass spectrometer (Agilent Technologies, Santa Clara, CA, USA). A Carboxen-1010 Plot capillary column (30 m by 0.32 mm) was used for separation (Supelco, Bellefonte, PA). Ten microliters of the headspace gas of each sample was injected manually using a 25-μl gastight syringe (Hamilton Company, Reno, NV). GC system conditions were as follows: He as the carrier gas at a flow rate of 10 ml/min, split injection with a split ratio of 5:1, an inlet temperature of 170°C, and an oven temperature maintained at 145°C throughout the analysis. The mass spectrometry ion source and quadrupole temperatures were 250°C and 200°C, respectively. Under these conditions, ¹³CH₄ and ¹³CO₂ were detected at 2.16 min and 2.86 min, respectively. Data were acquired in selected ion monitoring (SIM) mode, monitoring m/z 17 for ¹³CH₄ and m/z 45 for ¹³CO₂.