Biological methane oxidation proceeds either through aerobic or anaerobic pathways. The newly discovered bacterium Candidatus ‘Methylomirabilis oxyfera’ challenges this dichotomy. This bacterium performs anaerobic methane oxidation coupled to denitrification, but does so in a peculiar way. Instead of scavenging oxygen from the environment, like the aerobic methanotrophs, or driving methane oxidation by reverse methanogenesis, like the methanogenic archaea in sulfate-reducing systems, it produces its own supply of oxygen by metabolizing nitrite via nitric oxide into oxygen and dinitrogen gas. The intracellularly produced oxygen is then used for the oxidation of methane by the classical aerobic methane oxidation pathway involving methane mono-oxygenase. The present mini-review summarizes the current knowledge about this process and the micro-organism responsible for it.
The process of AMO (anaerobic methane oxidation) coupled to denitrification refers to the following reaction of methane with nitrite (or nitrate) as oxidant :
(ΔG°′=−928 kJ/mol of methane).
This reaction is highly exergonic and yields free energy values comparable with those with oxygen (O2) (−818 kJ/mol of methane) as electron acceptor . However, the high-activation energy required to break the C–H bonds of methane (439 kJ/mol) renders the reaction biochemically relatively difficult . Aerobic methanotrophs overcome this high activation energy by using O2 as a highly reactive co-substrate for the initial attack, in a reaction catalysed by MMO (methane mono-oxygenase) (reviewed in [4,5]). Hence, it was believed that methane was inert under anoxic conditions and that anaerobic oxidation of methane was biologically impossible.
This view gradually changed with the notion of methane being biologically oxidized in sulfate-rich sediments and water columns [6,7]. Proof for sulfate-dependent methane oxidation came from the discovery of natural enrichment cultures composed of consortia of ANME (anaerobic methanogenic archaea) and sulfate-reducing bacteria [8–10]. In these consortia, the ANME organisms, which are currently known to fall into three distinct phylogenetic groups (ANME I–III) , activate methane probably by a reversal of the terminal reaction in methanogenesis. This reaction is catalysed by a homologue of MCR (methyl-coenzyme M reductase), the methane-releasing enzyme of methanogens . Because methane and nitrite (and nitrate) often coexist in the environment and nitrogen oxides are much more powerful oxidants than sulfate, it was speculated that an organism or consortium capable of making a living on methane as an electron donor for denitrification under anoxic conditions could exist in Nature [13,14].
In 2006, the first evidence for AMO coupled to denitrification was reported in an enrichment culture from a Dutch canal sediment . At that time the enrichment consisted of a co-culture of a dominant bacterial phylotype, which made up to 80% of the population and a smaller fraction (up to 10%) of an archaeal species phylogenetically positioned between Methanosaeta and ANME-II. This finding and several other observations led to the hypothesis that a consortium of bacteria and archaea would drive the process through reverse methanogenesis and electron shuttling to the denitrifying partner, analogously to ANME archaea and sulfate-reducing bacteria. Later, however, it was conclusively shown that the process could also be carried out in the absence of archaea . Upon prolonged enrichment (>1 year), the archaeal population gradually declined while the culture was still capable of performing AMO coupled to denitrification. Consistently, the methane oxidation was not inhibited by bromoethane sulfonate, a specific inhibitor of MCR . These findings were intriguing since there was no known alternative mechanism other than reverse methanogenesis that could explain AMO. An activation of the C1-compound methane by the addition to fumarate catalysed by glycyl-radical enzymes, as known for longer-chain alkanes , would be energetically too costly .
In 2010, the genome of the bacterium responsible, named Candidatus ‘Methylomirabilis oxyfera’ (‘M. oxyfera’ hereafter), was published alongside with a mechanistic explanation for AMO coupled to nitrite reduction . It appeared that ‘M. oxyfera’ is capable of producing O2 via a new intra-aerobic pathway, which involves the dismutation of NO (nitric oxide) into O2 and N2 (nitrogen) (see the Central catabolic pathways and energy metabolism section below). With that, some key questions about ‘M. oxyfera’ and AMO coupled to denitrification were unravelled. However, the research is still in its infancy, and many interesting findings are to be expected from this unique organism.
No organisms have so far been isolated in pure culture which could couple AMO to nitrite reduction. The enrichment cultures described so far all contained significant amounts (30–80%) of NC10 phylum bacteria closely related to ‘M. oxyfera’, which probably represent different strains of the same species (>97.5% identity of the 16S rRNA gene), despite the geographical distance between the inoculum sources, The Netherlands [1,20] and Australia . It is unknown whether other species of the NC10 phylum, all found in potentially anaerobic methane-containing habitats , share a similar metabolism, or have a different lifestyle. In two of the cultures [1,21], archaea were also co-enriched, and they too shared approx. 98% identity with the 16S rRNA gene among each other. Although it was shown that nitrite-dependent methane oxidation can proceed without the presence of archaea , the questions remains whether it can, via a different mechanism, also be carried out by this type of archaea, or whether they are methanogens thriving on metabolic intermediates of ‘M. oxyfera’ or other organic compounds.
All the cultures described preferred nitrite over nitrate as an electron acceptor, and it remains to be shown whether nitrate can actually be used by ‘M. oxyfera’ as an electron donor for methane oxidation. Possibly it needs an electron donor other than methane for the reduction of nitrate to nitrite, or ‘M. oxyfera’ might depend on other community members for this step. The specific activity is low when compared with other denitrifying bacteria (3.6–5.6 nmol of NO2−/mg of protein per min [1,20,21]), and so is the growth rate, an estimated 1–2 weeks under laboratory conditions . These conditions (all studies used a defined freshwater mineral medium, temperatures of 25–35°C and a pH of 7.0–8.0) might not be optimal, as suggested by the cessation of apparent growth after reaching a certain cell density [15,20,21]. This indicates limitation by an unknown compound that may be supplied by other community members [e.g. PQQ (pyrroloquinoline quinol), see below]. Having many copper-containing enzymes (see below), copper availability is an important concern, but can be excluded as the limiting factor in the Dutch enrichment cultures (K.F. Ettwig, unpublished work). It is probably for these reasons (the low growth rate and unknown limitation, possibly in combination with inhibition by commonly used materials ) that isolation attempts have so far not been successful. The highly enriched (~80% ‘M. oxyfera’) cultures nevertheless were suitable to test the hypotheses formulated after genome annotation.
Central catabolic pathways and energy metabolism
The assembly and annotation of the complete genome of ‘M oxyfera’ from the enrichment culture, together with proteomic and transcriptomic analysis allowed the prediction of the central pathways involved in methane and nitrite conversions, which were subsequently validated by stable isotope labelling experiments (Figure 1) .
‘M. oxyfera’ central catabolism and energy metabolism
Known genes involved in alkane activation, such as fumarate-adding glycyl-radical enzymes, were not found  and, consistent with the absence of archaea, MCR was also absent . ‘M. oxyfera’ encoded, transcribed and expressed the full repertoire of genes for the aerobic methane oxidation pathway (Figure 1), including the H4MPT (tetrahydromethanopterin)-dependent C1-transfer module . Furthermore, all genes coding for the H4F (tetrahydrofolate)-dependent reactions were present in the genome, but only some of their gene products could be detected in the proteome. The first step of the methane oxidation pathway, the oxidation of methane to methanol, is catalysed by MMO, which in ‘M. oxyfera’ is represented by a single and phylogenetically divergent membrane-bound form . Methanol is further oxidized to formaldehyde by the periplasmic PQQ-dependent MDH (methanol dehydrogenase). In the genome, three sets of genes encoding MDH paralogues as well as all the proteins necessary for the catalytic function are found, but most of the genes required for PQQ biosynthesis (pqqABCDEFG) are absent (except for pqqE and pqqF). It remains to be established whether this organism has an alternative PQQ biosynthesis pathway, uses a modified form of this cofactor, or relies solely on PQQ excreted by other members of the enrichment culture, like some PQQ-deficient bacteria [22,23], or does not use PQQ altogether.
Perhaps the most astonishing aspect unravelled by the ‘M. oxyfera’ genome was the absence of a complete denitrification pathway (Figure 1) . On the basis of the established knowledge about of the enzymology of denitrification, the reduction of NO2− could not proceed further than intermediate nitrous oxide (N2O). The known genes encoding N2O reductase, the enzyme complex catalysing the terminal nitrogen-producing reaction, could not be identified. Thus the experimental evidence that N2 was the end-product of nitrite reduction [1,15,20] was in conflict with the genomic data. Subsequent stable isotope labelling experiments solved part of this apparent contradiction, and an alternative denitrification pathway was proposed. Following the reduction of NO2− to NO by a periplasmic cd1 nir (cd1 nitrite reductase; nirS), N2 and O2 would be formed by disproportionation of two molecules NO into O2 and N2, resembling the dismutation of chloride (ClO2) to Cl− and O2 in chlorate-reducing bacteria . The intracellularly produced O2 would be partly (3/4) used for methane activation, whereas the remainder (1/4) might serve other purposes. The genome also encodes three putative qnor (quinol-dependent NO reductases). Whether these enzymes function in NO detoxification, a role well known in other bacteria , or play a pivotal role as NO dismutases as suggested, is still unclear .
Being a respiratory organism, ‘M. oxyfera’ should conserve the energy derived from nitrite-dependent methane oxidation by a chemiosmotic mechanism. Inspection of the genome reveals the full potential for such a mechanism (Figure 1). First, a proton-pumping NDH [NAD(P)H dehydrogenase] (complex I) is present, which couples the oxidation of NADH, formed in the methylene-H4MP (and methylene-H4F) dehydrogenase and formate dehydrogenase reactions, to the reduction of quinone with the concomitant export of protons. Re-oxidation of the quinol is performed by the bc1 complex, again coupled to proton translocation. Reduced cytochrome c then serves as the electron donor in nitrite reduction by cd1 nir. The protonmotive force created in this way is utilized to drive ATP synthesis by the F1Fo ATP synthase. As mentioned, residual O2 is left after methane oxidation. In this respect, it is remarkable that the ‘M. oxyfera’ genome encodes four different putative terminal oxidases for oxygen respiration, of which at least one, a quinol-dependent oxidase is being functionally expressed .
Classification and C1 metabolism
‘M. oxyfera’ represents a new taxonomic group among bacterial methanotrophs (Figure 2). Together with the Verrucomicrobia [6,27–29], they are the only non-proteobacterial methanotrophs known so far. Phylogenetic analysis of the 16S ribosomal RNA showed that ‘M. oxyfera’ is a member of the NC10 phylum , which was so far only defined by environmental 16S rRNA gene sequences .
Phylogenetic tree of bacterial methanotrophs based on 16S rRNA gene sequences
The isolation of the first methanotrophs dates back a century ago . Later, the isolation and characterization of over 100 proteobacterial methanotrophic species by Whittenbury et al.  established the basis for the functional taxonomic classification of methanotrophs known today. The key features for their classification are their phylogeny, C1-assimilation pathways and the arrangement of their ICM (intracytoplasmic membranes) (reviewed in [4,5,29]). The latter are cytoplasmic membrane invaginations containing high amounts of the enzyme pMMO. On the basis of these criteria, methanotrophs were classified into two groups. Type I (Gammaproteobacteria) utilizes the RuMP (ribulose monophosphate) pathway and sometimes the CBB (Calvin–Benson–Bassham) cycle and some enzymes of the serine cycle pathway for C1 assimilation and possesses ICM arranged as bundles of vesicular disks. Type II (Alphaproteobacteria) utilizes the serine pathway for carbon assimilation and possesses ICM arranged in paired peripheral layers. The currently known Verrucomicrobia species are exceptional for the absence of ICM structures; there is still no experimental evidence for their C1-assimilation pathway. However, it was suggested to proceed through a combination of the CBB and the serine (or a variant) cycles or autotrophically via the CBB cycle .
Besides the taxonomic divergence, ‘M. oxyfera’ also appears not to fit in the above described categories, particularly with respect to the C1 assimilation. Regarding the ICM geometry, it is still to be established whether it is like the Type I or Type II, or simply does not develop ICM like the Verrucomicrobia .
The analysis of the genome showed that both the RuMP and the serine pathway for C1 assimilation are incomplete. The genome lacks the gene encoding the enzyme 6-hexulosephosphate synthase (hsp), which catalyses the first step of the RuMP pathway . Some of the key serine cycle enzymes are not found either, including malyl-CoA lyase and phosphoserine aminotransferase. Thus the C2 intermediate glyoxylate would be depleted during the cycle and genes encoding an enzyme or enzyme systems that alternatively converts acetyl-CoA into glyoxylate , such as isocitrate lyase and a glyoxylate-regeneration cycle , were also not found. Conversely, genes encoding a type I Rubisco (ribulose-1,5-bisphosphate carboxylase/oxygenase; EC 126.96.36.199), and all other genes of the CBB cycle are found in the genome and proteome . These findings suggest that this organism uses the CBB cycle as the primary C1-assimilatory pathway, whereas the incomplete serine cycle may serve as a means of formaldehyde detoxification and for the synthesis of intermediates. This is in accordance with the only very weak labelling of the dominant bacterial lipids after incubation with [13C]methane . Nevertheless, more research is required to validate these assumptions.
With the current development of new analytical techniques, there is no doubt that ‘M. oxyfera’ or its relatives will be found in environments other than the eutrophic freshwater sediments described as habitats so far [20,21]. This will then allow the study of the microbial diversity of bacteria performing AMO coupled to denitrification. Primers and probes based on 16S rRNA genes are already available for the detection of ‘M. oxyfera’ , and will soon be followed by other markers. Also, it would be interesting to explore whether ‘M. oxyfera’ or other micro-organisms using a similar oxygen-production pathway might use other substrates, such as longer-chain alkanes or aromatic hydrocarbons.
The publication of the genome provided a platform for many investigations. Now, genes can be cloned and expressed without the dependence on biomass. Heterologous protein expression can be used to raise antibodies against key enzymes. Furthermore, immunoelectron microscopy using antibodies against unique proteins would allow the distinction of the ‘M. oxyfera’ morphotype from other community members. Moreover, such studies will provide detailed ultrastructural and morphological information, as well as the subcellular localization of targeted enzymes.
There is no doubt that the key question in the biochemistry of ‘M. oxyfera’ is the nature of the oxygen-producing enzyme. To solve this enigma, an in silico analysis for the selection of candidate proteins combined with cloning and purification strategies, might prove successful.
Enzymology and Ecology of the Nitrogen Cycle: A Biochemical Society Focused Meeting held at University of Birmingham, U.K., 15–17 September 2010. Organized and Edited by Jeff Cole (University of Birmingham, U.K.), Rosa María Martínez-Espinosa (University of Alicante, Spain), David Richardson (University of East Anglia, Norwich, U.K.) and Nick Watmough (University of East Anglia, Norwich, U.K.).
anaerobic methane oxidation
anaerobic methanotrophic archaea
cd1 nitrite reductase
methyl-coenzyme M reductase
ribulose monophosphate pathway
We gratefully acknowledge all the co-workers and specially M.K. Butler, N. Kip, A. Pol and H.J.M. Op den Camp for discussions.
Our research was supported by various grants from the Netherlands Organization for Scientific Research (NWO), an NWO Horizon grant [grant number 050-71-058] and the European Research Council [grant number 232937].