Three types of aerobic methanotrophs are recognized. Type I methanotrophs are γ-proteobacteria that have stacked membranes harboring methane monooxygenase (pMMO), the enzyme for primary methane oxidation, and that use the ribulose monophosphate (RuMP) cycle, which converts formaldehyde into multicarbon compounds, for building cell biomass 5. Type II methanotrophs belong to the α-proteobacteria, have rings of pMMO-harboring membranes at the periphery of the cells, and use the serine cycle, an alternative pathway for converting formaldehyde into biomass; these bacteria also often contain a soluble (s) MMO in addition to pMMO 5. The third type, type X methanotrophs, belong to the genus Methylococcus (γ-proteobacteria) and combine features characteristic of the other two types: they have stacked membranes and the RuMP cycle, but they also have elements of the serine cycle and sMMO 5. The type X methanotroph Methylococcus capsulatus has been a favorite model for research because of its robust growth on methane and its relative ease of use as a genetic system 6789. Two almost identical gene clusters have been identified encoding the subunits of pMMO, which are expressed simultaneously and are functionally redundant 78, and another gene cluster encodes the subunits of sMMO 9. Copper has been shown to play an essential role in expression of the pMMO operons, whereas the sMMO operon appears to be expressed only in low-copper conditions 10. The catalytic mechanisms for both pMMO and sMMO 1112 are understood on a sophisticated level, but until recently no whole-genome sequence has been available for M. capsulatus or for any other methanotroph. Two recent studies 12 have used a whole-genome-shotgun sequencing approach to complement the mounting dataset on the biochemistry and regulation of aerobic methane oxidation.

In contrast, understanding of the process of anaerobic methane oxidation is in its infancy. Geochemical evidence points strongly towards a coupling of anaerobic methane oxidation with sulfate reduction 13. Microbes involved in this process have been identified recently as archaea related to Methanosarcinales that fall phylogenetically into two distinct groups, ANME-I and ANME-II; these are normally found in association with sulfate-reducing bacteria 13. There is no clear concept of how methane oxidation is linked to sulfate reduction; Figure 1 shows a possible model. This co-metabolism has to be viewed in the light of the thermodynamic constraints, however; the free energy (ΔG) for anaerobic methane oxidation in situ is estimated at -20 to -40 kJ/mol), the lowest value described that enables microbial growth 1314.

There is agreement on the hypothesis that reverse methanogenesis plays a key role in the methane oxidation process 1314: most enzymes of methanogenesis are easily reversible, and part of the methanogenesis pathway operates in reverse for energy generation in Methanosarcina species growing on such substrates as methanol or methylamine 1516. But the last step of methanogenesis and presumably the first in anaerobic methane oxidation (step 1 in Figure 1), catalyzed by methyl-coenzyme M reductase (MCR), presents a mechanistic challenge given the fact that methane is chemically unreactive. Nevertheless, data have been obtained showing that methanotrophic archaea have homologs of the genes for all three subunits of MCR, suggesting that MCR or a similar enzyme may indeed be responsible for anaerobic methane oxidation 17. Two recent studies 34 describe efforts to establish the roles of mcr homologs and of other genes potentially involved in reverse methanogenesis by directly assessing environmental DNA and protein pools.

In a paper recently published in PLoS Biology, Ward et al. 1 describe the complete genomic sequence of Methylococcus capsulatus (Bath). They annotate the genome in terms of the specific adaptations this organism has evolved in order to succeed at a lifestyle solely dependent on utilization of methane. The genome of M. capsulatus (3.3 megabases, Mb) is much smaller than the genome of a model facultative methylotroph, Methylobacterium extorquens AM1 (7 Mb), a bacterium with a much more versatile lifestyle 18, but is comparable in size to the genome of another obligate methylotroph, Methylobacillus flagellatus (2.9 Mb) 19, suggesting that the degree of specialization in methylotrophy may correlate with genome size. The cause of the obligate methylotrophy of M. capsulatus remains unresolved, however. The tricarboxylic acid (TCA) cycle is the pathway that converts acetyl-CoA to CO₂ and is the major source of reducing equivalents during growth on multicarbon compounds; the long-held hypothesis that M. capsulatus lacks a complete TCA cycle 20 has not been proven true by genome sequencing, as putative genes for all the enzymes of the cycle were identified in the recent study 1. In addition, the organism seems to encode an array of enzymes that could metabolize sugars, so the inability of M. capsulatus to grow on sugars remains enigmatic.

Analysis of the genes encoding enzymes involved in the metabolism of single-carbon compounds in M. capsulatus (Figure 2) has been greatly simplified by the addition of data available from pre-genomic analyses 78921 and from the initial analysis of the genome of M. extorquens 18. As expected, all the genes encoding enzymes of the RuMP pathway have been identified. In accordance with previous observations, most of the genes for the serine cycle were also found, as were the genes for the Calvin-Benson-Bassham (CBB) cycle, the pathway that reduces CO₂ and converts it into biomass (Figure 2f) 520. The potential to operate all three known pathways for the assimilation of single-carbon compounds that are found in various methylotrophs makes this organism unique, but further analysis involving knockout mutations is needed to understand the functions of each of the three pathways.

The first glimpses into the expression patterns of pathways enabling methanotrophy are coming from a proteomic analysis of M. capsulatus by a group that has independently sequenced the M. capsulatus genome to 8X coverage 2. In this work 2, quantitative proteomic analysis was performed in order to compare the response of M. capsulatus to low-copper and high-copper conditions. Kao et al. 2 identified a total of 682 differentially expressed proteins using a cleavable isotope-coded affinity tag (cICAT) technique. The authors 2 demonstrated that, as expected, pMMO is overexpressed in conditions of high copper whereas sMMO is expressed at low copper levels. Equally interesting data from this work concern the expression of proteins other than MMOs, indicating that, indeed, all three assimilatory pathways are simultaneously expressed. The oxidative pathway linked to tetrahydromethanopterin (H₄MPT) is one of the pathways by which formaldehyde can be oxidized to CO₂ (Figure 2b); all the enzymes in this pathway were identified 2, pointing to the importance of this pathway, as suggested previously by enzyme-activity measurements 22. Peptides for the oxidative branch of the RuMP cycle were also identified 2, suggesting that it is operational in M. capsulatus (Figure 2a).

Some of the major serine-cycle enzymes were found to be overexpressed under high-copper conditions 2. It is unlikely, however, that their expression would be directly regulated by copper; it is more likely that they are responding to the higher flux of formaldehyde that occurs during growth under high-copper conditions. It is important to note that the serine cycle cannot operate as a major assimilatory pathway in M. capsulatus unless the two-carbon compound glyoxylate that is depleted during the cycle can be regenerated 20, but no genes have been identified in the genome that potentially encode either of the enzyme systems that can convert acetyl-CoA into glyoxylate: the isocitrate lyase and the glyoxylate-regeneration cycle 23.

Given these considerations, what might the function of the serine cycle (and the interconnected TCA cycle) be in M. capsulatus? We suggest that a possible role for this pathway could be to handle the extra flux of formaldehyde that the organism may encounter under certain growth conditions (Figure 2c). The excess of formate generated in the H₄MPT-linked pathway (Figure 2b) could also be redirected into the serine cycle after reduction to methylene-tetrahydrofolate (methylene-H₄F; Figure 2d). Acetyl-CoA and other intermediates generated in this way could serve as building blocks for cell biomass.

The role of the CBB cycle in M. capsulatus (Figure 2f) is not clear at present. Given that the fixation of CO₂ is a far less efficient mechanism of carbon sequestration than the RuMP or serine cycles, a significant amount of carbon shunted through the CBB cycle would be predicted to decrease growth yield. It is possible, however, that it serves to reduce the local concentration of CO₂ and/or to generate intermediates for biomass production. Once again, further experiments are needed to establish the validity of these hypotheses.

To provide support for the hypothesis that reverse methanogenesis is important in anaerobic methanotrophy, a consortium of researchers focused on identifying the enzyme potentially involved in the initial step of anaerobic methane oxidation; this enzyme is hypothesized to be similar to the bacterial MCR (Figure 1, step 1). A microbial mat in the Black Sea largely consisting of ANME-1-type archaea was chosen as a source of this hypothetical enzyme. As described in Nature in 2003 by Krüger et al. 4, a conspicuous protein consisting of three subunits similar to the α, β, and γ subunits of MCR is abundantly present in this microbial mat (7% of the total extracted protein), suggesting that it has an important role in anaerobic methane oxidation. The protein contains a variant of F₄₃₀, a cofactor used by the classical MCR, but the two cofactors differ in molecular weight as determined by mass spectrometry. The genes encoding this protein were sequenced as a part of an insert detected in an environmental DNA library 4. Alignment of amino-acid sequences translated from these genes with the respective sequences of methanogen MCR subunits showed that residues involved in active-site formation in the β and γ subunits were conserved, but one of the important residues in the active site of the α subunit was substituted. It is interesting to speculate that this modification of the active site and the use of a modified F₄₃₀ cofactor could provide a mechanism for the biochemical activation of methane and could make the first step of reverse methanogenesis thermodynamically and kinetically possible. Further in-depth mechanistic studies of this enzyme will be of great interest.

In a recent paper published in Science, Hallam et al. 3 describe a large environmental sequencing effort which aimed to provide further evidence for the hypothesis of reverse methanogenesis. The group 3 isolated DNA from a 52O-meter-deep sediment of Eel River Basin in California, known for a high abundance of ANME-1 and ANME-II archaea, and used it for both whole-genome shotgun analysis and fosmid 'walking' (fosmids are large-insert plasmids). A total of 111.3 Mb of non-redundant sequence was generated by shotgun sequencing and another 4.6 Mb more were generated by fosmid-end sequencing. Fosmids containing either 16S rRNA genes belonging to ANME-I or ANME-II archaea or homologs of the mcrA gene were analyzed in detail, producing an additional 7.4 Mb of sequence.

The main conclusion from this work 3 is that ANME archaea contain most of the genes involved in methanogenesis, with one exception: mer, the gene encoding methylene-H₄MPT reductase (step 3 of reverse methanogenesis; see Figure 1) 15. On the basis of the apparent lack of mer, the authors propose a model in which parts of the methanogenesis pathway function in two opposite directions: a novel MCR-like enzyme oxidizes methane to methyl-CoM (step 1), and methyl-H₄MPT:CoM methyl-transferase catalyzes a reverse reaction to produce methyl-H₄MPT (step 2), while the rest of the enzymes reduce CO₂ to methylene-H₄MPT (steps 4 to 7 in reverse); that is, contrary to previous models 1314, methane is not oxidized to CO₂ by ANME archaea. This proposed scenario creates some metabolic difficulties, however. Firstly, the model aggravates the thermodynamic constraints mentioned earlier, given that reduction of CO₂ to formyl-methanofuran (step 7) is an energy-consuming reaction (ΔG° = +16 kJ/mol) 15. Secondly, the fate of the methylene-H₄MPT produced in steps 4 to 7 is proposed to involve either the assimilatory serine cycle or formaldehyde oxidation, but the high energy cost of such schemes would suggest they could operate only as minor pathways, not as major assimilatory or detoxification pathways. Thirdly, there is no discussion by Hallam et al. 3 of how net CO₂ would be produced from methane.

Thus, although the schemes presented by Hallam et al. 3 are an attempt to explain how methane metabolism might function in the absence of mer, they highlight the many aspects of this metabolic mode that are still unknown. Two different explanations might be that either mer has simply not been detected because of incomplete sequence data, or that the function of Mer is fulfilled by a novel enzyme (a non-homologous substitution), possibly involving a cofactor different from F₄₂₀, so the reverse-methanogenesis pathway might in fact be complete (as in Figure 1). An example of such a non-homologous substitution is seen in methylotrophic bacteria, in which a version of the 'reverse methanogenesis' pathway has been found to operate where an NAD(P)-linked methylene-H₄MPT dehydrogenase acts in place of unrelated F₄₂₀-linked or H₂-forming enzymes 24.

In conclusion, recent studies involving both organismal and environmental genomics shed new light on the biochemical details of the two processes important for methane balance on Earth - aerobic and anaerobic methane oxidation - and suggest that these processes have more in common than just the substrate, methane, and the final oxidation product, CO₂. Both processes involve common cofactors, such as H₄MPT, common single-carbon intermediates bound to H₄MPT, and common or similar enzymes for core reactions. Although some enzymes involved in reactions that shift single-carbon compounds between different levels of oxidation are evolutionarily related in both processes, the primary methane oxidation enzymes, MMO and the newly identified MCR homolog, must have evolved independently and are fundamentally different.