Environmental Transcription of mmoX by Methane-Oxidizing Proteobacteria in a Subarctic Palsa Peatland

Abstract

Methane-oxidizing bacteria (MOB) that possess the soluble form of methane monooxygenase (sMMO) are present in various environments, but unlike the prevalent particulate methane monooxygenase (pMMO), the in situ activity of sMMO has not been documented. Here we report on the environmental transcription of a gene (mmoX) for this enzyme, which was attributed mainly to MOB lacking a pMMO. Our study indicates that the sMMO is an active enzyme in acidic peat ecosystems, but its importance for the mitigation of methane releases remains unknown.

TEXT

Methane-oxidizing (methanotrophic) prokaryotes hold an important function in ecosystems around the globe because they mitigate the release of the greenhouse gas methane (CH₄). Most methanotrophs belong to the Proteobacteria and Verrucomicrobia (1). Additionally, “Candidatus Methylomirabilis oxyfera” (2) and some methanogenic archaea (3) can oxidize CH₄ in the absence of oxygen, using nitrate and sulfate as electron acceptors, respectively. The key enzyme, methane monooxygenase (MMO), which oxidizes methane to methanol at the expense of NAD(P)H, exists in two forms. A particulate, membrane-bound enzyme (pMMO) occurs in nearly all methane-oxidizing bacteria (MOB) with the exception of Methyloferula and Methylocella. These latter two genera feature only a soluble, cytoplasmic enzyme (sMMO). While the pMMO has a narrow substrate range, oxidizing only C₁ to C₄ alkanes and alkenes, the sMMO can also utilize C₅ to C₉ alkanes, alkenes, alicyclic and aromatic compounds, and the chlorinated compounds trichloroethylene (TCE) and chloroform (4–6), exposing it to interest for bioremediation and biotechnology applications. The well-conserved genes pmoA and mmoX encode subunits of the pMMO and sMMO, respectively, and are used as functional marker genes for both enzymes (7). To date, environmental transcripts of mmoX have failed to be detected, even in acidic peatlands (8, 9) where pMMO-lacking methanotrophs thrive (10–12). This has led to the suggestion that the pMMO may be involved in methane oxidation here (13), leaving the ecological relevance of the sMMO unclear.

Various MOB also possess genes encoding a nitrogenase and can carry out N₂ fixation (14). The nifH gene is highly conserved, widely used as a functional marker gene for N₂ fixation, and suitable for phylogenetic analysis. Among the four phylogenetic clusters of nifH sequences, MOB-related nifH sequences fall into the “conventional” Mo-containing nifH cluster I (15). Despite the widespread ability among MOB species to fix nitrogen, little attention has been devoted to N₂ fixation by MOB in nitrogen-limited, remote ecosystems.

In this study, we investigated the presence of functional transcripts for the genes pmoA, mmoX, and nifH among MOB along the succession of permafrost in a palsa peatland. Palsa peatlands occur at the marginal zone of permafrost distribution and were widespread throughout Scandinavia until a very few decades ago (16). These usually pristine ecosystems offer heterogeneous substrate conditions for microbial activity and display frozen peat mounds (palsas), thermokarst ponds, hollows, and hummock-like structures in close proximity. At present, palsa formation and collapse are no longer in their natural balance since the permafrost is thawing faster than new palsas can build up. In some areas of northern Scandinavia, climate changes had already 10 years ago caused a decrease in the palsa area by two-thirds (17), and palsa regression continues to be significant.

Our study site was located in northern Norway (universal Mercator time: 69.694 N; 29.383 E) at the transition from the Arctic to the sub-Arctic using the 10° July isotherm as the border. Annual average temperatures and precipitation from 1965 to 2011 were −0.6°C and 435 mm, respectively (Norwegian Meteorological Institute, Stations Veines/Neiden and Kirkenes Airport). In this period, a positive trend of both annual mean temperature and precipitation was observed (see Fig. S1 in the supplemental material). Three different successional palsa stages were selected as sites for sampling and analysis. They covered a currently degrading palsa (DP), a thermokarst pond (TP) adjacent to the DP, and a hollow, which represents an old successional stage of a previously collapsed palsa (CP). At the palsas (elevated sites), the vegetation was dominated by Ledum palustre, Empetrum sp., Pleurozium sp., and Rubus chamaemorus, while the mire sites were dominated by Eriophorum vaginatum, Andromeda polifolia, Carex rotundata, Carex cannescens, Carex lapponica, Sphagnum riparium (TP), and Sphagnum lindbergii (CP). We sampled duplicates of pore water for the analysis of vertical concentration profiles of methane, ammonium, and nitrate and triplicates of soil cores/blocks per site for the determination of carbon and nitrogen content and for molecular analysis. In addition, determinations of plot-scale methane emissions were conducted in triplicate per site. The procedures for sampling and analyzing pore water and for methane emission measurements are described in detail elsewhere (18). Ammonium concentrations were determined by an analytical laboratory (TosLab, Tromsø, Norway). For molecular analysis, the top layer of fresh plant material was removed and the blocks were sectioned into an upper 10-cm layer and a lower ∼15- to 25-cm layer. Subsamples of those sections were pooled, distributed to sterile 50-ml tubes, and stored in a liquid-nitrogen-saturated dry shipper on site. The environmental data of the sampling sites are presented in Table 1. Briefly, the pH varied between 4.2 and 4.6 and methane emissions and soil methane concentrations were significant in the thermokarst pond, lower in the collapsed palsa site, and negligible from the palsa itself. Nitrogen was grossly limited in particular in the top 10-cm layer of the wet sites dominated by Sphagnum. Here, the C/N ratios varied between 65 and 97 (compared to 42 in the palsa), and pore water ammonium and nitrate concentrations were below the detection limits of 0.56 μM and 2.4 μM, respectively. Below a depth of 10 cm, the C/N ratios varied between 30 and 62. Unlike nitrate, ammonium could be detected here but did not exceed 2.6 μM.

Table 1.

Environmental data from the DP, TP, and CP sampling sites obtained in July 2010

Sampling site	Mean pH ± SD	Mean CH4 concn ± SD (μM)	Mean CH4 emission ± SD (mg m−2 day−1)
DP	4.2 ± 0.11 (n = 5)a	Pore gas, 0.06 ± 0.08 at depth 0–10 cm (n = 5)	BDb
		Pore gas, BD at depth 20 to 25 cm
TP	4.2 ± 0.3 (n = 4)	Pore water, 9.15 ± 7.76 at depth 0–10 cm (n = 4)	623 ± 419 (n = 3)
		Pore water, 403.51 ± 58.01 at depth 20–35 cm (n = 4)
CP	4.6 ± 0.6 (n = 4)	Pore water, 0.04 ± 0.04 at depth 0–10 cm (n = 4)	31 ± 2 (n = 3)
		Pore water, 96.15 ± 106.83 at depth 20–35 cm (n = 4)

^aMeasured in July 2011.

^bBD, below detection.

The samples preserved for RNA analysis were ground in liquid N₂ to a fine powder. Subsequent extraction of total nucleic acids was carried out in duplicate. Approximately 0.3 g of sample was mixed with 0.5 ml of extraction buffer (5% cetyltrimethylammonium bromide [CTAB], 120 mM K₃PO₄ [pH 8]) and subjected to bead beating for 45 s. After phenol-chloroform extraction, nucleic acids were precipitated by incubation with linear acrylamide and 2 volumes of 30% polyethylene glycol 8000 (PEG-8000) for 120 min at room temperature, collected subsequently by centrifugation for 60 min at 4°C, and resuspended in diethyl pyrocarbonate (DEPC)-treated water. To retrieve only the RNA for downstream cDNA generation and analysis, the solutions were treated with RNase inhibitor while the DNA was digested with the Turbo DNA-free kit (Ambion). Synthesis of cDNA was carried out using 100 to 500 ng RNA as the template, random hexamers (Invitrogen) at a final concentration of 500 nM, and SuperScript III reverse transcriptase (Invitrogen). The obtained cDNA was first used as the template for universal bacterial 16S rRNA gene amplification using RNA as the no-template control (NTC) to check for successful cDNA synthesis and complete DNA digestion. Then, the different functional genes were targeted. Amplification of pmoA, mmoX, nifH, pxmA, and 16S rRNA gene fragments, cloning, sequencing and bioinformatics were performed as described in the supplemental material.

We were able to show that pmoA gene products are present in all sites, while environmental transcripts of mmoX and nifH were found only in the two wet sites (TP and CP). This is, to our knowledge, the first study that reports the detection of environmental transcripts of mmoX. Considering the presence of transcripts as an indication for activity, this strongly points toward the environmental relevance of the soluble methane monooxygenase in acidic peat ecosystems. The majority of mmoX transcripts were assigned to the group of Beijerinckiaceae, presumably to relatives of Methylocella (see Fig. 2B) and thus to species that lack a pMMO. Methylocystis was the dominant group, based on the total numbers of both pmoA and mmoX sequences retrieved from DNA, but did not transcribe mmoX. In contrast, the identification of Methylocystis-related pmoA transcripts that belong to a cluster of pmoA sequences originating solely from peat ecosystems (19) (Fig. 1B) indicates that Methylocystis utilize the pMMO rather than the sMMO. This was expected for MOB hosting both forms of MMOs. In addition to mmoX transcripts from the Beijerinckiaceae, transcripts of this gene detected in the collapsed palsa were assigned to an (OTU) distantly related to Methylomonas and thus belong to the group of type I MOB. There are two possible scenarios that could explain this exciting finding. The first is that in contrast to current beliefs, type I MOB that lacks a pMMO does exist. The alternative explanation would be that methanotrophs preferentially transcribe mmoX under certain conditions, although they host both the soluble enzyme and the particulate enzyme. Thus, our findings pose interesting questions regarding the competition between “sMMO-dependent” methanotrophs and MOB possessing a pMMO with regard to enzyme kinetics, in situ substrate preferences, and the general importance of species lacking a pMMO for mitigation of methane emissions. Recalling its broad substrate range and the diverse pool of potential compounds in northern peatlands (20), pMMO lacking MOB could even utilize alternative substrates rather than methane. Similarly to Methylocystis, Methylobacter- and Methylobacter-related sequences (type Ia) were detected in all sites and one OTU was represented also by pmoA transcripts. Furthermore, active species were identified among Methylomonas and Methylocapsa in the thermokarst pond and in the collapsed palsa, respectively. In the palsa site (DP), cDNA synthesis and detection of mmoX were not successful. Also, the amplification of pmoA and 16S rRNA was problematic, with only a nested approach yielding pmoA products. This is indicative for a low abundance of MOB in the palsa site, where methane concentrations and emissions were negligible, and it is consistent with quantitative PCR (qPCR) data (unpublished data).

Fig 2.

(A) Venn diagram comparing OTUs at the MmoX level. Numbers in parentheses refer to cDNA. (B) Neighbor-joining tree of partial mmoX sequences based on deduced amino acid residues retrieved from the three sampling sites, degrading palsa, thermokarst pond, and collapsed palsa (in boldface), compared with public database sequences. The two numbers in parentheses next to the OTU assignment refer to the numbers of DNA and cDNA sequences retrieved, respectively. Solid circles mark nodes that were verified by a maximum likelihood tree. unc, uncultured.

Fig 1.

(A) Venn diagram comparing OTUs at the PmoA level. Numbers in parentheses refer to cDNA. (B) Neighbor-joining tree of partial pmoA sequences based on deduced amino acid residues retrieved from the three sampling sites, degrading palsa, thermokarst pond, and collapsed palsa (in boldface), compared with public database sequences. The two numbers in parentheses next to the OTU assignment refer to the numbers of DNA and cDNA sequences retrieved, respectively. Solid circles mark nodes that were verified by a maximum likelihood tree. unc, uncultured.

Targeting PmoA, altogether 12 OTUs were assigned, with the highest species richness in the latest successional stage of palsa degradation (CP) and the lowest in the palsa itself (Fig. 1A). The largest number of OTUs based on MmoX was also detected in the late successional stage of CP (Fig. 2A). A summary of the number of DNA and cDNA sequences of each site used for phylogenetic and diversity analysis and the respective amount of OTUs is given in Table S2 in the supplemental material. Targeting the 16S rRNA gene using MOB-specific primers, only 8 OTUs were revealed, indicating primer-based failure to detect some MOB. Nevertheless, all dominant groups that were identified based on functional genes were also found by targeting the 16S rRNA gene (see Fig. S2 and S3 in the supplemental material), and rarefaction analysis in general revealed a good coverage of species richness (Fig. 3). Six OTUs were assigned based on the MmoX, of which at least 3 most likely lack a pMMO (pmoA gene), increasing the total number of detected MOB to 15. In order to define the MmoX OTU cutoff on the species level, we used a distance of 4%, which was based on the correlation between MmoX and 16S rRNA gene sequence distances of selected species (Fig. 3B). Plotting of pairwise distances also proved that mmoX is an appropriate phylogenetic marker within MOB. In comparison with mmoX homologues (Fig. 3C), mmoX seems to have evolved within MOB species and is presumably an essential enzyme. Overall, the diversity of palsa MOB is moderate and ranges between the MOB species numbers of rice paddies (21) and Arctic soils (22–24). However, only a very few species were observed to be active, which most likely is a result of the low pH (25–27).

Fig 3.

(A) Rarefaction analysis of deduced PmoA, MmoX, and 16S rRNA gene sequences. (B and C) Correlation of MmoX versus 16S rRNA gene sequence distances of 32 methanotrophic species (B) and of 43 species, including MmoX homologues (C).

In general, the composition of the MOB community of this palsa peatland and the dominance of Methylocystis related sequences are representative of what has been reported for acidic Sphagnum-dominated peat (8–12, 28–32). Most cultivars from acidic Sphagnum peat are known to be capable of N₂ fixation, which was reported for Methylocapsa acidiphila (11), Methylocella tundrae (10), and Methylocella palustris (12). Also, species of the genotypes Methylomonas, Methylocystis (14) and Methylobacter (e.g., see reference 33) are known to carry out N₂ fixation. This set of MOB reflects the palsa community. Thus, it supports the assumption that nitrogen availability influences soil bacterial communities (34) in particular since in our study MOB-related nifH sequences made up ≥10% of all sequences on both the DNA level and the cDNA level (Fig. 4). The presence of MOB-related nifH transcripts suggests a direct compensation for nitrogen deficiency through N₂-fixating MOB. So far, environmental transcripts of nifH related to MOB were reported neither in Sphagnum peat nor in any other pristine and oligotrophic habitat. Our findings point to an important ecosystem function carried out by MOB in both the carbon and the nitrogen cycles of acidic peatlands. The contribution of MOB to N₂ fixation in acidic peatlands could be substantial, considering the pronounced number of MOB-related nifH transcripts detected and should attract more attention.

Fig 4.

Neighbor-joining tree of partial nifH sequences based on deduced amino acid residues retrieved from the studied palsa peatland (in bold) compared with public database sequences. The two numbers in parentheses next to the cluster labeling refer to the numbers of DNA and cDNA sequences retrieved, respectively, and indicate affiliation with existing and new clusters. The detailed view shows clusters 1.11 and 1.7, which primarily consist of methanotroph-retrieved nifH sequences. Closed circles mark nodes that were verified by a maximum likelihood tree.

Nucleotide sequence accession numbers.

Sequences have been submitted to the GenBank database under accession numbers KC202439 to KC202796 and KC261068 to KC261278.