Identification of the Functionally Active Methanotroph Population in a Peat Soil Microcosm by Stable-Isotope Probing

Abstract

The active population of low-affinity methanotrophs in a peat soil microcosm was characterized by stable-isotope probing. “Heavy” ¹³C-labeled DNA, produced after microbial growth on ¹³CH₄, was separated from naturally abundant ¹²C-DNA by cesium chloride density gradient centrifugation and used as a template for the PCR. Amplification products of 16S rRNA genes and pmoA, mxaF, and mmoX, which encode key enzymes in the CH₄ oxidation pathway, were analyzed. Sequences related to extant type I and type II methanotrophs were identified, indicating that these methanotrophs were active in peat exposed to 8% (vol/vol) CH₄. The ¹³C-DNA libraries also contained clones that were related to β-subclass Proteobacteria, suggesting that novel groups of bacteria may also be involved in CH₄ cycling in this soil.

It is estimated that 400 to 640 Tg of methane (CH₄) is produced annually in anoxic environments (9), much of which is oxidized at the anaerobic-aerobic interface, thereby mitigating the global emissions of this greenhouse gas into the atmosphere. There are two distinct populations of organisms responsible for aerobic CH₄ oxidation. One population exhibits high-affinity CH₄ oxidation kinetics and acts as a sink for atmospheric concentrations of CH₄ in many soils (2, 8, 10, 15, 17, 20, 24). The second population (methanotrophs) can grow on CH₄ as a sole source of carbon and exhibits low-affinity CH₄ oxidation kinetics (2, 17), and extant strains can be divided into type I (γ-Proteobacteria) or type II (α-Proteobacteria) methanotrophs (4, 5, 7, 40).

Methane monooxygenase (MMO) and methanol dehydrogenase (MDH) are key enzymes involved in CH₄ oxidation. There are two forms of MMO, both of which oxidize CH₄ to methanol. A membrane-bound, particulate enzyme (pMMO) (32) has been reported in nearly all methanotrophs (12, 32). A soluble, cytoplasmic enzyme (sMMO) is only found in certain methanotrophs (32). The pMMO genes exhibit high similarity to genes encoding ammonia monooxygenase (23) from ammonia-oxidizing nitrifying bacteria (nitrifiers). Nitrifiers also oxidize CH₄ but probably do not use it as a carbon source (1). MDH oxidizes methanol to formaldehyde, and it is present in all known methylotrophic bacteria within the Proteobacteria (30). Genes encoding MMO and MDH (functional genes) and 16S rRNA genes have been used to assess the diversity of methanotrophs in the environment (11, 14, 18, 24, 31, 33, 44). Phylogenetic analysis of 16S rRNA, sMMO, pMMO, and MDH gene sequences obtained from these environments has suggested that a greater diversity of methanotrophs exists than we have seen in culture, a fact demonstrated by the continued isolation of novel methanotroph taxa (12, 25, 45).

A major goal of microbial ecology is to identify the microorganisms that are actively involved in specific processes in the environment. This has recently been addressed by using the natural abundance of stable isotopes in biomarkers (22, 34) and techniques including fluorescent in situ hybridization (FISH) coupled with microautoradiography (27, 35), analysis of phospholipid fatty acids by gas chromatography-isotope ratio mass spectrometry (3, 36), FISH coupled with mass spectrometry, and stable-isotope probing (SIP) (37, 41), which use substrates enriched with stable isotopes or radioisotopes. SIP exploits the fact that DNA of an organism growing on a ¹³C-enriched carbon source becomes ¹³C labeled (“heavier”), enabling it to be resolved from the total community DNA by density gradient centrifugation (37). The ability to isolate DNA from microorganisms involved in the metabolism of ¹³CH₄ allowed us to characterize the microbial population actively involved in low-affinity CH₄ oxidation in a peat soil microcosm without any prior knowledge of the organisms involved or the need to isolate them into culture.

Sample site and microcosms

The study site was drained fenland peat soil (49 to 51% carbon; pH 6.8) from Suffolk, United Kingdom [42]). The peat sample (0 to 5 cm in depth) was collected in May 1999 from an area not cultivated for 4 years. The in situ soil water content was 65% water-holding capacity, which was optimal for low-affinity (5% [vol/vol]) CH₄ oxidation (S. A. Morris, unpublished data).

Two microcosms were established for the SIP experiment. Each consisted of freshly collected peat soil (10 g at 65% water-holding capacity) in a 125-ml crimp-top serum vial, sealed with a butyl rubber stopper, and injected with 0.4 mmol (10 ml) of CH₄. One microcosm was exposed to ¹³CH₄ (99% pure, 99% ¹³C atom enriched; Linde Gas), and the second microcosm was exposed to an identical amount of ¹²CH₄ (98% pure; Linde Gas). The ¹³CH₄ microcosm was used for extraction of ¹³C-labeled DNA and characterization of the microbial population actively involved in low-affinity CH₄ oxidation.

Microcosms were incubated in the dark at 25°C, and headspace CH₄ concentrations (200-μl samples) were determined every 2 to 5 days by gas chromatography. After >90% of the CH₄ was consumed (7 to 12 days), the vials were flushed with air to remove ¹³CO₂ and ensure that the microcosms remained aerobic. A further 0.4 mmol of ¹³CH₄ or ¹²CH₄ was added, and the vials were incubated until 1.6 mmol of CH₄ (four injections of 10 ml of CH₄) had been consumed by each microcosm (40 days).

DNA extraction. DNA was extracted from 10 g of soil exposed to ¹³CH₄ and 3 g of soil exposed to ¹²CH₄ by a bead-beating method (46) that was scaled up to process 3-g soil samples in 12-ml glass bead beater tubes (Braun). Soil samples were shaken for 5 min in a CO₂ cooled bead beater (Braun). Extracts from the ¹³CH₄ microcosm were pooled, and large particles were removed by centrifugation for 5 min at 120 × g. Proteins in the supernatant were precipitated with potassium acetate (7.5 M; one-sixth of the sample volume) and removed by centrifugation for 5 min at 15,000 × g. To the supernatant, an equal volume of binding matrix (Bio 101) diluted 1:5 with 6 M guanidine isothiocyanate was added. The tube was inverted regularly for 5 min and centrifuged for 5 min at 15,000 × g, and then the supernatant was discarded. The binding matrix was washed twice by resuspension in an equal volume of wash buffer (70% ethanol, 100 mM sodium acetate) and centrifugation at 15,000 × g for 1 min. DNA was eluted in 3 ml of Tris-EDTA (TE) buffer, and 1 g of CsCl was added per ml of DNA solution.

DNA fractions were resolved by equilibrium centrifugation (265,000 × g, 16 h, 20°C) in CsCl-ethidium bromide density gradients (13-by-51-mm polyallomer tubes; Beckman). A single DNA band was observed from the soil exposed to ¹²CH₄, whereas an additional faint “smear” of more dense DNA was observed up to 1 cm below the bright band from the soil exposed to ¹³CH₄ (Fig. 1). Assimilation of both ¹³C- and ¹²C-labeled substrates would result in intermediate-density DNA, as observed in the ¹³C-DNA gradient. The most dense DNA fraction (lower 0.25 cm; ca. 0.35 ml) from the ¹³C gradient (¹³C-DNA) and all DNA from the ¹²C gradient (¹²C-DNA) were collected with a syringe and needle (19 gauge). The ¹³C-DNA fraction was subjected to a second ultracentrifugation to remove any small amounts of coextracted ¹²C-DNA (not visible), providing highly enriched ¹³C-DNA for PCR. DNA fractions were extracted three times with an equal volume of 1-butanol, dialyzed against TE buffer, ethanol precipitated, and dissolved in 100 μl of TE buffer.

FIG. 1.

Equilibrium centrifugation of DNA extracted from soil that oxidized ¹²CH₄ (a) or ¹³CH₄ (b). CsCl-ethidium bromide (100 μl at 10 mg ml⁻¹) density gradients were centrifuged at 265,000 × g for 16 h, and DNA was visualized with long-wavelength (365-nm) UV light. The positions of the ¹²C-DNA (indicated by “[”) and ¹³C-DNA (indicated by “]”) fractions collected after primary ultracentrifugation are indicated. Bar, 1 cm.

PCR amplification, cloning, sequencing, and analysis

The ¹³C-DNA and ¹²C-DNA fractions were used as a template for PCR. Primers (Table 1) specific for bacterial (Eubac27F and 1492R), archaeal (Arch21F and 1492R), and eucaryal small-subunit rRNA genes (EukF and EukR) (13) were used to determine the active methanotroph population at the domain level. With ¹²C-DNA as the template, PCR products of the expected size were obtained with primers that amplified Bacteria, Archaea, and Eucarya small-subunit rRNA genes. However, with ¹³C-DNA, PCR products were only obtained with bacterial 16S ribosomal DNA (rDNA) primers, indicating that ¹³CH₄ was incorporated into a restricted bacterial community. Specific products of the expected size were amplified from both the ¹³C- and the ¹²C-DNA fractions with primers encoding key enzymes in CH₄ metabolism (Table 1).

TABLE 1.

PCR amplification primers used to characterize the ¹³C- and ¹²C-DNA fractions

Amplified gene(s)	Primers	Product (bp)	Library constructeda	Reference
16S rRNA of Bacteria	Eubac27F	1,450	13C-DNA (100)	13
	1492R		12C-DNA (50)
16S rRNA of Archaea	Arch21F	1,450	NAb	13
	1492R
18S rRNA of Eucarya	EukF	1,500	NA	13
	EukR
pmoA and amoA (putative active site subunit of pMMO and AMO)	A189	525	13C-DNA (50)	23
	A682	12C-DNA (50)
pmoA (pMMO active site subunit)	A189	500	13C-DNA (50)	11
	mb661
mxaF (MDH active site subunit)	mxa f1003	550	13C-DNA (50)	30
	mxa r1561
mmoX (sMMO active site subunit)	mmoX F356	350	13C-DNA (10)	16
	mmoX R898

^aDNA template used for PCR amplification and number of clones in the library (in parentheses).

^bNA, not applicable.

Amplification products were cloned with the TOPO-TA cloning kit (Invitrogen). Libraries were constructed from the ¹³C-DNA fraction for the bacterial 16S rDNA (100 clones), pmoA-A189/A682 (50 clones), pmoA-A189/mb661 (50 clones), mxaF (50 clones), and mmoX (10 clones) PCR products. Libraries were also constructed by using the 16S rDNA (50 clones) and pmoA-A189/A682 (50 clones) PCR products from the ¹²C-DNA fraction. Clones were grouped into operational taxonomic units (OTUs) by restriction fragment length polymorphism (RFLP) analysis by using restriction endonucleases as follows: EcoRI/RsaI and EcoRI/Sau3A for 16S rDNA clones, EcoRI/RsaI and EcoRI/PvuII/HincII for pmoA clones, and EcoRI/HincII and EcoRI/RsaI for mxaF clones.

For the 16S rDNA library, full sequence data (ca. 1,450 bp) between Escherichia coli positions 8 and 1511 (6) were obtained for one clone from each OTU containing five or more clones. A partial sequence (positions 375 to 890) was determined for each OTU that contained two to five clones and for selected OTUs with single clones. Complete sequence information (between the PCR primers) was obtained for each OTU in the pmoA libraries, eight mxaF clones, and ten mmoX clones. To verify that each unique restriction pattern represented a single clone type, partial sequence data were obtained for at least 10% of the clones within each OTU.

The ARB package (http://www.mikro.biologie.tu-muenchen.de) was used for sequence alignment and phylogenetic analysis. 16S rDNA sequences were aligned by using the ARB automatic alignment tool (Aligner v2.0) and corrected according to secondary structural constraints. Functional gene sequences were aligned manually to sequences from GenBank, and deduced amino acid sequences were used for the analyses. To evaluate tree topology, phylogenies were reconstructed by using evolutionary distance (DNA, Jukes and Cantor model; amino acids, Dayhoff PAM model), maximum parsimony (default parameters for ARB and DNAPARS or ARB and PROTPARS), and maximum likelihood (default parameters for ARB and fastDNAml or ARB and Protein_ML) analyses in conjunction with various filters and sequence subsets (28). Similarities between 16S rDNA sequences were determined by using the similarity matrix option within ARB. No chimeric 16S rDNA sequences were indicated with the Chimera Check program version 2.7 (http://rdp.cme.msu.edu). Inspection of the DNA and amino acid alignments of functional gene sequences did not identify any potential chimeric sequences. The sequences were deposited in GenBank (AF357990 to AF358055).

16S rDNA libraries

RFLP analysis of 100 16S rDNA clones from the ¹³C-DNA fraction (Fig. 2) assigned most of the library (70 clones) to 13 OTUs. The remaining 30 clones had unique restriction profiles (30 OTUs), and 10 were partially sequenced. Phylogenetic analysis revealed that, with the exception of five OTUs (containing seven clones), all ¹³C-DNA clones clustered in the α, β, or γ subclass of the Proteobacteria (40, 21, and 12 clones, respectively) (Fig. 2). In stark contrast to the ¹³C-DNA library, the 50 16S rDNA clones from the ¹²C-DNA contained 49 unique RFLPs (indicating wide diversity), of which only two were identical to an RFLP from the ¹³C-DNA. This difference was highlighted by sequencing 10 16S rRNA clones from the ¹²C-DNA; these were related to a wide variety of Bacteria (Fig. 2). One additional clone (LO12.10), with an RFLP identical to LO13.19, was sequenced and identified as an α-proteobacterial methanotroph.

FIG. 2.

Maximum-likelihood tree of 16S rDNA sequences amplified from ¹³C-DNA (•) or ¹²C-DNA (○) fractions. From the ¹³C-DNA library (100 clones), 23 OTUs (containing 80 clones) were sequenced, whereas the ¹²C-DNA library (50 clones) contained 49 OTUs (11 sequenced). The number of clones assigned to each sequenced OTU by RFLP analysis is shown in parentheses. OTU sequences were added to the tree of almost full-length sequences (>1,300 bp) in ARB by using maximum-parsimony analysis. Sequences most closely affiliated to each OTU, as well as sequences (>1,300 bp) of known methanotrophs and nitrifiers, were selected. Partial sequences (<1,300 bp) were chosen from the ARB or GenBank databases when no closely related full-length sequences were available. The tree was constructed by using Thermotoga maritima as an outgroup and a 50% conservation filter (962 nucleotides; this excludes positions of ambiguity, missing data, and positions where the frequency of a nucleotide occurring was <50% [28]). Multifurcations indicate branch points where the topology between maximum-likelihood, evolutionary-distance, and maximum-parsimony analyses of the data set were not supported by a strict consensus rule (28). With the exception of LO12.5, LO12.3, and LO13.20, the various treeing analyses consistently recovered all sequences in the groups depicted. Sequences of <1,300 bp (dashed line) were added to the tree by using a maximum-parsimony option within ARB. (A) Cytophagales; (B) δ-Proteobacteria; (C) Acidobacterium-Holophaga group; (D) uncultivated group; (E) Verrucomicrobia; (F) green nonsulfur group; (G) Actinobacteria.

Twenty-six clones in the ¹³C-DNA library were related to extant methanotrophs (96 to 99% 16S rDNA identity). These OTUs were most similar to the α-proteobacterial methanotrophs Methylocystis and Methylosinus (LO13.7 and LO13.19) and Methylocella palustris (LO13.9 and LO13.10) or to the γ-proteobacterial methanotroph Methylobacter (LO13.6 and LO13.13). A further 24 clones were related (92 to 95% identity) to genera in the α subclass of the Proteobacteria. Known methylotrophs (Hyphomicrobium and Methylobacterium) clustered among some (LO13.3, LO13.17, and LO13.8) but not all (LO13.2 and LO13.18) of these OTUs. The third major group of sequences (21 clones; LO13.1, LO13.5, LO13.11, LO13.12, and LO13.21) was closely related to genera within the β-Proteobacteria. Recovery of clones related to Bdellovibrio (LO13.14) and Cytophaga (LO13.4 and LO13.22) may have resulted from turnover of ¹³C due to predation (29, 39). The remaining sequenced clones clustered with the γ-Proteobacteria (LO13.15 and LO13.16), a group of uncultivated bacteria (LO13.20), and the Verrucomicrobia (LO13.23).

Functional gene libraries

Libraries of 50 pmoA clones were constructed with the A189-A682 primer set for both the ¹³C-DNA and the ¹²C-DNA fractions. RFLP and sequence analysis of the ¹³C-DNA library identified three OTUs (LOPA13.2, LOPA13.3, and LOPA13.5) that were similar to PmoA sequences of Methylocystis and Methylosinus (Fig. 3). Other clones formed a distinct group related to PmoA of type II methanotrophs (LOPA13.1), and four clones were similar to AmoA of Nitrosomonas europaea (LOPA13.4).

FIG. 3.

Maximum-likelihood tree of deduced amino acid sequences of pmoA and amoA genes amplified from ¹³C-DNA or ¹²C-DNA fractions. Clones obtained from the ¹³C-DNA fraction are prefixed with “LOPA13” when amplified with the A189-A682 (•) primer set or with “LOPB13” when amplified with the A189/mb661 (⧫) primer set. Clones obtained from the ¹²C-DNA fraction and amplified with the A189-A682 (○) primer set are prefixed with “LOPA12.” The tree was constructed with a filter (163 aligned amino acid positions; this excludes ambiguities and missing data), and the PmoA sequence of the environmental clone RA21 as an outgroup. Multifurcations indicate branch points where the topology between maximum-likelihood, evolutionary-distance, and maximum-parsimony analyses of the data set were not supported by a strict consensus rule. The number of clones (50 in each library) assigned to each OTU by RFLP analysis is shown in parentheses.

Analysis of the corresponding ¹²C-DNA library further demonstrated the different population within the ¹³C-DNA fraction. Thirty-six clones were assigned to OTUs closely related to AmoA of Nitrosomonas (LOPA12.1, LOPA12.2, LOPA12.5, and LOPA12.7) or Nitrosospira (LOPA12.3). OTU LOPA12.4 contained a distinct AmoA-like sequence. The remaining OTUs in the ¹²C-DNA library (LOPA12.6 and LOPA12.8) formed distinct groups that were related to the sequence of the environmental clones RA14 (24) and LOPA13.1, respectively (Fig. 3).

Since type I methanotroph sequences were detected in the 16S rDNA analysis of the ¹³C-DNA fraction, but not in the corresponding pmoA library with the A189-A682 primer set, we also used the methanotroph-specific primer set A189-mb661. Thirty-one clones (LOPB13.1 and LOPB13.3) contained a PmoA that was similar to that of type II methanotrophs (Fig. 3). A further 16 clones (LOPB13.2 and LOPB13.5) clustered with the PmoA of type I methanotrophs. The remaining OTU (LOPB13.4) was identical to LOPA12.6 and most closely related to the environmental clone RA14.

The 50 mxaF clones from the ¹³C-DNA grouped into 32 OTUs and, of the 20 clones sequenced, only 8 showed identity to MxaF. Phylogenetic analysis indicated that four clones (LOM13.1 to LOM13.4) clustered with the MxaF of type II methanotrophs and three (LOM13.5 to LOM13.7) clustered with other α-proteobacterial methylotrophs (Fig. 4). The different analyses of the pmoA and mxaF data sets consistently recovered the clades depicted in Fig. 3 and 4. Of the 10 mmoX clones from the ¹³C-DNA fraction, six different sequences were identified, all of which were similar to MmoX of the type II methanotroph Methylosinus (data not shown).

FIG. 4.

Maximum-likelihood tree of deduced amino acid sequences of mxaF genes amplified from the ¹³C-DNA fraction (•). The tree was constructed with a filter (171 aligned amino acid positions; this excludes ambiguities and missing data), with the amino acid sequence of the structurally related ethanol dehydrogenase of Pseudomonas aeruginosa (26) as an outgroup. Multifurcations indicate branch points where the topology between maximum-likelihood, evolutionary-distance, and maximum-parsimony analyses of the data set were not supported by a strict consensus rule.

Characterization of the population that assimilated ¹³C

In our study, ¹³C was incorporated into the DNA of metabolically active bacteria that used ¹³CH₄ as a sole source of carbon (as shown by the extra, more dense, DNA fraction in the ¹³C-DNA gradient). Interestingly, the yield of ¹³C-DNA was very low relative to that of a methanotroph grown on 1.6 mmol of CH₄ (not shown), suggesting cooxidation of CH₄ in the peat, rapid turnover of methanotroph biomass, or CH₄ production (not observed). Thus, it is probable that a variety of ¹³C-labeled compounds were produced during the 40-day incubation, including intermediates or products of methanotroph metabolism. Such ¹³C-labeled compounds could subsequently be assimilated by organisms other than methanotrophs. However, production of the corresponding ¹²C-labeled compounds (e.g., ¹²CO₂) and other trophic interactions would considerably dilute many non-primary substrates within the complex environment of a soil microcosm. Furthermore, since identical centrifugation conditions separated the ¹³C-DNA and ¹²C-DNA fractions of a pure methanotroph culture grown on ¹³CH₄ or ¹²CH₄ by 1 cm (data not shown), we estimate that the ¹³C-DNA fraction collected (Fig. 1) contained between 75 and 100% ¹³C. For DNA to possess such a high ¹³C content and buoyant density, any non-methanotrophs must have been in very close association with the primary ¹³CH₄ oxidizers. Therefore, it is most likely that the DNA collected in the “heavy” ¹³C-DNA fraction originated from microorganisms that assimilated ¹³C-labeled compounds as a primary carbon source.

Analysis of 16S rRNA and functional genes amplified from the ¹³C-DNA identified a high proportion of clones that were closely related to extant methanotrophs. These results indicate that a variety of methanotrophs had actively assimilated CH₄ in the peat soil microcosm. The 16S rDNA similarity between LO13.10 and the nearest extant methanotroph, Methylocella palustris (12), was <97%, which suggested that a novel methanotroph species was involved in CH₄ oxidation in the microcosm. Since PmoA/AmoA phylogeny of extant strains reflects that obtained with 16S rDNA sequences (23), the pmoA libraries support the activity of both type I and type II methano-trophs in the microcosm. The PmoA-like clones in the ¹³C-DNA fraction (LOPA.13.1 and LOPB13.4) also suggest that an unusual methanotroph within the α subclass of Proteobacteria was active in the peat soil. Recent studies of aerobic ¹³CH₄ or ¹⁴CH₄ oxidation have detected small amounts of uncharacteristic lipid fractions, which also suggests that unusual methanotrophs are active in soils and sediments (3, 8, 24, 38).

The distribution of clones in the libraries from the ¹³C-DNA fraction must be interpreted with great caution due to biases inherent in DNA extraction and PCR (43). However, one striking and unexpected feature of the 16S rDNA library is the large number of clones in OTUs that are not closely related to extant methanotrophs (Fig. 2). It is likely that these organisms had assimilated a high proportion of ¹³C into their DNA, even though the identity of the ¹³C substrate is unclear. Bacteria capable of growth on methanol and other excreted organic compounds copurify during attempts to isolate methanotrophs into culture (17). Indeed, 22 clones (LO13.2, LO13.3, and LO13.8) in the 16S rDNA library and 3 clones (LOM13.5 to LOM13.7) in the mxaF library were sufficiently related to sequences of Hyphomicrobium, Pedomicrobium, Rhodomicrobium, and Xanthobacter spp. to suggest that some methylotrophs within the microcosm may have assimilated ¹³CH₃OH excreted by CH₄-oxidizing bacteria.

The other major group of 16S rDNA sequences (LO13.1, LO13.5, LO13.11, LO13.12, and LO13.21) identified in the ¹³C-DNA library clustered within the β subclass of Proteobacteria. No functional genes indicative of β-proteobacterial methylotrophs were identified, a result that is similar to those of other environmental surveys done with mxaF PCR primers (18, 19, 21, 30, 37) and may be due to a primer bias. Unfortunately, therefore, it remains unclear which ¹³C substrate these organisms had assimilated in the peat soil microcosm. Isolates of β-proteobacterial methylotrophs are relatively rare (5, 40), and therefore their diversity might be underdescribed. Our data suggest that some β-Proteobacteria are actively involved in the cycling of carbon after CH₄ oxidation in soil. These may represent novel methylotrophs, distinct from most extant strains within the α and γ subclasses of Proteobacteria. This ability of SIP to characterize the community in the ¹³C-DNA fraction with phylogenetic and functional gene PCR primers makes it a powerful technique for resolving microbial structure-function relationships in complex environments such as soil.