Present addresses: Per Bengtson, Department of Microbial Ecology, Ecology Building, Sölvegatan 37, 223 62 Lund, Sweden. Marc G. Dumont, Max Planck Institute for Terrestrial Microbiology, Marburg D-35043, Germany. Nathan Basiliko, Department of Geography, University of Toronto Mississauga, 3359 Mississauga Road North, Mississauga, ON, Canada L5L 1C6.
Correspondence: Per Bengtson, Department of Forest Sciences, University of British Columbia, 2424 Main Mall, Vancouver, BC, Canada V6T 1Z4. Tel.: +46 46 222 3760; fax: +46 46 222 3800; e-mail: email@example.com
The main gap in our knowledge about what determines the rate of CH4 oxidation in forest soils is the biology of the microorganisms involved, the identity of which remains unclear. In this study, we used stable-isotope probing (SIP) following 13CH4 incorporation into phospholipid fatty acids (PLFAs) and DNA/RNA, and sequencing of methane mono-oxygenase (pmoA) genes, to identify the influence of variation in community composition on CH4 oxidation rates. The rates of 13C incorporation into PLFAs differed between horizons, with low 13C incorporation in the organic soil and relatively high 13C incorporation into the two mineral horizons. The microbial community composition of the methanotrophs incorporating the 13C label also differed between horizons, and statistical analyses suggested that the methanotroph community composition was a major cause of variation in CH4 oxidation rates. Both PLFA and pmoA-based data indicated that CH4 oxidizers in this soil belong to the uncultivated ‘upland soil cluster α’. CH4 oxidation potential exhibited the opposite pattern to 13C incorporation, suggesting that CH4 oxidation potential assays may correlate poorly with in situ oxidation rates. The DNA/RNA-SIP assay was not successful, most likely due to insufficient 13C-incorporation into DNA/RNA. The limitations of the technique are briefly discussed.
Methane oxidation occurs in most soils and is a major sink for CH4 worldwide, with upland soils responsible for >5% of annual removal of CH4 from the atmosphere (Schlesinger, 1997). The process has been known for decades, but the organisms involved and the factors determining the rate of the process are still not completely understood. Early studies suggested that two types of methanotrophs existed, those related to cultivated strains with a relatively low affinity for CH4, and uncultivated organisms with a high affinity, which are responsible for atmospheric CH4 consumption (Bender & Conrad, 1992). The theory that all cultivated methanotrophs have a low affinity for CH4 and are unable to oxidize atmospheric concentrations (1.8 p.p.m.v.) has been complicated by the discovery of some type II methanotrophs with the ability to oxidize CH4 at an atmospheric concentration for prolonged periods (Dunfield et al., 1999; Knief & Dunfield, 2005). These organisms also grow rapidly with high CH4 concentrations, and recent studies on Methylocystis spp. have revealed that this is achieved by producing both low- and high-affinity particulate methane mono-oxygenase (pMMO), i.e. one of the enzymes that converts methane to methanol (Baani & Liesack, 2008). It is believed that these organisms may possess the ability to sustain oxidation at low CH4 concentrations, but require additional energy sources, or alternatively nearby anaerobic microsites with elevated CH4 concentrations, for long-term survival (Knief & Dunfield, 2005; Baani & Liesack, 2008).
Cultivation-independent studies of the methanotrophs in upland soils exhibiting atmospheric CH4 consumption suggest they are dominated by pmoA sequence types from uncultivated bacteria, such as what is now commonly referred to as the upland soil cluster (USC) α and USCγ clades, and labeling of lipids with 13CH4 indicate that phospholipid fatty acids (PLFAs) are distinct from those present in cultivated methanotrophs (Holmes et al., 1999; Bull et al., 2000; Knief et al., 2003; Maxfield et al., 2006). These and other findings illustrate the limitation of cultivation-dependent studies in studying structure–activity relationships. For example, reduced methane oxidation rates as a result of fertilization seem to be caused by a decreased biomass of yet uncultivated high-affinity methane oxidizers (Maxfield et al., 2008).
Methane oxidation to methanol is catalyzed by particulate (p) and soluble (s) methane mono-oxygenase (MMO). pMMO (a subunit of which is encoded by pmoA) is present in all known methanotrophs except Methylocella sp., which uses an sMMO (the α subunit of which is encoded by the mmoX gene) and is common in northern wetlands (Dedysh et al., 1998; Dedysh, 2002). pmoA is useful for phylogenetic characterization of methanotrophs in soils because pmoA phylogeny is congruent with the 16S rRNA gene phylogeny in methanotrophs (Kolb et al., 2003). DNA-SIP, where the incorporation of 13C from 13C-labeled CH4 into pmoA or other signature nucleic acids is measured, can therefore be used to identify active methanotrophs involved in CH4 oxidation (Dumont & Murrell, 2005; Madsen, 2005). Microbial PLFAs are another group of biomarkers that have the potential to be used as indicators of the microorganisms involved in the turnover and flow of carbon in ecosystems. For example, the monoeneic PLFA 16:1ω8 is considered to be a biomarker of type I methanotrophs and 18:1ω8 of type II methanotrophs (Hanson & Hanson, 1996; Knief et al., 2006; Maxfield et al., 2008). This technique is less phylogenetically informative than nucleic acid-SIP because the same PLFA may occur in more than one group/species of methanotroph. Furthermore, one type II methanotroph possesses large quantities of the 16:1ω8c (Dedysh et al., 2007), previously though to be a type I methanotroph-specific biomarker. However, a major advantage of PLFA analyses is that several orders of magnitude lower levels of 13C enrichment can be detected in PLFAs compared with nucleic acids, and PLFA analysis has also proven to be a more powerful technique to detect variations/changes in microbial communities compared with many DNA-based techniques (Ramsey et al., 2006).
The objective of this study was to use culture-independent methods to identify the active CH4 oxidizers in three soil horizons from a lodgepole pine forest, in an attempt to establish structure–activity relationships for this group of microorganisms. This was achieved by exposing soil cores to 13C-enriched (>99 atom%) CH4, and then following the label into specific biomarkers (DNA, RNA, and PLFAs), enabling us to reveal the identity of the active methanotrophs.
Materials and methods
Study site and soil sampling
Lodgepole Pine (Pinus contorta) forest located approximately 100 km east of Prince George, BC, near Kenneth Creek, approximately 1000 m above sea level, and in the sub-boreal spruce biogeoclimatic ecosystem classification zone (Meidinger & Pojar, 1991) was sampled for this study. The site is well characterized and preliminary observations had shown that it has an active methanotroph community. The site was burnt as part of site preparation after the last harvesting and replanted in 1981. Prince George has had average January temperatures of −9 °C and average August temperatures of 15 °C and has received between 600 and 700 mm precipitation per year (Weather Network). The site had very little understory consisting of small graminoid plants and shrubs. The soil belonged to the Eluviated Dystric Brunisol subgroup (Soil Classification Working Group, 1998), had a 4-cm-deep organic (LFH) horizon, a distinct 4–8-cm-deep eluvial (Ae) horizon, and a weekly colored Bf horizon <10 cm deep. Eight soil cores were collected on April 27, 2006, in 25-cm-long PVC pipes. The soil cores were 15 cm long × 7.7 cm inner diameter, with 10 cm headspace (c. 500 mL). The headspace was filled with a Styrofoam plug to keep soil intact, PVC end caps were sealed with vacuum grease and plastic tape, and cores were transported by air to the University of Victoria. The top end caps and plugs were removed and cores were kept at room temperature for 24 h.
13CH4 enrichment and CH4 oxidation measurements
Cores were sealed and oxidation of ambient CH4 was measured by sampling 0.5 mL of headspace gas through butyl-rubber septa in the top end caps every 15 min over 1 h, and was analyzed with a Varian 3800 gas chromatograph (Palo Alto, CA) equipped with a packed column and flame ionization detector. After calibration relative to a CH4 standard of 1.05 p.p.m. and calculation of masses of CH4 using the ideal gas law, oxidation rates were calculated as the linear decrease in mass of CH4 in headspace over time per area of soil. Immediately following measurements of ambient CH4 oxidation rates, end caps were removed, cores were aerated with room air, and resealed using vacuum grease and a Parafilm M (American National Can, Chicago, IL) seal around end caps. Septa were inserted after caps were in place to ensure no pressurization occurred during capping. To achieve headspace concentrations of 50 and 10 000 p.p.m.v., 2.5 mL of a 10 000 p.p.m. CH4 stock or 5 mL of pure CH4 containing either 12C or 13C (>99%13CH4 from Cambridge Isotopes, Andover, MA) was added to each core after the same volume of headspace air had been removed. Two of each core received 50 p.p.m.v. 12CH4, 50 p.p.m.v. 13CH4, 10 000 p.p.m.v. 12CH4, and 10 000 p.p.m.v. 13CH4. One empty control core was incubated with 10 000 p.p.m.v. CH4 to measure potential leakage from the cores. Initial incubated oxidation rates were measured as above based on concentration measurements taken four times over the first 24 h. Incubations were run for 31 days and cores were opened, aerated, and CH4 was reloaded into headspaces every 3 days. Oxidation rates were measured eight times over the 31-day period by sampling 500 μL of headspace gas daily and using GC as described above. After the final rate measurement on day 31, cores were opened, aerated for 2 h in a fumehood, and the rate of oxidation of ambient (c. 2 p.p.m.) CH4 in room air was measured over a 2-h period. Cores were carefully dissected into three depths: the surface organic horizons (c. 4 cm), the 5-cm upper mineral horizon (containing mainly the Ae horizon), and the remaining 9-cm lower mineral horizon. Samples from each depth were placed in sterile bags and homogenized by hand and then subdivided into samples for CH4 oxidation potential assays, DNA and RNA extraction, and freeze drying for lipid extraction.
CH4 oxidation potential assays in vitro
CH4 oxidation potentials (Basiliko et al., 2004; Glatzel et al., 2004) of individual soil depths were determined in 25-mL crimp-top serum vials each with approximately 1 g of field moist soil, 5 mL of sterile deionized water to ensure a homogeneous slurry, and pure CH4 added to achieve concentrations of 50 and 10 000 p.p.m.v. for each depth segment. The eight cores were dissected into 24 samples, and each depth segment was incubated in separate vials at two concentrations. Incubations occurred over 24 h in the dark at 20 °C and rates of oxidation were measured based on three concentration measurements as described above, except that rates were calculated per gram dry soil after oven drying vials and soils at 105 °C.
PLFA extraction and analysis
Lipids were extracted from ∼1 g (freeze dried) organic soil or ∼3 g mineral soil using a modified Bligh and Dyer method, and fractionated into neutral lipids, glycolipids, and phospholipids on a silica-bonded phase column by elution with chloroform, acetone, and methanol, respectively (Frostegård et al., 1991). A known amount of methyl nonadecanoate (19:0) was added to the fraction containing the phospholipids and the phospholipids were then transmethylated to their fatty acid methyl esters using mild alkaline methanolysis (Dowling et al., 1986).
The PLFAs were separated and identified on an Agilent 6890 N GC and connected to an Agilent 5975 Inert XL Mass Selective Detector (Agilent Technologies Inc., Santa Clara, CA). The column used was a 30-m J&W HP-5 column (5% phenyl-methylpolysiloxane) with an ID of 0.32 mm and a film thickness of 0.25 μm kept at a constant pressure of 76 kPa. The injector was kept at 225 °C and the samples were injected in a pulsed splitless mode. The temperature program was as follows: 50–150 °C at 15 °C min−1, 150–190 °C at 4 °C min−1, 190–200 at 1 °C min−1, 200 °C for 5 min, 200–270 °C at 5 °C min−1, and then kept at 270 °C for 10 min. The PLFAs were identified by means of a combination of mass spectra and retention times relative to authentic standards and the internal standard 19:0. The total and 13C concentration in the individual PLFAs were determined on an Isoprime™ stable-isotope ratio mass spectrometer (GV Instruments) connected to an Agilent 6890A gas chromatograph (Agilent Technologies Inc.) via a combustion interface, under the same chromatographic conditions as above. The PLFAs are designated in terms of total number of carbons:number of double bonds, followed by the position of the double bond from the methyl end of the molecule. The prefixes a and i indicate iso and aniso branching, and cy refers to cyclopropane fatty acids.
Nucleic acid extraction, characterization of CH4 mono-oxygenase genes, and nucleic acid stable-isotope probing (SIP)
Total soil DNA was extracted from each soil depth sample following Basiliko et al. (2003), using a modified protocol included with the FastSpin Kit for Soil (Qbiogene Inc., Carlsbad, CA) involving four additional washes with 400 μL of 5 M guanidine thiocyanate to remove humic substances and proteins. PCR amplification of partial pmoA and amoA genes was carried out using 189f/682r primer set (Holmes et al., 1995). The soluble CH4 mono-oxygenase gene (mmoX) was targeted using 945f/1401r, 206f/886r, and 166f/1402r PCR primer sets, generally following Dedysh et al. (2005), Hutchens et al. (2004), and Auman et al. (2000), respectively. PCR amplicons were visualized using agarose gel electrophoresis. PCR was attempted under altered conditions when amplicons of expected sizes were not amplified, including use of FailSafe™ PCR System (Epicentre Biotechnologies, Madison, WI) with 12 PCR PreMixes, varying primer annealing temperatures, and using touchdown temperature programs. PCR products were excised from agarose gels and cleaned using the QIAQuick Gel Extraction Kit (Qiagen, Hilden, Germany) following the manufacturers' instructions. Cloning and clone screening of pmoA and putative mmoX fragments followed protocols outlined in the TOPO TA Cloning® kit with TOP10 competent cells (Invitrogen Corp., Carlsbad, CA). Ligation reactions were carried out with gel-purified PCR products mixed from the same depths across replicate cores that were incubated under the same CH4 concentration before transformations. Sequencing of 12 pmoA and 12 purported mmoX clones reamplified with M13f/M13R PCR primers and purified with the QIAQuick PCR Purification Kit, was carried out on an ABI PRISM 3130 xl Genetic Analyzer by the University of Warwick Molecular Biology Service using the vector-specific M13f primer in the sequencing reactions. Nucleotide blast® 2.2.18+searches (Altschul et al., 1997) were used to retrieve most similar sequences amplified from other soil environments and from closest isolated methanotrophs. Sequences were aligned using the bioedit sequence alignments editor v 220.127.116.11 and phylogenetic trees were built and edited using embedded phylip and tree view (http://taxonomy.zoology.gla.ac.uk/rod/rod.html) applications (Hall, 1999). Neighbor-joining and maximum likelihood (dnaml v3.5, Felsenstein, 1989) tree-building algorithms produced nearly identical results. One sequence was chosen to represent multiple sequences when there was >98.5% similarity/identity calculated using bioedit. Sequences have been deposited in GenBank under the accession numbers FJ638419–FJ638421. DNA-SIP was carried out in eight of the 24 DNA extracts representing organic and surface mineral soils from four cores (13CH4 and 12CH4 incubated at 50 and 10 000 p.p.m.v.) that had oxidized the largest amount of CH4 over the 1 month incubation following the protocol of Neufeld et al. (2007). Briefly, 2.5 μg of DNA was mixed with gradient buffer and CsCl to a final mixture density of 1.71 g mL−1. Tubes were sealed and then centrifuged at 177 000 g for 48 h, and for each tube, 12 gradient fractions were isolated using a syringe needle and peristaltic pump and slowly displacing gradient/DNA solution with water and collecting fractions from a hole in the bottom of each ultracentrifuge tube. The resulting density gradients spanned from 1.68 to 1.74 g mL−1. DNA was precipitated overnight with glycogen and polyethylene glycol and pelleted by centrifugation at 4 °C (Neufeld et al., 2006). The CsCl solution was removed through aspiration and pellets were washed with 80% ethanol, dried, and then dissolved in molecular biology-grade water. DNA from gradient fractions from paired 12C and 13CH4 samples was used in PCR-denaturing gradient gel electrophoresis (DGGE) with bacterial 16S rRNA gene primers 341fGC and 906r (Muyzer et al., 1993) on a DCode system (Bio Rad Laboratories, Hercules, CA) and gels were analyzed visually for additional bands in the denser fraction of 13CH4 incubated samples relative to 12CH4 controls. RNA-SIP was performed with samples from the most active 10 000 p.p.m.v. incubations. Total soil RNA was extracted using the method of Griffiths et al. (2000). RNA clean-up, CsTFA ultracentrifugation, and the analysis of 16S rRNA gene transcripts from the gradient were performed as described previously (Hamberger et al., 2008).
Differences in 13C incorporation into PLFAs among the different horizons and between CH4 concentrations were assessed with a two-way anova followed by ShefféF-test. Differences in the community composition of the soil microorganisms among the different horizons and between CH4 concentrations were assessed with a principal component analysis (PCA). The abundance of the individual PLFAs (expressed as percent of total PLFA-C) was used as input data. Differences in community composition of the CH4 oxidizers (i.e. the microorganisms incorporating the 13C label) were assessed in the same way, only that the abundance of 13C in individual PLFAs (expressed as percent of total PLFA-13C) was used as input data. The score on the first principal axis (PC1) was recorded and the influence of the community composition and CH4 concentration on the total 13C incorporation (log10 transformed) into PLFAs was determined by a two-way anova with CH4 concentration and PC1 as predictors. A second, independent, test of the influence of the community composition on CH4 oxidation rates was conducted by performing an ancova, where the CH4 concentration was entered as a categorical predictor and the absolute or relative concentration of individual highly enriched PLFAs as continuous predictors, by a forward stepwise regression procedure (F to enter=4.000, F to remove=3.996). All statistical analyses were performed with the statistical package statistica 6.1 (StatSoft Inc., Tulsa).
The potential CH4 oxidation rates of the soils were measured after the labeling, and it varied between 0.64 ng CH4 g−1 soil dw day−1 in the upper mineral horizon (mineral 1) at 50 p.p.m.v. CH4 and 243 ng CH4 g−1 soil dw day−1 in the organic horizon at 10 000 p.p.m.v. CH4 (Table 1). The methane oxidation potential was the highest in the organic horizon at both 50 and 10 000 p.p.m.v. CH4, and similar in the two mineral horizons (Table 1). In the two mineral horizons, the potential oxidation at CH4 concentrations of 50 p.p.m.v. was higher in samples that had originally been incubated at 10 000 p.p.m.v. CH4 than in samples that had been incubated at 50 p.p.m.v. CH4 (Table 1). The difference in potential oxidation between the two concentrations was much lower and not significant in the organic soil. The difference between samples that had originally been incubated at 10 000 p.p.m.v. CH4 and samples that had been incubated at 50 p.p.m.v. CH4 was only observed when the potential oxidation was measured at 50 p.p.m.v. and not at 10 000 p.p.m.v. CH4 (Table 1). Consistent with the stimulation observed for the mineral soil, CH4 oxidation rates in cores over the last 24 h of the 1 month enrichment were higher than over the first 24 h of enrichment in those that had been incubated at 10 000 p.p.m.v. CH4 (Table 2). Conversely, there was no change in the rate of CH4 oxidation rate at 50 p.p.m.v. CH4 between the initial and final measurements (Table 2).
Table 1. Average potential CH4 oxidation rates (ng CH4 g−1 soil dw day−1) measured in serum vials over the first 24 h of enrichment with 50 and 10 000 p.p.m.v. CH4 in headspaces with soil taken from three depths from cores
CH4 in vial (p.p.m.v.)
Original CH4 concentrations in 1-month core incubations
10 000 p.p.m.
SDs of four replicates are given in parentheses. Fluxes were the slopes of statistically significant (P<0.05) and strongly correlated linear regressions with at least four measurements/calculations of decreasing headspace CH4 over time adjusted for CH4 removed through sampling.
Significant differences (P<0.05) between samples that had been incubated with 50 and 10 000 p.p.m. CH4.
Table 2. CH4 oxidation rates (μg CH4 m−2 h−1) in cores at ambient CH4 concentration before enrichment, over the first 24 h of enrichment with 50 and 10 000 p.p.m.v. CH4 in headspaces, over the last 24 h of the 1-month enrichment, as well as the total oxidation in each core (μg C) over the 31-day incubation
First 24 h (μg CH4 m−2 h−1)
Final 24 h
Total (μg C)
50 p.p.m. 12CH4
50 p.p.m. 12CH4
50 p.p.m. 13CH4
50 p.p.m. 13CH4
10 000 p.p.m. 12CH4
10 000 p.p.m. 12CH4
10 000 p.p.m. 13CH4
10 000 p.p.m. 13CH4
Between 0.2 and 42.9 ng 13C g−1 soil dw was incorporated into PLFAs during the experiment. Significantly more 13C was incorporated into PLFAs at high CH4 concentrations (10 000 p.p.m.) than at the lower concentration of 50 p.p.m.v. (Fig. 1, P<0.001, two-way anova). The degree of 13C incorporation also differed between horizons, with low incorporation in the organic soil and much higher in the two mineral horizons (Fig. 1, P<0.001, two-way anova). Thus, the 13C incorporation exhibited the opposite pattern to potential oxidation. The highest 13C incorporation occurred in the upper mineral horizon (mineral 1, P<0.05, SchefféF-test). In the two mineral horizons, br18:0 and 18:1ω7 were most enriched, followed by i17:0 and the unidentified PLFA × 2 (tentatively identified as br19:0) (Fig. 2). The 13C label was much more evenly distributed between PLFAs in the organic horizon (Fig. 2), but CH4 oxidizers in both the organic and the mineral horizons incorporated most 13C into the PLFA 18:1ω7 (Fig. 1). Rapid CH4 oxidation occurred instantaneously in the cores under the incubation conditions and a total of between 450 μg CH4-C (core 4 at 50 p.p.m.v.) and 33.9 mg CH4-C (core 8 at 10 000 p.p.m.v.) was oxidized during the 31-day incubation (Table 2).
PCA analysis of the PLFA data revealed that the community composition of the microbial community differed between horizons. The first principal component separated the organic horizon from the two mineral horizons, while the second principal component separated the two mineral horizons (Fig. 3b). The community composition of the CH4-oxidizing community, i.e. the microorganisms incorporating the 13C label into the different PLFAs, strongly differed between the organic and mineral horizons (Fig. 3a), The concentration of CH4 (50 or 10 000 p.p.m.v.) did not seem to have an influence on the community composition.
The observation that both CH4 oxidation rates and the composition of the microbial community incorporating the 13C label differed between organic and mineral soil does not necessarily enable us to infer that the differences were connected. We therefore performed an anova with CH4 concentration and PC1 as predictors. The results showed that total 13C incorporation was determined not only by the CH4 concentration (P<0.001, two-way anova), but that it was also related to the community composition of the CH4-oxidizing community as expressed by PC1 (P<0.01). In order to further elucidate these results, we performed an ancova where the CH4 concentration was entered as a categorical predictor and the absolute or relative concentration of individual highly enriched PLFAs as continuous predictors by a forward stepwise regression procedure. We chose to include only the four most enriched PLFAs in the analysis, i.e br18:0, 18:1ω7, i17:0, and × 2 (br19:0), because these are most likely to be representative of the CH4-oxidizing community and it has been demonstrated that recycling of 13C within the microbial community may occur (Lueders et al., 2004). Even though most 13C was incorporated into 18:1ω7c (Fig. 1), the relative concentration of this PLFA in the microbial community was not related to 13C incorporation (Table 3), probably because 18:1ω7c occurs in several microbial groups apart from CH4 oxidizers (Zelles, 1999). In contrast, 13C incorporation was positively related to both the absolute and relative concentration of br18:0 in the whole microbial community, as well as the relative concentration of br18:0 in the CH4-oxidizing community (Table 3), reinforcing the observation that CH4 oxidation rates were related to the community composition of the CH4 oxidizers.
Table 3. Summary of ancova analyses testing the dependency of the total 13C incorporation (ng g−1 soil dw day−1) into PLFAs during the experiment on the following factors: (1) the relative (%) abundance of total C in the highly enriched PLFAs (expressed as percent of total PLFA-C); (2), the absolute (ng g−1 soil dw day−1) abundance of the enriched PLFAs (expressed as total C in individual PLFAs); and (3) the relative (%) abundance of 13C within the highly enriched PLFAs (expressed as percent of total PLFA-13C)
Variables in final model
Significance full model
Methane concentration was entered as a categorical predictor and the absolute or relative concentration of individual highly enriched PLFAs as continuous predictors by a forward stepwise regression procedure (F to enter=4.000, F to remove=3.996). The total 13C incorporation was log10 transformed before the analysis.
Percent of total PLFA-C
× 2 (br19:0)
Total C in individual PLFAs
× 2 (br19:0)
Percent of total PLFA-13C
× 2 (br19:0)
Using the A189F–A682R PCR primers that target amoA and pmoA, it was possible to obtain PCR products of the correct size from mineral soil DNA; however, a product could not be amplified from organic soil under any conditions tested. [Bacterial 16S rRNA gene was easily amplified and generated similar amplicons (in terms of abundance) from both organic and mineral bulk DNA extract, excluding the possibility of PCR problems.] In addition, mmoX fragments could not be amplified from any sample. Identification of the pmoA genes present in the mineral horizons from cores incubated under 50 and 10 000 p.p.m.v. CH4 indicated that they were similar to the as yet uncultivated USCα sequences characteristic of many upland soils that oxidize atmospheric CH4 (Fig. 4 and references cited within). All clone sequences analyzed were highly similar, and resulted in only three unique phylogenetic sequences within the 98.5% similarity threshold and were most similar to each other and a pmoA sequence isolated from a Danish spruce forest soil (Bourne et al., 2001, Fig. 4). The closest related cultivated isolate was Methylocapsa acidiphilia B2 (Fig. 4). We attempted to recover 13C-labeled nucleic acids using SIP methods (Neufeld et al., 2007), but were unable to detect any enrichment, and bacterial 16S rRNA gene primers did not result in any unique DGGE banding patterns between 12CH4 and 13CH4 incubated soils (data not shown).
Although it is generally assumed that the compositions of soil communities are related to soil function, evidence to support this hypothesis is scarce. Our results show that a combination of SIP of PLFAs and pmoA gene sequencing could be successfully used to demonstrate structure–activity relationships between CH4 oxidation rates and methanotrophic community structure in a pine forest soil. In fact, two independent analyses suggest that at a given CH4 concentration, the rate of CH4 oxidation is a direct result of, or at least related to, the community composition of the methanotrophs, and that even PLFA patterns of the whole microbial community might be indicative of the degree of CH4 oxidation. More specifically, the abundance of a single PLFA, br18:0, seems to be proportional to CH4 oxidation rates. High levels of 13C incorporation into branched PLFAs such as i17:0, but not br18:0, has been reported in other PLFA-SIP experiments (e.g. Knief et al., 2003; Crossman et al., 2005, 2006; Maxfield et al., 2006). Bull et al. (2000) found an enrichment pattern similar to that of the above-mentioned studies after incubation at atmospheric 13CH4 concentrations, although the 13C label was recovered in a different branched 17:0 PLFA. When the same soil was incubated at 100 p.p.m.v. CH4, the most heavily labeled PLFA was a br18:0. It is possible that this PLFA is representative of a novel methanotroph. However, because br18:0 occurs only at minute concentrations close to the detection limit (Bull et al., 2000; this study), it is more likely that it has not been noted/detected before in PLFA-SIP experiments performed at low or atmospheric CH4 concentrations.
In accordance with findings by, for example, Crossman et al. (2005) and Neufeld et al. (2008), high substrate concentrations did not seem to alter the community composition of the methanotrophs in this study. Long-term exposure to high CH4 concentrations did, however, result in higher potential CH4 oxidation rates in the mineral horizons. Bender & Conrad (1992) noted that upland soils that oxidize atmospheric CH4 could be stimulated to rapidly consume elevated CH4 by incubation with 20% CH4 in air for 3 weeks. After the incubation, the soils exhibited biphasic CH4 oxidation kinetics, suggesting the enrichment of a ‘low-affinity’ CH4 uptake characteristic, with a high Km and high Vmax, which is characteristic of many methanotrophs in culture (Bender & Conrad, 1992; Dunfield et al., 2007). The low-affinity methanotrophs were stimulated only when soils were incubated with CH4 exceeding 100–1000 p.p.m.v. (Bender & Conrad, 1995). This stimulation does not always occur (Bradford et al., 2001; Reay & Nedwell, 2004). In this study, some stimulation was observed for soil cores incubated under 10 000 p.p.m.v. and not 50 p.p.m.v. CH4 (Table 2). The amount of stimulation, of less than threefold, does not nearly match the level of stimulation, suggesting the growth of ‘low-affinity’ methanotrophs, which would correspond to oxidation rates that are orders of magnitude greater than those observed here. The lack of a significant increase in the potential CH4 oxidation as a response to long-term exposure to high CH4 concentration in the organic horizon, in combination with very low 13C incorporation into PLFAs and failed attempts to extract and amplify pmoA, again suggests that the oxidation was disconnected from production of ‘low-affinity’ methanotroph biomass in this horizon. Potential oxidation assays may, therefore, not necessarily be proportional to in situ biomass/activity of bacterial methanotrophs. Alternatively, our findings reflect an uncoupling of methane oxidation and biomass production due to, for example, nutrient limitation (Radajewski et al., 2002) or the use of additional energy sources (Knief & Dunfield, 2005; Baani & Liesack, 2008).
Although we used CH4 concentrations as high as 10 000 p.p.m.v. to produce DNA or RNA sufficiently 13C enriched to be separated by density centrifugation, our attempts were unsuccessful. Buckley et al. (2007) demonstrated that a particular DNA species must have at least 30 atom%15N to be conclusively resolved from its unlabeled counterpart; given that the shift in buoyant density is approximately two times higher for 100 atom%13C-labeled DNA compared with 100 atom%15N labeled DNA, approximately 15 atom%13C should be required using a 13C substrate. The maximum 13C enrichment of PLFAs in our experiment was 1.9 atom%. If we assume that an equivalent incorporation of 13C into target nucleic acids was achieved, the labeling would be almost 10 times lower than necessary to sufficiently increase the buoyant density of the DNA or rRNA. In contrast, an enrichment of 1.9 atom%13C into lipids is orders of magnitude higher than the detection limit of the 13C PLFA-SIP technique. Although less informative, PLFA-SIP is the preferred choice for identifying metabolically active microorganisms and for trying to determine structure–activity relationships in complex microbial communities when the amount of labeling is insufficient for DNA- or RNA-SIP.
The authors graciously acknowledge funding from the Natural Resource and Engineering Research Council of Canada (NSERC), primarily through a Strategic Research Grant to S.J.G. and colleagues and a UK Royal Society Grant to N.B. and J.C.M. Work in the laboratory of JCM was supported by the Natural Environment Research Council (UK). P.B. is funded by the Swedish Research Council. Paul Sanborn (University of Northern British Columbia) and Rob Brockley (BC Ministry of Forests) kindly assisted with, and allowed access to, their long-term research site near Kenneth Creek, BC.