Introduction

Most climate models predict that by 2080, summer temperatures in Arctic regions will increase by 4.0–7.5 °C, and there will also be increased precipitation (Anisimov and Fitzharris, 2001). Motivated by the demonstrated importance of temperature on emissions of the greenhouse gas methane from Arctic and subarctic wetlands (Christensen et al., 2003), several studies have addressed the effect of temperature on the methane production potential in these soils (Svensson, 1984; Dunfield et al., 1993; Kotsyurbenko et al., 1996, 2004; Metje and Frenzel, 2005, 2007; Ganzert et al., 2007).

Recently, archaeal community analysis was included in studies of temperature effects on methanogenesis and methanogenic pathways in subarctic permafrost peat from Northern Finland (Metje and Frenzel, 2005) and Siberia (Metje and Frenzel, 2007). Both the archaeal community structure and the dominating methanogenic pathways differed significantly between the studied sites. In samples of acidic (pH 4.1) peat from Northern Finland, the community was dominated by hydrogenotrophic methanogenesis and all methanogenic sequences belonged to Methanobacteriales (Metje and Frenzel, 2005). In contrast, in samples of weakly acidic (pH 5–6) peat from Siberia, acetoclastic methanogenesis accounted for about 70% of the methanogenesis, and the archaeal community included Methanosarcinales and Methanobacteriales (Metje and Frenzel, 2007). For both sites, increased temperature resulted in few changes in the archaeal community structure, and the methane production rates at 4 °C corresponded to as much as 10 and 17% of the rates at the optimum temperature (25 °C and 26–28 °C, respectively).

The archaeal communities in peat from the weakly acidic, 6.0 in MgCl2 (pH) wetland at Solvatnet, Spitsbergen, include representatives of a wide range of methanogenic groups, including groups with potential for hydrogenotrophic methanogenesis (Methanomicrobiales, Methanobacteriaceae), acetoclastic methanogenesis (Methanosaeta) and both acetoclastic, hydrogenotrophic and methylotrophic methanogenesis (Methanosarcina) (Høj et al., 2005). Previous studies of rice paddy soils, which harbor similar methanogenic groups, have demonstrated changes in the archaeal community structure with temperature (Fey and Conrad, 2000). No information is however available on how increased temperature will affect the archaeal community structure in Arctic peat where all these methanogenic groups are represented.

Previous studies of archaeal communities in subarctic and Arctic wetlands have repeatedly detected archaeal groups not expected to be involved in methane production (Kotsyurbenko et al., 2004; Metje and Frenzel, 2005; Høj et al., 2005, 2006). The studies indicate that these groups have a high relative abundance in the archaeal community, though little attention has been given to these findings as the studies were focused on the methanogenic community. In contrast, increasing attention has been given to the relatively high abundance of archaea in cold nonmethanogenic systems such as the Southern Ocean (DeLong et al., 1994; Murray et al., 1998; Church et al., 2003), the Arctic Ocean (Wells and Deming, 2003; Bano et al., 2004; Garneau et al., 2006; Kirchman et al., 2007) and an Arctic river (Galand et al., 2006). No information is however available on how temperature affects the diversity and abundance of these nonmethanogenic archaeal populations.

In this study, unamended slurries of peat from Solvatnet were incubated at six temperatures in the range 1–25 °C. After 4 weeks incubation, active archaeal cells were visualized and counted using fluorescent in situ hybridization (FISH) and the archaeal community structure was studied in detail for two selected temperatures (5 and 20 °C) by cloning of archaeal 16S rRNA genes. Archaeal community denaturing gradient gel electrophoresis (DGGE) fingerprints for six incubation temperatures and two time points revealed whether apparent temperature trends changed with time, and hence were due to indirect rather than direct temperature effects. The effect of temperature on microbial activities was addressed by monitoring the accumulation of methane, CO2 and volatile fatty acids.

Materials and methods

Study site and sample collection

The site Solvatnet is located on a marine terrace outside the township of Ny-Ålesund (78°50′N–11°30′E) on the west coast of Spitsbergen. The site and peat characteristics, as well as the in situ archaeal community composition at two depths, have been described in detail elsewhere (Høj et al., 2005, 2006). The peat is classified as undecomposed (H1) and almost undecomposed (H2) using von Post's scale (Clymo, 1983). The peat has a high organic content (88% of dry weight), a high water content (>600% of dry weight) and is weakly acidic (pH 6.8 in H2O). The site is grazed by Barnacle Geese (Branta leucopsis). Previously recorded CO2 emissions were in the range 0.18–0.51 g m−2 h−1, and methane emissions were in the range 93–2801 μg m−2 h−1 (Høj et al., 2005). Peat samples for this experiment were collected in mid-August 2000 using polyvinyl chloride corers (inner diameter 5 cm, length 10 cm), as previously described (Høj et al., 2005). The cores were stored for 4 months at 2–4 °C until the experiment was commenced.

Preparation and incubation of soil slurries

Soil from the inner part (5–7 cm depth) of three cores was mixed to create a pooled composite sample, which was subsequently used to make slurries. The peat slurries were made under N2 atmosphere by mixing 50 g wet peat (corresponding to 4.5 g dry peat) with 15 ml sterile, anoxic water in a 125 ml serum bottle, which was then sealed and capped. The slurries were gassed with N2 and incubated in the dark without shaking in precisely thermostated (±0.01 °C) water baths at temperatures 1, 5, 10, 15, 20 and 25 °C. Two parallel slurries were made for each temperature. One parallel was terminated in week 4 to perform analyses of the archaeal community structure, while the other parallel was terminated in week 13. Due to the labor intensity associated with sample processing, samples from no more than two temperatures were processed per day. All temperatures were sampled within 5–6 days, with the highest temperatures being sampled first.

Analytical techniques

Before sampling the headspace gas, bottles were shaken vigorously to allow equilibration between the liquid and gas phases. The methane concentrations in 1 ml of headspace gas was measured using a Shimadzu GC-14A gas chromatograph equipped with a Porapak Q stainless steel column (1.5 m × 1/8 inch) and a flame ionization detector, with N2 as carrier gas. Injection, detection and column temperatures were 200, 240 and 45 °C, respectively. The hydrogen concentration in 2 ml headspace gas was measured using a Varian Aerograph model 920 gas chromatograph equipped with a Stainless Steel Washed Molesieve (Alltech, Deerfield, IL, USA) 60/80 mesh column (2.0 m × 1/8 inch) and a thermal conductivity detector (125 mA), with Argon as carrier gas. Injection, detection and column temperatures were 100, 100 and 50 °C, respectively. Carbon dioxide concentration in 2 ml headspace gas was determined using an infrared CO2 analyzer (Binos 100, Rosemount GmbH & Co., Geschaftsbereich Analysentechnik, Hanau, Germany).

Samples of slurry water (1 ml) were stored frozen at −20 °C until analysis for volatile fatty acids by a Schimadzu GC 14A gas chromatograph equipped with a flame ionization detector, with N2 as carrier gas. Thawed samples were acidified by addition of 10 μl HCl so that C2–C5 volatile fatty acids could be separated as free acids on a 6 ft × 2 mm i.d. glass column packed with 10% SP-1200 and 1% H3PO4 on 80/100 Chromosorb W AW. Injection, detection and column temperatures were 200, 200 and 125 °C, respectively.

Calculations

Calculations of methane accumulation were based on the concentration in the gas phase, since the amount of dissolved methane can be assumed to be low (Wilhelm et al., 1977). The partial pressure was calculated based on the amount of methane in the headspace using the ideal gas law. The CO2 partial pressure in the headspace was calculated as described above for methane.

Total counts and FISH analysis

Soil samples (2.5 g) were fixed in 4% paraformaldehyde (10 ml) at 4 °C with shaking overnight. The fixed soil was transferred to a 250 ml Waring-blender steel jar (Waring, CT, USA) with 12.5 ml ice-cold 0.2 μm filtered water and homogenized 10 times for 1 min, with 5 min cooling between each blending. After the last blending, the homogenate was allowed to sediment for 1 min, and two parallels of 1 ml samples were taken with a 1 ml plastic syringe without needle. The maximum amount of sample that could be filtered while still keeping the linear relationship between PicoGreen counts and sample volume was determined for each sample (Fægri et al., 1977). Appropriate amounts of soil homogenate was then filtered onto white polycarbonate filters (pore size 0.2 μm, 25 mm diameter; Osmonics Poretics Products, Minnetonka, MN, USA), and air-dried filters were stored at −20 °C (Glöckner et al., 1999). For each sample, two parallel filters from each parallel were processed, giving a total of four filters analyzed per sample. The dry weight of soil samples was determined gravimetrically.

The DNA-specific stain PicoGreen (Molecular Probes, Invitrogen, Carlsbad, CA, USA) was chosen for determination of total counts of bacteria and archaea. This was based on initial experiments, which showed that PicoGreen was superior to 4′-6-diamidino-2-phenylindole (Porter and Feig, 1980) in discriminating between cells and soil particles, and caused less background on white polycarbonate filters than SYBRGreen I (Molecular Probes). Filter sections were stained with PicoGreen (1:100 dilution in 1 × Tris-acetate-EDTA) in the dark for 15 min (Weinbauer et al., 1998) and mounted with a drop of antifade solution (0.1% p-phenylenediamine, 50% glycerol, 0.5 × PBS) (Noble and Fuhrman, 1998).

The oligonucleotide probes used in this study were labeled with a CY3 fluorochrome on the 5′ end (Interactiva Biotechnologie, Ulm, Germany). The probes and used stringency conditions are presented in Table 1. For each probe the stringency conditions were optimized using control cultures of Escherichia coli, Archaeoglobus fulgidus, Methanococcus voltae, Methanosaeta concilii, Methanosarcina siciliae, Methanobacterium formicicum, Methanospirillum hungatei and Sulfolobus solfataricus. Each filter was cut into six sections prior to hybridization with the labeled oligonucleotide probes. All hybridizations were performed at 42 °C overnight. Otherwise the hybridization and washing steps were performed as outlined by Glöckner et al. (1999). The NaCl concentrations of the washing buffer was adjusted according to the stringency condition of each probe as described in Pernthaler et al. (2001). Filter sections were mounted with antifade solution as described above.

Table 1 Oligonucleotide probes used for FISH
Full size table

Filter sections were inspected using a Zeiss Axioplan epifluorescence microscope (Zeiss, Jena, Germany) equipped with a 50 W high pressure mercury bulb. The specific filter set Chroma HQ 41007 (Chroma Tech. Corp., Rockingham, VT, USA) was used for CY3-labeled probes and the Zeiss filter set 09 was used for PicoGreen. For each sample and probe combination, at least 10 fields of view were counted per filter, corresponding to a total of at least 40 fields of view. For determination of PicoGreen total counts at least 2500 cells were counted per sample. The statistical significance of differences in total counts and fractions of cells detected by individual FISH probes were analyzed using the nonparametric Mann–Whitney U-test (confidence level 0.05). Counts with FISH probes were regarded as statistically significant if they (on average) were significantly higher than the negative control probe counts (NON338, Mann–Whitney U-test, P>0.05).

DNA extraction from soil slurries

Samples for DNA extraction were frozen at −80 °C until processed. DNA extraction and purification was performed with the FastDNA SPIN Kit for soil (BIO 101/Qbiogene, Irvine, CA, USA) in combination with a freeze-thaw lysis protocol (Nakatsu et al., 2000; Høj et al., 2005). Wet peat (0.4 g) and 1 ml lysis buffer (0.12 M NaP (pH 8), 5% sodium dodecyl sulfate) were combined in a MULTIMIX 2 Tissue Matrix Tube provided in the kit. The tube was vortex mixed for 1 min, incubated at 65 °C for 45 min, followed by three quick freeze-thaw cycles (−80 and +65 °C). After this combined mechanical and freeze-thaw lysis cycle, the DNA was purified using the kit solutions as described by the manufacturer. Two parallel extractions were performed per sample, and the resulting eluates were mixed.

PCR amplification, cloning and sequencing

Two clone libraries were constructed using the DNA extracts from slurries incubated for 4 weeks at 5 and 20 °C, respectively. A portion of the 16S rRNA gene was amplified using the archaea-specific primers A109f and A934b (Grosskopf et al., 1998a). The optimal template concentration was found using 10-fold serial dilutions of the DNA extracts. The reaction mixture of 25 μl contained 1 μl, 0.2 μM of each primer, 200 μM of each deoxyribonucleotide triphosphate, 0.1% (w/v) bovine serum albumin, PCR buffer 1 × (Tris-Cl, KCl, (NH4)2SO4, 15 mM MgCl2, pH 8.7) (Qiagen, Hilden, Germany) and 1.25 U μl−1 DNA polymerase (Qiagen). Amplification was performed with the following protocol: initial melting at 94 °C for 3 min, followed by 32 cycles of 94 °C for 45 s, 52 °C for 1 min, 72 °C for 1 min, followed by a final extension at 72 °C for 7 min. PCR products were purified by gel extraction using the Qiaquick Gel Extraction Kit (Qiagen), following the protocols recommended by the manufacturer. Purified PCR products were cloned using a TOPO TA cloning kit, version K2 (Invitrogen), as recommended by the manufacturer. For each library, 50 clones were randomly selected, and the cloned insert was PCR amplified using the M13 primer supplied with the TOPO vector. Purified PCR products were sequenced using the ABI 3730XL Genetic Analyzer with A109f as sequencing primer.

Analysis of phylogeny and diversity of clone sequences

To identify potentially chimeric molecules the sequences were analyzed using the computer program Mallard version 1.02. (Ashelford et al., 2006). Sequences were initially aligned with 16S rRNA gene sequences of their closest relatives available in the GenBank database using the BLAST search facility of the National Center for Biotechnology Information (NCBI). Sequences derived from this study and their relatives with the highest BLAST scores were added to a prealigned database of 936 complete and partial (>1000 bp) archaeal 16S rRNA gene sequences using the aligning tools from the ARB program package (Technical University of Munich, Munich, Germany; http://www.arb-home.de). The aligned sequences were checked and corrected manually where required. Sequences were added to a preexisting maximum likelihood tree supplied with the ARB database using a special parsimony tool, and nodes not to be included in the tree were removed. Sequences with 97% or higher sequence identity were assigned to the same phylotype, and one representative sequence for each phylotype was included in the presented tree. Library coverage was calculated as (1−(a/b)) × 100, where a was the number of phylotype represented in the library, and b was the number of sequences in the library. Diversity indices were calculated based on the defined phylotypes using the computer program Estimates 8.0.0.

PCR amplification and DGGE analysis

A nested protocol was used for amplification of the 16S rRNA gene of archaea using the primers PRARCH112F and PREA1100R in the first reaction, and the primers PARCH340-GC and PARCH519R in a second touchdown reaction as previously described (Høj et al., 2005). DGGE and sequencing of bands was performed as described previously (Høj et al., 2005). Sequences were analyzed using the BLAST tool at the NCBI (http://www.ncbi.nlm.nih.gov/blast). Peak histograms for all profiles were determined using the computer program Gel2k (Svein Norland, Department of Biology, University of Bergen, Norway). Relative band intensities (Pi) were calculated for each profile as the relative peak area in the profile (Pi=ni/Ni, where ni is the area under peak i and Ni is the sum of all peak areas in the profile). For double bands, the relative band intensity was determined by relating their combined peak area to the sum of all peak areas in the profile. Potential correlations between relative band intensities and temperature were analyzed statistically using Pearson's product moment correlation (confidence level 0.05). The statistical significance of other trends in relative band intensities was analyzed using Student's t-test (confidence level 0.05).

Nucleotide sequence accession numbers

All unique sequences generated in this study have been deposited in the EMBL database under accession numbers AM712493–AM712556.

Results

Process-related data

Unamended slurries of Solvatnet peat showed increasing methane partial pressures with temperature (Figure 1). The length of the apparent lag phase before onset of methane accumulation and the period with almost exponential methane accumulation decreased with increasing temperature (Figure 1), as previously reported for slurries of rice soil (Fey et al., 2004) and temperate peat (van Hulzen et al., 1999). The ratio between CO2 and CH4 accumulating in the headspace decreased with increasing temperature and with time (Table 2), indicating that the contribution of methanogenesis to the overall organic matter decomposition increased with temperature and time.

Figure 1
figure 1

Methane accumulation in slurries of Solvatnet peat incubated at 1 °C (•), 5 °C (○), 10 °C (▾), 15 °C (▽), 20° (▪) and 25 °C (□). The accumulation curves for 1 and 5 °C could not be distinguished on this plot.

Full size image
Table 2 Ratio of accumulated CO2 to CH4 and accumulation of fatty acids in all slurries after 28, 51 and 83 days incubation
Full size table

At the lowest temperature (1 °C) no fatty acids were present at concentrations above the detection limits throughout the experiment (Table 2). At 5 °C, isobutyrate was the only fatty acid detected, accumulating to 1430 μM at the end of the experiment. Accumulation of this fatty acid decreased with increasing temperature. At 10 °C, isobutyrate was still the predominant fatty acid (up to 799 μM), but toward the end of the experiment there was also accumulation of acetate (680 μM) and propionate (95 μM) (Table 2). At higher temperatures (15 °C and higher), acetate and propionate accumulated transiently and maximum values were measured at day 51. Butyrate was detected transiently at low levels (<70 μM) in slurries incubated at 10–20 °C. No hydrogen was detected throughout the experiment, probably due to the low sensitivity of the method (detection limit 18 Pa).

Total counts and FISH-specific counts

The relative abundance of bacteria and archaea in soil slurries incubated for 4 weeks was examined by performing FISH reactions with a bacteria-specific probe (EUB338), an archaea-specific probe (ARCH915) and three probes specific for subsets of the domain Archaea (Table 1). FISH revealed little variation in the percentage of cells binding the EUB338 probe between slurries incubated at temperatures of 5 °C and higher (40–50%), but a lower percentage of cells was detected at 1 °C (33%) (Mann–Whitney U-test, P<0.05) (Table 3). A relatively high proportion of archaea (11–12% of total count) was detected for the two lowest temperatures (1 °C and 5 °C). In the 10 °C slurry, this percentage was significantly lower (1%) (Mann–Whitney U-test, P<0.05), while at temperatures of 15 °C and higher 3–6% of the microbial cells were identified as archaea (Table 3). The probe EURY498 hybridized with 1–3% of the microbial cells in all slurries (Table 3). For some slurries the counts with this probe were at or below (10 °C) the detection limit of the method and some counts had large standard deviations (Table 3). Hence, differences in the percentage of EURY498-positive cells were in most cases not statistically significant. The fraction of archaea binding the EURY498 probe was however shown to be higher at 20 °C (52%) than at 5 °C (19%) (Mann–Whitney U-test, P<0.05).

Table 3 Total counts of bacteria and archaea (PicoGreen-stained cells) and percentages of cells detected with FISH probes EUB338, ARCH915 and EURY498
Full size table

The morphology of cells detected by the probes ARCH915 and EURY498 differed with temperature. Rods dominated the community at 1 °C and 5 °C, while cocci, sarcina-like clusters and highly organized spheres dominated at higher temperatures. The number of cells binding the group-specific methanogen probes MSMX860 (Methanosarcinales) and MB310 (Methanobacteriaceae) (Raskin et al., 1994) was below the detection limit of the method. Nevertheless, by examining a large number of fields of view, a few positive cells were identified in slurries incubated at temperatures of 10 °C or higher. The probe MB310 detected rod-shaped cells, and the probe MSMX860 detected sarcina-like and sphere-formed cell aggregates. The size and the probe signal intensity of the aggregates were lower at 10 °C than for the higher temperatures.

Clone libraries

A total of 100 archaea-specific 16S rRNA clones were analyzed from slurries incubated at 5 °C and 20 °C for 4 weeks (50 for each). The computer program Mallard identified one chimeric sequence in the 20 °C library, which was removed from further analyses. In the subsequent statistical analyses clones representing known methanogenic groups and other archaea were separated since they displayed contrasting trends with temperature (Table 4). For simplicity the two library subsets are referred to as the methanogenic subset and the nonmethanogenic subset, although the second subset can theoretically include organisms with a hitherto unrecognized capability of methane production.

Table 4 Diversity statistics for clone libraries (50 clones each) of slurries incubated at 5 °C and 20 °C
Full size table

The fraction of clones representing known methanogenic groups differed between the libraries (28% at 5 °C and 54% at 20 °C) (Figure 2). The methanogenic subset had a relatively low diversity (H′ 0.88) and species richness (Chao1 3.0) in the 5 °C library as compared to the 20 °C library (H′ 1.83, Chao1 11.0) (Table 4). Hence, the coverage of the methanogenic subset was higher in the 5 °C library (78.6%) than in the 20 °C library (69.2%) despite a lower number of clones in this subset (Table 4). In contrast, the nonmethanogenic subset had highest diversity in the 5 °C library (H′ 2.29 at 5 °C, H′ 1.68 20 °C) (Table 4). The rarefaction curve for the nonmethanogenic subset of the 5 °C library did not approach its asymptote showing that further sequencing of clones would have revealed additional diversity (data not shown). For both temperatures the estimated species richness was higher for the nonmethanogenic than the methanogenic library subset (Table 4).

Figure 2
figure 2

Archaeal community structure in 16S rRNA clone libraries derived from peat slurries incubated at (a) 5 °C and (b) 20 °C for 4 weeks. The community structure is represented by the relative abundance of clone sequences belonging to different phylogenetic clusters. Abbreviations: Mmi, Methanomicrobiales; Msar, Methanosarcina; Msae, Methanosaeta; Mb, Methanobacterium; RC-V, rice cluster V; LDS, Lake Dagow sediment; Cr1.3b, Group 1.3b Crenarchaeota.

Full size image

Sequences not affiliated with known methanogenic groups generally clustered within three clusters: Group 1.3b Crenarchaeota (Ochsenreiter et al., 2003) and the euryarchaeal clusters rice cluster V (RC-V, Grosskopf et al., 1998b) and Lake Dagow sediment (LDS, Glissmann et al., 2004) (Figure 3). An exception was seen for phylotype Sv-37, which was not closely related to any sequences in GenBank but was distantly related to clusters RC-V and LDS (Figure 3). In general, the sequences were closely related to sequences previously retrieved from a particle-rich Arctic river (Galand et al., 2006) (Figure 3). The same nonmethanogenic archaeal groups have however been detected also in other wet and cold soils (Kemnitz et al., 2004; Kotsyurbenko et al., 2004; Metje and Frenzel, 2005; Høj et al., 2006). While Group 1.3b Crenarchaeota was represented by two closely related phylotypes, clusters LDS and RC-V were represented by 10 and 7 phylotypes, respectively, of which most were represented by only 1–3 sequences (Figure 3). For all of these clusters their relative abundance was higher in the 5 °C than in the 20 °C library (Figure 2). Sequences in the methanogenic subsets were affiliated with the order Methanomicrobiales and the genera Methanobacterium, Methanosarcina and Methanosaeta (Figures 2 and 3). In general, the sequences were closely related to sequences previously retrieved from Siberian tundra (Metje and Frenzel, 2005, 2007; Ganzert et al., 2007) and Spitsbergen wetlands (Høj et al., 2005, 2006). Methanomicrobiales was the numerically dominant order in both libraries, constituting 57–58% of the methanogenic clones in each library (Figure 2). One of five Methanomicrobiales phylotypes (Sv-78) was affiliated with the family Methanospirillaceae, while the others did not cluster with any recognized family (Figure 3). Temperature affected the Methanomicrobiales phylotypes that were present; in the 5 °C library all Methanomicrobiales-affiliated sequences belonged to phylotype Sv-19, while in the 20 °C library only 1 of 15 Methanomicrobiales sequences belonged to this phylotype (Figure 3). The genera Methanobacterium, Methanosarcina and Methanosaeta were each represented by a single phylotype (Figure 3). While the Methanobacterium phylotype constituted a larger fraction of clones in the 5 °C library than in the 20 °C library, the opposite trend was seen for the Methanosarcina and the Methanosaeta phylotypes (Figure 2).

Figure 3
figure 3

Maximum likelihood tree showing the phylogenetic affiliation of sequences recovered from clone libraries of slurries incubated at 5 and 20 °C and from denaturing gradient gel electrophoresis (DGGE) profiles of slurries and field samples from the same site (Solvatnet) (Høj et al, 2005). Each phylotype is represented by one sequence, and the number of unique sequences in the 5 and the 20 °C libraries, respectively, is indicated in parenthesis for each phylotype.

Full size image

PCR-DGGE analysis of archaea

PCR-DGGE analysis was used to analyze and compare the archaeal communities in slurries incubated at 1, 5, 10, 15, 20 and 25 °C for 4 weeks and again after 13 weeks (Figure 4). The primers used for this analysis had mismatches with most known sequences in clusters RC-V and LDS, so the method mainly provided information on temporal changes in methanogenic populations. The DGGE profiles were similar or identical to profiles retrieved from field samples from the same site (Solvatnet) and sequences recovered from excised DGGE bands (148–152 bp) were identical to previously recovered sequences (Høj et al., 2005). As discussed previously, double bands were present in the profiles, which could not be separated by repeated excisions and reamplification and yielded pure sequences (Høj et al., 2005). All sequences recovered from DGGE bands were represented also in the clone libraries. Banding patterns were relatively stable with temperature, but some bands showed systematic changes in relative intensity with temperature and time (Figure 4 and Table 5). While band 4 (Msae) showed increasing relative intensity with increasing temperature at both time points, the opposite trend was seen for band 5 (Mmi4), which had decreasing relative intensity with temperature at both time points (Figure 4 and Table 5). Band 2 (Mmi3) and band 6 (Msar1) showed more complicated responses to temperature. After 4 weeks both bands showed positive correlations with increasing temperature (up to 25 °C for Msar1 and up to 15 °C for Mmi3) (Figure 4 and Table 5). In contrast, there was no longer a direct correlation with temperature for these bands (P>0.05) after 13 weeks but their relative intensity was higher at low (1–10 °C) as compared to high (15–25 °C) temperatures (Student's t-test, P<0.05, Figure 4).

Figure 4
figure 4

DGGE profiles of archaeal 16S rRNA genes amplified from Solvatnet peat slurries in weeks 4 and 13. Bands marked in the gel were successfully sequenced. The bar on the right shows a schematic diagram of the identified DGGE bands, with double bands indicated. Double bands could not be separated by repeated excisions and reamplification, and they yielded pure sequences. A sequence was successfully obtained for the band labeled * in our previous study (Mmi1) (Høj et al., 2005) though a sequence was not recovered in the present study. The band labels are consistent with this previous study of field samples from Solvatnet (Cr, Crenarchaeota, Mmi1, Mmi3, Mmi4, Methanomicrobiales; Mb, Methanobacterium; Msae, Methanosaeta; Msar1, Methanosarcina).

Full size image
Table 5 Correlations with temperature for relative DGGE band intensities
Full size table

Discussion

This study showed that temperature affected both the diversity and structure of archaeal communities in unamended slurries of high Arctic peat from Solvatnet (Spitsbergen), with contrasting effects seen for known methanogenic groups and other archaea. A polyphasic approach demonstrated that after 4 weeks of incubation, 10 °C was an important temperature for the selection of different archaeal populations. At lower temperatures the archaeal community was dominated by nonmethanogenic groups, while methanogenic groups showed increased diversity and relative abundance at higher temperatures. This was in accordance with increasing methane production rates with increasing temperature.

After incubation for 4 weeks at 1 °C and 5 °C the archaeal community in Solvatnet peat was relatively abundant, representing 11–12% of the total count, or 20–25% of all cells detected by FISH. This high abundance of archaea at low temperatures is in accordance with previous field studies of cold environments such as Siberian tundra soil (Kobabe et al. 2004), and marine plankton in Antarctic waters (Murray et al, 1998) and the Beaufort Sea (Wells and Deming, 2003; Garneau et al., 2006). FISH demonstrated that at these temperatures most of the archaeal cells had rod morphology, while clone libraries showed that at 5 °C the community was dominated by Group 1.3b Crenarchaeaota and the euryarchaeal clusters RC-V and LDS. While previous studies have not specifically targeted the ecophysiology of these groups, it has been shown that Group 1.3b Crenarchaeota is relatively abundant in permanently flooded or wet soils from Spitsbergen (Høj et al., 2006), and that in temperate soils this groups is relatively more abundant during winter (Kemnitz et al., 2004). Since the absolute abundance of archaea dropped significantly at 10 °C, this study indicated that some archaeal populations in Solvatnet peat had a competitive advantage in slurries incubated at 1 and 5 °C for 4 weeks. It is not known which physiological attributes that renders this advantage, as populations that are well adapted to low temperatures do not necessarily have a low optimum temperature for growth (Cavicchioli, 2006). It is intriguing however that a similar drop in the number of archaea at 10 °C was reported also for acidic peat from Northern Finland (Metje and Frenzel, 2005), suggesting that 10 °C may be an important temperature for selection of different archaeal populations in Arctic soils.

The methanogenic community in slurries of Solvatnet peat was diverse and included members of the order Methanomicrobiales, and the genera Methanobacterium, Methanosarcina and Methanosaeta, in accordance with field studies from the same site (Høj et al., 2005). Group-specific FISH probes identified a few methanogenic archaea in slurries incubated at 10 °C and above. The signal intensity from positive cells and the size of cell aggregates were generally lower at 10 °C than at higher temperatures, in accordance with relatively low methane production rates in this slurry (<0.2 μmol CH4 (g dw)−1 per day). Although the overall diversity and relative abundance of methanogenic archaea increased with temperature, individual methanogenic populations showed differences in their temperature response. While some populations showed little or no response to temperature, other populations showed systematic changes in their relative abundance with temperature and time. This was reflected in clone libraries constructed for slurries incubated at 5 °C and 20 °C for 4 weeks, but was more clearly showed by trends in the DGGE profiles obtained for all slurries (1–25 °C) at two time points (4 and 13 weeks). Since the same inoculum was used for all slurries and all samples were processed in parallel, it could be assumed that changes in relative intensities of DGGE bands were robust. In addition, for samples where both DGGE profiles and clone library data were available, the two methods, which depended on different primer sets, generally showed consistent trends.

The Methanosaeta population Sv-95 (Msae) increased in relative abundance with increasing temperature at both time points, in accordance with a previous study of unamended rice soil slurries (Fey and Conrad, 2000). In contrast, one of the Methanomicrobiales populations (Sv-19, Mmi4) had an apparent competitive advantage at low temperatures, suggesting that it was well adapted to the conditions prevailing in slurries incubated at low temperatures. This trend was supported both by differences in clone frequencies in the two clone libraries and by DGGE profiles at both time points. Again, it is unknown which traits that give this population a competitive advantage at low temperature as proteomics studies of the cold-adapted methanogen Methanococcoides burtonii have identified a range of cellular processes that are important for cold adaptation (Cavicchioli, 2006). DGGE profiles after 13 weeks incubation showed that the Methanomicrobiales population Sv-64 (Mmi3) and the Methanosarcina population Sv-60 (Msar1) could also grow at temperatures as low as 1 °C despite having relatively low abundance after 4 weeks, and after 13 weeks their relative abundance was highest at the temperatures (1–10 °C) (Student's t-test, P<0.05). The temporal change in their temperature response suggested that their relative abundance was controlled indirectly rather than directly by temperature. Since the transient accumulation of at least two fatty acids (acetate and propionate) was also delayed at low temperatures, it can be hypothesized that the observed indirect effect of temperature could be related to the availability of substrates or growth factors. This study was however not designed to elucidate the exact mechanism behind indirect temperature effects, and no direct link to a specific fatty acid could be demonstrated.

The results from this study are in accordance with previous studies of rice soil slurries (Chin et al., 1999a, 1999b; Fey and Conrad, 2000), which also showed changes in the archaeal community structure with temperature. In contrast, no significant temperature-induced changes in the archaeal community structure were observed in acidic peat from Northern Finland (Metje and Frenzel, 2005) or from Siberia (Metje and Frenzel, 2007), where Methanobacteriales only, or Methanobacteriales and Methanosarcinales, respectively, were predominant at all incubation temperatures. We note that the stimulating effect of temperature on methane production rates was higher for Solvatnet peat than reported for the acidic peat from Northern Finland (Metje and Frenzel, 2005) and from Siberia (Metje and Frenzel, 2007), but it is unknown whether this can be linked to differences in the temperature characteristics of the methanogenic communities. So far, no studies have been able to link methane emissions or methane production rates to the archaeal diversity and community composition (Høj et al., 2005; Ganzert et al., 2007).

In conclusion, this study demonstrated contrasting effects of temperature on known methanogenic archaea and other archaea present in peat from the high Arctic wetland Solvatnet, Spitsbergen. The diversity and relative abundance of nonmethanogenic archaeal groups were highest at low temperature, and at least some members of Group 1.3b Crenarchaeota, and the euryarchaeal clusters RC-V and LDS appeared to have a competitive advantage at low temperatures. The overall diversity and relative abundance of methanogenic archaea increased with increasing temperature, in consistency with a strong stimulation of methane production rates. Nevertheless, individual methanogenic populations showed differences in their temperature response. While the relative abundance of some populations showed the same temperature trend (increase or decrease) after both 4 and 13 weeks, the relative abundance of other populations was controlled by indirect effects or did not respond to temperature.