• Arctic-alpine tundra;
  • microbial communities;
  • terminal restriction fragment length;
  • phospholipid fatty acid;
  • 16S rRNA gene;
  • Acidobacteria


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The seasonal and spatial variations of microbial communities in Arctic fjelds of Finnish Lapland were studied. Phospholipid fatty acid analysis (PLFA) and terminal restriction fragment analysis (T-RFLP) of amplified 16S rRNA genes were used to assess the effect of soil conditions and vegetation on microbial community structures along different altitudes of two fjelds, Saana and Jehkas. Terminal restriction fragments were additionally analysed from c. 160 cloned sequences and isolated bacterial strains and matched with those of soil DNA samples. T-RFLP and PLFA analyses indicated relatively similar microbial communities at various altitudes and under different vegetation of the two fjelds. However, soil pH had a major influence on microbial community composition. Members of the phylum Acidobacteria dominated especially in the low pH soils (pH 4.6–5.2), but above pH 5.5, the relative amount of terminal restriction fragments corresponding to acidobacterial clones was substantially lower. Both T-RFLP and PLFA analysis indicated stable microbial communities as the DNA and fatty acid profiles were similar in spring and late summer samples sampled over 3 years. These results indicate that differences in microbial community composition could be explained primarily by variation in the bedrock materials that cause variation in the soil pH.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Finnish Lapland is mainly situated in the taiga (boreal forest) zone although Arctic-alpine tundra (treeless heaths) dominate the environment in the far northwestern part of Lapland. Arctic tundra soils are characterized by extreme conditions such as long periods of low temperatures, multiple episodes of freezing and thawing during spring and fall and relatively wide annual temperature variation. It is often assumed that below 0°C microbial activity ceases, and in boreal and Arctic environments the soil microbial activity is minimal outside the growing season. However, microbial activity has been reported in soil at subzero temperatures (Clein & Schimel, 1995; Mikan et al., 2002) and even at temperatures down to−20°C (Rivkina et al., 2000). Larsen et al. (2002) reported substantial carbon mineralization in soils of northern Sweden during the winter months as compared to the annual carbon flux. In tundra and taiga soils it was estimated that microbial respiration during winter accounted for 10–30% of the annual carbon loss (Clein & Schimel, 1995).

Global warming is expected to affect the Arctic ecosystems severely as both increased air temperature and increased snow accumulation will lead to significant warming of the soil and subsequently increased microbial activity and CO2 release from the Arctic tundra (Welker et al., 2000). Microbial activity of Arctic tundra soils has consequently received keen interest during the past decade. Studies of cold soils have, however, concentrated mainly on analyzing microbial respiration, carbon and nitrogen mineralization, while the structures of microbial communities are poorly characterized. Potential applications for cold-active microorganisms and microbial by-products have also been a strong motivation for the exploration of extremely cold environments and numerous novel bacterial species have been isolated and described from the polar environments (for a review, see e.g. Cavicchioli et al., 2002). Microbial community structures of alpine tundra soils have been studied in the Niwot Ridge, Colorado (Lipson et al., 2002; Schadt et al., 2003; Lipson & Schmidt, 2004; Monson et al., 2006) and a few reports exist on the microbial diversity of high Arctic tundra in Siberia (Zhou et al., 1997) and Canada (Neufeld et al., 2004). Amplified 16S rRNA genes of alpine soils of the Niwot Ridge were dominated by Acidobacteria, Alpha- and Betaproteobacteria, members of the Verrucomicrobia division and Bacteroidetes (Lipson & Schmidt, 2004). Siberian permafrost tundra was dominated by Alpha- and Gammaproteobacteria (Zhou et al., 1997) while Deltaproteobacteria were reported to dominate in the high Arctic of Canada (Neufeld et al., 2004). Very little is known, however, on the microbial community structures of sub-Arctic and Arctic-alpine environments of the Fennoscandia, where the microorganisms are challenged by not only low temperatures but also repeated freeze–thaw cycles and a relatively wide annual temperature fluctuation. Fennoscandian tundra environments extend from the mountainous (fjeld) areas of southern Norway to the far northernmost parts of Finland. In Finland, the Scandinavian Caledonides extend to the Kilpisjärvi region (Fig. 1), which is considered one of the coldest places in Continental Europe with an annual temperature of −2.2°C and a growing season of c. 100 days. The Fennoscandian tundra differs from the arctic tundra of Canada and Siberia by the absence of permafrost. Consequently these soils experience a larger annual temperature variation as well as variation in the soil water content when there is no permafrost to trap the moisture (Richardson et al., 2003). Moreover, the Fennoscandian tundra differs from the more temperate alpine soils by e.g. a shorter plant growth season and differences in the photoperiods.


Figure 1.  Map of northern Finland showing the location of Kilpisjärvi.

Download figure to PowerPoint

The aim of this study was to investigate the major bacterial groups dominating in the Arctic-alpine tundra soils of Lapland. Particular interest was paid to the spatial and seasonal stability of these microbial communities. Therefore, the variability of the microbial community of tundra soils was studied in two fjelds with two culture-independent community profiling methods, phospholipid fatty acid analysis (PLFA) and terminal restriction fragment analysis (T-RFLP), with the aim of assessing the effect of altitude, vegetation and soil conditions, as well as the seasonal variation in the microbial community composition. By comparing the data obtained from T-RFLP analyses from soil DNA, cultivated isolates and a 16S rRNA gene clone library, it was possible to link the community profiles to a more detailed phylogenetic analysis of the bacterial community.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study site and soil sampling

The fjelds located in the Kilpisjärvi region (69°01′N, 20°50′E) in the northwestern part of Finland (Fig. 1) belong to the Scandinavian Caledonides. This part of Lapland is unique in Finland, with peaks reaching 1000 m or higher, including Saana, Malla and Jehkas fjelds. The bedrock differs from other parts of Finland as it contains calcareous dolomite, which makes the vegetation of the northwestern-most Finland exceptionally lush. The area is characterized by mountain birch (Betula pubescens ssp. czerepanovii) forests which reach up to c. 600 m above sea level (a.s.l.) and treeless tundra heaths with a mosaic vegetation of dwarf shrubs and alpine grasses. The climate in Kilpisjärvi is suboceanic due to the proximity of the Arctic Ocean. The growing season is c. 100 days with a mean annual temperature of −2.2°C. The annual precipitation is c. 500 mm and snow accumulation in the town of Kilpisjärvi c. 1 m in March. On the fjeld slopes, snow accumulation varies within short distances depending on wind exposure and topography. The wind-exposed ridges may remain essentially snow free, while depressions can receive more than 1 m of snow. Under the thick snow pack, the soil temperature remains close to 0°C all winter, while temperatures down to −25°C were measured in the wind-exposed soils during the winters of 2004–2005 and 2005–2006 (M. Männistö, unpublished data).

For T-RFLP and PLFA analysis, soil was sampled from Saana and Jehkas fjelds from 2001 to 2003 from different sites varying in vegetation and soil conditions (Table 1). For bacterial isolation, soil was sampled additionally from Pikku-Malla fjeld in 2001 and 2002. All three fjelds are located within 5 km of each other in close proximity to the town of Kilpisjärvi (Fig. 1). Microbial community structures of Saana fjeld were followed from soil samples obtained in August 2001, June 2002, September 2002 and June 2003, whereas soil from Jehkas was sampled only once in June 2003. In 2001, soil samples were obtained from five altitudes (Table 1) ranging from the mountain birch treeline (600 m a.s.l.) to top of the fjeld (1020 m a.s.l.). From each altitude, four soil cores were obtained from the top 5 cm of soil. These samples were used to test DNA and PLFA extraction methods as well as to obtain preliminary data of the microbial community composition at different elevations. In 2002–2003, soil was sampled from eight sites of Saana and nine sites of Jehkas, as indicated in Table 1. The 2002–2003 samples were pooled by combining 10 soil cores to two composite samples of each sampling site. The samples were transported to the laboratory where they were homogenized by hand. Plant debris, large roots and stones were removed and the sample was mixed and stored at −20°C for DNA, PLFA and soil analysis. For cultivation of bacteria, soil was stored at close to ambient temperature (4°C) for no more than 4 weeks.

Table 1.   Sampling sites and physico-chemical characteristics of soil samples from Saana and Jehkas fjelds
Sampling site*Altitude (m a.s.l)pHTotal N (mg g−1)NH4-N (mg g−1)Other
  • *

    S and J in the sampling site codes refer to Saana and Jehkas fjelds, respectively.

  • S01 and S02 refer to Saana samples obtained in years 2001 and 2002–2003, respectively.

  • Total-N and NH4-N concentrations are given as mg g−1 soil dry weight.

S01_110204.9 (± 0.03)     
S01_39605.2 (± 0.3)     
S01_48004.8 (± 0.29     
S01_56704.9 (± 0.1)     
S01_66006.1 (± 0.2)     
S02_110205.1 (± 0.1)
S02_210205.0 (± 0.3)23.732.88.211.7Grass-dominated
S02_39505.1 (± 0.3)18.514.34.52.6 
S02_48504.9 (± 0.2)8.932.92.310.8 
S02_57505.7 (± 0.1)16.423.111.613.4Grass-dominated
S02_67505.0 (± 0.3)
S02_76504.8 (± 0.3) 
S02_86204.7 (± 0.1)37.354.413.611.4 
J1/20039605.1 (± 0.05)ND17.8ND2.7 
J2/20039406.2 (± 0.04)ND37.9ND21.6Grass-dominated
J3/20039405.2 (± 0.05)ND10.3ND2.5Shrub-dominated
J4/20038555.2 (± 0.05)ND22.4ND14.0 
J5/20037905.5 (± 0.04)ND10.0ND1.8 
J6/20037555.8 (± 0.05)ND20.5ND7.6 
J7/20036505.2 (± 0.05)ND39.0ND11.2 
J8/20036204.8 (± 0.05)ND13.2ND1.1 
J9/20035804.6 (± 0.03)ND15.3ND1.3 

Soil pH was measured in 25% (v/v) soil–water suspension. Total and ammonium nitrogen were measured from 0.5 M K2SO4 extracts. For the extracts a subsample of c. 2 g fresh soil was extracted with 50 mL of 0.5 M K2SO4 and the NH4-N concentration in the extracts was determined by flow injection analysis (FIA 5012, Tecator). Total extractable N in the extracts was determined by oxidizing all extractable N to NO3 (Williams et al., 1995), and then analyzing it as NO3 by automated flow injection (FIA 5012, Tecator).

Phospholipid fatty acid analysis

Phospholipids were extracted from 5 g of homogenized soil as described by Ruess et al. (2005). Soil samples were extracted with 15 mL of a one-phase mixture (1 : 2 : 0.8 v/v/v) of chloroform, methanol and 0.05 M sodiumphosphate buffer (pH 7.4) overnight. The extraction was repeated with 7.5 mL extraction solvent for 3 h and 2.4 mL each of H2O and methanol was added to the solvent phase. After overnight separation at 4°C, the chloroform fraction was removed and phospholipids were fractionated with silicic acid columns (Ruess et al., 2005). Fatty acids were saponified and methylated following the MIDI protocol (Miller & Berger, 1985) and the fatty acid methyl esters analyzed as described earlier (Männistö & Häggblom, 2006). Fatty acids are designated by the total number of carbon atoms : number of double bonds following the position (when known) and geometry (cis/trans) of the double bond from the aliphatic (ω) end of the molecule. Prefixes i, a and cyc refer to iso- and anteiso branched fatty acids and cyclopropane fatty acids, respectively. 10- and 11-Me refer to a methyl branch in the 10th or 11th carbon from the carboxyl group.

Molecular analyses of soil samples

In 2001 soil DNA was extracted from 0.3 g (wet weight) of individual homogenized soil cores using two methods: the method developed by Zhou et al. (1996) and the Mobio UltracleanTM soil DNA extraction kit (MoBio Laboratories, Inc., Carlsbad, CA). DNA extracted by the Zhou method was purified using Sepharose 4B filled columns as described by Jackson et al. (1997). As both methods gave comparable T-RFLP patterns, only the Mobio Ultraclean™ soil DNA extraction kit was used for the subsequent samples. PCR amplification of 16S rRNA genes was performed using forward primer 27f and reverse primer 1522r (Johnson, 1994). PCR was carried out in a final volume of 50 μL using 0.5 μL template DNA, 200 μM of each dNTP, 0.3 μM of each primer and 1 U DyNAzyme II polymerase with 1 × concentration of the supplied buffer (Finnzymes, Espoo, Finland). PCR amplification was performed in a PTC-100 thermal cycler (MJ Research, Waltham, Mass.) as follows: initial denaturation at 95°C for 5 min, 30 cycles at 94°C for 0.5 min, 55°C for 1 min and 72°C for 2 min, and the final extension step at 72°C for 10 min. For T-RFLP analysis the forward primer was labeled with FAM (6-carboxyfluorescein). Three PCR amplification products from each DNA preparation were pooled and purified using a High Pure PCR product purification kit (Roche). Approximately 100 ng of the purified PCR product was digested with restriction enzymes MspI (Fermentas) and HhaI (Roche). From the 2001 samples, the T-RFLP of each soil DNA extract was analyzed, but from the 2002–2003 samples, the restriction digestions of DNA extracts of three soil samples were pooled for T-RFLP analysis. Consequently, each T-RFLP profile is a mixture of restriction digestions of three replicate DNA samples. T-RFLP profiles were analyzed by the Fragment analysis service of University of Turku using an ABI PRISM 377 automatic sequencer and GeneScan -500 TAMRA (Applied Biosystems) DNA fragment length standard.

Identification of dominating soil bacteria

A total of c. 200 bacterial strains were isolated from the tundra soil of Kilpisjärvi, of which 150 were reported earlier (Männistö & Häggblom, 2006). The 50 strains isolated in this study were cultivated from soil sampled from the north ridge of Saana using 20% R2A agar, polysaccharide agar, sugar agar and extract agar. R2A agar was prepared with 3.6 g of R2A per liter (20% of the recommended strength, Difco) and 6 g of agar per liter (Difco). Polysaccharide agar contained: 0.5 g each of carboxy methylcellulose, xylan and starch per liter and 0.3 g of yeast extract per liter prepared in 1/4 strength DMS465 mineral salts solution (pH 6.0, Sugar agar contained: 0.5 mM each of glucose, xylose, lactose, and cellobiose prepared in 1/4 strength DSM465 mineral salt solution (pH 6.0), 0.2% soil extract (Männistö & Häggblom, 2006) and 15 g agar per liter. The extract agar contained: 50 mL of each soil, moss and plant extract per liter amended with proteose peptone (0.3 g L−1), yeast extract (0.3 g L−1), glucose (0.2 g L−1), Tween80 (0.02% v/v) and agar (15 g L−1). Soil and moss extracts were prepared as described earlier (Männistö & Häggblom, 2006). Plant extract was prepared from 500 g of plant material (shrubs and grasses) that was collected from the Kilpisjärvi tundra. The leaves and grasses were simmered for 30 min in 1000 mL of tap water. After cooling, the plant material was removed by filtering the extract through cotton and the extract was sterilized by autoclaving. The pH of the extract agar was c. 5.5.

The bacterial isolates were identified by their whole cell fatty acid compositions and partial 16S rRNA gene sequences as described earlier (Männistö & Häggblom, 2006). Based on these results, 60 different isolates were selected for T-RFLP analysis. T-RFLP analysis of the bacterial strains was performed from 2 to 5 ng of DNA as described for the soil samples.

For cultivation-independent analyses the 16S rRNA gene PCR product amplified from the Saana soil DNA sample (S02_1) was inserted into a pCR4-TOPO vector and transformed to Escherichia coli TOP10 cells using TOPO TA Cloning Kit for Sequencing (Invitrogen). Putative positive colonies were selected by blue/white screening and plasmid DNA was extracted using the Qiaprep spin miniprep kit (Qiagen).

T-RFLP sizes of 100 clones were analyzed from 1 to 2 ng of PCR products that were amplified and digested using the same protocol as described for the soil DNA samples. Based on the T-RFLP analyses, two clones representing 13 of the major peaks in the soil T-RFLP profiles were selected for partial sequencing of the 16S rRNA genes using primer 27f on an ABI3700 DNA sequenzer. Similarity searches of sequences were performed using blast (Altschul et al. 1997). T-RFLP fragments of clones and bacterial isolates corresponding to the 13 soil T-RFLP peaks were analyzed from 1 to 2 ng of PCR products similarly as described for the soil samples. For each T-RFLP peak, between one and six corresponding sequences were obtained.

Data analysis

The T-RFLP data of each sampling period (August 2001, June 2002, September 2002 and June 2003) was analyzed separately. TRFs of 70–500 bp in length and with heights of >50 fluorescent units were included in the final analysis. TRF peaks that differed by more than 0.5 bases were considered unique. T-RFLP data was normalized to the total fluorescence of each sample, as described by Dunbar et al. (2001). PLFA peaks that represented at least 0.5% of total peak areas were considered for the analysis. The mole fraction of each fatty acid was used for the PLFA patterns.

The PLFA and T-RFLP patterns of various soil samples were compared using nonmetric multidimensional scaling (NMS) using the program package pc-ord (McCune & Mefford, 1999). NMS ordination was used as it is well suited for data that is nonnormal or arbitrary, discontinuous or with otherwise questionable scales and does not assume linear relationships among variables (McCune & Grace, 2002). NMS ordination was calculated using the Sorensen (same as Bray-Curtis) distance method and autopilot mode on medium speed. The graphical presentation of the NMS ordination was used to find the most (dis)similar sample points and to visualize any relationships between environmental conditions and PLFA/T-RFLP compositions. PLFA and T-RFLP patterns were correlated to the soil physicochemical properties using Pearson correlation. For this the ordination scores (component 1 and 2) of each sample was correlated with soil total-nitrogen, ammonium, pH, organic matter content and moisture using bivariate correlation with Pearson correlation coefficient and two-tailed test of significance on the spss 14.0 program package. The effect of DNA extraction method and sample altitude on T-RFLP ordination scores of the year 2001 samples was tested using two-way anova followed by a posthoc Tukey's test.

Accession numbers

The partial 16S rRNA gene sequences have been submitted to GenBank under accession numbers AM397026AM397057.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Assessment of soil microbial community structures

We analyzed the microbial community structure of arctic fjeld soils using culture-independent methods to assess seasonal variation and the effects of altitude, vegetation and soil conditions on community dynamics. In addition, the TRF sizes of isolated bacterial strains and cloned sequences were compared to the soil T-RFLP profiles as an attempt to identify dominating bacterial species in the tundra soils. Two fjelds in the Kilpisjärvi region were compared, Saana and Jehkas, with soil samples taken from different vegetation regions from the treeless tundra heaths to the mountain birch forest.

The Zhou DNA extraction method was comparable to the Mobio UltracleanTMsoil DNA extraction kit as the T-RFLP profiles were not significantly affected by the DNA extraction method. This is indicated in Fig. 2, which shows the NMS ordination of the T-RFLP profiles (HhaI digestion) of Saana soil sampled in 2001. A similar trend was observed when using MspI for digestion (data not shown). The Mobio soil kit was used for all subsequent experiments. T-TRFLP profiles indicated relatively similar microbial communities above the tree line that differed substantially from the microbial community of the higher pH soils in the mountain birch zone. NMS ordination grouped samples S01_1-S01_5, which were from 670 to 1020 m, relatively close together while T-RFLP samples of S01_6 (600 m) separated from the higher elevations (Fig 2). Two-way anova of the NMS ordination scores confirmed that elevation significantly affected T-RFLP profiles. The T-RFLP profile of the sample S01_6 (600 m) differed substantially from all the higher elevations in respect to many terminal restriction fragments (TRF) which grouped it apart all other samples. Furthermore, Tukey's test indicated that sample S01_6 was significantly different (P<0.01) from all other samples while sample S01_5 differed from S3 to S6. Apart from the altitude, the vegetation was substantially different at site S6 compared to all other sampling sites. The site was located in the mountain birch zone and the ground vegetation was notably richer than in the higher sites which presumably had a strong influence in the soil microbial communities. In addition, the soil pH of S01_6 (6.0) was notably higher than that of the other soil samples (4.7–5.3; Table 1), reflecting the high carbonate content of the soil.


Figure 2.  NMS ordination of T-RFLP data (HhaI digestion) of soil collected from the treeline (S01_6) to the top (S01_1) of Saana fjeld during August 2001. See Table 1 for sample codes. Letter M after the sample code refers to DNA extracted with the Mobio Ultraclean™ soil DNA extraction kit whereas letter Z refers to DNA extracted by the Zhou method.

Download figure to PowerPoint

In 2002 and 2003, soil samples were obtained from eight sites of Saana and nine sites of Jehkas fjelds (Table 1). Sites S02_1-S02_4 were located on the relatively sheltered summit plateau of Saana, which receives about 0.5–1 m of snow during the winter and becomes snow-free generally in early June. Sites S02_5and S02_6 were from a more wind-exposed ridge at an elevation of 750 m. These two sites, S02_5 and S02_6, were only 10–20 m apart but differed noticeably in their vegetation. Site S02_5 was a grass-dominated area whereas S02_6 samples were from shrub (mainly Empetrum nigrum ssp. hermaphroditum)-dominated soil. S02_7 and S02_8 were sampled further down the northwestern ridge, and S02_7 a little above and S02_8 below the treeline. S02_8 represented as similar a vegetation type as the higher sites as possible and was not as fertile and lush in vegetation as the site S01_6 of the 2001 sampling. Soil samples of Jehkas were obtained from the treeline (sample J9, 578 m) to the top of the fjeld (J1, 960 m). The vegetation was similar to that of Saana and was dominated by dwarf shrubs such as Empetrum nigrum ssp. hermaphroditum, Cassiope tetragona, Betula nana and Vaccinium vitis-idaea, alpine grasses, mosses and lichens.

Soil pH varied on Saana from 4.7 to 6.1 and on Jehkas from 4.6 to 6.2 (Table 1). A large variation in pH was observed within short distances on both fjelds which was most likely due to the presence of local basic bedrock materials (dolomite) and/or seepage waters which bring electrolytes from the dolomite rich areas. This was apparent for instance at the elevation of 750 m on Saana fjeld. A small stream of melting waters with pH close to 7 flowed between sample sites S02_5 and S02_6, which most likely was the source of nutrients and carbonate for site S02_5 that contained higher pH and nutrient levels (Table 1). A similar pH difference within a short distance was observed in Jehkas at 940 m between samples J2 and J3.

To determine whether altitude and the accompanying changes in vegetation and soil conditions affect microbial community structure, T-RFLP profiles along the northwestern slopes of Saana and Jehkas of samples from 2003 were compared. The NMS ordination patterns were highly similar with both restriction enzymes (MspI and HhaI) and only the MspI results are shown. In both fjelds, the samples with the highest (S02_5, J2) and lowest (S02_8, J9) pH (Table 1) separated most clearly from each other by NMS ordination (Fig. 3). Sampling site S02_5 separated from the other Saana sites also in the June and September 2002 samples (data not shown), indicating that the differences in the soil microbial communities were stable. When the scores of components 1 and 2 of the NMS ordination were correlated with pH, total nitrogen, ammonium, organic matter and soil moisture, pH correlated significantly (P<0.01) with component 1 and generally also with component 2 in all sample sets. Other soil properties did not correlate or correlated weakly and/or inconsistently with T-RFLP profiles indicating that pH was the most important factor controlling the T-RFLP profiles. To further analyze the significance of pH to the microbial community structures, all samples obtained from Saana and Jehkas were combined and classified in three pH categories: pH<5, pH=5.0–5.5 and pH >5.5. NMS ordination of all samples classified in the three pH categories further indicate that pH was the main factor controlling the T-RFLP profiles (Fig. 4a). Soil samples with high pH (>5.5) separated from the lower pH soil T-RFLP profiles of soil sampled from both fjelds.


Figure 3.  NMS ordination of T-RFLP (MspI digestion) profile of Saana (a) and Jehkas (b) samples. See Table 1 for sample codes. Sample altitudes are indicated in parentheses.

Download figure to PowerPoint


Figure 4.  NMS ordination of T-RFLP (MspI digestion) profiles (a) and PLFA profiles (b) of all soil samples obtained in 2002–2003 from Saana and Jehkas fjelds classified in three pH categories: pH<5.0, pH 5.0–5.5 and pH>5.5.

Download figure to PowerPoint

Phospholipid fatty acid analysis of the soils resulted in 35–40 distinct FAME peaks. The main fatty acids were straight chain C16 and C18 fatty acids and iso, anteiso and methyl branched C14–C18 fatty acids (Fig. 5c and d). Component 2 of the NMS ordination of Saana and Jehkas soil samples correlated (P<0.01) significantly with pH, indicating that also the soil PLFA profiles were influenced by soil pH (Fig. 4b).


Figure 5.  Relative abundance of dominating TRFs (a, b) and PLFAs (c, d) of different pH soil sampled from Saana (a, c) and Jehkas (b, d). Only the fragments detected at c. 2% or more are shown. The asterisks indicate TRFs that correlated significantly (*=P<0.05, **=P<0.01) with soil pH. The letters indicate the clones and isolates that matched with the soil TRFs (Table 2). See Table 1 for sample codes. The error bars indicate SD (n=2).

Download figure to PowerPoint

A comparison of the most abundant TRFs and PLFAs of soils which differed in pH and separated thus in the NMS ordination is shown in Fig. 5. In general, the dominating TRFs and PLFAs were similar in all soil samples but their relative abundance varied. The TRFs (MspI digestion) 265, 288/289 and 305 were present at much higher abundance in all low pH soils than in the high pH soils, whereas TRFs 138, 140 and 489/490 were present at higher proportion in the high pH samples of Saana and Jehkas (Fig. 5a and b). The highest variation in the PLFA profiles was generally observed in the C18 fatty acids. There was a high variation in the fungal fatty acid 18 : 2 in some replicate samples, which may at least partially be due to technical reasons. The samples were not sieved, which may have resulted in heterogeneous samples with varying amounts of fine roots and fungal biomass. Nevertheless, the relative proportion of 18 : 2 and 18 : 1ω9c clearly increased at low pH, whereas 18 : 1ω7c was present at higher amounts at higher pH (Fig. 5c and d). Variation was observed also in the relative proportion of 16 : 0, 16 : 1, a-15 : 0 and cyclopropane fatty acids. Generally cyc-17 : 0, a-15 : 0 and 16 : 1 fatty acids increased as pH increased, while the opposite was true for cyc-19 : 0 which was present at higher amounts in the low pH soils.

Comparison of T-RFLP and PLFA profiles of samples from early (June 2002 and June 2003) to late (August 2001 and September 2002) summer showed very little seasonal variation in the PLFA and T-RFLP profiles (data not shown). There were no substantial differences in the major T-RFLP peaks in early and late summer samples, indicating that the microbial communities were stable. Some variation was observed in the straight chain and branched fatty acids. Fatty acids iso-16 : 0, 18 : 2 and 18 : 1ω7c were slightly elevated in the early summer where as 16 : 0, 16 : 1ω7c, 16 : 1ω5c and 18 : 1ω5c increased towards late summer (data not shown).

Identification of dominant bacteria of the fjeld soils

As an attempt to identify the most dominating bacterial species in the tundra fjeld soils, TRF sizes of 100 clones and 60 isolated strains were analyzed. The 60 isolates were selected to represent different taxa based on fatty acid profiling and 16S rRNA gene sequencing of a total of 200 isolates. The most abundant TRF sizes of the cloned sequences were different from those observed most frequently in isolated strains, indicating that the cultivated strains were different from those obtained by molecular methods. T-RFLP analysis of the cloned sequences resulted in 56 different TRF sizes. The most common TRF size of the clones matched with the highest TRF peak A (Fig. 5a and b). This TRF was found in 19% of the clones which is in good agreement with the abundance of this TRF in the soil profile (15–20% of total fluorescence). Other clone TRFs matched peaks I (five clones), B (four clones), F (four clones), G (three clones), E (two clones), C (two clones) and K (two clones). In general, all TRFs that were represented by two or more clones matched with the dominating (>2% of total peak areas) soil TRF peaks further indicating good correlation with the T-RFLP analysis and cloning. The most common TRF sizes in isolated strains were 489–491 bp matching the soil TRF peaks M and N. In addition several isolates matched TRF peaks C, D and F. One of the isolates, strain SP1PR4, corresponded to one of the highest soil T-RFLP peaks (peak I). This strain was isolated from the extremely wind-exposed north ridge of Saana fell. The isolate grew extremely slowly and produced minute pale red colonies on R2A agar. Partial (700 bp) 16S rRNA gene sequence analysis placed it among the phylum Acidobacteria with 95.7% sequence similarity to Acidobacterium capsulatum.

Sequence analysis of two clones indicated the dominance of Acidobacteria in the tundra soil, as 16 of the sequences were most closely related to sequences of this phylum (Table 2). Comparison of the TRF sizes of the isolates and cloned sequences to the soil T-RFLP profiles further indicated the dominance of Acidobacteria. Three of the most dominating terminal fragments (peaks A, I and J in Fig. 5a) could be matched to nine acidobacterial sequences (Fig. 5a, Table 2). In addition, TRF peaks B, E and F matched Acidobacteria, but could also be attributed to a Firmicutes clone (peak B) and isolates identified as Bacillus sp. (peak E) and Sphingomonas spp. (peak F). TRF peaks C and G matched clones that fell in to Alpha- and Betaproteobacteria, closest to Burkholderia spp. and Bradyrhizobium spp. Soil TRF peaks H and K were more abundant in the Jehkas soil than in Saana and matched with clones that fell in the Planctomycetes and Actinobacteria, respectively. Among the isolates, the most common TRFs (490±1 bp) could be assigned to several beta- and gammaproteobacterial taxa, including Pseudomonas spp., Collimonas spp. and Janthinobacterium spp. Peaks C and D matched three isolates that were identified as Burkholderia spp. and three isolates matching TRF peak F were identified as Sphingomonas spp. The TRF size of a Bacillus sp. and a Paenibacillus sp. isolate matched peaks E and G, respectively. The Acidobacteria sp. SP1PR4 isolate matched peak I together with four acidobacterial clones. In addition to the two clones in Table 2, we sequenced several clones that did not match any dominating TRF peak of the soil samples. Among these were sequences related to Firmicutes (three clones), Actinobacteria (two clones) Chloroflexi (one clone), Cyanobacteria (two clones) and Alphaproteobacteria (two clones).

Table 2.   Phylogenetic assignment of tundra soil clones and isolated strains that matched the dominating soil T-RFLP peaks
TRF peakCorresponding clone(s) or isolate(s)Accession numberPhylogenetic affiliationClosest cultured relative*Accession numberSimilarity (%)
  • *

    The uncultivated relatives are shown to sequences for which no close relatives (>90% sequence identity) were found among the cultivated organisms but for which a close relative was found among uncultivated organisms.

  • The sequence lengths compared were between 400 and 1000 bp.

AClone Cl1.59AM397026AcidobacteriaBacterium K-5b2AF52486096.7
Clone Cl1.63AM397027AcidobacteriaAcidobacteria Ellin7137AY67330397.0
Clone Cl1.102AM397028AcidobacteriaBacterium Ellin5095AY23451295.7
Clone Cl1.138AM397029AcidobacteriaAcidobacteria Ellin7137AY67330394.9
BClone Cl1.130AM397030AcidobacteriaUncultured soil bacteriumAB24025196.1
Clone Cl1.121AM397031AcidobacteriaBacterium Ellin7184AY67335096.2
Clone Cl1.116AM397032FirmicutesUncultured soil bacteriumAY91347990.3
CClone Cl1.54AM397033BetaproteobacteriaBacterium Ellin6095AY23474794.9
Clone Cl1.109AM397034BetaproteobacteriaBurkholderia glathei isolate Hg 19AY15437991.5
Burkholderia sp. S2J1AM397046BetaproteobacteriaBurkholderia sp. SB5AJ97135499.2
DBurkholderia sp. SP4MA1AM397047BetaproteobacteriaBurkholderia glathei isolateAY60569599.8
Burkholderia sp. M1S2AM397054BetaproteobacteriaBurkholderia sp. SB5AJ971354100.0
EClone Cl1.57AM397035AcidobacteriaBacterium Ellin7184AY67335093.3
Bacillus sp. S5R4DQ234527FirmicutesBacillus sp. NIPHL090904/B1AY74891299.5
FClone Cl1.110AM397036AcidobacteriaSolibacter usitatus Ellin6076AY23472893.7
Clone Cl1.141AM397037AcidobacteriaSolibacter usitatus Ellin6076AY23472892.9
Clone Cl1.23AM397038AcidobacteriaSolibacter usitatus Ellin6076AY23472893.1
Clone Cl1.49AM397055AcidobacteriaBacterium Ellin5121AY23453890.9
Sphingomonas sp. M2J2DQ234495AlphaproteobacteriaSphingomonas sp. Pmxh3DQ31473499.7
Sphingomonas sp. SP1PSA7AM397048AlphaproteobacteriaSphingomonas sp. Enf2DQ33961099.0
GClone Cl1.96AM397039AlphaproteobacteriaBradyrhizobium elkanii strain NW-6AY56851395.7
Paenibacillus sp. M1TS4DQ234493FirmicutesGlacier bacterium FJS42AY31517099.8
HClone Cl1.81AM397040PlanctomycetesUncultured Planctomyces sp.AB01552797.5
IClone Cl1.88AM397041AcidobacteriaSolibacter usitatus Ellin6076AY23472892.7
Clone Cl1.104AM397042AcidobacteriaBacterium Ellin7137AY67330396.1
Clone Cl1.16AM397043AcidobacteriaBacterium Ellin5056AY23447397.6
Clone Cl1.30AM397056AcidobacteriaBacterium Ellin7137AY67330397.8
Acidobacterium sp. SP1PR4AM397049AcidobacteriaAcidobacteriaceae bacterium TAA43AY58722896.8
JClone Cl1.41AM397044AcidobacteriaBacterium Ellin7184AY67335090.4
KClone Cl1.103AM397045ActinobacteriaActinobacterium WJ25AY49595489.9
LSphingobacteriaceae sp. SP4AG4AM397050BacteroidetesBacteroidetes bacterium CHNCT12DQ33755895.6
MJanthinobacterium sp. S3T4DQ234521BetaproteobacteriaJanthinobacterium sp. HHS7AJ84627299.2
Pseudomonas sp. M1TS5AM397051GammaproteobacteriaPseudomonas gingeriTAF32099198.4
Pseudomonas sp. S1M1DQ234511GammaproteobacteriaPseudomonas sp. BE1dilAY26347199.9
Collimonas sp. M2T1aDQ234496BetaproteobacteriaCollimonas sp. CTO 291AY28114399.0
Pseudomonas sp. SP4PAG1AM397052GammaproteobacteriaPseudomonas sp. SE22#1aAY26347799.9
Pseudomonas sp. S3B2DQ234519GammaproteobacteriaPseudomonas frederiksbergensisTAJ249382100.0
NPseudomonas sp. SP1PLEA1AM397053GammaproteobacteriaPseudomonas borealisAJ01271299.7
Collimonas sp. S5TS4DQ234529BetaproteobacteriaCollimonas sp. CTO 291AY28114399.2

The effect of soil pH on microbial community composition was apparent. Of the TRFs that were more abundant at low than high pH soils (Fig. 5a and b) on Saana and/or Jehkas, three (peaks A, I and J) matched exclusively with acidobacterial sequences. One major peak (305 bp) did not match any of the clones or isolates. Peaks B, C, H and M, which were more abundant at the high pH soils, matched two acidobacterial clone sequences (B), a betaproteobacterial sequence and an isolate closely related to Burkholderia spp. (C), a Planctomycetes sequence (H), and several beta- and gammaproteobacterial isolates (M) (Table 2). TRF peak B matched also a clone that fell closest to Firmicutes. In addition, peak G matched with an alphaproteobacterial sequence and an isolate closest to Bacillus sp. These results indicate that low pH enriches Acidobacteria while at higher pH (>5.5) the share of Proteobacteria increases. Furthermore, phylogenetic analyses indicated that the acidobacterial clones and the SP1PR4 isolate obtained in this study were widely distributed in four subdivisions of the division Acidobacteria (Fig. 6). Acidobacterial clones that matched peaks A and I belonged generally to subdivision 1, while those corresponding to peak F belonged mostly to subdivision 3. Clones matching TRF peaks B, E, F and J belonged to subdivision 2 of Acidobacteria, except clone Cl1.130, which was closest to subdivision 7 sequences (Fig. 6).


Figure 6.  Phylogenetic relationship of the acidobacterial sequences and isolate SP1PR4 and their closest relatives based on partial 16S rRNA gene sequences. The phylogenetic tree was constructed using the neighbor-joining method with Jukes-Cantor corrections of 436 nucleotide positions. The corresponding TRF peaks (letters A–H) that matched each clone and the isolate are indicated. Closed circles indicate bootstrap support over 90%, open circles indicate bootstrap support over 60%. The scale bar indicates 0.1 changes per nucleotide. Accession number of Cl1.79 is AM397057. All other accession numbers are shown in Table 2. The accession numbers for the reference sequences are shown in parenthesis.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Microbial community structures of Arctic tundra soils are poorly characterized and while it is known that these environments harbor cold-adapted microorganisms that are active also at subzero temperatures (Schmidt & Lipson, 2004), very little is known about the species diversity and seasonal variation of the communities. We have analyzed bacterial community compositions of the Arctic-alpine tundra soils by culture-based methods (Männistö & Häggblom, 2006) which identified several novel bacterial taxa that were active at 0°C or below. This study is the first description of the Fennoscandian tundra microbial communities using molecular methods.

Based on molecular analysis, the division Acidobacteria dominated the Kilpisjärvi tundra fjeld soils. This was anticipated as Acidobacteria are widely distributed in the terrestrial environments (Barns et al., 1999). Acidobacteria have been reported as dominating bacteria in various soils (for a review see Janssen, 2006), including tundra soils of the Colorado Niwot Ridge (Lipson & Schmidt, 2004) and Siberia (Zhou et al., 1997). The diversity of the Acidobacteria in the tundra soils of Kilpisjärvi was high as the 17 cloned acidobacterial sequences represented at least four of the seven subdivisions of Acidobacteria (Hugenholtz et al., 1998). While Acidobacteria are commonly detected in soil environments by molecular methods, they are rarely cultivated. We isolated one Acidobacterium strain during this study. The strain SP1PR4 was cultivated on 20% R2A at pH 7 indicating that general medium and neutral pH can support the growth of this taxon. However, a long incubation time (2 months at 5°C) was required for the initial growth of this strain. Besides Acidobacteria, Gamma- and Deltaproteobacteria have been reported to be abundant in Arctic tundra, while Verrucomicrobia have been detected in high frequency in the Alpine tundra soils of Niwot Ridge (for review, see Nemergut et al., 2005). T-RFLP and clone analysis indicated that Alpha- and Betaproteobacteria were also abundant in the Kilpisjärvi tundra soil. Sequences related to Burkholderia and Bradyrhizobium matched with dominating TRF peaks.

The microbial community composition of the Kilpisjärvi fjelds was very resilient as indicated by stable T-RFLP profiles in the beginning and end of the growing season and over a 2-year sampling period. This was unexpected, as our initial assumption was that the freezing of soil and freeze–thaw events during the spring would alter the microbial communities that flourish during the warm summer months. A slight increase in straight chain C16 and C18 fatty acids and a decrease in iso-16 : 0 and 18 : 2 fatty acids was observed in the PLFA profiles of the late summer samples. Differences in the PLFA profiles may, however, be attributed to thermal adaptation of microbial membranes (e.g. Suutari & Laakso, 1994) rather than changes in microbial community composition as the temperature may be up to 10°C lower in early compared to late summer. Decrease of the fungal fatty acid 18 : 2 towards late summer is in agreement with other studies (Lipson et al., 2002) that have reported higher abundance of fungi in cold vs. warm seasons. Seasonal changes in the microbial activity and community compositions have been detected in other studies (Lipson & Schmidt, 2002, Lipson & Schmidt, 2004, Monson et al., 2006). In Niwot Ridge alpine soil, different bacterial species were reported to dominate in winter and summer soils (Lipson et al., 2004; Monson et al., 2006) indicating that there was a shift in the microbial community composition as soil freezes. In Niwot Ridge, the relative proportion of Acidobacteria was most abundant during spring, Betaproteobacteria and Verrucomicrobia were most abundant during summer and Bacteroidetes peaked in the winter soils (Lipson & Schmidt, 2004). In addition to seasonal changes, differences in topography-driven soil moisture appears to have a strong impact on microbial communities. The share of Acidobacteria and Alphaproteobacteria was higher in dry meadows while Chloroflexi sequences were more abundant in the wet meadows of Niwot Ridge (Nemergut et al., 2005).

In addition to the stable T-RFLP profiles over the 2-year sampling period, the microbial communities were relatively similar at different elevations and even different fjelds as long as the soil pH was similar. Although snow depths were not monitored at the sampling sites, observations during the winter 2005–2006 indicated that snow accumulation at wind-exposed sites S5–S7 was considerably less (between 5 and 30 cm in April 2006) than at the other sites of Saana (>50 cm in April 2006). Snow accumulation has a significant effect on winter soil temperatures. We have measured temperatures down to −25°C in top soil with <5 cm of snow, while under thick (50 cm or more) snow pack the temperature is relatively independent of air temperature and remains close to 0°C throughout the winter months (M. Männistö, unpublished data). We did not detect considerable differences in T-RFLP profiles of the samples from wind-exposed vs. sheltered ridges, suggesting that the winter soil temperature did not have drastic effects on the fjeld tundra microbial communities. More detailed studies on the effects of soil temperature variation and freeze–thaw cycles on microbial community structures are currently underway.

Based on both T-RFLP and PLFA profiles, out of the environmental conditions analyzed, pH influenced most clearly the microbial community compositions in the Kilpisjärvi tundra soil. The Kilpisjärvi region is unique in Finland as the Scandinavian mountain range extends only to this part of Finland. Due to the presence of limestone in the bedrock of the Kilpisjärvi region, soil pH is higher there than in many other parts of Finland. Due to the stratification of different bedrock materials, pH varies within short distances in the fjeld soils. Fierer & Jackson (2006) reported recently that bacterial diversity and richness in soils sampled across North and South America could largely be explained by soil pH, while temperature, latitude and other environmental conditions had little influence on the bacterial diversity. Furthermore, there was no clear relationship between plant diversity and soil bacterial diversity (Fierer & Jackson, 2006). This is in agreement with our observations that bacterial T-RFLP profiles do not correlate well with plant species composition (M. Männistö, unpublished data) and soil pH is the most significant factor influencing soil bacterial diversity. Our results indicate that low pH enriches especially for members of the phylum Acidobacteria.

PLFA results indicated that pH influences also the relative abundance of fungi in the tundra soil, as low pH soils of Kilpisjärvi were enriched especially with fatty acids 18 : 2 and 18 : 1ω9c while the share of a-15 : 0 and 16 : 1 fatty acids was higher in the high pH soils. The higher abundance of 18 : 2 and 18 : 1ω9c indicates a higher proportion of fungal biomass at low pH (Olsson, 1999; Ruess et al., 2002). Soil pH has been reported to significantly influence the microbial community compositions also in forest soils (Hackl et al., 2005; Bååth and Anderson, 2003). Similarly to what we observed in the fjeld soils, a higher abundance of fungi in low pH soils was reported in forest soils (Bååth & Anderson, 2003; Hackl et al., (2005), while a higher share of bacterial 16 : 1 and cyc-17 : 0 fatty acids at higher pH was found (Bååth and Anderson; Hackl et al., 2005). Arao (1999) studied the effect of pH in microbial community composition by incubating soils with 13C acetate and found that at low pH 13C was incorporated mainly to 18 : 2, 18 : 1ω9c and 16 : 0, while increased incorporation of 13C into bacterial PLFAs was detected at pH 7-8. Bååth & Anderson (2003) detected a decrease of i-15 : 0 and i-16 : 0 as the pH increased while no pH effect on anteiso fatty acids was reported. We observed an increase of a-15 : 0 at higher pH, while i-15 : 0 was not affected by pH, similar to what was observed by Hackl et al. (2005). Iso- and anteiso branched fatty acids are generally considered as biomarkers for Gram-positive bacteria. Although they are present in most Gram-positive taxa, several Gram-negative bacteria also contain significant amounts of iso/anteiso fatty acids. Moreover, the assignment of signature fatty acids is based on cultivated species, although it is well known that the majority of environmental bacteria remain to be cultivated and the fatty acid composition of the great majority of soil bacteria is still unknown. Both Acidobacterium capsulatum (Kishimoto et al., 1991) and the Acidobacterium isolate of this study (M. Männistö, unpublished data) contain high amounts of iso-branched fatty acids (especially i-15 : 0), indicating that Acidobacteria may be important contributors to the iso-branched fatty acids of soil PLFA profiles. T-RFLP analysis indicated a significant decrease of Acidobacteria at soil pH higher than 5.5. There was, however, no significant differences in the share of iso-branched fatty acids at high and low pH soils. This may indicate that while the share of Acidobacteria decreased at high pH, the share of some other iso-branched fatty acid containing taxa increased. The relatively high amount of 10-methyl branched fatty acids suggested high abundance of Actinobacteria in the Kilpisjärvi tundra soils (Kroppenstedt, 1985). Actinobacteria were not, however, connected to the major T-RFLP peaks of the soil profiles although three cloned sequences fell in to this phylum. It is possible that actinomycetes were resistant to the DNA extraction method and hence underrepresented in T-RFLP profiles.

Of the soil T-RFLP profiles, eight major TRF peaks (present at >5% of total fluorescence) could be matched with TRF sizes of cloned sequences or isolated strains. As several phylogenetically disparate organisms may have identical terminal restriction sites, a single soil T-RFLP peak may correspond to multiple species. The major TRF peaks A and I matched exclusively with four and five acidobacterial sequences, respectively, suggesting that at least these TRF sizes are strongly associated with the phylum Acidobacteria. In contrast, peak M matched with several beta- and gammaproteobacterial genera suggesting that this terminal restriction site is ubiquitous among several taxa. T-RFLP analysis of soil samples coupled to analysis of cloned sequences and isolates suggested that the abundance of Acidobacteria was dependent on soil pH. Subdivision 1 Acidobacteria have recently been reported to be significantly more abundant in low pH soils (Sait et al., 2006). We detected a strong correlation with pH in the abundance of TRF peaks that matched with subdivision 1, 2 and 3 acidobacterial clones. These results indicate that in addition to subdivision 1 (Sait et al., 2006), the occurrence of also other subdivisions of Acidobacteria may be strongly influenced by pH. In the higher pH soils (S5, J2, J6) the relative abundance of TRFs that matched beta- and gammaproteobacterial sequences and isolates was higher than at low pH suggesting that the Beta- and Gammaproteobacteria that were frequently isolated from the soils may be more common in the high pH than low pH soils. More studies are, however, needed to confirm the relationship of soil TRF sizes with various bacterial taxa. The cultivation of most of the isolates was done at pH 6–7 which most likely enhanced the growth of the Proteobacteria. We have analyzed growth pH range of several beta- and gammaproteobacterial isolates which has indicated that most of the Proteobacteria grow poorly at pH below 5 (M. Männistö, unpublished data). The high abundance of Acidobacteria at low pH may thus result from the inability of proteobacterial species to compete in the low pH soils.

In summary, similarities in the soil microbial communities at the two fells (Saana and Jehkas), as indicated by T-RFLP, the clonal library and PLFA analyses, as well as cultivation, suggested that certain bacterial taxa may be ubiquitously distributed in the Arctic tundra. Moreover, many clones of this study were closely related to GenBank sequences from Alaska (Toolik Lake) tundra soils indicating circumpolar distribution of these species in the tundra environments. Furthermore, these microbial communities appear to be stable with little variation observed during the summer season or from year to year. Variation in structure of the soil microbial communities appears to be controlled by pH more than temperature fluctuations. The results of this study indicate that soil temperature may have a limited impact on bacterial community composition. Climate change and subsequent increase of temperature fluctuations and freeze–thaw events may therefore not modify the soil bacterial communities as much as generally thought. More detailed research is, however, needed to assess the microbial community structures in frozen soil and after various freeze–thaw cycles.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study was funded by the Academy of Finland and an EU regional Development Fund from the National Technology Agency (Tekes). We thank Riitta Nielsen for assistance in DNA extractions and cultivating the isolates.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, Miller W & Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25: 33893402.
  • Arao T (1999) In situ detection of changes in soil bacterial and fungal activities by measuring 13C incorporation into soil phospholipid fatty acids from 13C acetate. Soil Biol Biochem 31: 10151020.
  • Bååth E & Anderson T-H (2003) Comparison of soil fungal/bacterial ratios in a pH gradient using physiological and PLFA-based techniques. Soil Biol Biochem 35: 955963.
  • Barns SM, Tkala SL & Kuske CR (1999) Wide distribution and diversity of members of the bacterial kingdom Acidobacterium in the environment. Appl Environ Microbiol 65: 17311737.
  • Cavicchioli R, Siddiqui KS, Andrews D & Sowers KR (2002) Low temperature extremophiles and their applications. Curr Opin Biotechnol 13: 253261.
  • Clein JS & Schimel JP (1995) Microbial activity of tundra and taiga soils at sub-zero temperatures. Soil Biol Biochem 27: 12311234.
  • Dunbar J, Ticknor LO & Kuske CR (2001) Phylogenetic specificity and reproducibility and new method for analysis of terminal restriction fragment profiles of 16S rRNA genes from bacterial communities. Appl Environ Microbiol 67: 190197.
  • Fierer N & Jackson RB (2006) The diversity and biogeography of soil bacterial communities. Proc Natl Acad Sci USA 103: 626631.
  • Hackl E, Pfeffer M, Donat C, Bachmann Z & Zechmeister-Boltenstern S (2005) Composition of the microbial communities in the mineral soil under different types of natural forests. Soil Biol Biochem 37: 661671.
  • Hugenholtz P, Goebel BM & Pace NR (1998) Impact of culture-independent studies on the emerging phylogenetic view of bacterial diversity. J Bacteriol 180: 47654774.
  • Jackson CR, Harper JP, Willoughby D, Roden EE & Churchill PF (1997) A simple, efficient method for the separation of humic substances and DNA from environmental samples. Appl Environ Microbiol 63: 49934995.
  • Janssen PH (2006) Identifying the dominant soil bacterial taxa in libraries of 16S rRNA and 16S rRNA genes. Appl Environ Microbiol 72: 17191728.
  • Johnson JL (1994) Similarity analysis of rRNAs. Methods for General and Molecular Bacteriology (GerhardtP, MurrayRGE, WoodWA & KriegNR, eds), pp. 683700. American Society for Microbiology, Washington, DC, USA.
  • Kishimoto N, Kosako Y & Tano T (1991) Acidobacterium capsulatum gen. nov., sp. nov: an acidophilic chemoorganotrophic bacterium containing menaquinone from acidic mineral environment. Curr Microbiol 22: 17.
  • Kroppenstedt RM (1985) Fatty acid and menaquinone analysis of Actinomycetes and related organisms. Chemical Methods in Bacterial Systematics (GoodfellowM & MinnikinDE, eds), pp. 173199. Academic Press, London, UK.
  • Larsen KS, Jonasson S & Michelsen A (2002) Freeze–thaw cycles and their effects on biological processes in two Arctic ecosystem types. Appl Soil Ecol 21: 187195.
  • Lipson DA & Schmidt SK (2004) Seasonal changes in an alpine soil bacterial community in the Colorado Rocky Mountains. Appl Environ Microbiol 70: 28672879.
  • Lipson DA, Schadt CW & Schmidt SK (2002) Changes in soil microbial community structure and function in an alpine dry meadow following spring snow melt. Microb Ecol 43: 307314.
  • Männistö MK & Häggblom MM (2006) Characterization of psychrotolerant heterotrophic bacteria from Finnish Lapland. System Appl Microbiol 29: 229243.
  • McCune B & Mefford MJ (1999) Multivariate Analysis of Ecological Data, Version 4.25. MjM Software, Gleneden Beach, Oregon, USA.
  • McCune B & Grace JB (2002) Analysis of Ecological Communities. MjM Software Design, Gleneden Beach, Oregon, USA.
  • Mikan CJ, Schimel JP & Doyle AP (2002) Temperature controls of microbial respiration in Arctic tundra soils above and below freezing. Soil Biol Biochem 34: 17851795.
  • Miller L & Berger T (1985) Bacterial Identification by Gas Chromatography of Whole Cell Fatty Acids. Hewlett-Packard Application Note 228-41. Hewlett-Packard co, Palo Alto, CA, USA.
  • Monson RK, Lipson DL, Burns SP, Turnipseed AA, Delany AC, Williams MW & Schmidt SK (2006) Winter forest soil respiration controlled by climate and microbial community composition. Nature 439: 711714.
  • Nemergut DR, Costello EK, Meyer AF, Pescador MY, Weintraub MN & Schmidt SK (2005) Structure and function of alpine and arctic soil microbial communities. Res Microbiol 156: 775784.
  • Neufeld JD, Yu ZT, Lam W & Mohn WW (2004) Serial analysis of ribosomal sequence tags (SARST): a high-throughput method for profiling complex microbial communities. Environ Microbiol 6: 131144.
  • Olsson PA (1999) Signature fatty acids provide tools for determination of the distribution and interactions of mycorrhizal fungi in soil. FEMS Microbiol Ecol 29: 303310.
  • Richardson MI, Moore IT, Soma KK, Lei F-M & Wingfield JC (2003) How similar are high latitude and high altitude habitants? A review and a preliminary study of the adrenocortical response to stress in birds of the Ainghai-Tibetan plateau. Acta Zool Sinica 49: 119.
  • Rivkina M, Friedmann EI, McKay C & Gilichinsky DA (2000) Metabolic activity of permafrost bacteria below the freezing point. Appl Environ Microbiol 66: 32303233.
  • Ruess L, Häggblom MM, Garcia Zapara EJ & Dighton J (2002) Phospholipid fatty acids of fungi and nematodes – possible biomarkers in the food chain? Soil Biol Biochem 34: 745756.
  • Ruess L, Schütz K, Haubert D, Häggblom MM, Kandeler E & Scheu S (2005) Application of lipid analysis to understand trophic interactions in soil. Ecology 86: 20752082.
  • Sait M, Davis KER & Janssen PH (2006) Effect of pH on isolation and distribution of members of subdivision 1 of the phylum Acidobacteria occurring in soil. Appl Environ Microbiol 72: 18521857.
  • Schadt CW, Martin AP, Lipson DA & Schmidt SK (2003) Seasonal dynamics of previously unknown fungal lineages in tundra soils. Science 301: 13591361.
  • Schmidt SK & Lipson DA (2004) Microbial growth under the snow: implications for nutrient and allelochemical availability in temperate soils. Plant Soil 259: 17.
  • Suutari M & Laakso S (1994) Microbial fatty acids and thermal adaptation. Crit Rev Microbiol 20: 285328.
  • Welker JM, Fahnestock JT & Jones MH (2000) Annual CO2 flux in dry and moist Arctic tundra: field responses to increases in summer temperatures and winter snow depth. Climatic Change 44: 139150.
  • Williams BL, Shand CA, Hill M, O'Hara C, Smith S & Young ME (1995) A procedure for the simultaneous oxidation of total soluble nitrogen and phosphorus in extracts of fresh and fumigated soils and litters. Commun Soil Sci Plant Anal 26: 91106.
  • Zhou J, Bruns MA & Tiedje JM (1996) DNA recovery from soils of diverse composition. Appl Environ Microbiol 62: 316322.
  • Zhou JZ, Davey ME, Figueras JB, Rivkina E, Gilichinsky D & Tiedje JM (1997) Phylogenetic diversity of bacterial community determined from Siberian tundra soil DNA. Microbiology (UK) 143: 39133919.