The impact of different soil parameters on the community structure of dominant bacteria from nine different soils located on Livingston Island, South Shetland Archipelago, Antarctica


  • Editor: Max Häggblom

Correspondence: Lars Ganzert, Alfred Wegener Institute for Polar and Marine Research, Research Unit Potsdam, Telegrafenberg A45, 14473 Potsdam, Germany. Tel.: +49 331 288 2162; fax: +49 331 288 2188; e-mail:


Microorganisms inhabit very different soil habitats in the ice-free areas of Antarctica, playing a major role in nutrient cycling in cold environments. We studied the soil characteristics and the dominant bacterial composition from nine different soil profiles located on Livingston Island (maritime Antarctica). The total carbon (TC) and total nitrogen (TN) values were high for the vegetated soils, decreasing with depth, whereas the values for the mineral soils were generally low. Soil pH was more acidic for moss-covered soils and neutral to alkaline for mineral soils. Numbers of culturable heterotrophic bacteria were higher at vegetated sites, but significant numbers were also detectable in carbon-depleted soils. Patterns of denaturing gradient gel electrophoresis (DGGE) revealed a highly heterogeneous picture throughout the soil profiles. Subsequent sequencing of DGGE bands revealed in total 252 sequences that could be assigned to 114 operational taxonomic units, showing the dominance of members of the Bacteroidetes and Acidobacteria. The results of phospholipid fatty acid analysis showed a lack of unsaturated fatty acids for most of the samples. Samples with a prevalence of unsaturated over saturated fatty acids were restricted to several surface samples. Statistical analysis showed that the dominant soil bacterial community composition is most affected by TC and TN contents and soil physical factors such as grain size and moisture, but not pH.


Ice-free areas represent only 0.32% of the Antarctic continent (Ugolini & Bockheim, 2008) and they are characterized by extreme climate conditions, with subzero temperatures almost throughout the entire year, low water availability and low nutrient concentrations. The tip of the Antarctic Peninsula is determined by a polar maritime climate, resulting in positive summer air temperatures, repeated freeze–thaw events and a distinctly higher precipitation than in continental areas. Liquid water can be found during the summer months when temperatures rise above 0 °C. These conditions allow the presence of scarce vegetation (lichens, mosses, grasses) that has a direct influence on the composition of soil microbial communities by providing degradable polymeric organic matter in the form of plant litter or root-based low-molecular substances. These root exudates can be used by the soil microbiota as a carbon source (MacRae & Castro, 1966; Rovira, 1969; van Veen et al., 1989; Ström et al., 2003). Thus, vegetation and microbial activity have a small-scale influence on the soil chemical parameters and therefore also on the structural soil microbial composition.

Data on microbial communities in Antarctic terrestrial habitats are still quite rare. Nevertheless, during the last two decades, it became evident that these extreme habitats harbor a unique community of microorganisms, and an increasing number of studies have focused on unraveling the structural network, the adaptation mechanisms and the physiological properties of these microbial communities, also in the face of possible biotechnological applications (Nichols et al., 1999, 2004; Aislabie et al., 2006, 2008; Smith et al., 2006; Novis et al., 2007; Yergeau et al., 2007, 2009; Niederberger et al., 2008; Pointing et al., 2009). Furthermore, the response of bacterial communities on climate change has become the focus of investigations in the last few years (Bokhorst et al., 2007; Rinnan et al., 2009).

The application of culture-dependent approaches describes only a limited number of taxa due to the selection of certain microorganisms when using artificial growth conditions (e.g. medium, substrate, temperature, oxygen concentrations). Therefore, culture-independent methods have become an important alternative in environmental microbiology to investigate complex microbial systems in various Antarctic habitats for example lake and marine sediments (Purdy et al., 2003; Li et al., 2006), dry mineral soils (Aislabie et al., 2006; Smith et al., 2006; Shravage et al., 2007), sea ice (Brown & Bowman, 2001; Brinkmeyer et al., 2003), freshwater and sea water (Pearce et al., 2003; Gentile et al., 2006) and microbial mats (Taton et al., 2006; Fernández-Valiente et al., 2007). A number of investigations used the method of denaturing gradient gel electrophoresis (DGGE) to assess the diversity and community structure of bacteria in different antarctic and subantarctic terrestrial environments (Chong et al., 2009a, b; Foong et al., 2010). DGGE can visualize only numerically dominant bacteria, which make up at least 1% or more of the in situ population of the overall bacterial community (Muyzer et al., 1993).

Microbiological investigations of different Antarctic mineral soils showed that these environments can harbor high bacterial numbers of up to 109 cells g−1 dry soil (Ramsay & Stannard, 1986; Roser et al., 1993; Bölter, 1995; Cowan et al., 2002; Aislabie et al., 2006, 2008; Xiao et al., 2007; Cannone et al., 2008), which are comparable to numbers from temperate forest (Ekelund et al., 2001), glacial forefield soils (Sigler et al., 2002) and Siberian arctic tundra lowlands (Schmidt, 1999; Kobabe et al., 2004).

In this study, we investigated and compared different maritime Antarctic soils sites located on Livingston Island to identify the dominant phylogenetic bacterial groups as well as compositional patterns. Sites with and without vegetation were chosen for analyzing different soil parameters to show their impact on the dominant soil bacterial community and how the bacterial community could possibly be altered when plant growth increases in the maritime Antarctic due to future climate change.

Materials and methods

Study area

Hurd Peninsula is located in the southern part of Livingston Island, one of the largest islands in the South Shetland Archipelago (Fig. 1). Ice-free areas can be found mainly along the coast, whereas the central and northwestern part of the peninsula is covered by glaciers. Climatic conditions can be defined as cold maritime, with mean summer temperatures of ∼2 °C, whereas daily maxima can be up to 10 °C. During the winter, temperatures are always below 0 °C, with minima of −35 °C (reviewed in Toro et al., 2007). The mean annual precipitation is much higher than in most continental areas of Antarctica, with values between 500 and 1000 mm (Bañón, 2001; Vieira & Ramos, 2003).

Figure 1.

 (a) Location of Livingston Island within the South Shetland Archipelago (black rectangle). (b) The satellite image shows the Hurd Peninsula (white rectangle).

The main study site was located in an ice-free area north-east of the Bulgarian Antarctic Base characterized by basaltic and andesitic bedrock at 64 m a.s.l. The soils and sediments were mainly characterized by coarse grain-sized (>2 mm) material and soil horizons can only be differentiated in the course of initial soil formation (e.g. humus accumulation, brownification), which was observed at some sites. Beyond this, sites were characterized by loose sediments, till deposits covered by lapilli-sized volcanic ashes and permafrost. Volcanic ashes precipitated by eruption of the nearby Deception Volcano (Calvet et al., 1993) were prevalent in the profiles, either as sole layers or mixed into the surface debris. Vegetation was found at different sites, with lichens preferring rocky ground, whereas mosses were restricted to small depressions that contain finer soil material.

The last major glacial retreat on Livingston Island started about 5–4 ka bp (Björck et al., 1991, 1996; Björck & Zale, 1996). In the ice-free areas, continuous permafrost can be found in altitudes above 35 m a.s.l., whereas below this altitude, the distribution of permafrost is isolated (Ramos & Vieira, 2003; Schwamborn et al., 2008).

Soil sampling

During the austral summer period of 2005, soil samples from ice-free areas of Hurd Peninsula (Livingston Island) were taken (Fig. 2). Sites were chosen by different habitat properties (e.g. with or without vegetation, surface characteristics) and exposition. Altogether, nine soil profiles were excavated and described and 38 samples were taken from defined layers for soil chemical, soil physical and soil microbiological analyses (Table 1). Five profiles (T1-1 to T1-5) were situated along a transect from the top of a hill to a small depression. The remaining profiles (SP-A to SP-D) were chosen by distinct site characteristics. Soil samples of each depth were taken from an area of 25 × 25 cm and homogenized directly in the field. Subsamples for soil chemical and soil physical analyses were transferred into sealable plastic bags and stored at 4 °C until further processing. Subsamples for microbiological analyses were placed in 250 or 500 mL sterile Nalgene® boxes, which were frozen immediately. Continuous cooling at −20 °C was ensured for microbial samples by the transfer with the research vessels MS Las Palmas and RV Polarstern to Germany.

Figure 2.

 Detailed picture of the Hurd Peninsula with the position of the sampling sites.

Table 1.   Sampling site description
SamplesCoordinatesSite description
T1-1S62°38′24.2″, W60°21′41.7″Silty sand in a bedrock depression; covered with a moss layer
T1-2S62°38′24.2″, W60°21′41.2″Silty sand with gravel and stones; lichens adherent on rocks
T1-3S62°38′24.2″, W60°21′40.2″Silty sand with gravel and stones; bands of gravel between 5–9 cm and 20–23 cm depth
T1-4S62°38′24.3″, W60°21′39.5″Silty sand with the watertable at 15 cm depth; moss cover on top
T1-5S62°38′24.1″, W60°21′39.4″Sandy volcanic ashes with gravel; till starting at a depth of 40 cm
SP-AS62°38′28.3″, W60°21′12.1″Sandy volcanic ashes with gravel; snowbed nearby, permafrost starting at 20 cm depth
SP-BS62°38′24.5″, W60°21′36.4″From 0 to 5 cm a volcanic ash layer with gravel and stones; from 5 to 35 cm sandy silt, permafrost occurred at 35 cm depth
SP-CS62°38′31.2″, W60°21′45.8″From 0 to 10 cm silty sand with gravel and stones, followed by sandy silt with gravel; lichens on top
SP-DS62°38′40.1″, W60°22′12.8″From 0 to 10 cm sandy volcanic ashes with gravel and stones; from 10 to 15 cm depth fine, silty-clayey material with gravel; from 15 to 20 cm sandy silt with gravel; lichens on top

Soil analysis

Grain size distribution was determined using freeze-dried, homogenized and sieved soil samples (<2 mm) of about 15 g (Schlichting et al., 1995). Water content was determined after freeze-drying the soil samples. The total carbon (TC) and total nitrogen (TN) contents were analyzed on a milled sample using an automatic element analyser (Elementar Vario EL Bundy III). Conductivity and pH were measured on a 1 : 2.5 soil : deionized water slurry that was shaken for 1 h in the dark using a MultiLab 540 (WTW, Germany).

The numbers of cultivable heterotrophic bacteria were determined by plating serial soil solutions (until 10−6) on modified BR agar plates (Ganzert et al., 2011) with incubation at 10 °C for 7 days.

Analysis of phospholipid fatty acid (PLFA) and phospholipid ether lipids (PLEL)

Lipids were extracted from 10 to 20 g sample material, according to the Bligh–Dyer method as described elsewhere (Knief et al., 2003). The resulting lipid material was fractionated into neutral lipids, glycolipids and phospholipids on a silica-bonded phase column (SPE-SI; Bond Elute, Analytical Chem International, CA) by elution with chloroform, acetone and methanol, respectively. An aliquot of the phospholipid fraction was taken for PLFA analysis. After mild alkaline hydrolysis, the lipid extract was supplemented with nonadecanoic acid methyl ester as an internal standard. The analyses of the fatty acid methyl ester extracts were performed using GC–MS as described previously (Lipski & Altendorf, 1997).

An aliquot of the phospholipid fraction was used for PLEL analysis according to Gattinger et al. (2003). After the formation of ether core lipids, ether-linked isoprenoids were released following the cleavage of ether bonds with HI and reductive dehalogenation with Zn in glacial acetic acid. The extracts were supplemented with nonadecanoic acid methyl ester as an internal standard and subjected to GC/MS analysis under operating conditions described by Gattinger et al. (2003).

DNA extraction and PCR amplification of 16S rRNA genes from soil samples

DNA was extracted in triplicate from 0.5 to 1.0 g of well-homogenized soil material using the UltraClean Soil DNA Isolation Kit (MoBio Laboratories Inc.), following the manufacturer's instructions. DNA triplicates were pooled and 16S rRNA gene fragments were generated using PCR with the general primer pair GC_341F-907R for Bacteria (Muyzer et al., 1993, 1995). A 50 μL reaction mixture contained PCR buffer, 0.25 mM of each dNTP, 2 mM MgCl2, 0.4 μM of each primer, 1.25 U Taq DNA Polymerase (Eppendorf) and ∼20 ng of DNA template. PCR was performed using an iCycler Thermal Cycler (Bio-Rad). The amplification conditions consisted of an initial denaturation step at 94 °C for 4 min, followed by 36 cycles of 94 °C for 1 min, 55 °C for 1 min and 72 °C for 1:30 min, with a final elongation step of 20 min at 72 °C. PCR products were checked on 2% agarose gels stained with SYBR Gold (Molecular Probes, Invitrogen).

DGGE and sequencing of environmental DNA bands

All samples were separated on 8% polyacrylamide gels in 1 × TAE buffer using a D-Code System (Bio-Rad). The denaturing gradient ranged from 20% to 60% (100% denaturant consisted of 7 M urea and 40% (v/v) deionized formamide). The gels were run at 60 °C at a constant voltage of 100 V for 14 h. After electrophoresis, the gels were stained for 35 min with SYBR Gold (1 : 10 000 diluted) and visualized under UV light using a GeneFlash system (Syngene). DGGE runs were repeated to ensure an identical banding pattern for further processing. For sequencing of DGGE bands, selected DGGE bands were cut out with a sterile scalpel and were transferred to sterile 0.5 mL Eppendorf tubes. DNA was eluted overnight in 30 μL of sterile water at 4 °C. The DNA fragments were then amplified for sequencing using the same conditions as described above, except using the forward primer without a GC clamp. After purification, using a QIAquick PCR Purification Kit (Qiagen), DNA bands were commercially sequenced by Eurofins MWG GmbH (Martinsried, Germany) and GATC Biotech (Konstanz, Germany) using the forward primer without a GC clamp.

Phylogenetic analysis

Phylogenetic analysis of partial 16S rRNA gene sequences (≥150 nucleotides) was performed using the arb software package (Ludwig et al., 2004; The automatic alignment tool of the SILVA database (Pruesse et al., 2007; was used for automatic sequence alignment and alignments were then corrected manually. Using the neighbor-joining method (Saitou & Nei, 1987) of the arb program, distance matrices were created to group the sequences into operational taxonomic units (OTUs) using the program dotur v. 1.53 (Schloss & Handelsman, 2005). One OTU is described as a cluster of DNA sequences that share ≥97% sequence similarity (Roselló-Mora & Amann, 2001). To assess the affiliation of the dominant soil bacteria, the RDP classifier with a confidence threshold of 95% was used (Wang et al., 2007).

Statistical analysis of data

For data processing, OTU numbers and environmental properties of the different depths of each sampling position were grouped to make the different sites comparable. The OTU numbers were normalized by log transformation, whereas the environmental parameters were normalized using a standard normal distribution (zero mean and unit variance) according to the following equation:


where z′ is the normalized value, x is the parameter that was measured, inline image is the mean and S is the SD (Leyer & Wesche, 2007). A matrix was compiled with the transformed data for both OTUs and environmental properties. The matrices were then processed to perform a canonical correspondence analysis (CCA) using the program canoco v. 4.5 (ter Braak, 1986). This analysis can provide us with information regarding the way in which the community is influenced and shaped by certain environmental parameters.

For the pairwise comparison of the dominant community structure of the different soils, the Sørensen index (S) was calculated according to the following equation:


where a and b are the number of OTUs obtained from the DGGE profiles of samples a and b, and ab is characterized by the number of OTUs shared by both samples.


Soil characteristics

The soil characteristics of nine soil profiles (Table 1) of Livingston Island (maritime Antarctica) were analyzed with regard to moisture, grain size fractions, TC and TN content, pH, conductivity (EC) and the number of culturable heterotrophs (see Table 2). Most of the investigated mineral soils showed a sandy texture interspersed with gravel. Soil profiles SP-B, SP-C and SP-D were dominated by silt from a depth of ≥10 cm. Clay content was clearly under-represented in most of the soil profiles. Only profiles T1-2, T1-3, T1-5 (40–50 cm) and SP-D (10–20 cm) showed elevated clay fraction values (10–20%). Profiles T1-1 and T1-4 were covered by mosses and showed an incipient humus accumulation in the uppermost part. The TC and TN contents were low within most of the soil profiles, with values ranging from<0.10% to 1.36% (TC) and <0.10% to 0.13% (TN), respectively, except for the two uppermost layers of the profiles T1-1 and T1-4 due to the moss layer. The values for pH ranged from neutral to alkaline, except for profile T1-1 and the top layer of soil profile T1-4, where pH values ranged from pH 4.8 to 6.1. Conductivity was low in all the soil samples. The moisture content of the investigated mineral soils ranged from 2.6% to 15.6%, with differences depending on their exposure and site characteristics. Profiles SP-A and SP-B were characterized by underlain permafrost, whereas for the other profiles, no frozen ground was observed.

Table 2.   Soil geochemical and soil geophysical descriptions of the sampling sites
SiteDepth (cm)Moisture (%)Sand (%)*Silt (%)*Clay (%)*Total C (%)Total N (%)pHEC (μS cm−1)Number of culturable heterotrophs (g−1 dry soil)
  • *

    Part of the grain size fraction <2 mm.

  • ND, no data.

T 1-10–47.1NDNDND26.500.844.81ND1.2 × 108
4–74.773.920. × 105
7–145.556.<0.106.10407.8 × 104
T 1-20–53.344.136.419.50.11<0.107.22387.2 × 104
5–124.159.625.714.7<0.10<0.107.80345.5 × 104
12–243.949.131.419.5<0.10<0.107.82374.7 × 104
T 1-30–93.660.728.410.90.13<0.107.58492.3 × 104
9–166.549.640.010.40.11<0.107.16344.1 × 104
16–233.673.818.08.2<0.10<0.107.92311.7 × 104
23–308.350.538.011.5<0.10<0.107.23311.3 × 104
T 1-40–510.0NDNDND9.340.325.97ND1.7 × 106
5–89.275.415.78.92.310.196.771125.4 × 103
8–135.479.316.44.30.28<0.107.53485.2 × 104
13–185.983.912.53.60.14<0.107.83392.5 × 104
T 1-50–103.388.410.31.3<0.10<0.107.81305.0 × 104
10–203.886.312.11.6<0.10<0.108.01281.9 × 104
20–304.581.816.71.5<0.10<0.108.16308.7 × 103
30–405.084.514.31.2<0.10<0.107.94302.3 × 103
40–503.145.836.317.90.23<0.108.451004.3 × 103
SP-A0–56.680.115.84.0<0.10<0.107.71361.0 × 104
5–107.081.615.03.4<0.10<0.108.04312.0 × 104
10–157.489.78.32.0<0.10<0.108.24301.5 × 104
15–209.992.46.41.2<0.10<0.108.61391.9 × 104
20–2515.691.67.01.4<0.10<0.108.49571.8 × 104
SP-B0–52.693.<0.107.48421.6 × 103
5–109.957.539.53.00.19<0.107.69347.4 × 102
10–157.846.750.23.10.11<0.107.83343.1 × 102
15–2010.341.356.42.4<0.10<0.107.78331.9 × 103
20–256.431.464.14.60.10<0.107.84373.9 × 103
30–3510.336.959.63.6<0.10<0.107.91371.5 × 101
SP-C0–102.769.<0.107.07532.2 × 104
10–205.641.350.87.90.38<0.107.48442.0 × 104
20–305.345.<0.107.49351.1 × 104
SP-D0–54.575.618.95.50.45<0.107.26658.4 × 104
5–105.171.920.18.10.48<0.107.41745.0 × 104
10– × 104
15–205.442.047.310.71.360.137.241014.5 × 103

In about half of the investigated mineral soil profiles (T1-1, T1-4, SP-A, SP-B, SP-D), the number of culturable heterotrophic bacteria decreased with increasing depth, whereas in T1-2, T1-3, T1-5 and SP-C, the numbers of culturable heterotrophs were relatively consistent throughout the entire soil profiles (Table 2). The highest numbers of culturable heterotrophic bacteria were observed in the surface layer of profiles T1-1 and T1-4 (1.4 × 108 CFU g−1 dry soil and 3.0 × 106 CFU g−1 dry soil, respectively).

Lipid biomarker analyses

The overall PLFA content decreased with increasing depth in almost all the profiles studied (Table 3). Only profiles SP-C and SP-D showed high PLFA concentrations for the bottom samples. The highest values were obtained in profiles T1-1, T1-4 and T1-5, respectively. While most top layers showed a predominance of unsaturated fatty acids, all the other samples were dominated by straight-chain saturated fatty acids (Table 3). For most profiles, the even-numbered compounds tetradecanoic, hexadecanoic and octadecanoic acid (14:0, 16:0 and 18:0) were constantly present throughout the profile (Supporting Information, Table S1). Branched-chain fatty acids could be detected throughout the entire profiles T1-1, SP-C and SP-D and within the uppermost layer of the profiles SP-A and SP-B. For deeper soil layers of profiles SP-A and SP-B as well as for profiles T1-2 and T1-5, no or only low concentrations of branched-chain fatty acids were detectable. PLEL characteristic for Archaea could not be detected in any of the samples.

Table 3.   Fractions of straight-chain saturated/unsaturated and iso-/anteiso-branched-chain fatty acids (FA) from the PLFA profiles of selected Antarctic mineral soil samples (per gram wet soil)
Site locationDepth (cm)Straight-chain saturated FAs (ng FA g−1)*Straight-chain monounsaturated FAs (ng FA g−1)Iso-branched-chain FAs (ng FA g−1)Anteiso-branched-chain FAs (ng FA g−1)§
  • *

    10:0, 12:0, 13:0, 14:0, 15:0, 16:0, 17:0, 18:0.

  • 16:1cis7, 16:1cis8, 16:1cis9, 16:1cis11, 17:1cis9, 17:1cis11, 18:1cis7, 18:1cis9, 18:1cis11.

  • 13:0 iso, 14:0 iso, 15:0 iso, 16:0 iso, 17:0 iso.

  • §

    § 13:0 anteiso, 15:0 anteiso, 17:0 anteiso.

  • ND, no data.


DGGE analysis of nine mineral soil profiles

Nine Antarctic mineral soil profiles from Livingston Island were investigated using DGGE fingerprints to classify the prominent bacterial representatives existing in these habitats. The DGGE banding pattern showed a heterogeneous and diverse picture within the vertical profiles and between the different sites (Fig. 3). Distinct DNA bands were detectable (up to 15), even in soil layers with very low contents of carbon and nitrogen. Within one profile (SP-D), the pattern of DNA bands showed almost no change with depth. Within the other profiles, a clear vertical difference in the number and position of DNA bands could be observed. Even for neighboring soil layers with similar soil chemical and soil physical properties (see Table 2), for example profiles T1-5 (from 0 to 40 cm), SP-A and SP-B, different DNA banding patterns in terms of number and position of the DNA bands could be observed. In profile SP-C, the number of DNA bands decreased with increasing depth.

Figure 3.

 DGGE fingerprints of nine different maritime Antarctic soil profiles. Circles indicate DNA bands that were cut for sequencing and that were subsequently used for phylogenetic analyses. Numbers in the DGGE picture indicate the depth in cm. (a) Soil profiles T1-1 to T1-5; (b) soil profiles SP-A to SP-D.

Phylogenetic analyses of dominant soil bacteria

To investigate the microbial life in mineral soils of Livingston Island, altogether 292 DNA bands of different intensities were cut out from the DGGE gels and were reamplified for sequencing. In total, 252 sequences of ≥150 bp could be obtained from the nine soil profiles. Sequences from mineral soils were affiliated to Alpha-, Beta- and Gammaproteobacteria, Bacteroidetes, Planctomycetes, Gemmatimonadetes, Acidobacteria, Nitrospirae, Firmicutes, Actinobacteria, Chloroflexi and unclassified bacteria (Table 4). Sequences affiliated with Bacteroidetes were detectable in almost all soil profiles (except profile T1-2), displaying the major group in profiles T1-1 and T1-4 (Fig. 4). Sequences related to the phylum Acidobacteria were also present in all investigated soil profiles and were most prominent in profiles T1-2, SP-C and SP-D. A much more diverse affiliation of sequences belonging to different bacterial phyla could be observed for the other soil profiles (T1-3, T1-5, SP-A, SP-B). Interestingly, sequences related to methanotrophic bacteria were found in different soil profiles, although no methane (CH4) emission could be detected at the sampling sites (data not shown) and no archaeal PLEL markers were found. A sequence related to the genus Methylocystis was found in only one mineral soil profile (T1-5), whereas Methylobacter psychrophilus sequences could be found in four different profiles (T1-2, T1-3, T1-5, SP-A). Altogether, 114 OTUs could be defined. Most OTUs obtained from the DGGE patterns belong to the Bacteroidetes (36 OTUs), Acidobacteria (29 OTUs) and unclassified bacteria (20 OTUs), followed by Firmicutes (10 OTUs), Actinobacteria (5 OTUs) and Betaproteobacteria (5 OTUs). The remaining groups were represented to a much smaller extent, with two OTUs for the Gammaproteobacteria, Gemmatimonadetes and Chloroflexi, and with one representative for the Alphaproteobacteria, Planctomycetes and Nitrospirae, respectively. Members of the Bacteroidetes, which represented the largest group of OTUs, were mostly assigned to the class Sphingobacteria, with most sequences clustering in the family Chitinophagaceae. OTUs of the Acidobacteria could be grouped into five different subdivisions (Gp4, 6, 7, 13, 17), in which 66% of the Acidobacteria were affiliated to subdivision Gp4. For the unclassified bacteria, OTUs can be assigned distantly to Acidobacteria, Actinobacteria, Bacteroidetes, Chloroflexi, Firmicutes, group OD1 and Proteobacteria. The Firmicutes sequences obtained were related to the families Lactobacillaceae, Staphylococcaceae, Streptococcaceae and Clostridiaceae.

Table 4.   Closest phylogenetic affiliation of sequences obtained from DGGE bands of each single soil profile
Closest phylogenetic affiliationT1-1T1-2T1-3T1-4T1-5SP-ASP-BSP-CSP-D
  1. Numbers in parentheses indicate the number of OTUs.

Alphaproteobacteria    1 (1)1 (1)   
 Rhizobiales     1   
 Methylocystis    1    
Betaproteobacteria    2 (2) 3 (3)3 (2) 
 Polaromonas      1  
 Comamonadaceae      1  
 Burkholderia       2 
 Burkholderiales       1 
 Oxalobacteraceae    1    
 Unclassified Betaproteobacteria    1 1  
Gammaproteobacteria 1 (1)6 (2) 1 (1)4 (1)   
 Methylobacter 16 14   
Bacteroidetes21 (9) 4 (3)14 (12)8 (3)13 (7)8 (4)2 (2)1 (1)
 Chitinophagaceae21 3478421
 Chitinophaga    1    
 Sphingobacteriales  11     
 Unclassified Bacteroidetes   6  4  
 Ferruginibacter   2 3   
 Terrimonas   1     
 Tannerella   1 2   
Planctomycetes     2 (1)   
 Planctomycetaceae     2   
Gemmatimonadetes 2 (2)  2 (1)   1 (1)
 Gemmatimonas 2  2   1
Acidobacteria1 (1)17 (8)9 (4)7 (3)13 (5)2 (1)9 (7)8 (3)24 (7)
 Group 419571226724
 Group 6 44   2  
 Group 7 3  1 1  
 Group 13       1 
 Group 17 1       
Nitrospirae   1 (1)  1 (1)  
 Nitrospiraceae   1  1  
Firmicutes  1 (1) 9 (4)4 (2)7 (2)  
 Lactobacillus    4    
 Bacillales    3    
 Staphylococcus  1   5  
 Staphylococcaceae      1  
 Clostridiaceae    1    
 Unclassified Firmicutes    1 1  
 Streptococcus     4   
Actinobacteria  6 (3)  5 (3)1 (1)1 (1) 
 Propionibacterium  5  311 
 Corynebacterium  1      
 Acidimicrobiales     1   
 Unclassified Actinobacteria     1   
Chloroflexi   1 (1)1 (1) 1 (1)  
 Anaerolineaceae   11 1  
Unclassified bacteria1 (1) 1 (1)2 (2)4 (4)3 (3)10 (7)1 (1)1 (1)
Total number of sequences (OTUs)23 (11)20 (11)27 (14)26 (19)41 (22)34 (19)40 (26)15 (9)27 (10)
Figure 4.

 Percentage abundance of the dominant community 16S rRNA gene sequences belonging to each taxonomic group for each individual soil profile, revealed by DGGE analysis.

Statistical analyses

Overall, the Sørensen similarity index was low for all investigated sites, with values of S≤0.4, even for sites where the environmental parameters were comparable (Table 5).

Table 5.   Comparison of the dominant soil community OTUs (with a sequence similarity ≥97%) using the Sørensen index S
  1. The numbers in parentheses indicate the sequence similarity at the genus level (95%).

T1-20.00 (0.13)1       
T1-30.08 (0.21)0.16 (0.19)1      
T1-40.00 (0.26)0.07 (0.08)0.13 (0.14)1     
T1-50.06 (0.15)0.18 (0.28)0.17 (0.19)0.15 (0.22)1    
SP-A0.07 (0.16)0.13 (0.15)0.30 (0.27)0.11 (0.18)0.20 (0.21)1   
SP-B0.11 (0.13)0.16 (0.19)0.25 (0.29)0.18 (0.26)0.13 (0.14)0.13 (0.20)1  
SP-C0.20 (0.27)0.10 (0.12)0.35 (0.40)0.15 (0.17)0.13 (0.14)0.21 (0.23)0.23 (0.26)1 
SP-D0.00 (0.15)0.29 (0.27)0.17 (0.33)0.07 (0.09)0.19 (0.15)0.07 (0.08)0.06 (0.07)0.21 (0.14)1

The CCA plot shows clusters of the different sites as well as the weights and orientation of the environmental factors (Fig. 5). The first two axes accounted for 18.3% (PC1) and 15.4% (PC2) of the variation. TC, TN and pH were correlated with PC1, whereas the other factors [moisture, electric conductivity (EC), sand, silt, clay] were correlated with PC2. Soil profile T1-1 is strongly correlated with TC and TN and the soil profiles T1-2, T1-5, SP-C and SP-D were correlated with clay. The other soil profiles (T1-3, T1-4, SP-A and SP-B) are mainly correlated with sand and moisture.

Figure 5.

 CCA plot of the investigated soil profiles. Circles indicate the different sampling sites in relation to different soil physical and soil chemical parameters (T1-1 to T1-5; SP-A to SP-D). Arrows indicate the different weights and orientations of the investigated soil factors in the CCA plot.


Although the Antarctic continent is characterized by extreme climate conditions, recent studies in Antarctic soil environments revealed a large bacterial diversity, with a high proportion of sequences that were only distantly related to known sequences, indicating the existence of novel species or genera (Shivaji et al., 2004; Saul et al., 2005; Aislabie et al., 2006, 2008; Smith et al., 2006; Shravage et al., 2007; Yergeau et al., 2007; Niederberger et al., 2008). Within this study, we identified the dominant bacterial groups in a broad variety of different soils from Livingston Island (maritime Antarctica) and identified the soil parameters affecting the community structure of dominant bacteria existing in these habitats.

Antarctic mineral terrestrial habitats were mostly poorly developed and initial processes in soil formation such as humus accumulation, acidification and brownification could be observed only at a few sites. The different soil profiles revealed a highly heterogeneous picture with regard to bacterial composition and site-specific properties (e.g. water and carbon content, grain size distribution). Water content was generally much higher than in mineral soils from continental Antarctica, for instance the Larsemann Hills, Prydz Bay (unpublished data), Victoria Land (Aislabie et al., 2006) or the Dry Valleys (Smith et al., 2006). The higher moisture results from higher precipitation rates due to the frontal systems bringing warmer and moister air to the western part of the Antarctic Peninsula. Temperatures can change rapidly in soils or moss beds (up to 42 °C), forcing the microorganisms to design adaptation mechanisms for a survival over a wide temperature gradient (Lewis-Smith, 1986; Bölter et al., 1989). The input, quantity and availability of organic compounds derived by plants, animals or autotrophic microorganisms are of importance for establishing a microbial food web. Thus, habitat conditions can vary on a micrometer scale, leading to distinct changes in bacterial community structures. However, higher carbon values were only found in two sites on Livingston Island, coinciding with a moss cover and high numbers of culturable heterotrophs. In general, cell numbers were comparable with the plate counts obtained from other Antarctic sites, ranging from 1.7 × 103 to 6.1 × 107 g−1 dry soil (Ramsay & Stannard, 1986; Saul et al., 2005; Aislabie et al., 2006, 2008).

In contrast to previous analyses of soil samples from other cold terrestrial habitats such as Siberian permafrost (Wagner et al., 2005), no increase in the ratio of anteiso- to iso-fatty acids and a lack of unsaturated fatty acids for most of the samples were observed. Samples with a prevalence of unsaturated oversaturated fatty acids were restricted to several surface samples (T1-4 and SP-profiles). Adaptation mechanisms other than unsaturation that maintain the fluidity of the cell membrane can be changes in lipid polar head groups, changes in the protein content, changes in the carotenoid composition of the membranes and changes of the length of the fatty acids (Chintalapati et al., 2004; Mangelsdorf et al., 2009). The composition of the PLFA profiles suggested the dominance of a bacterial community; marker lipids for eukarya, such as linoleic acid (18 : 2 cis9,12), were detected in significant amounts only in the top layer of profiles T1-4, T1-5 and SP-B (Table S1), but were absent in deeper layers. Other eukaryotic polyunsaturated fatty acids such as linolenic acid were not found in any sample. Marker lipids for several bacterial taxa indicated the presence and spatial allocation of these groups. Tuberculostearic acid (18:0 10-methyl) is characteristic for several genera within the Actinobacteria and was found in the top layers of profiles T1-1 and T1-4.

DGGE patterns of the investigated soils showed a highly diverse bacterial community in most of the soil profiles. The sequences were related to Acidobacteria, Actinobacteria, Bacteroidetes, Chloroflexi, Firmicutes, Gemmatimonadetes, Planctomycetes and Proteobacteria, revealing an establishment of a complex microbial community structure in different mineral soils under maritime Antarctic climate conditions. Most of these bacterial phyla, recently identified to be predominant in soils (Janssen, 2006), were detected with a high proportion of sequences that could not be affiliated to known genera. This is consistent with recent studies on other Antarctic soil environments (Shivaji et al., 2004; Saul et al., 2005; Aislabie et al., 2006, 2008; Smith et al., 2006; Shravage et al., 2007). Yergeau et al. (2007) found all nine of the proclaimed dominant soil phyla in soil samples spanning a latitudinal transect from the Falkland Islands (South Atlantic) to the Ellsworth Mountains (West Antarctica). The majority of the sequences they detected were affiliated to Acidobacteria, Actinobacteria, Bacteroidetes and Proteobacteria. Niederberger et al. (2008) were also able to detect the nine ‘soil’ phyla, with a dominance of Acidobacteria (up to 50% of the clone library) and Actinobacteria in all samples.

Soil texture and mineral composition as well as nutrient content and nutrient quality can influence the microbial colonization of raw mineral soils. It is known that these soil characteristics can affect the metabolic activity of soil microorganisms and thus the microbial community (van Loosdrecht et al., 1990; England et al., 1993; Costerton et al., 1995). Most of the sequences obtained from initial soils of Livingston Island belonged to the diverse phyla of the Acidobacteria and Bacteroidetes. Representatives of these groups were found in numerous terrestrial Antarctic habitats (Shivaji et al., 1992; Bowman et al., 1997; Brinkmeyer et al., 2003; Abell & Bowman, 2005; Yergeau et al., 2007; Niederberger et al., 2008; Foong et al., 2010). On Livingston Island, DGGE sequences related to Acidobacteria were restricted to soils with low carbon and nitrogen contents and neutral to alkaline pH values, whereas DGGE sequences related to Bacteroidetes could be found in almost all of the investigated profiles with a dominance in the moss-covered sites, where the pH is more acidic. Possibly, the input of organic compounds, provided by bryophytes (Bölter et al., 2002), leads to the dominance of Bacteroidetes in the vegetated soils. This would be in accordance to Fierer et al. (2007), who found that additional carbon led to a positive correlation between the Bacteroidetes abundances and the carbon mineralization rates, while the abundance of Acidobacteria was negatively correlated with carbon mineralization rates. For the Bacteroidetes, this implies an important role in bryophyte-influenced Antarctic soils by degrading polymeric organic matter and providing low molecular substances for the microbial food web. Sequences belonging to Bacteroidetes were also found to make up an important part of the bacterial community in other cold habitats such as Sphagnum-covered permafrost-affected soils in Siberia (Pankratov et al., 2006; Liebner et al., 2008; Wagner et al., 2009). However, a comparison of these data with those obtained by PLFA analyses shows that the characteristic iso-/anteiso fatty acids of the Bacteroidetes phylum did not dominate the lipid profiles of the samples. Therefore, this group may not represent the dominating bacterial population of the samples analyzed. This could be due to the different lysis efficiencies of bacterial cells during the DNA extraction, which could therefore lead to a preferential amplification of certain groups. Members of the phylum Acidobacteria, in which few isolates have been described so far (Kishimoto et al., 1991; Eichorst et al., 2007; Koch et al., 2008; Kulichevskaya et al., 2010), are widely distributed in a variety of soils (Barns et al., 1999; Janssen, 2006) and other habitats, for example wastewater, sediment or hot springs (reviewed in Eichorst et al., 2007). Sequences related to Acidobacteria were also detected in Antarctic cold-desert soils (Aislabie et al., 2006, 2008; Smith et al., 2006; Niederberger et al., 2008). They were exposed to similar geochemical conditions (low C and N values, neutral to alkaline pH) as in the bare maritime mineral soils of this study, where Acidobacteria represent a major portion in the overall community. Therefore, it is conceivable that members of the Acidobacteria play an important role in nutrient turnover in these cold, oligotrophic soils. Although somewhat speculative, it is possible that Acidobacteria are involved in a broad range of metabolic pathways, as they are phylogenetically highly diverse (Barns et al., 2007), suggesting a large physiological potential. However, their relevance and function in the ecosystem is still unknown, even though the abundance can make up to 50% in soils (Dunbar et al., 2002; Lipson & Schmidt, 2004; Niederberger et al., 2008). Because only a few representatives of the Acidobacteria had been isolated so far, the lipid data available for this group are limited and therefore, interpretation of the PLFA profiles obtained with respect to the abundance of this group is not possible.

Additionally, a distinct number of OTUs, with up to 25% of the dominant bacterial phylotypes in the single soil profiles, could only be assigned to unclassified bacteria, leading to the suggestion that these cold soil habitats harbor a yet undescribed part of the microbial community with currently unknown physiological and ecological functions.

In four of the oligotrophic soil profiles (T1-2, T1-3, T1-5, SP-A), we were able to detect sequences that were affiliated to Methylobacter psychrophilus and Methylocystis sp., previously known from Arctic tundra wetlands (Omelchenko et al., 1996; Liebner et al., 2009). Accompanying CH4 flux measurements at the same sites did not reveal any CH4 emission, not even after intense rain, whereas Gregorich et al. (2006) reported a noticeable emission from lakeshore soils in an Antarctic dry valley. Lacking CH4 emission from Livingston Island soil profiles could be firstly explained by complete oxidation of produced CH4 through methanotrophic activity indicated by the presence of sequences related to CH4-oxidizing bacteria or secondly methanogenic archaea might not be particularly abundant or even absent as a result of the aerobic conditions in most parts of the mineral soils, which are unfavorable for an anaerobic metabolism. This would be in accordance to our results of PLEL determination, where no archaeal biomarkers were detectable. The detection and description of methanotrophic bacteria that are able to use atmospheric CH4 was shown in several studies (Holmes et al., 1999; Bull et al., 2000; Henckel et al., 2000; Kolb et al., 2005). Therein, sequences related to the genera Methylobacter and Methylocystis, as observed in our data, were found in very different soil types (tropical and temperate forest soils). An indication for the existence of such methanotrophs could be the negative average flux of CH4 as it was observed for several nonornithogenic soils from maritime Antarctica (Sun et al., 2002; Zhu & Sun, 2005). In contrast, ornithogenic soils, which are characterized by bird excrement (penguin guano), showed a positive CH4 net flux comparable to values similar to mesic sites from northern high latitudes (>55°N; Bartlett et al., 1992; Christensen et al., 1995; Nakano et al., 2000; Wille et al., 2008). Further measurements as well as investigations regarding the activity and structure of methanotrophic communities could answer the question as to whether the southern ice-free polar regions, including the subpolar islands, can act as a sink of atmospheric CH4.

The statistical analysis of the dominant soil phylotypes and the soil parameters determined showed a strong influence of the nutrient content on the community structure of the soil profile T1-1, with an obvious favoring of Bacteroidetes in moss-covered, acidic soils. In the other soil profiles, soil-physical parameters (clay, sand, moisture) became important for the community structure. Interestingly, pH is not positively correlated with the TC and TN content, although it is known to have an influence on the composition, conservation and availability of soil organic and inorganic components (Kemmitt et al., 2006). This might be due to the low carbon and nitrogen values that favor a specific soil microbial community, which is not regulated by a large nutrient input derived from plants. In future scenarios, the predicted increase of the mean annual air temperatures in the maritime Antarctic could result in an increase in plant coverage together with a higher nutrient input and could therefore possibly lead to a change in the soil microbial community shifting from an oligotrophic, Acidobacteria-dominated or more diverse bacterial community to a less diverse microbial collective dominated by members of copiotrophic bacteria, for example representatives of the Bacteroidetes phyla.


We wish to thank all Bulgarian, Spanish and German colleagues for supporting fieldwork and logistics during the expedition in 2005, with special thanks to Christo Pimpirev for leading the expedition. Special thanks are due to Susanne Liebner (University of Tromsø) for helpful comments and critical reading of the manuscript as well as to Francisco Fernandoy (Alfred Wegener Institute for Polar and Marine Research) for drawing of the map. We highly appreciate the help and technical assistance of Quynh Wang-Lieu, Friederike Bruns, Felizitas Bajerski, Ute Bastian and Daniel Niggemann. We also wish to thank three anonymous reviewers for their comments. This work was supported by the Deutsche Forschungsgemeinschaft (DFG) in the framework of the priority program ‘Antarctic Research with Comparative Investigations in Arctic Ice Areas’ by a grant to D.W. (WA 1554/4) and A.L. (LI1624/2).


GenBank accession numbers: HQ153156–HQ153407