Change of microbial communities in glaciers along a transition of air masses in western China

Authors

  • Shu-Rong Xiang,

    1. Key Laboratory of Tibetan Environment Changes and Land Surface Processes, Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing, China
    2. State Key Laboratory of Cryospheric Science, Cold and Arid Regions Environmental Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, China
    Search for more papers by this author
  • Yong Chen,

    1. School of Life Sciences, Lanzhou University, Lanzhou, China
    Search for more papers by this author
  • Tian-Cui Shang,

    1. Chemistry and Biology College, Yili Normal University, Yining, China
    Search for more papers by this author
  • Ze-Fan Jing,

    1. State Key Laboratory of Cryospheric Science, Cold and Arid Regions Environmental Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, China
    Search for more papers by this author
  • Guangjian Wu

    1. Key Laboratory of Tibetan Environment Changes and Land Surface Processes, Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing, China
    2. State Key Laboratory of Cryospheric Science, Cold and Arid Regions Environmental Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, China
    Search for more papers by this author

Abstract

[1] Microbial community dynamics across glaciers in different climatic zones provide important information about the sources, transportation pathways, and deposition of microorganisms. To better understand the possible driving forces of microbial community shifts in glacier ice at a large spatial scale, 16S rRNA gene amplification was used to establish clone libraries containing 95 bacterial sequences from three different habitats in the Qiangyong Gacier in 2005. The libraries were used in phylogenetic comparison with 149 previously reported sequences from the surface samples collected from the Kuytun 51, and East Rongbuk glaciers in the same year. The results showed the presence of cosmopolitan and endemic species, and displayed a tendency of zonal distribution of bacterial communities at genera and community levels, corresponding to the geographic placement of the three glaciers. Data also showed a significant difference in the proportion of dominant phylogenetic groups in the three glaciers. Comamonadaceae/Polaromonas (Betaproteobacteria) and Flexibacteraceae (Bacteroidetes) were dominant in the Qiangyong Glacier, Cyanobacteria, Comamonadaceae/Polaromonas, and Rhodoferax (Betaproteobacteria) were dominant in the Kuytun 51 Glacier, and Acinetobacteria (Gammaproteobacteria) were dominant in the Rongbuk Glacier. In conclusion, the current study provides evidence of microbial biogeography in glacier ice at both the fine lineage and whole community levels. The biogeographical patterns were generally associated with the hydrological transition over the glaciers in the northern periphery and southern part of the Tibetan plateau. This supports our hypothesis of air mass behavior being one of the main drivers determining the zonal distribution of microbial communities across the mountain glaciers in western China.

1. Introduction

[2] Microbial community dynamics across glaciers in different climatic zones reflect microbial responses to global climatic and environmental variation on a large geographic scale, and thus provide important information about the sources, transportation pathway and deposition processes leading to the presence of microorganisms in the glacier ice. Earlier culture-based analyses of ice cores showed differences in taxonomical composition of bacterial isolates in spatially separated polar and nonpolar glaciers [Christner et al., 2000]. Investigations of bacterial diversity and bacterial density in the subsurface snow from four glaciers on the Tibetan plateau (Guoqu, Zadang, East Rongbuk and Palong number 4) recently revealed differences in the proportion of dominant bacterial phyla at the high altitude [Liu et al., 2009a, 2009b]. Phylogenetic analysis of a Greenland glacier indicated the predominance of bacteria from the phyla Proteobacteria, Firmicutes, and Actinobacteria, in contrast to glaciers on the Tibetan Plateau where Firmicutes were less frequently reported [Miteva et al., 2009]. These results emphasize the microbial ecological theory “everything is everywhere, the environment selects” [Beijerinck, 1913]. Despite the considerable diverse microorganisms in glacier ice, there is currently no direct evidence of geographic patterns of microorganisms in glacier ice at a whole community level on a large spatial scale. Our most recent preliminary statistical analyses showed a significant difference in bacterial community composition between the Kuytun 51 Glacier and Rongbuk Glacier [Xiang et al., 2009b]. The community shift between the Kuytun 51 Glacier and Rongbuk Glacier may reflect a true biogeographical distribution of organisms in glacier ice, or imply an important clue to microbial sources via atmospheric circulation over the glaciers. To increase our knowledge of biogeography of microorganisms, it is necessary to analyze the already available sequence data sources, such as those deposited GenBank (http://www.ncbi.nlm.nih.gov/), investigate the geographic patterns of microbial communities in glacier ice, and relate them to the changes of global climates and environments.

[3] The mountain glaciers in northwestern and southwestern China can be categorized into 3 types: temperate maritime, extremely cold continental (polar) and cold-based subcontinental glaciers [Shi and Liu, 2000]. Previous investigations have showed that spatial variations in chemical components in precipitation from these mountain glaciers are controlled primarily by two air mass patterns (Figure 1). One is westerly, in which the dry and cold hydrological flux derived from the arid, and semiarid regions of central Asia plays a dominant role in the northern periphery of the Tibetan plateau throughout the year [Wake et al., 1993; Shi and Liu, 2000], as well as in southern Tibetan plateau in winter [Wake et al., 1993]. The other is the summer monsoon, in which the moist and warm hydrological influx from the Bay of Bengal is prevalent on the southern Tibetan plateau in summer [Wake et al., 1993; Shi and Liu, 2000]. The Qiangyong Glacier (90°13′E; 28°51′N) is located in the southeastern region of the Tibetan Plateau where precipitation is derived from monsoonal air masses during the summer, the neighboring depressions [Dregne, 1968; Chen and Bowler, 1986], and from westerly depressions during the winter [Murakami, 1987; Wake et al., 1993; Shi and Liu, 2000]. The East Rongbuk Glacier (28°01′N; 86°57′E) on the north slope of the Himalayas receives its precipitation mainly from the south Asia summer monsoon [Kang et al., 2004; Liu et al., 2007]. The Kuytun 51 Glacier (84°24′E; 43°43′N) is in the northern periphery of the Tibetan plateau, and located in the Tienshan Mountain regions that are surrounded by vast deserts: the Gobi desert to the east, the Taklimakan desert in the Tarim basin to the south, the Peski Muyunkum and Peski Sary-Ishikotrau deserts to the west, and the Gurbantunggut desert in the Junggar basin to the north [[Xiang et al., 2009b; Li et al., 2003]. Dust storms occur frequently in this region and around the arid and semiarid regions of central Asia during the spring [Middleton et al., 1986]. In addition, precipitation in the Tienshan Mountain regions is derived from westerly air masses throughout the year [Wake et al., 1993; Shi and Liu, 2000]. Such an apparent transition in precipitation across the three glaciers in western China could lead to a geographic pattern of microbial community like zonal distribution of macro organisms across a large geographic area. This provides a unique opportunity to investigate how the air mass pattern influences the distribution of microbial communities in the glaciers across western China.

Figure 1.

Map showing the locations of three glaciers discussed in the current study: Kuytun 51, Qiangyong, and Rongbuk.

[4] In the current study, our main objective is to investigate the geographic patterns of microorganisms across the mountain glaciers in western China, and relate them to the precipitation pattern over the glaciers. First, to investigate the biogeography of bacteria at fine lineage level, new bacterial 16S rRNA gene sequences from the Qiangyong Glacier were pooled together with those previously reported from the Kuytun 51 Glacier [Xiang et al., 2009b] and used for further phylogenetic analyses. Second, to analyze geographic patterns of microbial communities in glacier ice at a large spatial scale, the bacterial sequences from the Qiangyong Glaciers were pooled with those previously reported from the Kuytun 51 [Xiang et al., 2009b] and East Rongbuk Glacier [Liu et al., 2007]. All pooled sequences used were from the samples colleted from the three glaciers in 2005. The principal coordinate analysis (PCoA) and hierarchical clustering of the microbial communities across the three glaciers was plotted using the UniFrac technique [Lozupone and Knight, 2005]. Finally, we interpreted the patterns of taxonomic diversity across the mountain glaciers in the context of possible driving forces of microbial community shifts in glacier ice at a large spatial scale.

2. Methods

2.1. Sample Collection From the Qiangyong Glacier

[5] Due to differences in the density of fine brown dust and light intensity, and wind-blown effects [Takeuchi et al., 2005; Takeuchi and Li, 2008], the surface snow in the Qiangyong Glacier (90°13′E; 28°51′N) in September 2005 was not homogenous, and presented a mosaic of brown and white areas. Surface glacial samples (around 5–20 cm deep) were scraped with a sterilized stainless steel scoop from three randomly selected points in a 50 cm by 50 cm plot, in three habitats: the dry snow zone at 5710 m a.s.l., the firn zone around the equilibrium line at 5640 m a.s.l., and the ablation zone at 5610 m a.s.l. A composite sample of approximately 500 ml was prepared from the replicate points at each site. To avoid contamination, autoclaved gloves, mask and a special sterile suit from head to toe were worn and changed at each sampling site during the whole sampling operation. All the collected snow/ice samples were separately kept frozen in 500 ml sterile plastic containers (Nalgene) with screw-cap in a freezer (below −4°C during transportation) and transported back to the cold room (air temperature between −18°C and −24°C) at the State Key Laboratory of Cryospheric Science and Environment of Chinese Academy of Science in Lanzhou, China. Around 450 ml snow samples from the Qiangyong Glacier were slowly melted in the sterile plastic containers at 4°C, and used for analysis.

2.2. Bacterial Density Measurement and Clone Library Establishment of Bacterial 16S rRNA Gene Amplified From the Qiangyong Glacier

[6] Density of bacteria in the different habitats such as the dry snow, firn, and ablation zones in the Qiangyong Glacier was determined by flow cytometric (FCM) analysis. FCM, DNA extraction and clone library establishment of the bacteria 16S rRNA gene PCR (Polymerase Chain Reaction) products from the Qiangyong Glacier were conducted by following the protocols previously used in microbial analysis of the Kuytun 51 Glacier [Xiang et al., 2009b]. The 16S rRNA gene amplicons used for the establishment of clone libraries from the Qiangyong Glacier were generated by PCR amplification with the universal bacterial primer pair 8f (5′-AGAGTTTGATCATGGCTCAG) and 1492R (5′-CGGTTACCTTGTTACGACTT). A total of 100 clone sequences were obtained from the Qiangyong Glacier. A naming convention was followed for each sequence, using the initial of Qiangyong Glacier (QY), and sampling site elevation followed by the clone reading numbers (1 to 162). For example, clones QY5610–125, 5640–49, QY5710–19 were the representatives from the elevation 5610 at the ablation zone, 5640 at the firn zone and 5710 m at the snow zone in the Qiangyong Glacier, respectively. The accession numbers of the cloned sequences obtained from the Qiangyong Glacier in GenBank are: GU246731-GU246737, GU246739- GU246765, GU246767- GU246777, GU246779-GU246798, GU246800-GU246812, GU246814-GU246830.

2.3. Sequence Comparison of Bacteria Between the Qiangyong and Kuytun 51 Glaciers

[7] To investigate the biogeography of bacteria in ice from the Qiangyong and Kuytun 51 glaciers, all 95 sequences from the Qiangyong Glacier surface were pooled with the 112 sequences from the Kuytun 51 Glacier surface. Reference species were identified using BLAST [Altschul et al., 1990] and all sequences were aligned using ClustalX [Thompson et al., 1997]. A Neighbor-Joining (NJ [Tamura et al., 2004]) phylogeny for the aligned sequences was constructed using MEGA 4.0 [Tamura et al., 2007] (http://www.megasoftware.net/) with pairwise deletion mode for gaps and with the Maximum Composite Likelihood (MCL) method for substitutions. The 16S rDNA sequences from Methanosaeta harundinacea strain 8Ac (accession number AY817738) and Methanosaeta concilii strain NW-1 (accession number DQ150255) were used as outgroup references on all trees. Sequence identities were assigned based on >97% similarity to known species. All clones were related to recognized cultivated genera or recognized genus clones (e.g., Clavibacter sp., Frigoribacterium sp. and Cryobacterium sp.).

2.4. Statistical Analyses of Bacterial Communities Among Four Glaciers in Western China

[8] The diversity (Shannon-Wiener index H′) and evenness (E) indices of bacteria in the Qiangyong Glacier were based upon the distribution of unique sequence OTUs (operational taxonomic units) obtained from the clone libraries using equations: H′ = -SUM{pi*ln(pi)} and E = H/ln(S), respectively [Hill et al., 2003], where pi = the proportion of the ith clone in the total clones in each individual library, and S is single unique sequence richness. Bacterial phylotype richness and coverage estimators were calculated with the software program EstimateS (http://www.aslo.org/lomethods/free/2004/0114a.html [Kemp and Aller, 2004]).

[9] To further investigate the biogeographical patterns of bacterial communities across the mountain glaciers in western China at a large spatial scale, the representative bacterial sequences from the current study were compared to Kuytun 51 (with representative sequence accession numbers EU263676–EU263787 in the work of Xiang et al. [2009b]), and East Rongbuk (with accession numbers EF190114–EF190144, DQ675500, DQ675465, DQ675487, DQ675469, DQ675470 and DQ675471 in the work of Liu et al. [2007]) glaciers. Bacterial community structure was compared using the UniFrac software package [Lozupone and Knight, 2005].

3. Results

3.1. Difference in Bacterial Density and Diversity of Dominant Bacteria Between the Qiangyong and Kuytun 51 Glaciers

[10] Bacterial density and dominant bacterial diversity were investigated in the different habitats on the Qiangyong Glacier surface and compared with those previously reported in the Kuytun 51 Glacier. The results showed a great variability in the distribution of bacterial density and diversity not only in the geographically different glaciers but also in the different habitats across the glacier surface (Table 1). The maximum values of microbial density and diversity index in the Kuytun 51 Glacier were slightly higher than in the Qiangyong Glacier. Compared with the dry snow (snow accumulation zone) and ablation zone, the firn zone contained higher bacterial density and diversity (Shannon index) in both glaciers. Six clone library curves from the Kuytun 51 and Qiangyong glaciers approached or reached an asymptotic level (data not shown), indicating more complete coverage (with >80% coverage, Table 1), and allowing a direct comparison of the bacterial communities of the two glaciers at a fine linage level.

Table 1. Bacterial Density and Diversity as Assessed by 16S rRNA Gene Sequence Analysis
 Qiangyong GlacierKuytun 51 Glaciera
Dry Snow (5710 m Altitude)Firn (5640 m Altitude)Ablation (5610 m Altitude)Dry Snow (3725 m Altitude)Firn (3601 m Altitude)Ablation (3505 m Altitude)
Total cells (104 cells ml−1)24.01 (±6.25)79.14 (±12.53)37.44 (±8.12)16.96105.40 (±23.54)34.92 (±24.59)
Number of OTUs predicted (Schao1)423562944513
Coverage CACE (%)888889829296
Number of OTUs observed252947333012
Shannon index2.793.213.642.253.491.9
Eveness0.870.950.950.651.040.77

3.2. Comparison of the Main Bacterial Phylogenetic Groups Between the Kuytun 51 and Qiangyong Glaciers

[11] The results showed that bacteria in the Kuytun 51 and Qiangyong Glacier ice and snow were all related (with >88% similarity) to reported species (Figures 3a–3e), with Proteobacteria and Bacteroidetes dominating the communities in both glaciers (Figure 2). Comparative sequence analyses showed significant phylogenetic differences between the bacterial species between the Kuytun 51 and Qiangyong Glacier (Figures 2 and 3a3e). Cyanobacteria were common across the Kuytun 51 Glacier surface, but only found in the Qiangyong Glacier snow-ice at 5640 m a.s.l. (Figure 2). Moreover, 62% of the total bacteria from the Qiangyong Glaciers clustered together, and the bacteria from the Qiangyong Glacier were significantly different from those in the Kuytun 51 Glacier at species level (indicated by the gray shaded areas in Figures 3a3e). In the Comamonadaceae group within the betaproteobacterial family, the Qiangyong Glacier clones from the different habitats clustered together, while the Kuytun 51 Glacier clones more closely grouped together. For example, the Qiangyong Glacier clone QY5610–97 from the ablation zone at 5610 m a.s.l., clone QY5640–6 from the firn zone at 5640 m a.s.l., and clone QY5710–27 from the dry snow zone at 5710 m a.s.l. more closely clustered than those from the Kuytun 51 Glacier (indicated by the gray shaded areas in Figure 3a), whereas five Kuytun Glacier clones KuyT-ice-30, KuyT-ice-19, KuyT-IWPB-36, KuyT-IWPB-107, and KuyT-water-20 grouped in a separate individual branch. The same trend was apparent in the Janthinobacteria and Panaciterramonas group within the family members of Betaproteobacteria (indicated by the gray shaded areas in Figure 3b), Flexibacteraceae and Flavisolibacteria group within the phylum Bacteroidetes (Figure 3c), Frigoribacteria-Leifsonia and Cryobacteria group of the phylum Actinobacteria (indicated by the gray shaded areas in Figure 3d), and in the phylum Deinococcales and Cyanobacteria (in Figure 3e).

Figure 2.

Relative abundance of the main phylogenetic groups (bacterial phyla) based on BLAST results of 16S rRNA gene sequences in each of the clone libraries from the different habitats on the glacial surface of the Kuytun 51 (data adapted from Xiang et al. [2009b]) and Qiangyong Glacier. IWP, ice, meltwater, and particle associated sample from ablation zone of the Kuytun 51 Glacier.

Figure 3a.

Phylogenetic analysis of the 16S rRNA genes for Burkholderiales (Betaproteobacteria) and Epsilonproteobacteria clones from the Kuytun 51 and Qiangyong Glacier and the closest relatives. The tree was generated by the neighbor-joining method after sequence alignment, and rooted with two Methanosaeta harundinacea strains (accession numbers AY817738 and DQ150255). Bootstrap values (100 replications) were specified for each Node. Numbers of the obtained similar snow clones (had the same amplified ribosomal DNA restriction analysis (ARDRA) pattern to the sequenced representatives listed on the tree) and relative sequence affiliations corresponding to GenBank accession number were provided in parentheses. The typical endemic bacterial species were indicated in the gray shaded areas. Scale bar indicated 0.05 substitutions per nucleotide. Sequences from the Qiangyong and Kuytun 51 Glacier are noted in bold. See a detailed description of the assigned sequence references and numbers in section 2.

Figure 3b.

Phylogenetic analysis of the 16S rRNA genes for the Beta-, Gamma-, Delta-, and Alpha-proteobacteria clones from the Kuytun 51 and Qiangyong Glacier and the closest relatives. The tree was established by following the same protocol as described in Figure 3a. Scale bar indicated 0.05 substitutions per nucleotide.

Figure 3c.

Phylogenetic analysis of the 16S rRNA genes for the Bacteroidetes and Chloroflexi clones from the Kuytun 51 and Qiangyong Glacier and the closest relatives. The tree was established by following the same protocol as described in Figure 3a. Scale bar indicated 0.05 substitutions per nucleotide.

Figure 3d.

Phylogenetic analysis of the 16S rRNA genes for the Actinobacteria, Firmicutes, Candidate division OP10, TM7, and Deinococcales clones from the Kuytun 51 and Qiangyong Glacier and the closest relatives. The tree was established by following the same protocol as described in Figure 3a. Scale bar indicated 0.05 substitutions per nucleotide.

Figure 3e.

Phylogenetic analysis of the 16S rRNA genes for the Cyanobacteria and Planctomycetes clones from the Kuytun 51 and Qiangyong Glacier and the closest relatives. The tree was established by following the same protocol as described in Figure 3a. Scale bar indicated 0.05 substitutions per nucleotide.

[12] There were also similarities and differences in the proportion of the main phylogenetic groups between the Kuytun 51 and Qiangyong Glacier (Figure 4). The genus Comamonadaceae was common (with 5–30% clonal frequency) across the surface snow in both glaciers, and Flavisolibacter were found across the surface snow in both glaciers, although they only accounted for a small percentage (<8% clonal frequency) of the bacteria as indicated by dotted lines in the Figure 4. However, there was a significant difference in the proportion of Cyanobacteria, Rhodoferax and Flexibacteraceae in the total clone numbers between the two glaciers (Figure 4). Cyanobacteria appeared across the surface snow in the Kuytun 51 Glacier, but rarely in the Qiangyong Glacier. Rhodoferax frequently appeared (with 4–62% clonal frequency) across the surface snow in the Kuytun 51 Glacier, but accounted for below 2.5% of the total clones in each of the habitats in the Qiangyong Glacier (Figure 4, dashed lines). Flexibacteraceae was less common (<10% clonal frequency except for the firn habitat with 29% clonal frequency) in the Kuytun 51 Glacier, but accounted for a majority (with 26–46% clonal frequency) of the total clones across the Qiangyong Glacier surface (Figure 4, dashed lines).

Figure 4.

Proportion of the main phylogenetic groups (genera) based on BLAST results for 16S rRNA gene sequences in each of the clone libraries from the different glacial surface habitats of the Kuytun 51 (data adapted from Xiang et al. [2009b]) and Qiangyong Glaciers.

3.3. Geographic Pattern of Bacterial Communities Among the Geographically Isolated Glaciers

[13] Spatial biogeography of microbial community was investigated with hierarchical clustering (Figures 5a and 5c) and principle coordinates analysis (PCoA, Figures 5b and 5d) based on the UniFrac distance matrix of the pooled 16S rRNA gene sequences from the geographically isolated glaciers Kuytun 51, Qiangyong and East Rongbuk. The results showed a clear geographical pattern with regard to bacterial communities. The geographical pattern revealed by sequence analysis corresponded to the spatial pattern of the Kuytun 51, Qiangyong and East Rongbuk glaciers. In the hierarchical clustering plot (Figure 5a), eleven clone libraries from the different glaciers generally formed three groups, one corresponding to each of the three glaciers and indicating the zonal distribution of microbial communities across the transition of the three glaciers. There was only one exception to the pattern seen in the hierarchical clustering of the microbial communities. The “dry-snow zone” community from the Qiangyong Glacier clustered with the firn-zone community within the Kuytun Glacier communities. The geographical microbial community pattern was particularly apparent in the PCoA plot of all pooled communities (Figure 5b), and hierarchical clustering analysis and PCoA plot of communities without the precence of the Qiangyong community from the dry-snow zone (Figures 5c and 5d). The bacterial communities from within a glacier grouped more closely than those from different glaciers, creating three separated ellipses in the PCoA plot, each indicating one of the three glaciers. The variation among the bacterial communities was only 30.27% to 35.9% explained by the second and third principal coordinates.

Figure 5.

UniFrac analysis of bacterial 16S rRNA gene sequences showing the overall zonal distribution of microbial communities in three glaciers Qiangyong, East Rongbuk and Kuytun 51. (a) Hierarchical clustering analysis and (b) principle coordinates analysis (PCoA) of all pooled microbial communities in glaciers. (c) Hierarchical clustering analysis and (d) PCoA plot of microbial communities without the presence of clone library from dry-snow zone at an elevation of 5710 m on the Qiangyong Glacier. In the hierarchical clustering analysis, a sequence jackknifing technique was applied to each cluster to determine the sensitivity of the relationships to sample size. Scale bar indicated difference among the bacterial communities. Percentage of variance explained by the two principal coordinates is shown in each axis label of PCoA plots. Both hierarchical clustering and PCA plots were based on UniFras distance matrix with the weighted Unifrac algorithm [Lozupone and Knight, 2005].

4. Discussion

[14] Previous investigations have shown variations in the composition of the main bacterial phyla present in different glaciers, and this trend is also apparent in the current study. This suggests the biogeography of microbial community in glacier ice. Moreover, this current study presents direct evidence of zonal tendence of microbial community distribution in the glacier surface snow/ice, corresponding to the spatial placement of three geographically isolated glaciers in western China. This supports our hypothesis of the air mass behavior being one of main forces driving the distribution of microbial community in glaciers.

4.1. Methodological Considerations

[15] Under the same project guideline for microorganisms in glacier ice and the relation to climatic and environmental changes, snow and ice samples were collected from the different habitats on the glacier surface from the East Rongbuk Glacier in April of 2005 [Liu et al., 2007], Kuytun 51 Glacier in August of 2005 [Xiang et al., 2009b] and Qiangyong Glacier in September of 2005 (this study). The sampling design ensures that the snow/ice samples represent the new snowfall deposited on the surface of the middle to upper part of the glaciers in 2005, and that the ice/meltwater from the ablation zones reflects the current habitats although they may also contain the previously melted snow. Moreover, the maximum regional annual temperature on the snow surface can only reach up to 0°C [Li et al., 2007]. The organic matter concentrations reported else in the high-altitude glaciers were extremely low, at a range of 10–350 ng gram melted-water−1 or 30 g m−2 [Takeuchi and Li, 2008]. By collecting snow and ice samples from three high altitude glaciers during the same season, we were able to compare and contrast the microorganisms present that year. We were thus able to make inferences about the influences of precipitation behaviors on the distribution of microbial communities in three geographically isolated glaciers in western China.

4.2. Biogeographic Effect on Microbial Density in the Mountain Glaciers in Western China

[16] Bacterial cell counts varied greatly across the three glaciers. The Kuytun 51 Glacier contained the highest maximum total bacterial density with 1.3 × 106 cells/ml [Xiang et al., 2009b], while the Rongbuk, and Qiangyong Glaciers contained a maximum value of bacterial density with 9.4 × 104 cells/ml [Liu et al., 2007] and 7.9 × 105 cells/ml (Table 1), respectively. The high maximum bacterial density in the Kuytun 51 Glacier may be attributed to the precipitation over the glacier and its geographic location. It neighbors many deserts where dust storms are common [Li et al., 2003], and is frequently influenced by dry air mass derived from the westerly [Wake et al., 1993]. Dust particles carry abundant microorganisms [Abyzov et al., 1998; Yao et al., 2006], and may be responsible for seeding the glacier. The relatively low maximum value of bacterial density in the Rongbuk and Qiangyong Glaciers could be attributed to the difference in the source of precipitation, which is mainly derived from monsoonal air masses [Wake et al., 1993; Shi and Liu, 2000]. Qiangyong Glacier contained an order of magnitude higher than the Rongbuk Glacier (Table 1). The difference in bacterial density between the Qiangyong and East Rongbuk glacier could be partially explained by the influences from the westerly and the neighboring dry air masses. While both glaciers are located on the southern Tibetan plateau, the precipitation over the Qiangyong Glacier is derived mainly from the monsoonal air masses and the neighboring depressions during the sampling season [Dregne, 1968; Wake et al., 1993; Shi and Liu, 2000], while the Rongbuk Glacier is most affected by the relatively clean with regard to dust particles and mild monsoonal air masses [Wake et al., 1993; Shi and Liu, 2000], and therefore contain the lowest value of maximum bacterial density. Taken together, these bacterial density data strengthen our hypothesis that air mass behaviors on the Tibetan plateau regulate the influxes of microorganisms into the glacier surface.

4.3. Cosmopolitan and Endemic Bacteria in Glacier Snow-Ice

[17] Sequence data showed the presence of cosmopolitan bacteria such as Comamonadaceae and Flavisolibacter in both Kuytun 51 and Qiangyong Glaciers (Figure 4). The presence of these clades was consistent with previous reports on the similarity of bacteria in glacier ice-snow to those from the outside environments such as the agriculture soil, lakes and plants [Xiang et al., 2005; Liu et al., 2009a, 2009b]. The ubiquity of microorganisms in geographically isolated glaciers was attributed to the extremely small cellular volume, wide spread of microbes [Prospero et al., 2005; Christner et al., 2008], and the flexibility of microbial metabolism in various environments [Seshu et al., 2002; Groudieva et al., 2004]. These conclusions reinforced the concept of aeolian deposition controlling the distribution of microorganisms in snow-ice through microbial load associated with aerosol, dust, and precipitation onto the glacier surface [Price et al., 2008; Xiang et al., 2009a]. However, sequence analysis also showed three biogeographic characteristics of microorganisms in glacier ice/snow (Figures 3a3e). The first is the frequent appearance of the phyla Proteobacteria, Bacteroidetes and Actinobacteria in glaciers worldwide and their close phylogenetic relationship to those from other cold environments [Xiang et al., 2005; Liu et al., 2009a, 2009b; Miteva et al., 2009; Zhang et al., 2009]. The second is the frequent occurrence of endemic species in a specific glacier. Comamonadaceae/Polaromonas sp. frequently occurred in the Kuytun 51 and Qiangyong Glacier (Figure 4), Guoqu (Geladaidong, e.g., representative sequence G6–215 EU153034) [Yao et al., 2008], Zadang in the southern Tibetan plateau (with representative sequence zd5–31 EU527154) [Liu et al., 2009a, 2009b], and Rongbuk Glaciers (with representative sequence RBL9–51 DQ323115) [Liu et al., 2009a, 2009b]; Flavisolibacter sp. frequently occurred in the Kuytun 51 and Qiangyong Glacier (Figure 4); Oscillatoria (Cyanobacteria) frequently occurred in the Kuytun 51 Glacier (Figure 4) and Alaska Glacier [Takeuchi, 2002], Tindal [Takeuchi and Koshima, 2004], and four glaciers at Taylor Valley, Antarctica [Mueller and Pollard, 2004]; Phormidia (Cyanobacteria) were found in the Kuytun 51 Glacier (Figure 4), four Svalbard glaciers [Stibal et al., 2006], Antarctic glacier [Christner et al., 2003] and other icy environments in Antarctica [Taton et al., 2003; Priscu et al., 2005]; Chamaesiphon subglobo (Cyanobacteria) were reported in the glacier Qiangyong (Figures 3e and 4), Rongbuk (with representative sequence RBL3–13 DQ323090 in the work of Liu et al. [2009a, 2009b]), and Zadang Glacier (with representative sequence zd1–20 EU527173 [Liu et al., 2009a, 2009b]); Acinetobacter sp. (Gammaproteobacteria) were frequently reported in the Muztag Ata Glacier [Xiang et al., 2005] and Rongbuk Glacier [Liu et al., 2009a, 2009b]; Rhodoferax (Betaproteobacteria) frequently appeared in the Kuytun 51 Glacier, but rarely in the Qiangyong Glacier (Figure 4, dashed lines). Finally, 62% of the total bacterial species recovered from the Qiangyong Glacier form a group (indicated by the gray shaded areas in Figures 3a3e) that is distinct from those reported in the Kuytun Glaciers and other different glaciers. The endemic characteristics of microorganisms reinforced the importance of postdeposition selection on the community of microorganisms in glacier ice-snow ecosystem [Price et al., 2008; Xiang et al., 2009a], and imply the importance of microorganisms as bio-indicators for climatic and environmental changes in glacier regions.

4.4. Biogeography of Microbial Communities in Glacier Ice and Ecological Implications

[18] The results showed distinct communities on the surface of three mountain glaciers in western China in 2005. The difference between the geographically isolated glaciers was more distinct than the difference between different habitats on the same glacier surface (Figures 5b–5d), despite within habitat heterogeneity of the microbial communities [Takeuchi, 2002; Takeuchi and Koshima, 2004; Xiang et al., 2009b]. Sequence analysis using UniFrac showed that bacterial communities from the same glacier form closer clusters than those from other glaciers (Figures 5a–5d). The communities exhibited a geographic pattern, which was consistent with the spatial isolation of three glaciers (Figures 5b–5d). Data also showed a significant difference in the relative abundance of the main phylogenetic groups in the geographically isolated glaciers (Figure 4) [Liu et al., 2009a, 2009b]. The results demonstrated a zonal distribution of microbial communities across the geographically isolated glaciers in western China.

[19] Two possible mechanisms can explain the zonal distribution of microbial communities across the mountain glaciers in western China. One is aeolian (wind) depositional processes, which may strongly control the microbial influx onto the glacier surface in different climatic and environmental zones [Price et al., 2008; Xiang et al., 2009a, 2009c]. Such aeolian processes are influenced by global atmospheric patterns, and thus microbial deposition may be linked to global dust and precipitation patterns. The Kuytun 51 Glacier, located along the northernmost periphery of the Tibetan plateau and surrounded by vast deserts (Figure 1), receives its precipitation from the westerly circulation throughout the year [Wake et al., 1993; Li et al., 2003]. On the contrary, the Rongbuk Glaciers on the southeastern Tibetan plateau are mostly influenced by monsoonal air masses [Murakami, 1987], while the Qiangyong Glacier on the southern Tibetan Plateau is affected by the monsoonal air masses in summer [Murakami, 1987], the neighboring depressions, and the westerly in winter [Dregne, 1968; Wake et al., 1993; Shi and Liu, 2000]. The dramatic transition in the distribution of air masses across the mountain glaciers in western China may lead to significant changes in the source of microbial influx to the glacier surface. This partially explains the differences in the glacier community structure between the northernmost periphery and southern part of the Tibetan plateau as observed in the current study.

[20] The second possible mechanism affecting the distribution of microorganisms on the glaciers of western China is a postdepositional selection mechanism. Postdepositional effects are governed by local climatic and environmental conditions including ambient temperature in the surface, prevailing wind direction, light intensity, and hydrological conditions, with the end result being the regulation of the taxonomic components of microbial community in a glacier [Christner et al., 2000; Takeuchi and Koshima, 2004; Mueller and Pollard, 2004; Xiang et al., 2009a, 2009c]. The Kuytun 51 Glacier is a cold-based subcontinental (subpolar) type glacier in the Tienshan mountain region with low equilibrium altitude about 3650 m [Xiang et al., 2009b], mild climate with an average annual temperature −9.7°C [You et al., 2006] and maximum temperatures up to 0°C [Li et al., 2007], and nutrient rich conditions [Takeuchi and Li, 2008], which could favor the growth of cold tolerant microorganisms in the surface snow. On the contrary, Qiangyong, and Rongbuk Glaciers are continent (polar) type of glaciers with comparatively low temperature on the glacier surface, and high-altitude equilibrium lines >5640 m [Kang et al., 2004]. The harsh climate and environmental conditions in the high-altitudinal glaciers Qiangyong and Rongbuk Glaciers could cause extremely low growth of microorganisms on the glacier surface, and therefore cause the relatively low bacterial density and different community composition in the glacier ice at these extremely high altitudes.

[21] Additionally, both mechanisms (aeolian deposition and postdeposition selection) most likely interact to regulate the distribution of microbial communities in the snow/ice of glaciers. Aeolian deposition leads toward a large difference in the fluxes of organic and biological components onto the surface of glaciers because of the transition from the prevailing westerly precipitation regimes in northwestern China to monsoonal regimes in southwestern China (Figure 1). The metabolic activities of microorganisms following deposition onto the glacier surface snow [Vincent et al., 1993; Fritsen and Priscu, 1998; Paerl and Priscu, 1998] would lead a change in glacier microbial consortia because of differences of microbial tolerance to the extremely cold environment and differential abilities to utilize available substrates [Xiang et al., 2005; Liu et al., 2009a, 2009b].

[22] The changes of microbial communities in glacier ice and across the geographically isolated glaciers have profound impacts on the glacier system since the specific spectral absorption of pigments generated by various microorganisms can reduce the surface albedo, increase melting of snow and ice, and then strongly influence the heat budget and mass balance of the glacier system [Kohshima et al., 1993; Thomas and Duval, 1995; Hoham and Duval, 2001]. Previous results demonstrated that the intact glacier surfaces with an extensive cover of cryoconite, a dark-colored biogenic material derived from snow algae and filamentous Cyanobacteria had a higher melting rate than that of the surfaces without cryoconite [Kohshima et al., 1993; Takeuchi, 2002]. The current results found that Cyanobacteria were much more abundant in the Kuytun 51 Glacier than in the Qiangyong Glacier (Figure 2), and below the detection limit in the East Rongbuk Glacier [Liu et al., 2007], which may be related to the differences in the climatic and environmental conditions on the surface of glaciers. More data on the combined measurements of microbial taxonomic and metabolic activities and their influences on the surface albedo of a glacier will be helpful for our improved understanding of microorganisms in glacier ice and the relation to global climatic and environmental changes.

5. Conclusions

[23] Microbial communities in glacier ice in western China showed a general zonal distribution associated with the precipitation pattern across the mountain glaciers. Most representative species from the same glacier form a closer cluster than those from other glaciers. Moreover, the dominant phylogenetic clusters were also different in the geographically isolated glaciers. The zonal distribution of bacterial communities across the mountain glaciers in western China appears to involve two different precipitation scenarios: the westerly and Asian monsoonal air masses, which may provide distinct transportation pathways for microorganisms into the glaciers between the northernmost periphery and southern part of the Tibetan plateau. Additional local climatic and environment conditions including ambient temperature, meltwater availability and light intensity on the snow-ice surface on the northern and southern Tibetan plateau glaciers may also cause differences in the community composition of microbial communities in glacier systems. The results support our hypothesis of air mass behaviors on the Tibetan plateau regulating the transportation pathway of precipitation and thus determining the zonal distribution of microbial communities across the mountain glaciers in western China. This has significant implications for microorganisms as indicators revealing global and regional climatic and environmental changes.

Acknowledgments

[24] We would very like to thank to Trista Vick for her kind help on the improvement in English used in this paper. We thank all of the members of the Qiangyong Glacier expedition for assistance in the field sample collection. This work was supported by the NSF project of China (grants 40471025 and 40871046).

Ancillary