Introduction

There has been increasing interest in the role of the microbial flora of glaciers in geochemical circulation and climate change. The cryosphere has been suggested to be a new biome uniquely dominated by microorganisms (bacteria, viruses, algae and prokaryotes) (Anesio and Laybourn-Parry 2012). Active biogeochemical processes in the cryosphere have both local and global impacts (Anesio and Laybourn-Parry 2012), and a study of all the bacteria retrieved from a deep ice core may shed light on this topic.

Previous studies based on 16S rRNA sequencing have shown that bacteria culturable from snow on the Tibetan Plateau are mainly affiliated with five groups: Actinobacteria, Firmicutes, Bacteroidetes, Alphaproteobacteria and Gammaproteobacteria (Christner et al. 2000; Xiang et al. 2005a, b; Zhang et al. 2008, 2010). The concentration of cultivable microorganisms in ice cores has ranged from 0 to 760 colony-forming units (CFU) mL−1 (Christner et al. 2000; Xiang et al. 2005a, b; Zhang et al. 2008, 2010). The predominant group of bacteria differs depending on the geographical location of the sample site—for example, culturable bacterial communities at Mt Cho Oyu, Puruogangri and Guliya are dominated by Actinobacteria (Christner et al. 2000; Ma et al. 2009a; Zhang et al. 2008), whereas Alphaproteobacteria and Betaproteobacteria are the dominant groups at Jade Dragon Snow Mountain and Mt Tiandong, respectively (Ma et al. 2009a, b). Despite the different sample sizes (~ 10 to 150 isolates), the presence and absence of Bacteroidetes is the most striking difference between geographically separated glaciers (Christner et al. 2000; Xiang et al. 2005a, b; Zhang et al. 2008, 2010). A layered distribution of cultivable bacteria was detected within single ice cores from East Rongbuk and Muztagata (Shen et al. 2012; Xiang et al. 2005a, b).

An increase in membrane fluidity at low temperatures is one of the main characteristics of bacteria found in ice and snow (Shivaji and Prakash 2010). Another adaptation strategy of bacteria that colonize glaciers may be their ability to produce pigments. Organisms that produce colonies colored with various pigments are common in cold environments (Bowman et al. 1997; Christner et al. 2000; Foght et al. 2004). Foght et al. (2004) suggested a relationship between the cold adaptation of bacteria and the production of pigments, possibly linked to the increased rigidity of the cytomembrane. In addition, pigments can shield organelles from excessive ultraviolet irradiation, which is one of the challenges of living on the glaciers of the Tibetan Plateau (Borić et al. 2011; Remias et al. 2010).

Many previous studies have revealed the diversity and numerical characteristics of bacteria from partial sections of ice cores. However, there have been few studies in which a statistical analysis has been made of all the colonies obtained from a whole ice core. We chose bacteria recovered from a 120 m ice core drilled in the Yuzhufeng Glacier on the Tibetan Plateau as our object of study to reveal the distribution pattern of the total number of colonies and the number of pigmented colonies. We investigated the variation in the abundance, diversity and pigmentation of culturable bacteria with depth in the ice core.

Materials and methods

Sample collection and pretreatment

A 120-m ice core was drilled in the Yuzhufeng Glacier (35°39.64′N, 94°14.77′E, 3800 m above sea-level), in the northeast of the Tibetan Plateau, in 2009 (Fig. 1). The frozen ice core was transported to the State Key Laboratory of Cryosphere Science in Lanzhou, China, and cut longitudinally into four parts. One of the parts was sliced into 1253 sub-sections at intervals of ~ 10 cm, and these sub-sections were then used for the microbial analysis. A 1-cm-thick layer of ice on the surface of the samples was chipped off with a sterilized blade to avoid any contamination from transportation and the sectioning process. All the preparation steps were performed at − 10 °C. The divided ice core sections were eluted with cold double-distilled water and 75% alcohol before being placed in sterilized beakers and melted slowly at 4 °C. The samples were processed under a sterile, positive-pressure laminar flow hood. Sterile gloves, clean laboratory clothing and hair covers were worn during all handling procedures.

Fig. 1
figure 1

Topographic map showing the location of the sampling site on the Tibetan Plateau

Full size image

Cultivation of bacteria

A 150-mL volume of thawed water from the ice core was separated into two sterilized Petri dishes containing R2A solid medium, which is the optimum medium for the isolation of bacteria from glacial ice cores (Christner et al. 2001). The inoculated Petri dishes were sealed with Parafilm (PM-996) and the Petri dishes that had not been inoculated served as control blanks to guard against contamination. The cultivation process was conducted in microbiological safety cabinets. The cultures were inverted and incubated aerobically at 4 °C for 6 months at an ambient relative humidity of 85%. Bacterial colonies from quadrant streak plates were re-streaked several times for purification and checked microscopically for purity. The number of colonies on each of the 1253 plates was counted and the colony morphology (size, color and edge of the colony) was noted. Purified isolates were maintained at – 80 °C in R2A broth supplemented with 25% (v/v) glycerol. CFU values were calculated using the formula CFU = colonies × 0.15−1. The quadrant streak method was used repeatedly to obtain pure cultures.

Extraction and analysis of pigments

Pigments were extracted from cells of the isolates harvested from the R2A medium after 7 days of incubation at 25 °C. The extraction was performed in acetone and the extracts were purged through a disposable filter unit (0.45 µm) into dark-colored autosampler vials. The production of pigments was measured using high-performance liquid chromatography (HPLC) (Agilent 1260 infinity column, Agilent Technologies, Inc.) (Schmid et al. 1998).

16S rRNA gene analysis

The genomic DNA of the strain was extracted based on the method of (Lowe and Eddy 1997). A 100-mL volume of the 48-h culture was extracted and then centrifuged at 4 °C and 10,000 rpm for 15 min. The cells were washed three times with 5 mL of sterile water. The washed cells were resuspended in 1128 µL of 10 mM Tris–HCl buffer containing 1 mM EDTA (pH 8.0) and 20 μg of lysozyme, and then incubated at 37 °C for 2 h. A 6-μL volume of proteinase K (20 mg/mL), 4 µL of DNase-free RNase (10 mg/mL) and 100 µL of sodium doceylsulfate (20% w/v) were then added and the cell suspension was incubated at 55 °C for 3 h. The cell lysate was extracted twice with phenol/chloroform/isoamyl alcohol (25:24:1) and once with chloroform/isoamyl alcohol (24:1). The aqueous layer was separated after centrifugation at 12,000 rpm for 15 min. The DNA was precipitated with 1 volume of frozen anhydrous ethanol.

The purity of the genomic DNA was assessed using a NanoDrop 2000c ultraviolet–visible spectrophotometer (Thermo Scientific) using an OD260/OD280 ratio of 1.8–2.0. The DNA was stored in TE buffer (pH 8.0) for 16S rRNA gene sequencing. The resulting DNA was amplified by polymerase chain reaction (PCR) using the universal bacterial primers 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-CGGTTACCTTGTTACGACTT-3′) (Lane 1991). PCR was carried out in a final volume of 50 mL using 2 mL of template DNA, 3 mL of MgCl2, 4 mL of each dNTP, 0.5 mL of each primer, 0.2 mL of Taq DNA polymerase and 35 mL of double-distilled water. Reactions were performed in an ABI PCR System Thermo Cycler with the following cycling parameters: 94 °C for 5 min for an initial denaturation, followed by 30 cycles of 94 °C for 30 s, 55 °C for 1 min, 72 °C for 1 min and a final extension at 72 °C for 10 min.

A BLAST search in GenBank (http://www.ncbi.nih.gov) was carried out for the bacterial 16S rRNA gene sequences obtained in this study. The blasted sequences and selected closest references were pooled and aligned using ClustlW. Phylogenetic analysis was performed using the distance-based neighbor-joining method with MEGA version 5.0 (Tamura et al. 2011). The sequences determined in this study were deposited in the GenBank database under accession numbers KF295220-KF295319, KF295540-KF295555 and KF295693-KF295742.

Results

Among the 1253 Petri dishes, 401 (32%) formed colonies and a total of 25,449 colonies were obtained from these plates. The number of colonies per dish ranged from 1 to 1000 (~ 7 to 6667 CFUs). The other 852 (68%) Petri dishes did not form any colonies during the observation period of 4 months. The number of colonies per Petri dish increased significantly with increasing depth in the ice core (r = 0.16, P < 0.01, n = 401). The sample section with the highest number of colonies (1000) was located at the bottom of the section of Y1239 at a depth of 119.29 m (Figs. 2, 3).

Fig. 2
figure 2

Variation in the number of total colonies with depth in the Yuzhufeng ice core

Full size image
Fig. 3
figure 3

Variation in the number of pigmented colonies with depth in the Yuzhufeng ice core. (a) Stacked area chart of the variation in the number of pigmented colonies with depth. Composition of colonies with different colors in (b) the whole ice core, (c) the upper 41.91 m of the ice core and (d) the bottom 78.47 m of the ice core. The pigmented colonies made up 89% of all isolates. The percentage of pigmented colonies increased significantly from 79 to 95% from the top to the bottom of the ice core

Full size image

Pigmented colonies made up 89% of all the isolates and, based on their color, the colonies were divided into six groups (orange, reddish-orange, yellow, pink, brown and white; Supplementary Figure S1). Yellow-colored colonies were predominant throughout the ice core and made up 47% of all colonies, followed by reddish-orange (24%), orange (16%), white (11%), pink (2%) and brown (23 colonies, < 1%) (Fig. 3b). The percentage of pigmented colonies increased markedly (from 79 to 95%) with increasing ice core depth (Fig. 3c, d).

One hundred and sixty-six representative strains were selected to perform 16S rRNA phylogenetic analysis based on the colony morphology and the vertical position in the ice core. The taxa of the sequenced strains were determined using the BLAST program at the National Center for Biotechnology Information. Based on the 16S rRNA phylogenetic analysis, these strains were affiliated to Proteobacteria (56.02%), Actinobacteria (28.92%), Bacteroidetes (14.46%) and Firmicutes (0.60%). The dominant 19 genera of strains were identified as Pseudomonas (55 strains, 33% of total isolates), followed by Cryobacterium (27 strains, 16% of total isolates), Psychrobacter (16 strains, 10% of total isolates) and Flavobacterium (13 strains, 8% of total isolates), with Janthinobacterium, Massilia, Lysobacter, Stenotrophomonas, Paracoccus, Brevibacillus, Rhodococcus, Arthrobacter, Leifsonia, Salinibacterium, Salinibacterium, Curtobacterium, Microbacterium, Dyadobacter, Sphingobacterium and Chryseobacterium making up the remaining 34% of the strains. Figure 4 provides further details on the composition of the bacterial communities in the Yuzhufeng ice core. Pseudomonas and Cryobacterium repeatedly occurred at different ice depths and 36 and 24 times in different layers (Supplementary Table 1).

Fig. 4
figure 4

Neighbor-joining phylogenetic tree for strains isolated from the Yuzhufeng Glacier ice core based on 16S rRNA gene sequence analysis. The numbers at the nodes indicate the bootstrap percentages (based on 1000 replications); only values > 50% are shown. Bar, 0.05 accumulated changes per nucleotide. The numbers in parentheses indicate how many isolates were compressed into one cluster

Full size image

HPLC analysis showed that 21 of the 25 strains produced pigments and 4 of these had no matched peak. Carotenoids and a complex of undetermined pigments were detected, namely, α- and β-carotene, 19′-butanoyloxy-fucoxanthin, fucoxanthin, diatoxanthin, peridinin and zea/lutein (Table 1). Forty percent of the isolates produced α-carotene, making it the dominant carotenoid produced by bacteria from the Yuzhufeng Glacier. The second most common pigment was diatoxanthin, which was produced by 28% of the isolates. Species in the same genus had similar pigment profiles—for example, strains Y1243-1 and Y1243-2 in the genus Massilia.

Table 1 Pigment profiles of strains of bacteria from the Yuzhufeng Glacier, Tibetan Plateau
Full size table

Discussion

The taxonomic groups of the strains isolated from Yuzhufeng Glacier were consistent with those obtained from glaciers in different geographical areas. All the strains belonged to four phyla: Proteobacteria, Actinobacteria, Bacteroidetes and Firmicutes (Christner et al. 2000; Liu et al. 2008; Miteva et al. 2004; Xiang et al. 2005a, 2009). Isolates related to Proteobacteria formed the largest cluster in terms of diversity and abundance (89 strains; 56% of all isolates), with seven different genera. The dominant groups at both Mt Cho Oyu and the Zhadang Glacier were Firmicutes and Actinobacteria (Ma et al. 2009a; Shen et al. 2014), indicating that the taxonomic sorting of culturable bacteria in glaciers may be influenced by differences in their origin as well as the local environmental conditions. For example, Pseudomonas and Cryobacterium repeatedly occurred at different ice depths and were present about 36 and 24 times in different layers. This indicated that the environmental conditions on the Yuzhufeng glacial surface were conducive for Pseudomonas and Cryobacterium species most of the time. However, some common preponderant species, such as Cryobacterium, were shared between different glaciers and made up an important portion of the bacteria at both the Yuzhufeng (16%) and Puruogangri (26%) glaciers (Zhang et al. 2008), indicating that there was a wide distribution of some species of bacteria isolated from the cryosphere of different glaciers under the same environmental selection pressure. The two opposing theories in microbial ecology—endemism and cosmopolitanism—co-determined the distribution of microbial species in the cryosphere (Anesio and Laybourn-Parry 2012). However, the overall composition of the microbial community in this biome seemed to be determined by non-random factors (Anesio and Laybourn-Parry 2012). For example, Cryobacterium, Stenotrophomonas and Sphingobacterium occur in many cold habitats (Foght et al. 2004; Miteva et al. 2004; Simon et al. 2009; Suzuki et al. 1997; Xiang et al. 2005a; Zhang et al. 2008). A total of 19 genera isolated in this study represent the greatest number of genera found from a single ice core to date. This demonstrates the importance of high-resolution cultures in studies of microorganisms in ice cores.

Bacteria from different glaciers have specific characteristics, but this does not mean that they are randomly distributed. Proteobacteria, Actinobacteria and Firmicutes can be isolated from almost every glacier worldwide. The presence or absence of Bacteroidetes was the most significant difference between geographical locations (Table 2). This suggests that cultivable Bacteroidetes may be the most sensitive to environmental gradients caused by geographical location. However, the predominant bacterial group recovered from different glaciers was different depending on the geographical location of the sample site (Amato et al. 2007; Christner et al. 2000, 2001; Miteva and Brenchley 2005; Miteva et al. 2004; Shen et al. 2014).

Table 2 Community composition of culturable bacteria from glaciers in different geographical locations
Full size table

Even if an ice core harbors abundant culturable bacteria, this does not mean that bacteria can be isolated from every section. Thus, an insufficient number of samples or a low resolution may hinder our understanding of the culturable microflora found in ice cores. The results of the present study show that colonies of bacteria only formed on 32% of our 1253 Petri dishes; no colony was formed on the other 68% of Petri dishes. However, the fact that no bacterium was isolated from some sections of the ice core did not mean there was no culturable bacterium in the whole ice core. Local sampling of ice cores resulted in 0 CFU from Siple Station and Dyer Plateau (Christner et al. 2003), although, realistically, ice from these locations will harbor culturable bacteria. To the best of our knowledge, Christner et al. (2003) investigated culturable bacteria in a 120-m ice core with the highest resolution, at 10-cm intervals.

The ice core drilled from the Yuzhufeng Glacier harbored a relatively higher number of CFUs (~ 7 to 6667 CFU mL−1, average 424 CFU mL−1) than ice cores from Sajama (Bolivia, 0–17 CFU mL−1), Taylor Dome (10 CFU mL−1), Siple Station (2 CFU mL−1), Dyer Plateau (0 CFU mL−1), Summit (0 CFU mL−1) and other glaciers on the Tibetan Plateau (0–760 mL−1) (Christner et al. 2000; Xiang et al. 2005a; Zhang et al. 2008), but much lower than ice from permafrost, in which living bacteria reproduce at concentrations as high as 106 CFU mL−1 (Katayama et al. 2007). Non-polar glaciers yield significantly higher numbers of individual colonies or CFUs than polar glaciers because they are closer to major sources of airborne particles from soils (Anesio and Laybourn-Parry 2012).

Within the Tibetan Plateau, the significantly higher numbers of CFU from Yuzhufeng Glacier than from other sites may be attributable to the cultivation method used, the cultivation temperature, the cultivation time and the geographical location (Liu et al. 2013; Yao et al. 2008). However, another, more important, reason may be the different resolutions used in each study. Among the 1253 plates examined, the number of colonies from one Petri dish ranged from 0 to 1000 with large fluctuations. This large fluctuation was also found in the Puruogangri ice core (Zhang et al. 2008), where one dish had 1000 colonies, whereas the others yielded no more than 540 colonies. Petri dish Y1239 (section 1239 of the ice core, Supplementary Table S1) was not included in the investigation and the highest concentration of culturable bacteria decreased from 6667 to 3600 CFU mL−1 (reduced by almost 50%) as a result of this exclusion. There was no successful recovery of viable bacteria from ice on the Dyer Plateau (Christner et al. 2000). This does not mean that Dyer Plateau ice harbors no viable bacteria, but was probably due to the limited number of samples used in the analysis. This highlights the importance of taking an adequate number of samples from one ice core in studies of the bacterial community.

Yellow-colored colonies were predominant throughout the ice core and made up 47% of all colonies, followed by reddish-orange (24%), orange (16%), white (11%), pink (2%) and brown (23 colonies, < 1%). The pigmented colonies made up 89% of all the isolates. This conclusion was also supported by the HPLC analysis. The dominance of pigmented bacteria in the Yuzhufeng ice core is consistent with the proposal that pigmented microorganisms are more abundant in cold environments than in other terrestrial environments (Bowman et al. 1997; Christner et al. 2003; Miteva et al. 2004; Zhang et al. 2008). The percentage of pigmented bacteria in the Yuzhufeng ice core was higher than that in the East Rongbuk ice core, the Puruogangri ice core and the Greenland ice core, where the proportions of pigmented bacteria were 85, 80 and 79%, respectively (Miteva et al. 2004; Shen et al. 2012; Zhang et al. 2008). The percentage of pigmented colonies increased markedly (from 79 to 95%) from the upper part to the bottom of the ice core, indicating that the proportion of pigmented bacteria that could be recovered increased with depth. This was further evidence that pigments play an important part in the adaptation of bacteria to cold environments (Dillon et al. 2003; Fong et al. 2001; Ponder et al. 2005).

The number of colonies from a single Petri dish increased with increase in ice core depth. This may be attributable to the growth and division of the bacteria that were well adapted to the ice core. Direct evidence of bacteria reproducing at − 15 °C has been demonstrated under laboratory conditions (Mykytczuk et al. 2013) and in situ microbial activity has been reported at temperatures from − 10 to − 17 °C (Carpenter et al. 2000; Finegold 1996; Rivkina et al. 2000; Russell 2000). However, the in situ cell reproduction of ice core bacteria has not been rigorously tested and it has been suggested that isolates from permafrost are only survivors (Bakermans et al. 2003). Hence, we hypothesize that these isolates from the ice core were also only survivors and, among these well-adapted bacteria, the pigmented species are dominant. The apparent layered distribution of pigmented bacteria in the ice core sections may reflect a microbial response to niche conditions and past changes in climate.

The dynamics of the bacterial community were also manifested by the variation in the proportion of pigmented colonies from snow and ice. Statistics on cultivable bacteria from the Tibetan Plateau and polar regions have shown that pigmented colonies represent 61–70 and 80–89% of the isolations, respectively (Fig. 5) (Miteva et al. 2004; Shen et al. 2013; Shen et al. 2014; Zhang et al. 2008). The change in phenotypic composition is consistent with the difference in phylogenetic composition of the microbial assemblages within the snow layers and between snow and freshwater (Møller et al. 2013). Much previous work has reported that the microbial community is influenced by local environmental conditions (Margesin and Miteva 2011; Sommaruga and Casamayor 2009), but the specific environmental factors are rarely reported. In this study, we hypothesize that intense ultraviolet radiation may have played an important part in shaping the microbial community of the glacial habitat by acting as an environmental force.

Fig. 5
figure 5

Percentage of pigmented colonies of bacteria in a snow pit and an ice core on the Tibetan Plateau

Full size image

Seven types of carotenoids and some unknown pigments were detected in our study. These carotenoids were α- and β-carotene, 19′-butanoyloxy-fucoxanthin, fucoxanthin, diatoxanthin, peridinin and zea/lutein, with α-carotene as the dominant carotenoid. It is well known that pigmented microorganisms are abundant in snow and ice, but few studies have analyzed the composition of these pigments (Amato et al. 2007; D’Elia et al. 2008; Foght et al. 2004; Liu et al. 2008; Miteva et al. 2004; Zhang et al. 2008). Previous HPLC studies focused on one species or a certain genus and aimed to clarify the contribution of carotenoids to the adaptation of some species to cold conditions (Chattopadhyay 2006; Shivaji and Prakash 2010). Our study analyzed 19 genera and the proportion of each pigment in each. The results showed that α-carotene was the most common pigment and was produced by 40% of the isolates. The second most common pigment was diatoxanthin, which was produced by 28% of the isolates. Previous study revealed that C50-carotenoid (the synthesis of which is induced at low temperatures) and certain carotenoids are only present in the thylakoid membranes of Cylindrospermopsis raciborskii at low temperatures, possibly protecting the cyanobacterium from reactive radicals (Fong et al. 2001; Várkonyi et al. 2002). The predominance of α-carotene in strains isolated from Yuzhufeng ice core is consistent with this conclusion. The second most common pigment, diatoxanthin, is also worth studying further.

Conclusions

This investigation of culturable bacteria in a deep ice core at high resolution may help to expand our understanding of these bacteria beyond the number of CFUs. The total number and proportion of pigmented bacteria with recovery activity increased with depth, indicating that this distribution may be determined by ecological rather than random factors. It is hypothesized that ultraviolet radiation may be one of the environmental factors shaping the microbial community of glacial habitats.