Characterization of potential stress responses in ancient Siberian permafrost psychroactive bacteria


  • Monica A. Ponder,

    1. Center for Genomic and Evolutionary Studies on Microbial Life at Low Temperatures, Michigan State University, East Lansing, MI 48823, USA
    2. Department of Microbiology and Molecular Genetics, Michigan State University, East Lansing, MI 48823, USA
    Search for more papers by this author
  • Sarah J. Gilmour,

    1. Center for Genomic and Evolutionary Studies on Microbial Life at Low Temperatures, Michigan State University, East Lansing, MI 48823, USA
    2. Department of Crops and Soil Sciences, Michigan State University, East Lansing, MI 48823, USA
    3. MSU-DOE Plant Research Lab, Michigan State University, East Lansing, MI 48823, USA
    Search for more papers by this author
  • Peter W. Bergholz,

    1. Center for Genomic and Evolutionary Studies on Microbial Life at Low Temperatures, Michigan State University, East Lansing, MI 48823, USA
    2. Department of Microbiology and Molecular Genetics, Michigan State University, East Lansing, MI 48823, USA
    Search for more papers by this author
  • Carol A. Mindock,

    1. Department of Chemistry, Michigan State University, East Lansing, MI 48823, USA
    Search for more papers by this author
  • Rawle Hollingsworth,

    1. Department of Chemistry, Michigan State University, East Lansing, MI 48823, USA
    Search for more papers by this author
  • Michael F. Thomashow,

    1. Center for Genomic and Evolutionary Studies on Microbial Life at Low Temperatures, Michigan State University, East Lansing, MI 48823, USA
    2. Department of Microbiology and Molecular Genetics, Michigan State University, East Lansing, MI 48823, USA
    3. Department of Crops and Soil Sciences, Michigan State University, East Lansing, MI 48823, USA
    4. MSU-DOE Plant Research Lab, Michigan State University, East Lansing, MI 48823, USA
    Search for more papers by this author
  • James M. Tiedje

    Corresponding author
    1. Center for Genomic and Evolutionary Studies on Microbial Life at Low Temperatures, Michigan State University, East Lansing, MI 48823, USA
    2. Department of Microbiology and Molecular Genetics, Michigan State University, East Lansing, MI 48823, USA
    3. Department of Crops and Soil Sciences, Michigan State University, East Lansing, MI 48823, USA
    4. MSU-DOE Plant Research Lab, Michigan State University, East Lansing, MI 48823, USA
    Search for more papers by this author

*Corresponding author. Tel.: +517 355 0271ext.287; fax: +517 353 2917, E-mail address:


Past studies of cold-acclimated bacteria have focused primarily on organisms not capable of sub-zero growth. Siberian permafrost isolates Exiguobacterium sp. 255-15 and Psychrobacter sp. 273-4, which grow at subzero temperatures, were used to study cold-acclimated physiology. Changes in membrane composition and exopolysaccharides were defined as a function of growth at 24, 4 and −2.5 °C in the presence and absence of 5% NaCl. As expected, there was a decrease in fatty acid saturation and chain length at the colder temperatures and a further decrease in the degree of saturation at higher osmolarity. A shift in carbon source utilization and antibiotic resistance occurred at 4 versus 24 °C growth, perhaps due to changes in the membrane transport. Some carbon substrates were used uniquely at 4 °C and, in general, increased antibiotic sensitivity was observed at 4 °C. All the permafrost strains tested were resistant to long-term freezing (1 year) and were not particularly unique in their UVC tolerance. Most of the tested isolates had moderate ice nucleation activity, and particularly interesting was the fact that the Gram-positive Exiguobacterium showed some soluble ice nucleation activity. In general the features measured suggest that the Siberian organisms have adapted to the conditions of long-term freezing at least for the temperatures of the Kolyma region which are −10 to −12 °C where intracellular water is likely not frozen.


The majority of the Earth’ s surface is permanently cold, with approximately 70% of the surface covered by oceans with an average temperature of 4 °C and over 20% of the terrestrial area occupied by permafrost including, 85% of Alaska, 55% of Russia and Canada, 20% of China, and the majority of Antarctica. Soils, sediments and rock exposed to temperatures of 0 °C or below for a period of at least 2 years are defined as permafrost [1].

A variety of microorganisms have been isolated from buried permafrost of the Kolyma region of northeast Siberia indicating that these organisms can survive sub-zero temperatures (−10 to −12 °C), low water activity (aw= 0.9), low nutrient availability, and the cumulative effect from background γ radiation from soil minerals that ranges from 1 to 6 kGy [2]. Depending on geologic strata these microbes have been in a continuously frozen environment for 20,000 to 3–5 million years [3]. In contrast to many oceanic isolates, permafrost isolates are not psychrophiles, but psychrotolerant in that they can grow at 4 °C and above 20 °C [4,5]. Most previous studies of permafrost microbes have been done with strains that have been isolated and grown at room temperature and in nutrient rich media. We previously isolated 238 bacterial strains from different age layers of Siberian permafrost without exposure to temperatures above 4 °C and using several isolation strategies, including low nutrient media and cryo-protectants [3]. Previous characterization has indicated that many of these isolates are psychroactive in that they grow at −2.5 °C [6].

Physiological responses to low temperatures of some mesophilic bacteria, psychrotrophic food-borne pathogens, and environmental isolates have been characterized [7]. The majority of studies have focused on membrane composition changes and cold shock protein production in response to lower temperatures. While most studies have defined responses to temperatures below the organism’ s temperature optimum, only a few of these mesophilic bacteria are capable of growth at temperatures below 4 °C. Those that can are Listeria monocytogenes[8], several Pseudomonas[9] and Arthrobacter strains [10], and some unidentified Arctic isolates [11] from non-permafrost environments.

The permafrost environment in contrast to many surface environments, is very stable, with a constant set of stresses that may have acted as selective factors on the survivors, including adaptation to a very long-term frozen environment and the associated desiccation, and to cumulative effects of γ radiation which would be expected to damage cell DNA [2,12]. The Kolyma permafrost temperatures of −10 to −12 °C would not be expected to freeze the cytosol of bacterial cells [13] and hence continued biochemical catalysis could be expected, albeit the fluid would be viscous and reaction rates very slow. The objective of this study is to characterize features of a selected set of these isolates that may play a role in or reflect their adaptation to permafrost conditions. Particular emphasis is placed on a Gram negative Psychrobacter strain isolated from a 20,000–30,000 year old permafrost layer and a Gram positive Exiguobacterium strain isolated from a 2–3 million year old layer.

2Materials and methods

2.1Isolation and phylogenetic characterization

Permafrost samples were obtained from the polar region of the Kolyma-Indigirka lowland (152–162°E, 68–72°N), located adjacent to the East Siberian Sea by David Gilichinsky and team (Cryobiology Laboratory, Russian Academy of Sciences, Pushchino). The isolation conditions and characteristics of sampling sites and cores chosen for bacterial isolation, were detailed previously by Vishnivetskaya et al. [3]. The first 500 bp of the 16S rRNA genes of the isolates were sequenced to determine phylogeny [6]. The isolates included members of the order Actinomycetales (genus Arthrobacter and Family Microbacteriacea) division Firmicutes (the genera Exiguobacterium and Planomicrobium), genus Flavobacterium, and division Proteobacteria (Psychrobacter and Sphingomonas). Analysis of the complete 16S rRNA gene of Psychrobacter sp. 273-4 and Exiguobacterium sp. 255-15 isolates confirmed their identity and established their similarity to other previously studied members of their respective clades (H. Ayala del-Rio and D. Rodrigues, respectively, unpublished observation). BOX PCR profiles revealed that Psychrobacter sp. 273-4 and sp. 215-51 were different strains, likely different species. Isolates were chosen from the larger set for further studies based on ease of culturability at 4 °C, growth at −2.5 °C, age of permafrost strata and for being representative of different taxa found.

2.2Growth rates

Growth rates, as functions of temperature were measured for Exiguobacterium strain 255-15 and Psychrobacter strains 273-4 and 215-51 by optical density in shaken flasks (200 rpm) of tryptic soy broth (TSB) (Difco, Detroit, MI), at temperatures from 45 to 0.5 °C using three biological replicates per temperature. Arrhenius plots of Exiguobacterium sp. 255-15 and Psychrobacter sp. 273-4 and sp. 215-51 were prepared by plotting the natural logarithm of the growth rate against the reciprocal of absolute temperature [14] using Statview with Lowess fitted lines (SAS Institute, Cary, NC). A Bělehràdek growth model was also constructed by plotting the square root of the growth rate against the absolute temperature and the minimum temperature determined by the method of Ratkowsky et al. [15]. Growth rates at −2.5 °C were obtained by viable plate count in triplicate non-shaken flasks. This data was not included in the Arrhenius and Bělehràdek models because of the different growth condition. Growth rates were calculated from the slope of four or more time points.

Growth rates as a function of low water activity were measured by optical density in 1/10 TSB supplemented with NaCl. Colony formation was scored after one week on 1/10 TSA media supplemented with 0.125, 0.250, 0.5, 1.0, 1.5, 2.0, or 2.50 M NaCl. These concentrations correspond to internal osmotic pressures of 0.217, 0.433, 0.873, 1.78, 2.72, 3.70 and 4.70 osm, respectively, as extrapolated from Rand et al. [16]. Several isolates were chosen based on growth at high concentrations of NaCl to further examine growth using sucrose as an alternate osmolyte.

2.3Cell morphology and size

Psychrobacter sp. 273-4 was grown to an OD600 of 0.2 in 1/2 TSB and 1/2 TSB + 5% NaCl (1.61 osm) at 4 and 22 °C. Twenty microliters of the culture was spotted onto 10 agar coated microscope slides and viewed by light microscopy. Digital images were analyzed for cell size and morphology using the Center for Microbial Ecology Image Analysis software (CMEIAS) [17]. Approximately 250 cells were analyzed for each treatment.

2.4Effect of temperature and salinity on lipid and polysaccharide composition

Psychrobacter sp. 273-4 and Exiguobacterium sp. 255-15 were incubated at 24, 4 or −2.5 °C in 1/2 TSB or 1/2 TSB + 5% NaCl with shaking until an OD600 of 0.5 was obtained. All cells were harvested by centrifugation and washed four times in sterile 1X PBS. Total lipids were extracted by the method of Mindock et al. [18]. The aqueous layers were kept for polysaccharide analysis. The organic layers containing polar lipids were analyzed by NMR spectroscopy (using d4-methanol as the solvent) and by thin layer chromatography (TLC) on silica plates using 10:4:2:2:1 chloroform: acetone: methanol: acetic acid: water. The spots were sprayed either with orcinol or 10% phosphomolybdic acid in ethanol and heated at 120 °C to visualize the organic components. NMR spectroscopy provided confirmation of identities obtained by TLC mobility compared to standards. NMR spectra were recorded on a Varian VXR-500 spectrometer (500 MHz). Chemical shifts for samples in d4-methanol are quoted relative to the proton resonances at 4.78 and 3.30 ppm, and for samples in D2O relative to the proton resonance at 4.65 ppm. For the double quantum filtered-correlated spectroscopy (DQF-COSY) experiments, a total of 256 data sets with 24 transients at 2048 points each were acquired. The total correlated spectroscopy (TOCSY) experiments were also performed by using a total of 256 data sets with 24 transients at 2048 data points each and with a mixing time of 90 ms.

Fatty acid methyl ester analysis was done using a portion of the fatty acid containing organic layer of the total lipid extract. The extract was incubated in acidified methanol at 75 °C for 36 h, dried and resuspended in 2:1:0.7 hexane: chloroform: H2O. After vigorous shaking, the mixture was centrifuged, the organic layer removed and concentrated to dryness. The resulting fatty acid methyl esters were analyzed by gas chromatography (GC) and GC–MS using a 30-m DB1 column (0.32 mm inner diameter, 0.25 μm film). The temperature program increased from 50 to 150 °C at 5°/min, 150–185 at 2°/min, 185–250 at 5°/min. After an initial analysis 2 ng of 2-OH dodecanoate (Sigma–Aldrich, St. Louis, MO), not present in any of the samples, was added to whole cells as an internal standard.

Polysaccharides were isolated from the aqueous layers of the total lipid extracts by precipitation with ethanol. The precipitated polysaccharide was recovered with a glass rod and dissolved in 10 ml of water, 10 mg MgCl2, and 10 units each of RNase A and DNase (Sigma–Aldrich) and incubated overnight at 22 °C. The polysaccharides were then dialyzed against water for 4–5 h, changing the water twice during that time, lyophilized and the resulting solids were weighed. These polymers were analyzed by gas chromatography–mass spectrometry (GC–MS) after converting them to alditol acetate derivatives [19]. The remaining aqueous alcoholic solution was centrifuged to remove any solids. The supernatants which contained free amino acids and carbohydrates were removed and also analyzed by GC–MS.

2.5Freezing tolerance

To determine if temperature acclimation influences freezing survival, the strains were grown at 24 and at 4 °C to a cell density of 108 CFU/ml in 1/2 TSB and subsequently frozen slowly (0.2 °C/min) to −20 °C. Viable plate counts were performed initially before freezing and compared to those obtained after slow thawing (0.3 °C/min) of samples held at −20 °C for 1 year. Plates were incubated at the same temperature at which the strains were originally grown. An unpaired t test was used to determine statistical significance (Statview version 5.0, SAS institute).

2.6Temperature dependent nutrient utilization

Biolog plates (Biolog, Inc, Hayward, CA) were used to assess whether temperature affected the utilization of 95 different carbon sources. Psychrobacter sp. 273-4 and Exiguobacterium sp. 255-15 were grown at 24 or 4 °C in 150 ml flasks, with 40 ml of 1/10 TSB with shaking until an OD600 of 0.5 was obtained. All cells were harvested by centrifugation and washed four times in sterile 1X PBS. Exiguobacterium sp. 255-15 was aliquoted into GP inoculating broth (Biolog Inc.) to match the percent transmittance (±3%) of the GP standard prepared by Biolog. The same procedure was followed with Psychrobacter sp. 273-4 with the exception that GN (non-enteric) broth was used for the inoculation standard. After inoculating the GP (Exiguobacterium) and GN (Psychrobacter) Biolog plates, they were then incubated at the temperature of inoculum growth. Six biological replicates (with two technical replicates each) were used per strain and temperature.

The OD595 of the wells was determined with a microplate reader every 8 h until the absorbance did not increase (4 days at 24 °C and 3 weeks at 4 °C). Average OD595 values for each carbon source were determined for each set of replicates per time point. Replicates which showed a standard deviation >15% were excluded and the averages recalculated. Carbon sources which showed at least an average absorbance change of 0.2 or more from the blank plate were considered utilized.

2.7Antibiotic susceptibility

The effect of temperature on antibiotic resistance was assessed using the impregnated disk method at both 24 and 4 °C of triplicate experiments [20]. Antibiotic disks (BBL Microbiology) were placed on lawns of bacteria on Mueller Hinton agar, incubated for 2 days at 24 °C or 2 weeks at 4 °C. After this time the size of clearings around the disks was measured. Resistant cells grew significantly closer to the antibiotic disk than susceptible cells while clearing size cutoffs for each antibiotic were used to determine resistance categories described by Barry and Thornsberry [20].

2.8UVC survival

The effect of temperature on UVC survival was assessed using cultures of Psychrobacter sp. 273-4 and Exiguobacterium sp. 255-15 grown to late-log phase (OD600= 0.6–0.9) in 1/2 TSB at 25 and 4 °C. UVC was chosen because it causes similar damage to low doses of γ irradiation [21,22], and the ease of application of small doses. Escherichia coli B606 was used as a comparison strain, because it is a mesophilic γ-Proteobacterium related to Psychrobacter sp. 273-4. Bacillus subtilis PY79 was selected as a related mesophilic Firmicute for comparison to Exiguobacterium sp. 255-15. Both mesophilic strains were grown to late log phase (OD600= 0.9) in 1/2 TSB at 24 °C.

A 15 ml culture at late log phase was collected by centrifugation at their cultured temperature, re-suspended in 15 ml 0.85% NaCl and stored on ice until use. Cells were then pipetted into a sterile glass Petri plate and exposed to a UVC fluorescent lamp (equivalent dose 1.5 J m−2 s−1) with constant mixing. Psychrobacter sp. 273-4 and E. coli B606 were cumulatively exposed to 0, 25, 50, and 100 J m−2. Exiguobacterium sp. 255-15 and B. subtilis PY79 were cumulatively exposed to 0, 100, 250, and 500 J m−2. At each exposure level, plate counts were performed on 1/2 TSA plates and incubated in the dark at the respective culturing temperatures. All steps from UVC exposure through incubation were carried out in the dark and all solutions were kept at the same temperature at which cells were cultured. After 48 h at 24 °C or 2 weeks incubation at 4 °C, CFU/ml were estimated. Average percent survival of four replicates at each of the UVC dosage levels was used to compare Psychrobacter sp. 273-4 grown at 24 °C to both of the other two samples using a one-tailed two sample t test assuming unequal variances.

2.9Ice nucleation activity

Ice nucleation studies were undertaken to determine the effect of temperature on ice nucleation activity in seven permafrost strains. The strains were grown at 24 °C on 1/10 TSA and 1/10 TSA supplemented with 5% glycerol, which has been shown to optimize ice nucleation activity (INA) in many bacteria [23,24]. INA was measured by the freezing drop method [25]. Suspensions of bacterial cells were prepared in sterile buffer and 60 10 μl drops of the suspension were placed on an aluminum boat in a circulating ethanol cold bath set to −10 °C. The freezing of droplets at or above −10 °C indicated INA. After initial determination of INA in cultures grown at 24 °C, the suspensions were subjected to 3.5 h at 4 °C and 1 h at −10 °C and then reassessed for INA by the freezing drop method. Pseudomonas syringae ATTC 35421 and E. coli ATTC 39524 were used as positive and negative controls, respectively, for INA. The ice nucleation activity per cell was determined with the method of Pooley and Brown [26].


3.1Effect of temperature on the maximal growth rate and cellular morphology

The Exiguobacterium and two Psychrobacter strains tested both grew at −2.5 °C with generations times of 5.5 and 3.5 days, respectively, but the Exiguobacterium strain had a much higher growth maximum (42 °C) than Psychrobacter strains (26 °C) (Fig. 1). Arrhenius plots of the Exiguobacterium growth profile reveal a change in slope occurring at 24 °C, while the Psychrobacter showed changes in slope at 22 and 6 °C. Minimum growth temperatures of −15, −12, and −7 °C were estimated using the Bělehràdek growth model for Psychrobacter sp. 273-4 and sp. 215-51 and Exiguobacterium, respectively, and agreed with extrapolations from Arrhenius plots.

Figure 1.

An Arrhenius plot of growth of Exiguobacterium sp. 255-15 and Psychrobacter sp. 273-4 and 215-51 at a range of temperatures between 42 and −0.5 °C. No growth occurred in Psychrobacter isolates at temperatures above 28 °C.

Psychrobacter sp. 273-4 exhibited a change in cell size and shape when subjected to either 5% salt or low temperature (Table 1). In general, the cells were slightly larger when grown at 4 °C or in the presence of salt. At the optimum growth temperature of 22 °C, the cells are rod shaped and 1.8 μm in length. The introduction of 5% salt resulted in a significant increase in width and decrease surface/volume ratio of the cells. The average biovolume of Psychrobacter increased at low temperature but most significantly in the presence of salt (Table 1). Cells were somewhat more pleomorphic when grown in the presence of salt. No differences in cell size or morphology were seen in initial observations of Exiguobacterium sp. 255-15 in response to 5% NaCl or 4 °C so further CMEIAS analysis was not performed.

Table 1.  Effect of salt and temperature on the cellular morphology of Psychrobacter sp. 273-4
 MediumMedian length (μm)Median width (μm)Median length/widthMedian biovolume (μm3)Median biosurface (μm2)Median surface/volume
  1. All results reflect the average of 250 cells analyzed by CMEIAS.

  2. ap-Value below 0.05 for 5%NaCl.

4 °C1/2 TSB1.80 ± 0.120.77 ± 0.032.35 ± 0.151.00 ± 0.125.23 ± 0.485.00 ± 0.31
 +5%NaCl1.74 ± 0.090.84 ± 0.03a1.99 ± 0.09a1.15 ± 0.135.59 ± 0.504.77 ± 0.14a
22 °C1/2 TSB1.62 ± 0.130.73 ± 0.032.25 ± 0.150.85 ± 0.12a4.44 ± 0.485.26 ± 0.30
 +5%NaCl1.86 ± 0.090.82 ± 0.03a2.20 ± 0.09a1.17 ± 0.135.74 ± 0.594.94 ± 0.14a

3.2Effect of increased osmolarity on growth of two Siberian permafrost isolates

Given the low water activity of permafrost, salt tolerance was assessed in Psychrobacter sp. 273-4 and Exiguobacterium sp. 255-15 using NaCl as an osmolyte. Addition of salt dramatically reduced the growth rates of both strains at 24 °C and somewhat for Exiguobacterium at 4 °C (Fig. 2). In contrast, the growth rate was relatively constant as a function of salt concentration for Psychrobacter sp. 273-4 at 4 °C. Growth rate and lag time did not differ significantly between cells grown in sucrose or NaCl as an osmolyte ensuring that ion toxicity did not influence growth capabilities (results not shown). As expected, lag phase increased for both strains with exposure to increasing salt concentration and low temperature. For Exiguobacterium, the lag phase increased from 0.67 h at 24 °C to 6.7 h at 4 °C (1/2 TSB), while the lag phase in high salt (1.78 m) increased more dramatically at 24 °C to 21 h when compared to 4 °C where an increase to 82 h occurred. Psychrobacter lag times also increased at 4 °C from 2.5 h to 22 h in 1/2 TSB, and incubation in 1.78 m salt increased lag times to 21 and 144 h at 24 and 4 °C, respectively.

Figure 2.

Comparison of average growth rates of two permafrost isolates at different temperatures and osmolarities. The μmax at 24 °C of Exiguobacterium sp. 255-15 was 0.9.

3.3Membrane composition under stress

Requirements for membrane fluidity at permafrost temperature and water activity should be reflected in membrane composition for organisms adapted to these conditions. Fatty acid profiles for Psychrobacter sp. 273-4 included both straight chain and unsaturated fatty acids but branch chained fatty acids were not detected. The overall composition of unsaturated fatty acids increases at lower temperatures and when grown in presence of 5% salt (Table 2a). The dominant fatty acid was a C18, saturated at 24 °C and unsaturated in the presence of 5% NaCl and 4 °C (Table 2a). A C18:2 fatty acid was present between 1 and 3% under all the growth condition (not shown). An unsaturated C16 was dominant at subzero temperatures alone. An increase in the amount of C17 correlated with increasing salinity, at the expense of C18 at 4 °C. The second most abundant fatty acid species were the C16 methyl esters. Saturated C18:0 was predominant at 24 °C, while in the presence of NaCl or 4 °C growth unsaturation was favored. However, the combination of NaCl and 4 °C resulted in an increase in the amount of C17:1 at 4 °C rather than C16:1 as seen in the single stress conditions (Table 2). Phosphatidylglycerol and phosphatidylethanolamine were detected at 4 and 24 °C. Diacylglycerol and additional spots not corresponding to standards were seen at 4 °C under both conditions (data not shown). In addition, spots were present at the origins that were not soluble in the TLC solution.

Table 2.  Fatty acid methyl ester composition (% of average total peak area) of two permafrost isolates at different temperatures and osmolarities
24 °C4 °C−2.5 °C24 °C4 °C−2.5 °C 
  1. Average of three biological replicates.

  2. aStandard deviation < 0.5.

  3. b0.5 > Standard deviation < 2.0.

  4. c2.0 > Standard deviation < 5.0.

  5. Indicates trace amounts below 5% of total percentage.

[a) Average percent fatty acid methyl ester profiles of Psychrobacter sp. 273-4
Straight chain fatty acids
% of total72.513.760.22.53.926.3
Unsaturated fatty acids
% of total3.775.
[b) Average Exiguobacterium sp. 255-15 fatty acid methyl ester profiles
C15:0 Iso7.0a4.5a10.5a6.8b1.8a11.3b
C17:0 Anteiso4.0b7.0c1.6b4.2a4.7b0.9a
C17:0 Iso16.5a26.0b6.6a22.0b18.2b7.5b
Straight chain saturated fatty acids
% of total32.320.067.436.623.7597
Branched chain fatty acids
% of total27.537.518.733.024.7197
Unsaturated fatty acids
% of total6.43215.533

In Exiguobacterium sp. 255-15, the presence of 5% NaCl or low temperature conditions alone shifted the fatty acids from saturated to unsaturated as expected at 4 °C, however unsaturated fatty acids were not detected from three biological replicates grown at −2.5 °C (Table 2b). C16:0 was the predominant fatty acid in Exiguobacterium sp. 255-15 at mesophilic and subzero temperatures but at 4 °C a shift occurred to isoC17:0. The interaction between 4 °C and salt results in a further increase in unsaturation with the predominant fatty acid becoming C18:1. Fatty acids below 5% of the total were not included in Table 2 and included: isoC13:0, anteisoC15:0, isoC15:0, isoC16:0, C12:0, C13:0,C14:0, C17:0 and C22:0. The phospholipid profile contained phosphatidylglycerol, diacylglycerol and phosphatidylethanolamine at 24 and 4 °C, with additional unidentified spots. Unidentified carbohydrate containing spots were present when cells were grown at 24 °C. Only one spot, identified as phosphatidylglycerol, was shared in 4 and 24 °C grown cells (data not shown).

3.4Polysaccharide composition

The soluble polysaccharide composition of Psychrobacter sp. 273-4 and Exiguobacterium sp. 255-15 differed with temperature and presence of 5% NaCl (Table 3). Psychrobacter sp. 273-4 polysaccharides consisted of rhamnose, glucose, mannose, galactose, xylose, fucose, ribose and arabinose when grown at 24 °C. When the cells were grown in 5% NaCl arabinose increased and fucose decreased. Lower temperatures resulted in an increase in predominance of glucose, mannose and a decrease in rhamnose (Table 3).

Table 3.  Soluble polysaccharide composition of two permafrost isolates grown at different temperatures and osmolarities
SugarPsychrobacter sp. 273-4Exiguobacterium sp. 255-15
24 °C4 °C−2.5 °C24 °C4 °C−2.5 °C24 °C4 °C−2.5 °C24 °C4 °C−2.5 °C 
  1. Ua, unknown amine.

  2. aStandard deviation < 2%.

  3. b2% > Standard deviation < 5%.

  4. c5% > Standard deviation < 10%.

  5. dND, not detected.

Glucose 23b27.9c36c28b27.9c39.2c27.5b59c50.4c22.5c35.9b16.7b 

The predominant sugar in Exiguobacterium sp. 255-15 at all three temperatures was glucose. Rhamnose, galactose, mannose, ribose, fucose, arabinose, xylose and an unknown amine sugar were also present (in decreasing percentages). Ribose decreased in dominance with low temperature, while arabinose and mannose increased. The presence of 5% NaCl resulted in a shift with mannose and arabinose (−2.5 °C) increasing in percent composition at the expense of galactose and arabinose (Table 3).

3.5Freezing tolerance

Twelve different permafrost isolates (including Psychrobacter and Exiguobacterium) were examined for their ability to survive freezing at −20 °C after a period of 1 year (Table 4). Most striking is that all strains showed excellent survival rates with 105–108 CFU/ml found from 108 CFU/ml after 1 year. Prior cold acclimation (growth at 4 °C) increased freeze survival in 9 of 12 strains, although only 3 strains (Planococccus sp. 215-68, 45-18 and Rathayibacter sp. 190-4) were statistically significant. Both results suggest that these strains are already adapted to subzero environments.

Table 4.  Effect of growth temperature on survival of permafrost isolates after one year at −20 °C
Strain Average cell loss after 365 days at −20 °C (log CFU/ml)Significance of growth temperature (p-value)
 Growth temperature (°C) 
Arthrobacter sp.255-
Arthrobacter sp.LTER2.31.00.04
Arthrobacter sp.33-
Exiguobacterium sp.392-
Exiguobacterium sp.190-
Exiguobacterium sp.255-
Flavobacterium sp.309-371.31.70.58
Planococcus sp.109-
Planococcus sp.215-680.62.50.03
Planococcus sp.45-
Psychrobacter sp.215-510.71.40.35
Rathayibacter sp.190-

3.6Temperature dependent carbon utilization

Psychrobacter sp. 273-4 and Exiguobacterium sp. 255-15 were tested for their ability to use 95 different carbon sources at either 4 or 24 °C (Table 5). Psychrobacter sp. 273-4 utilized 32 different carbon sources at 24 and 4 °C. Six of these carbon sources were used only at 4 °C, while 12 were used only at 24 °C. Exiguobacterium sp. 255-15 was able to utilize 42 different carbon sources at 24 °C, while at 4 °C only 36 were used. Seven of these carbon sources were used only at 4 °C and 13 carbon sources were used exclusively at 24 °C (Table 5).

Table 5.  Carbon sources uniquely utilized at the indicated temperatures
Psychrobacter sp. 273-4Exiguobacterium sp. 255-15
24 °C4 °C24 °C4 °C
d-Arabitold,l-lactic acidl-FucoseInulin
d-GalactosePropionic acidα-Methyl d-glucosideTween 40
Gentibioseα-Ketoglutaric acidPalatinoseTween 80
m-Inositiolα-Ketovaleric acidβ-Hydroxybutyric acidAmygladin
α-d-Lactoseβ-Hydroxybutyric acidd-Malic acidRaffinose
d-Glucuronic acidγ-Aminobutyric acidl-Malic acidd-Xylose
d-Saccharic acid Pyruvic acidAcetic acid
Glucoronamide d-Alanine 
l-Aspartic acid l-Alanine 
l-Arabinose l-Asparganine 
d-Gluconic acid l-Glutamic acid 
Cellobiose Cellobiose 

3.7Temperature effect on antibiotic susceptibility

Six different permafrost isolates were tested for naturally occurring resistance at 4 and 24 °C to five antibiotics (Table 6). Psychrobacter sp. 273-4 isolates showed no resistance to any antibiotics tested. Arthrobacter sp. 45-3 resistance was maintained at 4 °C with the five antibiotics tested, while Arthrobacter sp. 255-12 showed decreased resistance to ampicillin, chloramphenicol and tetracycaline at 4 °C. Only Exiguobacterium sp. 7-3 and Planococcus sp. 215-68 showed decreased resistance to streptomycin at 4 °C. No strains showed differential resistance to erythromycin with temperature. Resistance was maintained at 4 °C only in those strains possessing a high level of resistance at 24 °C (Table 6).

Table 6.  The susceptibilities of selected permafrost strains to five classes of antibiotics
Strain AmpicillinChloramphenicolErythromycinStreptomycinTetracycline
 10 μg30 μg15 μg10 μg30 μg 
 4 °C24 °C4 °C24 °C4 °C24 C4 °C24 °C4 °C24 °C 
  1. All three replicates gave consistent categories.

  2. aS, susceptible.

  3. bR, resistant.

  4. cRS, intermediate resistance.

Arthrobacter sp.45-3SaSRbRRScRSRSRSRSRS
Arthrobacter sp.255-12SRSSRSRSRSRSRSSRS
Exiguobacterium sp.255-15SSSRSRSRSRSRSSRS
Exiguobacterium sp.7-3SSSRSRSRSRSRRSRS
Planococcus sp.215-68SSSSSSSRSSRS
Psychrobacter sp.273-4SSSSSSSSSS

3.8UVC resistance

Because ionizing and UV radiation share some similarities in DNA damage, we tested the capacity of two permafrost isolates and reference mesophiles to remain culturable after exposure to UVC. The percent survival of Psychrobacter sp. 273-4 grown at 24 °C was significantly greater than those grown at 4 °C at all dosages (Table 7). Psychrobacter and E. coli B606 had similar sensitivities to UVC at mesophilic temperatures, except at the highest UV dose where E. coli was more resistant to UVC (p < 0.05). Exiguobacterium sp. 255-15, as well as B. subtilis PY79, exhibited more UVC tolerance than Psychrobacter sp. 273-4, in cells grown at 24 and 4 °C. Exiguobacterium sp. 255-15 grown at 4 °C was more sensitive to UVC than when grown at 24 °C at both 100 (p < 0.005) and 250 J m−2 (p < 0.02). Exiguobacterium sp. 255-15 at 24 °C exhibited significantly greater survival than B. subtilis PY79 at both the 100 (p < 7.9× 10−5) and at the 250 J m−2 (p < 8× 10−4).

Table 7.  Average percent survival of permafrost isolates exposed to UVC
Strain and temperatureUVC dose (J/m2)
  1. ND, not determined for this UVC dose.

  2. ap-Value (24 °C > 4°C) < 0.1.

  3. bp-Value (E. coli> 273-4) < 0.05.

  4. cp-Value (24 °C > 4°C) < 0.01.

  5. dp-Value (24 °C >B. subtilis) < 0.02.

Psychrobacter sp. 273-4 4 °C10026.9 ± 13.9a13.3 ± 6.6a0.29 ± 0.2bND
Psychrobacter sp. 273-4 24 °C10074.7 ± 22.6a37.1 ± 9.6a1.46 ± 0.8bND
E. coli B606 24 °C10068.6 ± 8.434.2 ± 0.45.30 ± 1.0bND
Exiguobacterium sp. 255-15 4 °C100NDND2.20 ±. 32c0.01 ± 0.01c
Exiguobacterium sp. 255-15 24 °C100NDND91.1 ± 13.1c,d2.72 ± 1.97c,d
B. subtilis PY79 24 °C100NDND11.7 ± 1.6d0.08 ± 0.01d

3.9Ice nucleation activity

Nine different permafrost isolates were tested for their ice nucleation activity, which may provide protection from harmful intracellular ice accumulation. Five isolates (Flavobacterium sp. 309-37 and 23-9, Psychrobacter sp. 215-51 and 273-4, Sphingomonas sp. 190-14 and 3361-2) possessed weak ice nucleation activity when grown at 24 °C. Exposure to 4 °C further increased ice nucleation activity in all strains except Psychrobacter sp. 215-51 and Sphingomonas sp. 3361-2 (Table 8). This cold shock treatment also induced INA in Exiguobacterium sp. 7-3, which exhibited no measurable ice nucleation ability at 24 °C. Cold shock at −10 °C resulted in a further increase in ice nucleation of in all six of the above strains (Table 8).

Table 8.  Ice nucleation activity (×10−9) per cell of selected permafrost isolates after growth at 24 °C followed by exposure to 4 and −10 °C
Strain Exposure temperature
 24 °C4 °C−10 °C 
  1. Mean value of two replicates. If no standard deviation given there was no difference in number of drops frozen between replicates.

Exiguobacterium sp.7-3010.5 ± 0.118.2 ± 0.2
Flavobacterium sp.23-93.4 ± 0.110.4 ± 0.0561.5 ± 0.4
Flavobacterium sp.309-377.4 ± 0.417.5 ± 0.12.8 ± 0.04
Psychrobacter sp.273-4252 ± 4.3360 ± 18170 ± 12
Psychrobacter sp.215-5116.2 ± 3.40.8 ± 0.13.4 ± 0.05
Sphingomonas sp.190-1410.4 ± 1.2314 ± 2017.0 ± 0.8
Sphingomonas sp.3361-254.8 ± 5.619.4 ± 0.215.2 ± 0.1
Escherichia coliMigula000
Pseudomonas syringaeSijderius411411411


Permafrost provides an opportunity to obtain microbes that have experienced long term exposure to cold temperatures, decreased water activities, γ radiation and low carbon availability. The Psychrobacter and Exiguobacterium strains, as diverse representatives of the permafrost community, should carry traits that have allowed them to adapt to these conditions. The genomes of these two organisms are currently being sequenced and the characterization of physiological traits potentially important to cryo-adaptation is important for beginning understand these adaptations at the genome and proteome levels.

Psychroactivity is common for bacteria isolated from cold environments such as sea ice. One such isolate, Psychromonas ingrahamii, grows at a temperature of −12 °C with a generation time of 240 h, the lowest growth temperature of any organism authenticated by a growth curve [27]. Recently, growth of another Psychrobacter from Siberian permafrost was reported at −10 °C (0.016 day−1[28]), the temperature of the Kolyma permafrost. Tolerance to low water activity coupled with sub-zero growth, including the minimum growth temperature prediction of −7 and −15 °C suggests that Exiguobacterium sp. 255-15 and Psychrobacter sp. 273-4 could be active in their native habitat.

Our Psychrobacter and Exiguobacterium isolates grew over a moderate to broad temperature range, (−15 °C calculated Tmin) −5 to 26 °C and (−7 °C calculated Tmin) −2.5 to 42 °C, respectively, the later being especially interesting because of its temperature span of at least 45 °C. Permafrost strains typically must live for thousands of years in the surface (active) soil layer with annual freeze and thaw cycles of +10 °C to −20 °C, which may have been the selective force for the broad growth range typical of permafrost isolates.

These two isolates exhibit different growth patterns as determined by the linear portion of the Arrhenius plot. Switch points in the slope of growth rates, indicative of a physiological shift in metabolism, occurred at 22 °C in Psychrobacter sp. 273-4, 24 °C in Exiguobacterium sp. 255-15 and also at 6 °C in Psychrobacter. These two isolates exhibit different growth characteristic as has been reported in Pseudomonas fluorescens MFO where the “thermometer temperature” of 17 °C triggers a change in physiological characteristics such as increased membrane permeability to β-lactamine and an increase in enzymatic activity of extracellular protease, lipase and periplasmic phosphatases [29]. While cell sizes, antibiotic susceptibility and membrane composition were not assessed at the switch points in our study, marked differences in these traits were seen between cells grown at 4 and 24 °C.

Minor but significant cell morphology and size changes in Psychrobacter sp. 273-4 occurred with exposure to low temperature and increased salt. The decrease in surface to volume ratio observed in Psychrobacter sp. 273-4 with low temperature and salt is similar to that seen in other bacteria isolated from low water activity environments [30]. An increase in surface to volume ratio is seldom encountered in cells isolated from low water activity environments since the presence of compatible solutes, which balance the external and internal solute concentrations, tend to swell cells. In contrast, another permafrost isolate, Arthrobacter sp. 45-3, showed a large decrease in cell size (14-fold) after incubation at 4 °C [18]. Members of the genus Arthrobacter are known to change cellular morphologies with stress and development stages, suggesting that this may be a common adaptation in this genus [31].

Low temperature and high osmolarity are known to induce changes in membrane composition that maintain membrane fluidity [32,13]. A decrease in temperature resulted in a decrease in the saturation of fatty acids and acyl chain lengths in Psychrobacter sp. 273-4 and Exiguobacterium sp. 255-15, as occurs in other bacteria shifted to lower temperatures. Together these changes function to lower the gel-liquid crystalline transition temperature, i.e., homeoviscous adaptation [33]. Unsaturated membrane fatty acids are less abundant in these psychroactive Siberian permafrost isolates grown at mesophilic temperatures compared to the psychrotolerant, Oleispira antarctica. This difference in unsaturation may account for the Psychrobacters narrower range of growth temperatures [34]. These Siberian permafrost isolates show a more dramatic shift to unsaturated fatty acids than previously described for a permafrost isolate, Arthrobacter sp. 33-1 [18]. Exiguobacterium sp. 255-15 exhibits the same shift to shorter branched chain fatty acids, with an increase in anteiso-C15:0 at the expense of anteiso-c17:0, which was also seen in Arthrobacter sp. 33-1 and L. monocytogenes[35].

Exposure to increased osmolarity resulted in increased fatty acid unsaturation and acyl chain length in these Siberian permafrost isolates. Moderate halophiles have been described as showing increased membrane fatty acid saturation [36] rather than the increased unsaturation seen in this experiment. The present study concurs with the trend shown in Vibrio costicola indicating that the combination of high salinity and low temperatures results in increased unsaturation of fatty acids and an increased ratio of phosphotidylethanolamine: phosphotidylglycerol; both of these metabolic changes are commonly seen only under low temperature conditions [37]. The changes in sugar composition with temperature and salinity reflect additional adaptations, likely in the membrane-associated exopolysaccharide that may improve survival. Recent reports of exopolysaccharide accumulation in sea ice supports the stabilizing role for the cold active Cowellia ColAP, thought to aid environmental survival [38].

An important consequence of membrane changes at cold temperatures is the effect on membrane transport which could explain the differences in carbon source utilization and antibiotic sensitivity seen at low temperatures. Glucose utilization increased at 0 °C in an unidentified cold-adapted psychrotroph [39]. Uptake rates of inline image in psychrophilic Vibrio increased when the organism was grown at temperatures between 0 and 15 °C compared to 24 °C, resulting in an increased Vmax for the inline image transport system [40]. Alternatively, cold shock increased the activity of malate dehydrogenase and glucose-6-phosphate dehydrogenase in Lactococcus lactis and Rhizobium[41,42]. In addition to alterations in enzyme efficiency and in Vmax of uptake, a more universal change in membrane permeability may explain the differential use of certain carbon sources in the permafrost bacteria tested. Studies on glycerol uptake, which is transported both by facilitated and passive diffusion, indicate an increase in both types of uptake in E. coli when membrane fluidity is increased [43]. Molecules, such as cellobiose, may not enter the cell at 4 °C due to changes in transport-associated proteins which are influenced by membrane composition. Additional substrates may also be bound when membrane and protein flexibility increases as demonstrated recently for the arabinose binding protein [44]. An increase in flexibility commonly occurs in proteins at low temperatures raising the possibility that additional substrates could be bound by other binding proteins. If a single transporter handles a range of substrates then decreased efficiency of such an enzyme would decrease uptake at low temperatures. The presence of temperature dependent nutrient uptake efficiencies suggests that microbes have developed adaptations to low temperatures which will counteract the unfavorable effects of decreased diffusion, allowing adapted microbes to be more competitive for nutrients under unfavorable conditions.

Continual exposure to sub-zero temperatures within the permafrost for organisms with long-term freezing survival, a trait noted in all permafrost isolates tested. Cryotolerance has been reported in another Arctic-isolated Pseudomonas when pre-conditioned at 4 °C prior to freezing at −20 °C for 24 h [9]. All permafrost strains possessed excellent survival rates with bacterial numbers of 105–108 CFU/ml observed after 1 year at −20 °C from an original population of 108 CFU/ml. This is substantially higher than E. coli[45], although it is similar to the survival reported for the food-associated, L. lactis[46]. Pre-conditioning to cold temperatures is known to increase freezing survival in many food-associated microbes, including E. coli O157:H7 [45] and L. lactis[46] though after only 28 days a one log decrease in survival occurs. This effect may be due to expression of cold-responsive genes and cryoprotectant molecules that function to enhance the survival of the microbe through the stress of freezing and thawing conditions [47–49]. However pre-conditioning to cold temperatures, does not increase freeze survival in all bacteria, as seen in this experiment and for lactic acid bacteria [47]. Permafrost organisms could already be selected for life in this continuously frozen environment since they don’ t need to respond to temperature cycles.

Ice nucleation activity, enhanced by exposure to lower temperatures, may be another mechanism of survival in permafrost. Ice nucleation activity results from outer membrane proteins whose structure mimics an ice crystal, providing a lattice for crystallization of water [50] at higher temperatures, and preventing harmful intracellular ice formation because the crystals are too large to penetrate the membrane and initiate intracellular freezing [51]. The formation of these small, thermodynamically unstable molecules makes ice nucleation proteins unlikely to offer long-term survival due to the crystals’ tendency to reform into large, damaging structures [52]. The discovery of a protein capable of both ice nucleation and antifreeze activity in Pseudomonas putida may present a solution to this dilemma. The ice nucleation domain is believed to cause the formation of extracellular ice, while the antifreeze domain maintains the crystals at a non-damaging size [53]. The ability of some plant and animal antifreeze proteins to inhibit microbial ice nucleation activity suggests that antifreeze proteins may lower the supercooling point of water within the cells. However, no direct evidence supports this theory [54,55]. The presence of small amounts of thermal hysteresis activity in Psychrobacter sp. 273-4 suggests it may possess antifreeze proteins or a combination antifreeze/ice nucleation protein, as seen in γ-Proteobacterium relative, P. putida (J. Duman, personal communication). The detection of ice nucleation activity in the Gram positive Exiguobacterium sp. 255-15 is surprising since the ice nucleation protein is located within the outer membrane in all bacteria known to date. Ice nucleation activity was maintained after filtering (results not shown) suggesting a soluble protein may be responsible in Exiguobacterium sp. 255-15.

Exposure to UVC under two growth temperatures was used to evaluate the ability of Psychrobacter sp. 273-4 and Exiguobacterium sp. 255-15 to survive DNA damage resulting from reactive oxygen species and occasional single stranded DNA breaks such as would occur from ionizing radiation. A decrease in UVC survival at 4 °C relative to 24 °C was observed in both strains, and UVC survival during growth at low temperature was not greater than that of either mesophilic comparison strain. This result appears to indicate that unusual capacity to withstand DNA damage was not necessary for either organism to survive and be resuscitated in the lab after 104–106 years in permafrost. One must conclude that some level of repair must have been occurring in situ. Recently, Price and Sowers compiled data indicating that in situ survival metabolism, defined as a metabolic state in which cells “can repair macromolecular damage but are probably largely dormant”, is higher than theoretical rates of DNA depurination over decreasing temperature [56] their calculated rates of survival metabolism were found to be approximately 10−6 times lower than that required for measured metabolic rates required for growth at similar temperatures. This very low metabolic requirement could be met in situ by Psychrobacter sp. 273-4 or Exiguobacterium sp. 255-15 and hence allow them to repair DNA damage.

The ability of permafrost isolated bacteria to respond to laboratory- simulated permafrost conditions suggests that these organisms possess adaptations to low temperature, increased osmotica and have efficient repair mechanisms that allow for these and not other tundra organisms to continue to live in permafrost.


This research was funded by the National Astrobiology Institute of NASA. We would also acknowledge the assistance of Chia-Kai Chang, Gisel Rodriguez, Debora Rodrigues and Alexa Turke. We also thank Marcia Lee for assistance with the ice nucleation activity studies, Frank Dazzo for CMEIAS expertise, George Sundin for assistance with UVC exposure experiments and Rich Lenski and Lee Kroos for gifts of strains E. coli 606 and B. subtilis PY79 respectively. All GC/MS analysis was performed by Beverly Chamberlin at the Center for Mass Spectrometry of Michigan State University.