The deep cold biosphere: facts and hypothesis


Corresponding author. Pyrieva St., 5-13-147, Moscow 119285, Russia. Tel.: +7 (95) 939-3179; Fax: +7 (95) 938-0672; E-mail:


Deep subterranean layers may be regarded as the most stable environment for microorganisms where possible fluctuations should be explained by geological events only. The analysis of the total amount of microorganisms has revealed that in sedimentary deposits their number is only one order of magnitude lower than the same parameter in soil. Taking into account the depth of sediments the microbial biomass in subterranean rocks has to be considerably larger than that in soils. Permafrost is the most constant and stable environment among deep habitats. Microbial communities survive in permafrost for at least some millions of years. The diversity of organisms and of microbial activities after permafrost thawing displays distinct differences to those in soils. The abundance of the bacterial biomass assumed is comparable in frozen and unfrozen sediments. Therefore, the permanently low temperature in permafrost is a stabilizing factor that sustains life in deep cold biotopes. Studies of microbial communities in permafrost sediments of different lithology and age suggest that the level of subzero temperature and the duration of its influence define the ratio between the hypometabolic cells, readily reversible to proliferation, and the so-called viable but non-culturable cells (deep resting cells). To a certain extent, cell aggregates in the extracellular matrix may be regarded as an additional survival mechanism supporting the hypometabolic state of cells. There is indirect evidence for adaptive physiological and biochemical processes in microorganisms during the long-term impact of cold.


At present we have easy access to greatly advance our knowledge of subterranean environments concerning the specific techniques developed at the forefront of research in soil science and biogeochemistry. Microbial life is obviously widespread within the whole depth of the earth's crust and is essentially independent of the surface conditions [1–3].

Recent data indicate that microbial life at subterrestrial levels is much higher in mass and volume than that in soils. The total number of bacterial cells in deep sediments varies from 106 to 109 cells/g dw (Table 1), therefore in sediments with an average total bacterial number of cells (107–108 cells/g dw), the microbial biomass in a layer 20–50 m thick would be equivalent to the biomass (108–109 cells/g dw) in the upper soils in moderate climatic areas.

Table 1.  Number of bacteria in natural environments
EnvironmentNumber of bacteriaReferences
 Total counts (cells/g dw)Viable counts (cfu/g dw) 
Soils 108–1010 105–108[4, 5]
 up to 10 m 107–1010 102–106[6]
 up to 550<106–108<102–108[2]
Permafrost sediments   
 Arctic (>300 m) 107–109 102–108[3, 7, 8]
 Antarctic (20 m) 107–108 10–104[9]
Permafrost buried soils 109 104–106[7]
Bottom sediments of lakes   
 upper layers 108–109 104–105[10]
 layers at 3–8 m 108 102–103 
Central part of the oceancells/mlcells/ml[11]
 water 102–106  0–10 
 upper layer of sediments 107 102–106 
Groundwater (up to 2000 m) 103–106 10–105[2]
Antarctic glaciers (up to 1800 m) single cells/l[12]

The deep microbiocenoses differ from those in soils: the former are predominantly bacterial, whereas in soils the fungal biomass is usually much larger than that of bacteria. Although yeasts, actinomycetes and algae exist in deep layers, they are not important for the microbial activity in these sites. Deep rocks and sediments are bacterial habitats, which have been stable for a long time span. It is not known whether these bacterial communities are active and well adapted, whether they just survive in an anabiotic state or how long they remain viable.

A more reliable indication may be drawn by comparing ‘open’ and ‘closed’ deep environments, e.g. some clay sediments, loesses, may be regarded as ‘closed’ systems. But the most intriguing sites are permafrost sediments that have not thawed for thousands or millions of years. Theoretically, these sites should be largely different from sediments and present strong arguments supporting the anabiotic way of survival.

Permafrost areas are widespread on the earth and often extend to considerable depths. These types of environments seem to serve as potential and actual reservoirs of biogeochemical activity, concealing the biogeochemical adaptation to long-term freezing in nature. Nevertheless, until recently only few attempts have been made to investigate the microbial activity in frozen subsoil layers.

Numerous facts supporting the survival of microorganisms in permafrost have been reported by Russian, American and Canadian researchers (see reviews [3, 7, 8]). Unfortunately, sampling techniques described in earlier reports are hardly adequate to exclude the possibility of sample contamination.

Since 1985, the systematic research on permafrost subsoil layers at different depths with techniques preventing sample contamination have provided new information on life preservation over long periods [3].

2Materials and methods

More than a thousand frozen sediments formed in the late Pliocene-Holocene were sampled from 100 bore holes down to 300 m in East Siberia, Yamal, Alaska, North Canada, and Antarctica (down to 20 m). Systematic studies of natural microbial systems at subzero temperatures have been carried out in the Kolyma-Indigirka Lowland, one of the world's last vast ecologically clean area still unaffected by human activity. This region is characterized by large areas of low-temperature (−9 to −12°C) permafrost 400–900 m in depth. The average geothermal gradient in the entire frozen layer is 1.5°C/100 m, and close to zero in the upper 200 m. There are no water horizons or infiltrations, and the influence of climatic factors is limited to the depth of the seasonal thaw layer (0.5 m). Distinct cryotextures confirm the lack of migration of water and other substances by diffusion.

The age of sediments was determined from radiocarbon, paleomagnetic, palynological and paleontological analyses. The late Cenozoic deposits in this region concern the most ancient signs of the beginning of the formation of the lithosphere in the present Northern Hemisphere. They date back to the second half of the Pliocene (3–5 million years) which is characterized by cryogenic disruptions and fossil pollens. Sediments of this age are silty, loamy sands and light loams, often containing peat, detritus layers, partially decayed plants, humus spots, and carbonaceous inclusions (the level of organic carbon ranging from 0.35 to 5.7%). Large numbers of spicules and diatoms document the aqueous origin of the sediments, whereas low salinity, calcium hydrocarbonate and sulfate composition point to a limnetic and alluvial origin. The absence of traces of marine transgressions eliminates the possibility of permafrost thawing under ocean waters. The climatic conditions prevented any significant degradation of frozen rocks. Therefore, the sediments frozen in the late Pliocene never thawed later as a result of climate fluctuations or under the impact of geological factors.

The late Pleistocene icy silty loamy sands and loams of the Edoma layer (QW 2–4), widespread from the surface to a depth of 30–50 m, with stratified and net cryotextures formed under conditions of distinct cryochrone 15–40 thousand years ago. Thick polygonal vein ices, accounting for about 50% of the overall rock volume, confirm the permanently frozen state of these layers.

Sample contamination was excluded by the absence of marker organisms in the central part of frozen cores when the drills, the samplers and frozen samples were contaminated with marker cultures. Other indications were: (a) mixing of layers with introduction of soil microflora did not happen because of the frozen grounds; (b) the presence of an anaerobic microflora in the samples which could not have been introduced from the air; (c) the isolation of rare strains; (d) the presence of samples lacking live microorganisms. The sampling techniques, storage and transportation of samples have been described previously [3, 7, 8, 13]. Antarctic sediments were collected in the Dry Valleys polar desert: in Taylor Valley (eolian and fluvial sediments, up to 150 thousand years, 17 m, −18°C), in Miers Valley (lacustrine sediments, age undetermined, 4 m, −21°C), and on Mt. Feather (ancient soil, Sirius formation, at least 2 million and possibly over 15 million years, 2–3 m, −25 to −27°C). The total organic matter varied from 0.0 to 0.43%, Eh from +260 to +480 mV and pH from 7.8 to 9.81.

Thus, the oldest permafrost samples (3–4 million years) were isolated in Siberia, the deepest ones (down to 300 m) in the Canadian Arctic; the lowest average temperature of sediments in situ was found in Antarctica.

The total count of microorganisms was obtained by microscopy and staining with acridine orange (AO). Microbial growth was analyzed using a wide range of nutrient-poor and nutrient-rich media at different pH values with additional supply of vitamins, yeast extract, and microelements. Noticeable growth was only observed on rich media (e.g. TSB).

Kinetics of colony formation on nutrient media made it possible to estimate the state and viability of microorganisms in permafrost [14]. A new type of growth parameter, λ, describes the kinetics of the proliferation of individual cells. This parameter explains the probability of a cell producing a colony per unit time. Another parameter, tr, is the time lag until a visible colony is formed. The model of colony formation is expressed by:


where N(t) is the number of colony-forming units (cfu) at time t, and N the final number of cfu.

The cfu were counted on five replicates every 2–3 h after the first appearance of a colony and expressed as the percentage of the total number at the end of the experiment. The succession of bacterial communities in thawed samples taken over 60 days was studied at constant humidity.

Most anaerobic processes were measured in suitable media by the products formed by gas chromatography (methane), ion chromatography, radioactive tracer techniques (sulfate reduction), and spectrometry (iron reduction).

The multisubstrate testing method [15, 16]was used to characterize the potential functional activity of microbial communities in permafrost sediments. Based on the BIOLOG technology, this method allows structure-functional studies of soil microbial communities by ‘fingerprinting’[15–17]. A substrate set of 47 carbon sources was used (Table 2).

Table 2.  Sole carbon sources for the Multisubstrate Testing method
Carbohydratesm-Inositol, l-arabinose, l-rhamnose, dulcite, d-sorbitol, α-lactose, d-mannitol, d-maltose, d-glucose, sucrose, xylose, pullulan, d-fructose, raffinose
Carboxylic acidsAcetic, aspartic, citric, succinic, maleinic, propionic, octanoic, valeric, malonic, d,l-lactic
Amino acidsNorleucine, l-cysteine, histidine, norvaline, threonine, alanine, asparagine, valine, serine, phenylalanine, lysine, glutamic acid, arginine, aminobutyric acid, aminopropionic acid
PolymersStarch, Dextran 500, Tween-80
AmidesCarbamide, acetamide

Homogenized 2-g portions of permafrost samples were suspended in 100 ml of phosphate buffer (pH 6.5), sonicated (1 min, 22 kHz, 0.1 A) and centrifuged. Aliquots (0.2 ml) of soil suspension were used to inoculate microplates containing carbon sources (1%), mineral salts and tetrazolium violet. The color development was measured 72 h after inoculation using a microplate scanner. The results were standardized [17]and the data processed by multivariate statistics, principal component analysis, multidimensional scaling and image recognition algorithms [15, 16, 28]. All the data were statistically tested at P=0.05 and compared with modern soil databases.

3Results and discussion

Permafrost sediments from polar regions contain large numbers of viable microorganisms. The total number of bacterial cells in samples of different lithology and age varied from 107 to 109 cells/g dw. Even in the Arctic and Antarctic samples lacking microbial growth, the AO method revealed 107 cells/g dw. This amount is regarded as the minimal number of microorganisms in all known terrestrial environments, thus permafrost differs neither from unfrozen and filtered sediments, nor from poor recent soils. Hence, the long-term cold stress has not been fatal for the native microbial communities, and microbial evolution under constant cold permafrost conditions through some millions of years can be retraced.

The viable cell number in permafrost samples varied significantly (10≤1–107 cfu/g dw). The analysis of Arctic permafrost samples showed that an average of 0.1–10% of viable cells accounted for by microscopy with AO showed growth on nutrient media. The increase in the permafrost age led to decreasing numbers of cultured cells from frozen sediments, and more samples which seemed to be ‘sterile’ were found. Despite the overall lower biodiversity in frozen rocks, various microbial communities were preserved at some sites of ancient permafrost, even exceeding the cell number found in more recent permafrost sediments.

In both ancient and young Antarctic sediments, only 0.001–0.01% of total cells grew on nutrient media. The low average temperatures of the Antarctic subsurface layers (−20 to −27°C) partially explains the deep resting state of bacterial cells.

Different morphological and physiological groups of aerobic bacteria have been identified. They are responsible for nitrification, nitrogen fixation, proteolysis, urea decomposition, and other processes after thawing [18, 19]. Aerobic microbial communities from permafrost grew within a wide range, from −10 to 50°C and even 70°C (only few samples were tested at temperatures below 0°C). The temperature resistance was a common characteristic of these communities. In the sediments with fine texture, the number of viable aerobic bacteria was generally 2–3 orders of magnitude higher than in sands and was characterized by a more diverse microbial spectrum. The library of isolated bacterial strains comprised more than 30 genera, including both spore-forming and non-spore-forming species of the following genera: Bacillus, Arthrobacter, Micrococcus, Cellulomonas, Rhodococcus, Flavobacterium, Pseudomonas, Aeromonas, Myxococcus, Exiguobacterium, Nitrobacter, Nitrosomonas, Nitrosospira, Streptomyces. In Antarctic sediments, heterotrophic bacteria of the genera Bacillus, Arthrobacter and Streptomyces were dominant.

Bacterial isolates were mostly psychrotrophic and mesophilic, some strains were even strict psychrophilic or non-strict thermophilic. Among the aerobic bacteria the non-spore-forming and non-halophilic forms were predominant.

The high ratio of total counts to viable counts in some permafrost sediments may be regarded as evidence of a high content of anaerobes. Field analyses revealed the reduced state (Eh from +40 to −256 mV) of most of the sediments. Compared to sterile controls, such sediments often contained ferrous iron and acid-soluble sulfide and displayed denitrifying activity and methane production from H2 and CO2 (MCH); methane production from acetate (MA) was detected as well [20]. The number of viable methanogenic bacteria varied from 101 to 104 cells/g dw, and in some permafrost locations touched even 105–107 cells/g dw.

Most of the Antarctic samples contained methane (up to 670 μl/kg) presumably of biological origin as judged from its isotopic composition (13C=54.8%); methanogens were also detected [9].

Iron-reducing bacteria (103 cells/g dw) were found only in sediments of moderate age along with iron-oxidizing bacteria (102–103 cfu/g dw). Despite the low number of sulfate-reducing microorganisms (101–104 cfu/g dw), radio-tracer experiments revealed a considerable potential activity of sulfate-reducing bacteria in the sediments. After 6 months of incubation the (H2S)/(SO2−4) ratio was 41.4% at 15°C and 27.0% at 4°C.

Culturable anaerobic bacteria exist in most permafrost samples in quantities comparable with aerobic bacteria.

Under natural conditions the freezing processes determine the activity of microbial communities by the level of exposure and the exposure time. The ratio between culturable and non-culturable microbial forms from permafrost is ideal for estimating the physiological response of microorganisms to extreme environmental conditions. Furthermore, the origin of the non-culturable portion of microbes is explained both by the selection of stable non-culturable forms during freezing and long-term temperature action and by the development of adaptation mechanisms in the cells. The mechanism of microbial survival in permafrost becomes clear by the comparative investigation of physiological parameters of the microbial communities when their activity is recovered after thawing of the permafrost sediments and the temporarily frozen tundra soils.

The kinetics of colony formation [14]reveals that the probability of cell proliferation and colony formation per unit of time (λ) both in melting cryogenic tundra soils and in permafrost sediments is very high (Fig. 1) compared to non-frozen sediments [21]and soils [14, 22]. In addition, multiplying cells from permafrost sediments show high values in the parameters of viability in spite of a relatively low number of cells.

Figure 1.

Kinetics of colony formation on nutrient media. t, time of incubation of melting samples (A: t=0, B: t=8 days); tr (h), time lag for visible colony formation; λ (h−1), probability of visible colony formation per unit time.

The hypothesis of ‘cryptic growth’ is not sufficient to explain these data. The ‘explosive’ growth of microorganisms when thawing manifests the metabolic activity and physiological adaptation of long frozen cells for a rapid recovery of essential metabolic processes after melting. These remarkable differences between the soil and permafrost samples were preserved during successions of microbial communities up to 60 days (Fig. 1B). Growth of microbial populations on nutrient media stimulated temporal fluctuations of viability parameters (λ), time lag of the formation of visible colonies (tr) and of the generation time. The positive correlation between the viability parameters (λ) and the cell generation time explains the way of the survival of microorganisms under permafrost conditions: when the rate of metabolic processes declines the generation times become geologically prolonged, and the inherent cell energy is increased due to endogenous metabolic recovery. Interactions of cells within populations (aggregates, microcolonies) may give additional opportunities for enhancing the viability. It starts to decline when cell interactions within communities reach their limit, and as a consequence die or transform into a deep resting state.

The analysis of the kinetics of colony formation reveals the joint influence of cryogenic factors, and in particular the effect of prolonged cold stress, on the viability of microorganisms under environmental conditions. The salient feature of the microbial communities after their release from the frozen state both in thawing permafrost and in thawing soils is the high rate of cell proliferation. It may be assumed that during freezing the vegetative cells pass to a reversible state of low metabolic activity (hypometabolism) under the growing influence of different stress factors (cryobiosis, osmobiosis, oxidizing stress) and develop anti-stress mechanisms ensuring rapid restoration of their biological activity after thawing. Further endogenous regulatory processes inside the cells stimulate the transition to the deep resting state or the transformation into the so-called viable but non-culturable (VBN) cells [23], a process probably independent of external factors. An increase in the number of non-culturable cells after freezing is expected, resulting in a prolonged impact of cold.

The relative amount of non-culturable microorganisms in Antarctic samples was higher than that in recent soils. In the Arctic permafrost, the percentage of microbes capable of restoring their activity is unusually high, one or two orders of magnitude larger than that in soils. Thus, the level of subzero temperature determines the ratio of hypometabolic (culturable) cells to VBN cells (resting forms) in natural environments. Freezing down to −10 to −12°C in some Arctic permafrost sediments may further stimulate the reverse transition of part of the non-culturable cells to the hypometabolic (culturable) state. It seems that a prolonged stable temperature is important for the adaptation. Under such conditions the hypometabolic cells will recover, and a balance between very low multiplying and dying cells is achieved. The development of this model over prolonged time periods is shown in the kinetics of the quantitative changes in the hypometabolic (low proliferating?), dying and VBN (resting, anabiotic) cells. When the energy resources are exhausted, the majority of hypometabolic cells revert to the deep resting state (VBN) conserving their genome.

In the Arctic permafrost, the smaller number of sites with a high content of culturable cells and the increasing amount of ‘sterile’ samples with increasing permafrost age confirm this hypothesis. At lower temperatures (−25 to −30°C in Antarctica) these processes clearly occur at higher rates, probably due to the low content of organic matter and the absence of plant residues in the Antarctic sediments.

The differences in the dynamics of microbial proliferation after thawing in permafrost sediments and soils result from the increase in the proportion of hypometabolic (culturable) to deep resting (VBN) cells at temperatures of about −9 to −12°C and the respective increase in the percentage of multiplying cells. Furthermore, the cause of such differences is probably the low metabolic reorganization inside the hypometabolic cells, as their physiological adaptation to long-term freezing ensures a prompt recovery of microorganisms and their high biological activity after thawing.

Our previous studies have revealed differences in the kinetics of RNA and DNA syntheses in the thawing permafrost and cryogenic soil samples [19]. The major feature of the culturable cells from permafrost is, in contrast to the soil-based microorganisms, that the systems of RNA synthesis are ready for immediate activation as the respective genetic structures are conserved.

Electron microscopy in situ [24, 25]showed that microorganisms in different types of subsoil sediments including permafrost have similar features. 70–80% of the cells did not reveal signs of lysis. The majority remained inside the soil particles, formed cell aggregates and had two specific layers: an electron-transparent layer surrounding the cell walls and an electron-dense surface of fibrillar appearance connecting the cells as aggregates. The ability to form conglomerates was lost after incubation on nutrient agar at room temperature, but the capsular surface layers resisted numerous passages of isolated strains. This property distinguishes the permafrost bacterial strains from the subsoil isolates. In sediments, some cells have thick walls and were capsular and cyst-like, others were L-shaped and lacked cell walls. There was a limited number of typical spore-forming cells or of mature spores. The main feature in all cases was additional surface layers of low electron density around the cell walls. These cells did not reveal signs of multiplication and are regarded as forms with slow metabolism (Fig. 2). The preservation of undamaged cell structures in frozen sediments correlated with the high resistance to freezing and thawing up to 100 cycles (Fig. 3A,B). Among the isolates from Arctic sediments still viable after such a stress, a few spores were detected, others were coryneforms or Gram-negative. Typical ultrastructures were the capsular surface layers around the cells and cell conglomerates. Capsular layers seem to play an important role in the protection of ‘permafrost’ strains to extreme factors. Antarctic sediments were characterized mainly by spore-forming bacteria that had a similar ultrastructure to known Bacillus strains (Fig. 4).

Figure 2.

Cell structure of Gram-positive bacteria isolated from permafrost, characterized by surface thick layers and well preserved internal cell structures. ×60 000.

Figure 3.

Cells after 100 cycles of freezing and thawing without visible lesions in cell structures (×60 000). A: Gram-negative bacteria. B: Gram-positive bacteria.

Figure 4.

Cell of the sporeforming bacteria isolated from Antarctic sediments. ×60 000.

The eukaryotic cells are less able to survive long-term cryoconservation, and the deep subterranean sediments are not suitable for their survival. Marked lesions in the cell ultrastructure of eukaryotes in the sediments correlated with a failure to obtain isolates from these samples on nutrient media.

Permafrost subsurface layers preserve eukaryotes in considerably lower numbers than prokaryotic cells. In contrast to bacteria, where viability is not exactly correlated with the age of the permafrost, the number of viable eukaryotic microorganisms considerably decreases in ancient sediments.

Micromycetes are easily cultured from recent Arctic and Antarctic permafrost sediments; mycelial fungi did not grow after inoculation with material from ancient frozen rock formations from the second part of the Pliocene.

The most frequently isolated eukaryotic microorganisms from permafrost were yeasts (Fig. 5), though they were not permanent components of these microbial communities. Statistically, yeast species were found more frequently in the young permafrost sediments with a high content of plant residues, in particular in peats. In some samples from the ancient Oleorian suite of syngenetic permafrost type (2–3 million years), up to 20–25% of culturable aerobic heterotrophs (103–104 cfu/g dw) were yeasts of the genera Cryotococcus (Cr. albidus, Cr. gilvescens), Rhodotorula (Rh. rubra), Saccharomyces. A similar type of yeast population with Cr. gilvescens as dominant and Cr. albidus as minor species is quite characteristic of modern tundra soils [15]. Our data confirm their outstanding resistance which allowed survival through at least several millions of years. The yeasts were mainly psychrotrophs. Repeated freezing-thawing cycles of samples partly reduced their abundance, but they were exceptionally resistant to low freezing (−30°C, −196°C).

Figure 5.

Cells of yeast with thick capsular layer isolated from the ancient Arctic sediment of 600 thousand years old. ×20 000.

Viable green algae were isolated from the Arctic deep sediments frozen for 5–7 thousand years (Table 3). All isolates grew at a low rate (20–25°C) and were sensitive to high light intensities. Photosynthetic pigments, chlorophyll a, chlorophyll b, and pheophytin were found in a wide range of sediments of different genesis and age, including the most ancient ones [8]. In 20 Antarctic samples no plant residues were found, and the attempts to isolate living algae were unsuccessful, although some traces of photosynthetic pigments were detected.

Table 3.  Preliminary identification of algal isolates from Arctic permafrost sediments (Kolyma Lowland, Siberia, well 1/95)
Depth (m)MorphologyOrganismOrder (genus)
  1. aIsolates were identified according to [26].

  2. bIsolates were identified according to [27].

  3. cisolates were not identified.

 1.6–1.65endospore forming, unicellularblue-green algaePleurocapsalesa
 2.4–2.45filamentous endospore formingblue-green algaeNostocales (Nostoc)b
 unicellularblue-green algaePleurocapsalesa
 unicellulargreen algaec
10.1–10.15endospore forming, unicellularblue-green algae
10.2–10.15unicellulargreen algaePleurocapsales
10.32–10.35unicellulargreen algae
 filamentousblue-green algaeNostocalesb
13.7–14.2filamentousgreen algae
14.8–14.9filamentousblue-green algaeNostocales (Anabaena)b

The functional diversity of microbial communities was analyzed using multisubstrate testing techniques. The data, when compared with a large database of sole-carbon-source utilization in different modern soil types by factor analysis and multidimensional scaling algorithms [14, 15], revealed significant differences between the sediment samples and the recent soils (Fig. 6c,d). According to the pattern recognition analysis [15], all thawed sediments were characterized by increasing aspartic acid and decreasing alanine utilization. All cryogenic samples (both tundra soils and permafrost sediments) were similar regarding the high lactic acid uptake and relatively low glycerol consumption (Fig. 6a,b). The spectrum of substrates utilized varied depending on the age of the deposits (Tables 4 and 5).

Figure 6.

a: Differentiation of permafrost samples in glycerol-lactate indices. b: Differentiation of permafrost samples in alanine-aspartate indices. c: Age group differentiation for permafrost samples. Multidimensional scaling in all 47 indices. Age of sediments: 1=<10 thousand years; 2=20–40 thousand years; 3=400–600 thousand years; 4=600–1800 thousand years; 5=3–5 million years. d: Differentiation of permafrost samples by multidimensional scaling in aspartic acid, glutamine, arginine, fructose, creatinine indices. For age of sediments see c.

Table 4.  Key rules for differentiation of age of permafrost samples
Age (years)Index
<10 000glucose >2.7
<40 000octanic acid >2.7
600 000mannitol=0
600 000–1 800 000asparagine >2.96
3–5 millionmaltose <0.3
tundra soilsaspartic acid <2.4
Table 5.  Key rules for differentiation of permafrost samples
permafrost sediments+tundra soillactic acid >2.7 and glycerol <1.2
permafrost sedimentsaspartic acid <2.44 and alanine <2.63

A relatively poor spectrum of substrate consumption and low levels of utilization of polymers and carbohydrates were characteristic of all permafrost samples (Fig. 7). The dispersion analysis [28]of the substrate uptake confirmed that microbial communities in permafrost should be regarded as highly stable (Fig. 8), except for one Arctic sample (600 000 years) that we considered a buried soil.

Figure 7.

Utilization of various substrate classes in permafrost samples.

Figure 8.

Stability of microbial communities from permafrost samples (estimation of substrate utilization data vector dispersion).

Free extracellular enzymes were also found in permafrost sediments (Fig. 9). Among these, invertase activity was found in all 120 Arctic and Antarctic samples. Amylase, protease and catalase activity were often present. Statistically significant data for deposits of different genesis and age showed that not only sediment-accumulated enzymes (invertase, catalase), but also the same enzymes synthesized de novo by resuscitated communities were extremely thermostable (active after heating of samples or mixed liquid cultures during 3 h) and had a high temperature optimum (Fig. 10). The stability of free extracellular enzymes in different soil types, including tundra soil, is lower [29]. Invertase and some catalase activity were detected in permafrost samples where no culturable microorganisms were found. This makes some enzymes attractive as prospective biomarkers.

Figure 9.

Enzyme activity (A=invertase, B=amylase, C=catalase) in modern tundra soil and in permafrost sediments of different lithology and age. 1=tundra soil; 2=silty loamy sands, 7–10 thousand years; 3=loams (edoma), 20–40 thousand years; 4=sandy sediments, 15–50 thousand years; 5=clay sediments, 150 thousand years; 6=alluvial sands and loamy sands (maastakh), 200–600 thousand years; 7=silty loamy sediments (oleor), 0.6–1.8 million years; 8=silty loamy sediments (oleor), 1.8–3 million years; 9=sandy sediments (thomus-yar), 3–5 million years.

Figure 10.

Invertase activity depending on the temperature. I: Temperature optimum of invertase activity in mixed microbial cultures: 1, liquid culture; 2, immobilised culture. II: Thermostability of invertase in cultures (activity in % after heating of cultures during 3 h). III: Thermostability (activity in % after heating of samples during 3 h) of sediment-accumulated invertase: 1, from tundra soil; 2, from permafrost.

These facts provide additional evidence for low metabolic activity and prolonged metabolic reconstruction in hypometabolic cells that exhibit an extremely high resistance and hence viability in permafrost. The suggested low activity of culturable cells in permafrost explains the lack of differences in the protein sequences of RNases secreted by two strains of Bacillus sp. from ancient (2 million years) permafrost as compared with bacilli from other sources [30].


Inside the subsoil sediment layers (down to a few hundred meters) the temperature varies a few degrees from above zero under warm and temperate soils down to −10 to −12°C in Arctic cryogenic regions and to −25 to −30°C in Antarctic sedimentary deposits. Isotherms apparently have considerable depressions with depth in the regions of permafrost that occupy large areas on the terrestrial surface, and consequently the biosphere boundary in these regions is expanded.

In most cases, deep subterranean layers may be regarded as the most stable environments for microorganisms where possible fluctuations are explained only by geological events. The total number of microorganisms present was only one order of magnitude lower in sedimentary deposits than the same parameter in soils of climatically moderate regions. As to the depth of the sediments, the total amount of microbial biomass in subterranean rocks is considerably larger than in soils.

Permafrost is the most static and balanced environment where microbial communities survive for at least some millions of years. The diversity of cells and microbial activities after thawing are different from soils, but similar to those in unfrozen subterranean deposits. The abundance of presumably bacterial biomass is also comparable in frozen and unfrozen sediments. Therefore, the permanently low temperature in permafrost may not be regarded as an extreme but as a stabilizing factor sustaining life in deep cold biotopes. Studies of microbial communities in permafrost sediments of different lithology and age suggest that the level of temperature and its duration define the ratio between the hypometabolic cells, readily reversible to proliferation, and the so-called viable but non-culturable cells (deep resting cells). Cell aggregates in the extracellular matrix form an additional survival mechanism.

It is impossible to keep permafrost in its natural state. Some degradation of microbial communities occurred during the preservation of samples. Indirect evidence confirms the presence of adaptive physiological and biochemical processes in microbial cells during the long-term impact of cold:

  • the large percentage of culturable cells in Arctic permafrost,
  • the high viability (probability of colony formation per unit of time) and, as a consequence, high rates of cell proliferation and of biochemical processes after thawing,
  • a high rate of RNA synthesis induced by immediate thawing,
  • extremely high thermostability of microbial enzymes synthesized de novo by mixed microbial cultures isolated from permafrost samples and high thermostability of these in permafrost samples (free extracellular enzymes),
  • microbial growth on nutrient media at temperatures below 0°C,
  • detection of unstable metabolites and their producers (e.g. detection of unstable nitrite- and ammonia-oxidizing bacteria).

Permafrost microbial and biochemical markers can be used in paleoenvironment reconstitution studies. The analyses of the Antarctic permafrost sediments formed some 20–30 million years ago surveying the entire age range of terrestrial permafrosts will be challenging and will provide information on microbial evolution as well as on geological and planetary history. Permafrost is the best deposit of paleolife, and the natural environment where microorganisms realize unknown possibilities for adaptation to prolonged cold action. Such possibilities are common to the Universe or at least to the Solar System.


  • 1

    This work was supported by the Office of Polar Program, Grant of the National Science Foundation of USA (PI, E.I. Friedmann). The samples were collected during the expedition to Antarctica in 1995.