Are permafrost microorganisms as old as permafrost?

ABSTRACT

Permafrost describes the condition of earth material (sand, ground, organic matter, etc.) cemented by ice when its temperature remains at or below 0°C continuously for longer than 2 years. Evidently, permafrost is as old as the time passed from freezing of the earth material. Permafrost is a unique phenomenon and may preserve life forms it encloses. Therefore, in order to talk confidently about the preservation of paleo-objects in permafrost, knowledge about the geological age of sediments, i.e. when the sediments were formed, and permafrost age, when those sediments became permanently frozen, is essential. There are two types of permafrost—syngenetic and epigenetic. The age of syngenetic permafrost corresponds to the geological age of its sediments, whereas the age of epigenetic permafrost is less than the geological age of its sediments. Both of these formations preserve microorganisms and their metabolic products; however, the interpretations of the microbiological and molecular-biological data are inconsistent. This paper reviews the current knowledge of time–temperature history and age of permafrost in relation to available microbiological and metagenomic data.

permafrost, syngenetic, epigenetic, age, microorganisms

INTRODUCTION

Earth is a planet of cryogenic type where all four elements (snow, ice/sea ice, glaciers and permafrost) of the cryosphere are present (Rivkina et al. 2018). Based on data from the U.S. Geological Survey (Williams and Ferrigno 2012), ∼15.6 × 10⁶ km² of Earth's surface is covered by glaciers; the extent of seasonal snow cover in the Northern Hemisphere varies from 46.9 × 10⁶ to 3.5 × 10⁶ km², and floating sea ice covers vast areas of the polar oceans, ranging in extent from ∼5 × 10⁶ to ∼15–18 × 10⁶ km² in both the Northern and Southern Hemispheres. Permafrost regions are geologically and ecologically diverse and occupy ∼18% of the Earth's exposed land surface (Zhang et al. 2000).

Available studies have demonstrated that Earth contains permafrost of different ages. The most ancient permafrost, formed 15.15 ± 0.02 million years ago (Ma), was found in Nibelungen Valley, Antarctica, as evidenced by a lack of authigenic clay minerals in the frozen volcanic ash deposits (Marchant and Denton 1996). Permafrost found in the Northern Hemisphere is younger. Specifically, permafrost in North America is not older than 800 000 years (ka). Permafrost in the Canadian High Arctic was estimated to be relatively young, ∼10 ka (Ziolkowski et al. 2012), followed by 19–33 ka permafrost from the CRREL Tunnel, Alaska (Shur et al. 2004; Mackelprang et al. 2017), up to 500 ka permafrost in northern Alaska Prudhoe Bay (Lunardini 1995) and ∼740 ka permafrost in central Yukon Territory, Canada (Froese et al. 2008). The most ancient, older than 1 Ma, relict permafrost in the Arctic was found in the Kolyma–Indigirka region of northeast Siberia, Russia (Sher 1971; Kaplina 1981).

The key features of permafrost, besides continuously negative temperatures, are a very low content of liquid water present as ∼5 × 10⁻¹⁰ m thick films of brine (Gilichinsky, Soina and Petrova 1993) and a possibility for conservation of indigenous microorganisms for thousands to millions of years. Permafrost is used as a paleoarchive to reconstruct the past conditions in accordance with preserved communities, to explore survival limits for viable cells and Deoxyribonucleic acid (DNA)/Ribonucleic acid (RNA) preservation, or to investigate the adaptation mechanisms that allow microorganisms to stay active in the permafrost and compare them to their counterparts thriving in contemporary environments at both positive and negative temperatures. Understanding the permafrost-related peculiarities is crucial for interpreting the viability of microbial communities found in subfreezing environments. It is also important for designing a specific research strategy for various biotechnological and exobiological approaches as well as for searching for the oldest perennially frozen soils on Earth.

The purpose of this article is to help microbiologists to understand a potential relationship between the age of microorganisms embedded in permafrost and the age of the surrounding permafrost. In order to provide an accurate interpretation of age for both syngenetic and epigenetic permafrost, we review records of consecutive periods of glacial epochs (cryochrons) and interglacial warm epochs (thermochrons), along with their duration and conception, and the time when the last freezing period occurred. We also summarize available data on deposit accumulation, permafrost formation and dating for the three best-studied sites located in the northeastern Siberian Arctic, where the most complete geological cross-sections from the late Pliocene to the Holocene have been discovered (Konishchev 2001; Melles et al. 2012). We then summarize quantitative data on the presence and number of microorganisms in permafrost of different ages and discuss cell adaptability, its subzero metabolic activity and maintenance to answer questions about the age of permafrost microorganisms.

PERMAFROST FORMATION AND EXISTENCE

Permafrost existence and its present-day distribution are a result of the past temperature history and the present equilibrium of the energy balance at the Earth's surface (Lunardini 1995). Previous glaciations have played an important role in the present existence of permafrost. Significant glaciation of the Earth's surface began in the late Oligocene (∼35 Ma) in eastern Antarctica, followed by Miocene (23–5.3 Ma) mountain glaciation in Alaska, Greenland, Iceland and Patagonia, and later glaciation in the Pliocene (5.3–2.6 Ma) in the Alps, the Bolivian Andes and Tasmania (Ehlers and Gibbard 2011) that resulted in deterioration of temperature. Ehlers and Gibbard (2011) showed that past continental glaciations occurred repeatedly and were responsible for the cyclic climate changes.

Conditions ideal for the formation and continuous existence of permafrost had developed in the terrestrial area bordering the Arctic Ocean in the Pliocene. This is evidenced by temperature decline (Horikawa et al. 2015) with the mean annual air isotherm being below −8°C (Osterkamp and Burn 2015), the absence of glaciation (Baranova and Biske 1964; Baranova et al. 1968) and the progressive changes in vegetation from a forest to fully developed tundra (Repenning and Brouwers 1992). Discoveries of pebbles with traces of frost cracking in Begunov suite and the fine sand and peat pseudomorphs (replacement of one mineral with another) along ice wedges (ground ice) in Kutuyakh suite provide evidence for the existence of permafrost in the northeastern Arctic in the late Pliocene ∼3–3.4 and 1.9–2.4 Ma, respectively (Mackay 1990; Konishchev 2001). Studies of paleovegetation and paleoclimate revealed sinusoidal temperature variations (Repenning and Brouwers 1992; Lea, Pak and Spero 2000) and indicated that the first cooling period ended abruptly ∼2.1 Ma with the returning of a warmer climate. The second cooling period began ∼1.7 Ma, and it culminated with the Ice Age, beginning 850 ka (Repenning and Brouwers 1992). Faunal, floral and depositional records indicate an oscillation of cold and warm periods during the second cooling cycle between 1.7 and 1.2 Ma with warmer climates about every 100 000–300 000 years (Sher 1971; Repenning and Brouwers 1992). High-resolution natural records reflecting glacial and interglacial climate intervals for the past 2.8 million years were reconstructed using geochemical and biological indicators in the sediment core from the Lake El'gygytgyn in northeastern Russia (Melles et al. 2012). The most significant warm periods described in the Lake El'gygytgyn borehole were at intervals of 1.08–1.06, 0.42–0.35, 0.13–0.12 and 0.01 Ma (Melles et al. 2012).

Perennially frozen deposits formed during the Pliocene continued to exist at depths through the Pleistocene in spite of repetitive warm epochs interspersed with long cold periods that triggered permafrost formation (Ehlers and Gibbard 2011; Melles et al. 2012). Much of today's Siberian Arctic permafrost originated during the late Pliocene and Pleistocene cold periods (Opel et al. 2017) when the climatic conditions were ∼5°C lower than today's temperature (Konishchev 2001; Melles et al. 2012). The late Pleistocene–early Holocene cold period changed ∼9–10 ka by the Holocene Climate Optimum, which induced the retreat of high-altitude glaciers, permafrost degradation and the development of numerous lake depressions formed as a result of thawing ground ice (Grosse et al. 2010). In the late Holocene ∼5 ka (Koshkarova and Koshkarov 2004; Marcott et al. 2013), climate deterioration, lake drainage and refreezing of lake sediments occurred, leading to the formation of young Holocene permafrost (Gubin and Lupachev 2008). The climate oscillations led to the formation of permafrost of different ages.

ESTIMATION OF PERMAFROST AGE DEPENDS ON TYPE OF PERMAFROST FORMATION

The age of permafrost is not always easy to determine because of iterative changes in paleoclimatic conditions and cycles of thawing and freezing over long intervals associated with them. Nevertheless, the age of permafrost is defined as the time that has elapsed since the freezing of the soil system (Lunardini 1995; French 2013). Therefore, the age of permafrost may be equal to the age of its deposits or less than the age of its deposits. Based on the relation between deposition and freezing, there are two main types of permafrost—syngenetic and epigenetic. The age estimation is different for syngenetic and epigenetic permafrost as evidenced by studies reviewed below.

Syngenetic permafrost

Syngenetic permafrost, which is characterized by the presence of massive wedge-like ice inclusions (Mackay 1990), was formed under cold climatic conditions as the soil material was gradually deposited at the surface during freezing conditions (Shur et al. 2004) at the mean ground temperature of −23 ± 3°C (Konishchev 1997, 2001). The syngenetically frozen silty sediments with large ice wedges, with a thickness of a few tens of meters, often referred to as ‘Ice Complex’, can be found in Alaska, western Yukon Territory, and are widely distributed in northeastern Siberia (Shur et al. 2004; Schirrmeister et al. 2013). Given that freezing occurred simultaneously with deposition (Sher, Virina and Zazhigin 1977; Gilichinsky et al. 2007), the age of syngenetically formed permafrost is considered to be equal to the age of the particular deposit. Therefore, the age of syngenetic permafrost is estimated by determining the deposit age using radiocarbon dating of soil organic matter (Schirrmeister et al. 2013) or dating of ice wedges using ³⁶Cl/Cl isotope ratios (Gilichinsky et al. 2007; Blinov et al. 2009). The age of syngenetic ice-rich permafrost in northeastern Siberia is not more than 60 ka as determined by the radiocarbon dating method (Schirrmeister et al. 2013). The syngenetic permafrost deposits along Dmitry Laptev Strait dated by ²³⁰Th/U and ³⁶Cl isotopes as well as by microtine biochronology have an age of 200–400 ka (Schirrmeister et al. 2002; Gilichinsky et al. 2007; Blinov et al. 2009; Tumskoy 2012). The oldest syngenetic permafrost with an age of 740 ± 60 ka was discovered in central Yukon Territory, Canada (Froese et al. 2008; Millar and Lambert 2013).

Epigenetic permafrost

Epigenetic permafrost, on the contrary, was formed from soil that had existed before freezing started (Lunardini 1995), sometimes with time lag of thousands or millions of years (French and Shur 2010). Therefore, the age of epigenetic permafrost is less than the age of the host deposits. The freezing of epigenetic permafrost starts from the ground surface progressing downward. Calculations showed that if surface temperature drops to −10°C for infinite time, it takes 2 years to form 28 m and 243 ka to form 332 m deep permafrost layer, respectively (Lunardini 1995). The thickness of epigenetic permafrost ranges from <200 to 650 m in Alaska (Osterkamp and Payne 1981), reaches down to 540 m in Nunavut, Canada (Onstott et al. 2009), and >1500 m in northeastern Siberia (Konishchev 2001). Epigenetically frozen lacustrine silt permafrost with ice lenses up to 5–10 cm in size was found in west-central Alaska (Kanevskiy et al. 2014) and in Nunavik, Northern Québec, Canada, where segregation ice makes up ∼65% of the permafrost volume (Calmels, Allard and Delisle 2008). The buried ice and massive ice bodies could be formed in epigenetic permafrost (Mackay 1990) if the downward freezing front encountered a significant source of groundwater (Gilbert, Kanevskiy and Murton 2016). In the Siberian Arctic, epigenetically frozen permafrost is represented by ice‐poor sediments formed during the Pliocene global cooling (Gilichinsky et al. 2007). Estimating the ratio of the cosmogenic radionuclide ³⁶Cl and stable Cl concentrations used for dating up to 600-K-old groundwaters (Love et al. 2000) appears to be a good method for dating ancient permafrost. The methods based on measurements of ³⁶Cl/Cl (Blinov et al. 2009) and ³⁶Cl/¹⁰Be (Gilichinsky et al. 2007) ratio were evaluated in regard to age estimation of old ground ice. However, application of such methods to quantitative dating of epigenetic permafrost is limited by (i) low volume of epigenetically formed ground ice and (ii) influence of NaCl brines within permafrost on the initial ³⁶Cl/Cl (precipitation controlled at time of deposition) ratio (Gilichinsky et al. 2007; Blinov et al. 2009). Therefore, for the age estimation of the epigenetic permafrost, scientists rely on climatic trends and indirect proofs. For example, the fact that formation of the syngenetic permafrost requires frozen ground supports a statement that epigenetic permafrost underlying the middle and late Pleistocene syngenetic permafrost is obviously older.

HISTORY OF PERMAFROST AT THREE SELECTED SITES ON THE KOLYMA LOWLAND

Data from the three best-studied sites located in the northeastern Siberian Arctic (Fig. 1) were summarized in this study to visualize and explain correlation between the age of sediments and the age of permafrost in this region.

Figure 1.

The location of key sites: A: Bykovsky Peninsula, Lena Delta area; B: Alazeya River, Kolyma–Indigirka Lowland; C: Chukochy Cape, Kolyma–Indigirka Lowland.

Bykovsky Peninsula, Lena Delta area

Ice-rich silty loams of the Yedoma suite (Fig. 2A) on Bykovsky Peninsula, Lena River Delta, are formed syngenetically and contain ice wedges (Sher et al. 2005; Grosse et al. 2010), providing evidence that those sediments have been never completely thawed. Formation of the Yedoma deposits and their paleoenvironmental significance were extensively reviewed previously (Schirrmeister et al. 2013). Extensive ¹⁴C chronology indicates age for these sediments to be between 60 and 12 K (Schirrmeister et al. 2011). The Yedoma sediments are overlaid by Holocene permafrost formed epigenetically ∼5 ka (Gubin and Lupachev 2008).

Figure 2.

General cross-sections and age of permafrost and deposits. (A–C) The sites shown in Fig. 1. Legends: 1: sandy loam; 2: loam; 3: sand; 4: sandy loam and sand layers; 5: gravel; 6: ice wedges; 7: cryopegs; 8: baydjarakhs; 9: marine shells remnants; 10: mammal bones; 11: peat; 12: outcrops; 13 and 14: types of permafrost; 13: syncriogenic; 14: epigenetic.

Alazeya River, Kolyma Lowland area

In the Alazeya River basin, the ice-rich silty loams of the Yedoma suite up to 42 ka (Shatilovich et al. 2018) form the upper part of the section (Fig. 2B) and overlay the older permafrost of the Olyor suite. The geological interpretation of alluvial-lacustrine sandy silt sediments uncovered by a drill hole (Fig. 2B) suggests the Olyor suite sediments were formed between 0.8 and 1.4 Ma (Kaplina 1981; Kaplina, Lahtina and Rybakova 1981). The age estimation was based on the research of mammal fauna associated with the Olyor sediments exposed in the middle section of the Alazeya River (Kaplina, Lahtina and Rybakova 1981) and the records of the microtine rodents found in the Olyor sediments of the Krestovka section (Repenning and Brouwers 1992). The age of the older part of the Olyor suite was suggested to be 1.5 Ma but not more than 1.67 Ma (Repenning and Brouwers 1992). The older stratigraphic units have been described as the Begunov suite of early Pliocene age and the Kutuyakh Beds of late Pliocene age (Sher 1980; Repenning 1992; Repenning and Brouwers 1992). The Begunov suite formed around 3–3.4 Ma are the oldest sediments with evidence of seasonal freezing (Konishchev 2001). The Tumus Yar and Kutuyakh Beds formed between 1.8 and 2.2 Ma, and 1.9 and 2.4 Ma, respectively (Konishchev 2001), are distinguished by emergence of periglacial landscapes and general cooling of the climate (Giterman 1975; Giterman, Sher and Matthews 1982; Repenning and Brouwers 1992) and contain evidences of permafrost (Repenning and Brouwers 1992). Low ice content in the Olyor suite and the underlying horizons of the Tumus Yar suite suggests that both Olyor and Tumus Yar permafrost horizons were formed epigenetically. Unfortunately, the lack of massive ice layers renders impossible the use of ³⁶Cl/Cl and ³⁶Cl/¹⁰Be isotope dating (Gilichinsky et al. 2007; Blinov et al. 2009) for these permafrost formations. Consolidated knowledge of climatic records available for northeastern Siberia reconstructed from the Lake El'gygytgyn 2.8 Ma record (Melles et al. 2012) allows for an age estimation of the epigenetic permafrost from the upper horizon of Tumus Yar suite of ∼1.1 ± 0.3 Ma and older for deeper horizons (Repenning and Brouwers 1992; Konishchev 2001).

Chukochy Cape, Kolyma Lowland area

The permanently frozen marine sediments of Kon'kovaya suite (Fig. 2C) at the Chukochy Cape were formed epigenetically ∼100–120 ka after the East Siberian Sea level had dropped and the marine sediments that were deposited in an anoxic environment of the shallow littoral lagoons had frozen (for review see (Gilichinsky et al. 2003, 2005). The cold climatic conditions with temperatures 5–7°C lower than present day (Gilichinsky et al. 2005; Melles et al. 2012) triggered downward freezing of saline marine sediments. The latter is supported by registering increased salt concentrations with permafrost depth that are often attributed to migration of highly mineralized solutions from upper layers downward during epigenetic freezing (Anisimova 1978). This caused formation of a 20-m-thick sodium chloride marine permafrost horizon containing methane gas hydrates and lenses of unfrozen brine known as cryopegs, with salt concentration of 170–300 g L⁻¹ at in situ temperatures of around −8°C (Gilichinsky et al. 2003). The epigenetically frozen marine sands are overlaid by ice-rich silty loams of the Yedoma suite formed syngenetically ∼26–43 ka (Gilichinsky et al. 2003).

TEMPERATURE OF THE ACTIVE LAYER AND ANCIENT PERMAFROST

Permafrost sediments occur under the layer of upper surface soil, which experiences seasonal freeze–thaw cycles, and is referred to as the active layer or permafrost-affected soil. Long-term monitoring of active layer temperature profiles from 1980 to 2014 in the Kolyma and Yana–Indigirka Lowlands and the Bykovsky Peninsula showed temperature fluctuations from maximum 15.2°C during summer to minimum −22.7°C during winter with mean annual temperature variations from −4.3 to −9.7°C at 20 cm below the surface (Fedorov-Davydov et al. 2018). Based on the data obtained from the Global Terrestrial Network for Permafrost Database, the mean annual permafrost temperatures registered during 2007–2010 inside six boreholes located in Kolyma Lowland ranged from −5.6 to −6.3°C. The annual temperature variations in these boreholes were from −1.2 to −4.5°C in upper 5 m permafrost layer whereas in deeper permafrost sediments seasonal fluctuations were smaller from −0.02 to −1.4°C. The mean annual temperature of permafrost in the region strongly depends on latitude and landscape conditions. For example, on the high points of the Yedoma watersheds, it varies from −12.3°C at 72°50′N to −9.9°C at 69°30′N, while in the Holocene depressions, temperature is higher, with −9°C at 71°40′N and −7°C at 68°50′N (Romanovsky et al. 2010). The reconstruction of ground temperatures is still a matter for continuing discussion. However, the preservation of the ice wedges in the middle and late Pleistocene permafrost overlaid by several meters of younger Holocene permafrost allows us to assume that temperature of the late Pleistocene sediments was never above zero.

PERMAFROST MICROORGANISMS

The presence of various microorganisms in both syngenetic and epigenetic permafrost sediments collected using aseptic drilling (Khlebnikova et al. 1990; Juck et al. 2005) and subsampling techniques (Bang-Andreasen et al. 2017) was shown by different methods: microbiological (Vorobyova et al. 1997; Rivkina et al. 1998; Vishnivetskaya et al. 2000), microscopic (Soina and Vorobyova 2004), phylogenetic (Vishnivetskaya et al. 2006; Steven et al. 2007), real-time Polymerase chain reaction (PCR) (Yergeau et al. 2009) and next-generation metagenomics (Yergeau et al. 2010; Rivkina et al. 2016).

Microorganisms are present in permafrost as estimated by cell counting or quantitative Polymerase Chain Reaction (qPCR)

Counting cells under the microscope proved that permafrost from different locations and of different ages contains abundant microbial populations that are nevertheless smaller compared to those found in permafrost-affected soil. Total cell counts in samples collected on the Kolyma Lowland varied from 10⁹ cells g⁻¹ in organic-rich buried soils (Vorobyova et al. 1997) to 1.2 × 10⁸ cells g⁻¹ in sediments with lower carbon content (Vorobyova et al. 1997; Rivkina et al. 1998; Vishnivetskaya et al. 2000). In samples of the active layer from Samoylov Island (Lena Delta, Siberia), the total number of microbial cells gradually decreased from 2.3 × 10⁹ cells g⁻¹ in the top 5 cm layer to 1.2 × 10⁸ cells g⁻¹ in the 40–45 cm layer (Kobabe, Wagner and Pfeiffer 2004). The determination of cell number using qPCR demonstrated a decrease in the abundance of bacteria from 3.1 × 10⁷ to 1.6 × 10⁶ Ribosomal ribonucleic acid (rRNA) gene copy number g⁻¹; of Archaea from 1.5 × 10⁴ to 1.5 × 10²; and of fungi from 8.6 × 10⁴ to 8.8 × 10³ for the active layer and the 2-m permafrost, respectively, in samples collected near Eureka, in the Canadian High Arctic (Yergeau et al. 2009). The similar decrease of bacterial 16S rRNA genes with depth from 7 × 10⁸ copies g⁻¹ at 5 cm of active layer to 1 × 10⁸ copies g⁻¹ at 80-cm-deep permafrost was shown using qPCR for samples collected in Axel Heiberg Island, Nunavut, Canada (Chauhan et al. 2014). These examples suggest that part of the microbial population, presumably cells sensitive to freezing, died during the transition from active layer soil to permafrost. Differences in total cell counts between permafrost samples of different ages were below 5% (Vishnivetskaya et al. 2000). However, abundance of cultivable microorganisms in Siberian permafrost decreased gradually with the increasing of permafrost age from 10³to 10⁶ CFU g⁻¹ (CFU: colony forming unit) in Yedoma (∼10–40 ka), 10³to 10⁴ CFU g⁻¹ in Olyor (600 ka–1 Ma) and 10² to 10⁴ CFU g⁻¹ in Tumus Yar (>1 Ma) (Khlebnikova et al. 1990).

Microscopy visualization of microbial cells

Microscopy approaches visualized the presence of spores, cyst-like cells, intact cells and cell aggregates in different permafrost deposits frozen for up to 2 million years (Soina et al. 1995; Soina and Vorobyova 2004). Another piece of evidence of the presence of intact bacterial cells in 21–33 ka permafrost was obtained using the flow cytometry and cell sorting (Sipes et al. 2020). The live–dead staining method revealed that 18–22% of all cells had intact cell membranes and those cells were classified as alive in permafrost samples of 19–33 ka (Mackelprang et al. 2017). The natural permafrost sediment enrichments for 12 weeks at 4°C monitored with fluorescent microscopy showed that increases in CFU over the duration of the experiment were caused by increased cultivability of the resident bacteria (Vishnivetskaya et al. 2000). This suggests that some microbial cells may exist in anabiotic, or resting state, with greatly reduced metabolism and that these cells require extended time or specific conditions for revival (Vorobyova et al. 1997).

Culturable microorganisms are survivors

The number of isolates and their taxonomic diversity depended on the origin and age of the permafrost, as well as on the permafrost's properties, including total organic carbon concentration, salinity and pH (Gilichinsky, Wagener and Vishnivetskaya 1995; Vorobyova et al. 1997), and on nutrient media and laboratory conditions used for microorganisms recovery (Palumbo et al. 1996). Diverse microorganisms from three domains—Archaea, Bacteria and Eukarya—and viruses were isolated from permafrost; however, only a small number of isolates were taxonomically identified (Rivkina et al. 2018). Successful culturing of those microorganisms is direct evidence of life preservation in permafrost (Khlebnikova et al. 1990; Vishnivetskaya et al. 2000; Steven et al. 2007). The fact that it's possible to isolate antithetical microorganisms from the same permafrost sample suggests that microorganisms survive in microniches, which could provide environmental conditions favorable for maintaining a life (Price and Sowers 2004). However, the ancient microniches with no influx of energy or new materials (Gilichinsky, Wagener and Vishnivetskaya 1995; Mackelprang et al. 2017) would not be able to provide microhabitats sufficient for continuous reproduction and growth. In the absence of direct proof, we want to provide some circumstantial evidence that microorganisms do not multiply in permafrost. First of all, the lack of dominance of anaerobic bacteria and archaea in the predominantly anaerobic conditions of the reduced permafrost samples with a redox potential of +40 to −256 mV (Rivkina et al. 1998) suggests that there was no cell proliferation. This is further supported by the fact that various anaerobic (Rivkina et al. 1998) and aerobic (Shi et al. 1997; Vishnivetskaya et al. 2000) microorganisms co-exist in the permafrost. Second, the presence of viable photosynthetic cyanobacteria and green algae (Vishnivetskaya et al. 2020), which require free water and sunlight—both lacking in the permafrost—for their active metabolism, indicates that during their stay in permafrost these microorganisms were not metabolically active. The notion that cells do not propagate in permafrost and that individuals do not pass advantageous characteristics to their offspring is further supported by the absence of true psychrophiles (Gilichinsky and Wagener 1995; Suetin et al. 2009), which are ‘cold-loving’ microorganisms with an optimal growth temperature below 15°C, a minimum temperature below 0°C and lack of growth above 20°C (Morita 1975). Then, the fact that the number of CFUs gradually decreases, and the number of barren samples increases as a function of permafrost age (Khlebnikova et al. 1990), provides additional evidence of the absence of microbial cell division in permafrost. Finally, the viability of protists such as amoebas, ciliates and flagellates (Shatilovich et al. 2009; Shatilovich, Stoupin and Rivkina 2015; Shmakova, Bondarenko and Smirnov 2016; Malavin et al. 2020), and nematodes (Shatilovich et al. 2018) in permafrost further indicates that the microorganisms merely survive in permafrost and may represent living fossils, in other words an ancient surface community preserved through time.

Permafrost as a supportive environment

Permafrost conditions, such as subzero temperatures and low water activity (a_w, 0.8–0.85), seem to be amenable to the preservation of a fossil archive. However, experimental evidences for the existence of thin, ∼5 × 10⁻¹⁰ m, films of unfrozen water in frozen clay–water mixture (Anderson 1967) and the presence of around 1–7% unfrozen brine films in permafrost deposits (Gilichinsky, Soina and Petrova 1993) raise a question about the cryoprotective properties of permafrost. The thin brine films can provide an opportunity for permafrost microorganisms to perform active metabolism, interact and evolve continuously after being buried. This question was thoroughly discussed previously (Gilichinsky and Wagener 1995; Gilichinsky, Wagener and Vishnivetskaya 1995); however, due to the contradictory interpretations it continues to be a major unresolved question in microbial ecology.

Adaptation to permafrost environment

Numerous laboratory studies have provided examples of how permafrost microorganisms cope with environmental stresses such as freezing, negative temperature, nutrient deprivation, etc. and how a negligible volume of liquid brine contributes to the protection of cells embedded in permafrost. During the process of natural slow-rate freezing, soil microbes experience physiological adjustments and adaptive behaviors (Walker, Palmer and Voordouw 2006), that may include cell volume reduction, modifications in lipids and fatty acids to maintain membrane fluidity, accumulation of polyols, antifreeze and ice-binding proteins, synthesis of cold shock proteins and cold-active enzymes, genome adaptations, etc. (McGrath, Wagener and Gilichinsky 1994; Ewart, Lin and Hew 1999; Phadtare, Alsina and Inouye 1999). The cell membrane that is first to be impacted by environmental changes, indeed, reveals fast increase of monosaturated fatty acids and decrease of saturated fatty acids as shown in permafrost bacteria grown at 4°C relative to 15°C (McGrath, Wagener and Gilichinsky 1994). Differential scanning calorimetry did not show the formation of intracellular ice in cells frozen slowly to temperatures as low as −150°C, suggesting that permafrost microbes can contain an unfrozen intracellular solution when encapsulated in the frozen permafrost (McGrath, Wagener and Gilichinsky 1994). Growth conditions before freezing may significantly improve cryotolerance and survivability. Thus, Exiguobacterium sibiricum isolated from permafrost, grown in a low-temperature liquid medium or on an agar medium, regardless of growth temperature, showed elevated cryotolerance (Vishnivetskaya et al. 2007).

To adapt to conditions of the low temperature and low water activity encountered in permafrost, bacteria demonstrate a substantial change in cellular energy metabolism. For example, microorganisms may slow down metabolism and become reversibly dormant. Thus, eukaryotic microorganisms (filamentous fungi Geomyces spp. and yeasts Leucosporidium spp.) after a short period of growth at temperatures as low as −35°C showed the decline of metabolic activity to zero, and entered a state of reversible dormancy (Panikov and Sizova 2007). Similar ability to enter a dormant state was observed in non-spore-forming permafrost bacteria of the genera Arthrobacter and Micrococcus. These bacteria were able to produce an extracellular factor responsible for cell transition to anabiosis (Mulyukin et al. 2001) that provided improved survival at unfavorable conditions under prolonged exposure to subzero temperatures, low water activity and unavailability of nutrients (Soina et al. 2004). Other bacteria, for example, permafrost bacterium Exiguobacterium sibiricum grown at 4°C (Qiu et al. 2009b; Qiu, Vishnivetskaya and Lubman 2009a) and Psychrobacter cryohalolentis K5 from cryopeg (Bakermans et al. 2006) grown at or below 4°C (Bakermans et al. 2007), showed the downregulation of enzymes involved in major metabolic processes (glycolysis, fermentation, sugar metabolism, etc.), and the upregulation of cellular proteins and enzymes involved in the transport of electrons, DNA/RNA repair and the salvaging of degraded DNA for nucleotide synthesis. Psychrobacter cryohalolentis K5 showed an increase of cellular ATP and ADP concentrations with decreasing temperatures between 22 and −80°C (Amato and Christner 2009). These results suggest that some bacteria maintain limited metabolic activity in permafrost. The data support views that microorganisms experience structural and metabolic changes during the freezing process and that these changes may lead to different survival strategies—dormant or inactive state and anabiosis or greatly reduced metabolism.

Growth and metabolic activity at subzero temperatures

The ambient temperature measured in permafrost sediments of the Kolyma Lowland ranges from −5.6 to −12.3°C (Romanovsky et al. 2010; Fedorov-Davydov et al. 2018) and lies within the lower limit for life on Earth (Clarke et al. 2013). Permafrost suspensions spread on solid nutrient agar medium yielded bacterial and fungal growth at −10°C within 4 weeks (Gilichinsky et al. 1992). The ability of some bacteria isolated from permafrost and permafrost-affected soils to grow at temperatures well below 0°C was shown in the laboratory. For example, Planococcus halocryophilus showed growth at −15°C in trypticase soy broth supplemented with up to 18% NaCl and 7% glycerol (Mykytczuk et al. 2013); Psychrobacter arcticus grew at −10°C in marine broth with 3% sea salts, 0.5% tryptone and 0.1% yeast extract (Bergholz, Bakermans and Tiedje 2009); and Exiguobacterium sibiricum grew at −6°C in trypticase soy broth supplemented with 0.7% yeast extract (Vishnivetskaya et al. 2007). Growth of eukaryotic microorganisms (filamentous fungi Geomyces spp. and yeasts Leucosporidium spp.) was recorded at −8°C, whereas respiratory and biosynthetic (¹⁴CO₂ uptake) activity was measured down to −39°C on solid media for 3 weeks (Panikov et al. 2006). In fact, these data could not serve as irrefutable evidence of in situ growth due to more stringent constraints within permafrost when low water content, low water activity, nutrient limitation and unavailability of frozen substrates happen simultaneously (Gilichinsky, Soina and Petrova 1993; Gilichinsky and Rivkina 2011).

Experiments with samples collected from cryopegs—lenses of unfrozen brine surrounded by marine permafrost formed 100–120 ka—showed that their native bacterial population was able to perform [¹⁴C]glucose intake ∼100 times higher at −15°C than bacteria from non-saline sediments added in to autoclaved cryopeg brine (Gilichinsky et al. 2003) providing evidence for in situ metabolic activity. A few other studies implied that indigenous permafrost bacteria could be metabolically active within frozen sediments (Vishnivetskaya et al. 2018; Liang et al. 2019). Potential metabolic activity under in situ permafrost conditions could be assumed from a positive correlation between methane concentration and presence of methanogenic archaea (Rivkina et al. 2016), phospholipids (Wagner et al. 2007) and total organic carbon (Koch, Knoblauch and Wagner 2009). Methane concentration did not correlate with permafrost age but rather with paleoclimatic conditions during permafrost aggradation (Rivkina et al. 2016; Holm et al. 2020). The native microbial community (with total cell number of 1.1 × 10⁸ cells g⁻¹ and 2 × 10⁵ CFU g⁻¹) from unamended Olyor permafrost showed incorporation of ¹⁴C-labeled acetate into lipids at −20°C during a period of 550 days, with an estimated doubling time of ∼160 days (Rivkina et al. 2000). Production of CO₂ at −10°C with mean rates of 0.142–0.794 µg of CO₂-C g⁻¹ day⁻¹ was measured in samples <600 ka but no CO₂ was detected in the 740-ka sample (Johnson et al. 2007). Perennially frozen loamy peat soils with average annual ambient temperature of −10 ± 5°C and estimated age of 2.9 ka showed methane production at −16.5°C from both H¹⁴CO₃⁻ (0.0013 µmol CH₄ kg⁻¹ day⁻¹) and ¹⁴CH₃CO₂⁻ (0.0005 µmol CH₄ kg⁻¹ day⁻¹) (Rivkina et al. 2002). Extensive review on the temperature dependence of metabolic rates has not elucidated any evidence of the minimum temperature for metabolism and estimated that metabolic rate at −40°C in ice corresponds to ∼10 turnovers of cellular carbon per billion years (Price and Sowers 2004).

Evidence for cell activity at molecular level

Measurable metabolic activity and the presence of intact microbial cells with visually undamaged cell structure demonstrated by microscopy approaches (Soina and Vorobyova 2004) indicate that some permafrost inhabitants have preserved cell integrity and may be alive. Similar conclusions could be drawn from the results of measuring amino acid racemization in permafrost microorganisms. The latter method evaluates the conversion of amino acids, which are the building blocks of proteins, from l to d forms and calculates ratio of d/l isomers leading to the estimation of how long ago the specimen died (Rutter and Blackwell 1995; Goodfriend et al. 2000). Analyses of samples from the Kolyma Lowland permafrost with temperature around −11°C and age up to 25 ka using the aspartic acid (Asp) racemization assay suggested active recycling of d-aspartic acid and showed that some permafrost organisms are capable of repairing molecular damage (Brinton et al. 2002). Though Asp in the older bulk permafrost sediments (0.8–1.1 Ma) underwent severe racemization relative to the youngest layer (∼0.01 Ma), the much lower d/l Asp ratio in the separated intact cells (0.05–0.14) in comparison to bulk sediment (0.31–0.41) indicated that indigenous microbial communities are viable and likely metabolically active in ancient permafrost up to 1.1 Ma (Liang et al. 2019).

The total community genomic DNA repeatedly extracted from a number of permafrost samples displayed a persistence of genetic material over geological timespans (Willerslev, Hansen and Poinar 2004a; Vishnivetskaya et al. 2006; Rivkina et al. 2016; Liang et al. 2019). High reproducibility of amplifications for short 16S rRNA gene fragments was obtained from samples up to 400–600 ka (Willerslev, Hansen and Poinar 2004a). However, some studies mentioned incoherent results for PCR amplifications and metagenomic sequencing for older permafrost samples and suggested that it may be caused by the formation of DNA interstrand crosslinks, chemical modifications, genetic damage, mutations or low abundance of DNA molecules (Willerslev, Hansen and Poinar 2004a; Vishnivetskaya et al. 2006; Yergeau et al. 2010). However, based on PCR results after genetic damage induced by treating aliquots of the permafrost DNA extracts with uracil-N-glycosylase before amplification of bacterial 4-kb DNA fragments, it was determined that ancient permafrost bacteria show evidence of DNA repair (Johnson et al. 2007). Nevertheless, mutation rate estimated for lineage of 16S rRNA genes derived from the seven different Lysobacter-like phylotypes originated from permafrost samples with age of 100 ka to 1.5 Ma was at ∼0.17% (or 2.45 fixed mutations) per 10⁶ years (Vishnivetskaya et al. 2006).

Age of permafrost microorganisms

The studies reviewed above demonstrate that diverse microorganisms found in permafrost show evidence of DNA repair, and are able to perform metabolic reactions at permafrost conditions simulated in a laboratory. In this section, we aim to integrate the results from those studies with the knowledge of permafrost sediments from the Kolyma Lowland in northeastern Siberia. The latter is the only region where thermal conditions during the past 3 million years led to formation of both syngenetic and epigenetic permafrost (Sher 1971; Gilichinsky and Rivkina 2011) and where permafrost of different ages can be found in the same location. The important features of frozen sediments are preclusion of any diffusion, infiltration and microbial penetration from the surface-active layer as well as impossibility of both lateral and vertical migration through permafrost (Ostroumov et al. 1998) that could lead to cryogenic preservation of the indigenous microorganisms (Khlebnikova et al. 1990; Vorobyova et al. 1997). Indeed, the permafrost microbial community has been described as ‘a community of survivors’ (Friedmann 1994) and viewed as the result of a continuous selection of organisms adapted to permafrost conditions and able to cope with low-temperature stress over the geologically long time periods (Gilichinsky 2002).

The interface between ice and mineral grain surfaces at temperatures of −10°C and lower contains a thin layer of unfrozen water ∼5 × 10⁻¹⁰ m or one to two molecular diameters thick (Anderson 1967; Gilichinsky, Soina and Petrova 1993; Wettlaufer 1999). Even though the microorganisms from frozen habitats are small (∼0.8 ± 0.3 × 10⁻⁶ m³) as determined by in situ transmission electron microscopy in Siberian permafrost (Soina et al. 1995) or ultrasmall (<0.1 × 10⁻⁶ m³) as detected by scanning electron microscopy in both Siberian permafrost (Gilichinsky, Wagener and Vishnivetskaya 1995) and silty ice core samples collected from 3043-m-deep Greenland glacier (Miteva and Brenchley 2005), they are still 10³-fold larger than water films. As such, they cannot move freely and there is not enough space for a progeny in case of cell division.

An important factor for long-term survival of microorganisms in dispersed systems at temperatures well below 0°C is the transfer of unfrozen liquid water and ions. Ion diffusion coefficients (D_i) detected in sediments with low ice content (D_i = 10⁻¹⁰ m² s⁻¹) and sediments with high ice content (D_i = 10⁻¹⁶ m² s⁻¹) (Ostroumov and Siegert 1996) supported the hypothesis that mass transfer in permafrost could result in exchange of liquid water and ions between living cells and the habitat permitting metabolism. Consequently, metabolic processes could occur in deep subsurface and permafrost, but evidently temperature-dependent metabolic rates of microbial communities in the environment are lower than those measured in the laboratory. Metabolic rates measured using the incorporation of tritiated thymidine into DNA in communities from unfrozen subsurface were 10³-fold lower than growth rates in surface sediments (Thorn and Ventullo 1988) and were consistent with metabolic rates sufficient for maintenance of functions but too low for growth (Price and Sowers 2004). Metabolic rate per cell estimated for microorganisms imprisoned in deep subsurface permafrost was 10⁶-fold lower than the metabolic growth rate measured in the laboratory and was probably sufficient for survival only when microorganisms can repair macromolecular damage at a greatly reduced rate but are largely dormant (Price and Sowers 2004).

If our assumptions are correct, then hypothetical communities that grow in a laboratory with availability of substrate (10 × 10⁻⁶ mol sodium acetate ml⁻¹) and at temperatures of −10 and −20°C with doubling time of 20 and ∼160 days (Rivkina et al. 2000) will slow down their growth in low-nutrient deep subsurface permafrost by 10⁶-fold and would double every 55 000 and 450 000 years, respectively. The ancient low-molecular-weight organic acids measured in permafrost samples collected from the CRREL Tunnel, Alaska, with age ∼34 ka were at 0.3 × 10⁻⁶ and 0.4 × 10⁻⁶ mol ml⁻¹ for acetate and butyrate, respectively (Drake et al. 2015), and they were >25-fold lower than the amount of these acids used for growth media in a laboratory. The lower nutrient concentration in the permafrost environment would further decelerate in situ growth.

What could we say about how old microorganisms isolated from permafrost are? First, microorganisms that resided in sediments or soils at the moment of freezing experienced a number of stresses associated with a rapid temperature downshift followed by chemical changes (supercooling, changes in pH and concentrations of solutes, denaturation of the cell proteins, etc.) and mechanical destructions (freezing of water and formation of ice crystals, disruption of cell surface, etc.) leading to the death of some microorganisms (Haines and Barcroft 1938) and triggering a physiological response in others to ensure their survival at changed conditions (Barria, Malecki and Arraiano 2013). Then, taking into account the permafrost properties, lack of microbial movement and cell division, insufficient substrate availability for microbial growth and negligible metabolic rate, resulting in the dormant state of a majority of permafrost inhabitants, we assume that microorganisms embedded in the ancient permafrost sediments are as old as permafrost itself. This argument is supported by the fact that microorganisms do not multiply in the permafrost. In the case of the syngenetic permafrost, accumulation of sediments and resident microorganisms happened simultaneously with freezing; therefore, the age of microorganisms is considered to be equal to the age of permafrost and to the age of sediments. However, in the case of epigenetic permafrost, sediments and native microbial inhabitants aggregated and evolved first well before freezing; thus, the age of microorganisms embedded into the epigenetic permafrost is assumed to be equivalent to the age of permafrost, though the age of permafrost, which is less than the age of sediments, reflects the time of last freezing and mirrors the last cooling period the sediments encountered.

CONCLUSIONS

The phenomenon of life preservation in permafrost, including microbial communities, extracellular metabolites, enzymes, nucleic acids and biogenic gases, has been shown by numerous researchers. Recent years have seen an increase in the number of studies about different aspects of permafrost microbiology. However, questions regarding the age of microbial cells in permafrost are rarely highlighted in the publications. Indeed, an accurate understanding of the age of microbial cells and how it relates to the age of the permafrost is critical for proper interpretation of microbiological and omics data in permafrost studies. Lastly, we would like to recap the relationship between ‘sediment age’ and ‘permafrost age’ for epigenetically frozen deposits from two locations on the Kolyma Lowland. Analysis and interpolation of the available geological, palynological and climatological data lead us to conclusions that (i) unfrozen deposits of the Olyor and Tumus Yar suites in the Alazeya River basin were formed 0.8–1.6 and 1.8–2.2 Ma, respectively; (ii) those sediments became a permafrost during the late Pliocene and the early Pleistocene at the cold continental climate; (iii) the Olyor and Tumus Yar permafrost have not ever degraded since the time they were last frozen; and (iv) existence of the late Pleistocene Yedoma suite (Ice Complex) sediments with syngenetic large ice wedges proves that the underlying sediments did not thaw. Therefore, the age of the Olyor and upper horizon of Tumus Yar permafrost was estimated to be ∼0.8–1.0 and 1.0–1.1 ± 0.3 Ma, respectively. Thus, the age of microorganisms from these permafrost deposits corresponds to estimated permafrost age. The remnants of relict Pliocene permafrost theoretically could be found at deeper levels, as all our samples were collected at no deeper than 80 m from the surface. The second location corresponds to the middle Pleistocene marine sediments of Kon'kovaya suite, which were discovered along the narrow strip of the East Siberian Sea coast. These sediments were epigenetically frozen ∼100–120 ka. Consequently, the age of the microbial cells in those sediments is comparable to the age of permafrost.

ACKNOWLEDGMENTS

We are grateful to Caitlin Billing and Karen Lloyd for the English proofreading.

FUNDING

This work was supported by the Russian Government Assignment #AAAA-A18-118013190181-6 to ER, TV and AA, the Russian Foundation for Basic Research (19-29-05003-mk) to ER, the National Science Foundation (DEB-1442262) to TV and partially supported by the U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research, Genomic Science Program under award number DE-SC0020369 to TV.