A Constant Flux of Diverse Thermophilic Bacteria into the Cold Arctic Seabed

See allHide authors and affiliations

Science  18 Sep 2009:
Vol. 325, Issue 5947, pp. 1541-1544
DOI: 10.1126/science.1174012


Microorganisms have been repeatedly discovered in environments that do not support their metabolic activity. Identifying and quantifying these misplaced organisms can reveal dispersal mechanisms that shape natural microbial diversity. Using endospore germination experiments, we estimated a stable supply of thermophilic bacteria into permanently cold Arctic marine sediment at a rate exceeding 108 spores per square meter per year. These metabolically and phylogenetically diverse Firmicutes show no detectable activity at cold in situ temperatures but rapidly mineralize organic matter by hydrolysis, fermentation, and sulfate reduction upon induction at 50°C. The closest relatives to these bacteria come from warm subsurface petroleum reservoir and ocean crust ecosystems, suggesting that seabed fluid flow from these environments is delivering thermophiles to the cold ocean. These transport pathways may broadly influence microbial community composition in the marine environment.

Microbial diversity surveys have revealed that species richness is determined by many low-abundance taxa—the so-called rare biosphere (13). In the ocean, certain members of this relatively unexplored biosphere comprise a dormant microbial “seed bank” that can be transported passively over great distances (1). Quantitatively tracking the migration of indicator taxa can highlight key factors that influence patterns of biogeography and may help evaluate the extent to which microorganisms exhibit a cosmopolitan distribution (4). Endosporulation allows certain bacteria to persist as dormant cells in hostile environments, explaining discoveries of viable thermophilic Firmicutes in inhospitably cold habitats (510). Quantitative studies of this phenomenon are scarce, and the origin and distribution of thermophiles in cold environments remain enigmatic (611). Thermophilic sporulating taxa such as certain Desulfotomaculum spp. may constitute only 0.001% of marine microbial populations (8, 12). Like the rare taxa, spores are less prone to viral lysis or predation, and are not detected by traditional diversity surveys (1, 2). A spore-forming Desulfotomaculum strain that can only grow between 26° and 47°C was recently isolated from permanently cold Svalbard fjord sediment in the Arctic (10). The present study assessed thermophile diversity, abundance, and distribution in Svalbard sediments to reveal insights into mechanisms governing biogeography in the marine environment.

Pristine sediment was sampled from Smeerenburgfjorden (80°N; fig. S1) and incubated over an experimental temperature range (13), which revealed two distinct sulfate-reduction regimes (Fig. 1A). The first has a temperature optimum (Topt) of 22°C and a maximum (Tmax) of 32°C, consistent with earlier studies in Svalbard sediments (14, 15) where the in situ temperature is –2° to +4°C year-round. This temperature-activity profile is characteristic of psychrophilic sulfate-reducing bacteria (SRB) that are well adapted to cold conditions (14, 16), which explains their activity and abundance (up to 16%) in this environment (17). A second Topt of 56°C (Fig. 1) indicates a previously unrecognized thermophilic community with a Tmax above 60°C. Pasteurization at 80°C killed the psychrophiles but did not adversely affect the thermophiles (Fig. 1B), indicating that the latter exist as endospores in situ and only germinate after heating. This was supported by successful polymerase chain reaction targeting the spore-forming SRB genus Desulfotomaculum only if the sediment was incubated at 50°C before DNA extraction.

Fig. 1

Temperature-gradient incubations. SRR were measured in Smeerenburgfjorden sediment samples incubated between 0° and 76°C. Distinct SRB populations were active between 0° and 32°C, and between 41° and 62°C. Replicates were either untreated (A) or pasteurized for 1 hour at 80°C (B) before incubation. Open and closed symbols correspond to experiments below and above 35°C, respectively. Incubations in the thermophilic range resulted in much higher SRR (fig. S3); therefore, data are plotted as the percentage of the maximum SRR in (A) at either Topt. Pasteurization killed the psychrophilic community and resulted in slightly higher activity by the thermophilic community [115% at the Topt (B)].

Psychrophilic and thermophilic communities were investigated further by incubating homogenized surface sediment at 50°C (13). Sulfate reduction rates (SRR) quickly dropped to below the detection limit (Fig. 2A) due to stimulation and subsequent death of vegetative psychrophiles as the sediment warmed up (compare to Fig. 1A). Thermophilic SRR increased exponentially between 20 and 96 hours. The transition from SRR below detection (up to 16 hours) to the onset of the exponential phase (at 20 hours) suggests a lag phase during which conditions became favorable for germination of thermophilic SRB spores. Endospore germination requires appropriate nutrients and substrates (18) such as volatile fatty acids (VFA) that are produced by microbial fermentation. VFA are typical electron donors for SRB and were generated rapidly during incubation at 50°C (Fig. 2B), creating conditions suitable for thermophilic sulfate reduction.

Fig. 2

Sediment incubation at 50°C. SRR (A) and concentrations of VFA (B) in Smeerenburgfjorden sediment (0- to 3-cm depth) incubated at 50°C. SRR represent 4- to 8-hour incubations with 35SO42− radiotracer. Open symbols correspond to the psychrophilic SRB community that was killed as the temperature increased to beyond their Tmax (compare to Fig. 1A). SRR were below detection during the 10 to 16 hours time interval (n = 2; vertical bar: standard error). Sulfate reduction by 20 to 24 hours indicates germination of Desulfotomaculum spores, which subsequently catalyzed an exponential increase in SRR. These data constrain the earliest onset of the exponential phase to 16 hours, as indicated by the dashed extension of the exponential trendline. The exponential phase corresponded to the consumption of butyrate and propionate, which increased steeply, together with acetate (B), before the onset of thermophilic sulfate reduction.

Several lines of evidence demonstrate that the VFA production was biologically mediated and not due to thermal alteration of the sediment organic matter. VFA did not accumulate in autoclaved sediment incubated in parallel at 50°C (concentrations never exceeded 0.3 mM). Amendment with the fluorescently labeled polysaccharides pullulan or arabinogalactan revealed extracellular enzymatic hydrolysis in sediment incubated at 50°C, whereas no hydrolysis was detected in sediment-free controls. Clone library analyses of 16S ribosomal RNA (rRNA) genes revealed enrichment at 50°C of different bacterial groups related to Desulfotomaculum, Caminicella, and Caloranaerobactor-Clostridiisalibacter-Thermohalobacter lineages within the Firmicutes phylum (Fig. 3) [see supporting online material (SOM) and table S3]. The latter lineages are represented by thermophilic anaerobes that convert carbohydrates and proteins to VFA, such as Caminicella sporogenes (19). Hydrolysis and fermentation at 50°C were therefore likely due to thermophilic spores that germinated at high temperature in the presence of complex natural organic substrates in the sediment.

Fig. 3

Phylogenetic analysis of putative spore-forming thermophilic bacteria in Svalbard sediment. Clone libraries constructed before and after incubation at 50°C differed significantly due to the enrichment of thermophilic Firmicutes (see SOM, tables S2 and S3). Dominant phylotypes (i.e., >5% of clones) from the 50°C sediment library are indicated in boldface with relative abundances shown in parentheses. These phylotypes were not detected before 50°C incubation. The consensus tree of 16S rRNA gene sequences of selected Firmicutes includes next relatives for each phylotype (top 10) plus additional type strains (for Desulfotomaculum). Habitats of origin for closest relatives are indicated. Filled and open circles indicate lineages with >90% and 80 to 90% parsimony bootstrap support (100 resamplings), respectively. The scale bar indicates 5% sequence divergence as estimated from maximum-likelihood analysis.

VFA production at 50°C stimulated growth of a thermophilic Desulfotomaculum population (Fig. 3), resulting in an exponential increase in SRR (Fig. 2A). The Desulfotomaculum cell density at the onset of sulfate reduction can be estimated by dividing the bulk rate (1.3 nmol cm−3 hour−1 at 16 hours based on the exponential function in Fig. 2A) by a mean cell-specific sulfate reduction rate of 2.0 fmol cell−1 hour−1 [based on 33 pure cultures during exponential growth (20), and representative of thermophilic Desulfotomaculum spp.; see SOM]. This calculation estimates a population of 6.3 × 105 SRB cm−3 at 16 hours, which corresponds to the in situ spore density given that SRB growth did not occur before 16 hours (Fig. 2A).

Exponential increases in SRR were also measured across a series of intact sediment cores (0- to 23-cm depth) incubated at 50°C (13). These data indicate similar numbers of thermophilic Desulfotomaculum spores, on the order of 105 cm−3, at all depths (fig. S2). Thermophiles thus constitute a stable component of the rare biosphere (~0.01%) in this Arctic marine habitat (17). Abundant taxa in Smeerenburgfjorden sediment are psychrophilic Cytophaga, Flavobacteria, Gammaproteobacteria, and Deltaproteobacteria (17) that catalyze hydrolysis, fermentation, and sulfate reduction during degradation of organic matter (15, 21). We induced the same processes at 50°C, which stimulated SRR about 30-fold higher than in sediment incubated at 4°C (fig. S3). It is interesting that the thermophilic bacterial community has metabolic potential that mirrors the dominant metabolic processes occurring in situ. However, these thermophilic spores make no contribution to local biogeochemical cycling (Fig. 1). Our experience with thermophiles cultivated from Smeerenburgfjorden that have no detectable metabolic activity below 25°C supports this. Thermophilic spores must therefore be supplied externally and can be considered akin to particles accumulating in the sediment.

To estimate the particle flux of thermophilic spores into the sediment, we determined an age model by measuring the 210Pb activity as a function of depth (see SOM). This revealed a sediment accumulation rate of 0.19 cm year−1 in Smeerenburgfjorden (figs. S2 and S4). Given a stable abundance of about 105 spores cm−3 with depth (fig. S2), the constant supply of thermophiles to this site is 2 × 108 m−2 year−1. Other fjords along the west coast of Svalbard have comparable thermophilic SRB densities in surface sediments (see SOM; fig. S1). The fjords and the adjacent coastal shelf in this area cover about 1000 km2 (fig. S1), hence an extrapolation of our estimate corresponds to an annual supply of ~1017 thermophiles in this Arctic region. These large numbers highlight not merely the occurrence of thermophiles in the Arctic but rather their unexpected quantity and the consistency of their flux. The warm, anaerobic environment where these bacteria originate must have sufficient geographic distribution, magnitude, and source strength to support the population sizes indicated by our data.

A limited number of marine habitats meet these criteria. Deeply buried sediments hundreds of meters below the sea floor represent warm habitats for anaerobic microbes (22). Advective flow of hydrocarbons or other fluids from these sediments can penetrate the sea floor (e.g., at mud volcanoes) and transport microscopic particles or cells up into the cold ocean (23). Seafloor pockmarks and active cold seeps are known around west Svalbard (24) (fig. S1) and are often associated with deep gas or oil-bearing deposits (23). 16S rRNA gene sequence comparisons revealed that taxa enriched in our 50°C experiments (Fig. 3) are most closely related to bacteria from subsurface petroleum reservoirs or oil production facilities (94 to 96% similarity; table S3). Another source of thermophiles could be nearby mid-ocean ridge spreading centers (fig. S1). Large volumes of fluids circulating through ocean crust (25) could transport cells away from warm anoxic niches in this seafloor habitat (26) and suspend them in abyssal currents. The closest relatives to the Arctic thermophiles also include an anaerobic thermophile isolated from deep, hot crustal fluid (94% similarity; table S3) (27).

Petroleum-bearing sediments and fractured ocean crust both host anaerobic heterotrophic microbial communities (26, 28). Areas of discharge connecting these habitats to the water column are widespread, and both processes expel large volumes of fluid into the oceans (23, 25). A combination of different point sources could explain the diversity and distribution of thermophilic taxa in Arctic sediments (Fig. 3 and fig. S1). Although our observations suggest that seabed fluid flow governs the biogeography of thermophilic spore formers, these passive dispersal mechanisms are unlikely to act only on these particular bacteria. Permeable conduits through sediments and ocean crust pass through several microbial niches with changing local temperature and geochemistry (22, 25, 26). Widespread seeding of the oceans by geofluids from deep biosphere habitats may therefore contribute broadly to the high microbial diversity observed in the marine environment.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S4

Tables S1 to S3

  • Present address: Department of Geology and Geochemistry, Stockholm University, Svante Arrhenius Väg 8C, 106 91 Stockholm, Sweden.

  • Present address: Institute of Biology, University of Southern Denmark, Campusvej 55, 5320 Odense M Denmark.

References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. We thank S. Henningsen, J. Mortensen, K. Barker, and A. Steen for assistance aboard R/V Farm and the Alfred Wegener Institute for kindly providing laboratory space in Ny Ålesund, Svalbard (project KOP56; RIS ID: 3298). We are grateful to K. Imhoff, M. Meyer, and A. Schipper for technical assistance and to A. Judd, M. Holtappels, N. Finke, and W. Bach for valuable discussions. This work was supported by the Natural Sciences and Engineering Research Council of Canada (C.H.), the Austrian Science Fund (P20185-B17; A.L.), the U.S. National Science Foundation (OCE-0323975 and OCE-0848703; C.A.), and the Max Planck Society. Nucleotide sequences have been deposited in the European Molecular Biology Laboratory database under accession numbers FN396615 to FN396795.
View Abstract

Navigate This Article