Most of the data discussed in the present review concern Desulfotomaculum species. Nevertheless, sulfate-reducing Gram-positive Bacteria also include Desulfosporosinus spp., Desulfovirgula sp., Desulfispora sp. and the candidate species “Desulforudis audaxviator.” The few available data dealing with the presence of these three other genera in the deep subsurface will be discussed later in this paper when necessary.

Desulfotomaculum spp. are anaerobic bacteria using sulfate as terminal electron acceptor, which is reduced to sulfide. They are members of the phylum Firmicutes, class Clostridia, order Clostridiales and family Peptococcaceae (Kuever and Rainey, 2009). When grown in pure cultures, Desulfotomaculum cells are straight or curved rods of dimensions 0.3–2.5 × 2.5–15 microns with rounded or pointed ends. Moreover, Desulfotomaculum spp. are spore-forming bacteria with central to terminal round or oval spores, often causing swelling of the cells. The genus is composed of 30 validly described species and one subspecies to date. Among these, 17 are thermophilic or moderately thermophilic, 3 are halophilic and one is alkaliphilic.

Beside sulfate, some species described to date can use other sulfur-containing inorganic compounds including thiosulfate, sulfite, and elemental sulfur as terminal electron acceptors (Kuever and Rainey, 2009). While the disproportionation of sulfur compounds (e.g., thiosulfate, elemental sulfur) has been demonstrated to frequently occur among the members of the Deltaproteobacteria class (Lovley and Phillips, 1994; Finster, 2008), there are only two reports on the ability of Desulfotomaculum spp to perform thiosulfate disproportionation. They include D. thermobenzoicum (Jackson and Mcinerney, 2000) and D. nigificans (Nazina et al., 2005). In addition, Desulfotomaculum reducens was shown to use metals [e.g., Mn(IV), Fe(III), U(VI) or Cr(VI)] as terminal electron acceptors (Tebo and Obraztsova, 1998). Numerous species including D. geothermicum, D. salinum and D. kuznetsovii (Daumas et al., 1988; Nazina et al., 1989, 2005), oxidize H₂, but also organic acids and long chain fatty acids. Some Desulfotomaculum spp. may grow autotrophically on hydrogen (Daumas et al., 1988; Nazina et al., 1989; Tasaki et al., 1991) or may oxidize it by reducing CO₂ into acetate. This metabolic process known as homoacetogenesis was only demonstrated for D. gibsoniae (Kuever et al., 1999), however, it has only been investigated in a few species. Several species use carbohydrates as electron donors. They include D. putei, D. carboxydivorans, and D. geothermicum (Daumas et al., 1988; Liu et al., 1997; Parshina et al., 2005). Substrates are either completely oxidized to CO₂ or incompletely oxidized to acetate. In the absence of sulfate, some species can also grow by fermentation of glucose, fructose or pyruvate. It was also recently shown that Desulfotomaculum sp. Ox39 utilized aromatic hydrocarbons (e.g., toluene, m-xylene, o-xylene) as carbon and energy sources (Morasch et al., 2004). This ability to oxidize monoaromatic hydrocarbons has also been reported for other undescribed members of this genus (Berlendis et al., 2010).

Different definitions of the extent of the deep subsurface (Fliermans and Balkwill, 1989; Sinclair and Ghiorse, 1989; Pedersen, 1993) have proposed an upper limit of between 10 and 100 m below the ground or seabed. From our point of view, deep environments should rather be defined as subsurface settings isolated from the current and direct influence of surface environments. Water column or surface sediments thus do not fit our definition, however, they are also discussed below in certain instances where they can be considered as gateway environments linked to truly deep environments, i.e., the deep geological formations, through sedimentation.

The deep environments show great differences regarding nutrient availability. Some of them are rich in organic matter which can be used as electron donors and/or carbon sources by the microorganisms while others are oligotrophic environments where carbon dioxide and hydrogen are sometimes the only carbon and energy sources available. These different environmental conditions are discussed separately below.

One important characteristic of subsurface environments is the availability of hydrogen to be used by autotrophic microorganisms. It can originate from the low but significant radioactivity of rocks by radiolysis of water, or from anaerobic mineral reactions (Stevens and Mckinley, 1995; Schrenk et al., 2013). There is also a low but constant flux of hydrogen from the Earth mantle to surface. Its presence is favorable to the development of autotrophic and homoacetogenic anaerobic microorganisms in the subsurface. Nevertheless, this is particularly important in oligotrophic environments. Where organic matter is abundant, like in sediments, even chemolithotrophs rely on energy that is released by the chemoorganotrophic breakdown of photosynthetically produced organic matter—nonetheless they are present and active.

The deep waters of lakes are not, strictly speaking, deep subsurface environments as defined in the introduction. However, some published work shows how bacteria can use this gateway to the subsurface thanks to their physiological adaptation to changing environments, and therefore these environments deserve to be mentioned in this review.

The surface and deep waters of the meromictic Lake Gek-Gel in Azerbaijan are mixed less than once a year, resulting in the formation of a deep anoxic zone (Karnachuk et al., 2006). Desulfotomaculum spp. were detected in the water column of this lake at depth below 31 m where there is no more dissolved oxygen. In addition, the proportion of Desulfotomaculum spp. in the water column determined by FISH increased with depth (0% of total microorganisms at 10 m 7% at 16 m, and 31% at 70 m), while species of the genera Desulfovibrio and Desulfomicrobium (Deltaproteobacteria) appear to follow a reverse trend. These results thus suggested that Desulfotomaculum spp. play a major role in the anoxic zone of the lake. Moreover, it is unlikely that the chemical treatment for FISH applied to the samples was efficient enough to allow the detection of Desulfotomaculum spores. It can therefore be assumed that the overall number of Desulfotomaculum could potentially be even higher than reported.

Strains related to the genus Desulfotomaculum have also been isolated from sediments of the oligotrophic Lake Stechlin located in the north of Berlin, Germany (Sass et al., 1998). This is a dimictic lake where the deep anoxic zone and the oxic surface layer mix twice a year, once in winter and the other in summer. In this study, 27 bacterial strains were isolated from sediments collected near the shore and approximately 30 m deep in the lake. Interestingly, all isolated Gram-positive bacteria were related to the spore-forming genera Desulfotomaculum and Sporomusa (Möller et al., 1984). They represented nearly 40% of all bacteria isolated at depth in this lake. Another study on these sediment samples again suggested that Desulfotomaculum isolates represented the dominant microbial community within the deepest layers of sediments (Sass et al., 1997). These results are consistent with data reported by Stahl et al. (2002) suggesting that Gram-positive sulfate-reducing bacteria (SRB) are the main sulfidogenic microorganisms in deep environments.

These two studies illustrated that not only sediments, but also lakes waters can be habitats for anaerobic microorganisms if they are deep enough to have a permanent or semi-permanent anoxic deep zone. Sass et al. (1997) stated that the presence of Desulfotomaculum spp. in fresh waters is not surprising given the low nutrient intake in these niches. Indeed, as mentioned above, Desulfotomaculum cells can sporulate and thus survive long deprivation periods. However, the spore-forming ability may not be the only reason for the resistance of these microorganisms in these environments since only 0.5% of Desulfotomaculum cells sampled in Lake Stechlin were shown to be present as spores and survived pasteurization (Sass et al., 1997). In contrast, Bak and Pfennig (1991) showed earlier that 50% of the Desulfotomaculum found in Lake Constance (on the border between Switzerland and Germany) were in the sporulated form. Nevertheless, it seems that such differences are difficult to explain until being able to link sporulation with the detailed physico-chemical conditions which could have influenced it in the studied lakes.

In this respect, the distribution of Desulfotomaculum in deep lakes can be explained by the adequacy of their physiological characteristics to particular and changing physicochemical conditions. Desulfotomaculum spp. are generally predominant in anoxic environments where sulfate concentrations are low (Ingvorsen et al., 1981; Widdel, 1988; Leloup et al., 2006). The selection of Desulfotomaculum species entering the sedimentation process should possibly also results from their metabolic capabilities. Castro et al. (2002) reported that complete oxidizers represented 95% of the Desulfotomaculum spp. in several human-impacted environments, whereas in pristine environments, with significantly lower concentrations in total phosphorus and carbon, all Desulfotomaculum-related sequences were related to incomplete oxidizing species. From these results authors hypothesized that the versatility of complete oxidizers allows them to proliferate in substrate-rich environments and those incomplete oxidizers were more adapted to use specific nutrients in pristine environments. These conditions correspond to those found at the bottom of deep freshwater lakes, but specific studies dealing with substrate oxidation in these environments are lacking. Anyway, we can speculate that during water mixing, although nutrients are renewed in depth, growth of Desulfotomaculum members should be limited due to the presence of oxygen. Once the favorable redox conditions are back in deep water and shallow sediments, Gram-positive bacterial spores can germinate and growth of selected Desulfotomaculum strains starts again.

Microbiological studies of deep subsurface samples are scarce. Sampling deep geological layers is most often closely linked to industrial activities, although specific national or international scientific programs dedicated to the study of the deep subsurface do exist. The reason is obviously that the cost of drilling at great depths can generally be assumed only if there is an industrial interest. Most of these studies are thus related to the industrial exploitation of the subsurface, including mining, oil production, aquifers, or different types of storages in the subsurface (e.g., natural gas, nuclear wastes). Although these environments have many similar physical properties such as temperature and pressure, they can be very different regarding the availability of nutrients.

In the Atlantic Ocean, microorganisms related to the genus Desulfotomaculum were generally not found at deep-sea hydrothermal vents (Ollivier et al., 2007). Nevertheless, the Lost City hydrothermal field near the mid-Atlantic ridge at a depth of 900 m represents a specific unusual case, where Brazelton et al. (2006) demonstrated that the cloned 16S rRNA genes related to D. halophilum (88% identity) and D. alkaliphilum (92% identity) represented the only sulfate-reducing Bacteria identified in this environment. These results are in agreement with those of Gerasimchuk et al. (2010), as the cloned 16S rRNA genes distantly related to D. alkaliphilum, D. halophilum and D. thermocisternum were found in sediments collected from the same site (with sequence identities between 88 and 91%). Furthermore, the potential ability to reduce sulfate was revealed by the detection of genes encoding the dissimilatory sulfate reductase (dsrAB) enzyme. These genes sequences were related with those of incomplete oxidizers Desulfotomaculum spp. as described by Klein et al. (2001). These informations are consistent with the limitation of carbon observed on the Lost City chimneys, where oxidized organic compounds originate from synthrophic anaerobic oxidation of methane (Bradley et al., 2009).

Mineralization of sedimentary organic matter occurs at the seabed where SRB play a major role in the carbon and sulfur cycles (Jørgensen, 1977). Many studies have shown the presence of Gram-positive SRB in deep marine environments and in shallow sediments under deep seawater columns (Reed et al., 2002; Zhuang et al., 2003; Teske, 2006). Such microorganisms seem to predominate among all sulfate reducers, as shown in the sediments of the 1900 m deep West Pacific Warm pool, where they represented 54% of the sulfate-reducing population at 1–12 cm depth (Wang et al., 2008). However, FISH counts showed that SRB only constitute 0.3–2% of active microorganisms in this ecosystem.

The presence of spores in surface sediments (0–25 cm deep) from the Arctic Ocean was revealed by endospore germination experiments (Hubert et al., 2009). Thermophilic Firmicutes including Desulfotomaculum spp. were identified in 16S rRNA genes clone libraries. More recently, Hubert et al. (2010) showed that spores of thermophilic microorganisms scattered on the Arctic seabed mineralize complex organic matter at 50°C in vitro. This activity was catalyzed via extracellular enzymatic hydrolysis, fermentation and sulfate reduction. Four distinct Desulfotomaculum phylotypes were identified. The most prominent was related to D. halophilum (91% identity), whereas three other lineages were related to D. geothermicum, D. reducens, and D. thermosapovorans (95–96% identity).

Very few information on the microbial populations in deep marine sediments are available, obviously because of the difficulty to collect samples through drilling under a deep water column. A metagenomic analysis of sediments of the Peru Margin (Ocean Drilling Program Leg 201, site 1229D) at depth from 1 to 50 m below the seafloor (mbsf) showed a minor contribution of Firmicutes to the microbial diversity, but did not give specific informations about Desulfotomaculum or related genera (Biddle et al., 2008). A metatranscriptomic study of samples from the same site at depths up to 159 mbsf was recently published. It showed that Firmicutes, and to a lesser extent Deltaproteobacteria, are the dominant phyla expressing dsr transcripts (sulfite-reductase, involved in sulfate respiration) at 5 and 30 mbsf. At greater depth, below the sulfate-methane transition zone where sulfate concentration is under detection limit, expression of dsr genes is linked to the activity of Gammaproteobacteria (Orsi et al., 2013). These data suggest that Gram-positive SRB are mainly active in shallow sediments. Interestingly the recently published culture-based study of sediments of the Juan De Fuca Ridge (IODP Site U1301) gave similar indications (Fichtel et al., 2012). Isolates of Desulfosporosinus lacus and Desulfotomaculum spp. were only retrieved from the upper 9.1 mbsf sediments, whether Desulfovibrio and Desulfotignum isolates were cultivated from deeper sediments up to 262 mbsf. Nevertheless, a culture-based study of ODP sites, 3 in the Open Pacific and 4 on the Peru Margin, including the 1229 site, did not reveal the presence of sulfate-reducing Firmicutes in samples collected at depth of 1 to 420 mbsf (D'hondt et al., 2004).

The few available studies summarized above suggest that the spore-forming SRB are found and are active only in the most superficial marine sediments, but are absent in depth, especially below the sulfate-methane transition zone. However, since too few studies of this type are currently available, it is premature to generalize these results.

One can speculate that the presence of most microorganisms inhabiting the deep subsurface is the outcome of a gradual burial of surface organisms over geological times. During this long term process, the physico-chemical conditions gradually change according to geological constraints (rock type, depth, pressure, thermogenic processes …) and organic matter is degraded or transformed through biological or abiotic processes. Spore formation is putatively one of the key characteristics of sulfate-reducing or fermentative Clostridiales helping them to survive transient periods of nutrient deprivation, unfavorable temperature or redox conditions during burial, and consequently allowing deep subsurface colonization (Kuever and Rainey, 2009). These “survival periods” could be extremely long, since it is well-documented that endospore-forming bacteria, such as Desulfotomaculum, remain viable in a wide variety of environments unfavorable for growth possibly during thousands or millions of years, like in amber (Cano and Borucki, 1995), halite crystals (Vreeland et al., 2000) or marine manganese nodules (Nealson and Ford, 1980). However, even if large amounts of spores (10⁷ endospores per cm³) can be found in several million year-old samples (Lomstein et al., 2012), many authors wonder about the possibility of spores activation and germination without DNA maintenance activity for very long periods (Lindahl, 1993; Johnson et al., 2007). A recent study of sediments from the Baltic Sea by De Rezende et al. (2013) used an MPN technique with a specific SRB culture medium. Interestingly, it was shown that the number of spores of thermophilic Desulfotomaculum decreased from 10⁴ spores per cm³ of sediment at the surface to one spore per cm³ at 6.5 m depth. Given that the sediment age at the deeper level is approximately 4500 years, the half-life of spores has consequently been estimated to be about 400 years. Nevertheless, if the survival of spores is as short as de Rezende and colleagues estimated, it seems difficult to explain the abundant presence of Desulfotomaculum in geological settings located several kilometers deep. Taking into account these contrasting results, further investigations are needed to better understand the distribution of Desulfotomaculum spp. between shallow and deep sediments, and connections between these different populations. On the other hand, microorganisms could also survive with low metabolic activity with generation times in the range of 1 × 10³ to 3 × 10³ years (Whitman et al., 1998; Lomstein et al., 2012). Sporulation could be periodically necessary to overcome extreme energy limitation because of a complete depletion of local nutrient supply (Fredrickson and Balkwill, 2006), or to facilitate passive dispersal (Isaksen et al., 1994; Jørgensen and Boetius, 2007; Hubert et al., 2009; De Rezende et al., 2013). At the end of the sedimentary process, extreme situations can be reached as in the case of the putative spore-forming Candidatus Desulforudis audaxviator that became the single bacterial autonomous species in its deep environment (Chivian et al., 2008). The detection of Desulfotomaculum spp. in northern sea sediments also illustrate how spore formation could be an advantage for the dissemination of microorganisms originating from the deep subsurface in surface environments, which may finally return to shallow sediments. Desulfotomaculum arcticum was isolated from sediments of Nordfjorden (under a 100 m deep water column) on the west coast of Svalbard in the Arctic Ocean (Vandieken et al., 2006). Since this species grows optimally at 42°C, it was suggested that it was present as spores in the cold sediments. D. arcticum share metabolic traits with most of species of the genus, including the use of H₂ as potential energy source. Two hypotheses were proposed to explain the presence of thermophilic spores in the cold seabed of the Arctic Ocean. The first refers to the oceanic crust fluids that harbor microbial communities including Firmicutes species (Cowen et al., 2003). These microorganisms might be transported to the oceanic ridge before being dispersed into the ocean via hydrothermal vents such as those of the Lost City (Brazelton et al., 2006). The second hypothesis suggests that spores come from oilfields in which Desulfotomaculum spp. have often been detected (see below) either from produced oil reservoirs (Magot et al., 2000), or from natural seeps. The physiologic traits of thermophilic bacteria found in the Arctic Sea support the idea of their adaptation to the hot subsurface conditions. It is thus suggested that the ocean floor would be connected to the oceanic crust and the oilfields by oceanic fluids. A similar idea has already been suggested by Stetter et al. (1993), assuming that hyperthermophiles from hydrothermal vents should have been transported into oil reservoirs by seawater injection.

It is noteworthy that many Desulfotomaculum spp. isolated from subsurface environments are autotrophic microorganisms. As indicated above, this metabolic ability allows them to take advantage of the recognized ubiquitous resource in hydrogen in the deep subsurface (Pedersen, 1997; Lin et al., 2005). Moreover, some of these SRB may also perform a homoacetogenic metabolism by oxidizing hydrogen and reducing CO₂ to acetate. There are more and more convincing results that such metabolism plays an important role in carbon cycling in the deep biosphere (Lever, 2011; Oren, 2012). In this respect, the delivery of acetate resulting from this metabolic process in the deep biosphere in particular, can be beneficial to the overall microbial community with a peculiar emphasis for (i) acetoclastic methanogenic archaea which have been recovered from the deep biosphere and (ii) other hydrogenotrophic SRB, or hydrogen-oxidizing methanoarchaea inhabiting subsurface ecosystems and requiring an organic carbon source for their growth. Only by taking into account these peculiar metabolic features of deep-living Desulfotomaculum spp. (autotrophic way of life and homoacetogenesis), it seems evident that they could be a significant, even major component for SLiME. Evidence has been provided at many occasions in the deep biosphere that sulfate-reducers coexist with homoacetogenic bacteria and methanogenic archaea. This is particularly true in oligotrophic ecosystems (e.g., deep granitic aquifers) (Pedersen, 1997), but also in organic compounds rich ecosystems (e.g., oil reservoirs) (Magot et al., 2000; Ollivier and Alazard, 2010) where these three prokaryotic groups have been recovered many times by cultural and molecular approaches. In the former ecosystems, hydrogen recovered from biological (fermentative process) or abiotic processes (e.g., radiolysis, mineral reactions, volcanic activity) has been recognized to drive an intra-terrestrial biosphere (Pedersen, 1997) and hydrogenotrophic autotrophic (i) homoacetogenic bacteria together with (ii) methanogenic archaea have been hypothesized to play a central role to deliver organic material in the deep biosphere (e.g., acetate for homoacetogens, methane for methanogens, and biomass). This makes the in situ bioenergetical systems independent of any photosynthetic activity (Pedersen, 1997).

Beside sulfate reduction, we must also consider that Desulfotomaculum spp may also use various other sulfur compounds (e.g., thiosulfate, elemental sulfur, sulfite) but also metals and metalloids as terminal electron acceptors (Tebo and Obraztsova, 1998) thus confirming their geomicrobiological significance in the deep biosphere where metals and metalloids in particular are quite widespread. Data on this point are nevertheless scarce, and deserve more investigations. Desulfotomaculum species were reported not only to use hydrogen as electron donor, but also a wide range of organic compounds (e.g., sugars, aromatic compounds, long chain fatty acids, organic acids) that may be also available but at less extent in some deep environments rich in organic matters. Clearly, from these metabolic traits that have been established for Desulfotomaculum spp, we may expect them to penetrate the deep biosphere, and to adapt to the in situ existing extreme physico-chemical conditions in terms of temperature, pressure, or low/rich nutrient availability. However, more exhaustive characterization of the physiology and metabolism of Gram-positive SRB should help to better understand their adaptation to their environments, or even their role in the emergence of different forms of life on Earth. As an example, while many sulfate-reducing members of the Deltaproteobacteria are recognized to disproportionate sulfur compounds such as thiosulfate, sulfite or elemental sulfur (Finster, 2008), there are very few evidences of Desulfotomaculum to perform such mineral oxidative-reductive processes. This is restricted to thiosulfate disproportionation (e.g., D. thermobenzoicum and D. nigrificans) and nothing is known so far on the ability of Desulfotomaculum spp. to disproportionate elemental sulfur into sulfide and sulfate, such reaction requiring the presence of oxidants (e.g., metal oxides) to be favorable thermodynamically (Finster, 2008). This is quite intriguing and merits further attention by the scientific community since the existence of an oxidative microbial sulfur cycling in which sulfur disproportionation playing a significant role has been postulated (Canfield, 1989; Thamdrup et al., 1993; Jørgensen and Nelson, 2004; Slobodkin et al., 2012) and even recently demonstrated (Riedinger et al., 2010) in the deep biosphere.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.