Microbial ecology of deep-sea hypersaline anoxic basins

Abstract

Deep hypersaline anoxic basins (DHABs) are unique water bodies occurring within fractures at the bottom of the sea, where the dissolution of anciently buried evaporites created dense anoxic brines that are separated by a chemocline/pycnocline from the overlying oxygenated deep-seawater column. DHABs have been described in the Gulf of Mexico, the Mediterranean Sea, the Black Sea and the Red Sea. They are characterized by prolonged historical separation of the brines from the upper water column due to lack of mixing and by extreme conditions of salinity, anoxia, and relatively high hydrostatic pressure and temperatures. Due to these combined selection factors, unique microbial assemblages thrive in these polyextreme ecosystems. The topological localization of the different taxa in the brine-seawater transition zone coupled with the metabolic interactions and niche adaptations determine the metabolic functioning and biogeochemistry of DHABs. In particular, inherent metabolic strategies accompanied by genetic adaptations have provided insights on how prokaryotic communities can adapt to salt-saturated conditions. Here, we review the current knowledge of the diversity, genomics, metabolisms and ecology of prokaryotes in DHABs.

DHABs, microbial diversity, microbial ecology, element cycles, Red Sea

INTRODUCTION

Among the most extreme places where microbial life has been reported are the so-called deep hypersaline anoxic basins (DHABs) (Mapelli et al. 2017a). Since the observation of the hot brine-filled Atlantis II Deep at the bottom of the Red Sea (Charnock 1964), many DHABs have been discovered in different locations in the Red Sea (Backer and Schoell 1972), the Mediterranean Sea (Scientific Staff of Cruise Bannock 1984-12 1985), the Gulf of Mexico (Shokes et al. 1977) and the Black Sea (Aloisi et al. 2004 ) (Fig.1). Despite the geology varying considerably for each specific case, in general, the majority of DHABs discovered so far originated from the re-dissolution of evaporitic deposits buried under layers of sediments and exposed to seawater because of tectonic activity (Cita 2006). This re-dissolution process generates the DHABs and a gradient of salinity, the brine-seawater interface (BSI), at the boundary between the hypersaline water mass and seawater. Such a salinity gradient is a halocline where the salt dissolution increases the water density determining a pycnocline. In parallel to the gradient of salinity and density a chemocline occurs with gradients of the many chemical species that form redox couples capable of supporting different microbial metabolisms.

Figure 1.

Locations of the most studied DHABs in the Gulf of Mexico (green dots), the Mediterranean Sea (blue dots), the Black Sea (black dot) and the Red Sea (red dots).

The Red Sea hosts the largest number of documented DHABs (Fig. 1) (Backer and Schoell 1972; Pautot et al. 1984). It has been proposed that during the Miocene period, the Red Sea became isolated from seawater influxes through the Bab El Mandeb Strait and the Gulf of Suez, possibly because of similar events that originated the Messinian Salinity Crisis in the Mediterranean Sea (Hsu, Stoffers and Ross 1978; Gargani, Moretti and Letouzey 2008; Garcia-Castellanos and Villasenor 2011). These events might have led to the precipitation of several km of evaporites in the central trough of the Red Sea (Searle and Ross 1975).

The Red Sea brines are categorized in two types depending on the postulated seismic origin and biogeochemical composition (Schmidt, Al-Farawati and Botz 2015). Type I, collapse-type DHABs, include Oceanographer Deep and Kebrit Deep. They present ranges of high salinity (167–182 mg Cl⁻ l⁻¹), low pH (5.5–5.6) and H₂S (15–24 ml l⁻¹), mild warm temperatures (23.4–24.9°C) and low traces of heavy metals (Schmidt, Al-Farawati and Botz 2015). The geochemical characterization of type I DHABs indicates that their composition is not influenced by input of hydrothermal fluid and water/volcanic rock interaction, but rather by evaporite dissolution and sediment alteration. Type II, intrusion-/extrusion-related DHABs, include the Conrad Deep, Shaban Deep and others located in the multi-deep regions of the Central Red Sea (Fig.1). Geochemically, these brines feature high concentrations of manganese (7–90 mg ml⁻¹) and other metal ions, possibly due to hydrothermal input and subsurface water/volcanic rock interaction (Schmidt, Al-Farawati and Botz 2015). Many of the DHABs in the northern (Conrad, Shaban and Nereus) and central regions (Chain, Valdivia, Discovery, Shagara, Wando, Albatros, Atlantis II, Erba, Suakin and Port Sudan) of the Red Sea might have originated from the re-dissolution of evaporitic deposits subjected to the seismic events during the early seafloor spreading (Bonatti 1985). For instance, the influence of hydrothermal fluids is evidenced - and continues to occur - in Atlantis II Deep and the interconnected Discovery Deep (Hartmann 1985; Hartmann et al. 1998; Swift, Bower and Schmitt 2012; Schardt 2015).

The deepest brine pools described to date are located in the eastern Mediterranean Sea, all lying at depths greater than three km below the sea level (Fig. 1; Cita 2006). Bannock, Discovery, Kryos, L’Atalante, Medee, Thetis, Tyro and Urania sit on the Mediterranean Ridge accretionary wedge, which lies from west to east in the south of the island of Crete. Other brine pools have been reported in the Nile Deep Sea Fan off the Nile river delta (Huguen et al. 2009). Geological data suggest that all the basins were produced by the dissolution of different layers of Messinian evaporites generated during the Messinian Salinity Crisis (Garcia-Castellanos and Villasenor 2011), with subsequent fluid migration in confined depressions on the seafloor (Bortoluzzi et al. 2011). The water masses that dissolved the evaporites can be of variable origins. For instance, in the Discovery DHAB the upmost layer of evaporites was dissolved by subsurface waters rather than seawater and remained entrapped in the sediments for unknown periods of time before reaching the actual Discovery basin (Wallmann et al.1997; Wallmann et al.2002). The variable processes of evaporites dissolution and brine generation have determined substantial chemical differences in the composition of Mediterranean DHABs (van der Wielen et al. 2005).

The brines in the Gulf of Mexico originated from the dissolution of evaporitic sheets, part of the Louann salt formation. Such a salt deposit was originally formed during the middle Jurassic (175–145 million years ago) and covered by sediments in the early Miocene. Gradual accumulation and differential loading of sediments caused the compression and generation of salt domes. Seismic events subsequently cracked the sediments leading to the dissolution of the evaporites (Cordes, Bergquist and Fisher 2009), originating several DHABs (Bouma and Bryant 1994; Roberts et al. 2007). Some studies have shed some light on the microbial ecology of DHABs of the Gulf of Mexico, reviewed by Antunes, Ngugi and Stingl (2011) and Mapelli, Borin and Daffonchio (2012), and references therein. The Orca basin (Fig.1), a 400 km² basin lying at about 2.25 km below sea level at the boundary between the Green Canyon and Walker Ridge protraction areas (Pilcher and Blumstein 2007), was the first DHAB discovered in the Gulf of Mexico (Shokes et al. 1977; van Cappellen et al. 1998; Shah et al. 2013) and was followed by other smaller, low-salinity brines/mud volcanoes such as NR-1/GC233 and GB425 (Brooks et al. 1979; MacDonald et al. 1990), which have received attention in the last few years (Joye et al. 2009).

The Dvurechenskii mud volcano in the Black Sea (Aloisi et al. 2004; Fig. 1) contains highly saline fluids enriched in Li, B, Ba, Sr, I and dissolved inorganic nitrogen. The analysis of the geochemical properties of the fluids indicated that they are not originated by the dissolution of evaporites, but from diagenetic processes (Aloisi et al. 2004).

UNIQUE FEATURES OF DHABs

Physical structure and productivity of DHABs

Due to the high density (up to 1.35 g cm⁻³ for the Discovery DHAB in the Mediterranean Sea; van der Wielen et al. 2005), and the high hydrostatic pressure, these brine pools hardly mix with the overlying deep seawater, and are often considered as lakes at the bottom of the sea. Some DHABs, such as Urania in the Mediterranean Sea (Borin et al. 2009) and Atlantis II Deep in the Red Sea (Bougouffa et al. 2013), encompass multiple vertically stratified brine bodies along the halocline that are stably separated by their different densities. Within the physical structure of single brine bodies, different layers can be recognized (Fig.2A), each contributing to the productivity and functioning of the entire system. The brine body, which can vary in depth from a few to several hundred metres (van der Wielen et al. 2005; Cao et al. 2015), typically shows uniform salinity and physicochemical characteristics. At the bottom of the brine, the brine-sediment interface often has greater environmental variability than the brine body due to fluid and gas emissions from the subsurface and the accumulation of organic matter sinking from the brine body (Yakimov et al. 2007a). The most productive layer of DHABs is the BSI between the oxygenated upper water column and the brine body, which constitutes the functional hotspot of the DHABs.

Figure 2.

General physical structure of a DHAB and fine-scale sampling strategy of the brine-seawater Interface (BSI). (A) Schematic drawing of a DHAB showing the different layers of the BSI, brine and brine-sediment interface. On the left side of the panel, a rosette-based approach for sampling the BSI is illustrated (Daffonchio et al. 2006). A rosette sampler equipped with a Conductivity, Temperature and Depth (CTD) sensor and with Niskin bottles is connected to the modus vehicle equipped with a live camera used to monitor the sampling operations. The modus vehicle is optional as the limits of the brine-seawater interface can be identified by the CTD on the rosette sampler. On the right side of the panel, a schematic representation of the brine-trapper used to sample the BSI in DHABs in the Gulf of Mexico is illustrated (Joye et al. 2005). The brine-trapper is a 3-mcylinder attached to the submersible Johnson Sea Link II in a horizontal position. Once above the brine, the instrument is lowered into the brine-seawater interface in a vertical position. The water enters in different compartments along the tube to entrap different layers of the interface. (B) Once on board, the BSI water sampled with the Niskin bottles is aliquoted in fractions of 0.5–2 lthat represent the salinity gradient in the BSI. (C) With the brine-trapper, the water in the different compartment (∼200ml) is extracted from the tube for further analysis.

The material sinking from the water column, which includes an array of diverse dissolved organic carbon molecules (Arrieta et al. 2015), accumulates in the BSI due to the density barrier, fueling microbial productivity (Shah et al. 2013). Peaks in the range of 20–25 nmol C_org l⁻¹ day⁻¹ of heterotrophic carbon biomass production (estimated using [³H]-leucine uptake) have been measured in the upper BSI of Medee DHAB (Yakimov et al. 2013). Similarly, the highest rates of carbon fixation were recorded in the upper BSI of L’Atalante (Yakimov et al. 2007b), Medee (Yakimov et al. 2013) and Thetis (La Cono et al. 2011). For instance, in the upper BSI of the Thetis DHAB, carbon fixation measured by [¹⁴C]-bicarbonate assimilation rates has been determined to be 315 ± 82 nmol C_org l⁻¹ day⁻¹ (given as rates of production of organic carbon), eight times higher than that in the overlying seawater (40 ± 8 nmol C_org l⁻¹ day⁻¹; La Cono et al. 2011). Quantitative data about the microbial populations showed that the interface is typically enriched in prokaryotic cell counts. In different DHABs, prokaryotic numbers increase from values in the order of 10⁴–10⁵ cells ml⁻¹ in normal seawater, to 10⁶–10⁷ cells ml⁻¹ in the BSI, then decrease again in the brine body (Borin et al. 2009; Joye et al. 2009; La Cono et al. 2011).

Depending on the chemistry of the brines, the BSIs and the chemoclines therein can be of variable thickness, ranging from one metre (Borin et al. 2009) to several tens of metres (Bougouffa et al. 2013). The thickness of the BSI is presumably determined by the brine chemistry and density, and/or by the currents within, above and around the DHAB. Within the BSI chemoclines, the changing pools of organic matter and concentrations of chemical species determine the vertical profiles of suitable redox couples, electron donors and acceptors likely feeding specific microbial metabolic groups. These characteristics determine microbial stratification along the chemocline. For instance, the highest biological activities in the DHABs have so far been measured in the BSIs relative to the proximal waters in the brine and in the oxic water above (Daffonchio et al. 2006). The microbial diversity and functionality of the communities that are thriving along the BSI chemoclines have been characterized at a 10-cmscale resolution. Specifically designed sampling strategies have been used, such as the fractioning of the BSI layers sampled with Niskin bottles at different depths within the chemoclines (Fig.2a and 2b) (Daffonchio et al. 2006), or the use of a specifically designed sampling bottle, the brine-trapper deployed by a submersible (Fig.2a and 2c) (Joye et al. 2005).

The presence of highly stratified microbial communities at the interface of DHABs (Daffonchio et al. 2006; Borin et al. 2009; Joye et al. 2009) is further reinforced by studies on the predicted effect of density stratification on microbial cells (Doostmohammadi, Stocker and Ardekani 2012). In stratified fluids, flow effects on small aquatic organisms (1 µm–2mm) are size-dependent. Prokaryotes (in the order of a few µm), phytoplankton and small protists (10–100 µm) are hardly affected by flow and therefore are predicted to move freely inside pycnoclines (Doostmohammadi, Stocker and Ardekani 2012). This suggests that the microbial stratification observed at the interface of Bannock and Urania (Daffonchio et al. 2006; Borin et al. 2009), or in the brine pools in the Gulf of Mexico (Joye et al. 2009), is likely determined by the changing environmental conditions and availability of suitable nutrients along the chemocline.

Constraining conditions on cell survival and macromolecule stability

A major feature of the DHABs is the extreme salinity that can reach values several orders of magnitude higher than in the overlying seawater. Ionic concentrations higher than 4 M have been measured in L’Atalante (4.7 M Na⁺) and Discovery (5 M Mg²⁺) in the Mediterranean Sea (van der Wielen et al. 2005). Such salinities significantly constrain the metabolic activities of cells (Oren 2011) by decreasing water activity and impairing cell turgor, and by exerting ionic, kosmotropic and chaotropic effects on the cells (Hallsworth, Heim and Timmis 2003; Hallsworth et al. 2007). In the absence of order-making kosmotropic solutes such as NaCl that stabilize the interactions of water molecules in the presence of the chaotropic MgCl₂, the molar concentration of this salt permissible for life has been estimated ~ 2.3 M. Under such conditions the indicators of microbial life, such as those based on DNA profiling of microbial communities, may be erroneous due to the high stability of remnant DNA material in chaotropic salts (Hallsworth et al. 2007). The informational content of bacterial plasmid DNA molecules exposed to Mediterranean Sea brines, including the Mg²⁺-rich Discovery basin, was preserved for more than one month with minor effects on plasmid conformation. The exposure to the brine waters of these plasmid DNA molecules did not affect the capability of the naturally competent Acinetobacter baylyi BD413 to uptake them and express the genes present on the plasmids (Borin et al. 2008). However, mRNA that is more labile than DNA has been successfully recovered from the brine body of the Kryos basin, characterized by MgCl₂ salt concentrations of 2.27–3.03 M, suggesting the presence of active microbial cells at salt concentrations previously considered non-permissive for life. The slightly elevated concentrations of Na⁺ and SO₄²⁻ ions were believed to compensate against the extreme chaotropicity of MgCl₂ due to their kosmotropic effect (Yakimov et al. 2015).

To cope with high salinity, halophilic prokaryotes employ the salt-in or the compatible-solute strategies (Oren 1999). The salt-in strategy is used by few halophiles because it consists of the accumulation of high ionic (mostly K⁺) concentrations in the cytoplasm and so requires the adaptation of the entire cellular machinery to such concentrations (Oren 2011). The compatible-solute (e.g. ectoine, glycine-betaine, glutamate, glutamine and proline) strategy has been described in different prokaryotes inhabiting the DHABs. Pathways for the synthesis of osmolytes were found in the genomes, obtained from DHABs, of ammonia-oxidizing archaea affiliated with the genus Nitrosopumilus (Ngugi et al. 2015) and a recently described anaerobic ammonium-oxidizing bacterium, Ca Scalindua rubra (Speth et al. 2017). However, exposure of moderately halotolerant bacterial isolates (including sporeformers) to brines from the Discovery basin, and three other DHABs from the Mediterranean Sea, induced cell death within 2 h to 144 days, depending on the strain and the brine type, with the Discovery brine being the most aggressive (Borin et al. 2008).

Although temperature varies in different DHABs, values have generally been found to resemble those in the overlying deep seawater column (~14°C in the Mediterranean Sea and 22°C in the Red Sea) (van der Wielen et al. 2005; Eder et al. 2001). Notable exceptions are the interconnected Atlantis II Deep and the Discovery Deep in the Red Sea, which have temperatures up to 68°C and 45°C, respectively, the highest temperatures ever recorded for DHABs (Swift, Bower and Schmitt 2012). Temperature was predicted to be a strong driver of microbial assemblage in these brine pools (Bougouffa et al. 2013). The capacity to synthesize thermoprotective molecules (e.g. hydroxyectoine) (Tanne et al. 2014) may have selected for thermophiles, as deduced from the reconstructed genome of key organisms present in Atlantis II Deep (Ngugi et al. 2015; Ngugi et al. 2016).

At the depth where typically the DHABs of the Mediterranean Sea are located (deeper than three km below sea level), hydrostatic pressures higher than 30 MPa may select piezophilic microorganisms (Kato and Qureshi 1999). While the method of sampling deep seawaters has been shown to variably affect the level of gene expression in prokaryotes and eukaryotes (Edgcomb et al. 2016), it was recently shown that even mild hydrostatic pressures such as 5–10 MPa might affect the response of microorganisms, determining loss of activities and inducing stress response (Scoma et al. 2016). Such effects, for which the term Alcanivorax paradox was coined, after hydrocarbon degraders of the genus Alcanivorax were found to be undetectable in the deep plume of the Deepwater Horizon oil spill (Mapelli et al. 2017b), might represent a differential factor of microbial selection between deep and shallower DHABs, such as those in the Mediterranean Sea and in the Red Sea, respectively.

A typical feature of DHABs is the depletion of oxygen usually already a few tens of cm inside the BSI and full anoxia in the brine bodies. Electron acceptors alternative to oxygen (e.g. iron, manganese, sulfate, elemental sulfur, carbon dioxide, nitrite and nitrate) are available in the DHAB layers, possibly allowing the existence and interconnection of a wide variety of anaerobic metabolisms (van Cappellen et al. 1998; Borin et al. 2013; Guan et al. 2015). However, the lack of oxygen, the most energy-yielding electron acceptor, in combination with the multiple stressors in the DHABs, represents a further challenge affecting the adaptation of novel microorganisms.

Enigmatic microbial inhabitants and higher organisms

Several studies indicate that DHABs are hotspots of many novel undefined taxa (Antunes, Ngugi and Stingl 2011), most of which remain uncultivated to date. Because of their differing chemistries (Table 1, Fig. 3), the microbial communities in individual DHABs are composed of unique phylotypes and functionally discrete groups of organisms (La Cono et al. 2011). Consequently, most of the metagenomic proteins deduced from these DHAB - both in the seawater-brine interface and the brine - encompass hypothetical proteins exhibiting little sequence homology to characterized proteins in public databases (Ferrer et al. 2012).

Figure 3.

Dendrogram grouping the different brines of DHABs of the Gulf of Mexico (green), Mediterranean Sea (blue) and the Red Sea (red) by hierarchical clustering using the Euclidean distance metric and average linkage based on the concentrations of Na⁺, Cl⁻, K⁺, Ca²⁺, Mg²⁺, SO₄²⁻ and H₂S. Data are summarized in Table 1.

Table 1.

Concentrations of chemical species available for different DHABs in the Red Sea, the Mediterranean Sea and the Gulf of Mexico.

	Category	Na+	Cl−	Mg2+	K+	Ca2+	SO42−	Sulfide
Red Sea
RS-SWa		547.33	584.18	65.75	9.41	12.4	33.37	n.a.
Albatrossa	T	4344.88	4608.75	48.84	55.96	111.98	14.89	n.a.
Atlantis II-Na	T	4615.09	5149.66	35.42	85.84	149.96	10.86	n.a.
Atlantis II-SWa	T	4677.73	5189.97	35.42	86.88	148.71	10.74	n.a.
Conrada	T	4295.43	4619.67	101.01	67.7	24.63	48.77	n.a.
Discoverya*	T	4635.53	5021.75	36.78	84.86	144.87	10.66	n.a.
Erbaa	T	2621.99	2678.14	71.43	30.92	29.57	47.15	n.a.
Kebrita	T	4805.74	5135.67	121.13	36.7	56.56	20.08	1.13
Oceanographera	T	4353.37	4716.86	280.64	77.09	86.21	10.32	0.71
Shaban-Na	T	4784.51	4900.52	97.96	49.29	19.74	51.57	n.a.
Shaban-Sa	T	4844.97	4970.92	84.96	48.77	22.16	48.47	n.a.
Nereusa	T	3556.62	4131.47	64.06	77.42	230.8	11.25	n.a.
Port Sudana	T	3784.68	3855.22	70.89	46.5	35.63	47.77	n.a.
Suakina	T	2416.77	2612.84	62.87	35.42	57.14	36.05	n.a.
Mediterranean Sea
MS-SWb		480	560	54.5	10.4	10.5	28.9	<0.002
Bannockb	T	4200	5378	643.9	126.3	16.3	135.2	2.9
Discoveryc*	AT	840.17	10154.29	5142.97	89.52	1	110.35	0.85
Kryosc	AT	1236.32	9054.24	4402.39	84.4	1	322.71	1.2
L’Atalante 1d	T	4670	5290	533	300	5.9	323	2.9
L’Atalante 2c	T	4654.24	5302.8	658.3	368.31	7.49	333.13	2.82
Medeee	T	4178	5259	788	471	2.8	201	1.64
Thetisd	T	4760	5300	604	230	9	265	2.12
Tyrob	T	5300	5350	71.1	19.2	35.4	52.7	2.1
Uraniad	T	3505	3730	315	122	31.6	107	15
Gulf of Mexico
GoM-SWf		462	564	11	43	11	29	∼0
GB425f	T	1790	2114	8.7	89	59	<1	0.004
GC233f	T	1751	2092	97	22	36	<1	0.002
Orca Basing	T	4240	4450	42.4	17.2	29	20	0.025

All values are mM. Abbreviations: n.a., data not available. N, North; S, South; SW, South West; RS-SW, Red Sea seawater; MS-SW, Mediterranean Sea seawater; GoM-SW, Gulf of Mexico seawater; T, thalassohaline; AT, athalassohaline. * Note that two different DHABs are named Discovery, one in the Red Sea and the other in the Mediterranean Sea. Data are from: ^a Schmidt, Al-Farawati and Botz (2015); ^b De Lange et al. (1990); ^c Yakimov et al. (2015); ^d La Cono et al., (2011); ^e Yakimov et al. (2013); ^f Joye et al. (2005); ^g Van Cappellen et al. (1998).

Animals are also conspicuous residents of DHABs (Danovaro et al. 2010; Neves et al. 2014). Danovaro and colleagues (2010) have presented evidence of the first ever described metazoan reported to permanently live anaerobically in the sediments of L’Atalante, even though the ability of such an organism to survive their entire lifecycle in the absence of oxygen is debated (Bernhard et al. 2015; Danovaro et al. 2016). Sea anemones, sponges, tubeworms, polychaetes, hydroids, clams, top snails, gastropods, crabs and mussels include some of the macrofauna lying at the edges of the DHABs and on the beaches surrounding the BSI (MacDonald et al. 1990; Nix et al. 1995; Batang et al. 2012; Niemann et al. 2013; Vestheim and Kaartvedt 2016). A detailed discussion of animal-associated or mat-forming microorganisms is beyond the scope of this review. Nevertheless, the prokaryote (endo)symbionts living together with their eukaryotic hosts (Dubilier, Bergin and Lott 2008), as well as the viral communities (Danovaro et al. 2005), represent an unchartered facet in the ecology of DHABs currently lacking systematic investigations.

THE MICROBIAL DIVERSITY AND ECOLOGY OF DHABs

Red Sea DHABs

The Red Sea DHABs have been intensively studied for their microbial communities with pioneering molecular descriptions of novel organisms (Eder, Ludwig and Huber 1999; Eder et al. 2002; Antunes, Ngugi and Stingl 2011). Except in Kebrit Deep and Atlantis II Deep, bacteria generally dominate over archaea, similar to the Mediterranean Sea DHABs (Guan et al. 2015). A rather wide diversity is also found in the BSI relative to the overlying deep seawater column (Bougouffa et al. 2013; Abdallah et al. 2014; Guan et al. 2015; Ngugi et al. 2015) with differences between the upper and lower portions of the BSI akin to Mediterranean Sea DHABs (Yakimov et al. 2007b; Borin et al. 2009).

In the Atlantis II Deep and Discovery Deep the bacterial community in the different layers was represented by lineages affiliated with Moritella, Ca. Scalindua, Nitrospina, Marinomonas, Planctomyces, Sulfurimonas and KB1 (Bougouffa et al. 2013; Guan et al. 2015). Several of these bacterial lineages belong to the following groups: Deferribacteres and Planctomycetes in the interfaces of Atlantis II Deep and Discovery Deep, respectively, and Gammaproteobacteria in the brine body of both (Bougouffa et al. 2013). Many of these groups, as in the case of Atlantis II Deep, are composed mainly of heterotrophic taxa potentially able to use aromatic compounds as suggested by the detection of aromatic compound degraders (such as Phyllobacterium) and enrichment of the related metabolic pathways for degrading hydrothermally generated aromatic compounds (Wang et al. 2011; Wang et al. 2013; Abdallah et al. 2014). Signatures of aerobic methane-oxidizing bacteria and unique uncultured taxa with divergent alkane monooxygenases were found in the brine-seawater interface of Atlantis II Deep and Kebrit Deep but not of Discovery Deep (Bougouffa et al. 2013; Abdallah et al. 2014).

In Kebrit Deep, Erba Deep and Nereus Deep, members of Deltaproteobacteria, in particular Desulfohalobiaceae and Desulfobacteraceae, predominated in the brine-seawater interface (Guan et al. 2015). In Kebrit Deep, aerobic methanotrophic taxa were also detected, likely driven by the juxtaposed presence of methane and oxygen in the BSI (Abdallah et al. 2014). The sulfidic Kebrit Deep is also home to the KB1 clade in both the BSI and the sediment-brine interface (Eder et al. 2001). KB1 also inhabits other brines from the Red Sea (Eder et al. 2002) and the Mediterranean Sea (van der Wielen et al. 2005).

Even though Archaea dominate in the deep ocean (Karner, DeLong and Karl 2001), their genetic and metabolic diversity was found to be lower in Atlantis II Deep and Discovery Deep (Wang et al. 2013). Mostly autotrophic archaea capable of CO₂ fixation and methane oxidation were identified in these brines, while the Thaumarchaeota (Stahl and de la Torre 2012) almost exclusively dominated the BSI (Ngugi et al. 2015; Guan et al. 2015). The major ammonia-oxidizing thaumarchaea are phylogenetically affiliated to the genus Nitrosopumilus and are distinct from bathypelagic Thaumarchaeota, sharing only about 54% of their predicted genetic inventory with bathypelagic thaumarchaea (Ngugi et al. 2015). Recent studies in the brine body of Atlantis II Deep and the interfaces of Kebrit Deep and Erba Deep also implicated methanogenic archaea in the dark primary production (Guan et al. 2015).

Mediterranean Sea DHABs

The majority of the Mediterranean DHABs are thalassohaline having as major dissolved ions those of seawater (Table 1). Two DHABs are athalassohaline, Discovery and Kryos, characterized by high Mg²⁺ concentrations (Table 1) possibly originating from the dissolution of bischofite.

In the Thetis DHAB aerobic heterotrophic halophiles of the archaeal family Halobacteriaceae occupy the upper BSI (up to 110 g l⁻¹ salinity). Low salinity and anoxia prevent their proliferation in the overlying seawater and in the lower part of the BSI and the brine body, respectively. Salinity restricts the distribution of Planctomycetes and Bacteroidetes to the upper BSI, whereas KB1 bacteria are distributed from the upper BSI to the brine body indicating adaptation to a wide range of salinities (up to 348 g l⁻¹ in the brine body). Enrichment of Deltaproteobacteria and high concentrations of sulfide (HS⁻, 2.12 mmol l⁻¹) in the brine body indicate the occurrence of active dissimilatory sulfate reduction (La Cono et al. 2011).

A study on the DHAB Medee gave interesting insights about the correlation between the geochemical milieu and the key microbial groups inhabiting the system (Yakimov et al. 2013): Mediterranean Sea Brine Lake 1 (MSBL1) and Candidatus Acetothermia (formerly Candidate Division KB1) are the dominant archaeal and bacterial prokaryotes, respectively. Syntrophic interactions between these two groups have been hypothesized based on the metabolism of glycine-betaine, a common osmoprotectant produced by a variety of moderate halophiles (Yakimov et al. 2013). According to this hypothesis, KB1, a deep phylogenetic lineage close to Thermotogales and Aquificales at the root of Bacteria domain (Eder, Ludwig and Huber 1999), could be able to degrade glycine-betaine to acetate and trimethylamine using H₂ as an electron donor. Trimethylamine in turn would support methylotrophic methanogenic activity of MSBL1, which would produce H₂. Measurements of glycine-betaine, trimethylamine, acetate and methane concentration profiles along the chemocline, together with methane production rates at different depths of the brine body and enrichment cultures on glycine-betaine and trimethylamine, support such a syntrophic interaction hypothesis. Recent reconstructions of the KB1 (from the Orca Basin; Nigro et al. 2016) and MSBL1 (Atlantis II Deep, Discovery Deep, Nereus Deep, Erba Deep and Kebrit Deep; Mwirichia et al. 2016) genomes, revealed that both organisms have the metabolic features to produce substrates for methanogens. KB1 can potentially use glycine-betaine not only for osmoregulation but also as a carbon and energy source (Nigro et al. 2016). MSBL1 does not present the core genes of methanogenesis, but the determinants for the Embden-Meyerhof-Parnas pathway together with the Wood-Ljungdahl pathway or the reductive TCA cycle suggest that it has a mixotrophic lifestyle, being capable of fermenting glucose when available, or, in the absence of organic carbon, fixing carbon dioxide (Mwirichia et al. 2016).

The Bannock and L’Atalante DHABs present some analogies in the major microbial groups. Among Archaea, ammonia-oxidizing Thaumarchaeota Marine Group 1 and MSBL groups were dominating the 16S rRNA gene libraries. Within bacteria, sulfate-reducing Deltaproteobacteria and sulfur-oxidizing Gamma- and Epsilonproteobacteria were among the most abundant clones together with the KB1 lineage (Daffonchio et al. 2006; Yakimov et al. 2007). The Urania basin was shown to be dominated by methanogenic archaea, sulfate reducers and sulfide oxidizers, supporting the dominance of methanogenesis and the sulfur-based metabolisms (Borin et al. 2009).

The two athalassohaline DHABs Discovery and Kryos presented, similarly to all the other thalassohaline brine lakes, dominance of the bacterial community over the archaeal one (van der Wielen et al. 2005; Yakimov et al. 2015). Bacterial representatives of Gamma-, Epsilonproteobacteria, Sphingobacteria and Halobacteria were detected only in the Discovery brine, whereas dominance of representatives of candidate division KB1 and Deltaproteobacteria was found in both DHABs. In the interface of Kryos, the population of KB1 was found to be uniform along the salinity gradient, whereas different groups of Deltaproteobactria were identified at different depths: sulfate reducers related to Desulfotignum and Desulfosalsimonas were dominant in the less saline layer, while signatures of unknown Deltaproteobacteria were found in the deeper layer of the BSI. Among Archaea, MG1 Thaumarcheota were only detected in the upper layer of the BSI of Kryos DHAB, whereas MSBL1 were dominant in the deeper layers of the interface of both Discovery DHAB and Kryos DHAB. Interestingly van der Wielen and co-workers(2005) retrieved in the Discovery DHAB sequences related to the archaeal genus Halorhabdus, which include the species H. tiamatea, isolated from the brine-sediment interface of Shaban Deep (Antunes et al. 2008a), able to tolerate up to 1 M of Mg²⁺.

Gulf of Mexico DHABs

In the NR1/GC233 brine pool and the GB425 mud volcano lying on the continental slope of the Gulf of Mexico, active microbial populations involved in sulfur cycling and methanogenesis were identified (Joye et al. 2009). In NR1, the existence of a dynamic sulfur-cycling microbial community was suggested by the correlation of geochemical data and molecular signatures of different Deltaproteobacteria sulfate-reducers (related to Desulfosarcinales, Desulfobacterium, Desulfobulbus and Desulfocapsa) and of sulfide-oxidizing Epsilonproteobacteria (Joye et al. 2009). The GB425 mud volcano community was instead characterized by a low diversity of sulfate-reducing bacteria, with only two phylotypes retrieved: one related to Desulfosarcinales and one to Desulfobacterium (Joye et al. 2009).

A diverse community of methanogens was observed in NR1 with signatures of the genera Methanolobus, Methanosaeta, Methanoculleus and Methanospirillum (Joye et al. 2009). However, low rates of acetoclastic and hydrogenotrophic methanogenesis were measured, suggesting that methanogenesis from substrates such as methanol or methylated amines was dominant rather than from acetate or H₂. On the contrary, the signatures of acetoclastic methanogens dominated over hydrogenotrophic and methylotrophic species (Joye et al. 2009).

Biogeochemical cycles in the DHABs

The steep gradients and extreme environmental conditions of each DHAB have created niches that lead to the selection of distinct prokaryotic communities peculiar to each DHAB and the different brine layers. Consequently, a wide biodiversity and unique metabolic features are expected along the halocline. The pycnocline-driven isolation of the different transition zones from the oxygenated overlying water column have given rise to complex biogeochemical cycles sustaining microbial life in the brines (Fig.4). The DHABs studied to date generally show evidence of sulfate reduction, acetogenesis, methanogenesis and heterotrophic activity, but also different processes of the nitrogen cycles.

Figure 4.

Simplified scheme of the biogeochemical processes occurring in the DHABs. On the left the redox potential E_h (continuous line) and salinity (dash line) profiles are taken as proxy of the BSI chemocline. The biogeochemical processes are schematically positioned in the BSI along the redox potential gradient. The names of the different processes in each schematic layer of the BSI are referred to in the reactions reported from left to right. MA, methylamine; GB, glycine-betaine.

Carbon cycle

CO₂ fixation

Initial hypothesis concerning microbial metabolic pathways in DHABs proposed that the majority of microbes most likely depend on organic material sinking from the overlying sea-water column as a source of carbon (Sass et al. 2001). However, the potentially limited availability of organic carbon, especially in the deepest DHABs, suggests chemoautotrophy likely features as a prominent metabolic lifestyle of communities thriving in the brine-seawater interface and the brine body. The first evidence of chemoautotrophy, namely methanogenesis, in DHABs was observed in the L'Atalante, Bannock, Urania and Discovery DHABs from the Mediterranean Sea. Methanogenesis was detected by both molecular methods (16S rRNA gene clone libraries) and measurements of methane production (van der Wielen et al. 2005). Even though the prediction of living microorganisms in the Discovery DHAB can be erroneous when based on DNA signatures, due to the high stability of remnant DNA in presence of chaotropic salts (Hallsworth et al. 2007), genes for the type I (cbbL) and II (cbbM) ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) have been later found in the BSI and in the brine of the Discovery basin. The cbbL genes were predicted to originate from chemoautotrophic Gammaproteobacteria, while the type II RuBisCO suggested the presence of chemoautotrophic facultative anaerobic bacteria (van der Wielen 2006), possibly capable of anaerobic sulfide oxidation with nitrate as an electron acceptor (Elsaied and Naganuma 2001). In the L’Atalante Basin, it was shown that dark primary production rates were higher in the BSI than in the overlying deep seawater and the brine body, which also coincided with the highest prokaryotic biomass (Yakimov et al. 2007b).

An active role of bacteria in autotrophic CO₂ fixation has been documented in the Thetis basin (La Cono et al. 2011). Evidence for aerobic and microaerophilic CO₂-fixing pathways were found only in the upper BSI, and were related to bacterial species such as Thiomicrospira halophile and members of the Epsilonproteobacteria. An eightfold autotrophic CO₂ fixation activity was measured in the upper BSI rather than in the overlying deep seawater. These findings were later confirmed by the detection of Thiomicrospira-like RNA transcripts of enzymes in the Calvin-Benson-Bassham cycle (Pachiadaki et al. 2014).

The reductive acetyl-CoA pathway (the Wood-Ljundahl Pathway (WLP)) was proposed to occur in the Thetis basin (Pachiadaki et al. 2014). WLP is typically found in sulfate-reducing bacteria, methanogenic archaea, acetogenic prokaryotes, and microorganisms performing anaerobic methane oxidation (Berg et al. 2010; Hugler and Sievert 2011). It should be noted that methylotrophic type of both methanogenic and acetogenic WLP are likely to be thermodinamically more relevant and can be indipendent from CO₂ fixation (Chistoserdova et al. 2009; Drake and Daniel 2004). Metatranscriptomic analysis of the upper and lower BSI documented the presence of transcripts belonging to sulfate-reducing bacteria of the family Desulfobacteraceae, the anaerobic methane oxidizers group 1 (ANME1) and the methanogenic archaea Methanoregula formicica (Pachiadaki et al. 2014).

Methanogenesis and anaerobic methane oxidation (AOM)

Given the thermodynamic limitations imposed by high salinity—that is, increasing salinity requires the uptake or synthesis of osmolytes—methylotrophic methanogenesis (from methylated amines, methanol and dimethyl sulfide) with higher energetic yield is thermodinamically more relevant in hypersaline environments in comparison to aceticlastic and hydrogenotrophic methanogenesis, with lower energetic yields (Oren 1999). The high concentrations of dissolved methane observed in numerous DHABs (reviewed in Antunes, Ngugi and Stingl 2011), the detection of high methanogenic activities (Joye et al. 2009) and few taxa related to methanogenic archaea (Daffonchio et al. 2006; Guan et al. 2015) support that methanogens contribute to the carbon biogeochemistry of DHABs. Because of the apparent thermodynamic constrains under high salinity it has been proposed that halophilic methanogens likely generate methane from methylamines, a fermentation product of the osmoprotectant glycine-betaine (Borin et al. 2009; Yakimov et al. 2013), but so far no methylotrophic methanogen has been isoalted from marine DHABs.

Recent molecular studies on methanogenesis in the Thetis basin led to the discovery of different groups of transcripts of the key enzyme, methyl coenzyme M reductase (mcrA; La Cono et al. 2011). The first group was specifically detected in the brine body and was closely related to Methanohalophilus, while the second group affiliated to ANME1 was also retrieved in the BSI. It was later found that anaerobic methane oxidation, rather than methanogenesis, dominates in the lower interface of Thetis based on the detection of ANME1 sequences and the lack of genes encoding coenzyme F420-dependent N5,N10-methenyltetrahydro-methanopterin dehydrogenase, an essential enzyme of methanogenesis (Pachiadaki et al. 2014).

In the brine body of Atlantis II Deep, and in the BSI of Erba Deep and Kebrit Deep in the Red Sea, 90% of mcrA sequences clustered with cultivated representatives of the genera Methanohalophilus and Methanococcoides, which are known to utilize methylated compounds as substrates (Guan et al. 2015). These observations lead to the inference that the disproportionation of methanol and methylamines are likely the main methanogenic pathways in these DHABs. In the Kebrit Deep BSI, one mcrA operational taxonomic unit (OTU) could not be assigned to any of the known methanogens, but showed 78% sequence similarity with that of Ca Methanoperedens nitroreducens, suggesting the presence of anaerobic methane-oxidizers in this DHAB.

A recent group of extremely halophilic methyl-reducing methanogens, a class-level lineage called “Methanonatronarchaeia” within the phylum Euryarchaeota was discovered in hypersaline sediments from deep lakes (Sorokin et al. 2017b). Strikingly, “Methanonatronarchaeia” is closely affiliated with environmental 16S rRNA gene sequences of the uncultured Candidate Division Shaban Archaea (SA1; Eder et al. 2002), in turn, providing the first clue of their metabolism and role in carbon cycling in DHABs. The genomic features of this newly-discovered methyl-reducing methanogen, i.e. the incomplete set of genes for part of the oxidation pathway of methyl group to CO2 and the membrane-bound cytochromes and heterodisulfide reductase, suggest that the strategies employed by methanogens to thrive in salt-saturating conditions are not limited to the classical methylotrophic pathway (Sorokin et al. 2017a). Another novel lineage distantly affiliated with “Methanonatronarchaeia” lacking key enzymes for methanogenesis but also affiliated with the SA1 Archaea was recently described based on single-cell genomics (Ngugi and Stingl 2018), suggesting the potential for metabolic diversity within members of the Candidate Division SA1 (Eder et al. 2002).

Nitrogen cycle

Dissimilatory nitrogen reduction represents one of the key metabolic features of DHABs. Evidence for different metabolic steps in the nitrogen cycle were recently detected in Thetis (Pachiadaki et al. 2014). Transcripts of enzymes involved in denitrification were found along the entire brine-seawater interface transition zone, while signatures of ammonia oxidation or anaerobic ammonium oxidation (anammox) were not found. Denitrification represents a major N₂ gas production pathway in Bannock and L’Atlante, accounting for up to 86% of the total N₂ production (Borin et al. 2013). Interestingly, despite high ammonia concentrations typically found in the brines relative to the deep seawaters above the DHABs (Borin et al. 2009; Ngugi et al. 2015), transcripts of nitrogenase enzymes involved in nitrogen fixation were overrepresented in the lower part of the brine-seawater interface in Thetis. The nitrogenase transcripts were attributed to archaeal clades, including ANME and methanogens, and it was speculated that nitrogen fixation and ammonia assimilation is performed in the lower Thetis BSI for the synthesis of osmoprotectants through the glutamine synthase and glutamate synthase pathways (Pachiadaki et al. 2014).

Ammonia-oxidizing archaea (AOA) dominate the brine-seawater interface microbial communities of several Red Sea DHABs. AOA were shown to be divergent from the mesopelagic and bathypelagic thaumarchaea (Ngugi et al. 2015; Zhang et al. 2016), confirming what was previously observed in Mediterranean DHABs (Daffonchio et al. 2006; Yakimov et al. 2007b; Borin et al. 2009). The dominant thaumarchaeal lineage is closer to epipelagic marine thaumarchaea (also called the Shallow Marine Group I clade; Francis et al. 2005), specifically to the genus Nitrosopumilus (Könneke et al. 2005). This genotype shows several features supporting a niche adaptation to the BSI environment (and the epipelagic zone), including acidic tuning of their membrane-bound proteins and the capacity to synthesize ectoine and hydroxyectoine, all of which are necessary for osmoregulation. The capacity to synthesize hydroxyectoine may confer tolerance to high temperatures such as those in Atlantis II Deep. The unique presence of key osmoregulatory mechanisms in thaumarchaea residing in saline ecosystems but absent in mesopelagic clades (Ngugi et al. 2015), supports the notion that salinity is a key factor determining the niche speciation of marine AOA (Erguder et al. 2009). Considering the extremely variable geochemistry of DHABs, it is likely also that other thaumarchaeal lineages exist and may have been shaped by their unique environments.

Nitrite oxidizers encompass the second group of microbes participating in nitrification. In the Red Sea, putative nitrite-oxidizing Nitrospina-like bacteria, named Ca. Nitromaritima RS were discovered in the brine-seawater interface of Atlantis II Deep. This group of ecologically important microorganisms constitutes up to on -third of the bacterial community and highly diverges from Nitrospina gracilis, one of the two cultured species of this widely distributed marine nitrite oxidizer (Ngugi et al. 2016). The osmoregulatory machinery of members of Ca. Nitromaritima includes high-affinity carriers for extracellular osmolytes and pathways for the biosynthesis of osmoprotectants.

The likelihood of anammox to occur in the Red Sea DHABs is supported by the detection of Anammox Planctomycetes sequences (Bougouffa et al. 2013; Abdallah et al. 2014; Guan et al. 2015). The ability of these organisms to thrive in seawater environments was previously demonstrated with the description of Candidatus Scalindua in many marine environments (van de Vossenberg et al. 2008; Woebken et al. 2008; Lüke et al. 2016). Apart from the challenge imposed by the high salinity, the environmental conditions of the DHABs are compatible with the requirements of the anammox metabolism, namely anoxic and ammonia-replete conditions, and the connection with an oxic interface providing nitrite. However, except for two studies, molecular data on the identity of resident anammox organisms and their ecological role in DHABs' nitrogen cycle remains poorly studied. The first molecular insights on the presence of Planctomycetes in DHABs were provided in a study of the highly sulfidic Urania deep hypersaline basin (Borin et al. 2009), and later the unequivocal detection of anammox activity and bacteria in the brine-seawater interface of the Bannock and L’Atalante basin (Borin et al. 2013). The detected anammox populations belonged to the Ca Scalindua clade (mostly Ca Scalindua brodae), which provides insights on the niches occupied by anammox including extremely saline sulfidic ecosystems. Recently, a planctomycetes genome phylogenetically distinct from 'Ca Scalindua brodae', named 'Ca Scalindua rubra', has been reconstructed from the metagenome of BSI samples collected from Discovery Deep in the Red Sea (Speth et al. 2017). Genomic analyses indicated that this scalindua species uses compatible solutes for osmoadaptation in contrast to other marine anammox bacteria that likely use a salt-in strategy.

Sulfur cycle

The presence of different sulfur chemical species and the detection of bacteria canonically involved in the sulfur cycle have suggested the importance of microbial sulfur cycling in different DHABs, especially the sulfidic ones. In Urania, one of the most sulfidic water bodies on earth, sulfate reduction was shown to be important in biogeochemistry of sulfur and the energetic metabolism of the whole microbial community (Borin et al. 2009). High sulfate reduction rates were measured at depths in the BSI where redox potential drastically decreased and the highest ATP concentration and bacterial cell numbers occurred. The 16S rRNA genes of the sulfate-reducing bacteria families Desulfobacteraceae and Desulfobulbaceae were abundantly detected along the Urania DHAB water column, from the BSI to the brine. Similarly, the presence of active dissimilatory sulfate reduction in the lower BSI was observed in Thetis, supported by a high abundance of rRNA transcripts from the family Desulfobacteraceae (Pachiadaki et al. 2014). Metabolic activity of sulfate-reducing Deltaproteobacteria was also found in the Kryos BSI following the detection of dsrAB gene transcripts distantly related to the ones of Desulfotignum balticum and the halophilic species Desulfosalsimonas propionicica (Yakimov et al. 2015).

Sulfate-reducing bacteria were also recently highlighted in the brines of Kebrit Deep, Nereus Deep, Erba Deep, Atlantis II Dee, and Discovery Deep where different communities were found (Guan et al. 2015). While in the first three DHABs dsrA sequences affiliated to known sulfate-reducing bacteria were detected, in the last two DHABs, characterized by higher temperatures, the dsrA sequences were instead affiliated to uncultured sulfate-reducing bacteria. This suggested that Atlantis II Deep and Discovery Deep harbour specific and novel sulfate-reducing communities (Guan et al. 2015).

Enzymes potentially responsible for three interconnected sulfide oxidation pathways were recorded from the BSI metagenome of Thetis, including i) the sulfur-oxidizing (SOX) multienzyme complex that catalyzes the oxidation of sulfide or thiosulfate to sulfate, with elemental sulfur and sulfite as intermediates, ii) a sulfide:quinone reductase (SQR), which oxidizes hydrogen sulfide to elemental sulfur and iii) polysulfide reductase (PSR), which reduces precipitated sulfur to sulfide (Friedrich et al. 2001). Accordingly, it was proposed that the interconnection of the second and the third pathways could allow microorganisms to yield maximal energy by switching from the complete oxidation of sulfur to sulfate (Sox system), to the production of elemental sulfur (sulfide:quinone reductase), which could then be reduced again by polysulfide reductase, avoiding S⁰ accumulation (Ferrer et al. 2012). Indeed, novel groups of strictly anaerobic sulfur-respiring haloarchaea with the capacity to reduce sulfur or thiosulfate using acetate, pyruvate, formate, or hydrogen as the sole electron donors have been isolated from various hypersaline lakes around the globe including Lake Medee from the Mediterranean Sea (Sorokin et al. 2016; Sorokin et al. 2017a). Thus, providing pioneer evidence of their potential role in biogeochemical sulfur cycling linked with anaerobic carbon mineralization in DHABs (Sorokin et al. 2017a).

Another important bacterial group for the sulfur cycle is the Epsilonproteobacteria encompassing several taxa capable of oxidizing sulfide and other sulfur chemical species. In the Urania water column it was observed that the number of sequences of the different epsilonproteobacterial taxa increased or decreased according to the change in salinity. Sulforovum and Helicobacteraceae abundances increased with salinity up to 18–20%, while Campylobacteraceae were dominating at lower salinities (Borin et al. 2009). Also, some species of the genus Arcobacter in the Campylobacteraceae family are involved in sulfur oxidation/reduction (Campbell et al. 2006; Sievert et al. 2007).

The involvement of representatives of the class Epsilonproteobacteria in other redox cycling processes has been considered, such as the cycling of manganese and iron (Campbell et al. 2006). Some works have highlighted the presence of Epsilonproteobacteria in the Bannock and the Urania basin in the Mediterranean Sea speculating on their potential involvement in a manganese cycle (Daffonchio et al. 2006; Borin et al. 2009). High concentrations of manganese have been observed in some DHABs such as Bannock (De Lange et al. 1990; Daffonchio et al. 2006) and other DHABs in the Mediterranean Sea (La Cono et al. 2011). Similarly, iron and manganese stratifications have been measured in the Gulf of Mexico (Trefry et al. 1984; van Cappellen et al. 1998) and Mediterranean Sea DHABs. For instance, in Bannock it has been shown that manganese exhibits a nonlinear slope, suggesting non-conservative behaviour and possible biologically mediated cycling (Daffonchio et al. 2006).

NEW RESEARCH DIRECTIONS

Despite the scientific effort of the last 20 years, many aspects of the microbial ecology and the metabolic traits of many dominant taxa specific to DHABs remain uncharacterized, highlighting the important challenges and the scientific questions yet to be answered.

The BSI emerges as the most metabolically active zone of DHABs. Due to the steep increases in salinity and density it is a particle trap for debris sinking through the water column, generating conditions that enhance microbial activity. Some studies demonstrated a precise stratification of the microbial communities along the chemocline with different prokaryote assemblages and networks resolved over depths of a few to tens of cm (Daffonchio et al. 2006; Yakimov et al. 2007b; Borin et al. 2009; Joye et al. 2009). However, most of these studies used fingerprinting and old-generation sequencing techniques of the small-subunit ribosomal RNA gene and few functional genes, thus exploring a limited amount of the existing genetic diversity. Attempts to circumvent this problem through application of whole-genome sequencing approaches (or metagenomics) have so far been done for a few brine pools only using bulk samples that disregard the micro-scale niches along the chemocline. Thus to date no studies exploiting high-throughput sequencing approaches (metagenomics and metatranscriptomics) at a fine spatial scale have been followed for a comprehensive in-depth understanding of the genetic and metabolic networks occurring in the BSI, even though the main conundrum of sampling approaches capable to maintain the integrity of the environmental gradients and the underlying microbial stratification has been addressed (Daffonchio et al. 2006; Joye et al. 2005).

Some studies to date have described the fauna in DHABs, but the linkage of predicted metabolic pathways to the actual intrinsic function of the symbiotic microbial communities remain circumstantial in the absence of transcriptomic or proteomic data. To reveal the functions occurring in DHABs, larger efforts exploiting meta-omics are desirable (Pachiadaki et al. 2014), complemented by the use of radiotracers and stable isotope techniques (Dumont and Murrell 2005; Borin et al. 2013; Yakimov et al. 2013).

The dramatic improvement of sequencing technology and of the downstream data analysis pipelines, and the decrease of sequencing costs allow efficient reconstructions of prokaryote genomes from metagenomic data and single cell genomes (Rinke et al. 2013). This has enlarged the breath of information achievable from small seawater or sediment samples, and such approaches are revealing novel metabolic functions of communities occurring in DHABs. For instance, several recent metagenomic and single-cell genomic studies revealed novel microorganisms, which include ammonia-oxidizing archaea (Ngugi et al. 2015), nitrite-oxidizing bacteria (Ngugi et al. 2016) and anaerobic ammonia-oxidizing bacteria (Speth et al. 2017), and unveiled the metabolisms of the uncultured MSBL1 archaeal clade (Mwirichia et al. 2016) and the KB1 bacterial clade (Nigro et al. 2016). We expect that the large differences in the environmental conditions and the biogeochemistry of the DHABs may have selected different genotypes and variants in the different brines, and we predict that novel genomes should have been evolved in these unique ecosystems.

The DHABs present a formidable set of multiple stress conditions that can potentially lead to the selection of unique phenotypes, metabolisms and enzymes. In this regard, DHABs represent an untapped genetic source of microbial extremophiles, extremozymes and extremolytes (Raddadi et al. 2015; Mapelli et al. 2016). Novel bacterial and archaeal taxa have been isolated from DHABs. These include, among others, extremely halophilic anaerobic archaea that metabolize elemental sulfur and acetate (Sorokin et al. 2016), a potential polysaccharide-degrading extremely halophilic archaeon (Antunes et al. 2008a; Werner et al. 2014), and a cell wall-less contractile bacterium representing a novel Haloplasmatales order (Antunes et al. 2008b). Besides being sources of novel enzymes, these novel prokaryotes can be directly used as catalyzers for biotransformations of fine chemicals. For example, bacterial isolates from the BSI of Mediterranean DHABs have been used for stereoselective hydrolysis of racemic esters that are used in the synthesis of prostaglandins (De Vitis et al. 2015). Versatile esterases equipped with different catalytic centres and the ability to function at prohibitive salt concentrations or hydrostatic pressures have been cloned and characterized from a metagenome of the Urania BSI (Ferrer et al. 2005).

Other applications of the microbial resources from the DHABs consider the microbial electrolysis technology for simultaneous treatment and energy generation from industrial high-temperature and high-saline wastewaters (Shehab et al. 2017). Moreover, the extreme conditions of the DHABs, together with the hydrocarbon inputs from the bottom sediments (Fig.4) (Borin et al. 2009), can serve as models for studying the processes that may lead to oil weathering in petroleum reservoirs (Bastin 1926; Bennett et al. 2013; Vigneron et al. 2017).

All these aspects should be addressed by future research efforts in order to (i) understand the functioning of the DHABs, (ii) assess the interaction of the microorganisms with the geochemistry of this systems, (iii) disentangle the structure-function relationships of microbes with their fine-scale environments, and (iv) elucidate the untapped biotechnological potential that DHAB microorganisms may have.

FUNDING

This work was supported by the Centre Competitive Funding (CCF) of the Red Sea Research Centre (RSRC) at the King Abdullah University of Science and Technology (KAUST).

Conflict of interest. None declared.