Since their discovery in the late 1970s, deep-sea hydrothermal systems have been considered as likely candidates for the origin and early evolution of life on Earth. However, while subsequent investigations have revealed a great diversity of modern deep-sea hydrothermal ecosystems, they have done little to shed light on the issues of the origin and early evolution of life, metabolism, cells, or communities. Phylogenetic, biochemical and geochemical clues all seem to point to the early evolution of hydrogenotrophic chemolithoautotrophy such as methanogenesis and sulfurreduction, thus pinpointing the availability of hydrogen as one of the key elements needed for the early evolution of earthly life. Hydrogen-driven, photosynthesis-independent communities are very rare on the contemporary Earth, however, being unambiguously found only in subsurface environments of H2-dominated hydrothermal systems. Such systems have been termed hyperthermophilic subsurface lithoautotrophic microbial ecosystems (HyperSLiMEs) (Takai et al., 2004; Nealson et al., 2005). The supply of abundant hydrogen and available inorganic carbon sources to power such communities is most likely coupled to hydrothermal serpentinization of ultramafic rocks and input of magmatic volatiles, both of which are related to specific geological settings. We propose here, on the basis of findings in the modern Earth and implications for the deep-sea hydrothermal systems in the Archean Earth, that “Ultramafics-Hydrothermalism-Hydrogenesis-HyperSLiME”, a linkage we refer to as Ultra H3, provided a suitable habitat for the early microbial ecosystem on the Archean Earth.
Introduction
Discovery of deep-sea hydrothermal systems along Mid-Ocean Ridges (MOR) toward the end of the 1970s provided an expanded view of the deep sea, which had been long believed to be a cold, dark, and barren environment (Corliss et al., 1979). Dense, abundant and diverse populations of animals were found associated with the active venting of superheated water, and it soon became apparent that these macrofaunal populations were strictly dependent on the primary production of symbiotic and free-living chemolithoautotrophic microorganisms, which obtain energy for carbon fixation, biosynthesis and other life activities from inorganic substances such as H2S, CO2, H2, CH4 and so on entrained by hydrothermal fluids from the Earth's interior (Rau, 1981; Rau and Hedges, 1979; Jannasch and Mottl, 1985). Subsequent to the initial studies of the MOR systems, many other types of hydrothermal fluid geochemistry and hydrothermal fluid-associated ecosystem have been explored in the subduction zones, so-called Volcanic Arcs (VA) and Backarc Basins (BAB), of the Western Pacific region (Takai et al., 2006).
The spectacular and unexpected properties of the deep-sea hydrothermal systems have attracted the interest of a diverse array of scientists, ranging from geophysicists (concerned with plate tectonics, mantle dynamics, magmatism, volcanology), to geochemists, microbiologists, and biologists. Not only are the systems a haven for extremophilic organisms, but they have inspired many ideas concerning their suitability as sites for the prebiotic chemistry required for the origin of life (Yanagawa and Kojima, 1985). Indeed, multiple lines of evidences from different fields (geology, chemistry, geochemistry, and molecular biology) have been used to argue that the deep-sea hydrothermal systems could have been the cradle of earthly life. For example, geologists have argued that the potentially most ancient microbial fossils are retrieved from a paleoenvironment that might be related to deep-sea hydrothermal systems in the early Archean (Rasmussen, 2000; Ueno et al., 2001 Ueno et al., 2004 Ueno et al., 2006). Chemists have defined a number of reactions potentially associated with prebiotic chemical evolution (synthesis of amino acids, nucleotides and hydrocarbon, and polymerization of these molecules) that occur readily under the physical and chemical conditions characteristic of the deep-sea hydrothermal systems (Imai et al., 1999; Yanagawa and Kojima, 1985).
In this paper, we discuss the possibility that irrespective of the origin of life, certain types of deep-sea hydrothermal systems might be havens for the types of communities that potentially represent the earliest microbial ecosystems. This discussion is somewhat associated with the last common ancestor concepts previously suggested as “progenote” (Woese, 1977; Woese et al., 1990), “cenancestor” (Pace, 1991), “the last universal common ancestor (LUCA)” (Forterre, 1996) or “commonote” (Yamagishi et al., 1991). Such thoughts derive for the most part from the interpretation of the molecular phylogeny of ribosomal RNA and other functional proteins, and lead to two contrasting notions of hyperthermophilic and mesophilic common ancestors, respectively. However, recent proteomic and protein engineering analyses have been used in support of the hypothesis for a hyperthermophilic common ancestor (Di Giulio, 2003; Iwabata et al., 2005; Miyazaki et al., 2001). Thus, it is reasonable for us to think that a certain high-temperature “window” occurred through which the ancient predecessors of today's surviving life forms must have passed. The Archean deep-sea hydrothermal environment could be the window.
While we focus here on the early evolution of life and the potential characteristics of the ancient microbial community, it is also reasonable to point out that the diversity of hydrothermal environments might be an important factor by which to predict the nature of the microbial community inhabiting those environments. This is particularly true with regard to metabolic abilities—each type of hydrothermal system might offer different metabolic opportunities and present different metabolic demands, leading to community formation of fundamentally different microbial metabolisms. With regard to the early microbial ecosystems in the Archean hydrothermal environments, one considers what kind of geological settings might host the ancient microbial ecosystems as predecessors of contemporary life. If one can come to some conclusion with regard to this question, then one can ask whether or not there are on the Earth today any environments that might serve as reasonable analogs to these ancient sites.
Here we propose a hypothesis that hydrogen-producing hydrothermal systems might favor the existence of the potentially earliest microbial communities. The fundamental ideas of this hypothesis are derived in part from recent results which suggest that there are metabolically dynamic communities existing in the subsurface that are geochemically isolated from any surface chemistry or biology, and that may serve as a good analogous model to the interrelationship between the Archean deep-sea hydrothermal environments and microbial ecosystems.
Hyperthermophilic subsurface lithoautotrophic microbial ecosystem (HyperSLiME) as a modern analog to the earliest microbial community
One approach to thinking about the hypothetical earliest microbial community is that of energy metabolism—what are the sources of energy available to different systems, and what kinds of communities or consortia might such systems have hosted? There are several different geochemical explanations for abundant energy sources in the Archean atmosphere and ocean (Habicht et al., 2002; Kasting & Ono, 2006; Kharecha et al., 2005; Tian et al., 2005). Almost certainly, however, the major oxidant available to the emerging life was CO2. Thus, early life forms were either fermentative, utilizing the limited amount of abiotically produced organic carbon and the inorganic fermentation sources of CO and S0 (no need for electron acceptors), and/or chemolithoautotrophic, utilizing the more abundant hydrogen (supposing that CO2 is the primary electron acceptor). Self-sustaining communities would surely have needed to utilize inorganic energy and carbon sources which would have been abundant and accessible substrates provided by the hydrothermal activities. The onset of such hyperthermophilic chemolithoautotrophy could excrete biosynthesized organic molecules that could be utilized by heterotrophic fermenting organisms, thus leading to what we imagine as the earliest communities or ecosystems. It is also an important part of our hypothesis that the nature of the hypothetical earliest ecosystem or community is not genomics-related, but metabolism-related: it must have been an assemblage of different metabolic abilities building to a self-sustaining community. To some extent it would be irrelevant whether these were the domain of Bacteria, Archaea, or Eucarya, despite the inability of contemporary Eucarya to grow chemolithoautotrophically.
The concept of the deep hot biosphere began with the speculations of Gold (1992), and the first such candidate for such a community called SLiME (subsurface lithoautotrophic, microbial ecosystem) was proposed by Stevens and McKinley for a subsurface microbial ecosystem in the Columbia River Basalt system (1995), a claim that was hotly contested by Anderson et al. 1998. Recently, Nealson et al. 2005 discussed the definition and criteria of H2-driven SLiME communities, and reviewed the evidence presented by several laboratories for and against the existence of SLiME communities. These authors also discussed the possibility that such SLiME communities might be adequate present-day analogs of ancient earthly ecosystems. To be brief, the criteria that must be met for a contemporary ecosystem are (i) that the energy driving the ecosystem should be of geological origin—this would include both the electron donors and electron acceptors (i.e., one thus imagines that a combination of H2 and CO2 should be the dominant and accessible redox couple for microbial chemolithoautotrophic metabolism), (ii) that neither electron donors nor electron acceptors produced by photosynthesis should be part of the ecosystem, (iii) that a community of appropriate organisms consistent with the proposed geologically driven metabolism should be present, and (iv) that this community should be shown to be active with regard to the metabolic hypothesis put forward for the ecosystem. An immediately apparent conclusion that seems to follow from these requirements is that specific subsurface environments are probably the only places in the modern Earth where it is operationally possible to escape the “pollution” of either the photosynthetically derived electron acceptors (O2, NO3 etc.), or the organic photosynthate (organic carbon) (Nealson et al., 2005).
A number of H2-driven SLiME communities have been recently proposed or considered: 1) subsurface groundwater environment in the Columbia River Basalts (Stevens and McKinley, 1995); 2) hot ground-water in the Lidy Hot Springs in Idaho (Chapelle et al., 2002); 3) a subseafloor environment of the Kairei hydrothermal field in the Central Indian Ridge (CIR) (Takai et al., 2004a); and, 4) a subseafloor environment of the Lost City hydrothermal field off the Mid- Atlantic Ridge (MAR) (Kelley et al., 2005). Of these, however, only the HyperSLiME found in the Kairei field meets the requirements laid down by Nealson et al. 2005 for a “true SLiME” community. This being said, however, some of the criteria were not met because measurements were not made, and there may, with some luck, be found in the future other “true-SLiME” communities.
Abundance of H2 in the hydrothermal activity is a key indicator for the presence of a HyperSLiME community
All the deep-sea hydrothermal systems generate abundant geochemical energy, as demonstrated by the model in Fig. 1, but only rarely is the output dominated by hydrogen. This is important with regard to the “true SLiME” criteria denoted above, as some energy sources such as H2S, while being excellent energy sources, are not able to form the biologically operative redox reaction with CO2 or other potential oxidants provided via hydrothermal fluids. Recent studies of different deep-sea hydrothermal fields have revealed that such hydrogen-rich systems are rather rare (Table 1). The highest levels of hydrogen were seen for the Rainbow field and the Logatchev field of the Mid-Atlantic Ridge (MAR) (Table 1), which are ultramafic rock-associated deep-sea hydrothermal systems (Charlou et al., 2002), and for the Kairei field (Table 1). The H2 concentration of hydrothermal carbonate fluids in the Lost City hydrothermal field is highly variable but can also represent extraordinarily high concentrations of H2 (Table 1).
Figure 1.
An original schematic model of a H2-driven, hyperthermophilic subsurface lithoautotrophic microbial ecosystem (Hyper-SLiME) sustained by H2 from degassing of magma and hydrothermal reaction with metal sulfides before Takai et al. 2004. As described in the text, for HyperSLiME to prosper requires much more abundant H2 than previously assumed. This leads to a new hypothesis, the “Ultramafics-Hydrothermalism-Hydrogenesis-HyperSLiME” linkage.
![i1342-8144-10-4-269-f01.jpg](ContentImages/Journals/jpal/10/4/prpsj.10.269/graphic/WebImages/i1342-8144-10-4-269-f01.jpg)
Table 1.
Comparison of reaction rocks, H2 concentration in hydrothermal fluids and presence of HyperSLiME or SLiME-like community among different deep-sea hydrothermal systems in the world.
![i1342-8144-10-4-269-t01.gif](ContentImages/Journals/jpal/10/4/prpsj.10.269/graphic/WebImages/i1342-8144-10-4-269-t01.gif)
Of the deep-sea hydrothermal fields in which the microbial communities have been investigated so far, strong evidence for a methanogenesis-dominated SLiME-like community was obtained only from the Kairei field (Kelley et al., 2005; Nakagawa et al., 2005a; 2005b; Shrenk et al., 2004; Takai and Horikoshi, 1999; Takai et al., 2001; Takai et al., 2003; Takai et al., 2004a; Takai et al., 2006). Kelley et al. 2005 have argued that the microbial communities in the carbonate structures of the Lost City field could be sustained by primary production via H2-based energy-yielding metabolism. However, all the data from geochemical analyses of the fluids, stable isotopic characterizations of total inorganic carbon (TIC), total organic carbon (TOC) and lipid biomarkers, and molecular phylogenetic characterizations of the microbial communities (Kelley et al., 2001, 2005; Shrenk et al., 2004), suggest that the microbial communities in the carbonate structures are energized by either (or both) anaerobic methane oxidation (CH4) or aceticlasticmethanogenesis (acetate) rather than by hydrogenotrophic methanogenesis (H2). For the carbon source, the Lost City microbial ecosystem perhaps uses inorganic carbon in the seawater, rather than from the earth's interior (Kelley et al., 2005). In addition, the molecular phylogenetic data suggest that the Archaea-dominated microbial communities in the Lost City carbonate structures are less thermophilic than HyperSLiME in the Kairei field. Thus, the Lost City microbial ecosystem does not represent a suitable analog for an early microbial ecosystem like Hyper-SLiME even though the occurrence of a H2-driven chemolithoautotrophic community in the Lost City field may be still possible.
Other than the Kairei field and Lost City field, a potential H2-driven SLiME-like community sustainedby methanogenesis is not suggested because of thelow levels of H2 (Table 1). As Lovley et al. 1982 and Lovley and Goodwin (1988) have pointed out, methanogenesis exhibits a very high range of kinetic constants for H2, and very high steady-state H2 concentrations when compared to other anaerobic terminal electron-accepting reactions such as sulfate reduction, Fe(III) reduction, and Mn(IV) or nitrate reduction for growth. This implies that if H2 concentration is low, hydrogenotrophic methanogenesis is prevented or outcompeted by other energy metabolisms. It thus seems likely that the occurrence and magnitude of HyperSLiME sustained by hyperthermophilic methanogens could be controlled by H2 concentration in hydrothermal fluids of deep-sea hydrothermal systems. This builds up portions of the UltraH3 linkage hypothesis, a “Hydrothermalism-Hydrogenesis-HyperSLiME” linkage.
Mechanisms and geological settings for hydrogenesis in hydrothermal systems
Several major mechanisms of generation of H2, i.e., hydrogenesis, in natural environments have been proposed: 1) degassing magma (Von Damm, 1995); 2) fault activity (Wakita et al., 1980); 3) radiolysis of water (Savary and Pagel, 1997; Vovk, 1982); 4) microbial fermentation (Jackson and McInerney, 2002); and, 5) serpentinization (Janecky and Seyfried, 1986; Coveney et al., 1987; Sleep, 2004). Although all these mechanisms may be to some extent involved in input of H2 to hydrothermal fluids, the extraordinary abundance of H2 in the hydrothermal fluids of the Rainbow, the Logatchev and the Lost City fields is explained by the serpentinization of oceanic peridotite, an ultramafic mantle rock (Table 1) (Charlou et al., 2002; Kelley et al., 2001).
Mantle peridotites and lower crustal plutonic rocks are significant components of the oceanic lithosphere. They are present in the shallow zones of ocean crust and are even exposed on the seafloor in many tectonic settings of the modern Earth (Früh-Green et al., 2004). The alteration of these rocks, particularly alteration of the mineral olivine, by the process of serpentinization involves the oxidation of Fe(II) to Fe(III), and results in the production of heat through exothermic reactions, generation of reduced, highly alkaline fluids, and hydrogenesis, and of course the formation of the mineral serpentine (Figure 2). In reality, serpentinization involves a series of continuous metastable reactions governed by local variations in the activities of elements such as Si, Mg, Fe, Ca, C and H+ in the fluids (Früh-Green et al., 2004). In addition, during the alteration of mantle peridotites and other ultramafic rocks, it is commonly believed that Fischer-Tropsch-type reactions have a great impact on the fluid chemistry associated with microbial energy metabolisms, that is, abiotic formation of CH4 and other hydrocarbons consuming H2 and CO2 with FeNi alloys, magnetite and chromite as catalysts (Fig. 2) (Yoshida et al., 1993; Berndt et al., 1996; Foustoukos and Seyfried, 2004). A number of experimental investigations of seawater-ultramafic mineral reactions at high temperatures have demonstrated the formation of abundant H2 and CH4, and even detectable ethane, propane and formate (Janecky and Seyfried, 1986; Berndt et al., 1996; Horita and Berndt, 1999). However, recent similar experiments have suggested that CH4 might be produced from H2 and reduced carbon sources (mainly graphite) in the minerals rather than aqueous CO2 (McCollom and Seewald, 2001). In either case of the carbon source, Fischer-Tropsch-type reactions are able to deprive the post-serpentinization fluids of H2 by abiotic methanogenesis, which may prevent the activity of a HyperSLiME or SLiME-like community based on microbial methanogenesis. In fact, the pore waters up to a depth of approx. 60 m below the seafloor of the serpentine mud volcano, South Chamorro Seamount, in the Mariana Forearc have no detectable H2 despite high concentration of CH4, and subsequent hydrocarbons are evidently produced from serpentinization-derived H2 through Fischer-Tropsch-type reactions (Salisbury et al., 2002; Mottle et al., 2003). At all depths of core samples, no apparent molecular signature and culture for methanogenic Archaea has been identified (Mottle et al., 2003; Takai et al., 2005).
Figure 2.
Simplified chemical reactions during hydrothermal alteration of ultramafic rocks such as serpentinization of olivine [(1)–(3)] and Fischer-Tropsch-type reactions [(4)–(X)].
![i1342-8144-10-4-269-f02.gif](ContentImages/Journals/jpal/10/4/prpsj.10.269/graphic/WebImages/i1342-8144-10-4-269-f02.gif)
The equilibrium of abundance of H2 and CH4 formed throughout serpentinization hydrogenesis and Fischer-Tropsch-type reactions is strongly influenced by physical and chemical conditions where the whole reactions proceed (Früh-Green et al., 2004). However, all the previous experimental investigations of serpentinization of ultramafics have demonstrated a higher yield of H2 than of CH4 (Janecky and Seyfried, 1986; Berndt et al., 1996; Horita and Berndt, 1999; McCollom and Seewald, 2001). In addition, all the vent and seep fluids associated with serpentinization of oceanic peridotites and Zambales ophiolite (other than the pore waters in the Mariana Forearc Seamounts) represent the predominance of H2 or similar amounts of H2 and CH4 (Abrajano et al., 1988; Charlou et al., 2002; Kelley et al., 2001; Mottle, 1992; Mottle et al., 2003). Thus, the extraordinarily high concentration of H2 in the hydrothermal fluids could be clearly attributed to hydrogenesis of hydrothermal serpentinization from ultramafics. This implies the rationale of portions of the UltraH3 linkage hypothesis, the “Ultramafics-Hydrothermalism-Hydrogenesis” linkage.
This “Ultramafics-Hydrothermalism-Hydrogenesis” linkage is unequivocal and ubiquitously occurs in the modern Earth, and seems to provide an ideal habitat for a H2-driven HyperSLiME or SLiME-like community, but the situation is complicated by the formation of extremely alkaline fluids during the hydrothermal serpentinization (the exact mechanism of the chemistry is still unclear). For example, the pore water samples from the serpentine mud volcanoes of the South Chamorro and the Conical Seamounts had an average pH of 12.5 (Mottle, 1992; Mottle et al., 2003), and the hydrothermal fluids in the Lost City were measured at pH 10–11 (Kelley et al., 2005), with even higher pH values of the end member fluids (Charlou et al., unpublished). These extremely high pH conditions approach the highest known pH at which an organism has been grown (Alkaliphilus transvaalensis was grown at pH 12.4 (Takai et al., 2001)), and surpass the usual limit for growth of pH ~11 (Wiegel, 2002). In addition, under such high pH conditions, most of the inorganic carbon is present as CO32− and thus unavailable for microbial energy and carbon metabolisms. Indeed, Kelley et al. 2005 suggested that TIC in the Lost City hydrothermal fluids was all CO32− and the microbial communities might utilize CO2 or HCO3− from the ambient seawater. Thus, the extreme pH conditions associated with the “Ultramafics-Hydrothermalism-Hydrogenesis” linkage potentially hinder the occurrence of a H2-driven SLiME-like community. However, this seems to be the case for both the Lost City hydrothermal field and the Mariana Forearc serpentine seamounts (Kelley et al., 2005; Schrenk et al., 2004; Takai et al., 2005).
In contrast, the H2-abundant hydrothermal fluids of the Rainbow and the Logatchev fields typically display a pH around 3. Such acidic hydrothermal fluids are common in deep-sea hydrothermal systems and are explained by input of magmatic volatiles, whether they are provided directly from magma itself or subsequently from the inclusion of magmatic volatiles in the heat-source rocks (Alt, 1995). In these deep-sea hydrothermal systems, therefore, the magmatic input might play a significant role in preparing the moderate pH conditions and the addition of inorganic carbon sources from the interior of the Earth. With the magmatic input, the microbial ecosystems in the ultramafic rock-associated deep-sea hydrothermal systems such as the Rainbow and the Logatchev fields might be energized by H2 from hydrothermal serpentinization of ultramafics. In such geological settings, the complete linkage of UltraH3 (Ultramafics-Hydrothermalisms-Hydrogenesis-HyperSLiME) could be operative.
Searching for an UltraH3 site favorable to life on the modern Earth
So, we have drawn a picture in which UltraH3 creates a problem for itself via the formation of high pH conditions, and have proposed a solution for this dilemma, namely, that the problem might be solved by the mixing of low pH waters from magmatic systems, thus providing abundant H2 at pH levels consistent with comfortable inorganic carbon chemistry. One such system may be the Rainbow hydrothermal field, which is located on a tilted ultramafic ridge in a non-transform offset of the AMAR segment on the Mid-Atlantic Ridge. The field, at water depths between 2270 and 2320 m, is based on a clear tectonic control by a network of faults generated both by the ridge and the non-transform system (Fouquet et al., 1997; Fouquet et al., 1998). The hydrothermal fluid chemistry reveals the considerable contribution of the hydrothermal serpentinization of peridotite (Charlou et al., 2002), but the vigorous black smoker ventings and the rare earth elements (REE) composition of the fluids strongly suggests the considerable contribution of deep magmatic intrusion and the gabbroic dykes to the hydrothermal reaction (Fouquet, personal communication). It seems likely that the hydrothermal activity of the Rainbow field stands on the deep subsurface structure with a combination of host peridotite and intrusive gabbroic magma: potentially an ideal setting for the UltraH3 linkage. However, the search for a HyperSLiME-like community has been not yet undertaken in the Rainbow hydrothermal field or other ultramifics-associated hydrothermal systems in the Mid-Atlantic Ridge.
The Kairei hydrothermal field is located in the first segment of the Central Indian Ridge (CIR-S1), which was the first deep-sea hydrothermal field discovered in the Indian Ocean (Hashimoto et al., 2001). It is situated at the eastern axial valley wall very close to the inside corner of the ridge-transform intersection (RTI) between the first and second segments of the CIR (Fig. 3). The local bathymetric topography and the dive surveys using DSRV Shinkai 6500 revealed that the hydrothermal activities of the Kairei field were distributed along the lava flow extending 100 m east to west and 40 m north to south. Based on the geochemical and stable isotopic characterizations of the hydrothermal fluids and the microbiological exploration of the subseafloor microbial communities, the extremely high concentration of H2 and significantly 13C-depleted CH4 in the fluids, the possible occurrence of HyperSLiME was demonstrated, thus establishing the elements of the UltraH3 linkage (Hydrothermalism-Hydrogenesis-HyperSLiME) in the Kairei hydrothermal field (Takai et al., 2004). What remains to be established is the contribution of ultramafic rocks to the hydrothermal reactions. This is strongly suggested by the extraordinary amount of H2 in the fluids, but the association of ultramafic rocks with the hydrothermal activity has not been directly identified according to geophysical and petrological criteria.
Figure 3.
Bathymetric map of the area including the Kairei hydrothermal field and vicinity (modified from Kitazawa and Nakanishi, personal communication). Light blue triangle indicates the Kairei hydrothermal field. Dotted pink curve indicates Ocean Core Complexes (OCC)-like structure found.
![i1342-8144-10-4-269-f03.jpg](ContentImages/Journals/jpal/10/4/prpsj.10.269/graphic/WebImages/i1342-8144-10-4-269-f03.jpg)
Recent bathymetric and geophysical surveys of the southern CIR around the ridge-ridge-ridge type triple junction, the Rodriguez Triple Junction (RTJ), to which the Kairei field is closely situated, suggest that the ultramafic rocks should be relevant to the hydrothermal activity in the Kairei field. As a whole, the CIR has been characterized as an intermediaterate spreading ridge, while the southern region of the CIR is based on variable morphological structures along the ridge axis, indicating spatial and temporal variation of magma supply (Honsho et al., 1996). For instance, the inflated axial high at the southern end of the third segment of CIR (CIR-S3) seems to be a morphological signature observed in a fast-spreading ridge, but the deep axial valley of the first segment (CIR-S1) is similar to the axial valleys in typical slow-spreading ridges like the MAR (Fig. 3) (Briais, 1995). In such slow-spreading ridges with insufficient magma supplies, some significant portions of mantle peridotite reside within the shallow oceanic crust (Cannat, 1996) and very remarkable massifs characterized by corrugation perpendicular to the ridge axis, known as “megamullions” or “Oceanic Core Complexes (OCC),” are frequently found (Cann et al., 1997). This OCC-like structure is found on the western flank of the CIR-S1, at a position opposite the Kairei hydrothermal field (Fig. 3). The relevant area shows not only the typical morphological signature of OCC but also a strong positive gravity anomaly (i.e., Bouguer anomaly), indicating the existence of dense materials beneath the massif (Wakabayashi, 2003). From a volcanological aspect, the very short length of the CIR-S1 (< 50 km) might be consistent with a smaller amount of magma supply and a rapid cooling of magma by adjacent older lithosphere known as a transform effect (Klein and Lungmuir, 1987). In addition, the plate reconstruction approach of the RTJ and the southern CIR suggests that the plasticity of the RTJ and the very short lifetime of the CIR-S1 have been associated with an unstable setting of the RTJ (Honsho et al., 1996; Mendel et al., 2000). All these observations strongly suggest the geological setting of the CIR-S1 and the Kairei hydrothermal field is in fact associated with ultramafic rocks in the form of mantle peridotites, as inferred from the extraordinary H2 concentration in the hydrothermal fluids. A series of geological, geophysical, and geochemical expeditions were conducted in early 2006 and the geological setting of the CIR-S1 and the Kairei hydrothermal field is now under extensive investigation.
UltraH3 linkage in the Archean Earth and implications for the early microbial ecosystem
The establishment of the UltraH3 linkage for the hydrothermal systems along slow-spreading MORs such as the Mid-Atlantic Ridge and the southern region of the Central Indian Ridge raises the question of whether similar UlraH3 linkages might have been continuously present over the long history of the Earth, and if so, whether they were important for the prosperity of early microbial ecosystems. That is, could similar habitats have existed on the early Archean Earth ~3:5 billion years ago?
What might the Archean Earth have looked like with regard to ultramafic-associated hydrothermal systems and UltraH3 habitats? Key factors are chemistry and temperature of the Archean mantle, which presumably controlled the structure and chemical composition of the Archean ocean crust. It is generally assumed that ultramafic rocks (as well as mafic rocks) were more abundantly present in the Archean ocean crust than in the present day. Particularly, ultramafic lava flows known as komatiites have been discovered from almost all Archean cratons (Condie, 1994). Among the Archean supracrustal rocks termed greenstones, the volume fraction of komatiite and basalt is 40~80% (Condie, 1994), and some of the komatiite/basalt successions contain up to 25% of the komatiite component (de Wit and Hart, 1993). On the one hand, although dunites and harzburgites comparable to modern abyssal peridotite have been discovered in the early Archean gneiss complex (Friend et al., 2002), mantle peridotite exposed onto the Archean ocean floor is assumed to be minor. The hotter Archean mantle probably conducted more extensive and active magmatism in the Archean MORs and formed a much thicker oceanic crust than in the modern Earth (McKenzie and Weiss, 1975; McKenzie and Bickle, 1988; Ohta et al., 1996). Under such conditions, tectonically controlled exposure of mantle peridotite to the ocean floor would be very unlikely.
Komatiite is a distinctive volcanic rock of the Archean and is common in Archean greenstone successions, whereas it is unusual in the Proterozoic and quite rare in the Phanerozoic (Condie, 1997). The Archean komatiite and basalt flows are commonly pillowed (Condie, 2005), indicative of subaqueous eruption onto the seafloor. Komatiitic volcanism is considered to be involved in mantle plume activities, because of the very high estimated eruptive temperature of komatiite (Arndt, 1994). In addition, the voluminous oceanic eruptions of komatiite and associated basaltic rocks are commonly referred to as an oceanic plateau (Kusky and Kidd, 1992; Condie, 1994; Barley and Picard, 1999). Thus, komatiite might have been predominantly and ubiquitously distributed in the Archean ocean floor in the manner of the present day's oceanic plateaus.
The Archean submarine komatiites and basalts are often altered (e.g., albitized, silicified, carbonatized) by hydrothermal fluids in a similar manner to that seen in the basalts of the modern MORs and VAs (Barley, 1993; de Wit and Hart, 1993; Kato and Nakamura, 2003). The degree of hydrothermal alteration is significantly greater in the Archean oceanic komatiite/basalt than in the modern counterparts, probably due to greater heat and fluid fluxes in the Archean seafloor hydrothermal systems (de Wit et al., 1987). Most of the Archaean komatiites were serpentinized by hydrothermal alteration immediately after eruption (de Wit et al., 1987; Rollinson, 1999; Sproule et al., 2002), and magnetite crystallization associated with serpentinization is also found in hydrothermally altered komatiites (Yoshihara and Hamano, 2004). Such petrological evidence strongly suggests that the seafloor hydrothermal serpentinization of komatiite which may have dominated in the Archaean ocean floor could have been linked with abundant hydrogenesis. The komatiite-derived H2, with CO2 provided from the komatiite volcanism and the CO2-rich Archean ocean (Grotzinger and Kasting, 1993), could have prepared widespread habitats suitable for early microbial ecosystems sustained by methanogenic primary production to prosper. In the Archean Earth, therefore, the UltraH3 linkage should be the komatiite-hosted version and it could have been much more abundant than the peridotite-hosted version in the modern Earth.
Concluding remarks
Here we have put forward arguments supporting an idea that we call the UltraH3 (Ultramafics-Hydrothermalism-Hydrogenesis-HyperSLiME) linkage hypothesis. This hypothesis claims: 1) that the ultramafic rock-associated hydrothermal system was the most plausible place for the existence of one of the earliest microbial ecosystems in the Archean Earth: 2) that H2 generated through hydrothermal serpentinization could energetically support the activity of the early microbial community; and 3) that the H2- abundant ultramafic-associated hydrothermal system has continuously harbored microbial ecosystems over the history of the Earth. The UltraH3 linkage is now being proven in the modern proxies situated in the MORs such as the Rainbow and the Kairei hydrothermal fields. However, direct exploration of the Archean version of the UltraH3 linkage based on komatiite is nearly impossible because of the lack of active komatiite volcanism in the modern Earth and significant differences in the ocean and atmosphere of the modern and Archean Earth. Indeed, it has been demonstrated that the mode of hydrothermal alteration in the Archean Earth should have been drastically affected by physical and chemical conditions of the Archean ocean and atmosphere (Kitajima et al., 2001; Nakamura and Kato, 2004). Experimental work under conditions relevant to hydrothermal fluid-rock interactions and Archean conditions will be required before the physical and chemical conditions can be constrained.
Figure 4.
Schematic models of mantle convections, mantle plumes, plate tectonics and volcanisms hosting UltraH3 linkage in the Archean and modern Earth. In the Archean Earth, a hotter mantle might have induced smaller scales of and faster mantle convections than in the modern Earth, probably resulting in much more activity on MORs and a smaller scale of and more separated oceanic plates. Simultaneously, komatiite-dominated volcanism on ocean plateaus might have been abundant in the Archean Earth.
![i1342-8144-10-4-269-f04.jpg](ContentImages/Journals/jpal/10/4/prpsj.10.269/graphic/WebImages/i1342-8144-10-4-269-f04.jpg)
Acknowledgments
We greatly appreciate Y. Fouquet and J. Charlou, IFREMER, France, for providing their unpublished data and instructive suggestions and discussions to this article. We also thank M. Kitazawa and M. Nakanishi for permitting us to use the unpublished data acquired by the KR00-06 research cruise.