Early bioenergetic evolution

Life is the harnessing of chemical energy in such a way that the energy-harnessing device makes a copy of itself. This paper outlines an energetically feasible path from a particular inorganic setting for the origin of life to the first free-living cells. The sources of energy available to early organic synthesis, early evolving systems and early cells stand in the foreground, as do the possible mechanisms of their conversion into harnessable chemical energy for synthetic reactions. With regard to the possible temporal sequence of events, we focus on: (i) alkaline hydrothermal vents as the far-from-equilibrium setting, (ii) the Wood–Ljungdahl (acetyl-CoA) pathway as the route that could have underpinned carbon assimilation for these processes, (iii) biochemical divergence, within the naturally formed inorganic compartments at a hydrothermal mound, of geochemically confined replicating entities with a complexity below that of free-living prokaryotes, and (iv) acetogenesis and methanogenesis as the ancestral forms of carbon and energy metabolism in the first free-living ancestors of the eubacteria and archaebacteria, respectively. In terms of the main evolutionary transitions in early bioenergetic evolution, we focus on: (i) thioester-dependent substrate-level phosphorylations, (ii) harnessing of naturally existing proton gradients at the vent–ocean interface via the ATP synthase, (iii) harnessing of Na+ gradients generated by H+/Na+ antiporters, (iv) flavin-based bifurcation-dependent gradient generation, and finally (v) quinone-based (and Q-cycle-dependent) proton gradient generation. Of those five transitions, the first four are posited to have taken place at the vent.

Ultimately, all of these bioenergetic processes depend, even today, upon CO2 reduction with low-potential ferredoxin (Fd), generated either chemosynthetically or photosynthetically, suggesting a reaction of the type ‘reduced iron → reduced carbon’ at the beginning of bioenergetic evolution.

Keywords: transition metals, acetogens, methanogens, sulfate reducers, origin of life, hydrothermal vents

Life is a net exergonic chemical reaction, it releases energy to go forward. Many settings have been proposed as the site for the chemical synthesis for life's building blocks [1] and the ignition of the continuous chemical reaction that includes us as its descendants. Such proposed settings include oceans or ponds of organic soup stocked either by ultraviolet light-dependent organic synthesis [2] or by organics delivered from space [3]; borate evaporites [4]; terrestrial zinc-rich hydrothermal settings [5]; ice [6] or pumice [7]—to name but a few. However, since their discovery, submarine hydrothermal vents stand out among the possible environments for life's origin, holding particular promise for understanding the transition from geochemistry to biochemistry [8–12]. Submarine hydrothermal vents harbour highly reactive chemical environments with far-from-equilibrium conditions, being rich in gradients of redox, pH and temperature. Furthermore, hydrothermal vents generate spontaneously, forming systems of inorganic microcompartments [9,13] that, in an origin-of-life context, could readily serve to concentrate the chemical reaction products and substrates that form there, at their site of their synthesis. The far-from-equilibrium condition is important because it harbours the potential for exergonic chemical reactions [14]. Concentration is important as a prerequisite for autocatalysis and for self-regulation, defining characteristics of living systems [15].

A number of submarine hydrothermal vents have been studied. They differ in the specific temperature and the chemical compositions of their effluent fluid [10,16]. Most are located at the seafloor spreading zone, that is, directly above magma chambers. As such, their effluent comes directly into contact with magma and emerges at the vent–ocean interface with a temperature often exceeding 300°C. Such ‘black smokers’ are not only very hot, but they are also short-lived, lasting in the order of decades. In contrast, off-ridge vents of the kind exemplified by Lost City [13,17] are situated several kilometres away from the spreading zone; hence the water circulating through them makes no contact with magma and emerges at a temperature of around 70–90°C. Unlike vent systems that reside directly on the spreading zone, off-ridge vents can be stably active over geological time; Lost City has been active for about 120 000 years [18]. The chemical disequilibrium at vents arises from the contact between their reducing, H2-containing effluent with CO2-containing ocean water, which is more oxidized. Hydrothermal vents harbour up to 50 mM H2 [10]. At Lost City, the H2 concentration of the alkaline effluent (approx. pH 9–10) is about 10 mM [19].

Hydrogen is a reducing agent: an electron donor. Given a suitable acceptor, of which there are many [20], hydrogen is a source of energy. The H2 in hydrothermal vents comes from a geochemical process called serpentinization [21–25]. Serpentinization takes place because most of the suboceanic crust consists of iron–magnesium silicates like olivine, in which the redox state of the iron is Fe2+. Seawater penetrates the ocean floor crust, reaching depths on the order of 3–5 km below the surface. There at high pressure and temperatures of around 150°C, water oxidizes the Fe2+ to Fe3+, the water being reduced to H2 with the oxygen in water being retained in the rock as iron oxides. Bach et al. [21] write the serpentinization reaction as in equation (2.1) whereas Sleep et al. [22] represent the main redox reaction as in equation (2.2).

Mg2.85Fe0.15Si2O5(OH)4 is serpentinite, the mineral from which the process is named, and 4Fe3O4 is magnetite, the mineral containing the oxidized iron. During serpentinization, a cubic metre of olivine eventually yields about 500 mol of H2 [24]. The heated water convects back up to the ocean floor, carrying H2 and other reaction products with it. The production of H2 via serpentinization has probably been taking place since there was liquid water on the Earth [22], and it continues today because, relative to the amount of water on the Earth, the amount of Fe2+ in the crust is inexhaustible. The Mg(OH)2 in equation (2.1) produces alkaline hydrothermal fluid in agreement with the pH observed for Lost City effluent [17].

Lost City effluent also contains about 1 mM methane of abiogenic origin [19,26], along with about 100 µM formate of abiogenic origin [27] and traces of longer hydrocarbons up to pentane [19]. This indicates that serpentinization in the crust at Lost City reduces CO2 to an organic product spectrum. The magma-hosted black smoker systems are less conducive to organic synthesis in the crust, because, above approximately 250°C, carbon in contact with water becomes CO2 [28,29], and magma has a temperature of around 1200°C. If Lost City harbours organic synthesis today, then organic synthesis in similar serpentinizing systems should also have been possible 4 billion years ago. While the composition of ocean water has changed since then [30], the composition of the seafloor crust, where serpentinization reactions take place, has remained roughly constant, as indicated by the zircon compositions of Early (Hadean) mantle-derived melts [31].

The chemical disequilibrium at off-ridge hydrothermal vents on the early Earth was, in all likelihood, much greater than that today. This is because there was much more CO2 in the atmosphere, perhaps up to 1000 times more [32,33], meaning that there would have been much higher CO2 concentrations in the ocean. Hence when H2 in the convective currents of serpentinizing systems made contact with marine CO2 (stemming inter alia from volcanos) there was disequilibrium and potential for organic synthesis.

How much potential do vents harbour for the synthesis of what kind of organic products? Shock and co-workers have studied the question of organic synthesis at hydrothermal vents from the thermodynamic standpoint, and what they find is encouraging from an origin-of-life perspective (reviewed in [14]). They find that CO2 reduction and organic synthesis is thermodynamically favoured—exergonic—when reduced vent fluid mixes with more oxidized ocean water at the vent–ocean interface, and hence can take place under the conditions presented by hydrothermal vents. This is true for the synthesis of carboxylic acids, alcohols and ketones [28,29], amino acids and proteins [34] and total microbial cell mass [35,36].

The nature and proportion of organic products that are thermodynamically most favoured depend upon the specific chemical conditions, for example, H2 availability, temperature and the redox state of the environment [14,37,38]. Under the conditions found at Lost City, for example, the overall synthesis of microbial cell mass from H2, CO2 and NH3 is exergonic in the temperature range 50–125°C [36].

Such findings are broadly consistent with the findings of Thauer et al. [39] that under conditions relevant for microbes, in the reaction of H2 with CO2, the equilibrium lies on the side of reduced carbon compounds, which is why acetogens can grow from the reaction

with ΔG°′ = –131 kJ mol−1 [40] as their sole source of energy, respectively. They harness (conserve) chemical energy from those reactions to push the life process forward. For a prokaryote, the main energetic cost in the life process is amino acid and protein synthesis, which consumes about 75 per cent of the cell's ATP budget, with RNA monomer and polymer synthesis consuming only about 12 per cent [41].

Because hydrothermal vents present conditions where the synthesis of proteins from H2, CO2 and NH3 would be an exergonic process [14,34], they truly appear to be special among the various settings that have been considered for the origin of life. Hydrothermal vents are particularly rich in chemical and thermodynamic similarities to the core energy releasing reactions of modern acetogens and methanogens [42,43], lineages that over 40 years ago—before the discovery of either hydrothermal vents or archaebacteria—were proposed to be the most ancient microbes, because they are anaerobic chemoautotrophs that live from the H2–CO2 redox couple [44].

The far-from-equilibrium conditions and favourable thermodynamic setting of hydrothermal vents do not indicate which chemical syntheses will occur, merely which are energetically possible [14]. The nature of catalysts present can also influence the kinds of products that are formed [45], depending on whether the reaction is thermodynamically controlled (the most stable products accumulate) or kinetically controlled (the most rapidly synthesized products accumulate). Here, we revisit a model for the origin of life (figure 1) as set forth previously in these pages while incorporating newer findings.

One can assume that off-ridge vents such as Lost City were more prevalent on the early Earth than today [11,22,23]. But the chemistry at a Hadean vent–ocean interface would have been slightly different, with abundant Fe2+ [30] and far more CO2 [32,33] in the ocean. Given that, and given the presence of sufficient sulfide in the hydrothermal fluid to steadily precipitate transition metal sulfides (FeS, NiS and the like) upon mixing of effluent and ocean water at the vent interface [9], the model posits that hydrothermal flux deposited a mound at the vent, consisting crucially of transition metal sulfides, among other possible minerals. The premise that a Hadean mound could have existed is underpinned by the report of fossilized metal sulfide hydrothermal mounds 360 Myr of age—the rocks that gave rise to this model in the first place [47]—found in Tynach, Ireland by Russell and co-workers [48]. These mounds reveal elaborate systems of naturally forming, inorganically walled microcompartments (on the order of 1 µm–1 mm in diameter). The walls of these compartments unite two very important properties for early chemical evolution: catalysis and compartmentation.

The catalysis was provided by the transition metals and transition metal sulfides themselves of the hydrothermal mound. That is, the microcompartments have inorganic walls made of essentially the same catalysts that are commonly found in redox-active enzymes of modern microbes [9,49–51] and commonly used in organometallic catalysis in the chemical industry [52,53]. This is an important similarity to modern anaerobic autotrophs, whose enzymes of core carbon and energy pathways are replete with FeS, FeNiS and MoS (WS) centres, FeS and NiS usually being incorporated into proteins via cysteine residues [54–58], Mo and W usually being complexed by thiols of the molybdopterin cofactor MoCo [59]. Importantly, such FeS and FeNiS centres are particularly common in H2-oxidizing and CO2-reducing proteins of anaerobic prokaryotes [60]. Because the nature of the catalysts at the interface of gases (H2, CO2, N2, H2S, SO2) and organic matter that (i) we posit to be present at the mound and (ii) occur in acetogens and methanogens is very similar, some congruence between the kinds of products obtained in the two systems is not an unreasonable proposition [45].

Interfacing with the CO2-rich seawater at the vent was highly reduced, H2-containing, vent effluent that probably contained some reduced carbon compounds itself, as Lost City does today [19,26,27], stemming from serpentinization reactions deep within the crust. How reducing might the effluent in an ancient serpentinizing system have been? A clue is provided by the mineral awaruite (Ni3Fe), an alloy of metallic iron and nickel, Fe±0 and Ni±0, that is quite commonly found in serpentinizing systems and that is produced from Fe2+ and Ni2+ minerals via reduction to the raw metals in situ by H2 [25,61]. Awaruite forms only when sulfide concentrations are very low, but that is not the point, nor is the point that Fe±0 readily reduces CO2 to acetate and formate under hydrothermal conditions [62]. The point is that in the temperature range 150–300°C, activities of H2 exceeding 200 mM are needed for awaruite to form [25]. To stress the point, 200 mM H2 represents a very large amount of hydrogen and a very reducing environment. The presence of awaruite in serpentinizing systems thus clearly indicates that at some point during their lifespan, serpentinizing systems regularly generate extremely reducing conditions on the order of greater than or equal to 200 mM H2, which would be conducive to organic synthesis, as Klein & Bach [25] point out. So while conditions at a Hadean-serpentinizing system cannot be pinpointed, we can say that serpentinizing systems do, in principle, have the reducing power at hand that would be needed for making organic compounds out of CO2 (or even elemental iron and nickel out of their salts).

For the building blocks of life to form, nitrogen is also essential. Where does reduced nitrogen come from? Some modern vent systems contain up to 1 mM NH3 [10], although it is not known that serpentinization is responsible for N2 reduction. However, simulated hydrothermal conditions of 300°C and high pressure readily convert nitrite into NH3 [63], and under very mild conditions, FeS minerals can reduce N2 to NH3 at yields approaching 0.1 per cent [64], whereby native Fe±0 is also an effective catalyst of N2 reduction under hydrothermal conditions [65], recalling once again awaruite as an indicator of serpentinization [25]. So it should be safe to assume that reduced nitrogen would have been available in a Hadean-serpentinizing hydrothermal system. Phosphate is more soluble at pH 6 than at pH 8 and was not likely to have been globally abundant in the early oceans [11,30], although there was surely considerable local phosphate input from volcanos [66,67]. Phosphate was possibly concentrated at hydrothermal vents, however, via the mineral brucite—Mg(OH)2—driven precipitation processes that Holm et al. [68] suggest to have involved sepentinization.

The foregoing would provide a mixture of H2, CO2 and NH3 similar to that assumed by Amend et al. [14] for synthesis of chemical constituents of cells, in an environment laden with transition metal catalysts, with abundant hydrothermal HS–. That is a starting point for the synthesis of compounds with high-energy bonds. Under mild laboratory conditions, Heinen & Lauwers [69] reduced CO2 to a spectrum of sulfur containing organic products, including methylsulfide (CH3SH), using FeS as the catalyst, and Huber & Wächtershäuser [70] reacted CH3SH and CO under mild conditions (ambient pressure, 100°C), using FeS, NiS or Ni2+ as catalysts, obtaining up to 40 mol% of the thioester methylthioacetate (or acetyl methylsufide, CH3COSCH3). Thioesters are energy-rich compounds because the thioester bond has a large free energy of hydrolysis, ΔG°′ = –32 kJ mol−1 [71]. Starting from such simple chemicals, a high spontaneous yield of a compound with an energy-rich thioester bond might seem surprising, but the result is expected from thermodynamics because even to the level of the energy-rich thioester, the reaction

is highly exergonic with ΔG°′ = –59 kJ mol−1 [72], whereby the pathway becomes endergonic at low H2 partial pressures, with ΔG°′ = +29 kJ mol−1 at approximately 10 Pa H2 [73]. Equation (3.1) entails the thiol group of coenzyme A (CoASH) rather than CH3SH, and it is the line reaction of the acetyl-CoA or Wood–Ljungdahl [74,75] pathway of microbial CO2 fixation. Such favourable energetics led Shock et al. [29, p. 73] to conclude that organisms using the acetyl-CoA pathway (acetogens, methanogens and many sulfate reducers) get ‘a free lunch that they are paid to eat’. Provided that transition metals in a hydrothermal vent could catalyse similar reactions as in the laboratory [51], then one would have, in principle, a route for sustained geochemical synthesis of acetyl thioesters (among many other products). That would constitute a critical/crucial link between geochemistry and biochemistry, because acetyl thioesters (acetyl-CoA) are the most central compounds in all of metabolism, as biochemical pathway maps will attest; and apart from the Calvin cycle, which is obviously a late evolutionary invention [76], acetyl-CoA is the direct product of all known CO2 fixation pathways [73]. The centrality of acetyl thioesters in modern metabolism is, we contend, a relic of their role in the chemical events that gave rise to metabolism.

Acetyl thioesters are furthermore the most probable entry point of phosphate into metabolism [42,77,78]. In a plethora of microbes, acetyl-CoA is enzymatically cleaved via phosphorolysis to yield acetyl phosphate [79,80], which, with its high free energy of hydrolysis (ΔG°′) of –43 kJ mol−1, can phosphorylate any number of substrates, including ADP to generate ATP, with a free energy of hydrolysis (ΔG°′) of –31 kJ mol−1. Weber [81] reported the non-enzymatic synthesis of acyl phosphates from a thioester in the presence of phosphate during the synthesis of pyrophosphate from thioesters, although the acyl phosphate itself was not isolated. A high free energy of hydrolysis of the acyl phosphate bond and simple phosphorylytic synthesis from thoiesters make acyl phosphates good candidates for the first universal currency of high-energy bonds [42], the precursors of ATP.

Provided there was a sustained geochemical source of chemically accessible methyl groups (e.g. methyl sufide), the first reactions relevant to the origin of carbon and energy metabolism might have been very similar to those catalysed by the bifunctional enzyme carbon monoxide dehydrogenase/acetyl-CoA synthase (CODH/ACS; figure 2b). The catalysis in CODH/ACS is performed by transition metals (Fe and Ni) coordinated as sulfide clusters [50,56,73]. In the enzymatic mechanism [56], electrons from a low-potential Fd are used by the enzyme to reduce one molecule of CO2 to CO at an FeNiS cluster called the C cluster; that CO migrates to a second FeNiS cluster in the enzyme (the A cluster), where it binds one of two Ni atoms at the Nip site to form a Ni-bound carbonyl group, where the two-electron donating property of Ni is important in the catalytic mechanism [56]. A methyl group, donated by the corrinoid iron–sulfur protein CoFeSP, binds the same Ni atom [56] in an unusual metal-to-metal methyl transfer reaction [88], and the carbonyl group inserts the Ni–CH3 bond to generate a Ni-bound acetyl group, which is then removed from the enzyme via thiolysis by CoASH to generate the thioester. Given the methyl group, this route of thioester synthesis involves only metals as catalysts and electron donors, no phosphate is involved, and given CO, the reaction works in vitro, without enzymes, using either Fe2+ or Ni2+ as catalysts [70].

CODH/ACS thus has hallmarks of a very ancient reaction, one that could well have proceeded readily before the advent of genes and enzymes. Across the divide that separates acetogens and methanogens, this primordial chemistry at CODH/ACS is homologous and conserved, as is the CoFeS protein [83], but the synthesis of the methyl group is fundamentally different [73], involving distinct, unrelated enzymes and different cofactors—tetrahydrofolate (H4F) and tetrahydromethanopterin (H4MPT)—both of which are, however, pterins (figure 2d), and chemically more similar than the different names suggest [85].

In acetogens, methyl synthesis entails reduction of CO2 to formate via an NAD(P)H-dependent formate dehydrogenase, an ATP-consuming step at the 10-formyl-H4F synthetase reaction, water elimination via 5,10-methenyl-H4F cyclohydrolase, reduction with NADH by 5,10-methylene-H4F dehydrogenase, and Fd-dependent reduction via 5,10-methylene-H4F reductase to yield 5-methyl-H4F which donates the methyl group to corrinoid iron–sulfur protein (CoFeSP) [73]. Of energetic consequence, the formation of the formyl pterin formyl-H4F is endergonic, requiring ATP hydrolysis, while the 5,10-methylene-H4F reductase step is highly exergonic (–57 kJ mol−1) [73] and is suspected to be a site of energy conservation in some acetogen species [89].

In methanogens, the synthesis of the methyl group involves different enzymes and to some extent different cofactors [83]. The first step is the formation of the carbamate formyl-methanofuran (formyl-MF) from CO2 and the primary amine of methanofuran. Notably, this reaction is spontaneous, requiring only the cofactor and the substrate, with no help from enzymes [90]. Fd-dependent formyl-MF dehydrogenase and a formyl tranferase yield the formyl pterin formyl-H4MPT. A cyclohydrolase, F420-dependent 5,10-methylene-H4MPT dehydrogenase and F420H2-dependent 5,10-methylene-H4MPT reductase yield methyl-H4MPT which donates the methyl group to CoFeSP in carbon metabolism, involving CODH/ACS. In energy metabolism, methyl-H4MPT donates its methyl group to coenzyme-M (CoM).

Thus, the reactions at CODH/ACS are catalysed by homologous enzymes, and the methyl-carrying corrinoid protein CoFeSP is homologous in acetogens and methanogens. But in the methyl synthesis branch of the two groups, the only homologous enzymes are at the CO2-reduction step. The MoCo-binding subunits of Mo-dependent formyl-methanofuran dehydrogenase, FmdB, and of the W-dependent enzyme, FwdB, in methanogens are related to formate dehydrogenase in acetogens [91], which also exist as Mo- or W-dependent enzymes [92]. The other enzymes are not related, use different cofactors, and in some cases catalyse different reactions. Thus, while it is generally agreed that the acetyl-CoA pathway is the most ancient among modern CO2-fixing pathways, the methyl synthesis branches of the Wood–Ljungdahl pathway in acetogens and methanogens are unrelated, even though the methyl groups serve a homologous function at a homologous enzyme, CODH/ACS, en route to thioester synthesis. This suggests three things: (i) the gene for CODH/ACS was present and functioning in their non-free-living common ancestor, (ii) the genes and proteins for the methyl synthesis pathways of acetogens and methanogens arose independently and subsequently to divergence between the archaebacterial and eubacterial stem lineages within the hydrothermal mound, and (iii) that a continuous and abundant supply of accessible methyl groups stemming from geochemical processes—either serpentinization or geochemical redox processes occurring directly at the vent–ocean interface [43]—served as the source of biochemical methyl moieties prior to the advent of the genes and proteins that underpin the methyl synthesis branch of the Wood–Ljungdahl pathway in acetogens and methanogens.

This is another way of saying that when the enzymes of methyl synthesis arose, they did not invent methyl synthesis, they just helped chemical reactions that tended to occur anyway, to occur more rapidly.

In this view, the thermodynamically favoured reaction of H2 and CO2 to methyl groups and methane, similar to that observed in the laboratory [93] or at Lost City [19], was taking place spontaneously (figure 2a), and the genes and proteins (figure 2b) just helped speed it along and add specificity to a highly selectable function (accelerated self-synthesis, given the necessity of methyl groups for self-synthesis). However, those genes arose in independent stem lineages, from which it follows that methyl groups, whether formed by serpentinization or in the walls of the mound, were abundant at the mound over geological periods of time.

If the reader will grant us, for the purpose of argument, a sustained source of acetyl thioesters at a hydrothermal vent setting, the spontaneous and continued accumulation of organic compounds, including peptides, would be thermodynamically possible. A reaction of this type is what Morowitz [45,94] has, rightly, always demanded. Thioesters could lead to some simple carbon chemistry as outlined in fig. 6 of Fuchs [73], some transition metal catalysed reductive aminations of 2-oxo acids as outlined by Huber & Wächtershäuser [95], and some peptide synthesis aided by transition metals and simple reactive compounds such as carbonyl sulfide [96,97]. Keeping in mind that the pyruvate-synthesizing enzyme of acetogens and methanogens, pyruvate synthase, is replete with FeS centres and uses acetyl-CoA, reduced Fd (Fdred) and CO2 as substrates [60], and also that reducing conditions of hydrothermal vents make energetics work in favour of amino acid and peptide synthesis [14], an alkaline hydrothermal vent would, in principle, be working in the right direction thermodynamically, because cells consist mainly (approx. 60% dry weight) of protein and to a lesser extent (approx. 25% dry weight) of nucleic acids.