Report

# α-Hydroxy and α-Amino Acids Under Possible Hadean, Volcanic Origin-of-Life Conditions

See allHide authors and affiliations

Science  27 Oct 2006:
Vol. 314, Issue 5799, pp. 630-632
DOI: 10.1126/science.1130895

## Abstract

To test the theory of a chemoautotrophic origin of life in a volcanic, hydrothermal setting, we explored mechanisms for the buildup of bio-organic compounds by carbon fixation on catalytic transition metal precipitates. We report the carbon monoxide–dependent formation of carbon-fixation products, including an ordered series of α-hydroxy and α-amino acids of the general formula R-CHA-COOH (where R is H, CH3,C2H5,orHOCH2 and A is OH or NH2) by carbon fixation at 80° to 120°C, catalyzed by nickel or nickel,iron precipitates with carbonyl, cyano, and methylthio ligands as carbon sources, with or without sulfido ligands. Calcium or magnesium hydroxide was added as a pH buffer. The results narrow the gap between biochemistry and volcanic geochemistry and open a new gateway for the exploration of a volcanic, hydrothermal origin of life.

The theory of a volcanic, hydrothermal, chemoautotrophic origin of life postulates a locally and temporally coherent, evolvable system of autocatalytic, synthetic carbon-fixation pathways, catalyzed by inorganic transition metal precipitates (14) and generating low molecular weight organic compounds from highly oxidized precursors. Here, this system of pathways is termed “pioneer metabolism.” In accordance with the principle of metabolic continuity, the theory assumes a step-by-step evolutionary changeover by evolution of ligand feedback from racemic ligands of the inorganic transition metal precipitates to homochiral metalloenzymes of extant organisms (3, 4). In a continuing effort to establish an experimental grounding for this hypothesis, we experimentally explored the viability of volcanic, hydrothermal carbon-fixation pathways using CO and CN as carbon sources.

We chose Ni or Ni,Fe precipitates as catalytic transition metals because of the catalytic roles of these biometals (as sulfide or hydroxide complexes) in our previous experiments (5, 6); (Ca, Mg)(OH)2 as source for hydroxy ligands and for buffering against acidification; Na2SorCH3-SNa as sources for sulfido or methylthio ligands; and CO and KCN as sources for carbonyl and cyano ligands in accordance with extant [Fe,Ni]- and [Fe,Fe]-hydrogenases (7).

The reaction conditions are listed in Table 1. Unless stated otherwise, the experiments were carried out in a slurry with 10 ml of H2O and 13C-labeled KCN (8). The alkaline pH range is in agreement with the pH requirement of peptide synthesis (6). The range of reaction temperatures was chosen in agreement with previous experiments (5, 6) and within the range of growth temperatures of hyperthermophiles. The CO gas pressure of 1 bar was chosen as in previous experiments (5, 6) and combined with a reaction time of 10 days (run 1). In other runs, the reaction time was shortened (and product yields increased) by an increase of CO gas pressure. The pH was measured at the end of the reaction. Products in the supernatant were analyzed after freeze-drying by gas chromatography–mass spectroscopy (GC-MS).

Table 1.

Formation of α-hydroxy acids and α-amino acids. M denotes metal and can be either Mg or Ca. d, day; h, hour; nd, product was not detected; tr, trace amount.

Run no. Base M(OH)2 M (g) Catalyst/carbon sourcesView inlineT (°C) Time pH (final) Products R-CHA-COOH (μmol)View inline
Fe2+ Ni2+ Na2S (mmol) CH3SNa KCN CO (bar) R H CH3 HO-CH2 CH3-CH2
A OH NH2 OH NH2 OH NH2 OH NH2
1 Ca (1) 0 2 0 0.5 2 1 100 10 d 12.4 0.02 0.03 0.01 (1.8) 0.001 (2) nd nd nd nd
2 Ca (1) 0 2 0 0.5 2 10 100 5 d 12.4 0.3 0.23 0.01 (30) 0.002 (53) 0.003 0.0008 tr nd
3 Ca (1) 0 2 0 0.5 2 75 100 2 d 12.5 0.6 0.08 0.10 (75) 0.04 (41) 0.05 0.01 0.005 tr
4View inline Ca (1) 0 2 0 0.5 2 75 100 2 d 12.5 1.1 1.05 0.12 0.09 0.05 0.02 0.02 0.02
5 Ca (1) 0 2 0 0.5 2 75 120 20 h 12.5 2.5 1.7 0.22 (89) 0.15 (93) 0.07 0.03 tr tr
6 Ca (0.25) 0 1 0 0.5 2 75 80 20 h 10 1.38 nd 0.07 (45) nd nd nd nd nd
7 Ca (1) 0 2 0 0 2 75 100 2 d 12.5 0.7 0.4 0.08 (72) 0.02 (88) 0.04 0.02 0.001 tr
8 Ca (1) 1 1 0 0.5 2 75 100 2 d 12.5 0.2 0.04 0.05 (85) 0.0004 (88) 0.02 0.005 0.02 tr
9 Ca (1) 1 1 0.67 0.5 1.33 75 100 2 d 12.5 0.4 0.05 0.10 (41) 0.0006 (82) nd nd nd nd
10 Mg (0.8) 1 1 0.67 0.5 1.33 75 100 2 d 9 1 0.6 0.20 (43) 0.10 (85) nd nd nd nd
11 Ca (1) 2 0 0 0.5 2 75 100 2 d 12.5 nd nd 0.06 (0) nd nd nd nd nd
12 Ca (1) 0 2 0 0.5 0 75 100 2 d 12.5 nd nd 0.02 (0) nd nd nd nd nd
13 Ca,MgView inline 0 2 0 0.5 2 0 100 2 d 12.8 nd nd nd nd nd nd nd nd
14 Ca,MgView inline 0 0 0 0 2 75 100 2 d 9.6 nd nd nd nd nd nd nd nd
• View inline* Fe2+ and Ni2+ as sulfates.

• View inline The number in parentheses signifies the value of 10013C3/(12C3 + 13C3) as a measure for the ratio of the 13C3 isotopomer to the 12C3 isotopomer.

• View inline 12C-labeled KCN, D2O.

• View inline§ Ca(0.45) + Mg(0.5).

• α-Hydroxy and a-amino acids as main products (Table 1) constitute ordered series defined by the general formula R-CHA-COOH, where R is H, CH3, CH3-CH2, or HO-CH2 and A is OH or NH2. Temperature increase correlates positively with product yield and with the ratio of amino acids to hydroxy acids. The replacement of H2ObyD2O leads to deuterated products. With 13C-labeled KCN, we discriminated cyano ligands from CO (and, optionally, methylthio) ligands as the carbon source. The resulting isotopomers revealed a rich and complex system of pathways involving allcyano, all-noncyano, and combined cyano/non-cyano ligands, as exemplified by ratios of 12C3:13C3 isotopomers for lactate and alanine. The contribution by noncyano pathways correlates positively with a decrease in CO gas pressure and an increase in temperature. This multiplicity of pathways may facilitate metabolic evolution and a stepwise changeover from a Ni-dependent use of CO and/or cyano ligands without energy coupling to a sole use of CO2 with energy coupling. We also detected a-hydroxy-n-valeric acid (run 4, 0.005 μmol), a-hydroxy-i-valeric acid (runs 4 and 5, trace amounts), and a-amino-n-valeric acid (run 4, trace amount). The progressive chain elongation suggests long-chain a-hydroxy or a-amino acids as primordial lipids.

Added 15N-NH3 enters the amino acids, which suggests the participation of a pool of ammonia. The replacement of KCN in run 3 by glycine and alanine generated a-hydroxy acids at very low rates. This means that a-amino and a-hydroxy acids are mainly competitive products and, to a minor extent, consecutive products. The detection of glycine amide by high-performance liquid chromatography (HPLC)–MS (8) suggests carboxamides as intermediates between CN and COOH groups. The detection of pyruvate (runs 3, 4, 5, and 10) having a similar isotopomer ratio as lactate and alanine suggests a-keto and a-imino groups as intermediates.

Acetate, propionate, and butyrate (in varying isotopomer ratios) have been detected with yields decreasing in that order. The detected ethylene glycol suggests a pathway via glycol aldehyde, a C2 sugar. Formate is formed from CO and cyano ligands, with ammonia and formamide as regular by-products.

The reactions are markedly selective. No tar is formed. This may be due to the combination of mild chemical energy and transition metal catalysis. By contrast, the input of harsh physical energy (for example, electric discharges) in the absence of transition metal catalysis results in low-specificity radical reactions with a predominance of tar production (9).

The Ni,Fe precipitate may range from amorphous polynuclear structures to nanocrystals with attached clusters. The precipitate has the nominal compositions (before carbonylation and methylthiolation) of Ni(OH)2 +Ni(CN)2 (runs 1 to 5, 7, and 13); Ni(CN)2 (run 6); Fe(OH)2 + Ni(CN)2 (run 8); (Fe,Ni)(OH)2 + (Fe,Ni)S + Ni(CN)2 + (Ca,Mg)2Fe(CN)6 (runs 9 and 10); Fe(OH)2 + Ca2Fe(CN)6 (run 11); and Ni(OH)2 (run 12). Ni(OH)2 and Fe(OH)2 are stable against dehydration (10) and are insoluble [solubility product constant (Ksp) ≅ 10–16 and 10–15, respectively]. Ni(CN)2 with bidentate cyano ligands is actually Ni[Ni(CN)4] with Ksp ≅ 10–9 for dissociating into Ni2+ and Ni(CN) 2–4 (11). Together with the stability constants β4 ≅ 1030 for Ni(CN) 2–4 and K6 ≅ 2 × 108 for Fe(CN) 4–6 (12), this implies that the concentration of free dissolved CN is effectively zero and that cyano ligands are the reacting species. In the absence of any transition metal (run 14), all cyanide exists as free dissolved CN or HCN, but none of the carbon-fixation products are formed. Therefore, our experiments are in sharp contrast to “prebiotic broth” experiments requiring aqueous HCN (13) or aqueous CN plus ammonia and formaldehyde for Strecker reactions (14, 15).

The redox reactions may be exemplified by the following reaction formulas: $Math$(1) $Math$(2) $Math$(3) $Math$(4) $Math$ $Math$(5) $Math$ where e may stand for reduced nickel centers. Besides being a carbon source, CO serves as a primary source of reducing equivalents (reaction1), akin to the CO dehydrogenase reaction. Hydrogen generated by the reaction (2) is available as a secondary source of reducing equivalents (reaction 3), akin to the hydrogenase reaction. The product CO2 reacts with Ca(OH)2 to produce calcium carbonate or with Mg(OH)2 to produce basic magnesium carbonate.

A recent theory (16) suggests as sites of a Hadean origin intensely fractured, reduced rock in the floors of submerged impact craters, with a wide range of hydrothermal conditions and of rates of hydrothermal quenching. In hot and strongly reducing Hadean magma, fluids have a high molar ratio of CO:CO2 [for example, 1.5:1 at 1400°C, 10 kbars, and log10 of oxygen fugacity relative to Ni-NiO buffer (ΔNNO log10f02) = 3.5] (17). Therefore, Hadean volcanic exhalations must have originated with a correspondingly high equilibrium ratio of CO:CO2. The equilibrium ratio of CO:CO2 decreases with decreasing temperature (due to cooling) along the flow path, as long as the composition can equilibrate by the water gas shift reaction (CO + H2O → CO2 +H2). In the course of this reaction, the reducing agent CO is replaced by H2, which may also function as a reducing agent according to reaction 3. The water gas shift reaction may be catalyzed by transition metals, or it may be noncatalyzed in the case of high water activity. In one special set of conditions with liquid water of maximum activity, the half-value time for noncatalyzed CO oxidation (CO + H2O → HCOOH → CO2 + H2) has been determined as 2 min at 350°C, 2 weeks at 150°C, and 3 years at 100°C (18). Therefore, if the rate of cooling (for example, by intense contact with cold water or ice) is high enough relative to the rate of CO oxidation, the volcanic fluid should be quenched with the result of high-disequilibrium CO and H2 concentrations (19) occurring in many places suitable as the site of emergence of the pioneer organism.

For an aqueous hydrothermal solution of 350°C, an equilibrium concentration of 0.13 mmol of CO per kilogram of solution has been determined with one set of conditions (18). Therefore, rapid quenching of hydrothermal fluids from a temperature sufficiently above 350°C down to ∼100°C may well result in an aqueous non-equilibrium CO concentration in the vicinity of 0.76 mmol/kg, which has been calculated for the CO gas pressure of 1 bar on the basis of water solubility data (20). Moreover, aqueous CO drives the carbonylation of the Ni or Ni,Fe precipitate to form reactive carbonyl complexes. These complexes, however, do not have to form in situ. Rather, under conditions of high pressure and high temperature, CO has been found to mobilize transition metal sulfides by forming carbonyl complexes, which would then be available for subsequent transport (21) by the volcanic fluid to a site of an origin by carbonylation reactions at low temperature. The proposal of primordial carbon fixation within a quenched flow of volcanic fluids contrasts with the proposal of primordial carbon-fixation reactions at the interface between aqueous, hydrothermal H2 and aqueous CO2 in an acidic ocean (22, 23). Finally, under the conditions of our experiments (below 300°C), the depletion of the chemical potential of CO by the formation of CH4 and graphite is kinetically inhibited (19, 24).

Cyano ligands of transition metals have been detected at extant volcanic sites (25) and may well have been abundant under Hadean conditions. H2S and methylmercaptan are found in volcanic gases (26). Nickel, together with iron, must have been abundant in the Hadean crust and in serpentinized crater floor material. Ca(OH)2 and Mg(OH)2 are formed by precipitation or by serpentinization of ultramafic material (27). Ca(OH)2 is also formed by decarboxylation of calcium carbonate to CaO and subsequent hydration (28). Therefore, beds of (Ca,Mg)(OH)2 may well have been ubiquitous as a base material for catalyst precipitation in floors of Hadean impact craters and subject to pH zoning (as well as sulfidization zoning and ligand zoning) (4).

The conditions of the present experiments are closely related to the conditions of previous studies in the context of a chemoautotrophic origin of life (3) and notably to the formation of methylmercaptan (29), to nitrogen fixation (30), to the formation of COS (5), and to the involvement of COS as a source for energy (6) and carbon (2). Therefore, all these reactions could cooperate with the newly found reactions in a locally and temporally coherent manner at volcanic, hydrothermal sites or along volcanic, hydrothermal flow channels. Our results free the hydrothermal origin-of-life debate from a narrow focus on Fischer-Tropsch reactions and obviate considerations (31) of combining the theory of a prebiotic broth of amino acids (9) with the theory of a chemoautotrophic origin of life.

It has been suggested that the formation of amino acids at hydrothermal sites would be under the thermodynamic control of metastable equilibria (32). Our results are consistent with kinetic control, because the a-amino acids are thermodynamically unstable and convert slowly into a-hydroxy acids; glycine is the most favored a-amino acid, whereas, under thermodynamic control, glycine should be least favored (32); yields of a-hydroxy and a-amino acids decrease with increasing carbon skeletons and with an increasing number of mechanistic steps, as expected from kinetically controlled reactions; yields and product ratios vary with the catalyst system. Such kinetic control is a necessary condition for the involvement of our reactions in the autocatalytic (reproductive) metabolism of a pioneer organism. Moreover, kinetic control makes room for an increase of catalytic activity by evolution from very low de novo rates to rather high autocatalytic rates and for the possibility of chiral symmetry breaking by autocatalytic ligand feedback.

This discussion concludes with the evolutionary context of our results. α-Hydroxy and a-amino acids are chelating ligands for transition metals. The synthesis of these compounds under presumptive Hadean, volcanic, hydrothermal conditions therefore supports the notion that the earliest mechanism of reproduction and evolution was based on positive (autocatalytic) feedback (14), whereby certain synthetic products led to ligand-accelerated transition metal catalysis (3, 4) that greatly increased the rates of the synthetic reactions, the activities of their products, and the spectrum of feedback possibilities, with an eventual emergence of metalloenzymes.

Supporting Online Material

Materials and Methods

## References and Notes

View Abstract