1. Introduction

Understanding the emergence and existence of biological complexity entails a detailed description of the sum reactions of metabolism. Consideration of processes that allow for cellular propagation (biosynthetic reactions), as well as the metabolic (energetic) processes that power them is paramount for understanding the origin of life. Significant efforts have been directed at elucidating which energy sources may have been harnessed in the first biochemical pathways (Martin and Russell, 2007; Martin, Baross, Kelley, and Russell, 2008; Deamer and Weber, 2010). However, challenges remain in deciphering the mechanisms of particular generative processes constrained by appropriate geochemical and environmental locations for the emergence of biochemical complexity and life.

Among the molecular transformations long recognized to be important to life's origins are the condensation reaction(s) that occur readily from hydrogen cyanide polymerization. First recognized over two hundred years ago (Proust, 1806, 1807), the polymerization of hydrogen cyanide (HCN) in aqueous solution has been shown to result in a number of molecules that feature in biological systems. These include the purines adenine, hypoxanthine, guanine, and xanthine (Oró, 1960; Sanchez, Ferris, and Orgel, 1967; Roy, Najafian, and von Ragué Schleyer, 2007) and some amino acids (Matthews, 1991) (see Figure 1). The production of purines from HCN has led to the proposal that purines may have served as the first nitrogenous bases involved in coding functions in biology (Joyce, 1989). Holm and coworkers (Holm, Dumont, Ivarsson, and Konn, 2006) speculated that boron stabilized ribose (Ricardo, Carrigan, Olcott, and Benner, 2004) formed in hydrothermal areas via HCN driven purine synthesis, may have lead to the formation of the first ribonucleosides.

In contrast to the exothermic process of cyanide polymerization, modern biology accomplishes purine biosynthesis in a complex energy expensive pathway starting with ribose-5-phosphate, an intermediate of the pentose phosphate pathway (Buchanan, 1951; Hartman and Buchanan, 1959). The molecule is first activated by a phosphoribosyltransferase, catalyzing the addition of pyrophosphate to the 1-carbon of the ribose. Following this, the carbon, nitrogen, and hydrogen atoms that form purines are added in several enzymatic amido-transfer and trans-formylation reactions. These enzymes utilize the following substrates progressively: glutamate, glycine, formyl-tetrahydrofolate, glutamine, CO₂, aspartate, and formyl-tetrahydrofolate; together, the reactions yield the purine derivative inosine 5'-monophosphate (IMP), the precursor for both adenosine-monophosphate and guanosine-monophosphate, both of which require two additional steps (and one ATP/GTP equivalent) from IMP for their formation.

The contrast in complexity between the two purine synthesis pathways (abiotic and biotic) suggests a scenario wherein purine synthesis may have been accomplished via simple HCN polymerization at or prior to the origin of life and as living systems evolved, the protein driven biosynthesis process observed today became dominant.

A major challenge to invoking cyanide polymerization for purine synthesis however is the availability of HCN itself. While HCN forms readily in the gas phase (Matthews, 1991; Minard, Hatcher, Gourley, and Matthews, 1998), the formation (or delivery) of HCN at biologically relevant locales is more problematic as the areas where molecular organization relevant to the origin of life are generally aqueous, not gaseous, environments. To satisfy environmental constraints on the directed delivery of HCN, the generation of HCN in the oceanic basement from the geochemically driven reaction of CO and NH has been proposed (Holm and Neubeck, 2009). These reactions would produce HCN as a precursor for generating the organic compounds discussed above. While moving closer to the realization of HCN delivery at relevant environments, this work does not address just how sufficient concentrations of HCN could be delivered specifically to areas where it could be of use in conjunction with other early reactions for further synthesis and a growing metabolism.

In light of recent biochemical observations that show the iron sulfur cluster dependent radical initiated formation of cyanide from tyrosine, we speculate that similar reactions on the early Earth may have been responsible for the formation of HCN in relevant geochemical environments. In these localities, transition metal sulfide mineral phases could result in the formation of carbon centered radical species. Molecules derived through this type of process would have contributed to the growing network of chemical reactions and proto-metabolism that eventually resulted in contemporary life.

2. Radical Chemistry in Enzymatic Catalysis

From energy transfer to chemical rearrangements, radicals (unpaired electrons) feature prominently in biology. Outside of their involvement in redox processes related to energy conservation, many biological radicals occur in catalytic processes (Frey, 2001, 1990; Stubbe, 1988), where the radical serves as a means to initiate reactions that would be difficult to accomplish by two electron transfer processes (Booker, 2009). Living systems generate these species in a variety of ways, but share in common the utilization of a transition metal capable of transferring a single electron.

In biology there are three general mechanisms for generating reactive radical species that carry out various reactions: oxygen dependent processes that often involve dioxygen reaction with an iron center, anoxic cobalt dependent processes, and anoxic iron sulfur cluster dependent processes (see Frey, 2001) and references therein). Considering the "first" biologically relevant radical chemistry reactions, those that are oxygen dependent are unlikely candidates on a low oxygen Earth. This leaves the cobalt dependent and iron-sulfur cluster dependent reactions as possibilities for early radical chemistry. Here, recent biochemical research of an iron sulfur radical dependent process is highlighted in possible connection to early HCN producing radical catalysis relevant to the origin of life.

3. Geochemical and Biological Constraints in Early Radical Chemistry in the Context of Cyanide Formation

In considering plausible biologically relevant chemistry on the early Earth, both the geochemical constraints of the Earth and known chemical properties of living systems must be considered. Iron sulfide based minerals are intrinsic components of the global sulfur and iron cycle, and are widely distributed on the Earth (Rickard and Luther, 2007),. Such minerals have been proposed to be relevant to the emergence of biochemistry especially at hydrothermal vent locations (Russell et al., 1989, 1994; Russell, 2007; Martin and Russell, 2003, 2007; Wachtershauser, 2007; Milner-White and Russell, 2010; Nitschke and Russell, 2010; Russell and Kanik, 2010) where the interaction of these ionic species (ferrous iron and sulfide) results in the propensity to form compartments in which catalytically derived reaction products may be sequestered (Milner-White and Russell, 2010; Nitschke and Russell, 2010; Russell and Kanik, 2010).

With respect to what can be gleaned from contemporary biology, recent biochemical advances have shown the chemical interconversion between amino acids to HCN and CO via transition metal sulfide catalysis. These observations have revealed the operation of iron sulfide derived radical chemistry in which a [4Fe-4S] cluster binding radical-S-adenosylmethionine (SAM) protein is responsible for the generation of cyanide from the amino acid tyrosine in a process where the cyanide is directed towards the formation of the [FeFe]-hydrogenase catalytic active site (Figure 2) (Driesener et al., 2010). In this case, an iron sulfur cluster-bound to the enzyme HydG initializes the formation of a deoxyadenosyl radical, which then abstracts a hydrogen from tyrosine, promoting alpha-beta bond cleavage of tyrosine to form CN^- and CO (Shepard et al., 2010). While existing as an example of an evolutionarily advanced reaction that produces a complex metalloenzyme active site, the mechanism implies that through radical-initiated chemistry, the interconversion between amino acids and the diatomic molecules CO and CN^- may have occurred in the context of the geochemical metal-sulfide milieu on the early Earth.

In contemporary biology, the formation of CO and HCN from tyrosine is directed at the maturation of the [FeFe]-hydrogenases, upon which CO and CN^- occur as ligands (Peters, Lanzilotta, Lemon, and Seefeldt, 1998). It is notable that other than the above described biological formation of cyanide, two other mechanisms of HCN production have been observed. One occurs in the biosynthesis of the other CN^-- generation proceeds by the action of the two proteins HypF and HypE which generate an enzyme-bound thiocyanate from carbamoyl phosphate in a two ATP requiring process, which can later be donated to iron. (Blokesch and Böck, 2002; Blokesch et al., 2004; Paschos, Glass, and Böck, 2001; Reissmann et al., 2003).

The other known process of cyanide formation in biology occurs by the action of HCN synthase that is present in some Pseudomonas, fungal, and algal species. In this case the cyanide is generated from the oxidation of glycine by a putative iron sulfur cluster-containing enzyme (Laville et al., 1998; Blumer and Haas, 2000). In these organisms, it is thought that the generation of cyanide occurs as a defense mechanism from pathogens. Interestingly, this enzyme is not related to HydG at the sequence level, and although it presumably shares the reaction intermediate dehydroglycine (Laville et al., 1998; Blumer and Haas, 2000; Driesener et al., 2010), the two enzymes most likely employ different mechanisms to generate cyanide. The observation of three independent systems for biogenic CN^- production prompts questions as to which (if any) can inform research as to the early formation of cyanide in hydrothermal settings. In terms of simplicity, the generation of HCN by HydG in the [FeFe]-maturation system is perhaps the simplest requiring only one protein. In addition, the wide distribution of iron sulfide minerals on the surface of the Earth (Rickard and Luther, 2007), suggests that iron sulfide centers analogous to contemporary protein-bound clusters may have been present on the early Earth as "ready made" (Russell, 2007). Abiotically derived mineral motifs may have thus acted in concert with hydrothermal pathways involved in the formation of amino acids (Huber, Eisenreich, Hecht, and Wächtershäuser, 2003; Huber and Wächtershäuser, 2003; Shock, 1992) to transform amino acids into cyanide through radical chemistry. This derived HCN could then carry out various synthetic reactions and as discussed further below, transition metal catalyst modification (Figure 3). It is important to note that while tyrosine was specifically observed to be responsible for the formation of cyanide by the HydG enzyme (Shepard et al., 2010), any amino acid could in theory provide the basis for such a conversion. For example, the degradation of glycine is a thermodynamically favorable process and could be expected to occur upon radical initiation (Peters, Szilagyi, Naumov, and Douglas, 2006). Thus, the formation of cyanide from amino acids could in principle be a general reaction where transition metal sulfides promoted homolytic degradation of amino acids and form a link between amino acids, cyanide, and its products (Figure 3).

4. Prospectus

The formation of cyanide in low oxygen conditions from amino acids may form a critical chemical link to the formation of HCN on the early Earth. As hypothesized by Joyce (1989), the purine nucleobases derived from HCN polymerization may have been the initial nucleobases for information storage and transfer. Together with the notion that ribose may have also been made in similar hydrothermal scenarios (Holm et al., 2006) (stabilized by borate minerals (Ricardo et al., 2004), the amino acid (through HCN) derived purine bases may have served as the first nucleobase polymers. In addition to the observation of borate stabilized sugar formation, the observation of silicate promoted sugar formation and sequestration provides further support for the co-occurrence of HCN derived nucleobases with ribose (Lambert, Gurusamy-Thangavelu, and Ma, 2010; Lambert, Lu, Singer, and Kolb, 2004). Addition of these bases to the ribose could occur as described by Mellersh and Smith in high pH solution via nucleophilic attack of a carbonyl group of a sugar such as ribose by a nitrogen atom of the base (Mellersh and Smith, 2010). This reactivity is consistent with oceanic hydrothermal scenarios, where silicate minerals have been observed to precipitate in vent simulation experiments which mimic an alkaline seepage flowing into an anoxic acidulous ocean (Mielke et al., 2010).

In such a scenario, this type and level of chemical evolution would presumably occur at the bottom of a hydrothermal mound (see for example Figure 5 in Martin and Russell, 2003). Here geologically derived reducing equivalents act to form various organic compounds (such as amino acids) and these would then be the subject of continuing modification by transition metal catalysts found in the hydrothermal mound. With the formation of the first CN^- derived purine nucleobases, chemical evolution in the mound make possible the eventual genetic takeover that subsequently occurred in chemical evolution.

Aside from the ability of CN^- to form polymers and purine bases, another activity important in the origin of life is in the ability of cyanide to modify transition metals by ligation. In this evolution of life-giving catalysts, this type of ligand modification at transition metal centers may have been a major step in the creation of a diverse repertoire of catalysts capable of participating in a growing set of biochemically relevant reactions (McGlynn, Mulder, Shepard, Broderick, and Peters, 2009; Haydon, McGlynn, and Robus, 2010). Thus the possible formation of CN^- as described may have been significant in making both the molecules responsible for genetic coding (purines), and metabolic/catalytic diversification possible. In this light, the early formation of CN^- answers some metabolic as well as replicative concerns, and lends further credence to the notion that the CN^- molecule played an important role in the emergence of biological complexity (Figure 3).

In the enzymatic formation of CN^- from tyrosine, electron transfer from a reduced protein-bound iron sulfur cluster results in the homolysis of a carbon-sulfur bond of the sulfonium (S⁺R₃) group of S-adenosylmethionine. While thiol group containing molecules have been shown to form in hydrothermal scenarios (Heinen and Lauwers, 1996), the formation of a sulfonium ion has not been observed in transition metal sulfide based origins of life experiments. However, anoxic radical formation has been observed in such experiments with the pyrite promoted formation of hydroxyl radicals, which could react with a host of organic molecules (Cohn et al., 2006; Cohn, Fisher, Brownawell, and Schoonen, 2010), perhaps in ways that could lead to amino acid degradation to CN^- as per the contemporarily observed enzymatic system. A major feature of radicals in biology however is the control of these reactive entities (Duschene, Veneziano, Silver, and Broderick, 2009), and thus a challenge to the origins of life field is rationalizing the means that these were harnessed for their chemical modification abilities in the first place. Given these constraints, we suggest that future experiments involving the analysis of possible sulfonium ion formation in reactor experiments, as well as those involved in probing for the formation of paramagnetic (free electron) species with techniques such as EPR would further the understanding of primordial radical reactions and their products.

Acknowledgments: S.E.M. is grateful for insightful discussions with Michael J. Russell, Nick Lane, Peter Roach, and Martin Challand. The authors acknowledge support from the NASA Astrobiology Institute-Montana State University Astrobiology Biogeocatalysis Research Center (NNA08CN85A). S.E.M was supported by a fellowship from the National Science Foundation Integrative Graduate Education Research Traineeship from the MSU Program in Geobiological Systems (DGE 0654336).