1. Introduction

Wolfe-Simon and co-workers have cultured a bacterial strain, GFAJ-1, of the Halomonadaceae from an arsenic-rich environment in Mono Lake, California (Wolfe-Simon et al., 2010). This bacterium can grow in the presence of high levels of arsenic with only trace amounts of phosphate present. The growth of bacteria under these conditions is a testament to the ability of microbial life to adapt to challenging environments. Under these conditions, arsenic is present in samples of the bacteria's macromolecules, notably DNA and proteins. Although the precise nature of the arsenic species is undetermined, the authors suggest that arsenate is incorporated in place of phosphate in these macromolecules. The proposal that arsenate can be a biochemical replacement for phosphate is an extraordinary claim that, as yet, lacks extraordinary evidence. By subjecting such claims to detailed critical scrutiny, and carefully analysing the underlying chemistry, our appreciation of the essential elements for life will be enhanced. However, whatever the details of the adaptation of GFAJ-1 to high arsenic concentrations, it is clear that, even for this organism, phosphate is used in preference to arsenic. Why does nature prefer phosphates? The utilization of chemical elements by cells is a function of their environmental availability and their chemical potential (Williams and Fraústo da Silva, 1996). Westheimer (1987) has eloquently described why phosphate was chosen for its chemical utility to biological systems. Here, Westheimer's analysis is used to shed light on why nature chose phosphate over arsenate as a foundation for earthly life.

2. The Chemical Utility of Biochemical Phosphates

Phosphate has diverse roles in biochemistry. In the form of polyphosphates, such as ATP, it is synonymous with biochemical energy. Attached to organic molecules, via phosphate ester links, it forms the backbone of nucleic acids. It also joins components within phospholipids and many co-factors. These functions are all based on the fact that phosphate can link other groups whilst remaining anionic. Anions can be retained within cell membranes and form the hydrophilic component of amphiphilic molecules such as nucleic acids and lipids. Their electrostatic properties underpin their adoption of biologically useful structures, e.g. bilayers and helices (Benner, 2004; Pratt, 2010). Whilst these physical properties are utilised in biochemical function, the intrinsic chemical reactivity of these phosphate derivatives is critical for biochemistry. Hydrolysis of these molecules is thermodynamically favourable, but occurs only slowly because of electrostatic repulsion between anionic phosphate and the electron-rich oxygen of water. It is this combination of thermodynamic instability and kinetic stability in water that is the key chemical property exploited by life (Lipmann, 1951; Westheimer, 1987). The capacity to utilise ATP as a water-compatible dehydrating agent, and to recycle phosphate derivatives via hydrolysis, is a critical feature of metabolism that is controlled by macromolecular catalysts (Cleland and Hengge, 2006).

3. The Chemistries of Phosphate and Arsenate

The chemistry of arsenate and phosphate differ only in degree. Like phosphoric acid, arsenic acid is a tribasic acid, having three replaceable hydrogens. The ease of removal of protons from the two acids is similar, with arsenic acids being marginally more acidic. Hence, under conditions where phosphate derivatives will be anionic, arsenate analogues will be likewise. However, this charge is not sufficient to engender the same level of stability towards hydrolysis. Hydrolysis takes place via the attack of a water molecule on the arsenic or phosphorus centre. The marginally larger size of arsenic allows better access to an attacking water molecule. Hence, the rates of hydrolysis of arsenate esters (Baer, Edwards and Rieger, 1981) and polyarsenates (Richmond, Johnson, Edwards and Rieger, 1977) are orders of magnitude greater than those of the corresponding phosphates. Arsenates are much less stable in aqueous solution and hydrolyse rapidly in the absence of catalysts. This undermines the accumulation of arsenate derivatives in water and is the critical point of chemical distinction that undermines the use of arsenate as a surrogate for phosphate.

Although the increased hydrolytic lability of arsenate derivatives subverts their role as phosphate surrogates, there is a further complicating factor which underpins the toxicity of arsenic to living systems: their redox chemistry. Oxidized arsenic species, such as arsenate, are more readily reduced than phosphates and the resulting arsenic (III) compounds are believed to be responsible for much of arsenic toxicity (Knowles and Benson, 1983; Thomas, Styblo and Lin, 2001). This greater redox instability of arsenate may act as a further factor to limit the use of arsenate as an alternative to phosphate.

4. Availability of phosphate and arsenate

The selection of chemical elements by living systems is a function of their availability and their chemical utility. Only if an element is available in the environment can an organism make use of it. Phosphate is the 11th most abundant element in crustal rock being present in at the level of parts-per-thousand. It is more than six hundred times more plentiful than arsenic from this source, the latter only being present at the parts-per-million level (Cox, 1989). Despite its relatively high abundance in surface rocks, phosphate is present at only low levels in aqueous systems due to precipitation with metal ions. In particular, the precipitation with iron probably limited its availability for life prior to 2 billion years ago (Bjerrum and Canfield, 2001). This was a fundamental challenge for the origin of phosphate-based life (de Zwart, Meade and Pratt, 2004); however it is unlikely that this problem was overcome by utilization of arsenate. The same iron-based concentration mechanisms that controlled phosphate levels would have acted in an analogous way on arsenate (Pichler, Veizer and Hall, 1999).

5. Choosing Phosphate Over Arsenate in the Contemporary Biosphere

In the contemporary biosphere, arsenate does not normally compete with phosphate because it is generally present at significantly lower levels in the environment. Competition becomes an issue when biological depletion lowers the phosphorus levels to make them comparable with the arsenic equivalent. This is the case in Mono lake, from which the arsenic-tolerant bacteria, GFAJ-1, were cultured (Wolfe-Simon et al., 2010). The similarity of phosphate to arsenate makes distinction between the two difficult, but there is evidence that organisms can distinguish between phosphate and arsenate under these conditions as a way of moderating arsenic toxicity (Morris, McCartney, Howard, Arbab-Zavar, and Davis, 1984).

The exact molecular details of how GFAJ-1 deals with high levels of arsenic remain to be elucidated. Whatever the nature of this interesting adaptation to challenging environmental circumstances, it is not a vestigial remnant of an alternative life. Repair mechanisms may make it possible for microbes to evolve the capacity to deal with some inadvertent incorporation of arsenate into DNA and other metabolites, but extensive incorporation would be problematic on simple chemical stability grounds. That the substitution of arsenic for phosphate as a nutrient compromises function, is clear from the improved viability of the microbes with higher level of phosphate. The report on the tolerance of GFAJ-1 to arsenic raises numerous questions about metabolites and metabolism that can be subject to experimental testing. For example, to what extent can arsenic replace phosphorus in different metabolites? How is the stability of such arsenic-containing analogues affected by the change? What are the adaptation mechanisms of these bacteria to environmental challenges?

6. Choosing Phosphate Over Arsenate At the Origin of Life

Would life have chosen phosphate if it had its origins in an environment rich in arsenate? Earthly metabolism relies on the accumulation of phosphate derivatives and the controlled catalysis of transfer of phosphates between metabolites. In contemporary organisms, this chemistry is effected by macromolecular catalysts (Cleland and Hengge, 2006). At the early stages of the emergence of life, the increased spontaneous transfer of arsenate in the absence of catalysts might have facilitated such processes. However, the increased lability of arsenate would have quickly become a liability in attempts to generate a complex proto-metabolism. The spontaneous hydrolysis of polyarsenates would have been a continuous drain on the 'biochemical energy' of the system – hydrolysing the dehydrating agents to no benefit. Furthermore the hydrolysis of oligomers based on arsenate would have limited their size, and hence information content. Ever since Eigen's insights into error thresholds (Eigen, 1971), we have been aware that the complexity of information that can be carried by a living system is limited by fidelity of reproduction of that information. A telling example is the evolution of DNA from RNA as the preferred carrier of genetic information. RNA hydrolyses more rapidly than DNA (Eigner, Boedtker and Michaels, 1961) by virtue of the 2'-hydroxyl group attacking the neighbouring phosphate centre. The removal of this hydroxyl group generated a polymer with greater hydrolytic stability. This provided an opportunity for increased fidelity in the reproduction of genetic information and hence increased genome size. This limiting mechanism for the stability of RNA is exactly the chemistry that occurs much more rapidly for arsenate derivatives. Hence, were an arsenate analogue of RNA to arise, there would be a strong selection pressure to convert to the phosphate version. The lower availability of arsenate, combined with this excessive hydrolytic instability would make arsenate inviable as an alternative to phosphate in the origin of life on earth.

7. Implications for Studies of The Origin of Life

It is always salutary to challenge anthropocentric prejudices about life 'in our own image'. The claims made in the report of the culturing of GFAJ-1 in arsenic-rich media provide an excellent opportunity to reflect on wider issues of earthly metabolism and the search for life in the wider cosmos. Once the claims of arsenic utilisation have been subjected to detailed critical scrutiny our appreciation of the essential elements for life will be enhanced. However, if life is fundamentally an aqueous phenomenon, it is hard not to believe that arsenic would be side-lined from the start.