Results

Thermodynamically driven peptide formation

The formation of peptides is a very slow reaction at pH values close to neutrality where is located the maximum stability of peptide bond. The abiotic formation of peptides can then result from two different kinds of processes. The first one relies on processes increasing the rates under conditions thermodynamically favouring peptide bond formation, i.e. in hot water, by drying cycles or by other physical dehydration means. The pool of (usually short) peptides formed in this way could then have played a prebiotic role. This approach is consistent with the prevalent, but not always clearly expressed, idea that a stepwise non-dynamic self-organisation process occurred: (i) monomer formed and accumulated, (ii) random polymers could be formed by drying or heating, and (iii) supramolecular organisation emerged by chance from active sequences of polymers. However, this approach is in contradiction with the requirements for protometabolisms expressed previously. It is based on different hierarchical degrees of organisation without connection between them and then self-organization only relies on the very low probability of specific arrangements to be formed without possibility of driving force to help the evolutionary process.

Activation of α-amino acids as N-carboxyanhydrides

The second approach consists in considering that peptide formation was a continuous process that took place from free amino acids and was driven by activating agents formed in the environment or by physicochemical processes capable of delivering energy to the system. In this view, a peptide proto-metabolism could have developed before the origin of life and co-evolved with ribonucleotide chemistry (RNA world) leading possibly to the emergence of translation and open-ended evolution at the same time. Unlike a process based on the occurrence of highly improbable contingent events, as proposed above, this view is consistent with the requirements for self-organisation mentioned in section 2.2. A system corresponding to this description has been proposed by Prof. A. Commeyras in Montpellier after the discovery of a potential prebiotic pathway for the formation of amino acids N‑carboxyanhydrides (NCAs) and was considered to display features capable of dissipating the free energy available in the environment for the selection of sequences with defined properties (Scheme 1).

A chemical system presenting some similarities and involving N-carboxyanhydrides formed by reaction of carbon monoxide on reducing minerals has also been proposed for driving a peptide proto-metabolism. The capabilities of the a-amino acid / peptide proto-metabolism shown in Scheme 3 have been illustrated by the APED model (Scheme 2), which has been devised starting from its architecture and which constituted the proof of principle that a non-racemic pool of a-amino acids may be formed abiotically and remain dynamically stable with no need of a contribution of living organisms. Further studies carried out in Montpellier and elsewhere gave additional information on the capabilities of N-carboxyanhydride chemistry. (i) An alternative pathway was identified for NCA formation that was observed by the reaction of cyanate by itself or its precursors such as urea (Scheme 3, path b). (ii) Leman, Ghadiri and Orgel identified the reaction of carbonyl sulfide (COS) with free amino acids as a source of NCA in the presence of iron complexes as prebiotically plausible oxidizing agents.

Activation of C-terminal aspartic acid residues

Alternatively to this N-terminal process of peptide elongation (i.e. a process in which the growing polymer is coupled to an activated monomer), the activity of cyanate as a C-terminus activation reagent was studied in a collaboration between Montpellier and Marseille’s groups. It turned out that cyanate is capable of activating peptides involving a C‑terminal aspartic acid residue leading to an anhydride (Scheme 3). This possibility of C-terminal activation is the result of neighbouring group assistance that served as a model of the process devised in the general case.

Activation of C-terminal residues in peptides as 5(4H)-oxazolones

Activating C-terminal residues in peptides in the general case turned out to be impossible with cyanate because of the insufficient degree of activation of this reagent. We considered that the amide oxygen of the preceding residue would be susceptible of giving rise to a 5(4H)-oxazolone intermediate with more efficient activating agents (Scheme 6). The formation of 5(4H)-oxazolones is a well known side-reaction in peptide synthesis that is a subject of concern for peptide chemists because it usually results in the epimerization of the activated residue. In the context of the activation of racemic mixtures of a-amino acids this side-reaction is no longer a drawback and may constitute a process leading to the selection of homochiral sequences in peptides, which constitutes one of the issues addressed in this project. This reaction is likely to take place with activating agents stronger than cyanate and to compensate for the low concentration of the adduct formed reversibly (Scheme 4). Usual reagents of peptide synthesis as carbodiimides are likely to behave in this way even though there was little information on the use of carbodiimides as activating agents for amino acids in aqueous solution. In fact, a water-soluble carbodiimide (EDC) has been proposed as an efficient activating agents in aqueous solutions of amino acids provided that N-acyl-amino acids are added in the mixture. Although this has not been noticed by the authors, we suspected that 5(4H)-oxazolones to be formed as intermediates.

Results confirming this hypothesis have been obtained through a collaboration of the groups Montpellier and Marseille. 5(4H)-Oxazolone formation was monitored by NMR studies starting from simple acylamino acids and results in a very efficient proton-deuteron exchange at the a-carbon when the reaction is carried out in heavy water illustrating the chiral instability of these intermediates. Although the carbodiimide reagents that are commonly used as activating agents in peptide synthesis are unlikely to have been present in prebiotic environments, the possibility that cyanamide could act as a substitute has been mentioned in the literature. This possibility has been demonstrated in our preliminary studies since cyanamide promoted the proton-deuteron exchange in a way very similar to EDC, which constitutes an evidence of the formation of a 5(4H)-oxazolone intermediate, although the reaction of cyanamide turned out to be tedious and required high temperatures (80°C). Cyanamide constitutes a key reagent that has emerged in the chemistries studied in the past by both partners. It has been identified in the interstellar medium. Its photochemical conversion into carbodiimide has been studied in Marseille. It has also led to results on the synthesis of peptides brought about by the pre-existing collaboration between Montpellier and Marseille’s groups. By the way, iIt is involved as a precursor in the prebiotically plausible nucleotide synthesis proposed by John Sutherland et al. Its formation is likely to occur by chemistry in the atmosphere and its long lifetime makes its distribution into other environments possible. It can also be formed by photochemistry of NH₄CN solutions by processes that are promoted by transition metal ions as demonstrated in the literature for Fe complexes. The literature also indicates that cyanamide can be formed by heating ferrocyanide salts, a pathway that may be compatible with the formation of these complexes under early Earth conditions.

Diastereoselection from peptide derived 5(4H)-oxazolones

The epimerization resulting from C-terminus activation could in principle constitute a source of stereoselectivity resulting from the presence of chiral centres in the activated peptide chain or in the nucleophile (Scheme 5). An experimental study of the process has been carried out through the Ph.D. investigations undertaken by Damien Beaufils since October 2012 and its first result have been published recently. This unprecedented study demonstrated that under prebiotically relevant conditions in aqueous buffers (MES or MOPS buffers pH 5.5–7.5) epimerization of the 5(4H)-oxazolones takes place at least as fast as subsequent hydrolysis or aminolysis processes. As a result significant diastereoselective excesses (d.e.) could be observed through activation/hydrolysis of acetylated dipeptides (with values of d.e. values ranging from 12 to 40%). More interestingly, the attack of chiral nucleophiles produced higher values of d.e. (up to 60%). It must be mentioned that we observed a preference for homochiral sequence in every case, which supports the likelihood of an involvement of this process to symmetry breaking under abiotic conditions.

Coevolution of peptides and nucleotides

Although the abiotic formation of ribonucleotides had been considered as impossible by many researchers, recent developments, mainly carried out in the group of John D. Sutherland, have demonstrated that abiotic chemistry can lead to ribonucleotides in a somewhat activated form through reactions of simple precursors (glycoaldehyde, glyceraldehyde, cyanamide, cyanoacetylene, phosphate), which led to a breakthrough, published in Nature in 2009, with the synthesis of two of the four monomers. But new indications that purine nucleotides may also be obtained through similar pathways have been reported. Moreover, consistently with the theoretical arguments developed above, photochemistry plays an important role in the overall process. These results show that the chemistry of ribonucleotides must not be discarded in the context of early supports of information, and that a scenario based on a linked origin of RNA and coded peptides can be considered as reasonable. Investigations have been carried out in this latter direction, which demonstrated the possibility of obtaining either amino acid adenylates (the most activated intermediates of peptide biosynthesis) or potentially important aminoacylated ribonucleotides. This spontaneous formation of aminoacyl adenylates from AMP and amino acid N-carboxyanhydrides through a chemical path that does not need catalyst shows (i) that abiotic chemistry can lead to essential intermediates of the translation process that are not formed spontaneously by the uphill biochemical reaction of amino acids and ATP, (ii) that early processes may be different from present day biochemical pathways, which can take advantage of the stabilization of unstable intermediates and transition states by enzymes, and (iii) that the role of ATP as an energy currency is likely to have evolved at a subsequent stage, at least after it could have served as an activated RNA monomer.

General scheme of our amino acid activation and peptide synthesis engine