Primordia Vita. Deconvolution from Modern Sequences.

The first principles of the evolution of the triplet code (Trifonov 2004), suggested by the consensus evolutionary temporal order of amino acids are: (1) Abiotic start, (2) Primacy of thermostability, (3) Complementarity of codons and of early mRNA, (4) Processivity of codon acquirements, each having its precursor, and (5) Codon capture at latest stages. The codon chart, its latest update, is shown in the Figure 1. It is built on the basis of total 82 different criteria for temporal order of amino acids in evolution, of which 60 have been used for construction of previous versions of the chart (Trifonov 2000, 2004), and 22 new criteria are taken from additional independent sources (references can be provided by request). As it was found in previous work (Trifonov 2000, 2004), the final chart is not sensitive to filtering procedures, so that even simple averaging of the chronology vectors results in the temporal order identical to the filtered data, within error bars of the estimates. Thus, for the latest update of the chart (Figure 1) a simple summation of the vectors is used. The only filtering applied is removal of earlier chronologies of the same author. The consensus temporal order of amino acids is shown on the top of the chart. Each line of the chart contains two complementary codons, appearing as pairs consecutively, following their descending thermostability. This is indicated on the left side of the chart as melting enthalpies of respective complementary codon pairs, calculated from original data for ribodinucleotide steps (Xia et al. 1998). Codons for valine and aspartate, according to the chart, are formed by transitions in the middle of the very first codons GGC (Gly) and GCC (Ala). All other codon pairs of the chart are formed by mutations in the third redundant codon positions, and by complementary synthesis of the mutated codon. The last six amino acids of the consensus chronology are in full correspondence with the idea of “codon capture” (Osawa et al. 1992): when all codons are already engaged, the new amino acids can be accommodated only by utilizing (capturing) the codons already assigned earlier.

Evolution of the triplet code. The chart is based on the consensus amino acid chronology averaged over 82 different amino acid temporal orders suggested in literature (see the text). For more detailed description of the chart see (Trifonov 2004).

Full size image

Complementarity principle in the evolution of the triplet code had been introduced first by Eigen and Schuster (1978). It is based on the original observation that alanine and glycine, the highest yield amino acids of the Miller’s mix (Miller 1953, 1987), are encoded today by the complementary codons, GGC and GCC, respectively. This complementary pair is also thermodynamically the most stable of all possible complementary codon pairs.

From the complementarity principle it follows (Trifonov et al. 2001; Trifonov 2005) that all amino acids can be divided in two families encoded either by n-Purine-n or n-Pyrimidine-n codons, Gly-family (G) and Ala-family (A), respectively. Every complementary pair of the codons (lines in the Figure 1) corresponds to two different amino acids from the “complementary” families G and A. In the first step (line) these are Gly(G) and Ala(A), in the second step Asp(D) and Val(V), third – Gly(G) and Pro (P), and so on. The Gly-family: G, D, E, R, S, Q, N, K, H, C, Y and W. The Ala-family: A, V, P, S, L, T, I, F, M. In the course of evolution of the codon table the codons corresponding to Gly family would all appear in the same strand as the original Gly codon GGC, while the Ala family codons – in the complementary strand, that carried the original codon GCC. Such division of the amino acids in two families, presumably, has been strictly maintained during duplex mRNA stage and, perhaps, all the way during evolution of the triplet code and beyond. Indeed, it turned out that amino acid replacements in modern proteins maintain the rule, so that Ala family amino acids are replaced almost exclusively by Ala family residues, and Gly family – by Gly family residues (Trifonov 2005). This confirmation of the fundamental complementarity principle for codons allowed to estimate, from the modern sequences, the size of the most ancient short peptides, at the very onset of the evolutionary history of the codons (Trifonov et al. 2001). For that purpose the sequences were rewritten in the binary G and A alphabet, and the ancient picture of alternating patches of A and G residues reconstructed, appearing as well detectable oscillation hidden in the sequences. The very first mini-genes are found to have contained only 6–7 codons, for mini-proteins A _6–7 and G _6–7 (Trifonov et al. 2001).

These short mini-genes could exist either in a duplex form, two complementary genes together, or as separate single-stranded molecules. In this case they could form complementary hairpin-like structures themselves, which would be more stable and, thus, advantageous. Could these hairpins of 18–21 bases, with 8–9 base pair stems be detected as molecular fossils in modern mRNA sequences? Massive mutational changes accumulated during billions of years leave no hope for survival of the hairpins in their intact original form. Considering, however, relative conservation in amino acid replacements (independent A and G families) one may expect partial survival of the original complementary middle-codon positions as purines “complementary” to pyrimidines. In this case one would expect that in mRNA sequences the distances between the middle-position purines and pyrimidines in the sequences could still reflect the complementary hairpin structure of the ancient mini-genes. In case of the 9-basepair stem hairpins with 3-base loops the distances 18, 12 and six bases may still appear as more frequent in the histograms of the complementary distances in the modern mRNA sequences. The hairpins of 18 bases (6-basepair stems and 6-base loops) would generate, similarly, the distances 15 and nine bases. That is, the histograms should contain a very characteristic feature – sharp drop in the occurrence of the complementary distances at 18 bases (for the mini-genes encoding 6–7 amino acids). As the Figure 2 demonstrates, indeed, the expected change is observed, exactly at 18 bases. This confirms the size of 6–7 amino acids for the earliest encoded peptides, as estimated above from hidden alternation of patches of A- and G-families of amino acids in modern proteins.

Distance histograms for purine/pyrimidine and pyrimidine/purine contacts in mRNA of four different prokaryotes. The histograms for respective shuffled sequences are shown for comparison, with the error bars, estimated as square roots of the occurrences.

Full size image

One way to detect the peptides is to locate in the protein sequences sections consisting of old amino acids. The age of a given amino acid may be described half-quantitatively by the number of steps in the consensus amino acid chronology (Trifonov 2004; Sobolevsky and Trifonov 2005), counting from the youngest, last amino acid, tryptophane, of age 1 (see Figure 1). An alternative way to evaluate the age of a given sequence fragment is to measure its degree of conservation. The most conserved sequences are omnipresent, that is found in all known proteomes. The degree of conservation may be measured by the proportion of species harboring the sequence fragment. The older the sequence, the more conserved it is, i.e., the more proteomes contain the fragment. As it has been recently shown (Sobolevsky and Trifonov 2005), the degree of omnipresence strongly correlates with the evolutionary age of the conserved sequences, as calculated from the numbers of steps in the amino acid chronology.

Modern functional involvement of the omnipresent peptides would not necessarily correspond exactly to their original functions. It is natural, however, to expect that in whatever capacity they excelled in the past, that capacity is still their strong characteristic today. If, for example, the motif LSGGQQQR, a typical element of ABC transporters today, plays some particular role in the transporters, its original role has been, likely, the same when the element acted in much simpler molecular environment, perhaps, as part of a very small molecule. The hypothetical small molecule could have been ancestral to the ABC transporters, or one of the modules of which the transporters are built. The modular structure of modern proteins is suggested by sequence conservation plots where the numerous highly conserved short motifs are indicated by sharp peaks at various sequence positions (Aharonovsky and Trifonov 2005).

The Table I lists those modern proteins where the highly conserved elements appear. These octamers are not omnipresent, except for the one on the top. This would mean that the octamers suffered some changes at various stages of evolution. Inspection of the list of respective protein functions reveals that the top conserved octamers are not involved in elementary syntheses. Only in the lines 28 and 31 the first activities of this kind appear (enolase and CTP synthase), followed by other enzymes involved in the synthesis of RNA monomers.

The primacy of RNA in early evolution has long been an accepted view. But where are the enzymes synthesizing amino acids? Inspection of the full list of the octamers (not shown) suggests that the first amino acid enzymes appeared only after several hundred steps (conserved in about 70% genomes). Reductase that reduces ribose to deoxyribose appears at the conservation level <50% of genomes. Thus, enzymatic production of RNA components appears to have been first, amino acid syntheses next, and synthesis of DNA monomers – last. Essentially, it is very close to what one would expect a priori. What is completely counterintuitive is seemingly full life, with replication, transcription and translation, but without any internal material resource at the earliest conserved stages of molecular evolution, as it is witnessed by the oldest omnipresent peptides. The simple suggestion bursts in: all the elementary syntheses in the beginning occurred abiotically, and the monomers had been supplied by environment. Wild as it sounds, the suggestion is not without good independent support. First, this is of course classical synthesis of amino acids in the imitation experiments of Miller (1953, 1987). Nine of 10 amino acids of the Miller’s mix appear at the top of the consensus chronology of amino acids (Figure 1). That is, with these 10 amino acids the Life could, indeed, proceed from its onset to the template syntheses without any domestic supply of the amino acids. As the Table 1 suggests, the first elementary building elements not readily supplied by the environment, had been nucleobases and ribonucleotides. In its earliest stages Life, apparently, could do well with the small amounts of the RNA monomers generated abiotically. One system of naturally catalyzed synthesis of cytosine and purine in appreciable amounts is TiO₂ catalysis in presence of formamide (Saladino et al. 2004). It is, thus, possible to make at least some parts of RNA abiotically. The sequence analysis data presented in this paper is an encouragement for further studies on the abiotic syntheses, which may well turn to be more generous and diverse than it is currently believed.