Evolutionary switches between two serine codon sets are driven by selection

Abstract

Serine is the only amino acid that is encoded by two disjoint codon sets so that a tandem substitution of two nucleotides is required to switch between the two sets. Previously published evidence suggests that, for the most evolutionarily conserved serines, the codon set switch occurs by simultaneous substitution of two nucleotides. Here we report a genome-wide reconstruction of the evolution of serine codons in triplets of closely related species from diverse prokaryotes and eukaryotes. The results indicate that the great majority of codon set switches proceed by two consecutive nucleotide substitutions, via a threonine or cysteine intermediate, and are driven by selection. These findings imply a strong pressure of purifying selection in protein evolution, which in the case of serine codon set switches occurs via an initial deleterious substitution quickly followed by a second, compensatory substitution. The result is frequent reversal of amino acid replacements and, at short evolutionary distances, pervasive homoplasy.tandem nucleotide substitutions | serine codon sets | purifying selection | positive selection | homoplasy P oint mutations are generally assumed to involve substitution of single nucleotides, an assumption that underlies many phylogenetic approaches (1-5). However, the presently characterized mechanisms of mutation do not exclude the possibility that two or more adjacent nucleotides change simultaneously in a single mutational event (6-13). There is experimental evidence that some replication and repair enzymes, for example, mammalian DNA polymerases η (14) and ζ (15, 16), produce excessive amounts of double (tandem) substitutions. Runs of consecutive mismatches between closely related genomes rapidly decrease in frequency with increasing length, following an exponential distribution and thus indicating that nucleotides are mostly substituted one at a time (17). Nevertheless, even for spontaneous mutations, there are indications that several adjacent nucleotides can change simultaneously in a single mutational event (2,(6)(7)(8)18). In particular, Averof and colleagues analyzed nucleotide substitutions in primate noncoding sequences and found a high frequency of double (tandem) substitutions in an approximately twofold excess over the expected frequencies (2).Averof and colleagues further analyzed evolutionary switches between TCN and AGY (where N is any nucleotide and Y is a pyrimidine) codons that encode strictly conserved serine residues in multiple protein families from diverse organisms (2). Serine is unique among amino acids in that it is encoded by two disjoint codon sets, TCN and AGY, which cannot be interconverted by a single-nucleotide mutation. Accordingly, switches between serine codons from the two sets can occur either directly, by simultaneous double (tandem) mutation (TC > AG), or indirectly, via two consecutive single-nucleotide substitutions (TC > AC > AG or TC > TG > AG). In the latter case, the switch would involve as an intermediate either threonine ACN or cysteine TGY, amin...

Multinucleotide mutations cause false inferences of lineage-specific positive selection

“…Further, the assumption that all mutations have the same transversion-transition rate might exacerbate the tendency to misinterpret MNM-produced nonsynonymous changes as evidence for positive selection. Although CMDs might also be driven to fixation by recurrent positive selection [40][41][42] , these considerations suggest that the possibility that failing to incorporate MNMs in likelihood models might make tests of adaptive evolution susceptible to false positive inferences. The BS and related tests might be particularly sensitive to this problem because they seek signatures of positive selection acting on individual codons on individual branches of the tree 11,43 .…”

Crossing fitness valleys via double substitutions within codons

“…From triplets of genomes with reliable phylogenetic relationships that were extracted from the ATGC database 17,27 , we obtained frequencies of double and single substitutions in codons, and in double synonymous controls (see Methods for details). The key difference between the present work and the previous studies is that all the analyses included comparison to double synonymous substitutions that served as null models for the double substitutions in codons.…”

“…The control for mutation biases was achieved by comparing the DF for codon double substitutions to those of double synonymous substitutions. Previously analyzed codon double substitutions in serine codons 17 and in stop codons 16 suggested that these changes are under positive selection due to elevated double substitution frequencies compare to the expectation from single substitutions. Our focus here was to infer the type of selection by using more adequate controls, namely equivalent synonymous double substitutions, in order to address the possibility that apparent selection affecting codon double substitutions was due to mutational biases as previously suggested 9,14 .…”

“…Thus, a second mutation in a codon that reverts a non-synonymous substitution to restore the codon for the original amino acid, generally, is expected to be advantageous. However, given that the apparent positive selection in codon double substitutions could be potentially explained by biased mutational processes that favor multi-nucleotide substitutions 6,9,14,17,19 , it is essential to compare codon double substitutions to an appropriate null model in order to accurately infer selection.…”

Self-Referential Encoding on Modules of Anticodon Pairs—Roots of the Biological Flow System

“…Difficulties in rationalization were reduced when only the Ser, Gly and Pro attributions could be interpreted as a coherent group. The 'enigma' of the two Ser codes, which could not be explained in any of the previous studies [25,26], now receives an understanding based on dimer complementariness. The indication that the Gly and Pro codes would also have belonged to an unforeseen pair (Table 1) found support only when the interpretation provided by the metabolic pathways of amino acid biosynthesis entered the argument.…”