A global perspective of codon usage

Codon usage in 2730 genomes is analyzed for evolutionary patterns in the usage of synonymous codons and amino acids across prokaryotic and eukaryotic taxa. We group genomes together that have similar amounts of intra-genomic bias in their codon usage, and then compare how usage of particular different codons is diversified across each genome group, and how that usage varies from group to group. Inter-genomic diversity of codon usage increases with intra-genomic usage bias, following a universal pattern. The frequencies of the different codons vary in robust mutual correlation, and the implied synonymous codon and amino acid usages drift together. This kind of correlation indicates that the variation of codon usage across organisms is chiefly a consequence of lateral DNA transfer among diverse organisms. The group of genomes with the greatest intra-genomic bias comprises two distinct subgroups, with each one restricting its codon usage to essentially one unique half of the genetic code table. These organisms include eubacteria and archaea thought to be closest to the hypothesized last universal common ancestor (LUCA). Their codon usages imply genetic diversity near the hypothesized base of the tree of life. There is a continuous evolutionary progression across taxa from the two extremely diversified usages toward balanced usage of different codons (as approached, e.g. in mammals). In that progression, codon frequency variations are correlated as expected from a blending of the two extreme codon usages seen in prokaryotes.AUTHOR SUMMARYThe redundancy intrinsic to the genetic code allows different amino acids to be encoded by up to six synonymous codons. Genomes of different organisms prefer different synonymous codons, a phenomenon known as ‘codon usage bias.’ The phenomenon of codon usage bias is of fundamental interest for evolutionary biology, and is important in a variety of applied settings (e.g., transgene expression). The spectrum of codon usage biases seen in current organisms is commonly thought to have arisen by the combined actions of mutations and selective pressures. This view focuses on codon usage in specific genomes and the consequences of that usage for protein expression.Here we investigate an unresolved question of molecular genetics: are there global rules governing the usage of synonymous codons made by genomic DNA across organisms? To answer this question, we employed a data-driven approach to surveying 2730 species from all kingdoms of the ‘tree of life’ in order to classify their codon usage. A first major result was that the large majority of these organisms use codons rather uniformly on the genome-wide scale, without giving preference to particular codons among possible synonymous alternatives. A second major result was that two compartments of codon usage seem to co-exist and to be expressed in different proportions by different organisms. As such, we investigate how individual different codons are used in different organisms from all taxa. Whereas codon usage is generally believed to be the evolutionary result of both mutations and natural selection, our results suggest a different perspective: the usage of different codons (and amino acids) by different organisms follows a superposition of two distinct patterns of usage. One distinction locates to the third base pair of all different codons, which in one pattern is U or A, and in the other pattern is G or C. This result has two major implications: (1) the variation of codon usage as seen across different organisms is best accounted for by lateral gene transfer among diverse organisms; (2) the organisms that are by protein homology grouped near the base of the ‘tree of life’ comprise two genetically distinct lineages.We find that, over evolutionary time, codon usages have converged from two distinct, non-overlapping usages (e.g., as evident in bacteria and archaea) to a near-uniform, balanced usage of synonymous codons (e.g., in mammals). This shows that the variations of codon (and amino acid) biases reveal a distinct evolutionary progression. We also find that codon usage in bacteria and archaea is most diverse between organisms thought to be closest to the hypothesized last universal common ancestor (LUCA). The dichotomy in codon (and amino acid usages) present near the origin of the current ‘tree of life’ might provide information about the evolutionary development of the genetic code.