Codon Usage Heterogeneity in the Multipartite Prokaryote Genome: Selection-Based Coding Bias Associated with Gene Location, Expression Level, and Ancestry

Prokaryotes represent an ancestral lineage in the tree of life and constitute optimal resources for investigating the evolution of genomes in unicellular organisms. Many bacterial species possess multipartite genomes offering opportunities to study functional variations among replicons, how and where new genes integrate into a genome, and how genetic information within a lineage becomes encoded and evolves. To analyze these issues, we focused on the model soil bacterium Sinorhizobium meliloti, which harbors a chromosome, a chromid (pSymB), a megaplasmid (pSymA), and, in many strains, one or more accessory plasmids. The analysis of several genomes, together with 1.4 Mb of accessory plasmid DNA that we purified and sequenced, revealed clearly different functional profiles associated with each genomic entity. pSymA, in particular, exhibited remarkable interstrain variation and a high density of singletons (unique, exclusive genes) featuring functionalities and modal codon usages that were very similar to those of the plasmidome. All this evidence reinforces the idea of a close relationship between pSymA and the plasmidome. Correspondence analyses revealed that adaptation of codon usages to the translational machinery increased from plasmidome to pSymA to pSymB to chromosome, corresponding as such to the ancestry of each replicon in the lineage. We demonstrated that chromosomal core genes gradually adapted to the translational machinery, reminiscent of observations in several bacterial taxa for genes with high expression levels. Such findings indicate a previously undiscovered codon usage adaptation associated with the chromosomal core information that likely operates to improve bacterial fitness. We present a comprehensive model illustrating the central findings described here, discussed in the context of the changes occurring during the evolution of a multipartite prokaryote genome.IMPORTANCE Bacterial genomes usually include many thousands of genes which are expressed with diverse spatial-temporal patterns and intensities. A well-known evidence is that highly expressed genes, such as the ribosomal and other translation-related proteins (RTRPs), have accommodated their codon usage to optimize translation efficiency and accuracy. Using a bioinformatic approach, we identify core-genes sets with different ancestries, and demonstrate that selection processes that optimize codon usage are not restricted to RTRPs but extended at a genome-wide scale. Such findings highlight, for the first time, a previously undiscovered adaptation strategy associated with the chromosomal-core information. Contrasted with the translationally more adapted genes, singletons (i.e., exclusive genes, including those of the plasmidome) appear as the gene pool with the less-ameliorated codon usage in the lineage. A comprehensive summary describing the inter- and intra-replicon heterogeneity of codon usages in a complex prokaryote genome is presented.