Evolution of differential codon usage preferences and subfunctionalisation in paralogous genes: the showcase of polypyrimidine tract binding proteins.
AO_SCPLOWBSTRACTC_SCPLOWGene paralogs are copies of an ancestral gene that appear after gene or full genome duplication. When two sister gene copies are maintained in the genome, redundancy may release certain evolutionary pressures, allowing one of them to access novel functions. Here, we focused our study on gene paralogs on the evolutionary history of the three polypyrimidine tract binding protein genes (PTBP) and their concurrent evolution of differential codon usage preferences (CUPrefs) in vertebrate species. PTBP1-3 show high identity at the amino acid level (up to 80%), but display strongly different nucleotide composition, divergent CUPrefs and, in humans, distinct tissue-specific expression levels. Our phylogenetic inference results show that the duplication events leading to the three extant PTBP1-3 lineages predate the basal diversification within vertebrates, and genomic context analysis illustrates that synteny has been well preserved over time for the three paralogs. We identify a distinct evolutionary pattern towards GC3-enriching substitutions in PTBP1, concurrent with an enrichment in frequently used codons and with a tissue-wide expression. In contrast, PTBP2s are enriched in AT-ending, rare codons, and display tissue-restricted expression. As a result of this substitution trend, CUPrefs are sharply different between mammalian PTBP1s and the rest of PTBPs. Genomic context analysis shows that GC3-rich nucleotide composition in PTBP1s is driven by local substitution processes, while the evidence in this direction is thinner for PTBP2-3. An actual lack of co-variation between the observed GC composition of PTBP2-3 and that of the surrounding non-coding genomic environment would raise an interrogation on the origin of CUPrefs, warrantying further research on a putative tissue-specific translational selection. Finally, we communicate an intriguing trend for the use of the UUG-Leu codon, which matches the trends of AT-ending codons. We interpret that our results are compatible with an scenario in which a combination of directional mutation-selection processes would have differentially shaped CUPrefs of PTBPs in Vertebrates: the observed GC-enrichment of PTBP1 in Mammals may be linked to genomic location and to the strong and broad tissue-expression, while AT-enrichment of PTBP2 and PTBP3 would be associated with rare CUPrefs and thus, possibly to specialized spatio-temporal expression. Our interpretation is coherent with a gene subfunctionalisation process by differential expression regulation associated to the evolution of specific CUPrefs. 1 Significance StatementIn vertebrates, PTBP paralogs display strong differences in gene composition, gene expression regulation, and their expression in cell culture depends on their codon usage preferences. We show that placental mammals PTBP1 have become GC-rich because of local substitution pressures, resulting in an enrichment of frequently used codons and in a strong, tissue-wide expression. On the contrary, PTBP2 in vertebrates are AT-rich, with a lower contribution of local substitution processes to their specific nucleotide composition, show high frequency of rare codons and in placental mammals display a restricted expression pattern contrasting to that of PTBP1. The systematic study of composition and expression patterns of gene paralogs can help understand the complex mutation-selection interplay that shape codon usage bias in multicellular organisms.