Proteome composition and codon usage in spirochaetes: species-specific andDNA strand-specific mutational biases

The genomes of the spirochaetes Borrelia burgdorferi and Treponema pallidum show strong strand-specific skews in nucleotide composition, with the lead ing strand in replication being richer in G and T than the lagging strand i n both species. This mutation bias results in codon usage and amino acid co mposition patterns that are significantly different between genes encoded o n the two strands, in both species. There are also substantial differences between the species, with T. pallidum having a much higher G+C content than B. burgdorferi. These changes in amino acid and codon compositions represe nt neutral sequence change that has been caused by strong strand- and speci es-specific mutation pressures. Genes that have been relocated between the leading and lagging strands since B. burgdorferi and T. pallidum diverged f rom a common ancestor now show codon and amino acid compositions typical of their current locations. There is no evidence that translational selection operates on codon usage in highly expressed genes in these species, and th e primary influence on codon usage is whether a gene is transcribed in the same direction as replication, or opposite to it. The dnaA gene in both spe cies has codon usage patterns distinctive of a lagging strand gene, indicat ing that the origin of replication lies downstream of this gene, possibly w ithin dnaN. Our findings strongly suggest that gene-finding algorithms that ignore variability within the genome may be flawed.