Accelerated gene evolution through replication–transcription conflicts

Bacteria promote faster evolution of individual genes through orientation-dependent encounters between DNA replication and transcription. Head-on collisions drive gene evolution Most bacterial genes are encoded on the leading strand of replication, potentially avoiding detrimental head-on collisions that can occur between the replication and transcription machineries when genes are encoded on the lagging strand. Here the authors report that 17% of core genes in Bacillus subtilis lie on the lagging strand, and that they have a higher rate of point mutations and of convergent changes at the same amino acid position than those on the leading strand. Taking into account gene length and gene expression levels, the authors propose that head-on replication–transcription conflicts are more mutagenic than co-directional conflicts and that bacteria — and perhaps other organisms — may use head-on replication–transcription conflicts as a mechanism for targeted adaptive evolution of certain genes. Several mechanisms that increase the rate of mutagenesis across the entire genome have been identified; however, how the rate of evolution might be promoted in individual genes is unclear. Most genes in bacteria are encoded on the leading strand of replication 1 , 2 , 3 , 4 . This presumably avoids the potentially detrimental head-on collisions that occur between the replication and transcription machineries when genes are encoded on the lagging strand 1 , 2 , 3 , 4 . Here we identify the ubiquitous (core) genes in Bacillus subtilis and determine that 17% of them are on the lagging strand. We find a higher rate of point mutations in the core genes on the lagging strand compared with those on the leading strand, with this difference being primarily in the amino-acid-changing (nonsynonymous) mutations. We determine that, overall, the genes under strong negative selection against amino-acid-changing mutations tend to be on the leading strand, co-oriented with replication. In contrast, on the basis of the rate of convergent mutations, genes under positive selection for amino-acid-changing mutations are more commonly found on the lagging strand, indicating faster adaptive evolution in many genes in the head-on orientation. Increased gene length and gene expression amounts are positively correlated with the rate of accumulation of nonsynonymous mutations in the head-on genes, suggesting that the conflict between replication and transcription could be a driving force behind these mu