Non-Random Inversion Landscapes in Prokaryotic Genomes Are Shaped by Heterogeneous Selection Pressures.

Jelena Repar,Tobias Warnecke

doi:10.1093/molbev/msx127

Abstract

Inversions are a major contributor to structural genome evolution in prokaryotes. Here, using a novel alignment-based method, we systematically compare 1,651 bacterial and 98 archaeal genomes to show that inversion landscapes are frequently biased toward (symmetric) inversions around the origin–terminus axis. However, symmetric inversion bias is not a universal feature of prokaryotic genome evolution but varies considerably across clades. At the extremes, inversion landscapes in Bacillus–Clostridium and Actinobacteria are dominated by symmetric inversions, while there is little or no systematic bias favoring symmetric rearrangements in archaea with a single origin of replication. Within clades, we find strong but clade-specific relationships between symmetric inversion bias and different features of adaptive genome architecture, including the distance of essential genes to the origin of replication and the preferential localization of genes on the leading strand. We suggest that heterogeneous selection pressures have converged to produce similar patterns of structural genome evolution across prokaryotes.

Highlights

In both eukaryotes and prokaryotes, genome architecture and its evolution are frequently non-random (Hurst et al 2004; Rocha 2008)
Starting from MUMmer (Kurtz et al 2004) alignments of two genomes, like the one shown in figure 1B, we make use of the fact that, if there is a single dominant axis around which inversions have occurred between two genomes, homologous sequence blocks will be located on one of the two diagonals that pass through that axis, generating the familiar X pattern
We find strong symmetric inversion bias in multiple bacterial phyla, including those that have shaped our understanding of replication-associated genome architecture—Proteobacteria and Bacillus–Clostridium—in line with previous observations from individual genome pairs and a more systematic comparison of recently diverged genomes (Khedkar and Seshasayee 2016)

Summary

Introduction

In both eukaryotes and prokaryotes, genome architecture and its evolution are frequently non-random (Hurst et al 2004; Rocha 2008). We go on to show that putatively adaptive features of genome architecture linked to the ori-ter axis, such as the fraction of genes encoded on the leading strand and the average distance of rRNA genes to the origin of replication, are predictive of symmetric inversion bias but in a cladespecific fashion.

Results

Conclusion