Adaptive evolution of chloroplast genome structure inferred using a parametric bootstrap approach

Background: Genome rearrangements influence gene order and configuration of gene clusters in all genomes. Most land plant chloroplast DNAs (cpDNAs) share a highly conserved gene content and with notable exceptions, a largely co-linear gene order. Conserved gene orders may reflect a slow intrinsic rate of neutral chromosomal rearrangements, or selective constraint. It is unknown to what extent observed changes in gene order are random or adaptive. We investigate the influence of natural selection on gene order in association with increased rate of chromosomal rearrangement. We use a novel parametric bootstrap approach to test if directional selection is responsible for the clustering of functionally related genes observed in the highly rearranged chloroplast genome of the unicellular green alga Chlamydomonas reinhardtii, relative to ancestral chloroplast genomes. Results: Ancestral gene orders were inferred and then subjected to simulated rearrangement events under the random breakage model with varying ratios of inversions and transpositions. We found that adjacent chloroplast genes in C. reinhardtii were located on the same strand much more frequently than in simulated genomes that were generated under a random rearrangement processes ( increased sidedness; p < 0.0001). In addition, functionally related genes were found to be more clustered than those evolved under random rearrangements ( p < 0.0001). We report evidence of co-transcription of neighboring genes, which may be responsible for the observed gene clusters in C. reinhardtii cpDNA. Conclusion: Simulations and experimental evidence suggest that both selective maintenance and directional selection for gene clusters are determinants of chloroplast gene order.