Evolutionary effects of translocations in bacterial genomes

Abstract

It has become clear that different genome regions need not evolve uniformly. This variation is particularly evident in bacterial genomes with multiple chromosomes, in which smaller, secondary chromosomes evolve more rapidly. We previously demonstrated that substitution rates and gene dispensability were greater on secondary chromosomes in many bacterial genomes. In Vibrio, the secondary chromosome is replicated later during the cell cycle, which reduces the effective dosage of these genes and hence their expression. More rapid evolution of secondary chromosomes may therefore reflect weaker purifying selection on less expressed genes. Here, we test this hypothesis by relating substitution rates of orthologs shared by multiple Burkholderia genomes, each with three chromosomes, to a study of gene expression in genomes differing by a major reciprocal translocation. This model predicts that expression should be greatest on chromosome 1 (the largest) and least on chromosome 3 (the smallest) and that expression should tend to decline within chromosomes from replication origin to terminus. Moreover, gene movement to the primary chromosome should associate with increased expression, and movement to secondary chromosomes should result in reduced expression. Our analysis supports each of these predictions, as translocated genes tended to shift expression toward their new chromosome neighbors despite inevitable cis-acting regulation of expression. This study sheds light on the early dynamics of genomes following rearrangement and illustrates how secondary chromosomes in bacteria may become evolutionary test beds.

Fig. 1.—

Relationships between substitution rates (A: dN and B: dS) of Burkholderia orthologs and gene expression measured in strain HI2424. Both values were cube root transformed. The overall regressions are significant (dN vs. expression: F = 384, dfs = 1, 4,079, P < 10⁻¹⁶, r² = 0.086; dS vs. expression F = 553, dfs = 1, 4,079, P < 10⁻¹⁶, r² = 0.120), and all pairwise comparisons among chromosomes (boxplots) differ significantly (expression: chromosome 1 (c1) to chromosome 2 (c2), t = −14.2, df = 2,885, P < 2.2 × 10⁻¹⁶; c2 to c3, t = −6.44, df = 261, P = 5.75 × 10⁻¹⁰; dN: c1 to c2, t = 11.12, df = 3,556, P < 2.2 × 10⁻¹⁶; c2 to c3, t = 3.51, df = 301, P = 0.000516; dS: c1 to c2, t = −13.72, df = 3,235, P < 2.2 × 10⁻¹⁶; c2 to c3, t = −6.56, df = 276, P = 2.711 × 10⁻¹⁰).

Fig. 2.—

Altered expression of translocated genes in Burkholderia cenocepacia matching their new chromosome neighbors. (A) Schematic of the translocation found in strain AU1054 in comparison to its nearest relative HI2424, which represents the consensus genome order of the species complex; dark red and dark blue represent the translocated genomic blocks. (B) Boxplots of expression ratios (AU1054/HI2424) of shared orthologs. The translocation from chromosome 1 (517 genes) was compared with nearby orthologs—the chromosome half nearer the terminus (1,056 genes)—because their expression in HI2424 better represents the translocated fraction. The resident (523 genes) and translocated (371 genes) regions of chromosome 3 in AU1054 exhibit equivalent expression in HI2424. Two sample t-tests were used; for chromosome 1, t = −2.338, df = 1,571, P = 0.0195, for chromosome 3, t = 2.974, df = 892, P = 0.0030.