Over-representation of repeats in stress response genes: a strategy to increase versatility under stressful conditions?

Abstract

The survival of individual organisms facing stress is enhanced by the induction of a set of changes. As the intensity, duration and nature of stress is highly variable, the optimal response to stress may be unpredictable. To face such an uncertain future, it may be advantageous for a clonal population to increase its phenotypic heterogeneity (bet-hedging), ensuring that at least a subset of cells would survive the current stress. With current techniques, assessing the extent of this variability experimentally remains a challenge. Here, we use a bioinformatic approach to compare stress response genes with the rest of the genome for the presence of various kinds of repeated sequences, elements known to increase variability during the transfer of genetic information (i.e. during replication, but also during gene expression). We investigated the potential for illegitimate and homologous recombination of 296 Escherichia coli genes related to repair, recombination and physiological adaptations to different stresses. Although long repeats capable of engaging in homologous recombination are almost absent in stress response genes, we observed a significant high number of short close repeats capable of inducing phenotypic variability by slipped-mispair during DNA, RNA or protein synthesis.

Figure 1

Scenarios of illegitimate recombination between close direct repeats and SSRs. Black boxes represent start and stop codons of the original gene, grey boxes represent strict repeats, and light grey boxes represent regions of weaker homology. Dashed lines indicate that deletions and duplication may induce frameshifts and therefore produce ORFs of very different length. (A and B) Duplication/deletion of the repeat and the region between occurrences. (C) The regions of non-strict similarity become similar after conversion. (D and E) Increase/decrease in the number of motifs of the SSR. Homologous recombination between long repeats closely follows the scenarios of (A), (B) and (C), except that large duplications are unstable and large deletions are strongly counter selected. Thus, conversions or reciprocal translocations are the most frequent outcome of homologous recombination between long distant repeats.

Figure 2

Histogram of the cumulated length of repeats present in E.coli genes, divided by the length of genes (density of repeat bases). Only 8% of the genes contain repeats, and most of them contain densities of repeats smaller than 0.5. The repeats concerning the stress response genes are drawn in black (they are all in the first bin of the histogram).

Figure 3

Distribution of observed/expected values for SSR densities for k-tuples from 1 to 4, in the stress subset (bars) and in remaining genes of the genome (lines). Expected values were calculated using the observed number of SSRs in 1000 random sequences of equal length and equal frequency of the motif (see Materials and Methods).

Figure 4

Distribution of the average number of close repeats in the stress subset (bars) and in remaining genes (lines).

Figure 5

Spatial distribution of close direct repeats in E.coli K12 top 10 genes + mutS. Black boxes represent repeats and grey boxes represent genes. Different occurrences of the same repeat are marked with the same letter.