Identification and Characterization of Domesticated Bacterial Transposases

Abstract

Selfish genetic elements, such as insertion sequences and transposons are found in most genomes. Transposons are usually identifiable by their high copy number within genomes. In contrast, REP-associated tyrosine transposases (RAYTs), a recently described class of bacterial transposase, are typically present at just one copy per genome. This suggests that RAYTs no longer copy themselves and thus they no longer function as a typical transposase. Motivated by this possibility we interrogated thousands of fully sequenced bacterial genomes in order to determine patterns of RAYT diversity, their distribution across chromosomes and accessory elements, and rate of duplication. RAYTs encompass exceptional diversity and are divisible into at least five distinct groups. They possess features more similar to housekeeping genes than insertion sequences, are predominantly vertically transmitted and have persisted through evolutionary time to the point where they are now found in 24% of all species for which at least one fully sequenced genome is available. Overall, the genomic distribution of RAYTs suggests that they have been coopted by host genomes to perform a function that benefits the host cell.

Keywords: IS200; RAYTs; REPINs; domesticated transposons; selfish genetic elements.

Fig. 1.—

Family size of ISs, RAYTs, and housekeeping genes.

Fig. 2.—

RAYT characteristics are more similar to those of housekeeping genes than IS elements. (A) Average copy number of each gene family per replicon (excluding replicons in which no family members occur). (B) Proportion of genes for which at least one duplicate (pairwise nucleotide identity of ≥95%) was found on the same replicon. (C) Proportion of family members found on plasmids. (D) Average frequency of the most abundant 16mer identified in the flanking 5′ and 3′ extragenic regions of each family member. All error bars show standard errors.

Fig. 3.—

Protein similarity network and genomic characteristics show significant divergence of RAYT groups. (A) RAYT protein relationship map constructed by Cytoscape using the organic layout (Cline etal. 2007). RAYT groups identified by MCL are highlighted according to the legend. Nodes represent RAYT proteins. Nodes are connected if the pairwise amino acid identity is ≥26%. Differences between the Cytoscape protein clustering (proximity of nodes) and the MCL algorithm (colour of nodes) are due to the difference in the information provided to MCL and Cytoscape: Cytoscape only receives information on whether or not nodes are connected, whereas MCL performs coloration based on the exact pairwise similarity between proteins. Graphs (B) to (F) contain the same information as figures 1 and 2, but for the six largest RAYT groups in (A) as opposed to the five gene families.

Fig. 4.—

Protein similarity network and genomic distribution of IS200 groups. (A) shows a protein similarity network of the IS200 family. The map was produced as described in the legend for figure 3A. Graphs (B) to (F) show the same genomic characteristics as in figures 2 and 3.

Fig. 5.—

RAYTs and IS200 gene families form monophyletic groups. (A) Phylogenetic tree. From each IS200 and RAYT group, three random proteins were selected and aligned with MUSCLE (Edgar 2004). We excluded IS200 Group 6, because the alignments could not be calculated correctly when members of this subgroup were included. The phylogeny was built using PHYML (Guindon and Gascuel 2003). All RAYTs are found in the same clade except for RAYT Group 6, which clusters within the IS200 family. This is plausible considering the IS-like genome distribution characteristics. Green bars indicate low duplication rates (<5%); red bars indicate high duplication rate and high presence on plasmids (>17%). RAYT Group 1 is marked with a red/green bar as within the RAYT family it has the highest duplication rate (6.2%). (B) Protein similarity network of the RAYT (left) and IS200 (right) gene family. We chose 30 random proteins from the six largest RAYT and IS200 groups and displayed them with Cytoscape as in figures 3F and 4F. RAYT Group 6 clusters within IS200 Group 1 in line with the phylogeny in (A) and the genomic distribution characteristics in figure 3.

Fig. 6.—

Comparison between RAYT phylogenies and housekeeping gene phylogenies shows vertical RAYT inheritance. (A) Shows the RAYT Group 2 phylogeny and its corresponding housekeeping gene tree of all RAYT Group 2 members that are found in the Enterobacteriaceae. Escherichia, Salmonella, Klebsiella, and Enterobacter cloacae are coloured in red, green, blue, and purple, respectively. The remaining genera are in black and their corresponding positions in the housekeeping gene tree are connected with black lines. The most recent common ancestor is estimated to have lived about 150 Ma from a 16S rDNA tree (Ochman and Wilson 1987). RAYT Group 2 genes occur mostly as single copy genes. (B) Shows the RAYT Group 3 phylogeny and its corresponding housekeeping gene tree for all RAYT Group 3 members that are found in Pseudomonas fluorescens, P. putida and P. syringae. They are coloured in red, green, and blue, respectively. The most recent common ancestor is estimated to have lived about 100 Ma from a 16S rDNA tree (Ochman and Wilson 1987). Small trees in between the housekeeping gene tree and the RAYT tree are corresponding housekeeping trees for the RAYT subtrees. The strain name in black indicates the outgroup. RAYT Group 3 genes occur often in multiple copies per genome. Their evolution is dominated by gene loss and duplication events.