Homologous recombination between tandem paralogues drives evolution of a subset of type VII secretion system immunity genes in firmicute bacteria

Abstract

The type VII secretion system (T7SS) is found in many Gram-positive firmicutes and secretes protein toxins that mediate bacterial antagonism. Two T7SS toxins have been identified in Staphylococcus aureus, EsaD a nuclease toxin that is counteracted by the EsaG immunity protein, and TspA, which has membrane depolarising activity and is neutralised by TsaI. Both toxins are polymorphic, and strings of non-identical esaG and tsaI immunity genes are encoded in all S. aureus strains. To investigate the evolution of esaG repertoires, we analysed the sequences of the tandem esaG genes and their encoded proteins. We identified three blocks of high sequence similarity shared by all esaG genes and identified evidence of extensive recombination events between esaG paralogues facilitated through these conserved sequence blocks. Recombination between these blocks accounts for loss and expansion of esaG genes in S. aureus genomes and we identified evidence of such events during evolution of strains in clonal complex 8. TipC, an immunity protein for the TelC lipid II phosphatase toxin secreted by the streptococcal T7SS, is also encoded by multiple gene paralogues. Two blocks of high sequence similarity locate to the 5' and 3' end of tipC genes, and we found strong evidence for recombination between tipC paralogues encoded by Streptococcus mitis BCC08. By contrast, we found only a single homology block across tsaI genes, and little evidence for intergenic recombination within this gene family. We conclude that homologous recombination is one of the drivers for the evolution of T7SS immunity gene clusters.

Keywords: Staphylococcus aureus; Streptococcus; T7SS; homologous recombination; immunity gene families.

Fig. 1.

Homologues of esaG encoded at the ess locus in RN6390. a. Illustration of the 3′ of the RN6390 ess locus which encodes the T7SS nuclease toxin, EsaD, its cognate immunity protein, EsaG and 11 further homologues of EsaG (numbered esaG2 – esaG12). Note that esaG4 is shown in hatched shading because it is annotated as a pseudogene. However, it does encode two predicted ORFs, EsaG4i and EsaG4ii. b. Sequence alignment of EsaG homologues encoded by RN6390. The black boxes represent regions of high sequence similarity based on this alignment.

Fig. 2.

Recombination within the RN6390 esaG homologues. a. Regions of high sequence similarity across the RN6390 EsaG protein sequences (middle panel) and the corresponding nucleotide sequences (bottom panel). The positions of homology blocks are shown in grey shading, with their relative positions along the gene sequence (top panel). The basepair positions that define the conserved regions are taken from the nucleotide sequences of esaG1. b. RDP4 was used to predict recombination events within the esaG homologues encoded by RN6390. The identity of each gene is given in black at the left, with regions of recombination labelled directly below, in the colour of the gene from which the recombinant section originated. ci-iii. For each of the three variable regions of the RN6390 esaG homologues a maximum likelihood tree was generated in IQTREE and visualised and annotated in iTOL. The variable regions consist of nucleotide 1–39, 96–251 and 358–465 of esaG respectively d. Illustration of regions of high sequence similarity in the esaG homologues in RN6390. Black bars represent conserved regions of the gene and the variable regions have been assigned a colour and corresponding number. Homologous regions are coloured with the same colour. Numbers were assigned based on the first gene in the series that had the unique variable region. White hatched shading indicates a pseudogene.

Fig. 3.

A recombination event in an epidemic lineage of USA300 results in loss of part of an esaG cluster and generation of a novel esaG gene. a. The esaD locus of USA300 FPR3757 and USA300 BKV_2. The dashed lines represent the region that is missing from the epidemic strain, USA300 BKV_2, when compared to the USA300 FPR3757 type strain. b. Nucleotide sequence alignment for SAUSA300_0295 and SAUSA300_0299 from USA300 FPR3757 and esaG4 from USA300 BKV_2. Coloured blocks indicated homology between BKV_2 esaG4 and the genes with which it is aligned.

Fig. 4.

Recombination between esaG paralogs results in variation at the ess locus in CC8 S. aureus strains from a historical culture collection. Two expansion events of the esaG repertoire in the NCTC CC8 dataset were identified in a. NCTC10724 and b. NCTC10702 as respective representatives. The solid black line depicts the predicted recombination event to give the recombinant sequence depicted below. Five unique loss events were also identified in this dataset, represented by strains c. NCTC9369, d. NCTC10652, e. NCTC12232, f. NCTC13140 and g. NCTC13196. The dashed line represents the region deleted in the recombinant. Genes were given an individual colour and sections were numbered based on esaG1-12, to make clear any recombination events.

Fig. 5.

Recombination also occurs between DUF5079 genes. Recombination events were detected between the two DUF5079 genes at the ess locus in the NCTC S. aureus CC8 dataset. Recombination occurred a. within the DUF5079 genes and b. in the intergenic regions directly upstream of these, resulting in the loss of esaG2-5 in several strains. c. An alignment of the region spanning the 5′ of esaG1 to the 3′ of DUF5079-1 with the 5′ of esaG5 to the 3′ of DUF5079-2. Coloured bars below the alignment represent esaG (orange), the intergenic region (white) and DUF5079 (blue). d. All instances of the region spanning the 5′ of esaG to the 3′ of DUF5079 were manually extracted from the NCTC CC8 dataset, aligned using MAFFT and visualised using Plotcon to identify regions of high similarity.

Fig. 6.

Homologues of tipC encoded at the telC locus of Streptococcus mitis BCC08. a. Genetic arrangement of tipC genes in S. mitis BCC08. b. An alignment of the encoded TipC homologues. The blue boxes represent regions of high sequence similarity, and the dashed line at the N-terminus of the aligned sequences indicates a probable lipoprotein signal peptide.

Fig. 7.

Recombination within the S. mitis BCC08 tipC homologues. a. Regions of high sequence similarity across the S. mitis BCC08 TipC protein sequences (middle panel) and the corresponding nucleotide sequences (bottom panel). The positions of homology blocks are shown in grey shading, with their relative positions along the gene sequence (top). The first 18 amino acids of TipC form a predicted lipoprotein signal sequence which is indicated by pale grey shading. The basepair positions that define the conserved regions are taken from the nucleotide sequences of D8786_RS05940. b. RDP4 was used to predict recombination events within the tipC homologues. The identity of each gene is given in black at the left, with regions of recombination labelled directly below, in the colour of the gene from which the recombinant section originated. c. A maximum likelihood tree was generated for tipC homologues in IQTREE and visualised and annotated in iTOL. d. The conserved (cyan) and variable (orange) regions of TipC were mapped to the crystal structure of S. intermedius TipC2 (pdb:6DHX; [36]).