Mosaic prophages with horizontally acquired genes account for the emergence and diversification of the globally disseminated M1T1 clone of Streptococcus pyogenes

Abstract

The recrudescence of severe invasive group A streptococcal (GAS) diseases has been associated with relatively few strains, including the M1T1 subclone that has shown an unprecedented global spread and prevalence and high virulence in susceptible hosts. To understand its unusual epidemiology, we aimed to identify unique genomic features that differentiate it from the fully sequenced M1 SF370 strain. We constructed DNA microarrays from an M1T1 shotgun library and, using differential hybridization, we found that both M1 strains are 95% identical and that the 5% unique M1T1 clone sequences more closely resemble sequences found in the M3 strain, which is also associated with severe disease. Careful analysis of these unique sequences revealed three unique prophages that we named M1T1.X, M1T1.Y, and M1T1.Z. While M1T1.Y is similar to phage 370.3 of the M1-SF370 strain, M1T1.X and M1T1.Z are novel and encode the toxins SpeA2 and Sda1, respectively. The genomes of these prophages are highly mosaic, with different segments being related to distinct streptococcal phages, suggesting that GAS phages continue to exchange genetic material. Bioinformatic and phylogenetic analyses revealed a highly conserved open reading frame (ORF) adjacent to the toxins in 18 of the 21 toxin-carrying GAS prophages. We named this ORF paratox, determined its allelic distribution among different phages, and found linkage disequilibrium between particular paratox alleles and specific toxin genes, suggesting that they may move as a single cassette. Based on the conservation of paratox and other genes flanking the toxins, we propose a recombination-based model for toxin dissemination among prophages. We also provide evidence that a minor population of the M1T1 clonal isolates have exchanged their virulence module on phage M1T1.Y, replacing it with a different module identical to that found on a related M3 phage. Taken together, the data demonstrate that mosaicism of the GAS prophages has contributed to the emergence and diversification of the M1T1 subclone.

FIG. 1.

Summary of differential hybridization results. Distribution of best BLASTN hits for the sequences that hybridized preferentially with M1T1 but not with SF370 DNA (raw data are provided in Table S1).

FIG. 2.

Mosaic nature of M1T1 prophages. The diagram shows the patterns and extent of similarity between different segments of SPhinX (A), MemPhiS (B), and PhiRamid (C) and their closest homologs among GAS prophages. Best BLASTN hits are shown below or above each phage segment, and—in some cases—the percentage of nucleotide identity is indicated. White boxes represent sequences with no BLASTN hits.

FIG. 3.

PCRs showing phage excision and integration. (A, upper part) Map of the different genes in the lysis and the lysogeny modules of GAS prophages. (A, lower part) Map of the same genes' relative positions when the phage is circularized. Genes are not drawn to scale. Positions of PCR primers are shown by small black arrows (a, b, c, and d). (B) PCRs show the presence of each phage (SPhinX, MemPhiS, and PhiRamid) in both attached and excised forms. All PCR products were sequenced and their sequences validated.

FIG. 4.

Paratox: a highly conserved ORF in M1T1 prophages. (A) A comparison between the lysogenic conversion modules and attachment sites of the three M1T1 prophages shows a highly conserved ORF that best matches a hypothetical phage protein located between each toxin and the phage attachment site. We named this hypothetical protein paratox (prx). Shaded areas indicate nucleotide similarity, and the percentage nucleotide identity is given. (B) Alignment of paratox protein alleles shows highly conserved amino acid sequence (represented by dots). Representative motifs linked to particular toxins or to phage attachment sites are boxed. All sequences are extracted from GenBank; in cases where the prx sequences were not annotated as ORFs, we picked them based on their similarity to the annotated ones. Each Prx will be referred to as (Prx_tox_Phi#), where tox is the name of the adjacent toxin and Phi# is the phage name and number (e.g., Prx_SpeA2_M1T1.X is the product of the paratox gene adjacent to SpeA2 in Phi M1T1.X, alias SPhinX). Serial numbers (1 to 11) were given to the distinct paratox alleles shown.

FIG. 5.

Putative model for toxin exchange between phages. Possible scenarios that may contribute to toxin exchange between different prophages by recombination are shown. Two recombination hot spots are shown on both sides of the toxin genes: one of them is the prx gene, and the other may be either lys, hol, or hylP.