Identification of a chimeric emm gene and novel emm pattern in currently circulating strains of emm4 Group A Streptococcus.

Group A Streptococcus (GAS) is classified on the basis of the sequence of the gene encoding the M protein (emm) and the patterns into which emm types are grouped. We discovered a novel emm pattern in emm4 GAS, historically considered pattern E, arising from a fusion event between emm and the adjacent enn gene. We identified the emm-enn fusion event in 51 out of 52 emm4 GAS strains isolated by national surveillance in 2015. GAS isolates with an emm-enn fusion event completely replaced pattern E emm4 strains over a 4-year span in Houston (2013-2017). The novel emm-enn gene fusion and new emm pattern has potential vaccine implications.

Data Summary

1. The Geneious software is available for purchase at https://www.geneious.com/.

2. The programs used to analyze single-nucleotide polymorphisms (SNP) and reconstruct whole-genome-sequencing-based phylogenies are publicly available at (a) kSNP v3.0 - https://omictools.com/ksnp-tool, (b) parsnp v1.2 - http://harvest.readthedocs.io/en/latest/content/parsnp.html and (c) FastTree 2 - http://www.microbesonline.org/fasttree/.

3. BioProject numbers for all reference genomes as well as the CDC collection of genomes used in the analyses are listed in the Data Bibliography section.

4. The short-read sequences associated with Chalker et al. can be found in the European Nucleotide Archive and are listed in the Data Bibliography section.

Group A streptococci (GAS) are a major cause of serious infections in humans, ranging from simple pharyngitis (strep throat) to life-threatening diseases, such as necrotizing fasciitis (the flesh-eating disease). One of the critical virulence factors of GAS is the hypervariable M protein, which is the primary target for the human immune system and a leading vaccine candidate. GAS strains are well known to slightly vary the sequence of the emm gene and hence the M protein in order to establish infections in humans. In this study, we characterize a novel emm gene which has resulted from fusion of the existing emm with its neighboring enn gene, theoretically resulting in a completely new M protein rather than the minor variations previously described. We found that GAS strains causing both invasive and superficial infections that carry this new emm gene have been replacing strains with the old emm version. Collectively, the outcomes of this study provide a framework to further investigate the contribution of the new emm gene to the emergent dominance of the new emm genotype in human populations.

Introduction

Group A Streptococcus (GAS) is among the most ubiquitous human pathogens and is classified into emm types based on the 5′ sequence of the emm gene that encodes the hypervariable M protein, a key GAS virulence determinant [1]. GAS infection is thought to engender serotype-specific immunity, and the M protein or its domains are leading vaccine candidates [2]. In addition to invariably containing the emm gene, the emm region can also contain one or two emm-like genes, typically designated as mrp and enn (Fig. 1a) [3]. The 3′ end of the emm and emm-like genes encode a peptidoglycan–spanning (PG) domain, and PG sequence variation determines four subfamilies (SF-1 through SF-4, Fig. 1a). The composition of emm-family genes along with their respective PG domain subfamily types divides GAS strains into five emm patterns (A–E), and for the overwhelming majority of GAS strains studied to date, there is strict concordance between emm type and emm pattern [3]. The emm pattern of GAS strains strongly correlates with the preferred epithelial site of infection i.e. pharynx versus skin [3]. Thus, emm pattern A–C strains are considered ‘throat specialists’, whereas pattern D are ‘skin specialists’, and pattern E are ‘generalists’ [4].

Fig. 1.

Characterization of GAS isolates. (a) Schematic diagram to compare the new emm pattern with existing patterns (adapted from reference [3]). The emm type-specific (M) determinant and the subfamily (SF) of the PG domain is indicated. (b) Diagram showing predicted binding sites for IgA, IgG, C4BP and albumin based on motifs [1] present in the deduced sequence of the M protein from MGAS10750 and SSGAS001. (c) Maximum-likelihood tree reconstructed using 814 core SNPs of emm4 GAS strains from CDC [11] along with publicly available emm4 GAS genomes (names are provided). The local support values are provided. The type of M protein is color coded as indicated in legend.

The majority of data establishing the five emm patterns and their relationship to emm type were generated from strains isolated in the 1980s and 1990s using targeted approaches like PCR and Southern blotting [4, 5]. Initial whole-genome sequencing (WGS) of one or a few GAS strains per emm type done in the 2000s supported these findings [6]. However, the past decade has witnessed a marked increase in WGS of large cohorts of GAS strains, which has revealed significant intra-emm type diversity [7]. Moreover, the highly antigenic and thus hypervariable nature of the M protein has long been recognized [8]. Therefore, with the additional density of sequencing data and the elapsed time since the recognition of the five emm patterns, one might predict that additional variations in the emm region would be recognized. Herein, we report the identification of genetic event in currently circulating emm4 strains, previously considered pattern E, that give rise to a chimeric emm gene and a novel GAS emm pattern.

Methods

The invasive GAS emm4 strain SSGAS001 was isolated from a point-source outbreak in 2015 and its genome described using the strain name ‘Duke B’ [9]. We accessed the complete genomes of the 2001 pharyngeal isolate MGAS10750, considered the emm4 reference strain, and the 2015 pharyngeal emm4 isolate MEW427 from NCBI [6, 10]. Short-read sequencing data of invasive emm4 strains collected by the Centers for Disease Control and Prevention (CDC) during 2015 [11] and in England in 2014 [12] were accessed from BioProject PRJNA395240 and the European Nucleotide Archive project ERP015112 respectively. Emm region sequences were extracted from annotated draft genomes and visualized through Geneious (Biomatters). Two core single-nucleotide polymorphism (SNP)-based methods were deployed to reconstruct WGS-based phylogenies: (a) kSNP v3.0 [13] identified core SNPs based on reference free k-mer analysis (k-mer size of 19, default settings). These SNPs were used to reconstruct a maximum-parsimony tree; and (b) parsnp v1.2, available in the Harvest suite [14], was used to align assembled emm4 genomes to the MGAS10750 reference, and SNPs identified in local collinear blocks were subsequently used for reconstructing an approximate maximum-likelihood tree using FastTree 2 [15] while incorporating the general time reversible (GTR) model of nucleotide substitution. The Shimodaira–Hasegawa test implemented in FastTree2 was used to assess the support for significant clustering in the observed phylogeny. Due to similar findings, only trees reconstructed by parsnp are provided. A total of 88 emm4 GAS were identified among 930 GAS isolates collected at Texas Children’s Hospital in Houston (Texas, USA) between 2013 and 2017 [16] under a protocol approved by the Institutional Review Board at Baylor College of Medicine. The emm region was amplified by PCR and subject to Sanger sequencing (primer sequences in Table S1, available in the online version of this article). Taqman qRT-PCR was performed as described previously [17]

Results

Emm4 strains commonly cause both invasive and non-invasive GAS disease and are considered a prototype emm pattern E GAS [1, 18]. We previously compared the complete genome of an invasive emm4 strain isolated in 2015 (strain SSGAS001) to that of the reference emm4 strain MGAS10750, which was isolated in 2001 [6, 9]. Closer examination revealed a fusion event between emm and the immediately downstream enn gene in SSGAS001 resulting in a chimeric emm gene, which we have designated as emm4. The emm4 gene is a fusion of the 5′ end of the emm and the 3′ end of the enn gene of MGAS10750 (Fig. 1a). Given that the 5′ of emm4 is unchanged, SSGAS001 is still categorized as emm4. However, the C-terminal half of the emm4-encoded protein, M4C, only shares 62 % identity with the C-terminus of the MGAS10750 M4 protein. Moreover, the chimeric M4C protein contains an SF-3 allele in the PG domain rather than the SF-2 allele found in the PG domain of the MGAS10750 M protein. Given that the majority of M protein function has been mapped to the N-terminal region [1], the M4C protein is not predicted to have alterations in binding IgA, IgG, or C4BP, but interaction with albumin could be affected (Fig. 1b). In addition to generating a new predicted chimeric M protein, the emm–enn gene fusion event observed in SSGAS001 also results in just two genes in the emm region (mrp and emm) instead of the three (mrp, emm and enn) that are characteristic of Pattern E GAS (Fig. 1a). Thus, the gene fusion event gives rise to an emm pattern heretofore undescribed in GAS strains. We analyzed the other fully sequenced emm4 GAS strain, MEW427, also isolated in 2015 [10], and found it to have the identical emm4 gene and emm pattern to strain SSGAS001.

Although study of enn is limited, there have been reports of low enn transcription and/or translation, perhaps due to an unusual start codon [19-22]. We found that enn4 transcript level was approximately 20-fold lower than that of emm4 in MGAS10750 (Fig. S1) and that enn4 in MGAS10750 is predicted to have a non-standard start codon (isoleucine rather than methionine). Thus, Enn in MGAS10750 is likely to be produced at low levels.

To determine the frequency of the gene fusion event amongst currently circulating emm4 GAS strains, we first analyzed published WGS data from 52 invasive emm4 isolates collected by the CDC in 2015 [11]. We identified the emm–enn gene fusion event in 51 out of 52 strains with only a single strain containing emm pattern E (Fig. 2b, Table S2). Whole-genome-based phylogeny revealed that strains with the chimeric emm4 gene clustered distinctly from strains with the canonical emm4 gene (Fig. 1c). Importantly, the chimeric emm4 genes are present in strains isolated from multiple locations in the USA rather than being confined to a single location (Fig. S2).

Fig. 2.

(a) Schematic diagram showing the two most common chimeric M protein variants. The fusion point of the emm and enn genes is indicated by the change in box shading with numbers indicating the last amino acid from emm and the first amino acid from enn respectively in M4C protein. The 42 amino acid (126 nucleotide) deletion that is the difference between M4C-392 and M4C-350 is noted. C-terminal repeat regions (CRRs) along with J14 sequences are displayed as color-coded boxes. (b) Percentage distribution of the canonical and chimeric M4 proteins in 52 emm4 GAS strains collected by the CDC in the USA in 2015 depicted as a pie-chart. Numbers indicate percentages of the total. (c) Distribution of the canonical M4 and the different M4C variants identified in 88 emm4 strains isolated from patients at Texas Children’s Hospital between 2013 and 2017 expressed as a percentage of the number of isolates per year. Total number of isolates in a given year is indicated on the right of the bars. There are relatively fewer isolates in 2013 because collection only began in May, 2013. For (b) and (c), the type of M protein is color coded as indicated in the legend.

We identified six variations of the predicted M4C protein (Figs 2a and S3). The 350 amino acid long variant, which we have named M4C-350, occurred most commonly and was identical to M4C in strains SSGAS001 and MEW427. M4C-350 differs from the other common M4C variant, M4C-392, by a 126 nucleotide (42 amino acid) deletion (Fig. 2a). Together M4C-350 and M4C-392 are present in 90 % of strains in the CDC dataset that have the emm–enn gene fusion event (Fig. 2b). The other four variants appear to be derived via insertions or deletions in the M4C-392 or M4C-350 protein (Fig. S3). To determine if the presence of the chimeric emm–enn fusion was limited to strains in the USA, we analyzed short-read data for 33 emm4 isolates collected in England in 2014 [12]. We detected the emm–enn fusion event in 84 % of these strains.

Although the N-termini of the predicted M4C proteins are identical to the canonical M4 protein, the M4C proteins have different C-terminal repeat regions (CRRs). The CRRs represent a domain within the cell surface-exposed portion of the M and Enn proteins that contain tandemly arranged blocks of direct sequence repeats which have been targeted as potential GAS vaccine candidates [23]. The sequence repeats are not always identical and can be further classified based on the sequence of the C-terminal 14 amino acids (J14) in a given repeat [24]. The MGAS10750 M4 protein contains two CRRs, both containing the J14.2 sequence. Unlike MGAS10750, M4C-392 contains three CRRs, but other M4C variants also possess two CRRs each. However, the composition of the CRRs in every M4C variant differs from that in MGAS10750 in terms of the J14 sequence (Figs 2a and S3, Table S2).

To assess temporal emergence of emm4 strains with a chimeric emm4 gene, we analyzed 88 invasive and non-invasive emm4 isolates collected at the Texas Children’s Hospital in Houston between 2013 and 2017 via Sanger sequencing of the emm region (Table S3) [16]. We identified the emm–enn gene fusion in 74 strains (84 %). While 66 % of the emm4 strains isolated in 2013 were still the canonical emm pattern E, there was total replacement by strains carrying emm4 by 2017 (Fig. 2c). Similar to the CDC strains, the M4C-350 and M4C-392 variants were most common among the Houston isolates. Consistent with the hypothesis that M4C-350 arose from M4C-392, M4C-350-containing strains were not detected until one year after the initial identification of strains with the predicted M4C-392 protein (Fig. 2c).

Discussion

Since its discovery some 90 years ago, M protein has been considered the key GAS virulence determinant and is currently a prime vaccine candidate [2]. Despite extensive investigations over the past decades, continued study of M protein and the emm region continues to uncover new findings. Herein, we demonstrate that over 80 % of recently circulating emm4 GAS strains in the USA and England contain a non-canonical chimeric emm4 gene and a previously unrecognized emm pattern.

The key finding of this study was identification of a gene fusion event that gave rise to multiple novel entities. First, it created a chimeric emm gene, emm4, that consists of the 5′ end of the reference emm4 gene and the 3′ end of the enn4 gene. Second, formation of the chimeric emm4 gene eliminated the enn gene, thereby producing an emm region that only contains mrp and emm, a pattern different from the five established ones. Finally, unlike previously reported M proteins, which harbor either a SF-1 or SF-2 allele in the PG domain [3], the predicted M4C carries the SF-3 allele of the PG subfamily.

Analysis of 173 recent emm4 isolates revealed that strains with the chimeric emm4 gene and the new emm pattern have almost completely replaced the previously circulating emm4 pattern E strains in the USA and England. This indicates that the emm–enn gene fusion may confer a selectively advantageous trait, by itself or together with additional genomic changes in emm4 strains, which drove the emergent dominance of the emm4 strains in the studied populations. Given the critical nature of M protein in GAS host–pathogen interaction, it is plausible that the gene fusion event is at least a partial catalyst for the observed replacement.

M protein sequence variation can confer changes in immune recognition and has been associated with clonal population expansion [7]. However, reports of M protein variation have described small-scale changes (i.e. insertions, deletions, tandem repeat variation) in the N-terminal region whereas we observed replacement of the entire C-terminus [7, 8]. Although the N-terminus is typically considered to be the portion of the M protein under selective immune pressure [1], there are also data supporting immunogenicity of the C-terminal region. Like other M4 strains, Enn in MGAS10750 contains an unusual start codon and low enn transcript level, indicating that Enn4 may not have engendered an immune response in older M4 strains, consistent with reports from other serotypes [20, 25]. Hence, these data indicate that the emm–enn gene fusion may have significant immunological effects on GAS–host interactions, but this needs to be experimentally tested.

In summary, we report the occurrence of a recombination event that has given rise to a chimeric emm gene and a novel emm pattern and provide evidence of its predominance among current emm4 strains. These data indicate that GAS mechanisms that alter the M protein are more varied than previously appreciated, which could affect the efficacy of M-protein-based vaccine strategies.

Galloway-Pena J, Clement ME, Sharma Kuinkel BK, Ruffin F, Flores AR, et al. NCBI BioProject accession #PRJNA300859.

Beres SB, Richter EW, Nagiec MJ, Sumby P, Porcella SF, et al. NCBI BioProject accession #PRJNA224116.

Jacob KM, Spilker T, LiPuma JJ, Dawid SR, Watson ME, Jr. BioProject accession# PRJNA308988.

Chochua S, Metcalf BJ, Li Z, Rivers J, Mathis S, et al. Population and Whole Genome Sequence Based Characterization of Invasive Group A Streptococci Recovered in the United States during 2015. BioProject accession# PRJNA395240.

Chalker V, Jironkin A, Coelho J, Al-Shahib A, Platt S et al. Genome analysis following a national increase in Scarlet Fever in England 2014. European Nucleotide Archive project number ERP015112.

The Illumina short read data for SSGAS001 is available as BioProject # PRJNA300859 with the strain name of SASM4-Duke.

The complete and annotated genome of SSGAS001 can be found in Genbank under the accession number CP031770.

Click here for additional data file.

Click here for additional data file.