https://orcid.org/0000-0002-6048-6829
ABSTRACT
Streptococcus suis is a significant and emerging zoonotic pathogen. ST1 and ST7 strains are the primary agents responsible for S. suis human infections in China, including the Guangxi Zhuang Autonomous Region (GX). To enhance our understanding of S. suis ST1 population characteristics, we conducted an investigation into the phylogenetic structure, genomic features, and virulence levels of 73 S. suis ST1 human strains from GX between 2005 and 2020. The ST1 GX strains were categorized into three lineages in phylogenetic analysis. Sub-lineage 3-1a exhibited a closer phylogenetic relationship with the ST7 epidemic strain SC84. The strains from lineage 3 predominantly harboured 89K-like pathogenicity islands (PAIs) which were categorized into four clades based on sequence alignment. The acquirement of 89K-like PAIs increased the antibiotic resistance and pathogenicity of corresponding transconjugants. We observed significant diversity in virulence levels among the 37 representative ST1 GX strains, that were classified as follows: epidemic (E)/highly virulent (HV) (32.4%, 12/37), virulent plus (V+) (29.7%, 11/37), virulent (V) (18.9%, 7/37), and lowly virulent (LV) (18.9%, 7/37) strains based on survival curves and mortality rates at different time points in C57BL/6 mice following infection. The E/HV strains were characterized by the overproduction of tumour necrosis factor (TNF)-α in serum and promptly established infection at the early phase of infection. Our research offers novel insights into the population structure, evolution, genomic features, and pathogenicity of ST1 strains. Our data also indicates the importance of establishing a scheme for characterizing and subtyping the virulence levels of S. suis strains.
Introduction
Streptococcus suis, an important swine pathogen and emerging zoonotic agent, is responsible for septicemia, meningitis, endocarditis, and arthritis in both pigs and humans [Citation1,Citation2]. Humans primarily contract infections via close contact with infected pigs and the consumption of raw or undercooked contaminated pork products. The majority of human infections occur in Thailand, Vietnam, and China [Citation3]. S. suis serotype 2 and sequence type (ST) 1 strains are the most frequently isolated from sporadic human cases across the globe [Citation3]. In China, two outbreaks (one in Jiangsu Province in 1998 and another in Sichuan Province in 2005) characterized by streptococcal toxic shock-like syndrome (STSLS) with high mortality were caused by pathogenicity increased ST7 epidemic strains [Citation4]. Since the last outbreak in 2005, ST1 and ST7 have been predominant in sporadic human infections in China (multilocus sequence typing [MLST] database, https://pubmLst.org/organisms/streptococcus-suis). According to the National Notifiable Disease Reporting System of China, the Guangxi Zhuang Autonomous Region (GX) has been the most dominant region for reported sporadic human S. suis cases since 2006. Furthermore, S. suis strains from patients worldwide were divided into three lineages. The lineage I was composed of ST7 strains from Jiangsu province and Sichuan province outbreaks, whereas the lineage II and III was primarily composed of ST1 and ST7 strains from GX, respectively [Citation5]. The genomic characteristics and evolution of ST7 GX strains were investigated in our previous study [Citation6]. It is significant and urgently needed to elucidate the population structure, genomic features, evolution, and pathogenicity of S. suis ST1 strains from GX. In the present study, we analyzed the phylogenetic relationship of 73 ST1 clinical strains isolated in GX between 2005 and 2020. We also determined the presence of integrative and conjugative elements (ICEs) carrying antibiotic resistance (AR) genes and assessed the antimicrobial susceptibility profile of ST1 GX strains. Furthermore, we evaluated the virulence levels of representative ST1 GX strains and conducted a deeper investigation into the mechanisms underlying the differences in virulence levels among these strains.
Methods and materials
Bacterial strains and epidemiological features of cases
In this study, a total of 73 S. suis ST1 strains isolated from different cases in GX between 2005 and 2020 were collected (). All strains were confirmed as S. suis by amplifying S. suis-specific 16S rRNA [Citation7] and recN [Citation8] genes. The serotype of each strain was determined through seroagglutination using serum purchased from Statens Serum Institut in Copenhagen, Denmark.
Table 1. The information of ST1 GX strains.
The gender, age, and geographic information was available in 69, 65, and 72 cases, respectively. The proportion of male patients (92.8%, 64/69) was predominant. The age of patients ranged from 23 to 89, and the median age was 55 (47.5–63.75). The cases were distributed in 12 cities of GX, including Baise City (n = 21), Nanning City (n = 19), Yulin City (n = 9), four each from Chongzuo City, Guilin City, Liuzhou City, and Hechi City, two each from Guigang City and Hezhou City, one each from Beihai City and Qinzhou City. The epidemiological information on the exposure was available in 39 cases. Twenty-nine patients (74.4%, 29/39) with S. suis infection recalled exposure to pigs or pork within one week of the appearance of the initial symptoms. Interestingly, 21 (72.4%, 21/29) patients were occupational exposure involving abattoir workers, chefs, pig breeders, and butchers /sellers/porters of raw pork.
Host clinical information was available for 63 cases. Among these, 47 patients were diagnosed with meningitis, eight with STSLS, six with sepsis, one with pneumonia, and one with arthritis. Additionally, two meningitis patients had pneumonia, and one patient had arthritis as a complication (). All patients with STSLS presented with thrombocytopenia, except related information of one patient with STSLS was not available. Five STSLS patients presented with skin manifestations, including petechiae and ecchymosis.
Investigation of antimicrobial susceptibility profiles
In this study, we employed the minimum inhibitory concentration (MIC)-test strip (Liofilchem, Roseto degli Abruzzi, Italy) to assess the antimicrobial susceptibility of ST1 strains from GX. The following antibiotics were tested: vancomycin (0.016–256 μg/mL), penicillin G (0.016–256 μg/mL), ceftriaxone (0.016–256 μg/mL), azithromycin (0.016–256 μg/mL), tetracycline (0.016–256 μg/mL), streptomycin (0.064–1024 μg/mL), gentamicin (0.016–256 μg/mL), kanamycin (0.016–256 μg/mL), and spectinomycin (0.064–1024 μg/mL). For penicillin G, ceftriaxone, azithromycin, and tetracycline, we used breakpoints as recommended by the Clinical and Laboratory Standard Institute guidelines 2022 (M100-Ed32) for Streptococcus spp. Viridans group. Breakpoint values for vancomycin, streptomycin, kanamycin, spectinomycin, and gentamicin were derived from previous studies [Citation6,Citation9,Citation10].
Chromosomal DNA preparation, genomic sequencing and bioinformatics analysis of ST1 GX strains
Genomic DNA extraction and purification were conducted using the Wizard Genomic DNA Purification kit (Promega, Madison, USA) following the method described in our previous study [Citation11]. The draft genomes of the strains were sequenced using Illumina NovaSeq PE150 and assembled according to previously described methods [Citation6].
| (1) | MLST and Minimum Core Genome (MCG) typing | ||||
The sequence type (ST) and MCG group of each genome were determined based on whole-genome sequencing data by searching the MLST 2.0 database (https://cge.food.dtu.dk/services/MLST/) and the Pathogen Genome and Metagenome Analysis Cloud Platform (https://analysis.mypathogen.org/workflow/config/chinacdc/Ssuis_CGT/1/) [Citation12], respectively.
| (2) | Identification of core-pan genes and Phylogenetic analysis | ||||
Genes were predicted using Glimmer 3.02, and gene orthologs were identified using Gene Ontology (GO) V20171011, Kyoto Encyclopedia of Genes and Genomes (KEGG) V20181107, and Clusters of Orthologous Groups (COG) Database V20171127. Core-pan genome analysis was conducted using the Roary pipeline [Citation13]. First, the gff. annotation file for each genome, which served as input for the Roary pipeline, was generated using Prokka v1.13 [Citation14] with default parameters. The pan-genome was constructed by considering a global match region at 80% and an identity at 90% of the blastp sequence. Core genes were defined as present in all ST1 GX the strains, whereas specific genes were uniquely present in one strain. Single-nucleotide polymorphisms (SNPs) in core genes were detected using MUMmer v3.23. The mutational SNP sites were selected based on the method described in our previous study [Citation15]. The core genome phylogenetic tree was constructed using the maximum likelihood (ML) method with FastTree v2.1.10. To root the tree, the complete genome of S. suis strain GZ1 [Citation16] (serotype 2 and ST1, accession no. CP000837), isolated from a sporadic meningitis patient in Guizhou province in 2005, was used as an outgroup. The genome of the epidemic S. suis strain SC84 (serotype 2 and ST7, accession no. FM252031) isolated from a patient with STSLS and meningitis during the Sichuan outbreak in 2005 was used as reference [Citation17]. The resulting tree was visualized using FigTree v1.4.0.
| (3) | Detection of AR genes and S. suis known virulence genes | ||||
Antibiotic resistance (AR) genes were identified when their proteins exhibited at least 80% amino acid identity to known resistance proteins across 80% of the protein’s length, as determined by searching the ResFinder database (https://cge.food.dtu.dk/services/ResFinder/). The distribution of 154 known S. suis virulence genes was investigated within the pan-genome of ST1 GX strains. Virulence genes showing a global match region at < 80% and an identity at < 80% of the nucleotide-acid sequence were considered absent.
| (4) | Identification and phylogenetic analysis of mobile genetic elements (MGEs) with AR genes | ||||
The present of signature proteins (integrase, relaxase, and VirB4) in downstream sequences of major insertion hot spots, including the rplL gene, rum gene, mutT gene, luciferase-like monooxygenase gene, and ADP-ribose pyrophosphatase gene for the integration of MGEs in S. suis were identified using the database from a previous study [Citation18]. Intact MGEs containing AR genes were extracted. Sequence comparisons were performed using the BLASTn programme (with an E-value cutoff of 1e−10) and visualized using a custom Perl script (https://github.com/dupengcheng/BlastViewer). The alignment of 89K-like PAIs present in ST1 GX strains was achieved using MAFFT 7.520 with the FFT-NS-i iterative refinement method (https://mafft.cbrc.jp/alignment/server/). The ML phylogenetic tree of 89K-like PAIs based on multiple sequence alignment was constructed using MEGA.
Conjugation assay
The presence of the circular extrachromosomal form of 89K-like PAIs in corresponding strains was initially detected using specific primers and amplification cycling parameters as described in a previous study [Citation19]. In mating experiments, we employed the S. suis derivative strain P1/7RIF (tetracycline-susceptible but rifampin resistant) as the recipient [Citation20]. Ten S. suis strains carrying representative 89K-like PAIs (tetracycline-resistant but rifampin-susceptible) were used as donors. Additionally, the S. suis ST7 epidemic strain SC84, which harboured the 89 K PAI, was used as a donor for comparison purposes. Mating experiments were conducted according to the method outlined in our previous study [Citation9]. Transconjugants, grown on Todd Hewitt Broth (THB) containing both tetracycline and rifampin (8 μg/mL for each antibiotic), were further confirmed using PCR analysis with the primer pairs provided in supplementary Table 3 and a previous study[Citation19]. Transconjugants were counted if they were confirmed to simultaneously harbour 89K-like PAIs and P1/7-specific SNPs. Conjugative transfer frequencies were calculated as the number of transconjugants divided by the number of strains grown on THB containing tetracycline.
Experimental infection
| (1) | Survival assay in C57BL/6 mice infection model | ||||
We investigated the survival rates of 37 representative strains and added the well-characterized highly pathogenic S. suis ST1 strain P1/7 and ST7 epidemic strain SC84 for comparison. C57BL/6 mice (6 weeks old, female) were intraperitoneally injected with 2 × 107 CFU of each strain in 1 mL THB or 1 mL THB only as a control group. Each infected group comprised ten mice, whereas the mock-infected group contained five mice. Mortality was recorded every 6 h within the first 24 h of infection, every 12 h from 24 to 48 h post-infection, and every 24 h from 48 to 96 h post-infection. The infection experiment of each strain was independently conducted at least twice. The mean survival rate of each strain was calculated using the Kaplan–Meier method (Table 3).
| (2) | Pro-inflammatory cytokine production and bacterial loads in peripheral blood of infected C57BL/6 mice. | ||||
C57BL/6 mice (6 weeks old, female) were intraperitoneally injected with 2 × 107 CFU of seven representative strains with different virulence levels in 1 mL THB. Each group comprised seven mice. As some mice from the infected groups succumbed around 8 h post-infection, mice were euthanized, and peripheral blood was aseptically collected at 4 and 7.5 h post-infection. Serial tenfold dilutions of peripheral blood from each infected mouse were plated onto THB agar plates, and colonies were counted and expressed as CFU/mL. The concentration of interleukin IL-6 and TNF-α in serum was measured using ELISA kits (R&D Systems, Minneapolis, MN, USA) following the manufacturer’s recommended protocols. The cytokine values for each infected group were expressed as the median in pg/mL.
| (3) | Survival assay in the zebrafish infection model | ||||
Fifteen adult zebrafish per group were intraperitoneally injected with 6 × 106 CFU of recipient S. suis strain P1/7RIF and five transconjugants in 1 mL phosphate-buffered saline (PBS) or 1 mL PBS only as the control group. Mortality was monitored every 12 h until 96 h post-infection, and the survival rate of each group was calculated using the Kaplan–Meier method.
Measurement of surface hydrophobicity and capsule detection by transmission electron microscopy
The relative surface hydrophobicity of ST1 GX strains was investigated by measuring their absorption to n-hexadecane, following the procedure outlined in a previous study with minor modifications [Citation21]. In brief, bacteria were cultured overnight in 10 mL of sterile LA broth at 37°C with 5% CO2, harvested by centrifugation, washed three times with 5 mL of PUM buffer, and resuspended into a bacterial suspension with an optical density (OD) value of 0.6 (OD660nm) with PUM buffer. Then, 3 mL of the bacterial suspension was transferred to round-bottom test tubes (BD Falcon, Shanghai, China) containing 400 µL of n-hexadecane and vortexed for 2 min. The aqueous phase was collected, and the OD value (OD660nm) was measured after standing for 10 min. The % hydrophobicity was calculated as (OD660 initial-OD660 final) / OD660 initial × 100.
The capsule of selected strains was examined by transmission electron microscopy using the method detailed in our previous study [Citation22]. Capsular thickness was measured between the inner and external edges of the capsular layer, with each value determined based on 20–25 measurements per strain.
Statistical analyses
The statistic programme IBM SPSS Statistics 24 was used in the present study. Survival rates of different infected groups were compared using the Gehan-Breslow-Wilcoxon test. Statistical analysis of pro-inflammatory cytokine concentrations and bacterial counts in different infected groups was performed using the Wilcoxon’s rank-sum test. A p-value < 0.05 was considered significant.
Nucleotide sequence accession numbers
The sequences of the genomes sequenced in the study were deposited in the GenBank under accession numbers listed in .
Ethical approval
The C57BL/6 mouse infection experiment was approved by the ethics committee of the National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention (Permit code: 20232-047). This zebrafish infection experiment was approved by the Laboratory Animal Monitoring Committee of Jiangsu Province (Permit code: S SYXK (Su) 2021–0086).
Results
Serotyping and MCG analysis of ST1 GX strains
Among the 73 ST1 GX strains, 69 were of serotype 2, and four were of serotype 14 (strains GX4, GX129, GX133, and GX136). MCG typing grouped the ST1 GX strains into MCG1.
Phylogenetic analysis of ST1 GX strains
High genomic heterogeneity was observed among ST1 GX strains. The pan-genome of the 73 ST1 GX strains consisted of 2,715 genes, including 1,719 core genes and 285 specific genes.
Based on the distribution of 7,652 mutational SNPs in the core genomes, the ST1 GX strains were categorized into three lineages (). Strains GX105 and GX96 formed their own distinct branches, whereas lineages 1 and 2 included five and seven strains, respectively. Fifty-nine strains clustered into lineage 3, primarily divided into sub-lineages 3–1 (n = 12) and 3–2 (n = 46), with strain GX139 standing apart from other lineage 3 genomes. Lineage 3–1 primarily comprised two sub-lineages, 3-1a (n = 4) and 3-1b (n = 8). Lineage 3–2 was predominantly composed of four sub-lineages, 3-2a (n = 2), 3-2b (n = 3), 3-2c (n = 18), and 3-2d (n = 22), although strain GX90 did not cluster with other lineage 3–2 genomes. Notably, lineage 3-2d included strains of both serotypes 2 and 14, suggesting potential horizontal transfer of two cps loci among strains in lineage 3-2d.
Figure 1. The maximum-likelihood phylogenetic tree of 73 S. suis ST1 strains from GX. The phylogenetic tree was constructed based on the mutational SNPs in the core genomes. The S. suis strain GZ1 was used as an outgroup to root the tree. The strains were coloured based on the serotype. Black was serotype 2, red was serotype 14. The scale is given as the number of substitutions per variable site.
To determine the evolutionary relationship of ST1 GX strains with the ST7 epidemic strain SC84, whole-genome SNPs were identified between each lineage/sub-lineage and the reference strain SC84. Sub-lineage 3-1a exhibited a closer phylogenetic relationship with the reference strain SC84, with 466 SNPs. This was followed by lineage 2 with 1,496 SNPs and lineage 1 with 1,520 SNPs. Sub-lineages 3-1b, 3-2a, 3-2b, 3-2c, and 3-2d were phylogenetically distant from the reference strain SC84, with 795 SNPs, 1,338 SNPs, 912 SNPs, 1,941 SNPs, and 2,819 SNPs, respectively.
Detection of AR genes and Antimicrobial susceptibility profiles of ST1 GX strains
Among the 73 ST1 GX strains, only strain BS30 did not harbour any AR genes. In the remaining 72 ST1 GX strains, a total of 168 AR genes were present (). These AR genes fell into six categories, including tetracycline, macrolide-lincosamide – streptogramin B (MLS), lincosamide, streptogramin, aminoglycoside, and oxazolidinone resistance genes. Notably, over 86% of ST1 GX strains (63/73) carried AR genes from at least two categories. Strains BS22 and BS23 were remarkable in that they harboured seven AR genes simultaneously (). Twenty-two and 72 ST1 GX strains carried MLS resistance gene ermB and tetracycline-resistant genes, respectively. The predominant tetracycline-resistant gene was tetM (n = 56), followed by tetO (n = 16) and tetW (n = 3). In addition, the streptomycin resistant gene ant(6)-Ia (n = 53), kanamycin resistant gene aph(3’)-IIIa (n = 5), the oxazolidinone-resistant gene optrA (n = 4), gentamicin resistant gene aac(6’)-Ie-aph(2')-Ia (n = 2), spectinomycin resistant gene spw (n = 2), lincosamide-resistant genes lnuB (n = 2) and lsaC (n = 2) were also found in ST1 GX strains.
Table 2. The AR genes and antimicrobial susceptibility profile of ST1 GX strains.
In this study, the antimicrobial susceptibility of ST1 GX strains was investigated (). All ST1 GX strains were susceptible to vancomycin, penicillin G, and ceftriaxone. Strains carrying tetracycline resistance genes were resistant to tetracycline, with MIC values ranging from 12 μg/mL to 32 μg/mL. Resistances to azithromycin (MIC value >256 μg/mL) were observed in all strains carrying the erm(B) gene. High degrees of kanamycin (MIC value >256 μg/mL), spectinomycin (MIC value >1,024 μg/mL), and gentamicin (MIC value >256 μg/mL) resistance were found in corresponding strains carrying the aph3-iiia, spw, and aac(6’)Ie-aph(2'‘)Ia genes, respectively. The MIC values of streptomycin ranged from 256 μg/mL to >1,024 μg/mL in the strains carrying the ant(6)-Ia gene (). These data indicate that these AR genes conferred corresponding antibiotic resistance phenotypes to their host.
MGEs with ARs in ST1 GX strains
Five ST1 GX strains, namely GX96, GX92, BS30, BS12, and GX56, did not carry ICEs. Intriguingly, four of these were clustered into lineage 1 of the core genome phylogenetic tree. There were 58 intact ICEs with ARs integrated at the rplL site and eight at the rum site in ST1 GX strains ( and ). Notably, the ICEs harboured the majority of AR genes (134/168) present in ST1 GX strains.
All ICEs integrated at the rplL site contained a 15-bp att sequence 5′-TTATTTAAGAGTAAC-3′ in the flanking regions and were classified into four types. Fifty-four ICEs were similar to the 89 K PAIs harboured by S. suis ST7 epidemic strains, denominated as 89K-like PAIs. Additionally, two ICEs (found in strains GX32 and GX139) were identical to ICESsuLP081102 (GenBank: KX077885), whereas ICESsuGX132 (present in strain GX132) and ICESsuGX152 (present in strain GX152) had no homologies in the NCBI database. Strains GX46 and BS15 carried non-intact 89K-like PAIs at the rplL site. Notably, ST1 GX strains carrying 89K-like PAIs were clustered into lineage 3 of the core genome phylogenetic tree ().
The ICEs with AR genes integrated at the rum site were present in eight ST1 GX strains: BS18, GX62, GX74, GX105, GX144, GX146, GX149, and GX151. Intriguingly, these eight strains were primarily clustered into lineages 1 and 2 of the core genome phylogenetic tree (). All ICEs integrated at the rum site possessed 14-bp att sequences: 5′-CACGTGGAGTGCGT-3′ and 5′-CACATAGAAGTTGT-3′ in the 5’ and 3’ side flanking regions, respectively.
The eight ICEs shared a high degree of sequence similarity. The sequences of ICEs in GX74, GX144, GX146, GX149, and GX151 were almost identical to ICESsu14ND1 (GenBank: MK211778.1). They exhibited 96% coverage and 99.45% identity with ICESsuGX62, 90% coverage and 94.92% identity with ICESsuBS18, and 88% coverage and 96.82% identity with ICESsuGX105 at the nucleotide level. Notably, ICESsuGX88 in ST7 sporadic strains from GX shared 100% coverage and 99.98% identity with ICESsuBS18 at the nucleotide level.
Characteristics of 89K-like PAIs in ST1 GX strains
| (1) | Bioinformatic analysis of 89K-like PAIs | ||||
In this study, the 54 intact 89K-like PAIs were categorized into four clades based on sequence alignment: clade 1 (73 Kb, n = 8), clade 2 (78Kb, n = 40), clade 3 (83 Kb, n = 2), and clade 4 (71 Kb, n = 4) (). The gene sequences and arrangements of 89K-like PAIs were highly conserved within the same clade.
Figure 2. The schematic regions of the 89 K PAI based on the absence and presence in the 89K-like PAIs. Different regions were marked by coloured shading. CR: conserved region; HVR: highly variable region; VR: variable region; the different coloured lines indicated the regions were absent in the corresponding clade of the 89K-like PAIs phylogenetic tree. The maximum-likelihood phylogenetic tree based on Multiple Sequence Alignment (MSA) of 89K-like PAIs was constructed using MEGA with bootstrap value 1000.
Clade 4 89K-like PAIs shared a higher phylogenetic relationship with the 89 K PAI harboured in S. suis ST7 epidemic strains compared to other clades (). Notably, ST1 GX strains carrying clade 4 89K-like PAIs also exhibited a closer phylogenetic relationship with the S. suis ST7 epidemic strain at the whole genome SNP level. Clade 1 89K-like PAIs only harboured the tetM gene, whereas clades 2, 3, and 4 89K-like PAIs contained both the ant(6)-Ia and tetM genes. Additionally, the ermB gene was found in clade 3 89K-like PAIs. ST1 GX strains carrying clades 1 and 4 89K-like PAIs were clustered into lineage 3-1b and lineage 3-1a of the core genome phylogenetic tree, respectively. ST1 GX strains carrying clades 2 and 3 89K-like PAIs were clustered into lineage 3–2 of the core genome phylogenetic tree ().
Comparing the 89 K PAI harboured in ST7 epidemic strain SC84, genetic absences were observed in all 89K-like PAIs. Based on the frequency of genetic variations in the 89K-like PAIs, the 89 K PAI was divided into several regions, including the highly variable region (HVR: absent in all 89K-like PAIs), variable regions (VRs: absent in some 89K-like PAIs), and conserved regions (CRs: present in all 89K-like PAIs) ().
HVR consisting of SSUSC84_0816, SSUSC84_0817 and SSUSC84_0818 genes was absent in all 89K-like PAIs. These genes encode hypothetical proteins, asparagine synthetase, and hypothetical proteins, respectively.
VR1 consisting of SSUSC84_0837 and SSUSC84_0838 genes was absent in clades 2 and 4 89K-like PAIs. Both genes were located in the Tn916 transposon and encoded hypothetical proteins.
VR2 consisting of SSUSC84_0839 (encoding DNA helicase) and SSUSC84_0840 (encoding exonuclease) genes was absent in clade 4 89K-like PAIs.
VR3 consisting of the salK gene (SSUSC84_0850), which encodes a two-component signal transduction system, was absent in clade 4 89K-like PAIs.
VR4 consisting of SSUSC84_0852, SSUSC84_0853, SSUSC84_0854, and SSUSC84_0855 may be related to lantibiotic transport. VR4 was absent in clades 2, 3, and 4 89K-like PAIs.
VR5 consisting of SSUSC84_0856, SSUSC84_0857, and SSUSC84_0858 genes may be related to lantibiotic modification. VR5 was absent in clades 1 and 4 89K-like PAIs.
VR6 consisting of eight genes (from SSUSC84_0859 to SSUSC84_0866) was absent in clade 1 89K-like PAIs. Notably, the streptomycin-resistant gene ant(6)-Ia was located in VR6.
The remaining genes of the 89 K PAI were present in all 89K-like PAIs and comprised four conserved regions (CRs): CR1 (genes from SSUSC84_0807 to SSUSC84_0815), CR2 (genes from SSUSC84_0819 to SSUSC84_0836), CR3 (genes from SSUSC84_0841 to SSUSC84_0849), and CR4 (genes from SSUSC84_0868 to SSUSC84_0890). CR3 primarily constituted the Tn916 component, whereas CR4 contained an intact type IV secretion system (T4SS). Notably, CR3 and CR4 were also widespread in ST1 GX genomes that did not harbour 89K-like PAIs.
Compared to the 89 K PAI, different insertions were also found in some 89K-like PAIs. The ISL3 transposase-ermB composition was integrated between VR1 and VR2 in the clade 3 89K-like PAIs.
A conserved ICE integrated into the 5′ end of three 89K-like PAIs (carried by strains BS22, BS23, and GX10) of clade 1 and four 89K-like PAIs (carried by strains GX1, GX20, GX23, and GX94) of clade 2, forming seven tandem ICEs. Three 15-bp att sequences (5′-TTATTTAAGAGTAAC-3′) were found in the flanking regions of these tandem ICEs. The homology of the ICE integrated into seven 89K-like PAIs was also present in tandem ICESsuYS19 (MK211811) and ICESsuZJ20091101-2 (KX077883).
| (2) | Transferability of 89K-like PAIs | ||||
The circular 89K-like PAIs, resulting from the integration of att sites, were detected in all ST1 GX strains carrying 89K-like PAIs. It is noteworthy that 89K-like PAI of seven tandem ICEs formed circular products and 89K-like PAI of tandem ICE present in strain GX23 transferred to P1/7RIF individually. Noticeable differences in transfer frequency were observed among 89K-like PAIs from different clades. For instance, the transfer of 83 K PAI-BS25 from clade 3 was detected at a frequency of 1.15 × 10−5 per donor, significantly higher than those of 78 K PAI-GX23 (2.1 × 10−6), 78 K PAI-GX4 (7.22 × 10−8), 78 K PAI-GX52 (5.86 × 10−8), 78 K PAI-GX34 (6.67 × 10−8), 78 K PAI-GX2 (1.07 × 10−8), 78 K PAI-GX31 (9.1 × 10−8), and 78 K PAI-GX53 (1.62 × 10−6) from clade 2, 73 K PAI-GX72 (4.74 × 10−8) from clade 1, and 78 K PAI-GX36 (3.37 × 10−7) from clade 4. In this study, the transfer of 89 K PAI-SC84 to P1/7RIF was detected at a frequency of 2.27 × 10−7.
| (3) | Biological properties of 89K-like PAIs | ||||
The acquisition of 89K-like PAIs conferred antimicrobial resistance (resistance to tetracycline, streptomycin, and MLS) to the recipients (data were not shown). Differences in virulence levels between the recipient strain P1/7RIF and its transconjugants were observed using zebrafish infection models. The mortality of zebrafish infected with transconjugants was significantly higher than that of zebrafish infected with the recipient strain P1/7RIF.
Over 90% of zebrafish infected with the reference transconjugant 89 K PAI-SC84 died within the infection period, whereas none of the zebrafish infected with recipient strain P1/7RIF died within the same period. Similarly, the mortality rates of zebrafish infected with transconjugants of 89K-like PAIs at 96 h post-infection reached or exceeded 80%, except for the transconjugants 78 K PAI-GX23, which reached 66.7% (Supplementary Table 2), respectively. The acquisition of 89K-like PAIs increased the virulence of the corresponding recipient strains. Notably, no differences in the survival curve were observed among the five transconjugants of 89K-like PAIs-infected zebrafish.
Differences in virulence levels among ST1 GX strains
In the survival assay, 37 ST1 strains were selected based on their positions in the core genome phylogenetic tree. Highly pathogenic ST1 strain P1/7 and ST7 epidemic strain SC84 were added as reference strains. Noticeable differences in survival rates were observed among mice infected with ST1 GX strains ( and ).
The survival levels of mice infected with GX53 and GX144 were significantly lower than those of mice infected with the highly pathogenic strain P1/7 and similar to those of mice infected with ST7 epidemic strain SC84. These two strains were classified as epidemic (E) strains.
The survival levels of mice infected with 10 strains were similar to those of mice infected with the highly pathogenic strain P1/7. Based on the mortality of infected mice at 12 h post-infection, they were classified as either E or highly virulent (HV) strains, respectively.
Mice infected with strains GX96, BS21, and GX72 had a survival rate of ≤10% at 12 h post-infection, similar to that of mice infected with ST7 epidemic strain SC84 and lower than that of mice infected with the highly pathogenic ST1 strain P1/7 at the same time point. These three strains were also classified as E strains.
The survival rates of mice infected with strains GX56, GX43, BS22, GX31, BS25, BS6, and GX4 were ≥20% at 12 h post-infection. These seven strains were classified as HV strains.
The survival levels of mice infected with 18 strains were significantly higher than that of the highly pathogenic strain P1/7-infected group and lower than that of the mock-infected group. Based on the survival rates of infected mice at 96 h post-infection, 11 and seven strains were classified as V+ and V strains, respectively. At 96 h post-infection, the survival rates of mice infected with V+ strains were ≤25% (the highest survival rate of mice infected with HV strains at 96 h post-infection), whereas those of mice infected with V strains were ≥60%.
No significant differences in survival curves were observed between mice infected with seven ST1 GX strains and mock-infected mice. These seven strains were classified as LV strains.
Table 3. The data of survival assay in C57BL/6 mice infection model.
Mechanisms related to differences in virulence levels among ST1 GX strains
| (1) | Pro-inflammatory cytokine production in serum correlated with S. suis virulence | ||||
To understand the mechanisms underlying the differences in virulence levels among ST1 GX strains, the ability of selected strains with varying virulence to induce pro-inflammatory cytokine production in vivo was evaluated. The following strains were included: one E strain (GX53), two HV strains (BS25 and GX31), two V+ strains (GX6 and GX20), one V strain (GX62), and one LV strain (GX23).
At 4 h post-infection, the levels of IL-6 in the serum were similar among mice infected with E, HV, and V+ strains but significantly higher than those of mice infected with V and LV strains. V strain GX62 induced higher IL-6 production compared to LV strain GX23 (A).
Figure 3. Production of pro-inflammatory cytokines IL-6 (A) and TNF-α (B) in sera and bacterial loads (C) in peripheral blood of C57BL/6 mice infected with 2 × 107 CFU of representative ST1 GX strains with different virulence level. Median values of each infected group were used to express cytokine levels in sera. Bacterial counts of individuals, including median with interquartile ranges, were presented. Colonies were expressed as CFU/ml. Statistical differences in pro-inflammatory cytokine concentration and bacterial counts among infected groups were determined by Wilcoxon’s rank sum test. *: significantly higher than those of HV strain BS25 and GX31 infected groups. #: significantly higher than those of V+ strain GX6 and GX23 infected groups. &: significantly higher than those of V strain GX62 infected group. §: significantly higher than those of LV strain GX23 infected group.
At 7.5 h post-infection, the levels of IL-6 in serum for E and HV strain-infected mice increased dramatically, significantly higher than those for V+, V, and LV strains. Importantly, no differences were observed between E and HV strains. V+ strains induced significantly higher IL-6 levels in serum compared to V and LV strains. Moreover, the IL-6 levels in serum for LV strain GX23 were significantly lower than those for V strain GX62 (A).
At 4 h post-infection, mice infected with E strain GX53 produced the highest level of TNF-α in serum, followed by those infected with HV strains, which induced higher TNF-α production than V+, V, and LV strains. The level of TNF-α in serum was significantly higher in V+ strain-infected mice compared to V strain-infected mice, and V strain-infected mice had higher TNF-α levels than LV strain-infected mice (B).
At 7.5 h post-infection, TNF-α levels in serum for V+ strains GX6 and GX10 significantly increased, similar to those for HV strains at the same time point. TNF-α levels were significantly higher in mice infected with E, HV, and V+ strains than in those infected with V strain GX62. The levels of TNF-α in serum was significantly higher in mice infected with V strain GX62 compared to LV strain GX23 (B).
| (2) | Bacterial loads in peripheral blood | ||||
We also investigated whether differences in virulence levels correlated with the bacterial capacity to survive and replicate in peripheral blood.
At 4 h post-infection, the bacterial burdens in peripheral blood of mice infected with E, HV, and V+ strains were similar and significantly higher than those of mice infected with V and LV strains. There were no differences in bacterial burdens between mice infected with V strain GX62 and LV strain GX23 at this time point (C).
At 7.5 h post-infection, the bacterial loads of E and HV strains-infected mice were similar and significantly higher than those of mice infected with V strain GX62 and LV strain GX23. The bacterial loads of V+ and V strains-infected mice were similar and significantly higher than those of mice infected with LV strain GX23 (C).
| (3) | The cell surface hydrophobicity of ST1 GX strains | ||||
Capsule polysaccharides (CPS) contribute to the colonization and immune evasion of S. suis strains in vivo. In this study, we evaluated the genetic characteristics of cps loci and encapsulation of ST1 GX strains.
All ST1 GX serotype 2 and 14 strains had intact cps loci, which closely resembled those of serotype 2 strain SC84 and serotype 14 reference strain 13730 (Genbank No. AB737822), with minor differences, particularly the absence of certain transposases in the cps loci of ST1 GX strains.
The assessment of surface hydrophobicity in seven selected strains with varying virulence levels showed consistently low values, all below 5% (). Furthermore, transmission electron microscopy revealed that the capsule of LV strain GX23 similar to the reference strain SC84 had thicknesses ranging from 40–50 nm, and V strain GX62 similar to the E strain GX53 had a thickness of thicknesses ranging from 30–40 nm (Supplementary Figure 1). These observations aligned with the surface hydrophobicity test results and confirmed that V strain GX62 and LV strain GX23 were effectively encapsulated.
Further assessment of the surface hydrophobicity () of the remaining ST1 GX strains revealed that the majority (86.4%, 57/66) were also well encapsulated, exhibiting very low hydrophobicity (below 5%). A smaller subset of six strains exhibited relative low hydrophobicity (10%−20%), and three strains displayed high hydrophobicity (above 30%) (). Notably, strain GX92, with a high hydrophobicity of 51.6%, had a thinner capsule measuring 10–20 nm, significantly less than strains with hydrophobicity below 5% (Supplementary Figure 1).
| (4) | Distribution of S. suis known virulence genes in ST1 GX strains | ||||
Among the seven ST1 GX strains, E, HV, V+, and LV strains harboured all 154 known virulence genes, except for the absence of the mrp gene in HV strain GX31. Additionally, V strain GX62 lacked six virulence genes, namely nisK, nisR, salK, salR, hhly3, and sp1 (Supplementary Table 3).
Of the 154 known virulence genes, a total of 142 were present in all ST1 GX strains. Twelve of these known virulence genes were classified as accessory genes. Seven of these were found in the 89 K PAI, which included nisK, nisR, hhly3, sp1, salR, salK, and traG genes. The “classical” virulence marker for S. suis strains, the mrp gene, was absent in only two strains of lineage 3-2a. However, the epf and sly genes were present in all ST1 GX strains. In addition, the genotype of the mrp gene found in ST1 GX strains was of the EU subtype. Furthermore, four genes (SSUSC84_1905, SSUSC84_1906, SSUSC84_1907, and SSUSC84_1908), encoding the major pilus subunit, were absent in strain GX139. Of note, two genes, 340 and 344, which are recognized as diagnostic markers specific for human-associated S. suis [Citation5], were absent in strain GX90 (Supplementary Table 3).
Discussion
GX has been the region with the highest number of human S. suis cases in China since 2006. The average incidence of human S. suis cases was 0.49 cases per million population between 2006 and 2022, obviously lower than those of Vietnam (5.4 cases/million population) and Thailand (8.21 cases/million population) [Citation3]. The annual incidence of human S. suis cases in GX was 0.08–1.58 cases per million population from 2006 to 2022, significantly increasing since 2019 (Supplementary Table 4). In total, 166 S. suis strains were isolated from patients in GX between 2005 and 2020. The proportion of ST1 strains in GX was 44% (73/166).
In our previous study, we hypothesized that epidemic ST7 strains evolved from ST1 strains by acquiring five “acquired islands” (AIs) [Citation16]. We aimed to further our understanding of the genomic epidemiology and pathogenicity of ST1 GX strains, shedding light on their evolutionary history and their potential connections with ST7 epidemic strains.
Core genome phylogenetic analysis of the ST1 GX strains revealed a high level of genomic diversity, dividing them into three distinct lineages. Notably, all ST1 GX strains that harboured 89K-like PAIs were clustered within lineage 3, which comprised six sub-lineages. Of particular interest, lineage 3-1a displayed a closer phylogenetic relationship with the ST7 epidemic strain SC84. A previous study categorized S. suis strains from patients worldwide into three lineages. Lineage I is composed of ST7 epidemic strains with 89 K PAIs and lineage II is primarily composed of ST1 strains with 89K-like PAIs. These two lineages were believed to have evolved independently [Citation5]. The specific genomic features of lineage 3-1a warrant further exploration to clarify its evolutionary ties to ST7 epidemic strains.
In this study, a major observation was the prevalence of 89K-like PAIs in ST1 GX strains, with these PAIs being classified into four clades based on sequence alignment. Notably, 89K-like PAIs of clade 2 (78 Kb) were predominant among ST1 GX strains, which was consistent with the finding of a previous study [Citation5].
The study revealed the spontaneous excision of 89K-like PAIs in ST1 GX strains, forming extrachromosomal circular structures through site-specific recombination between attL and attR sites. These circular 89K-like PAIs were capable of reintegrating into the rpIL site of P1/7RIF, but with varying transfer frequencies ranging from 10−5–10−8. The 89K-like PAIs were consistently identified in ST7 and ST1 human strains in Vietnam, GX, and Guangdong Province neighbouring to GX of China [Citation5,Citation6,Citation23]. The presence of the 89K-like PAIs in the S. suis ST658, ST869, ST951, ST1005 strains [Citation5,Citation24] reflects their high transferability and wide dissemination.
89 K PAI has been only identified in S. suis ST7 epidemic strains and has played critical roles in the increased pathogenicity of S. suis ST7 epidemic strains. It contributes to STSLS through mechanisms involving the T4SS-like system and two-component signal transduction systems salR/salK and nisR/nisK [Citation25–27]. These features are also present in all 89K-like PAIs, except for the absence of the salR gene in clade 4 89K-like PAIs. The absence of salR gene was also found in an 89K-like PAI from an S. suis ST1 strain isolated from an STSLS patient in Guangdong Province [Citation23]. Transferring representative 89K-like PAIs from the four clades into P1/7RIF resulted in significantly increased mortality of zebrafish infected with the corresponding transconjugants. This indicates that 89K-like PAIs contribute to the enhanced pathogenicity of host strains. Interestingly, no significant differences in virulence levels were observed among the transconjugants in the zebrafish infection model. The specific role of the salR gene in the pathogenesis of 89K-like PAIs requires further investigation.
In the current study, we noted the apparent diversity in virulence levels among ST1 GX strains. Compared to the highly pathogenic S. suis strain P1/7, significantly lower virulence levels were observed in the 17 strains harbouring 89K-like PAIs. Furthermore, the proportion of strains harbouring 89K-like PAIs (34.6% 9/26) classified as E/HV was similar to that of strains without 89K-like PAIs (27.3% 3/11). The presence of 89K-like PAIs was not a reliable indicator of highly pathogenic strains. Further studies are required to thoroughly elucidate the role of 89K-like PAIs in the pathogenesis of the S. suis infections.
In the current study, 142 of 154 putative virulence genes were identified as core genes in ST1 GX strains. Notably, the virulence-associated genes typically found in highly pathogenic S. suis serotype 2 strains, such as mrp, sly, epf, ofs, revS, nadR, SSU05_0473, neuB, and neuC [Citation28,Citation29], were present in all V and LV strains. However, up to this point, identifying a specific virulence marker to differentiate the virulence levels of S. suis strains remains impractical. Consistent with our previous result, it appears that the virulence level of S. suis strains might be determined by the expression of virulence-associated genes rather than their mere presence [Citation30]. There is an urgent need for the development of an effective scheme for characterizing and subtyping the virulence levels of S. suis strains.
The study delved into the mechanisms underlying the variation in virulence among ST1 GX strains. Proinflammatory cytokine levels in serum were used to gauge the severity of inflammation. The level of IL-6 in serum has an inverse correlation with survival time in patients with sepsis [Citation31]. TNF-α is one of the most crucial host mediators in the pathogenesis of septic shock [Citation32] and was responsible for the STSLS observed in the Sichuan outbreak [Citation16]. The study compared the ability of representative ST1 GX strains with different virulence levels to induce the production of pro-inflammatory cytokines. The E strain infection induced the highest level of TNF-α among the infected groups at 4 and 7.5 h post-infection. This early burst of TNF-α likely played a critical role in the significantly higher mortality of mice infected with the E strain compared to those infected with HV strains before 12 h post-infection. The levels of TNF-α at 4 h post-infection and IL-6 at 8 h post-infection in the serum of mice infected with HV strains were significantly higher than those of mice infected with V+ strains. The synergistic effects of the sequential and substantial induction of these two pro-inflammatory cytokines contributed to the higher mortality of mice infected with HV strains compared to those infected with V+ strains at 24 h post-infection. In contrast, mice infected with V+ strains exhibited a delayed burst of TNF-α compared to E and HV strains. This delayed burst of TNF-α may be responsible for the higher mortality of mice infected with V+ strains during the infection period than that of mice infected with V and LV strains.
Having observed significant variations in TNF-α levels among mice infected with ST1 GX strains with different virulence levels, the study further investigated whether these variations were related to the bacterial capacity to survive and disseminate in the peripheral blood. Similar bacterial loads in the peripheral blood were observed among mice infected with E, HV, and V+ strains at 4 and 7.5 h post-infection, indicating that the differences in the early induction of pro-inflammatory cytokines were not linked to variations in bacterial loads.
Compared to E, HV, and V+ strains, V strain GX62 exhibited a delayed capacity to establish infection. This delayed establishment of infection might contribute to the lower mortality of mice infected with V strain during the infection period compared to those infected with E, HV, and V+ strains.
It is worth noting that the bacterial burdens in mice infected with LV strain GX23 were significantly lower than in other infected groups during the infection period. Maintaining a consistently high level of bacteremia is closely linked to the pathogenesis of S. suis strains. The ability to evade the host’s immune system, establish colonization, and replicate in vivo is a crucial premise for the pathogenesis of S. suis. The lower bacterial loads in vivo might lead to a reduced potential for inducing lethal infections, resulting in decreased mortality among mice infected with LV strain GX23.
The presence of CPS and SLY have already been shown to be critical in resisting host clearance, allowing bacteria to establish infection [Citation33,Citation34]. The well-encapsulated LV strain GX23 possessed an intact sly gene. The virD4 gene located in the T4SS and the tstS gene located in Tn916, both carried by the 89 K PAI, enhanced the resistance of host strains to phagocytosis and killing by various host cells in vitro [Citation35,Citation36]. Both virD4 and tstS genes were present in the 89K-like PAI harboured by LV strain GX23. The mechanism underlying the vulnerability of LV strain to be cleared by the host immune system requires further investigation.
In conclusion, this study revealed significant diversity in phylogenetic structure and virulence levels among ST1 GX strains. The strains harbouring the 89K-like PAIs were predominant in ST1 GX strains. Although the 89K-like PAIs increased the pathogenicity of corresponding transconjugants in the zebrafish infection model, the role of 89K-like PAIs in S. suis infection requires further investigation. A significant correlation was observed between the TNF-α level in serum during the early phase of infection and the mortality of mice infected with S. suis ST1 GX strains. A high bacteremia level in the early phase of infection was also crucial for S. suis strains to initiate lethal infections. The lower mortality of mice infected with V and LV strains was partially due to their delayed capacity to establish infection or their higher susceptibility to bacterial clearance by the host. Identifying a virulence-associated marker to differentiate the virulence levels of S. suis strains remains challenging. Establishment of a novel strategy to effectively characterize and subtype the virulence levels of S. suis strains are urgently needed.
Supplementary_Tables
Download Zip (45.3 KB)Supplementary_Figure_1
Download TIFF Image (7.7 MB)Acknowledgements
We are grateful to Jingdong Song (National Institute for Viral Disease Control and Prevention, Chinese Center for Disease Control and Prevention) for the technical support of transmission electron microscopy. HZ and JX designed the project; HZ drafted the manuscript; MG and ZW reviewed the manuscript. WK, WM, XY, JW, XZ, ZW, and HS carried out the experiments and generated the data; WK, YW, ZW, and HZ analyzed the data. All authors have read and approved the final version of the manuscript.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Additional information
Funding
References
- Segura M. Streptococcus suis research: progress and challenges. Pathogens. 2020;9(9). doi:10.3390/pathogens9090707
- Goyette-Desjardins G, Auger JP, Xu J, et al. Streptococcus suis, an important pig pathogen and emerging zoonotic agent-an update on the worldwide distribution based on serotyping and sequence typing. Emerg Microbes Infect. 2014;3(6):e45.
- Huong VT, Ha N, Huy NT, et al. Epidemiology, clinical manifestations, and outcomes of Streptococcus suis infection in humans. Emerg Infect Dis. 2014;20(7):1105–1114. doi:10.3201/eid2007.131594
- Ye C, Zhu X, Jing H, et al. Streptococcus suis sequence type 7 outbreak, Sichuan, China. Emerg Infect Dis. 2006;12(8):1203–1208. doi:10.3201/eid1708.060232
- Dong XX, Chao YJ, Zhou Y, et al. The global emergence of a novel Streptococcus suis clade associated with human infections. Embo Mol Med. 2021;13(7):1–11.
- Wang ML, Du PC, Wang JP, et al. Genomic epidemiology of Streptococcus suis sequence type 7 sporadic infections in the Guangxi Zhuang Autonomous Region of China. Pathogens. 2019;8(4):1–17.
- Marois C, Bougeard S, Gottschalk M, et al. Multiplex PCR assay for detection of Streptococcus suis species and serotypes 2 and 1/2 in tonsils of live and dead pigs. J Clin Microbiol. 2004;42(7):3169–3175. doi:10.1128/JCM.42.7.3169-3175.2004
- Ishida S, Tien le HT, Osawa R, et al. Development of an appropriate PCR system for the reclassification of Streptococcus suis. J Microbiol Methods. 2014;107:66–70. doi:10.1016/j.mimet.2014.09.003
- Wang J, Qi K, Bai X, et al. Characterization of integrative and conjugative elements carrying antibiotic resistance genes of Streptococcus suis isolated in China. Front Microbiol. 2022;13:1074844. doi:10.3389/fmicb.2022.1074844
- Huang JH, Chen L, Li DW, et al. Emergence of a carrying and multidrug resistant ICE in zoonotic pathogen. Vet Microbiol. 2018;222:109–113. doi:10.1016/j.vetmic.2018.07.008
- Liang P, Wang M, Gottschalk M, et al. Genomic and pathogenic investigations of Streptococcus suis serotype 7 population derived from a human patient and pigs. Emerg Microbes Infect. 2021;10(1):1960–1974. doi:10.1080/22221751.2021.1988725
- Han N, Miao JJ, Zhang TT, et al. MDACP: a pathogen genome and metagenome analysis cloud platform. Front Genet. 2020;11:1–5.
- Sitto F, Battistuzzi FU. Estimating pangenomes with roary. Mol Biol Evol. 2020;37(3):933–939. doi:10.1093/molbev/msz284
- Seemann T. Prokka: rapid prokaryotic genome annotation. Bioinformatics. 2014;30(14):2068–2069. doi:10.1093/bioinformatics/btu153
- Chen C, Zhang W, Zheng H, et al. Minimum core genome sequence typing of bacterial pathogens: a unified approach for clinical and public health microbiology. J Clin Microbiol. 2013;51(8):2582–2591. doi:10.1128/JCM.00535-13
- Ye C, Zheng H, Zhang J, et al. Clinical, experimental, and genomic differences between intermediately pathogenic, highly pathogenic, and epidemic Streptococcus suis. J Infect Dis. 2009;199(1):97–107. doi:10.1086/594370
- Holden MT, Hauser H, Sanders M, et al. Rapid evolution of virulence and drug resistance in the emerging zoonotic pathogen Streptococcus suis. PLoS One. 2009;4(7):e6072. doi:10.1371/journal.pone.0006072
- Huang J, Ma J, Shang K, et al. Evolution and diversity of the antimicrobial resistance associated mobilome in Streptococcus suis: a probable mobile genetic elements reservoir for other streptococci. Front Cell Infect Microbiol. 2016;6:118.
- Li M, Shen X, Yan J, et al. GI-type T4SS-mediated horizontal transfer of the 89K pathogenicity island in epidemic Streptococcus suis serotype 2. Mol Microbiol. 2011;79(6):1670–1683. doi:10.1111/j.1365-2958.2011.07553.x
- Liang ZJ, Wu HZ, Bian C, et al. The antimicrobial systems of Streptococcus suis promote niche competition in pig tonsils. Virulence. 2022;13(1):781–793. doi:10.1080/21505594.2022.2069390
- Oehlmann S, Krieger AK, Gisch N, et al. D-Alanylation of lipoteichoic acids in Streptococcus suis reduces association with leukocytes in porcine blood. Front Microbiol. 2022;13:1–17.
- Zheng H, Ji S, Liu Z, et al. Eight novel capsular polysaccharide synthesis gene loci identified in nontypeable Streptococcus suis isolates. Appl Environ Microbiol. 2015;81(12):4111–4119. doi:10.1128/AEM.00315-15
- Ji LY, Chen ZG, Li F, et al. Epidemiological and genomic analyses of human isolates of Streptococcus suis between 2005 and 2021 in Shenzhen, China. Front Microbiol. 2023;14:1–12.
- Zhou Y, Dong X, Li Z, et al. Predominance of Streptococcus suis ST1 and ST7 in human cases in China, and detection of a novel sequence type, ST658. Virulence. 2017;8(6):1031–1035. doi:10.1080/21505594.2016.1243193
- Xu J, Fu S, Liu M, et al. The two-component system NisK/NisR contributes to the virulence of Streptococcus suis serotype 2. Microbiol Res. 2014;169(7–8):541–546. doi:10.1016/j.micres.2013.11.002
- Zhao Y, Liu G, Li S, et al. Role of a type IV-like secretion system of Streptococcus suis 2 in the development of streptococcal toxic shock syndrome. J Infect Dis. 2011;204(2):274–281. doi:10.1093/infdis/jir261
- Li M, Wang C, Feng Y, et al. Salk/SalR, a two-component signal transduction system, is essential for full virulence of highly invasive Streptococcus suis serotype 2. PLoS One. 2008;3(5):e2080. doi:10.1371/journal.pone.0002080
- Huang W, Wang M, Hao H, et al. Genomic epidemiological investigation of a Streptococcus suis outbreak in Guangxi, China, 2016. Infect Genet Evol. 2019;68:249–252. doi:10.1016/j.meegid.2018.12.023
- Zheng X, Zheng H, Lan R, et al. Identification of genes and genomic islands correlated with high pathogenicity in Streptococcus suis using whole genome tiling microarrays. PLoS One. 2011;6(3):e17987. doi:10.1371/journal.pone.0017987
- Wang J, Liang P, Sun H, et al. Comparative transcriptomic analysis reveal genes involved in the pathogenicity increase of Streptococcus suis epidemic strains. Virulence. 2022;13(1):1455–1470. doi:10.1080/21505594.2022.2116160
- Norrby-Teglund A, Pauksens K, Norgren M, et al. Correlation between serum TNF alpha and IL6 levels and severity of group A streptococcal infections. Scand J Infect Dis. 1995;27(2):125–130. doi:10.3109/00365549509018991
- Faulkner L, Cooper A, Fantino C, et al. The mechanism of superantigen-mediated toxic shock: not a simple Th1 cytokine storm. J Immunol. 2005;175(10):6870–6877. doi:10.4049/jimmunol.175.10.6870
- Chabot-Roy G, Willson P, Segura M, et al. Phagocytosis and killing of Streptococcus suis by porcine neutrophils. Microb Pathog. 2006;41(1):21–32. doi:10.1016/j.micpath.2006.04.001
- Smith HE, Damman M, van der Velde J, et al. Identification and characterization of the cps locus of Streptococcus suis serotype 2: the capsule protects against phagocytosis and is an important virulence factor. Infect Immun. 1999;67(4):1750–1756. doi:10.1128/IAI.67.4.1750-1756.1999
- Xu ZM, Chen B, Zhang Q, et al. Streptococcus suis 2 transcriptional regulator TstS stimulates cytokine production and bacteremia to promote streptococcal toxic shock-like syndrome. Front Microbiol. 2018;9:1–12.
- Jiang XW, Yang YK, Zhou JJ, et al. Roles of the putative type IV-like secretion system key component VirD4 and PrsA in pathogenesis of Streptococcus suis type 2. Front Cell Infect Mi. 2016;6:1–13.