Streptococcus suis is one of the leading causes of bacterial infection in swine and contributes to significant economic losses to the swine industry worldwide (Feng et al., 2014; Segura et al., 2014a,b, 2017). S. suis causes a spectrum of clinical disease outcomes in pigs including pneumonia, endocarditis, septicemia, and meningitis. S. suis is also a zoonotic pathogen capable of causing diseases in humans, mainly arthritis, sepsis, and meningitis (Wertheim et al., 2009a,b; Feng et al., 2014; Segura et al., 2014a,b). Human infections have been either sporadic, and thought to be acquired from penetrating injuries associated with occupational exposure, or epidemic, and associated with consumption of raw or undercooked pork products, primarily in southeast Asia (Gottschalk et al., 2007; Dutkiewicz et al., 2018).

Virulence mechanisms used by S. suis are not extensively characterized (Fittipaldi et al., 2012; Segura et al., 2017). Studies addressing S. suis virulence have been confounded because different isolates cause a spectrum of disease outcomes ranging from lethal systemic disease to asymptomatic carriage (Segura et al., 2017; Votsch et al., 2018). Virulence-associated factors generally regarded as the most important contributors to disease are the capsular polysaccharide (CPS), muramidase-released protein (mrp), extracellular protein factor (epf), and suilysin (sly) (Fittipaldi et al., 2012; Segura et al., 2014a,b, 2017). However, none of these factors alone correlate with the ability to cause invasive systemic disease (Fittipaldi et al., 2012; Segura et al., 2017). Additionally, invasive clinical isolates obtained from both people and pigs frequently do not harbor all of these factors (Fittipaldi et al., 2012; Feng et al., 2014; Segura et al., 2014a,b, 2017). In addition to being considered a major virulence-factor for S. suis, CPS is the basis for the typing system of S. suis (Liu et al., 2013). Genotyping isolates using epf + /mrp + /sly + as markers was historically used in epidemiological studies to predict the virulence potential of S. suis isolates, particularly serotype 2 isolates obtained from European countries (Gottschalk et al., 1998; Wisselink et al., 2000; Fittipaldi et al., 2012).

Serotyping and multilocus sequence typing (MLST) are the most common methods used to differentiate S. suis isolates. Isolates are serotyped based on structural differences in the CPS and to date there are 27 recognized serotypes based on the antigenicity of CPSs and 28 novel CPS loci (Pan et al., 2015; Qiu et al., 2016; Zheng et al., 2017; Huang et al., 2019; Bojarska et al., 2020). Among these, serotype, serotype 2 is regarded as the most virulent and the most frequently isolated serotype from sick pigs (Goyette-Desjardins et al., 2014; Zheng et al., 2014). Additionally, two large outbreaks of human S. suis serotype 2 infection have elevated public health concerns (Tang et al., 2006; Yu et al., 2006). A standard MLST scheme has been developed for S. suis based on the nucleotide sequences of seven housekeeping genes (King et al., 2002), with isolates subsequently assigned a sequence type (ST) based upon their combination of unique allelic sequences. Worldwide ST1, ST25, and ST28 are the most prevalent STs isolated from swine, the predominance of each ST tends to vary with geographic location (Fittipaldi et al., 2011). For example, in Europe and Asia ST1 strains are more commonly isolated from swine, whereas in North America, ST25 and ST28 are more prevalent (de Greeff et al., 2011; Fittipaldi et al., 2011; Onishi et al., 2012; Schultsz et al., 2012; Zhu et al., 2013; Estrada et al., 2019).

Phylogenetic studies employ whole-genome sequencing (WGS) and data analysis to better understand the epidemiology, origin, and evolution of bacteria (Eppinger et al., 2014; Sahibzada et al., 2017). WGS analysis can be used to assess isolate relatedness or determine genetic characteristics that define subsets of isolates allowing for the determination of phylogenetic relationships among bacterial isolates providing key data for epidemiological studies addressing biosurveillance and risk assessment (Eppinger et al., 2011, 2014; Rusconi et al., 2016; Sahibzada et al., 2017). Numerous studies have undertaken WGS in combination with comparative genomic approaches to evaluate the genomic diversity harbored by S. suis isolates (Weinert et al., 2015; Wileman et al., 2019; Hadjirin et al., 2021; Estrada et al., 2022). The U.S. is the third-largest producer and consumer of pork and pork products globally and unfortunately the vast majority of these studies utilize S. suis isolates derived from countries other than U.S. (Princivalli et al., 2009; de Greeff et al., 2011; Wang et al., 2011; Goyette-Desjardins et al., 2014; Athey et al., 2015, 2016b; Weinert et al., 2015; Wileman et al., 2019; Scherrer et al., 2020; Guo et al., 2021; Hadjirin et al., 2021; Liang et al., 2021; Estrada et al., 2022). The goal of the current study was to begin to fill this gap by evaluating the genetic relatedness, population structure, and genetic diversity, relating to antimicrobial resistance (AMR) and virulence-associated factors, of S. suis isolates obtained within the U.S. using a core genome (cg) MLST-based methodology. Additionally, S. suis isolates obtained from various countries outside the U.S. were included in our analyses to provide a broader perspective and examine how the population structure of U.S. S. suis isolates compares with isolates obtained from various countries outside the U.S.

A total of 148 S. suis from the U.S. and various countries outside the U.S. were comprehensively analyzed in this study. International isolates were included if complete genome assemblies were publicly available, allowing for the in silico screening for AMR- and virulence-factors, as well as sequence type (ST), CPS serotype, and clonal complex (CC) determination. Strain ISU1606 was used as the reference or seed strain because it shares numerous traits and characteristics with S. suis strain P1/7 (Holden et al., 2009), which is routinely used a reference strain. ISU1606 and P1/7 are ST1, serotype 2 isolates with similar chromosome size, total number of predicted protein coding sequences (CDSs), G + C content, global synteny, and a similar ability to cause similar clinical disease in swine (Nicholson et al., 2020). To begin evaluating population structure and relatedness among S. suis isolates obtained from the U.S. (n = 103) and from various countries outside the U.S. (n = 45), a neighbor-joining tree was constructed from cgMLST targets using the S. suis strain ISU1606 as the seed (Figure 1). The combined cgMLST target gene inventory was determined at 1,879 genes and was comprised of 565 core and 1,314 accessory genes (Supplementary Table 2). The cgMLST-based tree topology shows a high degree of genome plasticity among all the S. suis isolates, with the isolates obtained from countries outside the U.S. interspersed throughout the tree (Figure 1). The extensive genome plasticity depicted by the tree topology is reflective of the high genomic variation and diversity previously described for S. suis isolates (Princivalli et al., 2009; de Greeff et al., 2011; Weinert et al., 2015; Shelyakin et al., 2019; Guo et al., 2021). Additionally, the S. suis isolates obtained within the U.S. did not cluster by state or geographic location (Figure 1).

To examine the nucleotide-level relatedness among the isolates, ANI values were determined for all isolates when compared to S. suis strain ISU1606. The ANI values ranged from 99.99% (P1/7) to 94.42% (29995) with an average of 97.12% (Supplementary Table 5). The broad range of ANI values illustrates the extensive genomic variation among the sampled isolates given that ANI values of ≥95% are typically observed for organisms belonging to the same species (Richter and Rossello-Mora, 2009; Kim et al., 2014). The ANI values among the sampled S. suis isolates correlate with the phylogenetic group distribution of the cgMLST-based tree (Figure 1). For example, isolates with ANI values ranging 99.99 to 99.70% (P1/7, S10, 10, S735, A7, ZY05719, ISU2714, GZ1, LSM157, 40533, 40424, SC84, SS2-1, SC070731, LSM102, LSM178, JS14, BM407, 05ZYH33) from a distinct cluster with a single branch point (Figure 1 and Table 1). Another example includes the isolates with ANI values ranging 94.70 to 94.42% (ISU2912, 40426, 40445, LS9N, 29995) that are grouped together with only few branch points (Figure 1 and Table 2).

Two additional cgMLST-based trees were constructed based on the core genes (Supplementary Figure 1) and the accessory genes only (Supplementary Figure 2). The resulting core-genome based tree resulted in an overall tree topology similar to the cgMLST-based tree (Figure 1 and Supplementary Figure 1). This finding is not surprising as the phylogenetic input is dominated by the shared core genes rather than genes classified as accessory. The accessory-based tree showed in general similar relatedness of isolates, however, reposition of some isolates were observed in the same branches (Figure 1 and Supplementary Figure 2). For example, isolates ZY05719, SS2-1, and SC84 are clustered together from a branch point in the cgMLST-based tree (Figure 1) and in the accessory-based tree these isolates are also clustered together from a branch point but in a repositioned order (Supplementary Figure 2).

Overall isolates with similar CCs and serotypes, obtained from within and outside the U.S., were observed were observed to cluster together (Figure 1). For example, a group of 13 isolates mainly consisting of CC87 (ST19, ST87, and ST1181) were clustered together in Figure 1. Eleven of these 13 isolates were CPS serotype 8. The two non-serotype 8 isolates were T15 (serotype 2, ST19, and CC87) and 37904 (serotype 7, ST89, and CC87) (Figure 1). Additionally, serotype 2 isolates mainly clustered together and are partitioned in two distinct groups. One group largely consists of CC28 (ST28, ST620, ST787, ST961, and ST1180) isolates and a second group that mainly consisting of CC1 (ST1, ST7, ST658, ST665) isolates (Figure 1). The CC28- dominated group forms a large clade of 45 isolates, all serotype 2, except for nine (Figure 1). Of the nine non-serotype 2 isolates, D9 and 13-00285-02 (serotype 7, ST29, and CC29) cluster together and next to four serotype 3 and CC27 isolates: 40464, ST3, YB51, and 30815 (Figure 1). These two small clusters separate isolates ISU2514 and NSU1060 both serotype 2, ST25, and CC25 from the other thirty-one isolates, which were serotype 2 and CC28. Twenty-one of these serotype 2 isolates are ST28 and five are ST961. Four isolates within this subclade of CC28 isolates are non-serotype 2 isolates. These non-serotype 2 isolates are 30192, 40523, and 40461, all serotype 1/2, and isolate 40471, which harbor an undetermined serotype (Figure 1).

The other group of serotype 2 isolates consisted of twenty-five isolates, which is further divided into two subclades. The smaller subclade consists of five CC17 (ST17 and ST20) and CC16 (ST16) isolates and included isolates 861160 and GD-0001 (both serotype 2, ST20, CC17), isolate GD-0088 (serotype 9, ST16, CC16), as well as isolates 40439 and 40458 (both serotype 4, ST17, CC17). The larger subclade consists of 20 isolates, all CC1 (ST1, ST7, ST658, ST665) except for isolate LSM178, an ST1005 isolate not assigned to a CC and therefore listed as a singlet. Four of the twenty isolates are U.S. isolates (40424, 40533, ISU1606, and ISU2714), while the majority (16/20) of isolates within this subclade are non-U.S. isolates. Additionally, 17 isolates of this subclade were serotype 2 and three isolates were serotype 14 (40424, 40533, and JS14). Isolates within this subclade were mainly ST1 (9/20) or ST7 (6/20).

Genomes were screened for chromosomal mutations and genes conferring antimicrobial resistance. For the U.S. isolates, the number of AMR elements harbored by an individual isolate ranged from one to twelve (Figure 2 and Supplementary Table 3). The average number of AMR elements harbored by an individual U.S. isolate was three (Figure 2 and Supplementary Table 3). The number of AMR elements harbored by an individual isolate ranged from one to nine for the non-U.S. isolates (Figure 2 and Supplementary Table 3). The average number of AMR elements harbored by an individual non-U.S. isolate was two, which was significantly (P = 0.038) lower than number of AMR elements harbored by an individual U.S. isolate (Figure 2 and Supplementary Table 3). While numerous AMR elements were identified, the most prevalent were ble (100% in both U.S. and non-U.S.) followed by tetO (87% U.S. and 49% non-U.S.), and ermB (65%, U.S. and 49% non-U.S.). Both the tetO and ermB loci are absent in 27 genomes, eleven of which were serotype 2 isolates grouped within the larger subclade CC1 isolates in Figure 1.

Genes epf, mrp, and sly have historically been used to predict the virulence potential of S. suis (Gottschalk et al., 1998; Wisselink et al., 2000; Fittipaldi et al., 2012). Thus, all genomes were screened for the presence of these loci and the percent identity relative to the reference orthologs was determined. The epf gene is present in four U.S. isolates (40424,40533, ISU1606, and ISU2714) with a 99.9% identity to the reference ortholog. None of the four isolates encode the long form of extracellular protein factor (EF*), which can be expressed by some serotype 2 isolates and produce a high molecular weight variant (>110 kDa) (Galina et al., 1996; Gottschalk et al., 1998). This gene is further found in sixteen non-U.S. isolates with a sequence identity range from 91.37 to 99.9%. Two of these non-U.S. isolates, BM407 and S735, carry epf* (Galina et al., 1996; Gottschalk et al., 1998). The mrp gene was found in 86 U.S. isolates and the nucleotide sequence diversity ranged from 31.47 to 100%, with an average sequence identity of 80.21% (Supplementary Table 4). Among non-U.S. isolates, the mrp gene was found in 37 isolates and, similar to U.S. isolates, a high nucleotide sequence diversity was observed ranging from 29.76 to 100% with an average 73.72% sequence identity (Supplementary Table 4). The mrp gene for nine (13-00283-02, D9, JS14, LSM157, LSM178, NSUI060, SC070731, ST3, T15) of the 37 isolates was predicted to encode a frameshift resulting in a pseudogene. Extensive nucleotide diversity with the mrp gene has been previously reported (Vecht et al., 1991; Galina et al., 1996; Baums and Valentin-Weigand, 2009). The sly gene was found in 45 U.S. isolates and is highly conserved, with a sequence identity range from 99.2–100% and an average 99.78% sequence identity (Supplementary Table 4). Twenty-four non-U.S. isolates carry the sly gene and, similar to U.S. isolates, feature high nucleotide sequence diversity with an average 99.9% sequence identity (99.4–100%) (Supplementary Table 4).

Among the total 148 analyzed isolates, 19 [15 non-U.S. and four U.S. isolates (Table 2) carry all three epf, mrp, and sly genes]. These nineteen isolates form a larger subclade mainly comprised of serotype 2 isolates, as previously mentioned (Figure 1). As stated above, all isolates are CC1 except for isolate LSM178, an ST1005 isolate not assigned to a CC and therefore listed as a singlet (Table 2). Sixteen of these isolates were serotype 2 and three isolates were serotype 14 (40424, 40533, and JS14) and the majority of these isolates were ST1 (10/19) or ST7 (6/19) (Figure 1 and Table 2).

Diseases caused by S. suis are a substantial economic problem as well as an animal welfare concern for the global swine industry. Additionally, there are increasing public health concerns regarding the zoonotic potential of S. suis. While numerous studies have employed a combination of comparative genomic approaches to assess the genomic diversity and population structure of S. suis isolates globally, the overwhelming majority of these studies utilize S. suis isolates obtained from countries other than the U.S. (Princivalli et al., 2009; de Greeff et al., 2011; Wang et al., 2011; Goyette-Desjardins et al., 2014; Athey et al., 2015, 2016b; Weinert et al., 2015; Wileman et al., 2019; Scherrer et al., 2020; Guo et al., 2021; Hadjirin et al., 2021; Liang et al., 2021; Estrada et al., 2022). In this report we examined the genomic diversity, phylogenetic relatedness, and population structure of S. suis isolates obtained within the U.S. by utilizing a cgMLST approach. Additionally, S. suis isolates obtained from various countries outside the U.S. were included in our analyses to provide a broader perspective and examine how the population structure of U.S. S. suis isolates compares with isolates obtained from countries outside the U.S.

The resulting phylogenetic tree from the cgMLST-based typing method demonstrated the extensive genomic variation and diversity among all the sampled S. suis isolates, with isolates obtained from countries outside the U.S. found interspersed throughout the cgMLST-based tree. The observed extensive genome plasticity in the established cgMLST-phylogeny is reflective of the high genomic variation and diversity previously described for S. suis (Princivalli et al., 2009; de Greeff et al., 2011; Weinert et al., 2015; Shelyakin et al., 2019; Guo et al., 2021). In fact, S. suis is considered to have an incompletely defined or open pan-genome since the number of unique or accessory genes has increased and the number of conserved genes have decreased as more genome sequences become available (Wang et al., 2011; Shelyakin et al., 2019; Guo et al., 2021).

Average nucleotide identity has increasingly been utilized as a robust method for assessing genomic similarity between two genomes, with organisms belonging to the same species typically having ANI values of ≥95%. The ANI values determined for the S. suis isolates in this study were extensively broad, ranging from 99.99% (P1/7) to 94.42% (29995), and approach the recommended species demarcation value (Richter and Rossello-Mora, 2009; Kim et al., 2014). The broad range of ANI values observed illustrate the extensive genomic variation among the isolates and correlates with the group distribution in the cgMLST-based tree. The plasticity in genome composition for the sampled S. suis isolates is also evidenced in the genome size, ranging from ISU2912 with a size of 2,720,381 bp and thus 31.17% larger than ISU1606 to S735 with a size of 1,980,887 bp (4.49% smaller than ISU1606).

While the S. suis isolates obtained within the U.S. did not cluster by state or geographic location, in general all isolates with similar serotypes, regardless of their origins, cluster together in the cgMLST phylogeny. An exception to this general trend were the serotype 2 isolates, which clustered into two distinct groups. One group mostly consisted of ST28 and ST961 and contained three serotype 1/2 isolates and one isolate with an undetermined serotype. The second group, consisting of ST1 and ST7, had three serotype 14 isolates clustered within the group. The clustering of these isolates harboring the 1/2, 2, and 14 serotypes is reasonable and consistent with the high degree of similarly among these serotypes. Specifically, the CPS loci of serotypes 1 and 14 and serotypes 2 and 1/2 are almost identical and most current PCR methods are unable to differentiate these specific serotypes (Okura et al., 2013). Their CPS structure differs by the substitution of only a galactose (Gal) for a N-acetylgalactosamine (GalNAc) due to a single nucleotide polymorphism in the glycosyltransferase cpsK gene (Van Calsteren et al., 2016; Roy et al., 2017). Further, serotype switching of field strains of serotypes 2 and 1/2, and 14 and 1, solely by replacing the amino acid 161 of CpsK has been demonstrated (Roy et al., 2017). In contrast, the CPS loci of all other serotypes are greatly different (Pan et al., 2015; Zheng et al., 2017; Huang et al., 2019).

The clustering of serotype 2 isolates into two distinct groups is supported by previous work that inferred the population structure using MLST and CC approaches (Okura et al., 2013; Hatrongjit et al., 2020; Liang et al., 2021). For example, one branch of serotype 2 isolates within the cgMLST-based tree consisting of ST25, ST27, ST35 isolates further branched into ST28, ST620, ST961, and ST1180, which are the STs comprising one of the serotype 2 groups. Based on CC methods, CC25 (consisting of ST25 isolates), CC27 (consisting of ST27 and ST35 isolates), and CC28 (consisting of ST28, ST620, ST961, ST1180, and ST787) are all linked together (Okura et al., 2013; Hatrongjit et al., 2020; Liang et al., 2021). Additionally, CC1 (consisting of ST1 isolates) and CC7 (consisting of ST7 isolates) are linked (Okura et al., 2013; Hatrongjit et al., 2020; Liang et al., 2021). Consistent with this, the second branch of serotype 2 isolates is comprised of ST1 and ST7 isolates. The results demonstrate that overall, there is a high degree of similarity and congruence between S. suis population structure based on previously CC methods and the cgMLST-based strategy utilized in this study. However, the established cgMLST phylogenies provide additional resolution as evident by the observed expanded branching and network.

In addition to the phylogenetic relatedness and population structure of S. suis isolates, AMR elements and virulence-associated genes were analyzed. Numerous AMR elements were identified among both U.S. and non-U.S, the most prevalent were ble, tetO, and ermB. However, the tetO and ermB genes were not observed in the genomes of isolates grouped within the larger subclade CC1 isolates shown in Figure 1. The epf, mrp, and sly genes, historically used as markers for virulence potential, were also observed in the genomes of isolates grouped within the larger subclade CC1 isolates shown in Figure 1.

Taken together, the data provided in this report demonstrate that the extensive genomic variation and diversity observed among S. suis isolates globally is also reflected among S. suis isolates obtained within the U.S. A high-resolution phylogenetic framework for S. suis can help provide data needed to improve risk assessments and address public health concerns associated with occupational exposure to S. suis.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary material.

Ethical approval was not required for the study involving animals in accordance with the local legislation and institutional requirements because all isolates were either obtained from samples collected as part of previous studies or were obtained from samples submitted as part of field case investigations and did not require Institutional Animal Care and Use Committee (IACUC) approval.

All authors conceived and designed the experiments, performed the experiments, analyzed the data, contributed reagents, materials, and analysis tools, wrote the manuscript, approval of the final version to be published, and agreed to be accountable for all aspects of the work.