Population structure, genetic diversity and pathotypes of Streptococcus suis isolated during the last 13 years from diseased pigs in Switzerland.

Streptococcus (S.) suis is a globally important swine pathogen, which comprises certain zoonotic serotypes. In this study, a detailed characterization of 88 porcine S. suis isolates was performed by analyzing capsular (cps) types, multilocus sequence typing (MLST) and investigation of the minimum core genome (MCG). In order to focus on the virulence potential of presumable invasive disease-associated S. suis isolates, virulence-associated gene profiles were assessed followed by screening a chosen subset of S. suis strains with a molecular pathotyping tool. Results showed a high genetic variability within this strain collection. In total, seventeen cps types were identified with a predominance of cps type 9 (15.9%) and 6 (14.8%). MLST revealed 48 sequence types (STs) including 41 novel ones. The population structure of S. suis was heterogenous and isolates belonged to eight different clonal complexes (CCs) including CC28 (9.1%), CC1109 (8%), CC13/149 (6.8%), CC1237 (5.7%), CC1 (3.4%), CC17 (3.4%), CC87 (2.3%), and CC1112 (1.1%), whereas a significant portion of isolates (60.2%) could not be assigned to any described CCs. Virulence-associated markers, namely extracellular protein factor (epf), muramidase-released protein (mrp), and suilysin (sly), showed a link with STs rather than with cps types. With this study an expanded knowledge about the population structure and the genetic diversity of S. suis could be achieved, which helps to contribute to an optimal public health surveillance system by promoting a focus on strains with an increased virulence and zoonotic potential.

Introduction

Streptococcus suis (S. suis) is a Gram-positive bacterium recognized to be an important swine pathogen responsible for various diseases including septicemia with sudden death, meningitis, endocarditits, polyserositis, and arthritis [1]. S. suis also colonizes the upper respiratory tract with commensal strains giving rise to carrier pigs [2]. Moreover, S. suis can appear as an opportunistic pathogen in combination with other pathogens [3]. In addition, infection with S. suis not only leads to major economic problems due to important losses in pig production worldwide, but the bacterium is also capable of infecting humans. In Western countries, infection of humans usually involves a single person such as pig farmers, veterinarians, people who work at abattoir, or butchers. Infection mainly takes place through direct contact of skin wounds with contaminated pork even in cases where no obvious wound is present [4]. On the other hand, in Southeast Asian countries, the main route of infection seems to be the gastrointestinal tract where consumption of undercooked pork products or a traditional soup containing raw pig blood is common and humans frequently live in close contact to pigs [5]. The most frequently described manifestations in humans are meningitis and septic shock [6].

Currently, there are at least 29 S. suis serotypes [7] described based on a serological reaction against the capsular polysaccharide, which has been commonly reported to be a major virulence factor with antiphagocytic properties [8]. Strains of heterogeneous serotype 2, which are responsible for more than 80% of human cases, represent a globally emerging zoonotic threat [9]. To a lesser extent, serotype 14 is also associated with disease in both humans and pigs [10]. Worldwide, major S. suis serotypes isolated from diseased pigs are serotypes 2, 9, 3, 1/2 and 7 [9]. In some European countries, serotype 9 strains are emerging and are considered to be responsible for causing disease among pigs [11] and sporadically capable to infect humans as reported in one case report from Thailand [12].

The most commonly described virulence markers for S. suis include muramidase-released protein (MRP) [13], extracellular protein factor (EF) [14], and the hemolysin suilysin (SLY) [15], which were mainly associated with a virulence potential of S. suis serotype 2 strains [16]. MRP is a surface protein anchored to the cell wall [17] and has recently been shown to be a major fibrinogen-binding protein [18]. Human fibrinogen bound to MRP increases viability of S. suis serotype 2 in human blood and its capability of migrating across the human cerebral endothelial cell barrier, thereby promoting the development of meningitis [19]. EF is secreted independent of an interaction between bacteria and host cell involved in enhancing bacterial infection, invasion, and pathogenicity [20]. Both MRP and EF have variants of different molecular weights [21, 22]. Strains harboring the short form of epf are considered the most virulent form of the epf subtypes [21]. SLY is a thiol-activated toxin antigenically related to cholesterol-binding toxins, which forms transmembrane pores [23] and is described to be toxic to phagocytic, endothelial, and epithelial cells [24]. However, since isogenic mutants can be found, which do not express MRP, EFP, and/or SLY and are still virulent, it is more precise to term those factors as virulence-associated markers [14].

Multilocus sequence typing (MLST) is a useful method to examine the population structure and provide epidemiological tracing of pathogenic bacteria. It has been applied to various species and is a valuable tool that allows further investigation of the global distribution and evolution of several pathogens [25], including S. suis as reported in former studies [26, 27]. The classification of sequence types (STs) resulting from MLST allows grouping of genetically similar genotypes in clonal complexes (CCs), thus revealing the population structure of S. suis. Various CCs are identified globally, whereas the prevalence of certain genotypes can vary in different geographical regions. The most important CCs implicated in human infections are the following: CC1 (harboring most important ST1) in Europe, Asia, North- and South America, and Africa [28]; CC20 in the Netherlands [11]; CC25 and CC28 in North America, Australia, and Asia [28], although in the United States an increasing trend of the genetically diverse CC94 was observed [29]. In Europe, a high prevalence of virulent mrp + epf + sly + cps2 strains being part of CC1 and mrp*cps9 strains of CC16/87 can be noticed [30].

A molecular surveillance tool to predict the virulence potential of S. suis, by differentiating potentially invasive disease-associated strains from commensal or possibly opportunistic non-disease-associated strains is of advantage. Such a pathotyping tool was recently described by Wileman et al. [31] and, provided that its application is successful, could be used as a convenient and promising diagnostic assay enabling the detection of highly virulent emerging strains. Based on whole genome sequencing data, three pathotyping markers were identified [31]. First of all, a disease-associated marker was described, which is predicted to be a copper-exporting ATPase (SSU207) potentially playing an essential role in copper homeostasis in Gram-positive and -negative bacterial pathogens [31, 32]. The second proposed disease-associated marker is annotated as type I restriction-modification (RM) system S protein (SSU1589) in S. suis strain P1/7 (GenBank accession number NC_012925.1) and is supposed to be implicated in the defense of host bacteria [33]. The third genetic marker is thought to be non-disease-associated and the corresponding protein is suggested to be a putative sugar ATP-binding cassette (ABC) transporter (SSUST30534).

The aim of this study was to determine the genetic diversity of Swiss porcine S. suis isolates collected during the past 13 years and to assess their virulence potential. Observed phenotypes arising from pathological, histological and bacteriological examinations were compared to the virulence-associated genotype profiles obtained by analyzing the classical virulence-associated markers (mrp, efp and sly) and by classifying the S. suis population using a system based on minimum core genome (MCG). Moreover, a selected subset of S. suis strains, based on a comprehensive data availability of histopathological findings of the corresponding clinical cases, was screened with an available novel pathotyping tool [31]. The identification of major pathogenic strains could help to establish an optimal public health surveillance system by promoting a focus on strains with an increased virulence and zoonotic potential.

Materials and methods

Streptococcus suis isolates

In this study a total of 88 porcine isolates obtained from routine diagnostic submissions to the Department of Veterinary Bacteriology at the Vetsuisse Faculty, University of Zurich, within the time span of 2006–2019, were included. The samples were recovered from 77 independent farms located in three main regions of Switzerland (Central- and Eastern part, and Canton of Zurich) with one to three samples per farm always derived from different animals. No isolates came from the French speaking Western part of Switzerland and the Canton of Bern. Isolates of the same farm were included only if STs were different. Samples analyzed originated from the following sites: 34.1% (n = 30/88) from blood, 20.5% (n = 18/88) from brain, 14.8% (n = 13/88) from lung, 11.4% (n = 10/88) from joint, 8% (n = 7/88) from heart, 3.4% (n = 3/88) from genital tract, 2.3% (n = 2/88) each from abscess and kidney, and 1.1% (n = 1/88) each from abdominal cavity, nasal swab and liver (Table 1). In order to verify the relevance of the S. suis isolates, the results of the histological examination were consulted. S. suis infection was considered as relevant in cases where the histological diagnoses included meningitis, polyarthritis, endocarditis, or placentitis, and when histological features consistent with septicemia [34] were observed. In two cases no histopathology was performed, but S. suis was determined relevant due to clear bacteriological findings (pure culture of S. suis), appropriate anamnesis and clinical signs. Isolates, for which a clear link to S. suis could be ruled out histopathologically (due to inappropriate age or due to another etiological diagnosis), were termed as not relevant. Isolates with an incomplete record of data, no S. suis relevance could be assigned and were therefore excluded for screening with the pathotyping tool.

Table 1

Number of porcineisolates from corresponding site of isolation.

Site of isolation	Number of isolates
Blood	30
Brain	18
Lung	13
Joint	10
Heart	7
Genital tract	3
Abscess	2
Kidney	2
Abdominal cavity	1
Nasal swab	1
Liver	1
Total	88

DNA extraction and species identification

Isolates were cultured for 24 to 48 h at 37 °C on Columbia blood sheep agar plates (Thermo Fisher Diagnostics AG, Pratteln, Switzerland) under aerobic conditions. DNA was extracted using a standard heat lysis protocol [35]. Briefly, colony material was dissolved in lysis buffer containing Proteinase K at 60 °C for 45 min under constant shaking followed by a heat deactivation step at 96 °C for 15 min. DNA samples were diluted to 20 ng/µl for subsequent reactions.

The isolates were identified by matrix-assisted laser desorption ionization-time-of-flight mass spectrometry (MALDI-TOF MS, Bruker, Bremen, Germany). In addition, to confirm S. suis species, a recN (recombination/repair protein gene) PCR was performed as described earlier [36].

Capsular typing and MLST

Capsular typing was performed as previously described [37]. Universal primers targeting 16S rRNA gene were used for internal control of the PCR reactions [38]. PCR products were analyzed on a capillary electrophoresis QIAxcel Advanced device (Qiagen, Hilden, Germany) using a screening cartridge, QX 15 bp–3 kb alignment marker and QX 100 bp–2.5 kb size marker (Qiagen) according to the manufacturer’s instructions. The resulting electropherograms were viewed with the QIAxcel ScreenGel 1.2.0 software (Qiagen). Isolates with capsular (cps) type pair 2 or 1/2 and 1 or 14 were assigned to the corresponding cps type using a recently developed HRM assay [39].

MLST was performed targeting aroA, cpn60, dpr, gki, recA, thrA and mutS as described previously [26]. For strains with no PCR amplification using primer mutS, alternative mutS primers have been used as described [40]. Briefly, PCR was set up in a final volume of reaction mixture of 20 µl containing HotStartTaq Master Mix (Qiagen), 0.5 µM of each primer pair and 20 ng template DNA with cycling conditions as described [26]. Sequencing of the seven housekeeping genes was performed by Sanger sequencing (Microsynth, Balgach, Switzerland). CLC Main Workbench 7.0.1 was used for sequence analysis. STs were assigned based on PubMLST database (https://pubmlst.org/ssuis). New housekeeping gene alleles were submitted to the database in order to create new STs. For visualizing groups of related genotypes and for identification of CCs, Swiss isolates were clustered at triple locus variant level with all isolates from the S. suis database (2289 isolates on September 27, 2019). The goeBURST algorithm from Phyloviz software [41] was used to visualize CCs and to create a MLST-based minimal spanning tree. CCs comprised STs with at least six identical alleles. STs that did not form any group were defined as singletons and unrelated STs grouping only with individual STs were indicated as non-defined founders.

Virulence-associated gene profiling and MCG typing

All S. suis isolates were tested for three virulence-associated genes (mrp, epf, sly) by multiplex PCR using primers as described previously [22]. Additionally, a epf monoplex PCR [22] was conducted to ensure reliable detection of large variants of the epf gene designated as epf* [21]. To accurately differentiate size variants of mrp, a PCR was performed using the mrp variant primers as described previously [22]. The PCR assays were performed using HotStartTaq Master Mix (Qiagen) according to the manufacturer’s instructions. MCG sequence typing was performed as described previously [42].

Pathotyping

Pathotyping was carried out using a recently developed multiplex PCR [31] by analyzing a selected subset of 75 isolates, which could be linked without any doubt to either etiologically relevant S. suis diseased or not S. suis relevant diseased pigs (Table 2). Briefly, a multiplex PCR was conducted using primers targeting two disease-associated genes (copper exporting ATPase 1 and type I RM system S protein), one non-disease-associated marker (putative sugar ABC transporter), and S. suis specific primers (sporulation regulator WhiA). The multiplex PCR was executed using HotStartTaq Mix (Qiagen) applying the three-step thermal cycling program as described previously [31]. PCR products were analyzed on a capillary electrophoresis QIAxcel Advanced device (Qiagen) using a screening cartridge, QX 15 bp–3 kb alignment marker and QX 100 bp–2.5 kb size marker (Qiagen) according to the manufacturer’s instructions.

Table 2

75 porcineisolates evaluated with the pathotyping tool.

Isolates were chosen based on the fact of an unequivocal link of etiologically relevant S. suis diseased or non-disease-associated pigs. Pathotyping results, which did not correspond to the results obtained by histopathological findings, are indicated in red.* indicates two isolates, which were included in the evaluation although a record of histopathological findings was missing, but due to evident bacteriological findings and appropriate anamnesis, strains were evaluated as relevant. mrp variants identified: mrpS (747 bp), mpr (1148 bp), mrp* (1556 bp), mrp*** (2400 bp). epf* represents a large variant of epf, which was detected only in one isolate (19-1715-1). +* relates to the pathotyping tool representing a variant PCR product of copper exporting ATPase gene.

aNontypeable.

bNot determined (nd): one lacking housekeeping gene.

cNo clonal complex could be assigned for ST that occurred as singletons (S) or with no determined founder (ndf).

dUngroupable: allele missing.

eNovel: allel combination not described for MCG sequencing typing by Zheng et al. [42].

fWileman et al. [31].

Whole genome sequencing of a selected panel of isolates

Whole genome sequencing (WGS) of a selected panel comprising 13 Swiss clinical S. suis strains of porcine origin has been performed previously [43]. Strains were chosen based on the aim to cover isolates of the most prevalent cps types (cps6 and cps9) and cps types most likely supposed to be virulent (cps1/2, cps1 and cps2) taking into consideration different years of sampling. Briefly, genomic DNA was isolated with the DNA Blood and Tissue Kit (Qiagen) and sequencing was performed on a MiniSeq sequencer (Illumina, San Diego, USA) with 150-bp paired-end reads using the Nextera DNA Flex kit (Illumina). Sequences are available under GenBank SRA accession numbers SRR8290472-SRR8290486.

Results

Pathological and bacteriological findings

In total, 88 isolates were obtained. Detailed histopathological investigations and bacteriological examinations revealed 52 isolates as S. suis relevant and 23 isolates as not S. suis relevant. For 13 isolates no detailed records were available and therefore, were not included in the comparison with the pathotyping tool. The most frequently reported history was found to be occurrence of sudden death of one or more pigs on a farm (n = 29/88) and observations of neurological symptoms (n = 24/88). In terms of post-mortem findings, the most commonly identified reason of death found by pathologists was septicemia (n = 27/75) and meningitis (n = 27/75).

Characterization of S. suis isolates by cps typing and MLST

Cps typing of the isolates identified a distribution of seventeen different cps types comprising 89.8% of isolates, whereas in 10.2% of cases no cps type could be assigned and were therefore classified as nontypeable (NT) (Additional file 1). Capsular gene analysis yielded two main cps types: 15.9% cps9 (n = 14) and 14.8% cps6 (n = 13). Furthermore, the following cps types were identified: cps1 (n = 6), cps15 (n = 6), cps16 (n = 6), cps7 (n = 4), cps2 (n = 4), cps8 (n = 3), cps3 (n = 4), cps1/2 (n = 3), cps12 (n = 3), cps4 (n = 3), cps31 (n = 3), cps5 (n = 2), cps21 (n = 2), cps28 (n = 2), cps19 (n = 1), and NT (n = 9). Clinically observed S. suis relevant phenotypes confirmed by histopathological analysis was detected for isolates of cps1, cps2, cps6, and cps7. In contrast, non-relevant S. suis features could explicitly be associated with isolates of cps8.

MLST analysis revealed that 22.7% of isolates belonged to eight previously identified STs, while 77.3% had a new MLST profile encompassing 41 new STs. The predominant STs were ST1104 (cps6, n = 13/13) comprising 14.8% of isolates, followed by 9.1% of ST1105 (cps9, n = 8/14), and 6.8% of isolates comprised ST13 (cps1, n = 6/6). The newly assigned STs in this study were ST1099–ST1130, ST1133, ST1134, ST1237–ST1240, ST1243, and ST1244 (Additional file 2). Three isolates of cps2 (n = 3/4) were an atypical single locus variant (SLV) of ST1 (ST1103), whereas one cps2 isolate (n = 1/4) was identified as a relatively common ST28 (Figure 1). The distribution of STs in relation to cps type highlights the diversity of S. suis. Thirteen of the seventeen cps types identified enclosed multiple STs, ranging from two to six different STs per cps type. Cps1, cps4, cps6, and cps19 only harbored one ST, whereas cps16 showed the greatest variety of six different STs (Figure 1).

Figure 1

Sequence type (ST) distribution of porcinein relation to capsular type. Identified STs are shown in form of a stacked histogram and illustrated in the corresponding bar sections. Nontypeable (NT) isolates could not be identified by multiplex PCR. Isolates with one lacking housekeeping gene identification could not be determined (nd).

Distribution of STs in relation to year of isolation

Figure 2 illustrates a MLST-based minimal spanning tree of all tested isolates within a time span of 2006–2019. ST1104 (cps6, n = 13/13) and ST1105 (cps9, n = 8/14) were the most prevalent STs identified in respect to the temporal distribution recovered during eight and six different years, respectively. ST13 (cps1, n = 6/6) was collected in four different years. ST17 (cps4, n = 3/3), ST29 (cps7, n = 3/4), ST231 (cps9, n = 1/14; cps3, n = 2/4), ST1103 (cps2, n = 3/4), and ST1236 (cps9, n = 3/14) originated from three different years. ST1101 (cps15, n = 2/6) and ST1109 (cps8, n = 1/3; cps9, n = 1/14) were found within two different years. However, the majority of isolates, representing 43.2% of the tested collective, was isolated uniquely in 1 year.

Figure 2

MLST-based minimal spanning tree of 88 porcineisolates. The MLST-based minimum spanning tree is representing the temporal distribution of sequence types determined for S. suis isolates collected during the last 13 years from diseased pigs. The tree was calculated using the goeBURST full MST algorithm in Phyloviz 2.0. Sizes of nodes reflect the number of isolates with a specific MLST profile. Numbers within the nodes indicate the corresponding sequence type. Node colors refer to the year of isolation as represented in the legend. Numbers on lines indicate locus variants between nodes. Black lines indicate single locus variants and grey lines represent multi locus variants. x and y represent isolates with no determined ST due to lacking of housekeeping gene identification.

Association between genotype and CCs

STs of all S. suis isolates were assigned to CCs using the eBurst illustration (Figure 3). Among the 88 isolates analyzed, eight CCs were identified: CC1, CC17, CC28, CC87, CC13/149, CC1109, CC1112, and CC1237 comprising 3, 3, 8, 2, 6, 7, 1, and 5 isolates, respectively. The distribution of S. suis isolates in relation to ST revealed ST1104 (cps6, n = 13/13) with the highest prevalence of 14.8% and no association to any CCs. Further on, 9.1% of isolates representing ST1105 (cps9, n = 8/14) were found to be unrelated singleton-isolates not associated to any CCs.

Figure 3

Population snapshot ofin Switzerland. Groups at triple locus variants (TLV) level were created by goeBURST v1.2 software using the phyloviz software [41] applying a data set from the PubMLST database (https://pubmlst.org/ssuis). Grey lines define a link at double locus variant (DLV) or triple locus variant (TLV) between the CCs following eBURST rules. Numbers in nodes represent sequence types (ST), whereas light green represents founder groups, blue shows common nodes and red indicates STs identified in Switzerland. Clonal complexes (CCs) identified are indicated in bold. Association of Swiss isolates between capsular types (cps), ST and its corresponding CCs is shown. Identified virulence markers namely extracellular protein factor (epf), muramidase-released protein (mrp), and suilysin (sly) are indicated, highlighting its correlation with according STs and CCs. Isolates harboring mrp + epf + sly + , mrp + (epf +)sly + , sly + , and mrp + are represented in pink, green, orange, and brown, respectively. Variants of mrp are indicated with mrpS (small variant) and mrp*** (large variant); variant of epf is marked as epf* (large variant). Isolates appearing as singletons or with no determined founder are not represented. CC1109 and CC1237 are shown separately with no connections to the main CCs due to the absence of relation at TLV level. Corresponding minimum core genome (MCG) groups are highlighted in yellow.

A total of 6.8% of isolates belonged to CC13/149 harboring ST13 (cps1, n = 6/6). 3.4% of isolates belonged to CC28 comprising ST29 (cps7, n = 3/4). ST17, ST28, ST231, ST1103, and ST1236 represented each 3.4% of isolates associated with CC17, CC28, CC1109, CC1, and CC1237, respectively. Finally, all remaining isolates had different STs only occurring in few isolates illustrating a massive genetic diversity identified in diseased Swiss pigs (Additional file 3). Samples comprising CC1 included ST1103 (cps2, n = 3/4). ST17 (cps4, n = 3/3) was part of CC17. The most abundantly identified CC28 contained isolates of ST28 (cps1/2, n = 2/3; cps2, n = 1/4), ST29 (cps7, n = 3/4), ST856 (cps3, n = 1/4), and ST1133 (cps1/2, n = 1/3). CC87 contained ST87 (cps8, n = 1/3) and ST1130 (cps7, n = 1/4). All isolates of ST13 (cps1, n = 6/6) were part of CC13/149. CC1109 comprised isolates of ST231 (cps3, n = 2/4; cps9, n = 1/14), ST1106 (cps8, n = 1/3), ST1109 (cps8, n = 1/3; cps9, n = 1/14), and ST1113 (cps3, n = 1/4). CC1112 harbored only one isolate of ST1112 (cps15, n = 1/6). Finally, CC1237 comprised ST1236 (cps9, n = 3/14), ST1237 (cps9, n = 1/14) and ST1240 (cps21, n = 1/2) (Table 3).

Table 3

clonal complexes (CCs) in relation to capsular type and the according sequence types (ST).

Clonal complex (CC)	Caspular type																		Sequence Types (STs)
	1/2	1	2	3	4	5	6	7	8	9	12	15	16	19	21	28	31	NTb
CC1			3																1103d
CC17					3														17, 1112d
CC28	3		1	1				3											28, 29, 856, 1133d
CC87								1	1										87, 1130d
CC13/149		6																	13
CC1109				3					2	2									231, 1106d, 1109d, 1113d
CC1112												1							1112d
CC1237										4					1				1236d, 1237d, 1240d
no CCa						2	13			8	2	5	6	1	1	2	3	8	1099-1102d, 1104d, 1105d, 1107d, 1108d, 1110d, 1111d, 1114-1116d, 1116d, 1117-1129d, 1134d, 1238d, 1239d, 1243d, 1244d
Ungroupablec											1							1	–
Total no. of isolates (n)	3	6	4	4	3	2	13	4	3	14	3	6	6	1	2	2	3	9

aNo CCs assigned for STs observed as singletons or with no determined founder.

bNontypeable by multiplex PCR.

cST could not be assigned due to failure of one housekeeping gene identification.

dNovel STs.

Virulence profiling

CC1 harbored one cps2 isolate with a large variant of epf* (mrp+epf*sly+), whereas two cps2 isolates had a mrp+epf−sly+ genotype. Variable subtypes of mrp were identified in one cps2 isolate and two cps1/2 isolates with a genotype of mrpSepf−sly−; one cps1/2 isolate, one cps3 isolate, and three cps7 isolates contained mrp***epf−sly−. Furthermore, genotype mrpSepf+sly+ was found in all three isolates of cps4 belonging to CC17. Two isolates of cps15 had a mrp***epf+sly+ genotype, whereas one isolate of each cps15 and cps28 contained mrp*epf−sly+. 22.7% of isolates showed a mrp−epf−sly+ genotype, which was associated with CC13/149 consisting of cps1 (n = 6/6); CC1109 comprising cps3 (n = 3/4), cps8 (n = 2/3), and cps9 (n = 2/14); CC87 consisting of cps7 (n = 1/4), and cps8 (n = 1/3); CC1237 comprising cps9 (n = 1/14); and ultimately, singleton-isolates consisting of cps15 (n = 3/6) and cps28 (n = 1/2). Finally, a total of 56.8% of isolates studied had a mrp−epf−sly− genotype, which could not be associated to any CCs and were identified as singletons consisting of cps5 (n = 2/2), cps6 (n = 13/13), cps9 (n = 11/14), cps12 (n = 3/3), cps16 (n = 6/6), cps19 (n = 1/1), cps21 (n = 2/2), cps31 (n = 3/3), and NT (n = 9/9) (Additional file 2).

MCG typing

MCG typing revealed MCG group 1 for members of CC1, CC17, and CC1112; MCG group 2 for isolates of CC87; MCG group 4 for isolates of CC28; MCG group N (representing an ungroupable subset of isolates) for CC13/149 and CC1109; and a novel MCG group for isolates of CC1237. In total, 13.6% of isolates were classified into five novel MCG groups not previously described (Table 4). Almost 30% of isolates could not be assigned to any MCG group due to the failure of one housekeeping gene identification.

Table 4

Single nucleotide polymorphisms (SNPs) used for minimum core sequencing (MCG).

MCG groups as defined by Zheng et al. [42] including 5 novel MCG groups identified in the present study. Red highlights discriminative SNPs for each SNP profile.

aReference genome: strain GZ1 (GenBank accession number: CP000837).

bEpidemic (E) strains (ST7 strains) identified by Zheng et al. [42].

cUngroupable.

The following isolates resulted in novel MCG groups with a new SNP profile.

dMB1158 (cps type 5), MB1497, SS2097 (cps type 9), PP106 (cps type 19), MB1185 (cps type 21), SS921 (nontypeable).

eMB1629 (cps type 12).

fHuman pathogen strain (cps type 14).

gMB 1297 (cps type 31).

hSS 3919 (cps type 31).

Pathotyping for differentiation between invasive disease-associated and non-disease-associated isolates

Figure 4 shows the amplicon patterns obtained by multiplex PCR after capillary electrophoresis. Amplicons of S. suis specific genetic marker (SSU0577) were obtained by all isolates proving the correct species identification. Amplicons of 892 bp represented the putative sugar ABC reporter as a non-disease-associated marker (SSUT30534), whereas disease-associated gene markers were identified by amplicons of 347 bp and 211 bp illustrating a type I RM system S protein (SSU1589) and a predicted copper ATPase (SSU0207), respectively. Additionally, new variant forms of the copper ATPase gene marker appeared, illustrated by a 190 bp long amplicon (Additional file 4). In total, 29.3% (n = 22/75) of isolates analyzed harbored a gene variant of the copper ATPase gene containing partial deletions, represented by all isolates of CC28 (n = 7/75), CC1109 (n = 7/75), and cps9 isolates with ST1105 (n = 8/75). As a contrast, 4% (n = 3/75) of isolates belonging to CC1 showed a 211 bp long amplicon as comprehended by the virulent P1/7 strain (Table 2). The disease-associated marker type I RM system S protein-gene revealed a PCR amplification of 347 bp only for the three isolates comprising CC1. Sequence analysis of some exemplary isolates with cps1/2, cps1, and cps2 revealed a truncated gene version of the implicated gene, whereas the reverse primer could not bind to the target sequence due to the absence of the 3′end of the gene in most isolates analyzed (Additional file 4).

Figure 4

Capillary electrophoresis plots illustrating invasive disease-associated and non-disease-associated Swissisolates. On the left hand side the DNA size marker (100 bp–2.5 kb) is shown. The alignment marker (green) is representing the start and end of electrophoresis. Amplicons sizes obtained from the molecular pathotyping tool [31] are indicated on the right hand side. Lane 1 corresponds to the amplification reaction of a highly virulent isolate of cps type 2 (sequence type 1103) revealing both disease-associated markers (red) predicting a putative copper exporting ATPase 1 (SSU0207) and a type I restriction-modification (RM) system S protein (SSU1589). Lane 2 and 3 represent disease-associated isolates harboring the copper ATPase and a variant form with a 21 base pair (bp) deletion (lane 3), respectively. Lane 4, 5, and 6 demonstrate observed amplification patterns of non-disease-associated samples (blue). Lane 4 shows a PCR amplicon for the putative sugar ABC reporter (SSUST30534) also in the presence of the copper ATPase gene. Samples with no amplification of any markers are considered to be non-disease-associated (lane 5). A sporulation regulator (WhiA) serves as identification control for S. suis (SSU0577).

Pathotyping of the subset of 75 isolates unequivocally attributed as S. suis relevant or not S. suis relevant, revealed isolates of cps2 and cps9 to be in accordance with the pathotyping tool. However, comparing S. suis relevant isolates of cps1 and cps6 to results obtained by pathotyping, all isolates yielded a contradictory result. All other cps types could not clearly be classified by pathotyping due to the paucity of isolates having the same capsular type and therefore no statement can be made.

Discussion

In this study a detailed assessment of genetic characteristics of Swiss S. suis isolates of diseased pigs was performed and compared to an earlier analyzed human isolate. In general, seventeen different cps types were identified underlining high serotype diversity in accordance to a previously observed massive diversity in whole genome sequences of a chosen set of isolates of clinically affected pigs [43]. The most common cps types found among the examined strain collection were cps9 and cps6. Contrary to a German investigation with over 700 S. suis isolates, not a single cps6 strain was found, which is worth to highlight when keeping in mind that Germany is a main neighboring country of Switzerland [44].

In the current report, all 13 isolates identified as cps6 comprised a single clone of ST1104, which was not associated to any CCs. ST1104 is representing a clone with no determined founder, identified to be a single locus variant (SLV; differing only at cpn60) of a Danish invasive S. suis strain with ST55 [45]. Noteworthy, this identified clone reoccurred throughout eight different years (Figure 2) underlining its capacity of causing disease in pigs due to an invasive behavior of the involved strain. Interestingly, a study from Chile, investigating phenotypically and genotypically diseased pigs at nonrelated farms during the time span of 2007–2011, revealed a single clone of S. suis with n = 28/29 of isolates identified to be serotype 6. Although an invasive phenotype could be observed, no enhanced virulence could be proven in a murine model testing serotype 6 isolates from Chile including the Danish serotype 6 reference strain. No concomitant infections of the involved isolates could be observed, therefore, an enhanced virulence capacity was suggested [46].

Furthermore, it was notable that all isolates belonging to cps9 were rather heterogenic harboring five different STs, although one abundant ST (ST1105, n = 8/14) seemed to be predominant among the identified samples. The genetic diversity of cps9 isolates in China and Canada showed a high variety among examined isolates [47, 48], which is comparable to the observed genetic diversity of inspected Swiss isolates. Interestingly, the aforementioned clone comprising ST1105 reoccurred within six different years (Figure 2) emphasizing its importance among the pig population in Switzerland and its ability to induce infections. Similar reports of pigs with invasive disease have stated a predominance of mrp*cps9 belonging to CC16/CC87 for Europe including The Netherlands, Spain, Germany and Belgium [9, 16]. The observation that cps9 isolates of the present study did not harbor mrp together with the identification of novel sequence types, underlines the emergence of very unique S. suis strains with new genetic rearrangements rarely found in other countries.

Actually, Switzerland has very little traffic of piglets with foreign countries and the pig farming system is different in comparison to other European countries. Herds are much smaller and the utilization of farrowing crates for reproduction purposes is prohibited by law since 1997. The physiological behavior in nesting is impaired due to the narrowing conditions in farrowing crates and consequently, a more natural environment helps to avoid complications during birth of piglets thereby strengthening their immune status. In addition, Switzerland is considered to be free of porcine reproductive and respiratory syndrome virus (PRRSV) [49] and Mycoplasma hyopneumoniae [50]. Moreover, a majority of farms vaccinate piglets against porcine circovirus (PCV) [51]. The general health status of Swiss pigs is considered to be high, which could be explained by the absence of these pathogens considered to be immunosuppressive and therefore having an implication in the defense of S. suis. Nevertheless, stress caused by weaning at an age of 28 days and rehousing of piglets usually after 10–12 weeks frequently paired with a change of feed at the fattening farms can lead to infection caused by virulent S. suis strains. Fattening farms usually combine weaning piglets of different farms leading to a higher transmission rate of S. suis among piglets. These circumstances could explain the high genetic variability and the distinct composition regarding ST and capsular types of S. suis strains obtained in this study in contrast to other countries.

Most remarkably, isolates belonging to CC1 are considered as potentially zoonotic and were mostly described as virulent strains [26]. In the present study three out of four isolates comprising cps2 were identified belonging to CC1. Showing its current relevance, one cps2 isolate originating from a brain of a diseased pig, was collected very recently in fall 2019, highlighting an emerging S. suis clone with ST1103. Interestingly, the implicated strain had a mrp+epf*sly+cps2 genotype containing a large variant of epf, which is considered to be a moderately virulent pathotype [21]. In fact, this specific clone is a SLV of ST1, differing only in allele gki. The fourth cps2 isolate had a more common ST28 belonging to CC28. One highly virulent human ST1 (cps14) isolate from another study [52] was also part of CC1 harboring mrp+epf+sly+ underlining an association between virulence-associated markers epf, mrp, and sly with STs rather than cps types (Figure 3). Three cps2 isolates (ST1103) were classified as MCG group 1, whereas the fourth cps2 isolate (ST28) was part of MCG4. For the human isolate, MCG typing revealed a novel group not described before, which was uniquely attributed to this isolate. The fact that this isolate was classified as a novel MCG group, absent among the analyzed Swiss porcine isolates, proves an improbable transmission event of this highly virulent pathogenic isolate from Swiss pigs pointing out a different source of infection. The fact that the virulent strain was introduced through a person originating from Moldova importing meat to Switzerland [52], proves an independent introduction of this zoonotic strain. Furthermore, all cps4 isolates which are part of CC17 were classified as MCG group 1 with a genotype of mrpSepf+sly+ emphasizing a possible increased virulence potential. In addition, a majority of cps15 isolates belonged to MCG group 1, whereas two isolates had a mrp***epf+sly+ genotype. Moreover, most isolates of cps7 were ST29 and belonged to CC28, similar to an emerging virulent strain recently reported in Germany [53] highlighting that isolates classified to be part of CC28 clearly show invasive and highly virulent features.

To evaluate the usefulness of pathotyping [31], a subset of 75 isolates having an unequivocal link of etiologically relevant S. suis diseased or non-disease-associated pigs was screened by the pathotyping tool. Cps2 and cps9 showed a clear correlation between pathotyping results and histopathological findings. Remarkably, comparing observed phenotypes classified to be S. suis relevant, obtained by a combination of histopathological examinations and bacteriological analysis, to results obtained by pathotyping, isolates of cps1 and cps6 yielded a contradictory result since no gene marker indicating disease-association yielded a positive PCR result. Due to the perceived inconsistency of phenotypically observed manifestations determined to be S. suis relevant in contrast to the non-S. suis relevant results obtained with the pathotyping tool, sequence analysis of the involved gene markers (type I RM systems S protein- and copper ATPase-gene) was performed highlighting a high genetic variability of different isolates (Additional file 4). Worth mentioning, all cps type 1 isolates (ST13) were associated to CC13/149. Interestingly, in two previous studies [29, 54] isolates of ST13 were associated with a pathogenic phenotype and represented an indicator of virulence underlining its importance. Yet undiscovered potential virulence factors especially of cps1 and cps6 strains remain matter of further investigation.

Cps3, cps7, cps15, and cps16 showed partially contradictory results between histopathological findings and the pathotyping tool, which points out the difficulty of consistent histopathological examinations, considering the multifaceted properties of bacterial infections.

Conclusion

With this study an expanded knowledge about the population structure and the genetic diversity of S. suis could be achieved. The epidemiological situation of S. suis in Switzerland revealed a heterogeneous composition of strains with high genetic variability belonging to several CCs. Strikingly, cps9 and cps6 were among the predominant capsular types associated with S. suis diseased pigs. Genetic variant forms of implicated disease-associated genes can be explained partially with frequent genomic rearrangements in S. suis as shown previously [27, 55]. These genetically variable and distinct rearrangements could be facilitated by the unique pig farming of Switzerland with very little transfer of piglets from other European countries, however, with a frequent transport rate of weaning piglets to fattening farms.

In view of an affordable tool as a diagnostic application to rapidly discover highly invasive S. suis strains and to highlight a potentially threatening zoonotic pathogen, it would be advantageous to have a reliable pathotyping tool. The screening using our well characterized strain collection showed a good feasibility of the tool for cps2 and cps9, however, for cps1 and cps6 contradictory results were obtained. Therefore, we recommend using the pathotyping tool only in case of cps2- and cps9-identification in order to differentiate between invasive disease-associated isolates and non-disease-associated isolates.

An efficient identification of highly virulent strains could help to alert public health surveillance programs to take action in line with a one health approach. Keeping track of disease-associated isolates helps to reduce the risk of zoonotic infections and will allow promoting a high standard in animal welfare programs.

Additional file 1. Capsular type distribution among Swissisolates. Number of isolates in relation to capsular type is indicated. NT = nontypeable due to no successful capsular typing by multiplex PCR.

Additional file 2. Clinical data, genotyping, and determination of virulence markers of Swiss porcineisolates ( = 88).

Additional file 3. Distribution of porcineisolates in relation to sequence type and clonal complex.

Additional file 4. Alignment of target gene sequences and the corresponding amino acid sequences used by pathotyping. Sequence alignment of copper ATPase 1-gene (A and B) and partial gene sequence alignment of type I RM system S protein-gene (C and D) of invasive disease-associated isolates of Swiss S. suis in comparison to the highly virulent reference strain P1/7 are shown visualizing different gene variants and its corresponding protein sequences. Conserved, matching nucleotide residues are illustrated as blue dots, whereas red represents differences of nucleotide sequences. (A) Copper ATPase 1-gene sequences of S. suis PP463 (cps2, ST28), SS470 (cps1/2, ST28), PP423 (cps1/2, ST1133), and PP536 (cps9, ST1105) are represented. Primer sequences of the pathotyping tool are indicated in green. A duplication of a 54 bp long DNA segment in isolate SS470 and deletion of a 21 bp fragment in all represented Swiss isolates could be observed, illustrating a high genetic variability. (B) Corresponding amino acid sequence alignment of Copper ATPase 1 is shown. (C) RM system S protein gene sequences of S. suis PP463 (cps2, ST28), PP423 (cps1/2, ST1133), and PP269 (cps1, ST13) are represented. The forward primer is indicated in green, whereas the reverse primer could not be shown since illustrated Swiss isolates are truncated. (D) Corresponding amino acid sequence alignment of RM systems S protein is shown.