The capsule

Serotyping of S. suis is based on differences in composition and structure of the polysaccharide capsule. For S. suis serotype 2, it has been demonstrated by two independent laboratories that the capsule is an important virulence factor (Charland et al., Reference Charland, Harel, Kobisch, Lacasse and Gottschalk1998; Smith et al., Reference Smith, Damman, van der Velde, Wagenaar, Wisselink, Stockhofe-Zurwieden and Smits1999). However, unencapsulated S. suis strains might also invade host tissue, though to a lower degree. This was shown by isolation of isogenic unencapsulated mutants from serosa, joints and central nervous system after intranasal application (Smith et al., Reference Smith, Damman, van der Velde, Wagenaar, Wisselink, Stockhofe-Zurwieden and Smits1999). Large differences in virulence have been described for strains belonging to the same serotype (Vecht et al., Reference Vecht, Wisselink, van Dijk and Smith1992), indicating that the capsule of S. suis is not sufficient for full virulence and that other factors have important functions in the pathogenesis of S. suis.

The capsule of serotypes 1 and 2 strains contains N-acetyl neuraminic acid (sialic acid) and four additional sugars, namely glucose, galactose, N-acetyl glucosamine and N-acetyl galactosamine (serotype 1) or rhamnose (serotype 2) (Elliott and Tai, Reference Elliott and Tai1978). Smith et al. (Reference Smith, Damman, van der Velde, Wagenaar, Wisselink, Stockhofe-Zurwieden and Smits1999) sequenced the cps2 locus and identified genes encoding putative glucosyl-, galactosyl-, N-acetylglucosaminyl- and rhamnosyl-transferases (genes cpsE-K). Furthermore, strains of serotypes 1, 2, 14, 27 and 1/2 carry the genes involved in sialic acid synthesis (Smith et al., Reference Smith, de Vries, van't Slot and Smits2000). In many bacteria, surface-associated sialic acid contributes to pathogenesis by interfering with the activation of the alternative complement pathway by increasing the affinity of C3b to factor H (relative to factor B). This prevents formation of the C3 convertase C3bBb thereby limiting C3b deposition on the surface of the pathogen (Marques et al., Reference Marques, Kasper, Pangburn and Wessels1992). In S. suis, sialic acid contributes to the adherence of serotype 2 strains to murine macrophages (Segura and Gottschalk, Reference Segura and Gottschalk2002). However, adherence was not accompanied by phagocytosis. This finding supports the ‘modified Trojan horse theory’, as proposed earlier by Gottschalk and Segura (Reference Gottschalk and Segura2000). The ‘modified Trojan horse theory’ suggests that S. suis strains, at least suilysin negative strains (see below), may cross the blood–brain–barrier (BBB) through adherence to immune cells, in particular monocytes, which may function as vehicles for passage through the BBB. However, there is also evidence for invasion of choroid plexus epithelial cells and directed transport of S. suis through these cells from the basolateral to the apical site, which faces the cerebrospinal fluid compartment (Tenenbaum et al., Reference Tenenbaum, Papandreou, Gellrich, Friedrichs, Seibt, Adam, Wewer, Galla, Schwerk and Schroten2009).

In contrast to the wild-type, isogenic unencapsulated mutants of serotype 2 strains were efficiently eliminated by primary alveolar macrophages (Smith et al., Reference Smith, Damman, van der Velde, Wagenaar, Wisselink, Stockhofe-Zurwieden and Smits1999) as well as by the murine macrophage cell line J774 (Segura et al., Reference Segura, Cleroux and Gottschalk1998). Segura et al. (Reference Segura, Gottschalk and Olivier2004) found that, unlike the encapsulated wild-type strain, the unencapsulated mutant-induced phosphorylation of the phosphatidylinositol-3 kinase downstream kinases Akt and PKCα in J774 macrophages, indicating the importance of these signaling pathways in phagocytosis of unencapsulated S. suis strains. Furthermore, in macrophages deficient in a tyrosine phosphatase (SHP-1), phagocytosis of an encapsulated S. suis wild-type strain was significantly increased. This was associated with Akt phosphorylation, suggesting that capsule-mediated resistance to phagocytosis involves activation of SHP-1 (Segura et al., Reference Segura, Gottschalk and Olivier2004).

The capsule of S. suis might also have other important immunomodulatory functions. As cell wall components of S. suis induce a significant release of proinflammatory cytokines (tumor necrosis factor-α (TNF-α), IL-1β and IL6) and chemokines (IL-8 and monocyte chemoattractant protein 1 (MCP-1)) by host cells, the capsule might reduce the production of such cytokines by blocking the interaction of cell wall components with pattern-recognition receptors, most importantly Toll-like receptor (TLR)2. On the other hand, the capsule of S. suis serotype 2 itself induces macrophage MCP-1 release through a TLR2- and MyD88-independent pathway (Graveline et al., Reference Graveline, Segura, Radzioch and Gottschalk2007). The authors discussed induction of MCP-1 release as part of the mechanism used by S. suis to translocate the BBB, as brain microvascular endothelial cells (BMECs) were shown to alter expression of tight junction-associated proteins in response to MCP-1 in vitro (Song and Pachter, Reference Song and Pachter2004).

Adhesion of encapsulated serotype 2 strains of S. suis to epithelial cells is generally low, with values below 5% (Lalonde et al., Reference Lalonde, Segura, Lacouture and Gottschalk2000) or about 1% (Benga et al., Reference Benga, Goethe, Rohde and Valentin-Weigand2004). In contrast, unencapsulated strains, including isogenic cps2 mutants, show adherence of 15–60% (Benga et al., Reference Benga, Goethe, Rohde and Valentin-Weigand2004). Interestingly, capsule expression was reported to interfere with adhesion to BMECs of porcine (Vanier et al., Reference Vanier, Segura, Friedl, Lacouture and Gottschalk2004) but not of human origin (Charland et al., Reference Charland, Nizet, Rubens, Kim, Lacouture and Gottschalk2000). It was suggested that porcine, but not human BMECs, might express an important receptor molecule that is recognized by unencapsulated but not encapsulated S. suis strains (Vanier et al., Reference Vanier, Segura, Friedl, Lacouture and Gottschalk2004). However, such a receptor has not been identified and, furthermore, in such a case one would expect lower adhesion values of the unencapsulated strains to human than to porcine BMECs.

Norton et al. (1999) described invasion of the epithelial cell line HEp2 by encapsulated S. suis serotype 2 strains, in contrast to observations published for various epithelial cell lines by other laboratories (Lalonde et al., Reference Lalonde, Segura, Lacouture and Gottschalk2000; Benga et al., Reference Benga, Goethe, Rohde and Valentin-Weigand2004). However, unencapsulated S. suis strains, including an isogenic cps2 mutant, have been demonstrated to be invasive, most likely through receptor-mediated endocytosis. A plausible explanation for the different phenotypes of encapsulated and unencapsulated S. suis strains might be that S. suis down-regulates capsule expression for increased adhesion to epithelial cells and upregulates it for protection against phagocytosis as it enters the bloodstream (Gottschalk and Segura, Reference Gottschalk and Segura2000). Accordingly, an increased thickness of the capsule was demonstrated during cultivation of S. suis in liquid media supplemented with serum (Wibawan and Lammler, Reference Wibawan and Lammler1994). Furthermore, increased thickness of the capsule was also noted during growth in the intra-peritoneal cavities in rats and pigs (Quessy et al., Reference Quessy, Dubreuil, Jacques, Malouin and Higgins1994; Charland et al., Reference Charland, Kobisch, Martineau-Doize, Jacques and Gottschalk1996). Interestingly, Smith et al. (Reference Smith, Buijs, de Vries, Wisselink, Stockhofe-Zurwieden and Smits2001) identified an in vivo-induced gene that encodes a protein highly homologous to CpsY, a regulator of capsule expression in Streptococcus agalactiae (73% identity). Further evidence for capsule regulation in S. suis is provided by a recent study on the orphan response regulator CovR (Pan et al., Reference Pan, Ge, Li, Wu, Wang, Wang, Feng, Yin, Zheng, Cheng, Sun, Ji, Hu, Shi, Feng, Hao, Dong, Hu and Tang2009). The ΔcovR mutant showed a thicker capsule and longer chains. This was in agreement with the finding that cps2C and the gene encoding CMP-N-acetylneuraminic acid synthetase were among the upregulated genes in the ΔcovR mutant. The gene cps2C has been proposed to be involved in chain length determination and export of polysaccharides for capsule production (Smith et al., Reference Smith, Damman, van der Velde, Wagenaar, Wisselink, Stockhofe-Zurwieden and Smits1999).

The capsule of S. suis is poorly immunogenic. Sera from convalescent or bacterin immunized piglets generally have very low specific antibody titers directed against the capsular polysaccharides (Campo Sepulveda et al., Reference Campo Sepulveda, Altman, Kobisch, D'Allaire and Gottschalk1996; Wisselink et al., Reference Wisselink, Vecht, Stockhofe-Zurwieden and Smith2001; Baums et al., Reference Baums, Kock, Beineke, Bennecke, Goethe, Schroder, Waldmann and Valentin-Weigand2009). Immunization of piglets with polysaccharides elicited opsonizing antibodies only in the presence of Feunds' incomplete adjuvant (Elliott et al., Reference Elliott, Clifton-Hadley and Tai1980). We induced very high capsule-specific serum IgG titers through immunization with conjugated serotype 2 capsular polysaccharides. The respective sera caused opsonization of the homologous serotype 2, but not the heterologous serotype 9 strain, and subsequent killing by porcine neutrophils (Baums et al., Reference Baums, Kock, Beineke, Bennecke, Goethe, Schroder, Waldmann and Valentin-Weigand2009). Therefore, we think that a combination of conjugated polysaccharides from different serotypes might elicit protection in swine against S. suis, similar to immunoprophylaxis of pneumococcal infections in humans.

Peptidoglycan

In the cell wall of Gram-positive bacteria glycan strands consisting of alternating β-1,4-linked N-acetylglucosamine and N-acetylmuramic acid are cross-linked by short peptide chains (Vollmer et al., Reference Vollmer, Blanot and de Pedro2008). Fittipaldi et al. (Reference Fittipaldi, Sekizaki, Takamatsu, Dominguez-Punaro, Harel, Bui, Vollmer and Gottschalk2008a) sinvestigated muropeptides of S. suis using two different experimental approaches. The muropeptides had mainly D-isoglutamine at position 2 and L-lysine at position 3 of the peptide stem. In contrast to most other Gram-positive bacteria, S. suis peptidoglycan lacks interpeptide bridges but contains direct 3–4 cross-links. By high-sensitive Fourier transform mass spectrometry, the authors were able to detect N-deacetylated muropeptides in very low quantities in S. suis (probably below 1% of the total muropeptides). N-deacetylated peptidoglycan was not detectable in an S. suis ΔpgdA mutant confirming the proposed function of the protein encoded by the pgdA gene, which was originally identified as a gene upregulated upon interaction with porcine BMECs (Fittipaldi et al., Reference Fittipaldi, Gottschalk, Vanier, Daigle and Harel2007). The peptidoglycan polysaccharide deacetylase PgdA from S. suis shares homology to PgdA from Streptococcus pneumoniae, in particular with respect to catalytic residues, and has also a predicted N-terminal membrane anchor. In general, modification of peptidoglycan through deacetylation results in lysozyme resistance (Vollmer, Reference Vollmer2008). S. suis is very sensitive to lysozyme in vitro, most likely related to the low content of N-deacetylated muropeptides. Furthermore, inactivation of the pgdA gene does not result in increased lysozyme susceptibility. However, these phenotypes might be different in vivo, as pgdA is substantially upregulated upon interaction with host cells and within the host (Fittipaldi et al., Reference Fittipaldi, Sekizaki, Takamatsu, Dominguez-Punaro, Harel, Bui, Vollmer and Gottschalk2008a).

The S. suis pgdA mutant was attenuated in virulence in murine and porcine models of infection. The impairment in virulence was associated with an accelerated decrease in numbers of bacteria in the blood. In accordance, the pgdA mutant was killed efficiently by neutrophils in the presence of naïve serum in vitro, in contrast to the wild-type strain. Furthermore, cytokine production, in particular IFN-γ and IL-6, was severely diminished in CD1 mice infected with the pgdA mutant in comparison to wild-type infected animals (Fittipaldi et al., Reference Fittipaldi, Sekizaki, Takamatsu, Dominguez-Punaro, Harel, Bui, Vollmer and Gottschalk2008a).

Li et al. (Reference Li, Liu, Chen and Zhou2009) performed selective capture of transcribed sequences (SCOTS) to identify genes preferentially expressed under iron starvation. At least six of the 63 identified upregulated genes encode proteins homologous to enzymes involved in cell wall biosynthesis. The upstream sequences of the genes murG (SSU05_0477) and murA (SSU05_1170) contain putative Fur boxes, which suggests that these genes are regulated by the iron-dependent transcriptional regulator Fur (Li et al., Reference Li, Liu, Chen and Zhou2009), and, thus, might be regulated during infection.

Cell wall components of S. suis exacerbate the host inflammatory response to infection as they induce a significant release of proinflammatory cytokines (TNF-α, IL-1β and IL6) and chemokines (IL-8 and MCP-1) (Segura et al., Reference Segura, Vanier, Al Numani, Lacouture, Olivier and Gottschalk2006). Production of these molecules, in particular IL-6, IL-8 and MCP-1, is induced not only in inflammatory cells but also in BMEC (Vadeboncoeur et al., Reference Vadeboncoeur, Segura, Al Numani, Vanier and Gottschalk2003). This induction appears to be cell type specific as it does not occur in human umbilical vein endothelial cells. Production of chemokines by BMECs might be crucial for the inflammatory reaction induced by S. suis invasion of brain tissue and an important source of cytokines and chemokines in the cerebrospinal fluid.

Teichoic and lipoteichoic acids

Teichoic acid (TA) is a secondary polymer of α-glycerol phosphate. Lipoteichoic acid (LTA) contains a glycolipid moiety, resulting in an amphiphilic molecule. Elliot et al. (Reference Elliott, McCarty and Lancefield1977) proposed that LTA from S. suis has a backbone structure similar to TA from Streptococcus pyogenes, but with differences in glucosyl substituents. LTA and TA of S. suis were characterized earlier as antigens recognized by group D antiserum (Elliott et al., Reference Elliott, McCarty and Lancefield1977). Equivocal reactions described in Lancefield typing of S. suis, in particular classification as groups R, S and T, were mainly due to extraction of antigens from capsular material rather than from the cell wall as described by Gottschalk et al. (Reference Gottschalk, Segura and Xu2007) in more detail.

TA and LTA might contribute to the adherence of pathogenic streptococci and staphylococci to host cells. In S. suis, the adhesion to porcine BMEC was reduced by pretreatment of the cells with LTA suggesting that LTA is partially involved in the adhesive phenotype of S. suis to these cells, which form part of the BBB (Vanier et al., Reference Vanier, Segura and Gottschalk2007). Upon interaction with porcine BMECs S. suis upregulates expression of DltA, an enzyme that catalyzes the D-alanylation of LTA (Fittipaldi et al., Reference Fittipaldi, Gottschalk, Vanier, Daigle and Harel2007, Reference Fittipaldi, Sekizaki, Takamatsu, Harel, Dominguez-Punaro, Von Aulock, Draing, Marois, Kobisch and Gottschalk2008b). This modification of LTA is crucial for resistance against cationic antimicrobial peptides, most likely through reduction of the global negative charge of the bacterial envelope. Accordingly, a ΔdltA mutant strain, but not the wild-type, was efficiently killed by porcine neutrophils in the presence of naïve serum and was attenuated in virulence in the CD1 mouse model (Fittipaldi et al., Reference Fittipaldi, Sekizaki, Takamatsu, Harel, Dominguez-Punaro, Von Aulock, Draing, Marois, Kobisch and Gottschalk2008b).

Pili

Pilus formation has been detected in a number of Gram-positive bacteria including the streptococcal species S. pyogenes, S. agalactiae and S. pneumoniae (Telford et al., Reference Telford, Barocchi, Margarit, Rappuoli and Grandi2006). In these streptococci, the genes encoding pili are located on pathogenicity islands. In S. pneumoniae, pili have been demonstrated to be important for adhesion to lung epithelial cells and for virulence in an intranasal challenge experiment with mice (Barocchi et al., Reference Barocchi, Ries, Zogaj, Hemsley, Albiger, Kanth, Dahlberg, Fernebro, Moschioni, Masignani, Hultenby, Taddei, Beiter, Wartha, von Euler, Covacci, Holden, Normark, Rappuoli and Henriques-Normark2006). Ultrastructural investigations of S. suis revealed surface appendages with a length of approximately 250 nm and a width of 2 nm resembling pili of other Gram-positive pathogens (Jacques et al., Reference Jacques, Gottschalk, Foiry and Higgins1990). Accordingly, sequence analysis of the two serotype 2 reference strains P1/7 and 89-1591 indicated presence of a truncated homologue of the S. agalactiae pilus island 2b (PI-2b), which included the two genes encoding the ancillary and major pilin subunits (named srtF cluster by Takamatsu et al., Reference Takamatsu, Nishino, Ishiji, Ishii, Osaki, Fittipaldi, Gottschalk, Tharavichitkul, Takai and Sekizaki2009). Importantly, the sequence of the putative major pilin subunit includes an E-box motif, which has been identified to be crucial for pilus assembly in other bacteria (Ton-That et al., Reference Ton-That, Marraffini and Schneewind2004; Telford et al., Reference Telford, Barocchi, Margarit, Rappuoli and Grandi2006). The first gene (ssu0424 or sipF) of the series of genes that encode the truncated version of PI-2b in S. suis encodes a putative signal peptidase that is preferentially expressed by S. suis upon interaction with porcine BMECs (Fittipaldi et al., Reference Fittipaldi, Gottschalk, Vanier, Daigle and Harel2007). Several mutations are present in the pilus gene cluster of S. suis strains P1/7 and NCTC 10237. Therefore, these strains might not express functional pilus structures. In the future, it is important to clarify whether the gene products of the truncated version of PI-2b in S. suis lead to pilus formation and whether this pilus is important for interaction with host cells and pathogenesis of S. suis.

Takamatsu et al. (Reference Takamatsu, Nishino, Ishiji, Ishii, Osaki, Fittipaldi, Gottschalk, Tharavichitkul, Takai and Sekizaki2009) described a putative pilus gene cluster srtBCD with structural homology to the rlrA island of pneumococci. The cluster lacks a gene encoding a protein homologous to RlrA, but includes genes encoding putative pilin subunit proteins with E-box and LPXTG motifs and three class C sortases.

Sortases

Sortases are transpeptidases that anchor secreted proteins covalently to the cell wall with a C-terminal cell wall sorting signal (CWS). The CWS includes an LPXTG- or related motif followed by a hydrophobic domain and a positively charged tail consisting of at least one arginine or lysine residue. For protein A of Staphylococcus aureus, it has been demonstrated that sortase A specifically cleaves the precursor protein between the threonine and the glycine residues and catalyzes the transfer of the processed protein to the free amino group of a pentaglycine cross-bridge of peptidoglycan (Mazmanian et al., Reference Mazmanian, Liu, Ton-That and Schneewind1999; Ton-That et al., Reference Ton-That, Liu, Mazmanian, Faull and Schneewind1999). S. suis serotype 2 carries five to six genes (srtA–F) encoding putative sortases, which is exceptional for streptococci (Osaki et al., Reference Osaki, Takamatsu, Shimoji and Sekizaki2002; Wang et al., Reference Wang, Li, Feng, Zheng, Dong, Pan, Cheng, Dong, Hu, Feng, Ge, Liu, Wang, Cao, Hu and Tang2009). The gene srtA was detected in S. suis strains of 28 different serotypes indicating that the gene is present in most if not all S. suis strains. It shows substantial allelic variation, which has been proposed to be a result of evolutionary divergence rather than recombination (Osaki et al., Reference Osaki, Takamatsu, Shimoji and Sekizaki2003). The S. suis serotype 2 srtA gene displays high homology (65% identity) to the srtA of Streptococcus gordonii. The homology of the other srt genes (srtB–E) to known sortase genes is much lower. Furthermore, the putative S. suis sortases B–E are class C sortases that have a C-terminal hydrophobic domain serving most likely as a membrane anchor (Dramsi et al., Reference Dramsi, Trieu-Cuot and Bierne2005). Osaki et al. (Reference Osaki, Takamatsu, Shimoji and Sekizaki2002) found that more than 15 protein spots in 2D-PAGE of cell wall proteins depended on the presence of the gene encoding sortase A but were independent of other srt genes. Four putative sortase A substrates with a CWS were identified; among these was the muramidase-released protein (MRP, see also Table 1). Furthermore, the CWS-protein surface antigen one (SAO) was also not detectable on the surface of an srtA mutant (Wang et al., Reference Wang, Li, Feng, Zheng, Dong, Pan, Cheng, Dong, Hu, Feng, Ge, Liu, Wang, Cao, Hu and Tang2009).

Table 1. Confirmed (c) and proposed (p) S. suis proteins with a CWS

^a The prefixes SSU98 and SSU05 in the ordered locus names have been annotated in the genomes of the two sequenced STSS strains 98HAH12 and O5ZYH33, respectively. Factors are regarded as proposed, if no confirmation of protein expression has been reported.

^b The theoretical molecular weight is based on the genome sequence of serotype 2 reference strain P1/7 (www.sanger.ac.uk), if it is not provided by the authors.

^c Further genes and pseudogenes encoding putative pilin subunit proteins with CWS have been described (see Table S2 in Takamatsu et al., Reference Takamatsu, Nishino, Ishiji, Ishii, Osaki, Fittipaldi, Gottschalk, Tharavichitkul, Takai and Sekizaki2009, and Table 3 in Wang et al., Reference Wang, Li, Feng, Zheng, Dong, Pan, Cheng, Dong, Hu, Feng, Ge, Liu, Wang, Cao, Hu and Tang2009).

Table 2. Confirmed (c) and proposed (p) surface-associated proteins of S. suis without a CWSFootnote ^a

^a Integral membrane proteins are not included with the exception of a few selected proteins with a C-terminal membrane anchor.

^b The prefixes SSU98 and SSU05 in the ordered locus names have been annotated in the genomes of the two sequenced STSS strains 98HAH12 and O5ZYH33, respectively. Factors are regarded as proposed, if no confirmation of protein expression is available.

^c These factors were identified as surface-associated factors in S. suis serotype 9 based on an immunoproteomic analysis of cell wall proteins (Wu et al., Reference Wu, Zhang and Lu2008a).

Table 3. Proposed (p) and confirmed (c) secreted factors of S. suis (selection)

^a This ORF was annotated as an LPXTG-motif cell wall anchor domain protein in SsuiDRAFT_3692 of strain 89/1591. However, the putative Gram-positive anchor identified by Pfam 5 is localized between amino acids 67 and 106 at the N-terminus.

^b Cell-associated activity of DPP IV and Arg-aminopeptidase was also reported in the respective publications.

^c An ORF encoding a putative IgG-binding protein is annotated in the genome of S. suis 98015 (accession no. DQ410872), but experimental evidence regarding IgG binding of this factor has not been published.

^d Bacteriocin-like inhibitory substances.

Isogenic srtA serotype 2 mutants are attenuated in adherence to and invasion of host cells including porcine BMECs, indicating that CWS proteins are involved in this interaction (Vanier et al., Reference Vanier, Sekizaki, Dominguez-Punaro, Esgleas, Osaki, Takamatsu, Segura and Gottschalk2008; Wang et al., Reference Wang, Li, Feng, Zheng, Dong, Pan, Cheng, Dong, Hu, Feng, Ge, Liu, Wang, Cao, Hu and Tang2009). Accordingly, binding to plasma and cellular fibronectin as well as to collagen type I diminished in the srtA mutant. However, the binding of these host proteins is not completely abrogated in the srtA mutant, suggesting that the other sortases might also be involved in anchoring of adhesins or that non-CWS adhesins are also important for this phenotype (Vanier et al., Reference Vanier, Sekizaki, Dominguez-Punaro, Esgleas, Osaki, Takamatsu, Segura and Gottschalk2008).

Surprisingly, the isogenic srtA serotype 2 mutant was as virulent as the wild-type in the CD1 mice model (Vanier et al., Reference Vanier, Sekizaki, Dominguez-Punaro, Esgleas, Osaki, Takamatsu, Segura and Gottschalk2008). In contrast, intravenous infection of piglets with an srtA mutant of the Chinese STSS reference strain O5ZYH33 suggested attenuation in virulence. Competitive challenge experiments with the wild-type strain suggested that the srtA mutant is primarily attenuated in colonization of the brain and the lung (Wang et al., Reference Wang, Li, Feng, Zheng, Dong, Pan, Cheng, Dong, Hu, Feng, Ge, Liu, Wang, Cao, Hu and Tang2009). Because of the contradictory findings, it is difficult at present to assess the relevance of sortase A function in the pathogenesis of S. suis.

Proteins with a CWS (LPXTG-motif)

In silico analysis of the genome of the Chinese serotype 2 reference strain O5ZYH33 revealed 33 putative cell wall-anchored proteins containing an LPXTG- or related motif (Wang et al., Reference Wang, Li, Feng, Zheng, Dong, Pan, Cheng, Dong, Hu, Feng, Ge, Liu, Wang, Cao, Hu and Tang2009). As described above, LPXTG-motifs, in particular CWS, indicate covalent linkage of these proteins to the cell wall. In other streptococci, numerous LPXTG-proteins have been demonstrated to be involved in binding of extracellular matrix proteins and adhesion to host cells. Though it is accepted that S. suis interacts with extracellular host proteins and cells (Esgleas et al., Reference Esgleas, Lacouture and Gottschalk2005), conclusive evidence for such a function is lacking for any of the putative adhesins described in S. suis. Following, surface-associated factors of S. suis are described that contain a CWS (see also Table 1).

Proposed and confirmed surface-associated proteins lacking a known CWS

Hemagglutinins

S. suis expresses a lectin-like adhesin that binds to the disaccharide Gal-(α1–4)-Gal of the glycolipid trihexosylceramide (GbO₃) (Haataja et al., Reference Haataja, Tikkanen, Hytonen and Finne1996). This binding activity causes hemagglutination and is associated with an 18 kDa protein, which is classified into subtypes P_N and P_O (Tikkanen et al., Reference Tikkanen, Haataja, Francois-Gerard and Finne1995). N-acetylgalactosamine inhibits binding of P_N only, whereas galactose inhibits binding of P_N and P_O. The Gal-(α1–4)-Gal binding adhesin is present in strains of various serotypes (Tikkanen et al., Reference Tikkanen, Haataja and Finne1996). Encapsulation is thought to interfere with the accessibility of this adhesin. However, preliminary immunization experiments in mice with the purified Gal-(α1–4)-Gal adhesin indicated induction of opsonizing antibodies (Tikkanen et al., Reference Tikkanen, Haataja and Finne1996).

A different hemagglutinin expressed by only a limited number of strains binds to a sialylated poly-N-acetyllactosamine-containing component on the surface of erythrocytes (Liukkonen et al., Reference Liukkonen, Haataja, Tikkanen, Kelm and Finne1992). To the best of our knowledge, the genes encoding these hemagglutinins have not been identified.

Lipoproteins

Prolipoproteins contain a conserved sequence element in the signal sequence called ‘lipobox’, which leads to a specific processing pathway. A prolipoprotein diacylglyceryl transferase (Lgt) transfers a diacylglyceryl moiety from a membrane lipid onto the conserved lipobox cysteine forming a thioether linkage. Subsequently, a lipoprotein signal peptidase (Lsp) cleaves the prolipoprotein in the ‘lipobox’, leaving the lipid-modified cysteine at the N-terminus (Hutchings et al., Reference Hutchings, Palmer, Harrington and Sutcliffe2009). A putative lgt gene has been annotated in the genomes of the Chinese S. suis STSS strains O5ZYH33 (SSUO5_1605) and 98HAH33 (SSU98_1615), but so far it has not been characterized in further detail (Chen et al., Reference Chen, Tang, Dong, Wang, Feng, Wang, Zheng, Pan, Liu, Li, Song, Zhu, Sun, Feng, Guo, Ju, Ge, Dong, Sun, Jiang, Wang, Yan, Yang, Wang, Gao, Yang, Wang and Yu2007). A highly homologous open reading frame (ORF) is also present in the genome of the European serotype 2 reference strain P1/7 (our unpublished results). An Lsp was identified by de Greeff et al. (Reference de Greeff, Hamilton, Sutcliffe, Buys, van Alphen and Smith2003). Sequence analysis of Lsp suggested four transmembrane spanning regions. An lsp knock out mutant of a virulent serotype 2 strain showed at least four band shifts of metabolically labeled lipoproteins indicating accumulation of some prolipoproteins in the lsp mutant. Accordingly, a band shift in this mutant in comparison to the wild-type strain was also observed in Western blot analysis with an antibody directed against the pneumococcal lipoprotein PsaA, a component of a Mn²⁺ transporter. In a competitive experimental infection experiment the lsp knock-out mutant was not attenuated in virulence as compared to the parental strain. These results are, however, difficult to interpret as some of the lipoproteins might be processed by an alternative, yet unidentified pathway as suggested by band shifts of the metabolically labeled lipoproteins in the respective mutant (de Greeff et al., Reference de Greeff, Hamilton, Sutcliffe, Buys, van Alphen and Smith2003).

In another recent study, three putative divalent-cation-binding lipoproteins of ABC transporters were identified and shown to be upregulated upon divalent-cation deprivation (Aranda et al., Reference Aranda, Garrido, Cortes, Llagostera and Barbe2008). Accordingly, the predicted regulator of Zn²⁺ and Mn²⁺ uptake AdcR was demonstrated to bind to two of the three promoters of the putative lipoprotein genes in vitro. Furthermore, in a mutant of a different zinc uptake regulator (Zur), putative Zn²⁺-binding lipoproteins were also found to be differentially regulated (Feng et al., Reference Feng, Li, Zhang, Zheng, Han, Wang, Yan, Tang and Gao2008). Interestingly, one of the three putative lipoproteins (SsuiDRAFT 0103) described by Aranda et al. (Reference Aranda, Garrido, Cortes, Llagostera and Barbe2008) induced significant protection in mice (Table 4). A second putative protective lipoprotein was identified in an immunoproteomic analysis of cell-wall associated proteins. The protein SSU98_1094 is predicted to be a lipoprotein of an uncharacterized ABC-type transport system (Geng et al., Reference Geng, Zhu, Yuan, Zhang, Li, Wang, Zheng, Wei, Cao, Wang and Jiang2008). Immunization of rats with rSSU98_1094 elicited high titers of opsonizing antibodies based on a whole human blood killing assay. Furthermore, convalescent sera from humans (meningitis or STSS patients) and piglets as well as from bacterin immunized piglets contain high serum antibody titers against SSU98_1094. Based on the results of this comprehensive immunoproteomic analysis of S. suis serotype 2, SSU98_1094 was suggested, together with two other surface-associated proteins (SSU98_0197 and SSU1664), as a vaccine candidate and diagnostic target (Geng et al., Reference Geng, Zhu, Yuan, Zhang, Li, Wang, Zheng, Wei, Cao, Wang and Jiang2008).

Table 4. Evaluation of protective efficacies of S. suis proteins

^a Immunization experiments with three piglets suggested protection.

^b CFA, complete Freund's adjuvant; IFA, incomplete Freund's adjuvant.

^c Noteworthy, 50% mortality was observed in immunized piglets.

^d Annotations in the genome of the North American reference strain 89/1591.

^e The protective efficacies of eight other recombinant S. suis proteins were investigated in this study using the same protocol for immunization and challenge with the STSS strain 05ZYH33 (Liu et al., Reference Liu, Cheng, Wang, Pan, Cong, Pan, Wang, Zheng, Hu and Tang2009).

^f Murein-associated protein.

Suilysin

Suilysin is a pore-forming cholesterol-dependent cytotoxin (CDC) expressed by many but not all virulent S. suis strains. It shares 52% amino acid identity with pneumolysin, its closest known relative within the CDC protein family (Segers et al., Reference Segers, Kenter, de Haan and Jacobs1998). In contrast to pneumolysin it contains a signal sequence and is secreted. The mature protein has an estimated molecular weight of 54 kDa. The hemolytic activity of suilysin is lost upon oxidation and is inhibited by cholesterol (Jacobs et al., Reference Jacobs, Loeffen, vandenBerg and Storm1994). The susceptibility of erythrocytes to suilysin-mediated lysis shows some species specificity, with human group O erythrocytes being the most susceptible, followed by horse, sheep, cow and pig red blood cells (Gottschalk et al., Reference Gottschalk, Lacouture and Dubreuil1995).

Sequencing of S. suis genomes revealed additional genes encoding putative hemolysins (Chen et al., Reference Chen, Tang, Dong, Wang, Feng, Wang, Zheng, Pan, Liu, Li, Song, Zhu, Sun, Feng, Guo, Ju, Ge, Dong, Sun, Jiang, Wang, Yan, Yang, Wang, Gao, Yang, Wang and Yu2007). However, our Blast searches with the published S. suis genome sequences did not reveal a further putative CDC in S. suis (unpublished). In this respect, Table 1 of the publication by Chen et al. (Reference Chen, Tang, Dong, Wang, Feng, Wang, Zheng, Pan, Liu, Li, Song, Zhu, Sun, Feng, Guo, Ju, Ge, Dong, Sun, Jiang, Wang, Yan, Yang, Wang, Gao, Yang, Wang and Yu2007) listing ply and sly for S. suis is confusing. Apparently, both names refer to the suilysin gene (C. Chen, personal communication), and S. suis does not have a pneumolysin gene.

The sly gene displays very limited genetic diversity. Eighty percent of 208 investigated sly+S. suis isolates contained an identical sly allele as defined by restriction fragment length polymorphism (RFLP) analysis (King et al., Reference King, Heath, Luque, Tarradas, Dowson and Whatmore2001). Based on sequencing and Southern blot analysis it was suggested that the sly gene was acquired from a foreign source and laterally transferred from strain to strain by a mechanism including natural transformation and homologous recombination (Takamatsu et al., Reference Takamatsu, Osaki and Sekizaki2002).

In contrast to other virulence-associated genes such as mrp and fbps, expression of suilysin mRNA was found to be induced rather late during infection, i.e. at 48 h after intravenous application of the serotype 2 challenge strain (Tan et al., Reference Tan, Liu, Jin, Liu, Chen, Wu, Fu, Bei and Chen2008b). Noteworthy, manifestation of disease was observed at 20 h post infection. Depending on the tissue, a 2–4-fold induction of suilysin mRNA was described in comparison to in vitro expression. However, mRNA from in vitro cultures was collected at 3 h of cultivation. Therefore, it cannot be excluded that the postulated in vivo induction is simply due to the substantial upregulation of suilysin expression at the end of the exponential growth phase, as demonstrated by experiments measuring hemolytic activities (Gottschalk et al., Reference Gottschalk, Lacouture and Dubreuil1995) and sly promoter activity analysis (Lun and Willson, Reference Lun and Willson2005). Interestingly, transposon interruption of the gene manN endocing mannose permease IID of the mannose-specific phosphotransferase system resulted in a hyper-hemolytic phenotype due to increased suilysin expression. As this phenotype was, at least partially, reverted by complementation with manN, it was concluded that regulation of sly can be mediated by the mannose-specific phosphotransferase system (Lun and Willson, Reference Lun and Willson2005).

Suilysin has cytotoxic activity for various host cells, as shown for different epithelial cell lines (Norton et al., Reference Norton, Rolph, Ward, Bentley and Leigh1999; Lalonde et al., Reference Lalonde, Segura, Lacouture and Gottschalk2000), BMECs (Charland et al., Reference Charland, Nizet, Rubens, Kim, Lacouture and Gottschalk2000) and macrophages (Segura and Gottschalk, Reference Segura and Gottschalk2002). As determined by lactate dehydrogenase activity, more than 80% cytotoxicity was seen using 1–10 μg/ml suilysin (Charland et al., Reference Charland, Nizet, Rubens, Kim, Lacouture and Gottschalk2000; Lalonde et al., Reference Lalonde, Segura, Lacouture and Gottschalk2000; Segura and Gottschalk, Reference Segura and Gottschalk2002). However, suilysin also induces other cellular responses at sublytic concentrations (see below).

Invasion of epithelial cells by encapsulated serotype 2 strains is somewhat controversial, as already mentioned above. Norton et al. (Reference Norton, Rolph, Ward, Bentley and Leigh1999) described invasion of HEp-2 epithelial cells by encapsulated sly+ but not sly− S. suis strains at a multiplicity of infection that was too low to cause suilysin-induced cell lysis. The authors suggested that suilysin mediates invasion of upper respiratory tract epithelial cells by virulent S. suis serotype 2 strains. Whether or not suilysin may be involved in invasion of (encapsulated and unencapsulated) S. suis in epithelial cells remains to be further elucidated.

Suilysin seems to contribute to the ability of virulent encapsulated serotype 2 strains to resist killing by porcine neutrophils in the presence of complete serum lacking specific IgG (Chabot-Roy et al., Reference Chabot-Roy, Willson, Segura, Lacouture and Gottschalk2006; Benga et al., Reference Benga, Fulde, Neis, Goethe and Valentin-Weigand2008). Addition of suilysin at sublytic concentration restored the phenotype of the isogenic sly mutant suggesting that some other mechanism in addition to cytotoxicity was involved (Chabot-Roy et al., Reference Chabot-Roy, Willson, Segura, Lacouture and Gottschalk2006). Our study indicated that expression of suilysin results in reduced internalization by porcine neutrophils but not reduced adherence to these cells (Benga et al., Reference Benga, Fulde, Neis, Goethe and Valentin-Weigand2008). No differences in the phenotype of the attenuated sly mutant were found between opsonizing and non-opsonizing conditions, in contrast to the work by Chabot-Roy et al. (Reference Chabot-Roy, Willson, Segura, Lacouture and Gottschalk2006). Only the latter is consistent with the speculation that suilysin leads to reduced complement deposition on the bacterial surface, as has been demonstrated for pneumolysin in the pathogenesis of pneumococci (Yuste et al., Reference Yuste, Botto, Paton, Holden and Brown2005). Alternatively, suilysin might actively interfere with the ability of neutrophils to kill S. suis.

Dominguez-Punaro et al. (Reference Dominguez-Punaro, Segura, Plante, Lacouture, Rivest and Gottschalk2007) investigated inflammatory reactions in the brain of CD1 mice after experimental infection with a suilysin producing serotype 2 strain. In agreement with a previous in vitro study (Graveline et al., Reference Graveline, Segura, Radzioch and Gottschalk2007), they detected upregulation of TLR 2, and to a lesser extent also TLR 3 but not TLR 4 mRNA. As discussed by the authors, this finding suggests a different mechanism of pathogenesis for suilysin in comparison to pneumolysin. Activation of TLR 4 by pneumolysin has been demonstrated to be crucial for a protective innate immune response against pneumococci (Malley et al., Reference Malley, Henneke, Morse, Cieslewicz, Lipsitch, Thompson, Kurt-Jones, Paton, Wessels and Golenbock2003).

Expression of cytokines (IL-1, IL-6 and TNF-α) and chemokines (IL-8 and MCP-1) in a whole-blood culture system did not show significant differences between the wild-type strain and a suilysin knock-out mutant (Segura et al., Reference Segura, Vanier, Al Numani, Lacouture, Olivier and Gottschalk2006). In earlier studies, this group had found that purified suilysin did not induce production of TNF-α and IL-6 by murine macrophages. In contrast, Lun et al. (Reference Lun, Perez-Casal, Connor and Willson2003) reported that suilysin triggered production of TNF-α in human monocytes and IL-6 production by pig pulmonary alveolar macrophages and monocytes. Furthermore, human endothelial cells secrete increased amounts of IL-6 and IL-8, but not MCP-1 after stimulation with purified suilysin at sublytic concentrations (Vadeboncoeur et al., Reference Vadeboncoeur, Segura, Al Numani, Vanier and Gottschalk2003). Sublytic concentrations of suilysin induced also transcription of the chemokine genes encoding IL-8 and MCP-1 in HEp-2 cells (our unpublished observations).

As multiple functions have been postulated for suilysin in interaction with host cells, severe attenuation of suilysin knock out mutants might be expected. However, attenuation in virulence of a suilysin knock-out mutant has only been demonstrated in mice (Allen et al., Reference Allen, Bolitho, Lindsay, Khan, Bryant, Norton, Ward, Leigh, Morgan, Riches, Eastty and Maskell2001). In contrast, in piglets different isogenic suilysin serotype 2 mutants were as virulent as the wild-type (Lun et al., Reference Lun, Perez-Casal, Connor and Willson2003), or were only slightly attenuated in virulence (Allen et al., Reference Allen, Bolitho, Lindsay, Khan, Bryant, Norton, Ward, Leigh, Morgan, Riches, Eastty and Maskell2001). On the other hand, the sly gene is more frequent among isolates from pigs with meningitis, septicemia and arthritis than in those from pigs with pneumonia (Staats et al., Reference Staats, Plattner, Stewart and Chengappa1999; King et al., Reference King, Heath, Luque, Tarradas, Dowson and Whatmore2001; Silva et al., Reference Silva, Baums, Rehm, Wisselink, Goethe and Valentin-Weigand2006). Furthermore, geographical differences in prevalence of sly+ strains have been described. The most striking difference is the low prevalence of suilysin positive strains in North America in comparison to Europe and Asia (Gottschalk et al., Reference Gottschalk, Lebrun, Wisselink, Dubreuil, Smith and Vecht1998).

Immunization with purified suilysin completely protected BALB/c mice against challenge with a highly virulent serotype 2 strain (Jacobs et al., Reference Jacobs, Loeffen, vandenBerg and Storm1994). It is less clear whether suilysin is also a protective antigen in piglets. Challenge of three piglets immunized with purified suilysin suggested partial protection (Jacobs et al., Reference Jacobs, vandenBerg and Loeffen1996, see also Table 4). However, we observed no protection against challenge with virulent serotype 2 and 9 strains after intranasal immunization with a live vaccine, though seroconversion against suilysin and suilysin neutralizing antibodies were elicited (Kock et al., Reference Kock, Beineke, Seitz, Ganter, Waldmann, Valentin-Weigand and Baums2009).

In summary, suilysin is the only known secreted hemolysin of the CDC family in S. suis. Though suilysin is not essential for virulence of S. suis serotype 2 strains in pigs, in vitro experiments suggest that it may well have important functions in interactions with host cells.

Other hemolysins

Suppression subtractive hybridization analysis of the Chinese STSS strain HA9801 (used as tester) and the avirulent serotype 2 European strain T15 (driver) revealed, among other fragments, an ORF encoding a putative hemolysin type III. This DNA sequence was found in many virulent serotype 2 and 9 strains (Jiang et al., Reference Jiang, Fan and Lu2009). Further ORFs encoding putative hemolysins are annotated in the genomes of the sequenced S. suis serotype 2 strains (Chen et al., Reference Chen, Tang, Dong, Wang, Feng, Wang, Zheng, Pan, Liu, Li, Song, Zhu, Sun, Feng, Guo, Ju, Ge, Dong, Sun, Jiang, Wang, Yan, Yang, Wang, Gao, Yang, Wang and Yu2007, see also Table 3). However, at present there is no experimental evidence for expression of any other active hemolysin except suilysin.

Phospholipase C

Culture supernatants of S. suis serotype 2 strains induce degradation of phosphatidylcholine to diacylglycerol and choline phosphate (but not to phosphatidic acid), indicating that the tested strains secrete a phospholipase C (Jobin et al., Reference Jobin, Fortin, Willson, Gottschalk and Grenier2005a). Phospholipase C activity was identified in a search for potential factors in S. suis culture supernatants that are involved in the release of arachidonic acid from human BMECs observed in co-cultivation experiments with S. suis culture supernatants. Release of arachidonic acid was reduced in human BMECs treated with culture supernatants from a suilysin deficient mutant, but arachidonic acid was not released after treatment with suilysin. Based on these results the authors discussed a synergistic effect of suilysin and phospholipase C on the release of this important mediator by BMECs, which might play an important role in the pathogenesis of S. suis-associated meningitis.

EF

The 110-kDa EF encoded by the gene epf was identified as a protein associated with virulence in serotype 1 and 2 strains (Vecht et al., Reference Vecht, Wisselink, Jellema and Smith1991). The function of EF is still unknown. Isogenic epf mutants of serotypes 1 and 2 were as virulent as the wild-type in experimental infections indicating that EF is associated but not essential for virulence (Smith et al., Reference Smith, Vecht, Wisselink, StockhofeZurwieden, Biermann and Smits1996). So far, EF producing strains have been identified only in serotypes 1, 2, 1/2, 14 and 15 (Wisselink et al., Reference Wisselink, Smith, Stockhofe-Zurwieden, Peperkamp and Vecht2000).

Large variants (>110 kDa) of EF (EF*) are expressed by some serotype 2 strains. They are characterized by long C-terminal tandem repeats (each 76 amino acids long). These repetitive sequences are missing in the 110 kDa EF. In contrast, the N-terminus, namely the first 811 amino acids, of EF and EF* are nearly identical. Based on the repeats five different classes of EF* proteins have been differentiated (Smith et al., Reference Smith, Reek, Vecht, Gielkens and Smits1993). PCR assays are also available for differentiation of the different size variants of epf (Wisselink et al., Reference Wisselink, Joosten and Smith2002; Silva et al., Reference Silva, Baums, Rehm, Wisselink, Goethe and Valentin-Weigand2006).

High mortality and morbidity was observed after experimental intranasal infection of newborn germfree piglets with MRP+ EF+ (referring to the 110 kDa protein) but not MRP+ EF* serotype 2 strains (Vecht et al., Reference Vecht, Wisselink, van Dijk and Smith1992). This finding is contradictory to the high prevalence of MRP+ EF* serotype 2 (or the respective genotype) strains among invasive isolates in various epidemiological studies (Martinez et al., Reference Martinez, de Castro, Pagnani, Nakazato, da Silveira and Gottschalk2003; Silva et al., Reference Silva, Baums, Rehm, Wisselink, Goethe and Valentin-Weigand2006). Furthermore, intravenous application of 10⁸ CFU of an MRP+ EF* serotype 2 strain resulted in high mortality in piglets (our unpublished results).

Approximately 10% of the wild boars in Northwestern Germany carry mrp+ epf* cps2 strains on their tonsils, whereas mrp+ epf+ (encoding the 110 kDa protein) cps2 strains were undetectable among these animals, in contrast to control domesticated piglets from the same region (Baums et al., Reference Baums, Verkühlen, Rehm, Silva, Beyerbach, Pohlmeyer and Valentin-Weigand2007). Therefore, we speculated that factors associated with modern swine production have led to the selection of the highly virulent MRP+ EF+ (110 kDa protein) serotype 2 strains. The high virulence of these strains might, however, be specific for domestic pigs as S. suis serotype 2 isolates from humans in Europe frequently express the large EF* variants, but not the 110 kDa EF protein (Smith et al., Reference Smith, Reek, Vecht, Gielkens and Smits1993; Baums et al., Reference Baums, Verkühlen, Rehm, Silva, Beyerbach, Pohlmeyer and Valentin-Weigand2007).

Prevalences of various EF+ or EF* positive strains show substantial geographical variation. In contrast to Europe, MRP+ EF+ and MRP+ EF* serotype 2 strains have been detected only in single cases in North America (Gottschalk et al., Reference Gottschalk, Lebrun, Wisselink, Dubreuil, Smith and Vecht1998).

The 110-kDa EF and the high molecular variants of EF are immunogenic proteins. Convalescent phase sera from animals infected either with MRP+ EF+ or MRP+ EF* serotype 2 strains generally contain high antibody titers against EF (our unpublished results). Immunization with EF alone did not elicit protection against MRP+ EF+ serotype 2 challenges, but a vaccine containing purified MRP and EF together with a water-in-oil adjuvant was as protective as a bacterin (Wisselink et al., Reference Wisselink, Vecht, Stockhofe-Zurwieden and Smith2001). Nevertheless, a major problem of an EF-based vaccine is that many virulent S. suis strains, including all serotype 9 strains, do not express this factor.

Bacteriocin-like inhibitory substances (BLISs)

S. suis is not only an important invasive pathogen but also a successful colonizer of the upper respiratory, genital and alimentary tracts (Baele et al., Reference Baele, Chiers, Devriese, Smith, Wisselink, Vaneechoutte and Haesebrouck2001; Su et al., Reference Su, Yao, Perez-Gutierrez, Smidt and Zhu2008). As part of such microbial ecosystems S. suis interacts with numerous other bacteria including other S. suis strains. Melancon and Grenier (Reference Melancon and Grenier2003) demonstrated production of BLISs in 4 of 36 tested S. suis strains, which were active against other S. suis strains and other bacterial species associated with pigs. These BLISs were sensitive to proteinase K, pronase and elastase indicating that they are proteinaceous. BLISs are thought to be of advantage for the producing strain to compete in a microbial ecosystem, but this has not been reported for S. suis.

Immunoglobulin G-binding protein

S. suis strains of various serotypes express at least one IgG-binding protein (Serhir et al., Reference Serhir, Higgins, Foiry and Jacques1993). Serotype 2 strains secrete a 52-kDa IgG-binding protein that reacts with a large variety of mammalian IgG and IgA molecules (Serhir et al., Reference Serhir, Dubreuil, Higgins and Jacques1995). The binding of IgG does not depend on the glycosylation state of the immunoglobulin molecule. The S. suis protein exhibits lower IgG-binding affinities than protein A and protein G, especially with regard to human IgG. In contrast to protein A and protein G, the S. suis IgG-binding protein also targets chicken IgG.