Streptococcus suis serotype 2 binding to extracellular matrix proteins


  • Edited by R.Y.C. Lo

*Corresponding author. Present address: GREMIP, Faculté de médecine vétérinaire, Université de Montréal, 3200 rue Sicotte, P.O. Box 5000, Saint-Hyacinthe, Que., Canada J2S 7C6. Tel.: +1 450 773 8521x8374; fax: +1 450 778 8108, E-mail address:


Streptococcus suis serotype 2 is a major swine and human pathogen that causes septicemia and meningitis. The ability of S. suis serotype 2 to bind to different extracellular matrix (ECM) proteins was evaluated by ELISA. All 23 strains tested bound to plasma and cellular fibronectin and collagen types I, III, and V, some to fibrin, vitronectin, and laminin, and none to the other ECM proteins tested. An unencapsulated isogenic mutant bound to ECM proteins better than its parental encapsulated strain, suggesting that the polysaccharide capsule interfered with binding. Cross-inhibition was observed between soluble plasma fibronectin and collagens in the ECM adherence assay, indicating that binding domains for both proteins exist on the same or nearby bacterial surface molecules. On the other hand, pre-incubation with plasma fibronectin increased binding to collagen IV, suggesting that S. suis might use fibronectin as a bridging molecule. The results of heat treatment and proteolytic digestion suggest that adhesins for these ECM proteins are proteinaceous in nature.


Streptococcus suis is a major swine pathogen that causes septicemia, meningitis, endocarditis, and arthritis [1]. Of the 35 known serotypes, serotype 2 is most frequently isolated and associated with diseases [1]. This pathogen is also a zoonotic agent that causes similar diseases in humans [2–4].

Little is known about S. suis virulence factors. The capsule polysaccharide (CPS) is the only proven virulence factor since unencapsulated isogenic mutants are completely avirulent and are rapidly cleared from circulation in pig and mouse infection models [5,6]. Other putative virulence factors include a hemolysin (suilysin), a 136-kDa muramidase-released protein (MRP), a 110-kDa extracellular factor (EF) protein, a hyaluronidase, a superoxide dismutase, various proteases, and bacteriocins [7].

The extracellular matrix (ECM) is a stable macromolecular structure underlying epithelial and endothelial cells and surrounding connective tissue cells [8] that becomes exposed when tissue integrity is disturbed by lesions or traumas. The ECM is composed of glycoproteins and glycosaminoglycans (GAG). Its composition differs in various organs, but the main components are fibronectin, collagen, elastin, laminin and GAG, like heparin, and heparan sulfate [9]. Many of these proteins can serve as potential cell receptors for bacteria and participate in the infectious process [8,10–13]. In addition, bacterial binding to plasma proteins such as fibronectin and fibrinogen can prevent recognition of bacteria by the host immune system by masking immunogenic epitopes [14].

Binding of virulent bacteria to host cells is the first step in the colonization of mucosal surfaces. Binding may also be a first step in the invasion of host cells, a process that may lead to bacteremia and sepsis [15]. The pathogenesis of infections caused by S. suis is not fully understood and many steps are probably involved [16]. S. suis is able to bind to and, in some cases, invade endothelial and epithelial cells of human and porcine origin [16–20]. However, the mechanisms involved in these interactions are unknown.

While various streptococci specifically bind to host ECM and while these interactions play a role in disease pathogenesis, little is known about the ability of S. suis to bind to ECM proteins. Recently, a fibronectin–fibrinogen-binding protein (FBPS) protein has been proposed as a contributing factor in the colonization of organs due to its binding to ECM proteins of host cells [21]. However, interactions between S. suis and various ECM proteins have never been studied. The purpose of this study was to evaluate the ability of S. suis serotype 2 field strains to specifically bind to various ECM proteins.

2Materials and methods

2.1Bacterial strains and growth conditions

S. suis serotype 2 strain S735 (ATCC 43765) was used as the reference strain. An unencapsulated mutant strain (B218) obtained in our laboratory by allelic exchange [22] and corresponding to a transposon-derived mutant described previously [23] and its virulent parental strain (31533) were also included. Strain B218 was shown to possess the same characteristics that mutant 2A, with a complete absence of capsular material at the bacterial surface. In selected experiments, 23 field strains of S. suis serotype 2 were used. These strains originated from different species (human or swine), different pathologies and different geographical origins. A complete list is presented in Table 1. Strains were grown as previously described [24]. Late exponential-phase bacteria were washed three times in phosphate-buffered saline (PBS) (pH 7.3) and killed by suspending them in 0.2% (v/v) formaldehyde overnight at 4 °C.

Table 1.  Characteristics of the Streptococcus suis strains used in this study
StrainSourceGeographical origin
S735Pig: meningitisEurope
98-B575Pig: endocarditisCanada
99-1539BPig: endocarditisCanada
98-C462BPig: endocarditisCanada
98-8993Pig: endocarditisCanada
98-B099Pig: meningitisCanada
98-B719Pig: meningitisCanada
SS166Pig: meningitisEurope
D-282Pig: meningitisEurope
89-1591Pig: meningitisCanada
90-1330Healthy pigCanada
94-623Healthy pigEurope
24Pig: septicemiaEurope
T15Healthy pigEurope
91-1804Human: endocarditisCanada
98-3634Human: endocarditisCanada
Biotype2/hemoHuman: endocarditisEurope
ReimsHuman: spondylodiscitisEurope
94-3037Human: meningitisEurope
FRU95Human: meningitisEurope
LEF95Human: meningitisEurope
31533Pig: meningitisEurope
B218Unencapsulated mutant derived from strain 31533 


Commercially available human ECM proteins were used. Cellular fibronectin, collagen types III and V, laminin, plasma vitronectin, plasma fibrin, and fibrinogen were from Sigma Chemical Co. (St. Louis, MO, USA); plasma fibronectin was from Roche Diagnostics Corporation (Indianapolis, IN, USA); collagen types I and IV were from BD Biosciences (Bedford, MA, USA); and elastin was from EMD Biosciences Inc. (La Jolla, CA, USA). Porcine fibrinogen was from Sigma, pronase and proteinase K were from Roche, and trypsin was from Gibco (Burlington, Ont., Canada).

2.3Microtiter plate binding assay

Maxisorp? flat-bottom microtiter 96-well plates (Nunc, VWR, Mississauga, Ont., Canada) were coated with 100 μl (0–50 μg ml−1, depending on the experiment) of ECM protein (plasma fibronectin; cellular fibronectin; collagen types I, III, IV, and V; laminin; elastin; vitronectin; fibrin; human fibrinogen; and porcine fibrinogen) in 0.1 M carbonate coating buffer (pH 9.6) and were incubated overnight at 4 °C. The plates were washed three times with PBS containing 0.05% (v/v) Tween 20 (PBST, pH 7.3), and 200 μl of 3% (w/v) non-fat dry milk in PBST was added to each well to prevent non-specific bacterial binding. After 1 h at 37 °C, the wells were washed three times with PBST. Formaldehyde-killed bacterial suspensions (100 μl) of individual strains were added and the plates were incubated for 2 h at 37 °C. Different bacterial concentrations, incubation times, and temperatures were also tested. All unbound bacteria were subsequently removed by washing the wells three times with PBST. A 100 μl volume S. suis serotype 2-specific rabbit antiserum (diluted 1/3000 in PBST) prepared as previously described [25] was then added to each well. The plates were incubated for 1 h at 37 °C. The wells were washed three times with PBST and 100 μl of horseradish peroxidase-labelled anti-rabbit IgG (Jackson Immunoresearch Laboratories Inc., West Grove, PA,USA) (diluted 1/8000 in PBST) was added. The plates were incubated for 1 h at 37 °C with the secondary antibody. After washing three times with PBST, 3,3′,5,5′-tetramethylbenzidine (Zymed, San Francisco, CA, USA) was used as the enzyme substrate according to the manufacturer's instructions. The reactions were stopped by adding 25 μl of H2SO4 (1 N) and were read at 450 nm using a microplate reader (Uvmax; Molecular Devices, Menlo Park, CA, USA). Uncoated wells served as background controls. Casein-coated wells served as a control for non-specific adhesion of S. suis to protein-coated wells.

2.4Proteolytic, formaldehyde, and heat treatments

For the proteolytic treatment, 1 ml of live bacterial suspension (108 bacteria ml−1) was centrifuged and resuspended in the same volume of PBS containing one of following enzymes: trypsin, pronase, or proteinase K. All proteolytic enzymes were used at a concentration of 1 mg ml−1. The suspensions were incubated for 1 h at 37 °C for trypsin and proteinase K and 40 min for pronase. Controls with either an untreated suspension of live bacteria or 0.2% formaldehyde-treated bacteria were also incubated for 1 h at 37 °C. The suspensions were then washed three times in PBS and resuspended in 1 ml of PBS for use in microtiter plate binding assay. The heat sensitivity of the components involved in the adhesion of bacteria to the ECM proteins was evaluated by heating bacterial suspensions for 30 min at 60 or 100 °C.

2.5Inhibition assays

For the inhibition assays, bacterial suspensions were pre-incubated for 120 min with 150 μg ml−1 of various ECM proteins (plasma fibronectin, cellular fibronectin, collagen type I, collagen type III, or collagen type V) before being incubated with immobilized fibronectin, collagen, or fibrin. After the pre-incubation period, the suspensions were centrifuged to remove unbound protein, the bacteria were added to the wells, and the ELISA was performed as described above.

2.6Statistical analysis

ELISA tests were performed at least three times for each binding assay. Differences between strains (31533 vs. B218), different treatments (proteolytic, formaldehyde, and heat treatments), and fibronectin-mediated binding (pre-incubation with fibronectin vs. pre-incubation with PBS) were analyzed for significance using general linear models followed by the Tukey–Kramer post hoc test for differences between the various incubation temperatures (4, 20, and 37 °C).


3.1Binding of S. suis serotype 2 reference strain S735 to ECM proteins

As shown in Fig. 1, strain S735 bound to both fibronectins (plasma and cellular) and to collagen types I, III, and V. This strain did not bind to the other proteins tested (data not shown). No binding was observed in the casein-coated control wells, indicating that binding to the five ECM proteins was specific.

Figure 1.

(a) Effect of ECM protein concentrations on S. suis serotype 2 strain S735 binding. Plates were coated with different ECM protein concentrations (open symbols), bacteria were added at 108 bacteria ml−1, and binding was evaluated after 2 h. (b) Plates were coated with fixed protein concentrations (10 μg−1 ml−1 for plasma fibronectin, 25 μg−1 ml−1 for cellular fibronectin, and 12.5 μg−1 ml−1 for collagen types I, III, and V) (open symbols) and different bacterial concentrations (from 106 to 108 CFU ml−1) were added. The plates were then incubated for 2 h. Casein (10 μg−1 ml−1; closed symbols) was used as a negative control. Data are expressed as means ± SD of at least three experiments performed in triplicate.

3.2Binding of S. suis serotype 2 reference strain S735 as a function of protein and bacterial concentration, time, and temperature of incubation

Adherence positively correlated with bacterial and protein concentrations. The binding of strain S735 increased with increasing protein concentration and reached a plateau at 10 μg ml−1 for plasma fibronectin, 25 μg ml−1 for cellular fibronectin, and 12.5 μg ml−1 for collagen types I, III, and V (Fig. 1(a)). However, even with these proteins, no binding was detected with bacterial concentrations under 107 bacteria ml−1 (Fig. 1 (b)).

Strain S735-binding to both fibronectins and collagen types I and V reached half-maximum between 30 and 60 min and maximum binding at 120 min, while binding to collagen type III reach half-maximum between 60 and 90 min and maximum at 120 min (Fig. 2). Similar results were obtained with the non-encapsulated strain B218 (result not shown).

Figure 2.

Time course of S. suis serotype 2 strain S735-binding to different ECM proteins (open symbols). Wells were coated either with 10 μg−1 ml−1 of plasma fibronectin, 25 μg−1 ml−1 of cellular fibronectin, or 12.5 μg−1 ml−1 of collagen types I, III, or V and incubated with bacteria (108 CFU ml−1). Data are expressed as means ± SD of at least three experiments performed in triplicate. Casein (10 μg−1 ml−1; closed symbols) was used as a negative control.

Strain S735-binding to both fibronectins and collagen types I, III, and V was also measured at different temperatures (4, 20, and 37 °C). As shown in Fig. 3, the optimum incubation temperature was 37 °C for all five ECM proteins. In general, S735-binding to the collagens was significantly (p < 0.0001) more temperature-dependent than for the two fibronectins. Similar results were obtained with the non-encapsulated strain B218 (result not shown).

Figure 3.

Effect of incubation temperature on S. suis serotype 2 strain S735-binding to different ECM proteins. Wells were coated with either 10 μg−1 ml−1 of plasma fibronectin (pFn), 25 μg−1 ml−1 of cellular fibronectin (cFn), or 12.5 μg−1 ml−1 of collagen type I (Cn I), type III (Cn III), or type V (Cn V) and were incubated with bacteria (108 CFU ml−1) at different temperatures for 2 h [4 °C (white bars), 20 °C (grey bars), and 37 °C (black bars)]. Data are expressed as means ± SD from at least three experiments performed in triplicate.

3.3Binding capacity of various S. suis serotype 2 field strains to immobilized ECM proteins

All the strains tested (Table 1), bound to both fibronectins and collagen types I, III, and V. However, individual differences in the degree of binding of the various strains were observed (data not shown). Unlike strain S735, some field strains were able to bind to fibrin, and laminin. None of the strains bound to collagen type IV, elastin, human fibrinogen, or porcine fibrinogen. No correlation was noted between the ability of the various strains to bind to the ECM proteins and their source (diseased or healthy animals), the host from which they had been recovered (swine or human) (Table 2A), or their geographical origin (European or North American) (data not shown).

Table 2.  Adherence to ECM proteins of different S. suis serotype 2 field strains according to species of isolation or presence of disease (A) and the effect of the capsule on adhesion between S. suis serotype 2 and ECM proteins (B)
 Diseased pigsHealthy pigsHumans31533B218
  1. B218: unencapsulated mutant derived from S. suis serotype 2 strain 31533. pFn, plasma fibronectin; cFn, cellular fibronectin; Cn I, collagen type I; Cn III, collagen type III; Cn IV, collagen type IV; Cn V, collagen type V; pFg, porcine fibrinogen; hFg, human fibrinogen; Lam, laminin; Vitro, vitronectin; Elast, elastin. Casein was used as negative control. Table 2B data are represented as means (OD450) ± SD.

  2. *General linear model, (p < 0.0001).

pFn12/123/37/72.48 ± 0.21*3.16 ± 0.07*
cFn12/123/37/72.06 ± 0.09*2.34 ± 0.12*
Cn I12/123/37/71.83 ± 0.101.81 ± 0.07
Cn III12/123/37/72.34 ± 0.12*3.79 ± 0.19*
Cn IV0/120/30/70.07 ± 0.040.25 ± 0.02
Cn V12/123/37/71.72 ± 0.18*3.89 ± 0.07*
Fibrin5/122/31/70.15 ± 0.050.29 ± 0.01
pFg0/120/30/70.14 ± 0.010.32 ± 0.01
hFg0/120/30/70.12 ± 0.030.32 ± 0.04
Lam3/121/31/70.07 ± 0.01*1.32 ± 0.19*
Vitro0/120/30/70.10 ± 0.01*0.58 ± 0.01*
Elast0/120/30/70.13 ± 0.060.28 ± 0.03
Casein0/120/30/70.19 ± 0.020.25 ± 0.09

The unencapsulated mutant strain B218 bound significantly (p < 0.0001) more to both fibronectins, collagen type III, and collagen type V than the parental strain 31533. In addition, unlike 31533, B218 bound to vitronectin and laminin (Table 2B).

3.4Proteolytic, formaldehyde, and heat treatments

Strain S735-binding to ECM proteins was affected by both proteolytic and heat treatments, decreasing by almost 90% of adhesion for all five ECM proteins tested (plasma and cellular fibronectin, and collagen types I, III and V). The 0.2% formaldehyde treatment did not affect binding to ECM proteins (p < 0.0001) (results not shown). Similar results were obtained with the non-encapsulated strain B218 as well as with three different field strains randomly tested (results not shown).

3.5Inhibition assays

Plasma fibronectin and collagen types I, III, and V, but not cellular fibronectin (even at high concentrations), competed with the homologous protein used in the ELISA assay, confirming the specificity of the binding (data not shown). Collagen at a concentration of 150 μg ml−1 inhibited binding not only to the homologous collagen but also to the two other collagens (data not shown). Pre-incubation of the bacteria with 150 μg ml−1 of plasma fibronectin significantly (p < 0.0001) reduced binding to immobilized cellular fibronectin and collagen types I, III, and V (Fig. 4). Since fibronectin is a multifunctional protein that also carries recruitment domains for host factors such as collagen and fibrin, strain S735-binding to collagen type IV and fibrin after pre-incubation with plasma fibronectin was evaluated. The addition of soluble plasma fibronectin promoted binding of this strain to collagen type IV (p < 0.0001) (Fig. 4) but not to fibrin (data not shown).

Figure 4.

Effect of pre-incubating bacteria for 90 min with 150 μg−1 ml−1 of soluble plasma fibronectin on S. suis serotype 2 strain S735-binding to plasma fibronectin (pFn), cellular fibronectin (cFn), collagen type I (Cn I), collagen type III (Cn III), collagen type V (Cn V), and collagen type IV (Cn IV). Data represent the binding of S. suis pre-incubated with plasma fibronectin (white bars) or PBS (black bars). Bars represent means ± SD of at least three experiments performed in triplicate.


Many adhesins have been described for S. suis serotype 2 in the last decade, including an 18-kDa galactosyl-α 1–4 galactose binding adhesin, which is present on different S. suis serotypes; a 60-kDa IgG-binding protein, which reacts with a large number of IgGs; and a 39-kDa glyceraldehyde-3-phosphate dehydrogenase (GAPDH), which binds to various host proteins, including plasminogen and albumin [26–29]. GAPDH is known to be involved in the binding of S. suis to host tissues because mutants defective in the surface expression of this protein bind to a lesser degree to tracheal cells and porcine tracheal rings [30,31]. Lastly, a 64-kDa fibronectin–fibrinogen-binding protein (FBPS) with a high binding capacity for fibronectin and fibrinogen has recently been described [21]. While the fbps gene is present in most S. suis serotypes and pathotypes and while FBPS has been cloned and its binding activity verified [21], the expression of FBPS in the different serotypes and phenotypes was not studied, and no direct studies of whole S. suis bacteria or constructed fbps knockout mutant binding to fibronectin/fibrinogen have been carried out. The role of FBPS in bacterial pathogenesis is not well understood. However, studies with the fbps mutant suggest that this protein plays a role in S. suis colonization of various organs [21]. We report here for the first time that whole S. suis cells are able to bind to fibronectin, confirming published data using purified proteins [21]. However, we could not confirm that S. suis binds to fibrinogen, probably because we used whole bacteria rather than purified protein.

Fibronectin exists in two main forms. Cellular fibronectin, which is an insoluble glycoprotein dimer that serves as a linker in the ECM, is produced by fibroblasts, chondrocytes, endothelial cells, macrophages, and certain epithelial cells. Plasma fibronectin, which is a soluble disulfide-linked dimer, is produced by hepatocytes and released in an unbound form into the plasma. Our study shows that all the S. suis serotype 2 strains tested are able to bind to both plasma and cellular fibronectin. In addition, S. suis serotype 2 binds to immobilized cellular fibronectin but not to soluble cellular fibronectin. Similar results have been obtained with Group B streptococci (GBS) and plasma fibronectin. In fact, GBS are able to bind to immobilized but not soluble plasma fibronectin [32]. These results might be due to structural changes resulting in the exposure of determinants required for bacterial binding when fibronectin is attached to a solid surface. On the other hand, S. suis binds to both immobilized and soluble plasma fibronectin, probably due to the soluble disulphide-linked dimer conformation of plasma fibronectin, which allows a monomer conformation that exposes the determinants required for S. suis binding. Fibronectin appears to be one of the components mediating binding to and, in some cases, invasion of various host cell types by streptococci and staphylococci [33–35], but its role in S. suis binding to and/or invasion of epithelial, endothelial, and phagocytic cells [5,17,18,20,36,37] remains to be elucidated.

Collagen is a major constituent of the ECM and, as such, may be a major target for pathogenic bacteria. Collagen is found in basement membranes (type IV), bone, skin, cartilage, tendons, and joints [14]. This study demonstrated that S. suis serotype 2 is able to directly bind to collagen types I, III, and V, but not to type IV, probably due to structural differences [38] that might hide receptors required for S. suis binding. Interestingly, while S. suis is not able to bind directly to collagen type IV, it has the potential to adhere to this protein via a surface-bound fibronectin mechanism. Fibronectin-mediated collagen recruitment might represent a novel mechanism of colonization and immune evasion for Streptococcus pyogenes[14].

S. suis serotype 2 binds to both fibronectin and collagen in a bacterial and protein concentration-dependent manner. No binding was detected at bacterial concentrations under 107 bacteria ml−1. Interestingly, it has been previously demonstrated in vivo that the presence of clinical signs and symptoms in diseased animals infected with S. suis correlates with high levels of bacteria in the bloodstream [39].

Cross-inhibition was observed with soluble fibronectin and collagen in the ECM binding assay, indicating that fibronectin and collagen binding domains exist on the same or nearby bacterial surface adhesins. Similar results were obtained with GBS by Tamura and Rubens [32].

The bacterial ligand(s) that allows binding to ECM proteins could be either a protein [40] or a cell wall component [41]. The dramatic reduction in binding by trypsin-, pronase-, and proteinase K-treated bacteria indicates, in the case of S. suis, a protein-mediated binding mechanism. This conclusion is supported by the fact that heat treatments (60 and 100 °C) also decreased binding to ECM proteins.

Most of the S. suis serotype 2 field strains tested behaved in a fashion similar to that of the reference strain. However, some were also able to bind to laminin, and fibrin, suggesting that they possessed other adhesins. On the other hand, interference by the CPS of some strains cannot be ruled out. In fact, our study showed that unencapsulated mutant B218 not only bound better to both fibronectins and collagen types III and type V than the encapsulated parental strain 31533, but it also bound to vitronectin and laminin, unlike strain 31533. In addition, Tikkanen et al. [29] reported an inverse relationship between the binding activity of the galactosyl-α 1–4 galactose binding adhesin and the expression of capsular polysaccharide. It has already been reported that CPS also influences the binding of S. pyogenes, Staphylococcus aureus[34], and Lactobacillus acidophilus[42] to some ECM proteins.

In summary, this study shows that S. suis serotype 2 is able to specifically bind to major constituents of ECM such as fibronectin and collagens, and that the adhesin(s) responsible for these interactions are proteinaceous in nature. The identification of the S. suis surface protein(s) responsible for this binding phenotype and the creation of isogenic knock-out mutants of the associated gene(s) may give some insights about the role of these adhesins in the pathogenesis of the infection caused by S. suis. These efforts are ongoing in our laboratory.


This work was supported by the Canadian Research Network on Bacterial Pathogens of Swine and by NSERC Grant 154280-04. We thank Dr. M. Segura and Dr. D. Dubreuil for reviewing the manuscript.