Streptococcus pyogenes causes a broad spectrum of infectious diseases, including pharyngitis, skin infections and invasive necrotizing fasciitis. The initial phase of infection involves colonization, followed by intimate contact with the host cells, thus promoting bacterial uptake by them. S. pyogenes recognizes fibronectin (Fn) through its own Fn-binding proteins to obtain access to epithelial and endothelial cells in host tissue. Fn-binding proteins bind to Fn to form a bridge to α5β1-integrins, which leads to rearrangement of cytoskeletal actin in host cells and uptake of invading S. pyogenes. Recently, several structural analyses of the invasion mechanism showed molecular interactions by which Fn converts from a compact plasma protein to a fibrillar component of the extracellular matrix. After colonization, S. pyogenes must evade the host innate immune system to spread into blood vessels and deeper organs. Some Fn-binding proteins contribute to evasion of host innate immunity, such as the complement system and phagocytosis. In addition, Fn-binding proteins have received focus as non-M protein vaccine candidates, because of their localization and conservation among different M serotypes.Here, we review the roles of Fn-binding proteins in the pathogenesis and speculate regarding possible vaccine antigen candidates.
Streptococcus pyogenes is a Gram-positive pathogen that belongs to Lancefield serogroup A, also known as group A Streptococcus, and causes various diseases, such as pharyngitis, rheumatic fever and toxic shock syndrome. It has been reported that severe S. pyogenes diseases are responsible for over 500 000 deaths each year throughout the world, while there are more than 100 million prevalent cases of S. pyogenes pyoderma and over 600 million new cases each year of S. pyogenes pharyngitis (Carapetis et al., 2005). In addition, economic loss is caused by S. pyogenes infectious diseases and it has been estimated that the total cost of S. pyogenes pharyngitis among children in the USA ranges from $224 to $539 million per year (Pfoh et al., 2008).
Streptococcus pyogenes had generally been considered to be an extracellular pathogen. However, it was demonstrated that the organisms can invade host epithelial cells via their own fibronectin (Fn)-binding proteins (Molinari et al., 1997; Henderson et al., 2011). Currently, at least 11 Fn-binding proteins of S. pyogenes have been identified, of which there are two major types (Table 1). In the first type, Protein F1 (PrtF1)/SfbI, Protein F2 (PrtF2)/PFBP, FbaA (formerly Fba), FbaB, SfbII/Serum opacity factor (SOF), SfbX and Fbp54 each contain Fn-binding repeats, whereas in the second type, M1 protein, GAPDH/Plr, Shr and Scl1 do not (Hanski and Caparon, 1992; Pancholi and Fischetti, 1992; Talay et al., 1992; Courtney et al., 1994; Kreikemeyer et al., 1995; Rakonjac et al., 1995; Jaffe et al., 1996; Rocha and Fischetti, 1999; Cue et al., 2001; Terao et al., 2001; 2002; Jeng et al., 2003; Fisher et al., 2008; Caswell et al., 2010). Most strains possess Fn-binding proteins that contain Fn-binding repeat domains and an LPXTG motif in the C-terminal region, while the distribution of each Fn-binding protein correlates with M serotype (Terao et al., 2001; 2002). In addition, it was reported that highly virulent S. pyogenes strains possess one or more Fn-binding proteins (Rocha and Fischetti, 1999; Terao et al., 2002). Fn-binding repeats interact with the N-terminal domain of Fn, which includes five type I modules of Fn (FnI1–FnI5) (Fig. 1A). Fn-binding repeats are present in several bacterial species, such as staphylococci and streptococci (Henderson et al., 2011). However, the detailed molecular interactions between Fn and Fn-binding proteins that lack the repeats remain unknown.
Table 1. Reported Fn-binding proteins of S. pyogenes
Protein F1/S. pyogenes Fn-binding protein I
PrtF1/SfbI is composed of a signal sequence, aromatic domain, proline-rich region, functional upstream domain, Fn-binding repeats and a cell wall anchoring motif.
Protein F2/S. pyogenes Fn-binding protein
PrtF2/PFBP contain two Fn-binding domains in the C-terminal regions, one of which consists of three Fn-binding repeats and the other is a non-repeated domain.
Serum opacity factor/S. pyogenes Fn-binding protein II
SOF/SfbII is composed of an N-terminal opacification domain and C-terminal Fn-binding repeats region, and binds fibrinogen and fibulin-1.
S. pyogenes Fn-binding protein X
SfbX is encoded immediately downstream of SfbII/SOF and contains Fn-binding repeats on the C-terminus.
Fbp54 is calculated to have a molecular mass of 54 kDa, and has a high similarity to Fn-binding proteins in other streptococci.
Fn-binding protein of group A streptococci type A
FbaA contains a signal sequence, α-helical variable region, repeat domains and a cell wall anchoring motif. The fbaA gene is positively controlled by the mga gene.
Fn-binding protein of group A streptococci type B
FbaB contains a signal sequence, Fn-binding repeats and a cell wall anchoring motif. The fbaB gene is regulated by the msmR gene, as well as the prtF1/sfbI and prtF2/pfbp genes.
M1 protein is composed of two polypeptide chains that form an α-helical coiled coil configuration, while the chains are built of four repeat regions and a cell wall anchoring motif.
Streptococcal haemoprotein receptor
Shr contains two ‘near transporter’ domains that mediate Fn and laminin binding.
Streptococcal collagen-like surface protein 1
Scl-1 has been reported to be streptococcal collagen-like protein and contains a variable region, collagen-like region and cell wall anchoring domain.
Glyceraldehyde-3- phosphate dehydrogenase
GAPDH protein, located in the cytoplasm and on the bacterial cell surface, plays important roles in the glycolytic pathway and is essential for bacterial growth.
The initial phase of infection involves colonization, followed by intimate contact with the host cells, thus promoting bacterial uptake by them. S. pyogenes recognizes Fn through its own Fn-binding proteins to obtain access to epithelial and endothelial cells in host tissue. Recently, it was reported that several Fn-binding proteins contribute to bacterial evasion of host innate immunity. In the following, we present current understanding of Fn-binding proteins and their role in S. pyogenes infections.
Interaction between Fn and Fn-binding proteins
Fn is an extracellular matrix (ECM) glycoprotein that circulates in body fluids (approximately 300 μg ml−1) as a soluble compact dimer and is assembled by cells into the fibrillar matrix (Singh et al., 2010). It plays an essential role in dependent and ordered processes that require cell migration, such as wound healing and tissue remodelling. Fn ranges in size from 230 to 270 kDa depending on alternative splicing. Dimeric Fn contains three structural β-sheet motifs, the Fn modules type I, II and III (FnI, FnII and FnIII) (Fig. 1A). The Fn dimer is mediated by a pair of antiparallel disulfide bonds at the C-terminus. Cells mediate Fn matrix assembly via integrin binding to the RGD (Arg–Gly–Asp) cell-binding domain. The primary receptor for Fn matrix assembly is α5β1 integrin, which binds to the RGD sequence in III10 and the synergy site III9 (Singh et al., 2010). This integrin can bind to Fn lacking a synergy site, although such a site is required for α5β1 integrin-mediated assembly. In contrast to α5β1 integrin, other Fn-binding integrins, such as αvβ3, cannot mediate Fn to assemble fibrils without treatment to increase their activity.
Bacterial surface proteins bound to Fn form a bridge to α5β1-integrins, which leads to rearrangement of cytoskeletal actin in host cells and uptake of invading bacteria. The binding mechanisms of PrtF1/SfbI-Fn have been well analysed, while NMR spectrometry has determined the structures of FnI1–FnI2 and the Fn-binding repeat peptide complex (Schwarz-Linek et al., 2003). Fn-binding repeats form additional antiparallel β-strands on sequential FnI modules of Fn. The structure of the FnI module is characterized by a β-sandwich composed of two antiparallel β-sheets, a double-stranded sheet followed by a triple-stranded sheet. During the molecular interaction termed ‘tandem β-zipper’, the Fn-binding peptide contributes an additional fourth antiparallel strand to the triple-stranded sheet (Schwarz-Linek et al., 2003). Bacterial Fn-binding proteins that contain Fn-binding repeats seem to exploit the molecular structure of Fn by forming an extended tandem β-zipper. Using thermodynamic analysis, Norris et al. reported that the fifth Fn-binding repeat of PrtF1/SfbI is not a stably folded protein that is functional, but rather an intrinsically disordered region, and that the binding energy in the fifth repeat and Fn tandem β-zipper interaction are distributed across a large interface rather than concentrated in a few residues (Norris et al., 2011).
It has also been shown that the 49-residue FUD of PrtF1/SfbI interacts with the E-strands of FnI2–FnI5 and FnI8–FnI9 modules by the β-zipper mechanism, and that interaction causes conformational change of Fn to block Fn assembly by breaking the FnIII3–FnI4 interaction (Maurer et al., 2010; Marjenberg et al., 2011). In addition, Maurer et al. reported that a ‘high-affinity downstream domain’ (HADD) was also active in blocking Fn assembly as well as the FUD. HADD is a 49-residue polypeptide similar to SfbI-5 with its adjacent non-repetitive sequence, and binds by β-strand addition to the N-terminal FnI modules FnI1–FnI5 (Maurer et al., 2012). Inactive Fn is generally assumed to be defined by intramolecular interactions involving the N-terminal domain of Fn. To interact with host cell integrin, the RGD motif in Fn must be accessible, which requires a conformational change in the inactive form of Fn. Binding of multiple copies of Fn to PrtF1/SfbI causes conformational change, and RGD motifs of Fn are exposed at regular intervals (Fig. 1B) and allow integrins to bind to Fn.
Using a cell-free binding assay, Chabria et al. found that bacterial peptides bind significantly less to stretched than relaxed Fn fibres, thus demonstrating mechano-regulation of a cell binding site on the N-terminus of Fn (Chabria et al., 2010). FnI modules contain many other physiologically significant binding sites, such as heparin, collagen, tenascin and fibrin (Fig. 1A), which could also potentially be subject to mechano-regulation. That first demonstration of a mechano-regulated binding site raised intriguing questions regarding whether bacteria can distinguish healthy tissue from wound sites by sensing matrix tension that exerts an effect on Fn fibres and the authors speculated that this could regulate early adhesion events. The finding that the specific binding of adhesins may be regulated by the tension of ECM fibres provides a unique and new perspective on how the mechano-biology of ECM might regulate early bacterial adhesion and the subsequent course of infection (Chabria et al., 2010).
Signal transduction by Fn-binding proteins
Engagement of Fn and bacterial Fn-binding protein causes S. pyogenes adhesion to host cells, which is followed by cellular invasion. Two different morphological patterns of invasion by Fn-binding protein have been reported, one mechanism that features M1-dependent zipper-like uptake and the other PrtF1/SfbI-dependent caveolae endocytosis (Molinari et al., 2000; Wang et al., 2007). M1 and PrtF1/SfbI proteins bridge bacteria to integrins and active cellular signalling for ingestion. It has been suggested that those interactions activate signalling pathways that include phosphoinotide 3-kinase (PI3K), integrin-linked kinase (ILK), paxillin and focal adhesion kinase (Wang et al., 2006b). However, distinct cellular morphological changes during bacterial invasion suggest that different signals are induced. Wang et al. showed that paxillin phosphorylation is induced by both proteins, but only required for M1-mediated invasion (Wang et al., 2007). PrtF1/SfbI activates PI3K/ILK-dependent signalling to caveolin-1 phosphorylation by Src, then Src leads to upregulation of caveolae endocytosis, which is dependent on lipid rafts. These processes suggest that a bifurcation point, downstream of ILK and PI3K, accounts for the distinct morphological changes. In addition, M1 protein binds CD46, which mediates S. pyogenes invasion and is Fn-dependent (Rezcallah et al., 2005). In experiments with the type II pulmonary epithelial cell line A549, deletion of the CD46 cytoplasmic domain resulted in partial inhibition of invasion. CD46 interacts with other focal adhesion proteins to generate signals required for invasion of the epithelial cells by S. pyogenes (Rezcallah et al., 2005). In β1 null cells, the αvβ3 integrin promotes PrtF1/SfbI-mediated internalization of S. pyogenes (Ozeri et al., 1998). These interactions may activate different signal pathways and induce different morphological patterns. Interestingly, Rohde et al. showed that differences in the aromatic domain between PrtF1/SfbI and GfbA, the PrtF1/SfbI homologue of Group G streptococci, are responsible for triggering morphologically distinct invasion pathways (Rohde et al., 2011).
In contrast to those reports, Klenk et al. showed that the Ras superfamily and RhoA pathways are exploited in serotype M49 S. pyogenes-invaded HEp-2 cells, suggesting serotype-specific interactions with the host cell cytoskeleton (Klenk et al., 2005). In addition, Nerlich et al. reported that M3 S. pyogenes, which lacks M1 protein and PrtF1/SfbI, invades human endothelial cells in a PI3K-independent manner (Nerlich et al., 2009). M3 S. pyogenes invasion induces Src kinases and Rac1 activation, and Rac-1 and the actin nucleation complex Arp2/3 are accumulated with actin at bacterial entry sites (Nerlich et al., 2009). Then, TGF-β1 enhances the interaction of host cells in a manner dependent on Fn-binding proteins of S. pyogenes via upregulation of α5β1 integrin and Fn-expression by epithelial cells (Wang et al., 2006a). Also, S. pyogenes-induced production of TGF-β1 by primary tonsil fibroblasts has been reported (Wang et al., 2006a).
Evasion of host innate immunity through Fn-binding proteins
When S. pyogenes organisms spread into blood vessels and deeper organs, they must evade the host innate immune system that functions as the initial protective barrier against bacterial pathogens. Several Fn-binding proteins, such as M1 protein, PrtF1/SfbI, FbaA, Scl1 and GAPDH, allow S. pyogenes to evade phagocytosis by inhibiting complement activity independently of Fn-binding activity (Fig. 2). The complement system plays a major role in innate immunity and can be activated through classical, alternative and lectin pathways. These pathways share a common terminal pathway, formation of the membrane attack complex, which directly lyses Gram-negative bacteria. In addition, C3b labels pathogens and facilitates phagocytosis, while C5a attracts professional phagocytes as a chemoattractant (Laarman et al., 2010). M1 protein, FbaA and Scl1-binding factor H are regulators of the complement system and inhibit C3b deposition (Pandiripally et al., 2002; 2003; Reuter et al., 2010; Ma et al., 2011). Notably, M1 protein and FbaA bind factor H and factor H-like protein 1 via the short consensus repeat 7 (SCR7) of those proteins, but do not bind complement factor H-related protein 1 (CFHR1). In contrast, Scl1 binds factor H and CFHR1 via the conserved C-terminal region, factor H SCR18-20 and CFHR SCR3-5, and do not bind factor H-like protein 1 or other members of the factor H protein family (CFHR2, CFHR3 or CFHR4A) (Reuter et al., 2010). M1 protein can also bind fibrinogen, which inhibits the alternative complement pathway. It has also been shown that PrtF1/SfbI restricts C3 deposition on M1-mutant S. pyogenes (Hyland et al., 2007). GAPDH binds C5a and the supporting streptococcal C5a peptidase degrades C5a. As a result, GAPDH inhibits neutrophil infiltration and reduces C5a-induced H2O2 production by neutrophils (Terao et al., 2006).
Fn-binding proteins as vaccine antigens and genotypic markers
Historically, S. pyogenes vaccinology has focused on the major virulence factor M protein (Steer et al., 2009). Developments have led to successful phase I/II clinical trials of a multivalent vaccine containing N-terminal fragments of 26 different M proteins (McNeil et al., 2005). However, there are at least 150 emm types of S. pyogenes and the dominant serotype may easily change (Steer et al., 2009; Safar et al., 2011). Fn-binding proteins have received focus as non-M protein vaccine candidates, because of their localization and conservation among different M serotypes. Furthermore, FBP54, FbaA, PrtF1/SfbI and SOF have been investigated in animal studies (Table 2).
Table 2. Fn-BPs reported as vaccine candidates tested in mice
The deduced amino acid sequence of FBP54 is conserved among different M serotype strains. Mice immunized with FBP54 subcutaneously, orally or nasally survived longer following S. pyogenes challenge than non-immunized mice (Kawabata et al., 2001). Those data suggest that FBP54 may be a universal vaccine candidate against all serotypes of S. pyogenes. In contrast, FbaA immunization specifically protected mice from infection with M1 S. pyogenes, which is highly virulent and invasive (Terao et al., 2005). It was also reported that a monoclonal antibody against FbaA inhibited the binding of factor H to S. pyogenes (Ma et al., 2011) and mice vaccinated with PrtF1/SfbI were protected from intranasal S. pyogenes infections (Guzman et al., 1999; Schulze et al., 2001; 2006b; Olive et al., 2007). In addition, there was no significant difference in the degree of protection between mice challenged at 36 days and those challenged at 110 days after primary vaccination with PrtF1/SfbI (Schulze et al., 2003). However, that vaccination did not prevent subcutaneous S. pyogenes infections. To protect from a skin infection predisposed to develop into a severe S. pyogenes infection, a PrtF1/SfbI-based vaccine should be complemented with an additional antigen (McArthur et al., 2004). In another report, SOF vaccination prevented intraperitoneal but not intranasal S. pyogenes infection (Courtney et al., 2003; Schulze et al., 2006a). Together, these findings suggest that FbaA and PrtF1/SfbI could be useful as vaccine antigens against highly virulent and invasive S. pyogenes strains.
Streptococcus pyogenes has at least 11 Fn-binding proteins, which are distributed in an M-serotype-dependent manner and may be correlated with clinical conditions. In studies conducted in several countries, Fn-binding proteins have been suggested to be genotypic markers and the prtF1/sfbI gene has been reported to be present significantly more often among macrolide-resistant strains isolated in Germany, Italy and Japan (Haller et al., 2005; Baldassarri et al., 2007; Hotomi et al., 2009). In addition, the prevalence of the prtF1/sfbI gene was found to be associated with non-invasive S. pyogenes disease in Japan and Romania, but not in Italy or Australia (Delvecchio et al., 2002; Gorton et al., 2005; Baldassarri et al., 2007; Luca-Harari et al., 2008; Hotomi et al., 2009). Interestingly, Nyberg et al. reported that the prtF1 gene-introduced S. pyogenes strain AP1 lacking this gene had increased adherence to and invasion of epithelial cells, but became less virulent in normal mice, with virulence partly restored when the bacteria were used to infect mice lacking plasma Fn. PrtF1/SfbI also reduced bacterial spreading in mice by interacting with Fn, with the reduction dependent on the presence of the Fn-binding domain of PrtF1/SfbI (Nyberg et al., 2004). In contrast, Cho and Mosher (2006) reported that the effect on formation of stable thrombin by perfused Fn was lost if Fn deposition was blocked by co-perfusion with the FUD of PrtF1/SfbI. Furthermore, the prtF2/pfbp gene was reported to be associated with invasive S. pyogenes disease cases in Australia and the Netherland (Vlaminckx et al., 2007), whereas no such reports have been made in Italy or Japan. The prtF2/pfbp gene is designated as both the pfbp type and fbaB type. We reported that the FbaB protein is found exclusively on bacterial surface of M3 and M18 strains of S. pyogenes isolated from toxic shock-like syndrome patients, but not those isolated from pharyngitis patients (Terao et al., 2002). There is a possibility that the designation of prtF2/pfbp should be changed to pfbp and FbaB be used as an epidemiological marker. However, any conclusion regarding the role of Fn-binding protein genes as epidemiological markers will require additional worldwide findings.
Streptococcus pyogenes has a more compact genome as compared with other pathogenic bacteria, such as Escherichia coli or Staphylococcus aureus, although S. pyogenes infections lead to a variety of clinical conditions. One possibility for this phenomenon is that S. pyogenes utilizes host molecules and systems, and possesses a number of multiple role proteins, such as Fn-binding proteins. It is interesting that S. pyogenes organisms contain several kinds of Fn-binding proteins, even though the genome is compact. In addition, those proteins appear to have acquired the same functions through convergent evolution, based on the low protein sequence similarity seen among some Fn-binding proteins. These findings indirectly support the notion that Fn-binding proteins are important for development of S. pyogenes.
Recently, molecular interaction and structural analyses have been used to investigate the detailed mechanisms of PrtF1/SfbI and Fn. To prevent pathogenic bacteria from initiating local inflammation or becoming invasive, it is important to understand the mechanisms underlying bacterial adhesion to and invasion of host cells. However, the detailed molecular interactions and signal transduction pathways in S. pyogenes Fn-binding proteins without PrtF1/SfbI remain unclear. We expect that additional basic and translational research studies will be employed to investigate the S. pyogenes infectious mechanism, and provide improved strategies to prevent S. pyogenes infection in the near future.