vhp Is a Fibrinogen-Binding Protein Related to vWbp in Staphylococcus aureus

ABSTRACT Staphylococcus aureus can target a variety of tissues, causing life-threatening infections. The basis for this diversity stems from the microorganism’s ability to spread in the vascular system throughout the body. To survive in blood, S. aureus coats itself with a fibrinogen (Fg)/fibrin shield. The protective shield is assembled by the coordinated actions of a number of Fg-binding bacterial proteins that manipulate the host’s blood coagulation system. Several of the Fg binders appear redundant, sharing similar functional motifs. This observation led us to screen for the presence of novel proteins with significant amino acid identities to von Willebrand factor-binding protein (vWbp), a key component in the shield assembly machinery. One identified protein showed significant sequence identity with the C-terminal region of vWbp, and we consequently named it vWbp homologous protein (vhp). The vhp gene lies within a cluster of genes that encode other virulence factors in S. aureus. Although each isolate only contains one copy of the vhp gene, S. aureus has at least three distinct alleles, vhpA, B, and C, that are present in the core genome. All three vhp isoforms bind Fg with high affinity, targeting a site located in the D fragment of Fg. We further identified an ∼79 amino acid-long, conserved segment within the C-terminal region of vWbp that shares high sequence identities (54 to 67%) with the vhps and binds soluble Fg with high affinity. Further analysis of this conserved motif and the intact vhps revealed intriguing differences in the Fg binding behavior, perhaps suggesting that these proteins have similar but discrete functions in the shield assembly.

The vhpA, B, and C proteins are smaller than the 508 amino acid residue-long vWbp and are 173, 156, and 185 amino acid residues long, respectively ( Fig. 1A and Fig. S1A and B). The vhps lack the two D domains found in the N terminus of staphylocoagulase (Coa) and vWbp, the characteristic first two N-terminal amino acids Val-Val of vWbp (33), and the Ile-Val of Coa (34) required for prothrombin activation. Therefore, vhps likely do not act as coagulases. The vhps also lack the 26 amino acid-long vWF-binding motif that is present in vWbp ( Fig. 1 and Fig. S1A) (35).
The vhp gene is located in a gene cluster encoding known virulence factors. We located the gene encoding vhpA protein in the S. aureus strain N315 genome to a gene cluster that encodes known virulence factors. The gene cluster consists of clfA, vWb, emp, vhp, and nuc genes ( Fig. 2) (35)(36)(37). This location is also observed for the vhpB gene from strain USA300_FPR3757 and the vhpC gene of strain TCH60 (Fig. 2). Thus, all three vhp genes are encoded in the same genetic location in different S. aureus strains, demonstrating that vhpA, B, and C proteins are isoforms of each other.
vhp is present in S. aureus as distinct isoform groups. We analyzed the genome sequences of 30 published clinical isolates with the goal to determine the genetic variation of vhp in relation to vWb. The isolates were obtained from different types of S. aureus infections from various geographical regions and represent a variety of sequence types (ST) ( Table 1). Since the amino acid sequence differences between vhpA, B, and C were substantial, we referred to these sequences as prototypes of specific isotype groups (Table 1). We chose strain N315 as the prototype for vhpA, USA300_FPR3757 for vhpB, and TCH60 for vhpC, as these are also well-established clinical isolates. We assigned vhp isoforms to a particular group if they had an amino acid identity of 85 to 100% compared to the prototype of their respective isotype group (Table 1).
A vhp gene is present in all isolates examined, and a majority (26 of 30) of the isolates examined belong to one of the identified isotype groups (vhpA, B, or C) ( Table 1). Four of the 30 isolates, which included strains RF122, 08-02300, BB155, and 71193, do not fall into either group, suggesting the existence of additional isoforms besides vhpA, B, and C. The amino acid sequence from strain 71193 appears as a hybrid of vhpB and vhpC, in which the N-terminal region shows a relatively high similarity to the N-terminal region of vhpB, while its C-terminal region is nearly identical to the C-terminal segment of vhpC (Fig. S1A). Although the N-terminal amino acid sequence for both vhpAs of strains RF122 and 08-02300 are identical to vhpA, the C-terminal regions are identical to that of vhpB (Fig. S1A). Genome analysis of the strains revealed that only one vhp gene exists in each isolate.
We further examined a limited data set composed of 41 S. aureus ST5 isolates, 66 ST8 isolates, and 26 ST30 isolates. We found that all ST5 isolates contained vhpA, all ST30 isolates contained vhpC, and 66 of 68 ST8 isolates contained the vhpB isoform (Table S2).
We also analyzed the vWb gene to determine if a particular vWbp isoform is associated with any of the discrete vhp isoforms. Analysis of the 30 prototypic strains revealed at least 5 isoforms, which we termed as vWbpA, B, C, D, or E, respectively, based on an amino acid identity of 80 to 100% (Table 2; Table S3; Fig. S1B). The isolates also appear to carry only one vwb gene (Table 2) (35). We then examined whether there was a correlation between the vhp and vWbp isoforms ( Table 2). It is noteworthy that the vWbpC isoforms are present in the strains that also harbor the vhpA isoforms, and the vWbpD isoforms are found in the strains that have vhpB (Table 2). However, vWbpA, B, and E isoforms do not appear to be present in isolates harboring specific vhp isoforms. Examining a larger set of isolates would likely reveal additional isoforms of both vhp and vWbp. Based on our results, it is possible that the two S. aureus proteins vhp and vWbp have evolved independently of each other.
Recombinant vhpA, B, and C bind Fg with high affinity. The C-terminal region of vWbp contains an Fg-binding site (27,29). We therefore examined the Fg binding activity of the three major vhp isoforms using enzyme-linked immunosorbent assay (ELISA)-type binding assays ( Fig. 3A and B) (27). We expressed and purified recombinant full-length vhpA, B, and C, each with a His 6 -maltose-binding protein (MBP) tag fused to the N terminus (Fig. S2A). All three isoforms, when coated in microtiter wells, showed dose-dependent binding to soluble Fg with interactions that exhibited saturation kinetics. The calculated apparent K D values (concentration required for half-maximum binding) were similar (29 nM, 41 nM, and 87 nM for vhpA, B, and C, respectively [ Fig. 3A and Table 3]).
When tested for binding to Fg coated on the ELISA plates, the vhp isoforms also bound to immobilized Fg in a dose-dependent manner with apparent K D values of 67 nM, 50 nM, and 100 nM, respectively ( Fig. 3B and Table 3). As a negative control, a recombinant His 6 -MBP fusion protein, purified by the same method used for purifying the recombinant vhps, showed no binding to soluble or immobilized Fg ( Fig. S2B and C). Together, our data show that all three isoforms specifically bind to both soluble and immobilized Fg with similar high affinities.
The vhp isoforms bind to the Fg D fragment. To further characterize the interaction between Fg and vhp, we sought to identify the segment in Fg that binds to the   Thomas et al. vhps (Fig. 4A to C). Fg was digested with the fibrinolytic enzyme plasmin to obtain two lateral globular D fragments (Fg-D) and a central E fragment (Fg-E) (7-13). We coated microtiter wells with full-length Fg, Fg-D, or Fg-E and examined the dose-dependent binding of the vhp isoforms to the isolated Fg fragments (Fig. 4A to C). The vhp isoforms bound to immobilized Fg-D in a dose-dependent manner that showed saturation kinetics and showed no binding to immobilized Fg-E. Binding of vhp isoforms to full-length Fg was used as a positive control ( Table 3). The calculated apparent K D values for vhpA, B, and C binding to the Fg-D fragment were 810 nM, 272 nM, and 244 nM, respectively ( Table 3). The difference in apparent affinities of vhpA compared to vhpB or C could be due to the stability of the protein and the sequence variation noted among the proteins ( Fig. 1A and Table 3). However, the apparent K D values for the interaction between the Fg-D fragment and the vhp isoforms was around 10-fold higher than binding to full-length Fg, indicating that the Fg D fragment contains a partial binding site for the vhp proteins.
The Fg-binding motif of vWbp-C encompasses residues 386 to 482. We observed that amino acid residues ;386 to 482 of the C-terminal half of vWbp share high sequence identity with the vhp isoforms. Considering that both vWbp-C and vhp bind Fg, we speculated that amino acids 386 to 482 of vWbp might be responsible for Fg binding activity observed in the C-terminal half of the protein. We therefore expressed and purified the C-terminal region of vWbp, covering amino acid region 250 to 482 (vWbp-C) (27), and its two truncated versions, 250 to 386 (vWbp-C [250 to 386] ) and 386 to 482 (vWbp-C [386 to 482] ) ( Fig. 5A and Fig. S2A). Using the ELISA-type binding assay, we determined that both vWbp-C and vWbp-C (386 to 482) bind soluble Fg in a dose-dependent process that exhibited saturation kinetics ( Fig. 5B; apparent K D of 1.5 nM and 2.4 nM, respectively). Consistent with our earlier published data, vWbp-C showed weak binding to immobilized Fg ( Fig. 5C and Table 3) (27). As expected, vWbp-C (386 to 482) also showed little to no binding to immobilized Fg (Fig. 5C). Lastly, vWbp-C (250 to 386) did not bind to either form of Fg under the experimental conditions used (Fig. S2D and E). His 6 -MBP fusion protein was used as a negative control and showed no binding  to both soluble and immobilized Fg ( Fig. S2D and E). We concluded that the Fg-binding motif for the C-terminal half of vWbp is located at amino acid residues 386 to 482. Furthermore, this motif binds well to soluble Fg but not immobilized Fg, suggesting that vWbp binding to Fg requires a specific structural conformation of Fg that may not be available when Fg is immobilized in the microtiter well.

DISCUSSION
In this study, we identified a unique protein, vhp, which shows significant sequence identity to the Fg-binding C-terminal section of vWbp ( Fig. 1A and B). S. aureus expresses multiple Fg-binding proteins, most of which are classified as MSCRAMMs (3,17,18) or SERAMs (3,6,17,(19)(20)(21). With the exception of vWbp, vhp does not share significant protein sequence identity or a notable motif with other Fg-binding proteins in S. aureus. Analysis of publicly available genomes of S. aureus clinical isolates revealed the presence of three distinct vhp isoforms A, B, and C. We identified Fg as a binding partner for all three isoforms. An initial characterization of this interaction showed that the vhps bind Fg with high affinity. The Fg D fragment, but not the E fragment, bound to the vhps, an observation that locates a vhp interactive site in Fg.  The sequence in the N-terminal half of the vhps shows a high degree of sequence variation between the isoforms. However, within each isoform group, the N-terminal region is conserved across different isolates (Fig. S1A in the supplemental material). This suggests that the N-terminal region of the isoforms might have different functions, perhaps interacting with different targets. The sequences of the C-terminal half of the vhps are conserved among isoforms of isolates and are highly homologous to residues 386 to 482 of vWbp. This vWbp segment was shown to contain a Fg-binding site. Since the prototypes of each vhp isoform behave similarly in their binding to Fg, it seems reasonable to conclude that Fg binding activity is located to the C-terminal region of all vhp isoforms.
Despite the similarities between vWbp and vhp, there are some significant differences in Fg binding between the two staphylococcal proteins. The apparent K D of vWbp (386 to 482) for binding to soluble Fg is about 10-fold lower than that shown for the binding of the vhp isoforms to Fg ( Table 3). The vhp proteins bind the Fg D fragment with an apparent K D of 10 27 M, whereas vWbp (386 to 482) does not show any significant binding to the Fg D fragment under similar experimental conditions (Table 3) (data not shown). The vhps bind to Fg in the ELISA-type assay when Fg is adsorbed onto the microtiter plate and vhps are added in solution and when the assay is reversed such that the vhps are adsorbed on the plate and Fg is added in solution (Fig. 3). On the other hand, vWbp (386 to 482) only binds to Fg when Fg is in solution. Soluble vWbp (386 to 482) does not show significant binding to Fg adsorbed on the plate (Fig. 5). These observations could indicate that the vhps and vWbp (386 to 482) bind Fg by different mechanisms. An alternative explanation is based on earlier studies by us and others, suggesting that the Fg binding of the C-terminal section of vWbp involves significant conformational arrangements in the two proteins (10,27,33,39). This type of structural rearrangement could be inhibited when a protein is adsorbed in a microtiter plate in which protein flexibility may become restricted. To further define the Fg binding mechanism(s) in vhp and vWbp, the predicted conformational plasticity in the proteins needs to be examined.
vWbp along with Coa and extracellular fibrinogen-binding protein (Efb) appear to form a subfamily of functionally related secreted staphylococcal proteins where the members contribute to host defense evasion (3,6,20,33). These multidomain molecules share homologous domains among each other (40). Both Coa and Efb have similar Fg-binding motifs that map to the intrinsically disordered regions of the proteins (3,6). Coa and vWbp share a 30% amino acid identity at their N-terminal prothrombinbinding domains (30,33,34). They activate prothrombin by a similar mechanism, where the two N-terminal amino acid residues Ile 1 -Val 2 and Val 1 -Val 2 , respectively, are inserted into the Ile 16 pocket of prothrombin (33,34). The now described vhp also fits into this family, as its C-terminal domain is highly homologous to the C-terminal Fgbinding section of vWbp. In addition, the extracellular complement-binding protein (Ecb) is related to Efb and shows 33% sequence identity to the C-terminal C3-binding domain of Efb, covering amino acid residues 65 to 136 of Efb (41,42). Like vWb and vhp, the ecb and efb genes are located close to each other. Both efb and ecb genes are encoded in a mobile genetic element called IEC-2 (42).
The vhp gene is located within a gene cluster that includes genes coding for clumping factor A (ClfA), vWbp, extracellular matrix protein (Emp), and Nuc proteins (Fig. 2). The other proteins encoded in this cluster are identified as virulence factors that act by neutralizing the host defense system (1, 22,25,35,37,43). The location of vhp in the S. aureus genome raises a number of questions about the protein. Since all of the other proteins in this cluster are involved in evading host defenses, does vhp also participate in immune evasion? If yes, how does its interaction with Fg contribute to immune evasion? To what extent does vhp cooperate in its actions with the different proteins in this gene cluster? A possible functional cooperation among these proteins would require coordinated expression. As an example of this cooperation, ClfA and vWbp have been noted to interact with each other, resulting in enhanced adherence of was analyzed by SDS-PAGE and appeared as a single band with a molecular mass of 85 kDa. Human Fg E fragment (Fg-E) was purchased from Haematologic Technologies (HCI-0150E).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.