Molecular Characterization of the Multiple Interactions of SpsD, a Surface Protein from Staphylococcus pseudintermedius, with Host Extracellular Matrix Proteins

Staphylococcus pseudintermedius, a commensal and pathogen of dogs and occasionally of humans, expresses surface proteins potentially involved in host colonization and pathogenesis. Here, we describe the cloning and characterization of SpsD, a surface protein of S. pseudintermedius reported as interacting with extracellular matrix proteins and corneocytes. A ligand screen and Western immunoblotting revealed that the N-terminal A domain of SpsD bound fibrinogen, fibronectin, elastin and cytokeratin 10. SpsD also interfered with thrombin-induced fibrinogen coagulation and blocked ADP-induced platelet aggregation. The binding site for SpsD was mapped to residues 395–411 in the fibrinogen γ-chain, while binding sites in fibronectin were localized to the N- and C-terminal regions. SpsD also bound to glycine- and serine-rich omega loops within the C-terminal tail region of cytokeratin 10. Ligand binding studies using SpsD variants lacking the C-terminal segment or containing an amino-acid substitution in the putative ligand binding site provided insights into interaction mechanism of SpsD with the different ligands. Together these data demonstrate the multi-ligand binding properties of SpsD and illustrate some interesting differences in the variety of ligands bound by SpsD and related proteins from S. aureus.


Introduction
Staphylococcus pseudintermedius is a commensal and opportunistic pathogen of companion animals, especially dogs [1,2], mainly causing skin infections such as pyoderma as well as surgical wound infections, urinary tract infections and otitis externa. Cases of infections in humans have occasionally been reported [3][4][5][6]. Methicillin-resistant S. pseudintermedius occurs widely [7,8]. The complete genome sequences of two isolates of S. pseudintermedius are available [9,10]. The strains are predicted to encode many putative virulence factors including toxins, extracellular enzymes such as lipases and proteases and surface proteins designated S. pseudintermedius surface proteins A-R (SpsA-R) [11] some of which are known to promote adhesion of the bacterium to desquamated skin epithelial cells (corneocytes) [12][13][14] and to components of the extracellular matrix [11,15].
One such surface protein that is likely to be important in skin colonization and virulence is SpsD. The presence of SpsD on the bacterial cell surface promotes adhesion to fibrinogen (Fbg), fibronectin (Fn) and cytokeratin 10 (K10). Immunoglobulin G specific for SpsD occurs in dogs with pyoderma indicating that the protein is expressed during infection [11]. SpsD has many features that are typical of staphylococcal surface proteins called microbial surface components recognizing adhesive matrix molecules (MSCRAMMs) that are related to clumping factors (Clf) and fibronectin binding proteins (FnBPs) of S. aureus [11].
The primary translation product of SpsD has an N-terminal secretory signal sequence and a C-terminal sorting signal comprising an LPXTG motif for sortase A-mediated linkage to peptidoglycan, a hydrophobic membrane-spanning domain and positively charged residues. The N-terminus of the mature protein comprises an A domain, part of which has sequence similarity to the A domains of ClfA and FnBPs ( Figure 1).
The A domains of FnBPA and FnBPB show significant sequence similarity to ClfA and are predicted to fold into three independent subdomains, N1, N2 and N3. The ligand binding sites within ClfA, FnBPA and FnBPB are confined to the N2 and N3 subdomains and they each bind to the C-terminus of the cchain of Fbg most likely by the same ''dock, lock and latch'' mechanism [16][17][18]. X-ray crystal structure analysis of ClfA revealed that the apo form of the protein adopts an open conformation with an exposed hydrophobic trench located between separately folded N2 and N3 subdomains. Binding of the Fbg c-chain peptide is initiated by docking of the ligand peptide in the trench. A flexible extension of the N3 subdomain redirects towards N2, covers the trench-bound peptide and locks it in place by b-strand complementation with a b sheet in subdomain N2. Structural and mechanistic studies strongly suggest that the N2N3 subdomains of FnBPA and FnBPB bind to the Fbg c chain by the same mechanism. However differences exist in the repertoire of ligands bound in addition to Fbg, with ClfA recognizing complement factor I from human serum [19,20], FnBPA and FnBPB both additionally binding to elastin [17,21,22] and subdomains N2N3 of FnBPB also binding to fibronectin [23].
Clumping factor B is another MSCRAMM of S. aureus with an A domain that is similar in structure and function to ClfA but which binds different ligands to ClfA and FnBPs by the dock, lock and latch mechanism [24]. ClfB binds to the glycine and serinerich omega loops that occur in the C-terminal tail of cytokeratin 10 and throughout the corneocyte envelope protein loricrin [25,26]. It also binds to a related sequence in the aC region of the a chain of Fbg [24,27].
Located distally to the A domains of FnBPA and FnBPB is an extended unfolded region comprising 11 or 10 tandemly repeated domains, respectively, that bind to the N-terminal type I modules of Fn by the tandem b-zipper mechanism [28,29]. In ClfA and ClfB this region is occupied by multiple repeats of the dipeptide Ser-Asp which have no known ligand binding function [30].
SpsD has been reported to promote bacterial adhesion to Fbg, K10 and Fn. In this study we set out to dissect SpsD and to localize and characterize its ligand binding region(s). We identified a region that is most closely related to the A domain of FnBPB of S. The A domain modelled on the structure of ClfA spans residues 37-519 following the secretory signal sequence S. This is followed by a connecting region C (residues 521-867) and a repeat region R in strain ED99. The number of repeats varies from strain to strain resulting in proteins of slightly different sizes. A sorting signal (SS: LPXTG motif, hydrophobic domain and positively charged residues) occurs at the extreme C-terminus. The minimum ligand binding domain construct N2N3 spans residues 167-519. Residues comprising the 'lock' and 'latch' (grey box residues 490-499) are indicated (N23DLL). Residue F288 that is predicted to lie within the ligand binding trench is indicated by the dashed line. doi:10.1371/journal.pone.0066901.g001 Figure 2. Binding of SpsD proteins to extracellular matrix or plasma proteins. (A), ligands were coated onto the surface of microtiter plates (1 mg/well) and incubated with 17.5 nanomoles of SpsD 37-519 or SpsD 167-519 , followed by mouse anti-His IgG and HRP-conjugated rabbit anti-mouse IgG. A, absorbance. (B), purified, unreduced human Fbg, Fn, a-elastin, mouse K10, collagen type II, laminin and BSA were subjected to SDS-PAGE, transferred to nitrocellulose filter and then incubated with SpsD 167-519 , followed by incubation with mouse anti-SpsD 37-519 and HRP-conjugated rabbit anti-mouse antibodies. Molecular mass markers are indicated (kilodaltons). doi:10.1371/journal.pone.0066901.g002 aureus that bound to these ligands and we provide insights into the ligand binding mechanism.

Bacterial Strains and Growth Conditions
Escherichia coli strain TOPP3 (Stratagene, La Jolla, CA, USA) was used as host for expression of recombinant proteins. E. coli strain TOPP3 was grown in Luria broth (LB) containing ampicillin (100 mg/ml). S. pseudintermedius strain 326 was isolated from canine pyoderma [15]. Staphylococcal cells were grown for the indicated times at 37uC in BHI (Brain Heart Infusion) broth (BD, Sparks, MD, USA) with shaking.

Proteins
Fibronectin was purified from human plasma as previously reported [31]. The N-terminal fragment of Fn (N29) containing the five N-terminal type I modules, and the gelatin-binding domain (GBD) consisting of four type I modules and two type II modules were isolated following the protocol reported by Zardi et al [32]. The recombinant fragment corresponding to residues 607-1265 consisting of the first seven type III modules, the region including the amino acid residues 1266-1908 containing the extra domain B (EDB) plus type III modules 8-13 and the initial region of module 14, and C-terminal fragment 1913-2477 containing one type III module, the IIICS module, three type I modules and the site for interchain disulfide linkage of Fn were purchased from R&D Systems (Minneapolis, MN, USA). Human Fbg and aelastin, a heterogeneous mixture of soluble peptides obtained from human lung, were obtained from Calbiochem (Merck, Darmstadt, Germany). Fbg was further purified on a gelatin-Sepharose column to remove contaminating Fn. The a-elastin preparation has a molecular weight of 10-60 kDa and runs as a smear on SDS-PAGE.
Bovine, porcine, ovine and canine Fbg as well as bovine serum albumin (BSA) were from Sigma (St Louis, MO, USA). CultrexH murine sarcoma basement membrane laminin was purchased from Trevigen (Gaithersburg, MD, USA). The synthetic peptide corresponding to the C-terminus of the Fbg c chain (residues 395-411) and the scrambled peptide were a generous gift by Dr I. Margarit (Novartis Diagnostics & Vaccines, Siena, Italy).

SDS-PAGE and Western Blotting
SDS-PAGE was performed with a 10% polyacrylamide gel. The gels were stained with Coomassie Brilliant Blue (BioRad, Hercules, CA, USA). For the Western blot assay, human Fbg and Fn were subjected to SDS-PAGE and then electroblotted onto a nitrocellulose membrane (GE Healthcare). The membrane was treated with a solution containing 5% (w/v) dried milk in PBS, washed, and incubated with SpsD 167-519 (5 mg/ml) for 1 h at 22uC. Following additional washings with 0.5% (v/v) Tween 20 in PBS (PBST), the membrane was incubated for 1 h with PBST containing an anti-His tag mouse monoclonal antibody (7E8) (2 mg/ml). After several washings in PBST, the membrane was incubated for 45 min with HRP-conjugated rabbit anti-mouse IgG. Finally, the membrane was treated with ECL detection reagents 1 and 2 according to the procedure recommended by the manufacturer (GE Healthcare) and exposed to an X-ray film for 30-60 s.

Antibodies
Mouse polyclonal antisera against SpsD 37-519 , ClfB 44-542 and human Fbg were generated by injecting BALB/c mice intraperitoneally five times at 1-week intervals with 50 mg of the purified protein. The antigen was emulsified with an equal volume of complete Freund's adjuvant for the first immunization, followed by three injections in incomplete adjuvant. The mice were bled, and the sera were tested for reactivity to the purified antigen using ELISA and Western blot. The antibodies were purified by affinity chromatography on Protein A/G-Sepharose columns according to the recommendations of the manufacturer (GE Healthcare). Horseradish peroxidase (HRP)-conjugated rabbit anti-mouse antibody was from Dako (Glostrup, Denmark). 7E8 is an in-house generated monoclonal antibody recognizing recombinant His tagged proteins. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the Italian Ministry of Health according to the European Union guidelines for the handling of laboratory animals. The protocol was approved by the local authorithies (''Comitato Etico per la Sperimentazione Animale'' of the University of Pavia, Italy) (Permit Number: 2-2012). The animals were sacrified under lidocaine anesthesia, and all efforts were made to minimize suffering.

Binding of SpsD Proteins to Extracellular Matrix Proteins by Solid Phase Binding Assay
To screen binding of SpsD to extracellular matrix proteins, 1 mg of human Fbg, human Fn, human lung elastin, murine K10, bovine nasal septum type II collagen, murine sarcoma basement membrane laminin and BSA were coated onto microtiter wells overnight at 4uC in 0.1 M carbonate buffer. The wells were washed, blocked with 2% BSA for 1 h at 22uC and incubated with SpsD 167-519 or SpsD 37-519 (17.5 nanomoles/well). Complex formation was detected with mouse anti-His tag antibody 7E8 (0.2 mg/well), followed by addition of peroxidase-conjugated rabbit anti-mouse IgG (1:1000) (Dako). Binding of SpsD 167-519 to Fbg from different species in ELISA format was performed by incubating 1 mg of the recombinant bacterial protein with each surface-coated Fbg type (1 mg/well) and detecting the bound protein as reported above.

Surface Plasmon Resonance (SPR)
All experiments were performed at 25uC on a BIAcore X instrument (Pharmacia Biosensor AB, Uppsala, Sweden). Purified human and canine Fbg, a-elastin or human Fn were covalently immobilized on the dextran matrix sensor chip surface (CM5 chip) by using a protein solution (30 mg/ml in 50 mM sodium acetate buffer, pH 5.5 for human and canine Fbg or 50 mM sodium acetate buffer, pH 4.0 for human Fn) in a 1:1 dilution with Nhydroxy-succinimide and N-ethyl-N'-(3-dimethylaminopropyl)carbodiimide hydrochloride. The excess active groups on the dextran matrix were blocked with ethanolamine (1 M), pH 8.5 [38]. On another flow cell, the dextran matrix was treated as described above but without any ligand to provide an uncoated reference flow cell. The running buffer used was PBS containing 0.005% (v/v) Tween 20. Increasing concentrations of SpsD proteins were passed in succession over the ligand and reference surfaces without regeneration at the flow rate of 5 ml/min in running buffer.
To perform SPR experiments with GST-HK10 544-563 , goat anti-GST antibody IgG (30 mg/ml) (GE Healthcare) was diluted in 10 mM sodium acetate buffer, pH 5.0, and immobilized on carboxy-derivatized sensor chip. GST-HK10 544-563 (500 nM) was passed over a flow cell, while GST alone was passed over the other flow cell to provide a reference surface. Increasing concentrations of SpsD proteins were flowed over the surface of both flow cells without regeneration at a rate of 5 ml/min.
Sensorgrams from three sets of data for each of the protein concentrations were collected and averaged for each ligand. All of the sensorgram data presented were subtracted from the corresponding data from the reference flow cell. The data were analyzed using the BIAevaluation software version 3.0. A plot of the level of binding (response units) at equilibrium against concentration of ligand was used to determine the K D .

Inhibition of the SpsD-Fbg Interaction
A solution of Fbg (10 mg/ml) was used to coat microtiter wells overnight at 4uC. The wells were washed three times with PBST and blocked with 2% (w/v) BSA in PBS for 1 h at 22uC. The wells were washed and 100 ml SpsD proteins (10 mg/ml) added after preincubation for 1 h at 22uC with a range of concentrations of a synthetic peptide comprising the 17 C-terminal residues of the c chain of Fbg or a scrambled peptide. The plates were incubated for 1 h at 22uC and bound protein detected by incubation with anti-His tag mouse 7E8 monoclonal antibody (0.2 mg/well) for 1 h at 22uC, followed by horseradish peroxidase-conjugated rabbit anti-mouse IgG serum. The percentage inhibition was calculated relative to the level of Fbg-bound protein detected in the absence of the inhibitor peptide.

SpsD Binding to Immobilized Fn Fragments by ELISA and Dot Blot
Fn fragments were used to identify binding sites of SpsD. Fn (1 mg/well) and Fn fragments (250 ng/well) were immobilized onto microtiter wells and then incubated with indicated concentrations of SpsD 167-519 . Bound protein was detected using an anti-SpsD 37-519 mouse antibody followed by HRP-conjugated rabbit anti-mouse IgG. For dot blot experiments, different amounts of Fn and Fn fragments (1000-62.5 ng) were dotted onto a nitrocellulose membrane, probed with SpsD 167-519 (1 mg/ml) and complexes detected as described above.

Region A of SpsD
The sequence of the entire spsD gene of S. pseudintermedius ED99 is known. Here we describe the A domain of SpsD from strain 326 which was obtained following PCR amplification of genomic DNA. The A domains of SpsD from ED99 and 326 have the same number of residues and are 80% identical. They are most closely related to the A domain of FnBPB (40% and 42% identity to isoform I, respectively) [39] and they have 23% and 27% identity respectively to the A domain of ClfA from S. aureus Newman [40]. A 3D molecular model based on the crystal structure of ClfA allowed us to predict the coordinates of subdomains N1, N2 and N3 (Figure 1).

Binding of Recombinant SpsD to Extracellular Matrix Proteins
Previous work has demonstrated that SpsD can promote bacterial adhesion to Fbg, Fn and K10 [11]. Here we sought to determine if the ligand binding ability resides in the N-terminal domain of the protein. An ELISA type ligand-binding assay was used to study the interaction of the recombinant SpsD proteins with surface-coated extracellular matrix proteins. SpsD 37-519 and SpsD 167-519 bound to immobilized Fbg, Fn and K10, as well as to a-elastin, while no interaction was observed with collagen, laminin or BSA (Figure 2A). The interaction of Fbg, Fn, K10 and a-elastin with the recombinant SpsD 167-519 protein was confirmed by Western ligand blotting where host proteins were electroblotted onto a nitrocellulose filter, probed with SpsD 167-519 and complex formation detected with a monoclonal antibody targeting the Histag. Conversely, no interaction was observed when SpsD 167-519 was incubated with membrane-bound collagen type II, laminin or BSA ( Figure 2B). Identical results were obtained with SpsD 37-519 (data not shown). Together these results suggest that the minimal region of SpsD required for binding to Fbg, Fn, K10 and elastin is the N2N3 subdomain. The absence of signal when Fbg, Fn, a-  Next we determined the ability of SpsD 167-519 protein to recognize Fbg molecules obtained from different animals. Increasing amounts of SpsD 167-519 were added to surface-coated human, porcine, canine, bovine and ovine Fbg and the binding was measured. SpsD 167-519 bound to each of the different types of Fbg with similar affinities with exception of the ovine protein ( Figure 3).

SpsD Binds the C-terminus of the c-chain of Fbg
Western ligand blotting was employed to determine if SpsD bound to the a-, bor cchain of Fbg. Human Fbg was reduced with dithiothreitol and the three chains separated by SDS-PAGE ( Figure 4A). Efficient transfer to a control nitrocellulose membrane was shown by probing with anti-Fbg IgG ( Figure 4B). Another membrane was incubated with recombinant His-tagged SpsD 167-519 and when probed with a monoclonal antibody recognizing the His-tag only the c-chain reacted ( Figure 4C).
ClfA, FnBPA and FnBPB are known to bind to a short sequence at the extreme C-terminus of the c-chain of Fbg while ClfB binds to repeat number 5 of the a chain. To determine if SpsD bound the same c chain residues as ClfA, the ability of a 17 residue synthetic peptide comprising the C-terminal residues of the Fbg c- chain to inhibit binding of SpsD 167-519 to immobilized Fbg was tested. The peptide strongly inhibited the binding of SpsD to human and canine Fbg dose-dependently while a scrambled peptide had no effect ( Figure 4D and 4E).
Finally, direct binding of SpsD to the Fbg c-chain peptide coupled to GST that had been immobilized on the surface of an ELISA plate was tested. As shown in Figure 4F, SpsD bound dosedependently and saturably to GST-Fbgc 395-411 with a half maximum binding concentration of 0.3260.02 mM but not at all to the GST control. Thus, SpsD binds to the same C-terminal Fbg c-chain residues as ClfA and the FnBPs.

Region A of SpsD Interferes with Fibrin Clot Formation and Platelet Aggregation
Several adhesins not only interact with the target molecule in the host but also neutralize and even disrupt the biological function of the target. For example, ClfA, although it does not interfere with catalytic activity of thrombin, by binding to the Cterminus of c chain it affects fibrin clot formation by abrogating fibrin monomer polymerization and cross-linking of Fbg c chains [41]. Therefore, we examined the effect of SpsD on fibrin clot formation. Human Fbg was preincubated with different concentrations (1.25-10 mM) of SpsD 167-519 prior to the addition of thrombin. In these conditions SpsD inhibited coagulation as effectively as Bbp 270-599 , a recently identified Fbg-binding protein from S. aureus [33] (Figure 5A). Conversely, no effect was observed when Fbg was preincubated with 10 mM BSA. Consistent with the indication that SpsD recognizes the C-terminus of the c chain of canine Fbg, we also observed that preincubation of canine Fbg with 10 mM SpsD resulted in complete inhibition of clot formation.
It is well known that the C-terminus of Fbg c chain contains the platelet integrin GPIIb/IIIa recognition sequence (QAGDV 411 ) [42] and that a monoclonal antibody that recognizes the c chain C-terminus of Fbg exerts inhibitory effects on platelet aggregation [43]. Thus, SpsD was tested for its impact on ADP-induced platelet aggregation. As depicted in Figure 5B, ADP-stimulated platelet aggregation was inhibited in a dose-dependent manner by SpsD. In contrast, the collagen-binding domain of the S. aureus adhesin CNA (CNA 30-531 ), did not affect the process. Together these data demonstrate that SpsD could act as a modulator of thrombus formation.

Binding Sites in Fibronectin for SpsD
To locate the binding site(s) in Fn for the A domain of SpsD, the binding of SpsD 167-519 to the Fn N29 domain, the gelatin binding domain and the segments spanning residues 607-1265, 1266-1908 and 1913-2477 was tested. Native Fn and the purified fragments were immobilized onto microtiter wells, probed with SpsD 167-519 and complex formation detected with anti-SpsD IgG. A strong signal was observed when SpsD 167-519 was incubated with full length Fn, the N29 fragment and the segment spanning the residues 1266-1908 (Fn 1266-1908 ), while the other fragments failed to bind ( Figure 6A). A similar assay was performed with FnBPB 163-480 and confirmed its ability to bind to full length Fn and the N29 fragment and not to the C-terminal Fn 1266-1908 fragment [23]. The binding of SpsD to the two different regions of Fn was confirmed by incubating the Fn fragments with SpsD 167-519 in a dot blot assay ( Figure 6B). Finally, as shown in Figure 6C, binding of SpsD 167-519 to both the N29 and the Fn 1266-1908 fragment was concentration-dependent and saturable. The affinity for each domain appeared to be similar with half maximum binding concentrations of 2.2160.30 mM for N-terminal domain and 1.9760.3 mM for Fn  .

SpsD Binds to the K10 Tail Region
ClfB binds to the C-terminal tail region of K10 [25]. To determine if SpsD binds to the same domain, recombinant proteins comprising full length murine K10 along with the Nterminal head and C-terminal tail regions, were immobilized into microtitre wells and tested for their ability to support SpsD binding. SpsD 167-519 bound to full length K10 and the tail region dose-dependently and saturably and with half maximal binding concentrations of 1.0260.25 and 1.260.1 mM, respectively ( Figure 7A).
ClfB is known to bind to glycine (G)-serine (S)-rich domains within the omega loop regions in the C-terminal tail of K10, in loricrin and in the a-chain of Fbg [26]. To determine if SpsD could bind to the same sequences, recombinant SpsD 167-519 was tested in solid phase binding assays with GST-HK10 544-563 . The SpsD protein bound dose-dependently and saturably but failed to bind to the GST ( Figure 7B). This interaction was confirmed by Western blot ( Figure 7C). However, in contrast to ClfB, SpsD did not interact with Fbg a chain ( Figure S1). Consistent with this, GST-Fbga 316-367 inhibited binding of ClfB, but not that of SpsD 167-519 , while GST-Fbgc 395-411 blocked the interaction of SpsD 167-519 but not that of ClfB with Fbg ( Figure S2). Thus SpsD bound the same GS-rich omega loop in K10 as ClfB but surprisingly failed to recognize the Fbg a chain.

Measurements of Dissociation Constants of SpsD for Host Ligands by Surface Plasmon Resonance
Binding of SpsD 167-519 to human Fbg, Fn, a-elastin or GST-HK10 544-563 immobilized on a sensor chip was analyzed by SPR using BIAcore X. Increasing concentrations of SpsD 167-519 were passed over the surface of each ligand-coated chip: in these conditions SpsD 167-519 bound to the ligand in a dose dependentmanner. From analysis of the equilibrium binding data, the dissociation constant (K D ) for the interaction between SpsD 167-519 and Fbg was determined to be 0.36060.032 mM ( Figure 8A and Table 1). SpsD 167-519 showed a dissociation constant value for canine Fbg comparable to that reported for human Fbg (Table 1).
SpsD 167-519 bound Fn with a K D of 2.1660.70 mM, indicating that the affinity of the recombinant SpsD protein for Fn was lower than that exhibited for Fbg ( Figure 8B and Table 1). The K D values of SpsD 167-519 for a-elastin and GST-HK10 544-563 were 1.0260.15 and 0.8460.032 mM, respectively (Figure 8, panels C and D and Table1). Very similar K D values were obtained when SpsD 37-519 was passed over sensor chips coated with these proteins (Table 1).

Ligand Binding by SpsD 167-519 Mutants
One goal of this work was to determine whether the N2N3 subdomains of SpsD bind Fbg by the same mechanism as the A domains of ClfA and FnBPs. Our previous work demonstrated that the C-terminal residues of the N3 subdomain of FnBPB were crucial for its interaction with Fbg [23]. As SpsD interacts with the Fbg c-chain with a high affinity, it was proposed that the Cterminal 20 residues of the N3 subdomain of SpsD form the locking and latching segment and play a similar role in the interaction of SpsD with Fbg. To test this, a recombinant truncate of the SpsD 167-519 protein, which lacked the predicted locking and latching segment (SpsD 167-498 : DLL, Figure 1), was expressed and its ability to bind to immobilized Fbg was analysed by ELISA. No detectable interaction was observed when increasing concentrations of SpsD 167-499 were incubated with Fbg-coated microtiter wells ( Figure 9A), while a small but remarkable reduction in elastin binding was noticed ( Figure 9C). In contrast, no reduction in binding was observed when increasing amounts of SpsD 167-499 were incubated with surface-coated Fn or K10 ( Figure, 9 panels B and D). To investigate further if SpsD binds to the ligands at same or different sites, an amino acid substitution was introduced in SpsD 167-519 at residue F288. F288 is predicted to be located within the ligand binding trench and is in an equivalent position to the crucial residues F314 of FnBPB and Y338 of ClfA. SpsD F288A binding to Fbg and elastin was greatly reduced showing that F288 is indeed important for ligand binding and supports the notion that both Fbg and elastin bind to the same or overlapping regions of SpsD. Conversely, a small reduction was observed when SpsD F288A was tested for binding to Fn or K10.

Discussion
The ability of S. aureus cells to adhere to components of the extracellular matrix is promoted by several distinct surface proteins called MSCRAMMs. The mechanistic basis of many of the interactions has been very well characterized by structural analysis and by probing function with appropriate variant mutant proteins. By molecular modelling and mutagenesis it is possible to infer/predict that proteins with similar structural organization bind their ligands by the same mechanism as those that have been investigated in detail.
The A domain of ClfA is the archetype of a family of proteins that fold into three subdomains called N1, N2 and N3, the last two of which are the absolute minimum required for ligand binding by the ''dock, latch and lock (DLL)'' mechanism [18]. The A domain of ClfA binds to the c-chain of Fbg by DLL and it does not interact with Fn, elastin or K10 [44][45][46]. In contrast, the A domain of ClfB binds by DLL to the a-chain of Fbg and also to K10. In both cases the DLL mechanism was shown by X-ray crystal structure and supported by the loss of binding by mutants that had substitutions in the ligand binding trench (trench mutants) or which lack the C-terminal lock-latch domain (LL mutants) [24].
The Fn binding proteins FnBPA and FnBPB are composed of two distinct ligand binding domains. Their A domains are most closely related to ClfA and bind to the Fbg c-chain [16,23]. In contrast to ClfA, both FnBPA and FnBPB also bind to elastin [21] and in the case of FnBPB, Fn [23]. Binding to Fbg and elastin most likely occurs by DLL as indicated by studies with mutant proteins [17]. Ten or 11 tandemly repeated domains at the Ctermini of FnBPA and FnBPB bind to Fn with high affinity by the tandem b zipper mechanism [29].
The SpsD protein of S. pseudintermedius was previously shown to promote bacterial adhesion to Fbg, Fn and K10 [11,15]. In this study we have dissected the SpsD protein and have investigated how a single protein can bind to such a disparate array of ligands. Bioinformatic analysis predicted that the N-terminal region folded into a typical A domain composed of subdomains N1 N2 and N3 and was most closely related in sequence identity to FnBPB [11]. Although there is a repeat domain at the C-terminus of SpsD, the individual repeats are shorter than, and bear no similarity to, the Fn-binding repeats of FnBPs [11].
The recombinant A domain of SpsD was shown here to bind to the c-chain of Fbg (and not to the a-chain), to K10, to Fn and also to elastin. There are intriguing similarities to and differences from the MSCRAMM A domains of S. aureus.
To investigate if the DLL mechanism is involved in ligand binding, DLL and trench mutants of the N2N3 construct were expressed and purified. Binding to Fbg c-chain was completely abrogated in the SpsD DLL and trench mutants indicating that binding to this ligand occurred by DLL. However SpsD also bound to K10 which is a defining property of ClfB, while ClfB binds to the a-chain and not to the c-chain of Fbg [27]. The ligand peptide sequences that bind to the trench are completely dissimilar. Indeed structural biology showed that the Fbg a-chain peptide and the K10 omega loop peptide bound to the same part of the ligand binding trench of ClfB and a consensus motif was identified [24]. Therefore it was intriguing to discover that the DLL and trench mutants of SpsD bound to K10 with the same affinity as the wild-type while similar mutants of ClfB were previously shown to be defective in binding to both ligands [24]. We can only conclude that SpsD binding to K10 does not occur by DLL and propose a distinct K10 binding site in N2N3.
Like its closest relative FnBPB, SpsD also binds to elastin. This most likely occurs by DLL because the trench and the DLL mutants were both defective in binding to this ligand. Although a similar investigation was not carried out with FnBPB, detailed studies with the A domain of FnBPA supported the DLL mechanism for elastin binding [17]. This confirms the close relationship of the A domain of SpsD to those of the FnBPs.
It was not clear if the binding of FnBPB A domain to Fn occured by DLL because the DLL truncate bound Fn with the same affinity as the wild-type while a trench mutant was defective in both Fbg and Fn binding. It seemed that the Fn binds to FnBPB in the trench region but that lock latch aspect of DLL is not involved [23]. The SpsD A domain shares with FnBPB the ability to bind Fn but with the difference that it bound to two distinct domains within the host protein while FnBPB only recognized the N29 fragment [23]. This might explain why both the SpsD DLL and trench mutants bound Fn while the DLL mutant of FnBPB was defective.
A domains can be stitched onto distinct ''stalks'' (B repeats, Sdr repeats, Fn-binding repeats, or to a complete different sequence in SpsD). The DLL binding mechanism in N2N3 regions has evolved to binds to different ligands. Indeed even the orientation of a peptide ligand can vary [24]. Many have acquired the capacity to bind ligands by mechanisms other than DLL although the binding of a large folded ligand might occlude the trench region and block DLL mechanism of binding to other ligands.
In conclusion, SpsD appears to be the most promiscuous staphylococcal adhesin so far identified and possibly is a factor involved in S. pseudintermedius colonization of the host. In future studies the three-dimensional structure of the apo-protein and ligand bound SpsD may provide molecular details on these multiple interactions and give additional insights to better understand the role played by this microganism in human and animal pathogenesis.