Crystal structures of Bbp from Staphylococcus aureus reveal the ligand binding mechanism with Fibrinogen α

Bone sialoprotein-binding protein (Bbp), a MSCRAMMs (Microbial Surface Components Recognizing Adhesive Matrix Molecules) family protein expressed on the surface of Staphylococcus aureus (S. aureus), mediates adherence to fibrinogen α (Fg α), a component in the extracellular matrix of the host cell and is important for infection and pathogenesis. In this study, we solved the crystal structures of apo-Bbp273−598 and Bbp273−598-Fg α561−575 complex at a resolution of 2.03 Å and 1.45 Å, respectively. Apo-Bbp273−598 contained the ligand binding region N2 and N3 domains, both of which followed a DE variant IgG fold characterized by an additional D1 strand in N2 domain and D1′ and D2′ strands in N3 domain. The peptide mapped to the Fg α561−575 bond to Bbp273−598 on the open groove between the N2 and N3 domains. Strikingly, the disordered C-terminus in the apo-form reorganized into a highly-ordered loop and a β-strand G′′ covering the ligand upon ligand binding. BbpAla298–Gly301 in the N2 domain of the Bbp273−598-Fg α561−575 complex, which is a loop in the apo-form, formed a short α-helix to interact tightly with the peptide. In addition, BbpSer547–Gln561 in the N3 domain moved toward the binding groove to make contact directly with the peptide, while BbpAsp338–Gly355 and BbpThr365–Tyr387 in N2 domain shifted their configurations to stabilize the reorganized C-terminus mainly through strong hydrogen bonds. Altogether, our results revealed the molecular basis for Bbp-ligand interaction and advanced our understanding of S. aureus infection process.


INTRODUCTION
Staphylococcus aureus (S. aureus) has been one of the leading causes of bacterial infections worldwide. Each year, some 500,000 patients in United States' hospitals contract staphylococcal infections. S. aureus resides as part of the normal flora in the healthy human body until there is damage to skin surface or mucosal barrier, when it can gain access to tissues or the bloodstream, ultimately resulting in a wide range of infections and diseases, including impetigo, bacteremia, endocarditis, sepsis and arthritis (Lowy, 1998). Several antibiotics have been introduced to successfully treat S. aureus infections in patients over the past few decades. However, the infections became a growing concern lately due to the emergence of highly virulent and antibioticresistant strains, leading to increased morbidity and mortality (Zetola et al., 2005). Effective vaccines against S. aureus at early stages of infection are highly desirable, although all efforts to develop these vaccines have failed to date (Deresinski and Herrera, 2010;Vazquez et al., 2011).
S. aureus has evolved multiple strategies to promote colonization and evade the immune system. The initial adhesion of the pathogen to the extracellular matrix (ECM) components of the host is believed to be a critical step for successful infection. This is mediated by S. aureus surface adhesins called Microbial Surface Components Recognizing Adhesive Matrix Molecules (MSCRAMMs) (Gillaspy et al., 1998;Patti et al., 1994). Several structurally related proteins characterized by serine-aspartate dipeptide repeats (SD repeats) make up a family of MSCRAMMs (McCrea et al., 2000). The serine-aspartate repeat (Sdr) family include SdrF and SdrG in S. epidermidis and clumping factor A (ClfA), ClfB, SdrC, SdrD, SdrE and Bbp in S. aureus (Josefsson et al., 1998;McDevitt et al., 1994;Ni Eidhin et al., 1998;Tung et al., 2000). S. aureus bone sialoprotein-binding protein (Bbp) is an allelic variant of SdrE (Peacock et al., 2002). The members of Sdr family are predicted to adopt a similar structural pattern (Trad et al., 2004). A secretary signal sequence locates at the N-terminus followed by a ligandbinding A region and a characterized R region composed of SD repeats. The C-terminus features a cell wall-anchoring motif including the conserved LPXTG sequence (W), a hydrophobic membrane-spanning domain (M) and a short positively charged cytoplasmic tail (C) (Downer, 2002). In addition, SdrC, SdrD, SdrE and Bbp have different numbers of B repeats inserted between region A and R with the presence of a well-defined 12 residues cation-binding EFhand loop (Josefsson et al., 1998). Our recent work showed that B1 domain interacted with N2 domain and opened the ligand binding cleft between N2 and N3 domains in SdrD (Wang et al., 2013).
The gene identified from chromosomal DNA isolated from S. aureus subsp. aureus TCH60 encodes bone sialoproteinbinding protein (Bbp) with 1149 amino acids, containing SDrepeats of 154 residues and the ligand-binding A region from 53 to 601 residues further divided into N1, N2 and N3 domains. S. aureus isolated from patients suffering from septic arthritis and osteomyelitis specifically interacts with bone sialoprotein (BSP), a noncollagenous protein of bone and dentine extracellular matrix, mediated by Bbp (Ganss et al., 1999;Ryden et al., 1987;Tung et al., 2000). BSP is proposed to induce hydroxyapatite crystal formation and distributes predominantly in the newly formed bone, which is more likely to be infected (Hultenby, 1994;Hunter and Goldberg, 1993).
As a bifunctional MSCRAMM, Bbp also recognizes the human Fg α chain and inhibits thrombin induced blood coagulation (Vazquez et al., 2011). The molecular basis for Bbp-ligand interaction remains unknown. In our study, we solved the crystal structures of apo-Bbp 273−598 and Bbp 273−598 complexed with the peptide of Fg α 561−575 . We described the N2-N3 domains of Bbp similar to Dev-IgG fold (Deivanayagam et al., 2002). The Bbp 273−598 -Fg α 561−575 complex revealed the ligand-binding basis through the rearrangement of the C-terminus and the significant changes in four regions. These results advance our understanding of the ligand binding mechanism of Bbp during S. aureus infection. Moreover, our study should shed light on the further identification of the substrate or ligand of other closely related Sdr proteins, and provide novel targets for the development of potent antagonists, vaccines or antibiotics.

Structure of apo-Bbp 273−598
It was previously reported that on other MSCRAMMs of the Sdr family, the ligand-binding region was mapped to the N2 and N3 domains of the N-terminal A region. Based on sequence alignment of Bbp, SdrG and ClfA (Davis et al., 2001;Josefsson et al., 1998), we engineered a plasmid that would generate a recombinant fusion protein covering residues Asn 273 -Pro 598 of Bbp from S. aureus, a segment containing both N2 and N3 domains (Fig. 1A), with N-terminal GST-tag for purification purposes in Escherichia coli. We obtained apo-Bbp 273−598 protein and solved the structure at 2.03 Å resolution (Table 1), consisting of residues Asn 273 -Leu 584 and additional residues Gly and Ser at N-terminus, two of the remaining five amino acid residues (GPLGS) from digested GST-tag (Fig. 1C). No electron density was observed for the 14 residues at C-terminus in the apo-Bbp 273−598 structure.
The apo-Bbp 273−598 folds into two distinct domains N2 and N3, both of which have two layers of β-sheets and are structurally similar to the Dev-IgG fold (Fig. 1B), a variant of IgG fold (Deivanayagam et al., 2002). The two β-sheets of the N2 domain are composed of A, B and E strands on one side and C, D, D1, F and G strands on the opposite side. In N3 domain, C′, D1′, D2′, F′ and G′ strands form one principal sheet and A′, B′, D′ and E′ strands contribute to the facing sheet. The additional D1′ and D2′ strands present the featured Dev-IgG fold. One difference occurs with regard to the D strand in N2 domain, which parallels with E strand, although exhibiting an antiparallel orientation with the corresponding strand in the description of SdrG, ClfB and ClfA (Ganesh et al., 2008;Ponnuraj et al., 2003;Xiang et al., 2012).
In the apo-Bbp 273−598 structure, the C-terminus with poor electron density extends into the solvent region, thus leading to an open groove. Presumably, a ligand-binding site could exist in the groove. (All structural figures in this paper were generated by PyMOL).
Each crystallographic asymmetric unit contains two independent Bbp 273−598 -Fg α 561−575 molecules. The bound peptide Fg α 561−575 threads into the groove between the N2 and N3 domains in a snug conformation, with a well-defined 2Fo-Fc electron density map observed for the residues Lys 562 -Ser 569 of the N-terminus of the peptide (Fig. 2C). The six residues of the C-terminus of Fg α 561−575 have few interactions with the ligand binding groove and are not traceable in the density map. Due to the poor side-chain density, the N-terminal residue Fg α Ser561 was replaced with an Alanine during structure refinement.
Structure comparison of Bbp 273−598 between the apo-protein and in complex with the Fg α 561−575 shows that the RMSD for the Cα atoms is 0.679 Å. Even though the overall topology of Bbp 273−598 is similar in the two structures, significant conformational changes were observed around the peptidebinding groove, including the C-terminus of Bbp 273−598 and four additional regions containing Bbp Asp338-Gly355 , Bbp Thr365-Tyr387 , Bbp Ala298-Gly301 in N2 domain and Bbp Ser547-Gln561 in N3 domain, respectively (Fig. 3).

The structural basis for peptide binding
The disordered C-terminus encompassing residues Ser 585 -Pro 598 in apo-Bbp 273−598 rearranges in the Bbp 273−598 -Fg α 561−575 structure. The connecting loop between the G′ and G′′ strands spanning residues Thr 586 -Gly 589 runs across the central region of the groove and the following sequence forms a short β-strand G′′ which inserts into N2 domain (Fig. 3A), making regions Bbp Asp338-Gly355 and Bbp Thr365-Tyr387 deviate significantly from the apo-form (Fig. 3B). It shows a deviation of 5.6 Å for Bbp Pro347 in the region Bbp Asp338-Gly355 between the C and D strands in N2 domain. Bbp Gly590 and Gly592 of β-strand G′′ of N3 domain form two hydrogen bonds with Bbp Thr345 in this region. The region Bbp Thr365-Tyr387 in N2 domain contains β-strand E and the TYKFTDYVD sequence, a TYTFTDYVD-like motif conserved in Sdr protein family (McCrea et al., 2000). The β-strand E moves toward β-strand G′′ to stabilize the C-terminus of N3 through several hydrogen bonds. Bbp Tyr387 in this region interacts with Bbp Ser585 at the end of the G′ strand, which play an important role in redirecting the C-terminus. Bbp Asp373 and Arg374 in this region form two hydrogen bonds with Bbp Lys597 to stabilize the tail of the reordered C-terminus (Fig. 4A).
In addition, the binding of the peptide also induces a reorganization of the region Bbp Ala298-Gly301 between the A and B strands in N2 domain and a large movement of the region Bbp Ser547-Gln561 between the E′ and F′ strands in N3 domain toward the peptide binding groove. Two newly formed α-helices are observed in both of the two regions (Fig. 3B). The residues Bbp His299 and Gly301 in the first α-helix interact with Fg α Thr568 and Ser566 through two main-chain hydrogen bonds and the Bbp Asp300 forms the third hydrogen bond with the side-chain of Fg α Ser567 . In the second α-helix, Bbp Asp555 contributes a hydrogen bond with Fg α Lys562 and the side chain of Bbp Leu557 contacts with the aromatic ring of Fg α Phe564 mediated by a hydrophobic interaction (Fig. 4B).
The structural rearrangements and the direct protein-ligand interactions formed upon peptide binding result in an effectively stabilized Bbp-Fg α complex compared to its apoform.

Structural insights into Bbp 273−598 and Fg α 561−575 interactions
Apart from the interactions between the Fg α 561−575 peptide and the residues from the two newly formed α-helices we have described above, there are several contacts with distances of less than 4 Å marked (Fig. 4B). Among them, there are three pairs of antiparallel main-chain hydrogen bonds between residues Fg α Ser567, Thr565 and Gln563 and Bbp Thr582, Leu584 and Thr586 in the G′ strand and the connecting loop region. The carbonyl group of Fg α Ser567 interacts with the side-chain polar group of Bbp Asn581 forming another hydrogen bond. The residue Fg α Ser569 from the C-terminus of the peptide forms a hydrogen bond with Bbp Thr580 in the G′ strand. The side chain hydroxyl group of Fg α Ser566 interacts with the backbone atom of Bbp Phe433 in N3 domain and Fg α Ser566 forms the second hydrogen bond with the side-chain carbonyl group of Bbp Asp334 in N2 domain. Fg α Phe564 contributes two hydrogen bonds with Bbp Asp334 and Ile335 in the loop region between the C and D strands in N2 domain. The In allowed regions 2.9 1.1 In disallowed regions 0.6 0.0 Values in parentheses are for the highest resolution shell. R = Σ|F obs − F calc |/Σ Fobs , where F calc is the calculated protein structure factor from the atomic model (R free was calculated with 5% of the reflections selected). backbone atom of Bbp Asn581 interacts with the carbonyl group of Fg α Ser569 , which plays a significant role in anchoring the C-terminus of the Fg α 561−575 peptide. Two hydrogen bonds formed between the polar group of Fg α Gln563 and Bbp Asp588 from the loop region are involved in locking the peptide N-terminus.

Analysis of the interactions between Bbp 273−598 mutants and Fg α 561−575
Mutagenesis studies were conducted to further verify the binding of Fg α 561−575 to Bbp 273−598 . We mutated the residues Thr 582 and Leu 584 to Ala respectively which form two pairs of hydrogen bonds with the peptide. The mutant proteins were purified to homogeneity and tested for their interaction with the Fg α 561−575 peptide by surface plasmon resonance (SPR) (Fig. 5). The results indicate that the mutated protein Bbp T582A or Bbp L584A exhibits higher binding affinities with the peptide than wild type (WT) Bbp 273−598 protein. This is probably because the side chain of Ala occupies less space compared to Thr or Leu, which brings an alteration of steric hindrance. Thus, the alteration presumably makes the peptide more easily dock into the groove and contacts more tightly with Bbp 273−598 . Even though Bbp Thr582 and Leu584 could interact with the peptide by two pairs of main chain-main chain hydrogen bonds (Fig. 3C), the replacement to alanine might serve a similar role instead of breaking the interaction with the peptide according to the results here. Perhaps, we can speculate that the residues Bbp Thr582 and Leu584 are only involved in the binding with the peptide, but not showing specificities on ligand recognition.

RESEARCH ARTICLE
Xinyue Zhang et al.

DISCUSSION
In this study, we have solved the crystal structures of apo-Bbp 273−598 and the Bbp 273−598 -Fg α 561−575 complex. Based on the structural information, we analyzed the structural basis for ligand binding.
In our study, tight interactions between the protein and the ligand result in a stable binding state. Due to the "Dock" of the ligand, the rearrangements occur to C-terminus and additional four regions of Bbp 273−598 . The connecting loop covers the open groove, resulting in "Lock" of the ligand. And then the G′′ strand forms compact interactions with the E strand in the region Bbp Thr365-Tyr387 of N2 domain, which "Latch" the ligand binding site and thus stabilize the overall structure. Our structure further supports the DLL model described for the SdrG-Fg β complex (Fig. 6A) (Ponnuraj et al., 2003).
In the work of V. Ganesh et al. on the ClfB-ligand complex, the ligand binding mechanism was described as the "DL" model due to the absence of the "Latch" process . In their structure, peptide Fg α 336−347 in ClfB adopts a reverse orientation compared to the peptide Fg α 561−575 in our structure. The C-terminus of ClfB does not stretch into N2 domain to interact with the E strand but exhibits a different direction. Upon ligand binding, no rearrangements are observed in the region between the D and D′ strands in N2 and the region between the E and F strands in N3 (Fig. 6B). However, large movements occur to the corresponding regions Bbp Asp338-Gly355 in N2 and Bbp Ser547-Gln561 in N3 of Bbp 273−598 in our studies. The diversity of the ligands binding pattern of MSCRAMMs adds to the necessity for structural analysis of individual members of this family.
Altogether, our findings have promoted the understanding of the ligand binding mechanism of Bbp in Sdr family, a critical step in the S. aureus infection process. In addition, potential new target sites based on these pathogen-host interactions could be explored for development of potent antibiotics and new therapeutic methods.

Cloning, expression and purification of the recombinant proteins
The fragment of the Bbp gene (corresponding to 273-598 aa) was amplified using S. aureus ATCC 25923 genomic DNA by PCR. The gene fragments of mutated proteins Bbp T582A and Bbp L584A followed the same protocol. After digestion with BamHI and XhoI (NEB), the amplified fragments were cloned into the prokaryotic expression vector pGEX-6p-1 (GE Healthcare Life Sciences) to produce the GST-Bbp fusion protein and were confirmed by DNA sequencing. The recombinant protein was expressed in Escherichia coli strain BL21 with a high yield. The bacteria cells were harvested and resuspended in lysis buffer containing 1× PBS, 2 mmol/L DTT and 1 mmol/L PMSF. The cells were homogenized by sonification and cell debris was removed completely by centrifuging at 13,000 rpm for 50 min at 4°C.
The recombinant protein was firstly purified by GST-affinity column and digested with PreScission protease overnight. The eluant was collected and concentrated for further purification using gel filtration chromatography (Superdex-200 column, GE Healthcare) in buffer containing 20 mmol/L HEPES pH 7.5, 200 mmol/L NaCl, 2 mmol/L DTT on the FPLC system (GE Healthcare Life Sciences). The proteins in every step of purification were analyzed by SDS-PAGE.

Crystallization and structure determination
The apo-Bbp 273−598 and its complex were concentrated to 30 mg/mL in 20 mmol/L HEPES pH 7.5, 200 mmol/L NaCl, 2 mmol/L DTT. Crystals were screened by the sitting-drop vapor diffusion method (Jancarik et al., 1991) using sparse-matrix screen kits Crystal Screen I and II (Hampton Research), followed by optimizing the crystallization conditions through the variation of protein concentrations and pH. Crystals were grown at 18°C using the hanging-drop vapor diffusion method by mixing 1.0 μL protein solution with 1.0 μL reservoir solution and equilibrating against 200 μL reservoir solution. The apo-Bbp 273−598 crystals were grown in 0.2 mol/L calcium acetate hydrate, 0.1 mol/L sodium cacodylate trihydrate pH 6.5, 18% PEG8000. The synthesized Fg α 561−575 peptide was added into the concentrated protein samples at 10:1 ratio and the protein-peptide complex crystals are grown in 0.2 mol/L lithium sulfate, 0.1 mol/L Tris-HCl pH 8.2, 30% PEG4000. The apo-Bbp 273−598 and Bbp 273−598 -peptide complex crystals diffracted to 2.03 Å and 1.45 Å respectively. The data were collected at the Shanghai Synchrotron Radiation Facility (SSRF) BL17U using a MAR225 (MAR Research, Hamburg) CCD detector at 100 K and processed with the HKL2000 package (Otwinowski and Minor, 1997). Further processing was carried out using programs from the CCP4 suite (Collaborative Computational Project, 1994). The model building of the Bbp 273−598 molecules was conducted in COOT and the structure with peptide was determined by molecular replacement methods in CCP4 (Emsley and Cowtan, 2004). All the structures were refined with the PHENIX packages (Adams et al., 2002). Data collection and structure statistic are summarized in Table 1.

Isothermal titration calorimetry
ITC experiments were carried out at 25°C using a Microcal iTC200 instrument (GE Healthcare). The cell contained 50 μmol/L Bbp 273−598 and the syringe contained 500 μmol/L peptide in the buffer containing 200 mmol/L NaCl and 20 mmol/L HEPES pH 7.5. Injecting peptide into buffer was performed as the blank titration. The data were fitted and analyzed using the Origin 7 software package (Microcal).

Surface plasmon resonance spectroscopy
The interaction affinities between Fg α 561−575 and Bbp 273−598 protein were conducted by surface plasmon resonance (SPR) using BIAcore T200 instrument (GE Healthcare). The wild type Bbp 273−598 protein and two mutants were immobilized, respectively, on a CM5 sensor chip in 10 mmol/L sodium acetate, pH 4.5. The immobilization level was 3375RU. The synthetic peptide in 200 mmol/L NaCl and 20 mmol/L HEPES (pH 7.5) was injected on the protein-coated