Functional characterization and molecular modelling of FnFgBP, a surface protein from Streptococcus agalactiae

Abstract

Pneumococcal adherence and virulence factor A (PavA) were first identified in Streptococcus pneumoniae as an adhesin binding to fibronectin and studies suggested that it is an important virulence determinant. Homologs of PavA are found in many Gram-positive bacteria including S. agalactiae. However, to date, the details of its structure or mode of interaction with the host molecule(s) are not known. To characterize and identify the ligand binding region of GBS1263 of S. agalactiae (ortholog of PavA), two segments of GBS1263 namely seg-N (residues 1–264) and seg-C (residues 265–551) was cloned and expressed which resulted in accumulation of proteins in inclusion bodies. Seg-N and seg-C were solubilised using urea and subsequently the refolded proteins were purified. Circular dichroism studies and secondary structure prediction indicated that both the segments predominantly consist of helices. Molecular modelling also supports this data. Sequence comparison of these segments with a previously characterized fibronectin (Fn)/fibrinogen (Fg) binding protein, FBP54 from S. pyogenes showed that the residues 77–165 of seg-N have 81% similarity with the Fn/Fg binding region of FBP54 whereas seg-C has no significant similarity. Binding assays (dot blot, western blot and ELISA) and biolayer interferometry studies indicated that seg-N binds to both Fn and Fg whereas seg-C binds only to Fg. The present study has identified that GBS1263 could be a dual-ligand binding adhesin, binding to Fn and Fg using its different binding regions. This property is analogous to S. aureus adhesin FnBPA which binds to Fn and Fg using its different modules. Based on its dual-ligand binding property, we termed GBS1263 as FnFgBP (fibronectin/fibrinogen binding protein).

---

Introduction

Streptococcus agalactiae, a Gram positive, usually β-haemolytic, catalase negative bacteria, often referred to as Group B streptococcus (GBS), causes illness in people of all ages from newborn babies to pregnant or postpartum women, the elderly, and adults with chronic illnesses such as diabetes mellitus, malignancy and acquired immunodeficiency syndrome.¹ It causes life-threatening invasive diseases in newborns, such as bacterial sepsis and meningitis^2–4 by invading the alveolar epithelial⁵ and endothelial cells⁶ of lungs, endothelial blood–brain barrier⁷ etc. In non-pregnant adults, cellulitis, bacteraemia, urinary tract infection, osteoarticular infections, and pneumonia are the most common illnesses caused by GBS.⁸ It is also responsible for bovine mastitis, a disease of significant economical loss.⁹

Many factors are accountable in GBS infection¹⁰ of which the ability to adhere, colonize and invade are important. Bacteria have evolved surface associated proteins (adhesins) that enable their interaction with the host extracellular matrix (ECM) molecules like fibrinogen (Fg), fibronectin (Fn), laminin (Lm) etc.^11,12 Binding of certain bacterial adhesins [bacterial fibronectin-binding proteins (FnBPs)] to fibronectin, a 440 kDa dimeric ECM glycoprotein has been considered pivotal in bacterial colonization, bacterial virulence and bacteria–host interactions.¹³ FnBPs have been found in both Gram-negative and Gram-positive bacteria and many of them exhibit multiple ligand specificities.¹⁴ Similar to fibronectin, fibrinogen is also a large glycoprotein (340 kDa) and many pathogenic bacteria interact with Fg through its surface proteins termed as Fg-binding proteins (FgBPs).¹⁵ In Gram-positive bacteria, both FnBPs and FgBPs play additional functions such as evasion of immune responses and biofilm formation apart from adhesion to and invasion of host cells and tissues.¹⁶

The pneumococcal adherence and virulence factor A (PavA) was first identified in S. pneumoniae as an adhesin and infection determinant.¹⁷ Subsequently, it was characterized as a fibronectin binding protein mediating pneumococcal colonization in host cells. Based on its interaction with immobilized human fibronectin and studies involving truncation and isogenic mutants, PavA was suggested to play a direct role in the pathogenesis of pneumococcal infections.¹⁸ Although PavA is localized to the pneumococcal cell outer surface, it lacks the typical features of Gram-positive adhesins namely, a characteristic N-terminal leader peptide for protein export and a C-terminal cell wall anchorage sequence. Thus, PavA has been categorized as anchorless adhesin and this group includes some other bacterial surface proteins such as fibronectin binding protein (FBP54) of S. pyogenes, enolase (eno) of S. pneumoniae, surface dehydrogenase (SDH) of S. pyogenes and surface enolase (SEN) of S. pyogenes.¹⁹ PavA also lacks repeated GGX_3–4I/VDF motif responsible for fibronectin binding.^20,21 These features distinguish PavA from many other fibronectin binding proteins and suggest that there might be a distinct mechanism of PavA binding to fibronectin. Following the discovery of PavA in S. pneumoniae, a similar protein (GBS1263) was identified in S. agalactiae.²² The 63.5 kDa (551 amino acids) GBS1263 shows 71% sequence similarity with PavA of S. pneumoniae, 77% and 72% similarity to fibronectin binding proteins FBP54 of S. pyogenes^23,24 and FbpA S. gordonii²⁵ respectively.

The high sequence similarity exhibited by GBS1263 to PavA, FBP54 and FbpA prompts us to hypothesize that it can also bind to fibronectin. To understand the binding mechanism of GBS1263 to fibronectin, we have initiated the structural and functional characterization of this protein. This work describes the cloning, expression, purification, ligand binding and molecular modelling of two different segments of GBS1263. Two constructs, seg-N (N-terminal 1 to 264 amino acid residues) and seg-C (C-terminal 265 to 551 amino acid residues) were generated, and their affinity towards fibronectin and fibrinogen has been characterized using dot blot, western blot, ELISA and the Biolayer Interferometery technique. Further, investigations on the structure have been carried out using CD spectroscopy and homology modelling studies. Our study revealed that GBS1263 is a dual-ligand binding adhesin capable of binding to both fibronectin and fibrinogen and therefore we termed it as FnFgBP (fibronectin/fibrinogen binding protein).

Experimental

Bacterial strains, plasmids and culture conditions

Freeze-dried Type III S. agalactiae NEM316 strain was obtained from the American Type Culture Collection (ATCC). E. coli DH5α was used for plasmid cloning and BL21 (DE3) for protein expression. pET28a (Novagen) was used to make the vector-DNA construct. The streptococcal strain was grown in Todd-Hewitt Broth (THB) supplemented with 0.1% yeast extract and all other E. coli strains were grown in Luria Bertani (LB) broth and when required 50 μg mL⁻¹ of kanamycin was supplemented. All cultures were grown at 37 °C. Cloning and expression hosts were made competent by the CaCl₂ method.

Cloning and expression

Genomic DNA from S. agalactiae NEM316 was isolated and the region encoding for GBS1263 N-terminal (1–264 amino acids, FnFgBP seg-N) and C-terminal (265–551 amino acids, FnFgBP seg-C) segments were PCR amplified with an upstream oligonucleotide, 5′-CGTG̲G̲A̲T̲C̲C̲ATGTCTTTTGATGGATTTTTT-3′ and a downstream oligonucleotide, 5′-CGGC̲T̲C̲G̲A̲G̲CTCTTGGTAGTAATAATCTAACAATC-3′ for N-terminal segment and upstream oligonucleotide, 5′-CCAG̲G̲A̲T̲C̲C̲AAGGCTGAAAAAGATCGCA-3′ and downstream oligonucleotide, 5′ GCGC̲T̲C̲G̲A̲G̲CTATTTTTTTAGTTTAAGGCTATCTAT-3′ for C-terminal segment, containing the restriction sites for BamH1 and Xho1 (underlined) at 5′ and 3′ ends respectively. The PCR involves the following steps, denaturation 95 °C for 30 s, annealing at 50 °C for 1 min followed by elongation at 68 °C for 1 min for a total of 30 cycles.

The amplified PCR products were double digested with BamHI and XhoI restriction enzymes and ligated with pET28a which had been digested with the same restriction enzymes. The vector encodes a C-terminal hexahistidyl tag (6×His) using which one-step purification by Ni²⁺-affinity chromatography could be employed. The ligated products were transformed into DH5α cells and positive clones were selected on kanamycin-supplemented agar plate. Colony PCR was carried out for the colonies grown on the LB plates followed by plasmid mini preps and gene sequencing.

The positive constructs were transformed into competent cells of expression strain, E. coli BL21 (DE3) and cells were grown in LB medium containing 50 μg mL⁻¹ of kanamycin at 37 °C with continuous agitation at 180 rpm until an optimum optical density (OD₆₀₀ of 0.6) was reached. Protein over-expression was induced with 1 mM IPTG (isopropyl β-D-1-thiogalactopyranose) and the cells were incubated for an additional 4 h at 37 °C. The cells were harvested by centrifugation at 6000 rpm for 15 min at 4 °C and the expression of FnFgBP seg-N and seg-C was analysed on 12.5% SDS-PAGE.

Purification

The SDS-PAGE analysis revealed that both FnFgBP seg-N and seg-C were found in the insoluble fraction and therefore purified from the inclusion bodies involving a stepwise purification process. The cell pellet from 250 mL culture was resupended in a buffer consisting of 50 mM Tris, 1 M urea, 5 mM EDTA, 5 mM β mercaptoethanol and 1 mM phenylmethanesulfonylfluoride (PMSF) (pH 8.0). The resuspended cells were lysed by sonication and centrifuged at 10 000 rpm at 4 °C for 15 min. The supernatant containing the contaminants was discarded and the pellet containing FnFgBP protein was resuspended in 50 mM Tris, 6 M urea, 400 mM NaCl, 5 mM β mercaptoethanol and 5% glycerol (pH 8.0). This suspension was again sonicated and centrifuged for 15 min at 10 000 rpm. The FnFgBP in supernatant was refolded stepwise by dialyzing against buffer with a reducing concentration of urea. As a first step, the unfolded protein was dialysed against 50 mM Tris, 4 M urea, 400 mM NaCl, 5 mM β mercaptoethanol, 5% glycerol (pH 8.0) for overnight at 4 °C. Subsequently the protein was dialyzed against the same buffer, but the concentration of the urea was reduced to 2 M and later to 50 mM for 4 hours at 4 °C. As a final step, dialysis was carried out against 50 mM Tris, 400 mM NaCl, 0.1 M sucrose, 0.1% Triton X 100, 10 mM arginine, 10 mM urea, and 5% glycerol (pH 8.0) for overnight at 4 °C. FnFgBP seg-N and seg-C were loaded separately onto an Ni–NTA IMAC column (5 mL HisTrap column, GE Healthcare) pre-equilibrated in buffer containing 50 mM Tris, 400 mM NaCl, 0.1 M sucrose, 10 mM urea and 5% glycerol (pH 8.0) and eluted stepwise using 100 mM to 1 M imidazole. The eluted fractions were concentrated and subjected to size-exclusion chromatography using Superdex 75 column (GE Healthcare Biosciences) equilibrated with the same buffer. The eluted peak fractions were concentrated (Pierce; 10 kDa cutoff) and the protein concentration was measured by UV absorption at 280 nm at ambient temperature using a 10 mm path length quartz cell.

A western blot analysis was carried out to confirm if both FnFgBP seg-N and seg-C were expressed with a C-terminal His-tag. For this seg-N and seg-C were subjected to 12.5% SDS-PAGE gel and electroblotted onto nitrocellulose membrane. The nonspecific sites on the membrane were blocked with 5% Bovine-Serum Albumin (BSA) for 3 hours in Tris buffer saline containing 0.1% Tween20 (TBST). The membrane was washed thrice with TBST and incubated in the primary antibody (His-tag monoclonal antibody, Sigma Aldrich) at a dilution of 1 : 1000 for 2 h at room temperature with agitation. After washing, it was then incubated with goat anti-mouse IgG alkaline phosphatase conjugate (Sigma Aldrich) antibody for 2 h at room temperature. Then the membrane was washed with TBST and incubated with BCIP (5-bromo-4-chloro-3-indoyl phosphate disodium salt, 1% BCIP in 100% dimethylformamide)/NBT (nitro blue tetrazolium chloride, 1.5% in 100% dimethylformamide) solution until the colour development was achieved. BCIP/NBT in TBST was used to detect the alkaline phosphatase conjugated antibodies.

Circular dichroism

The purified FnFgBP seg-N and seg-C, both with a concentration of 0.5 mg mL⁻¹ in 20 mM phosphate buffer saline, pH 8 were used for CD spectra analysis. Far-UV CD spectra were recorded in 190–250 nm range using 1 mm path length quartz cell at 25 °C with a resolution of 1 nm in JASCO J-815 circular dichroism spectrometer. The obtained data were analysed using K2D2 ²⁶ and SOMCD²⁷ web servers for the determination of secondary structure composition.

Fibrinogen (Fg) and fibronectin (Fn) binding assays

The binding of FnFgBP seg-N and seg-C to Fg and Fn was evaluated using blotting assays, enzyme linked immunosorbent assay (ELISA) and biolayer interferometry (BLI) analysis.

Blotting assays

For dot blot, 1 μL of human Fg (250 μg mL⁻¹) and 1 μL of human Fn (250 μg mL⁻¹) were spotted onto nitrocellulose membrane and allowed to dry at room temperature. The membrane was blocked with 5% BSA in TBST for 2 h at room temperature. After washing thrice in TBS, the membrane strips were incubated with 0.5 mg mL⁻¹ FnFgBP seg-N and seg-C separately for 2 h at room temperature. SgrA from Enterococcus faecium²⁸ and PfbA from Streptococcus pneumoniae²⁹ were used as a positive control for Fg and Fn binding respectively. Carbonic anhydrase (CA) from Geobacillus kaustophilus³⁰ and Jack bean urease (JBU) was used as a negative control for the binding of both Fg and Fn. The blots were incubated with His-tag monoclonal antibody for 1 h followed by incubation with goat anti-mouse IgG alkaline phosphatase conjugate antibody for 1 h at room temperature to detect the bound protein. The membrane was then washed three times with TBST and incubated with BCIP/NBT to detect the colour development.

The reverse ligand blot was carried out by spotting 1 μL each of FnFgBP seg-N (0.5 mg mL⁻¹) and seg-C (0.5 mg mL⁻¹) on the nitrocellulose membrane and air dried. The membrane was blocked for 2 h at room temperature with 5% BSA and incubated with Fg (250 μg mL⁻¹) at room temperature for 2 h. Subsequently the membrane was incubated with monoclonal anti-fibrinogen antibody for 1 h, followed by goat anti-mouse IgG alkaline phosphatase conjugate antibody for 1 h. The membrane was washed thrice after each step with TBST and incubated in a solution containing NBT/BCIP for colour development.

For western blot, Fg and Fn were subjected to 10% and 8% SDS-PAGE separately. Since Fg is made up of three different polypeptide chains (α, β and γ), three bands were observed in the gel. The protein bands in the gel were electro blotted onto a nitrocellulose membrane. The membranes were blocked with 5% BSA for about 2 h in TBST and incubated with purified FnFgBP seg-N and seg-C separately for 2 h at room temperature. The membranes were washed three times with TBST and incubated with His-tag monoclonal antibody (1 : 5000 dilution in TBS) for 1 h at room temperature. They were washed again three times with TBST and incubated for 1 h with goat anti-mouse IgG alkaline phosphatase conjugate (1 : 10 000 dilution in TBS). BCIP/NBT in TBST was used as a substrate to detect the alkaline phosphatase conjugated antibodies. The bound protein was visualized by incubating the membrane in a developing solution (100 mM Tris–HCl pH 9.5, with 1 mM MgCl₂) for about 15 min in the dark.

ELISA

The microtiter plates were coated with Fg and Fn in TBS at 4 °C overnight. The nonspecific binding sites were blocked with 200 μL of 5% BSA in TBST per well for 2 h at room temperature. The purified FnFgBP seg-N and seg-C were added in different concentrations to the wells and incubated for 1 h at room temperature. Bound protein was detected by incubating the wells with His-tag monoclonal antibody for 1 h followed by goat anti-mouse IgG alkaline phosphatase conjugate. The alkaline phosphatase reaction was carried out by adding pNPP (para nitro phenyl phosphate) later incubated in dark for 15 min. The reaction was stopped by the addition of 50 μL of 2 N NaOH to each well. The absorption of the solution at 405 nm was quantified in an ELISA reader. After each incubation step, the wells were washed thrice with TBST. For positive control, the wells coated with Fg and Fn was incubated in different concentration of SgrA and PfbA respectively. For negative control, CA with His-tag and JBU which is devoid of a His-tag was used.

Bio-layer interferometry

The molecular interactions of FnFgBP seg-N and seg-C binding to Fg and Fn was measured with BlitzPro 1.1 bio-layer interferometer advanced kinetics. As a first step, the Ni–NTA biosensor was hydrated in PBS for 10 min prior to each experimental run. Next, the Ni–NTA biosensor was dipped in PBS to attain zero baseline for 30 s. Following this, the biosensor was activated with analyte FnFgBP seg-N/seg-C for 120 s and the unbound protein was removed from the biosensor by placing it in a buffer for 30 s. Subsequently in the association step, the ligand Fg or Fn was allowed to bind to the immobilised FnFgBP seg-N/seg-C in the Ni–NTA biosensor for 120 s. Finally in the dissociation step, the FnFgBP–Fg or FnFgBP–Fn complex forged in association was transferred to PBS for 120 s to dissociate. Standardisation was accomplished with an analyte concentration of 0.25–3 μM and the ligand concentration of 0.005–1 μM. As a control, the Ni–NTA biosensor loaded with FnFgBP seg-N/seg-C and the association and disassociation steps were carried out with PBS alone without any ligand. Based on the binding curves of FnFgBP seg-N with Fg and Fn, FnFgBP seg-C with Fg the association rate constant (k_a), dissociation rate constant (k_d) and affinity constant (K_D) for the complexes FnFgBP–Fg and FnFgBP–Fn were calculated using Forte Bio Data analysis package with the global fitting function.

Sequence comparison

Homologous sequence search of full length FnFgBP was carried out against non-redundant database from NCBI using PSI-BLAST.³¹ To identify the closely related structures of FnFgBP, blastp search against the protein data bank (PDB) was carried out at the NCBI.³²

Structure prediction

The SWISS MODEL³³ and RaptorX³⁴ servers were used for the homology or comparative modeling of the three-dimensional structures of FnFgBP seg-N and seg-C. Ramachandran plot assessment server RAMPAGE³⁵ was used to examine the quality of the obtained model. ProSA-web: the interactive web server³⁶ was also used to analyze the errors in three-dimensional structures of the homology models.

Results

The 1656 bp gene (gbs1263) from S. agalactiae encoding the putative fibronectin binding protein, was cloned as two different segments into the vector pET28a with a C-terminal His-tag. The first 794 nucleotides encode the N-terminal fragment of the protein (FnFgBP seg-N) while the remaining 862 nucleotides encode the C-terminal fragment of the protein (FnFgBP seg-C).

Protein expression and purification from inclusion bodies

IPTG induced cultures showed protein bands corresponding to 30.6 kDa for seg-N and 32.9 kDa for seg-C on a 12.5% SDS-PAGE (Fig. 1A and B). 250 mL culture pellets were lysed by sonication and the proteins (FnFgBP seg-N and FnFgBP seg-C) were observed in the insoluble fraction as inclusion bodies. Extensive trials were carried out to solubilise the protein by varying the pH of the buffer from 7 to 12.5. In addition, agents like Triton X 100, Tween20, CHAPS, guanidine hydrochloride and urea were added in different concentrations to the buffer and the solubilisation of pellets was performed. Among the various agents tried the protein was more soluble in urea. Both seg-N and seg-C were solubilised (unfolded) in a buffer containing 50 mM Tris, 6 M urea, 400 mM NaCl, 5 mM β mercaptoethanol and 5% glycerol (pH 8.0). During refolding, a stepwise dialysis to reduce the urea concentration was carried out. In this process, precipitation of the protein was observed which was controlled by the addition of sucrose, arginine and glycerol to the buffer. Both proteins were found to be stable in a buffer containing 50 mM Tris, 400 mM NaCl, 0.1 M sucrose, 0.1% Triton X 100, 10 mM arginine, 10 mM urea and 5% glycerol (pH 8.0). The refolded proteins were purified to near homogeneity using Ni–NTA column and the bound protein was eluted with an increasing concentration of imidazole. FnFgBP seg-N was obtained in 300 mM to 500 mM imidazole elutions whereas FnFgBP seg-C was eluted with a buffer containing 200 mM to 400 mM imidazole concentration. The obtained fractions were pooled, concentrated and subjected to size-exclusion chromatography. The peak fractions were again pooled, concentrated and a single band was observed as judged by SDS-PAGE analysis (Fig. 1C and D). The concentration of seg-N and seg-C was measured as 7 mg mL⁻¹ and 5 mg mL⁻¹ respectively. The presence of a His tag fused to the recombinant proteins was further confirmed by anti-His antibodies in the blotting analysis (Fig. 1E and F).

Fig. 1 The FnFgBP seg-N and seg-C expression, purification, western and dot blot analysis of the purified proteins. (A) 12.5% SDS-PAGE gel showing over expression of FnFgBP seg-N in BL21DE3. Lane 1: molecular weight ladder (kDa), lane 2: seg-N before induction, lane 3: seg-N after induction, lane 4: seg-N supernatant after lysis, lane 5: seg-N pellet after lysis. (B) 12.5% SDS-PAGE gel showing expression of seg-C in BL21DE3. Lane 1: molecular weight ladder (kDa), lane 2: seg-C before induction, lane 3: seg-C after induction, lane 4: seg-C supernatant after lysis, lane 5: seg-C pellet after lysis. (C) Purified seg-N and seg-C on 12.5% SDS-PAGE gel. Lane 1: molecular weight ladder (kDa), lane 2: seg-N at 1 mg mL⁻¹, lane 3: seg-N at 5 mg mL⁻¹, lane 4: seg-C (1 mg mL⁻¹), lane 5: seg-C (5 mg mL⁻¹). (D) SDS-PAGE of purified seg-N and seg-C used for western blot analysis. Lane 1: molecular weight ladder (kDa), lane 2: seg-N (1 mg mL⁻¹), lane 3: seg-C (1 mg mL⁻¹). (E) The purified seg-N and seg-C analyzed by western blotting using anti-His tag antibody. Lane 1: seg-N, lane 2: seg-C. (F) Dot blot detection of 6× His-tagged seg-N and seg-C using anti-His tag antibody. Lane 1: seg-N (100 μg mL⁻¹), lane 2: seg-C (100 μg mL⁻¹), lane 3: positive control, Fg binding His-tagged SgrA (100 μg mL⁻¹), lane 4: JBU without a His-tag, a negative control for the experiment (100 μg mL⁻¹).

Secondary structure analysis

Structural information of FnFgBP or any of its close homologs is not available till date. Therefore the CD spectroscopic analysis of FnFgBP seg-N and seg-C (Fig. 2) was carried out to obtain their secondary structural details. The estimation of secondary structural content from the obtained CD data was calculated using the online servers K2D2 and SOMCD. From this analysis, it is inferred that seg-N consists of nearly equal amount of alpha helices and random coils (about 40%). The beta sheet content is about 15%. In seg-C, higher helical content (59%) was observed when compared with seg-N. It also consists of 31% random coil and 9% beta sheet. The secondary structural content of seg-N and seg-C was also predicted from their sequence using the bioinformatics tools SOPMA³⁷ and HNN³⁸ available at the webserver NPS@: Network Protein Sequence Analysis. The results obtained from the experimental data (CD) and the theoretical predictions are in close concurrence (Table 1).

Fig. 2 Circular dichorism spectra of FnFgBP seg-N and seg-C. The CD spectrum of FnFgBP segments in the wavelength range of 190 nm to 240 nm. Both seg-N (green) and seg-C (red) are mostly composed of alpha helices.

Table 1 Secondary structural composition of FnFgBP seg-N & seg-C obtained from CD spectral data and bioinformatics tool

Protein	Alpha helix (%)			Beta sheet (%)			Random coils (%)
	CD data	Bioinformatics tool		CD data	Bioinformaticstool		CD data	Bioinformatics tool
		SOPMA	HNN		SOPMA	HNN		SOPMA	HNN
FnFgBP seg-N	44.15	40.53	35.98	14.45	17.80	20.08	41.40	41.67	43.94
FnFgBP seg-C	59.17	50.87	44.60	9.70	16.38	19.86	31.13	32.75	35.54

---

Binding of FnFgBP seg-N and seg-C to Fg and Fn

To determine the ability of different segments of FnFgBP to interact with Fg and Fn, dot blot and western blot assays were carried out. From the dot blot experiments, using anti-His antibody, it was observed that seg-N binds to both Fg and Fn (Fig. 3A) whereas seg-C binds only to Fg and not Fn (Fig. 3B). For a positive control Fg binding E. faecium adhesin SgrA (Fig. 3C) and Fn binding S. pneumoniae adhesin PfbA (Fig. 3D) were used. Jack bean urease devoid of His-tag and recombinant carbonic anhydrase with His-tag was used as negative controls where no binding to both Fg and Fn was observed (Fig. 3E and F). Dot blot assay using anti-fibrinogen antibody also affirmed the binding of seg-N and seg-C to Fg (Fig. 3G).

Fig. 3 Dot blot to confirm the binding of FnFgBP seg-N and seg-C to Fg and Fn. The nitrocellulose membrane was spotted with 1 μL of Fg and Fn separately and incubated with seg-N (A) and seg-C (B) and probed with anti-His antibodies. The binding of seg-N towards both Fg and Fn was observed (A) whereas seg-C to binds only Fg and not to Fn (B). For a positive control, the nitrocellulose membrane was blotted with 1 μL of Fg and Fn separately and incubated with Fg and Fn binding proteins SgrA (C) and PfbA (D) respectively. For a negative control, the nitrocellulose membrane spotted with Fg and Fn was incubated with JBU (devoid of His-tag) (E) and CA (contains a His-tag) (F) and no binding was observed. (G) In a reverse blot, the nitrocellulose membrane was spotted with 1 μL of seg-N (Left spot) and seg-C (Right spot) and incubated with Fg. The membrane was probed with anti-fibrinogen antibody and the resultant color development indicates both seg-N and seg-C binds to Fg.

The 340 kDa Fg is a dimer and composed of three different chains namely α, β and γ. On SDS-PAGE gel, under reducing condition, these three chains can be separated. Therefore to unravel which chain of Fg is being recognized by FnFgBP segments, Fg was subjected to 12.5% SDS-PAGE gel and the three chains were separated (Fig. 4A). Two such gels were transferred to nitrocellulose membrane and probed separately with FnFgBP seg-N and seg-C by western blot. This analysis showed that seg-N is capable of binding to β and γ subunits of Fg independently (Fig. 4B) and seg-C is able to bind to β-subunit alone (Fig. 4C). The 440 kDa Fn is a dimer and composed of two identical chains linked together by disulphide bonds. The SDS-PAGE gel of Fn (8% polyacrylamide) is shown in Fig. 4D. The gel was transferred to the membrane and probed with seg-N and seg-C separately by western blot. The resultant color development revealed the binding of seg-N to Fn (Fig. 4E) whereas seg-C does not bind to Fn (Fig. 4F).

Fig. 4 Western blot assay showing FnFgBP seg-N and seg-C binding to Fg and Fn. (A) Fg was size separated into three different subunits such as α, β, γ chains in order of decreasing molecular mass as shown in 12.5% SDS-PAGE gel. The Fg was transferred onto the nitrocellulose membrane and incubated with seg-N (B) and seg-C (C) followed by its incubation in anti-His tag antibody. The experiment unravels that seg-N has the ability to bind to β and γ subunits of Fg whereas seg-C binds prominently to β subunit of Fg alone. (D) Fn (220 kDa) observed on 8% SDS-PAGE gel. After electrophoresis, it was electroblotted onto the nitrocellulose membrane and incubated with seg-N (E) and seg-C (F) followed by incubation with anti-His tag antibody. As the color development indicates only seg-N binds to Fn and seg-C does not.

Quantification by ELISA

The presence of His-tag in FnFgBP seg-N and seg-C and the binding of FnFgBP segments to immobilized Fg and Fn was quantified using ELISA. As an initial experiment, the microtiter plates were coated with seg-N and seg-C separately. The bound protein was detected using His-tag monoclonal antibody followed by goat anti-mouse IgG alkaline phosphatase conjugate. The absorbance taken at 405 nm was plotted on an X–Y graph (Fig. 5A). The plot confirms the presence of His-tag in both seg-N and seg-C.

Fig. 5 Quantification of FnFgBP seg-N and seg-C binding to Fg and Fn by ELISA. (A) The microtiter plates were coated with His-tagged FnFgBP seg-N and seg-C proteins separately and incubated with anti-His tag antibodies. The efficiency of the binding was quantified at the absorbance of 405 nm. This experiment confirms the presence of His-tag in both seg-N and seg-C. JBU not having the His-tag was used as a negative control. (B) The Fg (1 μg mL⁻¹) was immobilized on microtiter plates and probed with varying concentration of purified seg-N and seg-C separately followed by incubation with anti-His tag and secondary antibodies. The plot demonstrates the binding affinity of both seg-N and seg-C to Fg. SgrA is used as a positive control. CA and JBU are negative controls. (C) The Fn (5 μg mL⁻¹) was immobilized on microtitre plates where probed with varying concentration of purified seg-N and seg-C separately followed by incubation with anti-His tag and secondary antibodies. The plot reveals the binding affinity of seg-N to Fn whereas seg-C lacks binding to Fn. PfbA is used as a positive control. CA and JBU are negative controls. Points representing the means of the triplicates and standard deviations are indicated.

In another experiment, the microtiter plates were coated with Fg at a concentration of 1 μg mL⁻¹. To the wells, seg-N and seg-C were added separately in a serial dilution at a concentration of 0.01–0.1 μg mL⁻¹. The bound protein was detected using His-tag monoclonal antibody followed by goat anti-mouse IgG alkaline phosphatase conjugate. The absorbance taken at 405 nm shows that seg-N and seg-C binds to Fg in a dose dependent manner (Fig. 5B). Similar experiment was carried out with microtiter plates coated with Fn (at a concentration of 5 μg mL⁻¹). In this experiment the binding of seg-N towards Fn was observed whereas seg-C does not bind to Fn (Fig. 5C). The results of the ELISA experiments are consistent with the dot blot analysis. The data obtained from ELISA experiments were analyzed by non-linear regression fitting using GraphPad Prism 6.1 for Windows (GraphPad software, San Diego California, USA).

Bio-layer interferometry

The Ni–NTA biosensor tip was pre-equilibrated and the baseline correction was carried out with PBS. FnFgBP seg-N and seg-C at a concentration of 250 nM and 300 nM was loaded to the Ni–NTA biosensor independently. The biosensor loaded with the protein was incubated with the analytes Fg and Fn for association and at the end of 120 s the tip was placed in PBS for dissociation. Sensorgrams were recorded as a function of binding response (binding in nm) versus time to calculate the binding affinity, rates of association and dissociation during the interaction. The quadruplicate results were globally analysed and the sensorgrams were plotted using GraphPad Prism 6.1. The analysis revealed that FnFgBP seg-N and seg-C bound to Fg with a K_D value of 6.457 × 10⁻⁹ M and 4.101 × 10⁻⁸ M respectively. The resultant graphs are shown in Fig. 6A and C. The binding of Fn with seg-N resulted in a K_D value of 4.486 × 10⁻⁸ M and the corresponding graph is shown in Fig. 6B. The seg-C binding to Fn resulted in a K_D value of 1.258 × 10⁻¹ M as a consequence of no binding affinity.

Fig. 6 Biolayer interferometry studies of the interaction between immobilized FnFgBP seg-N and seg-C with Fg and Fn. Each binding curve in response to the increasing concentration of the ligand against the analyte shows the interaction of FnFgBP seg-N and seg-C with Fg and Fn. Seg-N was immobilized in the Ni–NTA biosensor tip at a concentration of 0.45 μM and the association and disassociation kinetics were calculated for Fg (A) and Fn (B) at four different concentrations respectively. (C) The Ni–NTA biosensor tip was loaded with seg-C at a concentration of 0.5 μM. Sensogram showing the binding (binding in nm) of seg-C with increasing concentrations of Fg. The sensograms were fit globally to a 1 : 1 binding model. The binding affinity (K_D) and the rate of association and disassociation constants (k_a and k_d) were calculated with the ForteBio Data Analysis software.

Sequence comparison

A BLAST search of FnFgBP (GBS1263) amino acid sequence against non-redundant protein database provided 500 candidates with an identity ranging from 47 to 100%. The 25 closest homologs were identified as dihydroorotate dehydrogenase and putative fibronectin/fibrinogen binding protein of different streptococal species and none of which have been structurally or functionally characterized. In this search, two slightly distant but well characterized Fn/Fg binding homologs, namely FBP54 of S. pyogenes (77% identity with FnFgBP of S. agalactiae) and PavA of S. pneumoniae (71% identity) have been identified and their sequences were compared to FnFgBP of S. agalactiae. Previous studies of FBP54 have confirmed that the N-terminal 89 residues were responsible for both Fn and Fg binding.^23,24 On the contrary, in another study on PavA of S. pneumoniae the C-terminal region consisting of 189 amino acids was identified critical for Fn binding.¹⁷ In view of these observations, a multiple sequence alignment was carried out by aligning the sequences of FBP54 from S. pyogenes, FnFgBP from S. agalactiae and PavA of S. pneumoniae. The N-terminal Fn/Fg binding region of FBP54 (residues 1–89) exhibit 81% and 78% sequence homology with the region comprising of residues 77 to 165 of FnFgBP from S. agalactiae and PavA of S. pneumoniae respectively (Fig. 7). This comparison clearly shows that the 89 amino acids primarily located in the N-terminal region of the three Fn/Fg binding proteins is highly conserved and this conservation suggests that this region might be responsible for ligand binding. However, further studies are required to validate this observation. Sequence comparison of full length FnFgBP of S. agalactiae against protein data bank (PDB) yielded sequences with a very low identity. But this search identified the N-terminal region of FnFgBP (residues 1–273) to exhibit 41.06% sequence identity with residues 1–288 of the putative fibrinogen binding protein of S. aureus Mu50 (PDB 3DOA). However, the functional characterization of this protein has not been carried out.

Fig. 7 Full length sequence comparison of FnFgBP from S. agalactiae (GBS1263, UniprotKB Q8E4X9), FBP54 of M5 S. pyogenes (UniprotKB Q54858) and PavA of S. pneumoniae (UniprotKB Q9RNF3). Highlighted region denotes the similarity among the proteins. The region 1–89 of FBP54 was identified as Fn/Fg binding region which is indicated by a black box.

Structure prediction

The full length sequence of FnFgBP of S. agalactiae (GBS1263, UniprotKB Q8E4X9) was used for modelling the structure. The SWISS-MODEL homology modelling server was used for the modelling. The initial template search process ranked the S. aureus putative fibrinogen binding protein (PDB 3DOA) as a best template to model the N-terminal first 273 residues of FnFgBP. Subsequently this structure was selected as a template and the three dimensional structure of FnFgBP seg-N was obtained. The modelled structure consisted of an N-terminal β-domain made up of two β-sheets forming a β-sandwich structure, constituted by residues 20 to 167 and a C-terminal α-helical domain made up of residues 168 to 264. The N-terminal putative Fn/Fg binding region constituted by the residues 77 to 165 of FnFgBP (corresponding to 1–89 of FBP54) was mapped on the model and it is found that this region is formed by one of the two β-sheets followed by a long loop that connects to the helical domain (Fig. 8A).

Fig. 8 Molecular modelling of FnFgBP of S. agalactiae. (A) Modelled structure of FnFgBP seg-N obtained from the SWISS-MODEL server. Seg-N consists of residues 1 to 264 and residues 1–76, 77–165 and 166–264 are shown in magenta, yellow and green respectively. In FBP54 of S. pyogenes, the N-terminal region (1 to 89) was predicted as Fn/Fg binding region. The corresponding region in FnFgBP of S. agalactiae is constituted by residues 77 to 165 which is shown in yellow. (B). Full length model of FnFgBP obtained from the Raptor-X server. Seg-N (1–264 residues) and seg-C (265–551 residues) are shown in cyan and grey respectively. (C) Superposition of seg-N modelled from SWISS-MODEL server (magenta, yellow and green) and Raptor-X server (cyan). The model obtained from both the servers is highly identical.

The homology modelling of FnFgBP seg-C using SWISS-MODEL was not possible since suitable homologous structure(s) covering the full C-terminal segment are not available. Therefore RaptorX server was used to predict the three dimensional structure of the full length FnFgBP. The advantage of using RaptorX is that it uses one or multiple distantly related template proteins for the structure prediction. From the full length FnFgBP model obtained from the RaptorX (Fig. 8B), the structure of the N-terminal segment was compared with the model obtained from the SWISS-MODEL. The structure of the N-terminal segment obtained from both the servers superpose with an RMS deviation of 0.273 Å (Fig. 8C) which suggests both the models are highly homologous. The structure of C-terminal segment predicted by the RaptorX mainly consists of alpha helices with two distinct domains; a linear helical domain constituted by residues Asp269-Arg417 followed by a cap domain (residues Ala418-Lys551) consists of both helices and beta-sheet.

Discussion

Microbial surface components recognizing adhesive matrix molecules (MSCRAMMs) are proteins located on the bacterial surface mediating adherence of the bacteria to components of the extracellular matrix of the host which includes fibronectin, fibrinogen, laminin, collagen etc. Bacterial attachment to fibronectin, fibrinogen and other ECM molecules is considered to be an important virulence factor in streptococcal, staphylococcal and enterococcal infections.^13,16,39,40 In particular with reference to GBS, in a recent study it was demonstrated that the deletion of fbsC, a gene encoding a Fg-binding protein, decreases the ability of bacterium to adhere to and invade human epithelial and endothelial cells.⁴¹ A similar study has also been carried out on other Fg-binding adhesins of GBS such as FbsA^42,43 and FbsB⁴⁴ and the results indicate that these proteins play an important role in the entry of S. agalactiae into epithelial cells. In an another study, it was demonstrated that the binding of surface protein Srr1 of S. agalactiae to fibrinogen promotes the attachment of the bacterium to human brain microvascular endothelial cells.⁴⁵ These results, along with other studies⁴⁶ clearly suggests that GBS surface proteins are strongly associated with the infection process. Likewise in many bacterial species the role of surface proteins in the pathogenesis has been clearly established. Therefore, in general, structural and biochemical characterization of the interaction of MSCRAMMs with the host molecule will provide useful data for the development of antimicrobials and vaccines.

One of the conserved features in MSCRAMMs is the C-terminal LPxTG motif, which is recognized by the sortase enzyme. Sortase cleaves the LPxTG sequence and covalently links the C-terminal end of an MSCRAMM to the peptidoglycan layer. In contrast, some surface proteins do not possess the LPxTG motif but they appear attached to the surface of the bacteria through unknown mechanism. These proteins are termed ‘anchorless adhesins’ and GBS1263 belongs to this category. The protein GBS1263 is an ortholog of PavA from S. pneumonia which is characterized as a fibronectin binding protein.¹⁷ In addition, GBS1263 also shares high degree of sequence similarity with fibronectin binding proteins FBP54 of S. pyogenes and FbpA S. gordonii. Although the homologues of GBS1263, PavA and FBP54 are found in many Gram-positive and Gram-negative bacteria,¹⁴ very little structural information is available for this family of proteins. Therefore we initiated the structural and functional characterization of GBS1263 and in particular to identify its ligand binding regions and the binding mechanism. Towards this goal, GBS1263 has been cloned as N-terminal (1–264 residues, seg-N) and C-terminal (265–551 residues, seg-C) fragments.

To evaluate the binding of recombinant seg-N and seg-C with Fg and Fn, ligand affinity blotting, ELISA and BLI studies were employed. The dot blot assay confirmed the fibrinogen binding property of GBS1263 which is localized both in the N-terminal and C-terminal segments. The western blot analysis indicated that seg-N is capable of binding independently to β and γ subunits of Fg and seg-C is able to bind only to the β-subunit. Subsequently to calculate the kinetic rate constants (k_a and k_d) and affinity constant (K_D) of interaction, BLI experiments were carried out for seg-N and seg-C with the ligand Fg (Fig. 6A and C). This analysis revealed that the N-terminal segment exhibited higher affinity when compared with the C-terminal segment. The observed difference in the binding affinity between the two regions of the molecule can be correlated with the results obtained from the western blot where the seg-N binds to both β and γ chain of Fg in contrast to seg-C which binds only to β chain (Fig. 4).

In addition to Fg, the Fn binding affinity of seg-N and C was analysed by BLI experiments (Fig. 6B). From this study, the Fn binding is observed only for seg-N but not to seg-C. This finding correlates well with the results of the dot blot and ELISA experiments where it was observed only seg-N exhibited binding to Fn. The dot blot, ELISA and bio layer interferometry analysis thus reveal the dual-ligand binding ability of GBS1263 and therefore we termed it as FnFgBP.

A characteristic feature of Fg and Fn binding MSCRAMMs is the presence of a N-terminal leader peptide and the C-terminal cell wall anchorage sequence. FnFgBP and PavA lacks these elements¹⁸ and also devoid of GGX_3–4I/VDF motif, a repeat sequence in cell wall anchored proteins of streptococcal and staphylococcal species that is involved in Fn binding,^20,21 making it a unique class of adhesins. The secondary structure analysis using CD spectroscopy revealed that both FnFgBP seg-N and seg-C are mainly constituted of α-helices and random coils whereas the well characterized Fg binding adhesins (SdrG of S. epidermidis,⁴⁷ ClfA and ClfB S. aureus^48,49 and Srr-1 of S. agalactiae⁵⁰) and Fn binding adhesins (PfbA of S. pneumoniae⁵¹ and FnBPA of S. aureus⁵²) are predominantly β-stranded.

In FnBP and its homologs a common tandem β-zipper mechanism of binding with Fn was observed, in which the Fn binding motif of the adhesin, which is usually a disordered region, forms an antiparallel β-strand and binds to consecutive F1 modules in the N-terminal domain of Fn.^53–55 However FnFgBP does not show significant sequence similarity with the above mentioned Fg and Fn binding proteins. Interestingly, FnFgBP possibly adopts different structural fold from the Fg and Fn binding MSCRAMMs structurally characterized till date. All these differences infer that there might be a different mechanism associated with FnFgBP in recognition and binding of both Fg and Fn.

A previous study on FnFgBP's homolog FBP54 of S. pyogenes has identified that the N-terminal 89 residues were involved in both Fn/Fg binding.^23,24 This region shares 81% identity with a segment constituted by residues Met77-Pro165 of S. agalactiae FnFgBP (Fig. 7). The present study reveals that the N-terminal region of FnFgBP is capable of binding to both Fn and Fg while the C-terminal region only to Fg. This result thus supports the previous observation. In another study, in PavA of S. penumoniae, the C-terminal region was identified as critical for Fn binding and this contradicts our finding that FnFgBP seg-C only binds to Fg. One plausible reason for the difference in the binding of FnFgBP seg-C and PavA towards Fn could be the presence of an additional 98 residues in FnFgBP seg-C (total number of residues is 287) in comparison with the C-terminal 189 residues of PavA. In addition, there is a difference of about 25% in the sequence identity between these two regions which could also play a role in the recognition of the ligand.

The present study thus suggests that FnFgBP could be a dual-ligand adhesin, analogous to FnBPA of S. aureus which is capable of binding to both Fg and Fn.⁵² It was shown that the N-terminal A-region of FnBPA interacts with Fg by a ‘dock, lock and latch mechanism’⁴⁷ whereas the intrinsically unstructured repeats which follows the Fg binding region is responsible for Fn binding through a ‘β-zipper mechanism’.⁵³ However, recently it has been shown that the A-region of FnBPB binds to Fn in addition to Fg.⁵⁵ This is similar to our observation on FnFgBP seg-N which binds to both Fn and Fg. In FnFgBP, although both N and C-segments exhibits the Fg binding property, neither showed significant sequence similarity to characteristic Fg and Fn binding regions (A-domain and intrinsically unstructured repeat region respectively) of FnBPA. To gain structural insights, the three-dimensional structure of FnFgBP seg-N and seg-C were generated by molecular modelling where the former one is composed of both beta-sheets and alpha-helices and the latter one is primarily alpha-helical in nature. Among the Fg and Fn binding MSCRAMMs, significant structural details have been obtained only for Fg binding MSCRAMMs. The predicted structure of FnFgBP seg-N and seg-C (both binds to Fg) is significantly different from the structure of Fg binding MSCRAMMs FnBPA, SdrG, ClfA and ClfB. Hence, we hypothesize that FnFgBP and its homologs binds to Fg and Fn by using mechanisms different from the well characterized ones. Thus, this study paves way for further structural and biochemical studies which could help in mapping the exact residues vital for Fn/Fg binding and the mode of adhesin-host binding.

Acknowledgements

KP thanks the Department of Science and Technology (DST), Government of India for financial support.

References

1. K. Le Doare and P. T. Heath, Vaccine, 2013, 31, D7–D12.
2. K. A. Simonsen, A. L. Anderson-Berry, S. F. Delair and H. D. Davies, Clin. Microbiol. Rev., 2014, 27, 21–47.
3. J. R. Verani and S. J. Schrag, Clin. Perinatol., 2010, 37, 375–392.
4. K. M. Edmond, C. Kortsalioudaki, S. Scott, S. J. Schrag, A. K. Zaidi, S. Cousens and P. T. Heath, Lancet, 2012, 379, 547–556.
5. C. E Rubens, S. Smith, M. Hulse, E. Y. Chi and G. V. Belle, Infect. Immun., 1992, 60, 5157–5163.
6. R. L. Gibson, M. K. Lee, C. Soderland, E. Y. Chi and C. E. Rubens, Infect. Immun., 1993, 61, 478–485.
7. K. S. Doran, E. J. Engelson, A. Khosravi, H. C. Maisey, I. Fedtke, O. Equils, K. S. Michelsen, M. Arditi, A. Peschel and V. Nizet, J. Clin. Invest., 2005, 115, 2499–2507.
8. D. Blancas, M. Santin, M. Olmo, F. Alcaide, J. Carratala and F. Gudiol, Eur. J. Clin. Microbiol. Infect. Dis., 2004, 23, 168–173.
9. G. P. Keefe, Can. Vet. J., 1997, 38, 429–437.
10. V. Nizet and C. E. Rubens, Pathogenic mechanisms and virulence factors of group B streptococci, in The Gram-Positive Pathogens, ed. V. A. Fischetti, R. P. Novick, J. J. Ferretti, D. A. Portnoy and J. I. Rood, ASM Press, Washington, D. C., 2000, pp. 125–136.
11. C. Chagnot, A. Listrat, T. Astruc and M. Desvaux, Cell. Microbiol., 2012, 14, 1687–1696.
12. J. M. Patti, B. L. Allen, M. J. McGavin and M. Hook, Annu. Rev. Microbiol., 1994, 48, 585–617.
13. U. Schwarz-linek, M. Hook and J. R. Potts, Mol. Microbiol., 2004, 52, 631–641.
14. B. Henderson, S. Nair, J. Pallas and M. A. Williams, Curr. Opin. Microbiol., 2012, 15, 44–49.
15. J. Rivera, G. Vannakambadi, M. Hook and P. Speziale, Thromb. Haemostasis, 2007, 98, 503–511.
16. T. J. Foster, J. A. Geoghegan, V. K. Ganesh and M. Hook, Nat. Rev. Microbiol., 2014, 12, 49–62.
17. A. R. Holmes, R. McNab, K. W. Millsap, M. Rohde, S. Hammerschmidt, J. L. Mawdsley and H. F. Jenkinson, Mol. Microbiol., 2001, 33, 1395–1408.
18. D. Pracht, C. Elm, J. Gerber, S. Bergmann, M. Rohde, M. Seiler, K. S. Kim, H. F. Jenkinson, R. Nau and S. Hammerschmidt, Infect. Immun., 2005, 73, 2680–2689.
19. G. S. Chhatwal, Trends Microbiol., 2002, 10, 205–208.
20. J. Jaffe, S. Natanson-Yaron, M. G. Caparon and E. Hanski, Mol. Microbiol., 1996, 21, 373–384.
21. Q. Sun, G. M. Smith, C. Zahradka and M. J. McGavin, Infect. Immun., 1997, 65, 537–543.
22. P. Glaser, C. Rusniok, C. Buchrieser, F. Chevalier, L. Frangeul, T. Msadek, M. Zouine, E. Couve, L. Lalioui, C. Poyart, P. Trieu-Cuot and F. Kunst, Mol. Microbiol., 2002, 45, 1499–1513.
23. H. S. Courtney, J. B Dale and D. L Hasty, Infect. Immun., 1996, 64, 2415–2419.
24. H. S. Courtney, Y. Li, J. B. Dale and D. L. Hasty, Infect. Immun., 1994, 62, 3937–3946.
25. J. Christie, R. McNab and H. F. Jenkinson, Microbiology, 2002, 148, 1615–1625.
26. C. P. Iratxeta and M. A. A. Navarro, BMC Struct. Biol., 2008, 8, 8–25.
27. P. Unneberg, J. J. Merelo, P. Chacon and F. Moran, Proteins, 2001, 42, 460–470.
28. R. Nagarajan, A. P. Hendrickx and K. Ponnuraj, FEBS J., 2016, 283, 3039–3055.
29. M. Yamaguchi, Y. Terao, Y. Mori, S. Hamada and S. Kawabata, J. Biol. Chem., 2008, 283, 36272–36279.
30. P. Ragunathan, G. Raghunath, S. Kuramitsu, S. Yokoyama, T. Kumarevel and K. Ponnuraj, Acta Crystallogr., Sect. F: Struct. Biol. Cryst. Commun., 2013, 69, 162–164.
31. S. F. Altschul and E. V. Koonin, Trends Biochem. Sci., 1998, 23, 444–447.
32. S. F. Altschul, T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller and D. J. Lipman, Nucleic Acids Res., 1997, 25, 3389–3402.
33. M. Biasini, S. Bienert, A. Waterhouse, K. Arnold, G. Studer, T. Schmidt, F. Kiefer, T. G. Cassarino, M. Bertoni, L. Bordoli and T. Schwede, Nucleic Acids Res., 2014, 42, 252–258.
34. M. Kallberg, G. Margaryan, S. Wang, J. Ma and J. Xu, Methods Mol. Biol., 2014, 1137, 17–27.
35. S. C. Lovell, I. W. Davis, W. B. Arendall, P. I. de Bakker, J. M. Word, M. G. Prisant, J. S. Richardson and D. C. Richardson, Proteins, 2003, 50, 437–450.
36. M. Wiederstein and M. J. Sippl, Nucleic Acids Res., 2007, 35, 407–410.
37. C. Geourjon and G. Deleage, Comput. Appl. Biosci., 1995, 11, 681–684.
38. Y. Guermeur, Combinaison de classifieurs statistiques, Application a la prediction de structure secondaire des proteins, Ph.D Thesis, Université Paris 6, 1997.
39. D. Joh, E. R. Wann, B. Kreikemeyer, P. Speziale and M. Hook, Matrix Biol., 1999, 18, 211–223.
40. L. E. Hancock, B. E. Murray and J. Sillanpaa, Enterococcal cell wall components and structures, in Enterococci: from commensals to leading causes of drug resistant infection, ed. M. S. Gilmore, D. B. Clewell, Y. Ike and N. Shankar, Massachusetts Eye and Ear Infirmary, Boston, MA, 2014.
41. M. Buscetta, S. Papasergi, A. Firon, G. Pietrocola, C. Biondo, G. Mancuso, A. Midiri, L. Romeo, G. Teti, P. Trieu-cuot and C. Beninati, J. Biol. Chem., 2014, 289, 21003–21015.
42. A. Schubert, K. Zakikhany, G. Pietrocola, A. Meinke, P. Speziale, B. J. Eikmanns and D. J. Reinscheid, Infect. Immun., 2004, 72, 6197–6205.
43. T. Tenenbaum, C. Bloier, R. Adam, D. J. Reinscheid and H. Schroten, Infect. Immun., 2005, 73, 4404–4409.
44. H. Gutekunst, B. J. Eikmanns and D. J. Reinscheid, Infect. Immun., 2004, 72, 3495–3504.
45. H. S. Seo, R. Mu, B. J. Kim, K. S. Doran and P. M. Sullam, PLoS Pathog., 2012, 8, 1002947.
46. A. Tazi, S. Bellais, I. Tardieux, S. Dramsi, P. Trieu-cuot and C. Poyart, Curr. Opin. Microbiol., 2012, 15, 44–49.
47. K. Ponnuraj, M. G. Bowden, S. Davis, S. Gurusiddappa, S. Moore, D. Choe, Y. Xu, M. Hook and S. V. Narayana, Cell, 2003, 115, 217–228.
48. C. C. Deivanayagam, E. R. Wann, W. Chen, M. Carson, K. R. Rajashankar, M. Hook and S. V. Narayana, EMBO J., 2002, 21, 6660–6672.
49. V. K. Ganesh, E. M. Brabu, C. C. Deivanayagam, B. Le, A. S. Anderson, Y. V. Matsuka, S. L. Lin, T. J. Foster, S. V. Narayana and M. Hook, J. Biol. Chem., 2011, 286, 25963–25972.
50. R. Sundaresan, U. Samen and K. Ponnuraj, Int. J. Biol. Macromol., 2015, 75, 97–105.
51. D. S. Beulin, M. Yamaguchi, S. Kawabata and K. Ponnuraj, Int. J. Biol. Macromol., 2014, 64, 168–173.
52. V. Stemberk, R. P. Jones, O. Moroz, K. E. Atkin, A. M. Edwards, J. P. Turkenburg, A. P. Leech, R. C. Massey and J. R. Potts, J. Biol. Chem., 2014, 289, 12842–12851.
53. U. Schwarz-Linek, J. M. Werner, A. R. Pickford, S. Gurusiddappa, J. H. Kim, E. S. Pilka, J. A. G. Briggs, T. S. Gough, M. Hook, I. D. Campbell and J. R. Potts, Nature, 2003, 423, 177–181.
54. N. C. Norris, R. J. Bingham, G. Harris, A. Speakman, R. P. Jones, A. Leech, J. P. Turkenburg and J. R. Potts, J. Biol. Chem., 2011, 286, 38311–38320.
55. F. M. Burke, A. Di Poto, P. Speziale and T. J. Foster, FEBS J., 2011, 278, 2359–2371.

---