Novel structure of the N-terminal helical domain of BibA, a group B streptococcus immunogenic bacterial adhesin

The X-ray crystallographic structure of an N-terminal fragment of group B streptococcus BibA (BibA_126–398) and a low-resolution structure of the full-length N-terminal domain (BibA_34–400) determined using small-angle X-ray scattering are described. The association of the N-terminal domain of BibA was localized to the C4BP α-chain.

Abstract

BibA, a group B streptococcus (GBS) surface protein, has been shown to protect the pathogen from phagocytic killing by sequestering a complement inhibitor: C4b-binding protein (C4BP). Here, the X-ray crystallographic structure of a GBS BibA fragment (BibA_126–398) and a low-resolution small-angle X-ray scattering (SAXS) structure of the full-length N-terminal domain (BibA_34–400) are described. The BibA_126–398 fragment crystal structure displayed a novel and predominantly helical structure. The tertiary arrangement of helices forms four antiparallel three-helix-bundle-motif repeats, with one long helix from a bundle extending into the next. Multiple mutations on recombinant BibA_34–400 delayed the degradation of the protein, and circular dichroism spectroscopy of BibA_34–400 suggested a similar secondary-structure composition to that observed in the crystallized BibA_126–398 fragment. A model was generated for the 92 N-terminal residues (BibA_34–125) using structural similarity prediction programs, and a BibA_34–400 model was generated by combining the coordinates of BibA_34–126 and BibA_126–398. The X-ray structure of BibA_126–398 and the model of BibA_34–400 fitted well into the calculated SAXS envelope. One possible binding site for the BibA N-terminal domain was localized to the N-terminal CCP (complement-control protein) domains of the C4BP α-chain, as indicated by the decreased binding of BibA to a ΔCCP1 C4BP α-chain mutant. In summary, it is suggested that the GBS surface protein BibA, which consists of three antiparallel α-helical-bundle motifs, is unique and belongs to a new class of Gram-positive surface adhesins.

1. Introduction  

Group B streptococcus (GBS) causes sepsis and meningitis in neonates (Dermer et al., 2004 ▸), and is the second most common pathogen infecting infants (the first being Escherichia coli) during their first week of life (Ladhani et al., 2019 ▸). A worldwide meta-analysis estimates that GBS affects infants at a rate of 0.3–2 per 1000 lives (with higher incidences in Africa compared with those in other regions), with early-onset infections occurring more frequently than late-onset infections (Madrid et al., 2017 ▸). GBS is also responsible for infections in pregnant and peripartum women (Krohn et al., 1999 ▸) and is increasingly being recognized to cause severe infections among adults with diabetes, cancer and other chronic illnesses (Farley, 2001 ▸; Sendi et al., 2008 ▸; Farley et al., 1993 ▸).

Human complement, which eliminates microorganisms and other antigens from blood and tissues, plays a critical role in the host defense against pathogens (Wessels et al., 1995 ▸; Jarva et al., 2003 ▸). Many enzymes, cofactors, regulatory proteins and receptors constitute complement, and they act in concert to induce inflammatory and cytocidal reactions in response to invading microbes. The activation of complement, which is essential for its biological activity, is mediated through three pathways: the classical pathway, which is activated primarily by antibody–antigen complexes, the alternative pathway, which is triggered directly by pathogens, and the lectin pathway, which is initiated by carbohydrates on the surface of microorganisms. The complement plays an important role in host defense against GBS: while the classical pathway is implicated in early bloodstream clearance of GBS (Butko et al., 1999 ▸), it has been shown that antibody-mediated immunity requires C3 and not C4, thus also suggesting a role for the alternative pathway in opsonophagocytosis of GBS (Wessels et al., 1995 ▸).

Santi et al. (2007 ▸) reported BibA to be a surface protein that was expressed in all 24 GBS strains that they analyzed, and its role in GBS adhesion was first demonstrated by the impaired ability of a bibA-knockout mutant strain to adhere to both cervical and lung epithelial cells. In a complement-mediated opsonophagocytic assay, 50% of mice infected with the wild-type 2603/VR strain died within three days, while only 10% of mice infected with a bibA-mutant 2603/VR strain died in the same period, confirming the role of BibA in GBS virulence. More importantly, when tested for the ability of GBS to replicate in human blood, the bibA-mutant 2603/VR strain proliferated 80% less efficiently compared with the wild type, suggesting that the survival of GBS is better in human blood when BibA is surface-expressed than when BibA is absent (Santi et al., 2007 ▸).

BibA is a modular protein that is predicted to have a helix-rich N-terminal domain (34–399) followed by a proline-rich C-terminal region (400–543) with a canonical LPXTG cell-wall-anchoring motif LPSTG at the C-terminus (Fig. 1 ▸ a). The sequence of BibA is highly conserved in all clinical isolates, with the N-terminal domain being identical and the proline-rich region showing 98–100% similarity (Santi, Maione et al., 2009 ▸). The putative N-terminal helical domain protects GBS from phagocytic killing by attaching to the complement inhibitor C4b-binding protein (C4BP; Santi et al., 2007 ▸; Pietrocola et al., 2018 ▸). BibA protein was shown to be highly expressed on the GBS surface at pH 7.0 and to a lower extent at pH 5.5 (Santi, Grifantini et al., 2009 ▸), and such pH-regulated BibA expression is fully reversible upon exposure to a new pH environment. This also explains the induced opsonophagocytic killing of GBS by anti-BibA antibodies at pH 7.0, while no killing activity was observed against bacteria grown at pH 5.5 (Santi, Grifantini et al., 2009 ▸). Mice immunized with BibA were shown to be protected against the challenge of GBS, strengthening the argument for the beneficial effects of vaccination using BibA (Santi, Maione et al., 2009 ▸; Senn et al., 2011 ▸).

Figure 1.

(a) Predicted domain arrangement for the GBS surface protein BibA. The C-terminal sorting motif is shown. Different C4BP variants were used: plasma-purified C4BP (b), rC4BP (c) and C4BP α-chain deletion mutants (d). The main isoform of C4BP contains seven identical α-chains and one unique β-chain. The β-chain-containing C4BP in circulation is bound to the vitamin-K-dependent anticoagulant Protein S (PS), forming a C4BP–PS complex. The rC4BP contains six α-chains and lacks the β-­chain and the associated PS. C4BP α-chain deletion mutants lack single CCP domains (represented by gray circles with an X in each α-chain).

The complement inhibitor C4BP is a spider-shaped molecule (Fig. 1 ▸ b). The main isoform of C4BP contains seven identical α-chains and a unique β-chain. The β-chain-containing C4BP in circulation is bound to the vitamin-K-dependent anticoagulant Protein S (PS), forming a C4BP–PS complex. A less abundant form of C4BP, which is upregulated upon inflammation, is composed exclusively of the α-chains (Fig. 1 ▸ c). Each α-chain is made of eight CCP (complement-control protein) domains and has multiple binding sites for C4b. The binding interface of C4BP for C4b is contributed by a cluster of positively charged amino acids present in the linker between CCP1 and CCP2 (Leung et al., 2006 ▸; Blom et al., 1999 ▸; Blom, Zadura et al., 2000 ▸; Villoutreix et al., 1998 ▸; Blom, 2002 ▸). This same region of C4BP is also crucial for binding C3b (Blom et al., 2003 ▸), heparin (Blom et al., 1999 ▸) and several bacterial pathogens such as Haemophilus influenzae (Hallström et al., 2007 ▸), the gingipain HRgpA from Porphyromonas gingivalis (Potempa et al., 2008 ▸), the Moraxella catarrhalis UspA1/2 proteins (Nordström et al., 2004 ▸) and the Streptococcus pyogenes surface M protein (Blom, Berggård et al., 2000 ▸; Buffalo et al., 2016 ▸). An NMR structure of C4BP_CCP1–2 (Jenkins et al., 2006 ▸) and the crystal structures of the hypervariable regions of four different M proteins (M2, M49, M22 and M28) in complex with C4BP_CCP1–2 (Buffalo et al., 2016 ▸) are available and have revealed the significant contributions of residues Arg39, Arg64, Arg66 and His67 towards the binding of S. pyogenes M proteins to C4BP_CCP1–2. The N-terminal region of BibA has previously been implicated in its association with C4BP (Santi et al., 2007 ▸).

A structural view of the association of BibA and C4BP could help in developing strategies for the prevention and treatment of GBS infections. Towards this goal, we attempted to determine the crystal structure of the N-terminal domain of BibA (BibA_34–400) in complex with C4BP_CCP1–2. However, we could only obtain crystals of a proteolytic fragment of BibA (BibA_126–398), and its structure was determined using X-ray crystallo­graphic techniques. BibA_126–398 displays a novel helical structure that forms four antiparallel three-helix-bundle-motif repeats. Further, we performed a structural investigation into the intact BibA_34–400 protein using circular dichroism (CD) spectrometry and small-angle X-ray scattering (SAXS). Based on the above analyses, we propose BibA (BibA_34–400) to be a distorted rod-shaped molecule. We also analyzed the molecular basis of the interactions between BibA and recombinant rC4BP (α-chain) by performing surface plasmon resonance analysis and localized the binding sites for BibA_34–400 to the C4BP α-chain.

2. Materials and methods  

2.1. Cloning and mutagenesis  

The bibA gene from group B streptococcus strain 2603 (serotype V/R) was used as a template to express the recombinant forms of BibA. The gene segment corresponding to the N-terminal domain of BibA encompassing amino-acid residues 34–400, labeled BibA_34–400, was amplified by PCR with the respective primers (Table 1 ▸). The products were cloned into vector pMCSG7 (DNASU plasmid repository, USA), which harbors an N-terminal histidine tag and TEV protease cleavage site, using ligation-independent cloning (LIC). The Change-IT Multiple Mutation Site-Directed Mutagenesis Kit (USB Affymetrix, USA; Chen & Ruffner, 1998 ▸) was used to generate the mutants with their respective mutagenic primers (Table 1 ▸). The resulting constructs and mutations were confirmed by DNA sequencing.

Table 1. Recombinant production information.

Source organism	Streptococcus agalactiae
Forward primer	TACTTCCAATCCAATGCACACGCGGATACTAGTTCAGG
Reverse primer	TTATCCACTTCCAATGTTATTTTTTATCTTCTTTAGTACCTCT
Forward K49N mutation primer	[Phos]GGAATATCGGCTTCAATTCCTCATAAGAACCAAGTTAATTTAGGGGCGGTTACT
Reverse K70N mutation primer	[Phos]TACTCTACTTAAAAGTATAGCAATAGCGTTGTCATTACCACGATATTTAGAAAT
Forward K112N mutation primer	[Phos]ATTTTATATCAAGGACAAATTGGTAACCAAAATAAACCAAGTGTAACTACA
Cloning vector	pMCSG7
Expression vector	pMCSG7
Expression host	E. coli BL21 (DE3)
Complete amino-acid sequence of the construct produced (BibA34–400)†	MHHHHHHSSGVDLGTENLYFQSHADTSSGISASIPHKKQVNLGAVTLKNLISKYRGNDKAIAILLSRVNDFNRASQDTLPQLINSTEAEIRNILYQGQIGKQNKPSVTTHAKVSDQELGKQSRRSQDIIKSLGFLSSDQKDILVKSISSSKDSQLILKFVTQATQLNNAESTKAKQMAQNDVALIKNISPEVLEEYKEKIQRASTKSQVDEFVAEAKKVVNSNKETLVNQANGKKQEIAKLENLSNDEMLRYNTAIDNVVKQYNEGKLNITAAMNALNSIKQAAQEVAQKNLQKQYAKKIERISSKGLALSKKAKEIYEKHKSILPTPGYYADSVGTYLNRFRDKQTFGNRSVWTGQSGLDEAKKMLDEVKKLLKELQDLTRGTKEDKK

^†Mutated residues are underlined and the His₆ tag is shown in italics.

2.2. Expression and purification of recombinant proteins  

The recombinant plasmids were transformed into E. coli BL21(DE3) cells (Novagen, USA) for expression. An overnight seed culture was used to inoculate 1 l Luria–Bertani (LB) broth supplemented with 50 µg ml⁻¹ ampicillin. The cells were grown at 37°C until the OD_600 nm reached ∼0.7, at which point protein expression was induced overnight by the addition of 1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG). The cells were harvested by centrifugation at 3210g for 30 min at 4°C using a Beckman Allegra 6R centrifuge. The cell pellets were resuspended in lysis buffer consisting of 50 mM Tris–HCl pH 7.0, 200 mM NaCl, 5%(v/v) glycerol, 100 mM l-arginine and an EDTA-free protease-inhibitor cocktail tablet (Roche, USA). The bacterial cells were lysed by the freeze–thaw method and 1 mg DNase I (New England BioLabs, USA) was added to remove unwanted DNA from the cell lysate to improve the protein-extraction efficiency. The lysate was cleared by centrifugation at 48 384g for 1 h at 4°C using a Beckman Avanti J-25 centrifuge. The cleared lysate was passed through a 5 ml pre-packed HisTrap FF column (GE Healthcare, USA) pre-equilibrated with lysis buffer at a flow rate of 1 ml min⁻¹. The protein-bound column was washed with ten bed volumes of Ni column buffer A (the same as the lysis buffer) and ten bed volumes of buffer B [50 mM Tris–HCl pH 7.0, 200 mM NaCl, 5%(v/v) glycerol, 100 mM l-arginine, 250 mM imidazole] at a 15% concentration. Bound protein was eluted with a linear gradient of 15–100% buffer B. Eluted fractions with the protein of interest were pooled, treated with TEV protease to remove the N-terminal His tag and concentrated using an Amicon ultracentrifugation system to a final volume of 1 ml. This sample was further purified by gel-filtration chromatography (HiPrep 16/60 Sephacryl S-100 HR column; GE Healthcare) in 20 mM Tris–HCl pH 7.5, 150 mM NaCl, 10%(v/v) glycerol, 25 mM l-arginine, 5 mM EDTA, 5 mM β-mercaptoethanol (SEC buffer).

The substitution of BibA_34–400 with selenomethionine (SeMet) and its expression were carried out by inoculating a single colony carrying pMCSG7-BibA_34–400 into 50 ml minimal medium (Molecular Dimensions, USA) supplemented with 50 µg ml⁻¹ ampicillin and incubating overnight at 37°C. The cells were harvested by centrifugation at 3210g for 30 min at 4°C and were washed thoroughly with distilled water before resuspending them in 1 ml water. The resuspended cell pellets were decanted into 1 l minimal medium containing l-selenomethionine, which had been prewarmed to 37°C, and were allowed to grow until the OD_600 nm reached ∼0.5. At this point, the medium was cooled to 30°C and protein expression was induced by the addition of 1 mM IPTG when the OD_600 nm reached ∼0.7; the culture was then grown overnight. The cells were harvested and lysed and SeMet-substituted BibA_34–400 was purified using a method similar to that used for the native protein. For both BibA_34–400 proteins (native and SeMet-substituted) the purity was checked by 12% SDS–PAGE (Laemmli, 1970 ▸), and their concentrations were determined by measuring their UV absorbance at 280 nm using an ∊ value of 21 890 M ⁻¹ cm⁻¹.

The main isoform of C4BP (570 kDa) containing α- and β-chains as well as Protein S (Fig. 1 ▸ b) was purified from human plasma as described in Dahlbäck (1983 ▸), while recombinant wild-type C4BP composed of only α-chains (Fig. 1 ▸ c) and rC4BP mutants lacking single CCP domains in the α-chains (Fig. 1 ▸ d) were expressed in eukaryotic cells and purified by affinity chromatography as described previously (Blom et al., 2001 ▸).

2.3. Protein crystallization and X-ray diffraction data collection  

The purified BibA_34–400 protein was concentrated with an Amicon ultracentrifugation system (molecular-weight cutoff 30 kDa). The JCSG Core Suites I and II (Qiagen), The PEGs Suites I and II (Qiagen) and Crystal Screen HT (Hampton Research, USA) crystallization kits were used in initial crystallization screens. Robotic screening was performed with a Crystal Phoenix Robot (Art Robbins Instruments, USA) in a 96-well sitting-drop plate (Corning 3550) at 20°C. Optimization of the conditions to produce good-quality single crystals was performed by hand in 24-well hanging-drop plates (VDX plates; Hampton Research). The final buffer composition and the reservoir solution were as described in Table 2 ▸. X-ray diffraction data were collected at 100 K using 20%(v/v) glycerol as a cryoprotectant on the SER-CAT beamline 22-ID at the Advanced Photon Source (APS), Argonne National Laboratory, USA.

Table 2. Crystallization information.

Method	Hanging-drop vapor diffusion
Plate type	24-well plate (VDX plate, Hampton Research)
Temperature (K)	295
Protein concentration (mg ml−1)	40.8
Composition of protein solution	20 mM Tris–HCl pH 7.5, 150 mM NaCl, 10%(v/v) glycerol, 25 mM L-arginine, 5 mM EDTA, 5 mM β-mercaptoethanol
Composition of reservoir solution	25%(w/v) PEG 3350, 0.1 M bis-Tris pH 5.0, 0.2 M magnesium chloride
Volume and ratio of drop	2.2 µl, 1:1 (+0.2 µl crystal seed)
Volume of reservoir (ml)	1

2.4. Structure determination and refinement  

Data collected from SeMet BibA_126–400 crystals at the selenium edge and peak were processed using XDS (Kabsch, 2010 ▸) and iMosflm (Battye et al., 2011 ▸). The crystal structure was solved by multi-wavelength anomalous dispersion (MAD) phasing to a resolution of 3.03 Å using CRANK2 (Skubák & Pannu, 2013 ▸), which is integrated into the CCP4 suite (Winn et al., 2011 ▸). Model building and refinement were performed using Phenix (Liebschner et al., 2019 ▸) and Coot (Emsley et al., 2010 ▸). The final refined model consisting of residues Asp126– sp398 converged with R _work and R _free values of 22.7% and 26.2%, respectively, and showed an excellent Ramachandran plot with no residues in the disallowed regions.

2.5. Circular dichroism measurements  

BibA_34–400 protein in SEC buffer was diluted eightfold with 10 mM Tris–HCl pH 7.5, and the CD spectra were measured in a quartz cell with an optical path length of 0.1 cm on a JASCO J-815 Circular Dichroism Spectrometer (JASCO, USA). The final concentration of the protein was 0.15 mg ml⁻¹ and the spectra were obtained from three scans measured from 240 to 190 nm. The initial two scans were made at 20°C and 40°C, and the sample was then denatured at 80°C and cooled slowly to 20°C. The third scan was recorded from this cooled sample. The CD spectrum of the SEC buffer (blank) was measured in the same manner and was subtracted from the collected data. The input units of measured ellipticity (mdeg) were converted to Δ∊ and the secondary-structure analysis was performed using the BeStSel web server (Micsonai et al., 2015 ▸; Fig. 2 ▸). The normalized root-mean-square deviation (NRMSD) values of all the samples were 0.01 or less. The atomic coordinates from the BibA_126–400 crystal structure were used with the SP175 (soluble) reference database to generate CD spectra (Fig. 2 ▸ a) using an empirical-based approach with the online PDB2CD web server (Mavridis & Janes, 2017 ▸). The BibA_126–400 CD spectra were used as input for the BeStSel web server for comparison with the secondary-structure analysis performed with the CD spectra of the BibA_34–400 protein (Fig. 2 ▸ b).

Figure 2.

(a) Circular dichroism spectra of freshly prepared and intact BibA_34–400 sample at 20°C (red), at 40°C (orange) and denatured at 80°C followed by cooling to 20°C (green), the standard deviation error between the BibA_34–400 samples (gray dotted line) and the spectrum generated from the BibA_126–398 crystal structure (black). Two negative peaks at 208 and 222 nm typical of a α-helical secondary structure were observed for the BibA_34–400 sample. (b) Estimated secondary-structure content (%) of BibA_34–400 sample in solution (top) and the spectrum generated from the BibA_126–398 crystal structure (bottom).

2.6. Small-angle X-ray scattering (SAXS) measurements and data analysis  

SAXS data sets were collected for BibA_34–400 at BioCAT (beamline 18ID at the APS) using inline size-exclusion chromatography (SEC-SAXS). This method ensures optimal sample quality by separating the sample from aggregates and other contaminants. The sample was dialyzed against 25 mM Tris pH 8.0, 75 mM NaCl, 10%(v/v) glycerol and concentrated to 16 mg ml⁻¹ before loading 700 µl onto a Superdex 200 Increase 10/300 GL column. The sample was run at 0.7 ml min⁻¹ on an ÄKTA pure FPLC (GE Healthcare Life Sciences) and the eluate passed through the UV monitor flowed through the SAXS flow cell, which is modeled after a sheath-flow cell (the coflow method; Kirby et al., 2016 ▸). Scattering intensity was recorded using a PILATUS3 1M detector (Dectris), and a q-range of ∼0.004 to ∼0.36 Å⁻¹ was accessed by placing the detector 3.64 m from the sample. Data sets were acquired with 0.5 s exposures at 2 s intervals during elution and data were reduced using BioXTAS RAW 1.6.0 (Hopkins et al., 2017 ▸). Buffer blanks were created by averaging regions flanking the elution peak and were subtracted from exposures selected from the elution peak to create the I(q) versus q curves used in subsequent analyses. Evolving factor analysis was used to extract SAXS curves for single components (Hopkins et al., 2017 ▸). Guinier fit and molecular-weight analysis was performed using BioXTAS RAW 1.6.0 (Hopkins et al., 2017 ▸) and P(r) distributions were calculated using GNOM (Svergun, 1992 ▸) in ATSAS (Franke et al., 2017 ▸) through the BioXTAS RAW interface.

2.7. Ab initio and ensemble modeling  

Ab initio bead models (also known as dummy-atom models) were calculated using DAMMIF (Franke & Svergun, 2009 ▸). The bead models were averaged by DAMAVER (Volkov & Svergun, 2003 ▸) to build a consensus shape and produce a final model. The crystallographic structure of BibA_126–398 was fitted into the low-resolution SAXS model using SUPCOMB (Kozin & Svergun, 2001 ▸) and UCSF Chimera (Pettersen et al., 2004 ▸). An initial model for the missing N-terminal residues was generated with the help of the MR1 and MR2 motif repeats of the BibA_126–398 crystal structure, which were used as a template in the I-TASSER online structure-prediction server (Zhang, 2008 ▸) to generate the full-length BibA_34–400 model. The final model was idealized and refined using Coot. This final BibA_34–400 model was fitted into the SAXS envelope using UCSF Chimera. Theoretical profiles for the atomic coordinates of the full-length BibA_34–400 model were calculated using FoXS (Schneidman-Duhovny et al., 2013 ▸, 2016 ▸), which is available through UCSF Chimera.

2.8. Antibodies  

The polyclonal rabbit anti-human C4BP 9008 was a kind gift from Professor B. Dahlbäck (Lund University, Sweden). The 9008 antibody recognizes all of the forms of C4BP (plasma purified C4BP–Protein S complex, rC4BP and mutants) equally well (Avirutnan et al., 2011 ▸). Antibodies were purified using a Protein A Sepharose column. Peroxidase-conjugated swine anti-rabbit and rabbit anti-mouse IgG were obtained from Dako (USA) and the bovine serum albumin (BSA) used as a negative control was purchased from AppliChem (USA).

2.9. Direct binding assays  

Microtiter plates were coated with BibA (5 µg ml⁻¹) in PBS overnight at 4°C. Wells coated with BSA were used as a control. The plates were washed with 50 mM Tris–HCl pH 8.0, 150 mM NaCl, 0.1% Tween-20 between each step. The plates were then blocked with 250 µl blocking solution [50 mM Tris–HCl pH 8.0, 150 mM NaCl, 0.1% Tween-20, 3% fish gelatin (Norland)] for 2 h at room temperature. After blocking, the plate was incubated with increasing concentrations of plasma-purified C4BP–PS or rC4BP for binding to BibA in binding buffer (50 mM HEPES pH 7.4, 150 mM NaCl, 2 mM CaCl₂, 50 µg ml⁻¹ BSA) and was incubated for 90 min at room temperature. To assess the binding of various forms of C4BP (wild-type rC4BP and mutants lacking individual domains) to BibA, 30 µg ml⁻¹ of each C4BP variant was incubated overnight at 4°C with BibA immobilized at 5 µg ml⁻¹. Bound C4BP was detected using a specific polyclonal Abs (9008), followed by a swine anti-rabbit peroxidase-conjugated Abs. The plates were developed with o-phenyl­enediamine (Dako, USA) substrate and H₂O₂, and the absorbance at 490 nm was measured (Varian Cary 50 MPR Microplate Reader).

2.10. Surface plasmon resonance analysis  

To measure the kinetics of rC4BP binding to BibA, surface plasmon resonance analysis was performed using a Biacore 2000 (GE Healthcare, USA). rC4BP was diluted to 10 µg ml⁻¹ in 10 mM sodium acetate pH 4.0 and was immobilized on the surface of a CM5 sensor chip to reach 3000 response units (RU). All experiments were performed at a continuous flow rate of 30 µl min⁻¹ using Biacore buffer (150 mM NaCl, 10 mM HEPES, 2.5 mM CaCl₂, 0.002% Tween-20 pH 7.4). The analyte, BibA_34–400, was injected in a concentration gradient from 1.75 to 56 nM followed by two consecutive injections of 2 M NaCl and 100 mM HCl followed by 4 M MgCl₂ and 0.05% SDS for regeneration. The obtained sensorgrams were analyzed using the Bio-evaluation software version 3.0 (Biacore) using a 1:1 Langmuir binding model of interaction.

2.11. Other software  

Figures were generated using PyMOL (version 1.8; Schrödinger; http://www.pymol.org) and UCSF Chimera.

3. Results and discussion  

3.1. BibA_126–398 crystal structure  

Concentrated BibA_34–400 protein (Fig. 3 ▸ a) was subjected to initial crystallization trials, but the sample was very susceptible to proteolysis. Several attempts were made to produce an intact protein by modifying the purification protocols and using various protease inhibitors. Based on the N-terminal sequencing results of the proteolytically digested protein sample (data not shown), multiple mutations were introduced. The residues at positions 49, 70 and 112 were mutated to asparagine, delaying the breakdown of the protein (Fig. 3 ▸ b). This mutated BibA_34–400 sample with selenomethionine substitutions was concentrated to 40.8 mg ml⁻¹ and subjected to crystallization screening. Single good-quality crystals only appeared in the drops after several months (Fig. 3 ▸ c). N-terminal sequencing was performed on the crystals and revealed the loss of 92 amino-acid residues (His34–Ser125).

Figure 3.

(a) A schematic of GBS BibA_34–400 with N-terminal His tag and TEV protease cleavage site. (b) SDS–PAGE gel showing the breakdown of the BibA_34–400 sample. The marker, BibA_34–400 without any mutations and BibA_34–400 with K49N, K70N and K112N mutations are labeled M, 1 and 2, respectively. (c) Picture of a BibA_126–398 crystal formed after proteolysis.

GBS BibA lacking the N-terminal 92 residues (that is, BibA_126–400) crystallized in space group P2₁22₁ with one molecule in the asymmetric unit. The crystal structure of BibA_126–400 was solved by the selenomethionine multiwavelength anomalous diffraction (MAD) method to a resolution of 3.03 Å. The experimentally phased electron-density maps of BibA_126–400 allowed the building of a 273-amino-acid single polypeptide chain (Asp126–Asp398). The Lys399 and Lys400 residues on the C-terminus could not be built, so the visible fragment of the crystal structure is referred to as BibA_126–398. Compared with other similar resolution structures deposited in the Protein Data Bank, the final refined model has favorable crystallographic refinement parameters. The data-collection, processing and refinement statistics are summarized in Table 3 ▸.

Table 3. Data collection, processing and refinement.

Values in parentheses are for the highest resolution shell

Diffraction source	SER-CAT 22-ID, APS
Wavelength (Å)	0.9757
Temperature (K)	100
Space group	P21221
a, b, c (Å)	26.42, 68.81, 199.38
α, β, γ (°)	90, 90, 90
Mosaicity (°)	0.25
Resolution range (Å)	68.81–3.03 (3.14–3.03)
Total No. of reflections	15398 (1474)
No. of unique reflections	7702 (737)
Completeness (%)	100.0 (100.0)
Multiplicity	2.0 (2.0)
〈I/σ(I)〉	16.10 (3.46)
R merge (%)	2.81 (18.93)
R meas (%)	3.98 (26.78)
CC1/2	0.999 (0.905)
Overall B factor from Wilson plot (Å2)	78.73
Refinement statistics
Resolution range (Å)	47.8–3.03 (3.14–3.03)
Completeness (%)	100.0 (100.0)
No. of reflections, working set	7699 (737)
No. of reflections, test set	462 (44)
Final R work (%)	22.7 (46.1)
Final R free (%)	26.2 (40.6)
No. of non-H atoms	2158
No. of protein residues	273
R.m.s. deviations
Bond lengths (Å)	0.011
Angles (°)	1.09
Average B factors (Å2)
Non-H atoms	68.2
Ramachandran plot
Most favored (%)	96.7
Allowed (%)	3.3
Outliers (%)	0
PDB entry	6poo

The overall structure of BibA_126–398 has an ∼150 Å-long distorted rod shape composed of nine α-helices, a 3₁₀-helix and small loops connecting the helices. Interestingly, the nine α-helices adopt a unique pattern to form four antiparallel three-helix-bundle-motif repeats labeled MR1 (Asp126–Ser182), MR2 (Thr183–Lys235), MR3 (Glu236–Leu303) and MR4 (Gln304–Asp398) (Fig. 4 ▸ a). Topologically, the N-terminus first assembles into an 18-residue α-helix (α1; Asp126–Ser142) and a small loop (L1; Leu143–Ser147), followed by the second α-helix (α2; Ser148–Ser160), which runs antiparallel to helix α1. Loop 2 (L2; Ser161–Asp163) proceeds α2, allowing the third α-helix (α3; Ser164–Leu195) to run antiparallel to α2. Among the first three helices, α3 is more extended, comprising of 33 residues, of which residues Ser164–Ser182 are involved in forming the MR1 motif, while residues Thr183–Leu195 extend into the MR2 motif. This shared-helix pattern to form the motif is repeated to generate motifs MR3 and MR4, which involve the remaining α-helices (α4–α9) and loops (L3–L8). The C-terminal MR4 motif is composed of a longer stretch of residues (94) compared with MR1 (57 residues), MR2 (53 residues) and MR3 (67 residues). The α7 helix that is shared by MR3 and MR4 is the longest, with 54 residues (Ile281–Lys331). The longer MR4 motif harbors an additional 3₁₀-helix (Lys333–Ile335) between the α7 helix and loop 7 (L7), and an extended loop (L8, residues Leu357–Trp365) that is distinct from the other motifs. Helices α3, α5 and α7 bridge the three-helix-bundle motifs together to form an extended rod with a curved shape. This curvature in the protein perhaps assists target proteins such as C4BP by increasing its inter­action surface (Youn et al., 2017 ▸).

Figure 4.

(a) Cartoon representation of the BibA_126–398 crystal structure shown as a rainbow model with color changing gradually from the N-­terminus (blue) to the C-terminus (red). The length of the rod-shaped structure is depicted, and the antiparallel three-helix-bundle-motif repeats are labeled from the N-­terminus to the C-terminus as MR1–MR4, respectively. (b) Superposition of the antiparallel three-helix-bundle-motif repeats MR1 (dark blue), MR2 (cyan) and MR3 (forest green). Similar residues such as Leu and Ile are boxed and shown in a red font and the conserved serine residue is highlighted in red.

Although there was no significant sequence conservation between the three antiparallel three-helix-bundle-motif repeats MR1, MR2 and MR3, their structural motifs superimpose with r.m.s.d. values of 0.818 Å (MR1 and MR2) and 0.996 Å (MR1 and MR3). The loops L1 (Leu143–Ser147), L3 (Ile196–Ser200) and L5 (Lys251–Ser256) have a conserved serine residue just before their respective succeeding helices and structurally align with each other (Fig. 4 ▸ b).

3.2. Circular dichroism spectra of BibA_34–400  

Crystallization attempts using the BibA_34–400 protein always produced crystals lacking the N-terminal 92 residues, leading us to believe that this segment might be disordered. The X-ray crystal structure analysis of BibA_126–398 indicates a highly helical protein. So, to evaluate the structure of full-length BibA_34–400, the freshly purified protein was analyzed by CD spectrometry. It showed a typical spectrum for an α-helical secondary structure, with two negative peaks at 208 and 222 nm (Fig. 2 ▸ a). Estimation of the protein secondary structure from the CD spectrum using BeStSel (Micsonai et al., 2015 ▸) revealed a composition of about 80% α-helix, with about 20% categorized as ‘others’ (Fig. 2 ▸ b). This composition is similar to that of the BibA_126–398 crystal structure. Further, to validate the secondary-structure analysis of BibA_34–400 by BeStSel, we generated CD spectra with the help of the PDB2CD web server (Mavridis & Janes, 2017 ▸) using the atomic coordinates of the BibA_126–398 crystal structure (Fig. 2 ▸ a) and performed secondary-structure analysis with the help of the BeStSel web server (Fig. 2 ▸ b). The secondary-structure content estimated from the generated CD spectrum of BibA_126–398 was similar to that from the experimental CD spectrum of BibA_34–400, and based on this secondary-structure content analysis we suggest that the N-terminal 92-residue segment (His34–Ser125) is α-helical in nature and might adopt a conformation similar to the rest of the molecule, forming antiparallel three-helix-bundle-motif repeats.

3.3. Small-angle X-ray scattering (SAXS) envelope and the BibA_34–400 model  

To gain further insights into the structure of BibA_34–400, freshly purified and intact recombinant BibA_34–400 was analyzed by SEC-SAXS. The pair distance distribution function P(r), calculated using GNOM, suggests that the protein is an elongated cylinder in shape (Kajiwara & Hiragi, 1996 ▸; Putnam et al., 2007 ▸), as the asymmetric curve has a gradual slope after the peak (Fig. 5 ▸ a). As inferred from the CD spectra and SAXS analysis, the initial model for full-length BibA_34–400 was generated using Coot by mimicking the MR1 and MR2 motifs for the missing N-terminal residues. After mutating the residues to the original sequence of the protein, the model from Coot was used as a template in I-TASSER to generate the full BibA_34–400 model. The final refined model predicted is ∼200 Å long and contains 367 amino acids (His34–Lys400) with six antiparallel three-helix-bundle-motif repeats. The N-terminal motif repeats were labeled MR-N1 and MR-N2 (Fig. 5 ▸ c). Further, we have calculated theoretical profiles for the BibA_34–400 model using FoXS (Schneidman-Duhovny et al., 2013 ▸, 2016 ▸) and compared them with the experimental scattering profile of BibA_34–400 in solution (Fig. 5 ▸ b). The profiles fit reasonably, with a χ² of 2.251.86, possibly supporting the structure of the predicted BibA_34–400 model. Although the scaling of the atomic radius (c ₁), which is 1.05, is within the reasonable range (0.99 ≤ c ₁ ≤ 1.05), the parameter controlling the density of the water layer around the molecule (c ₂) is higher than the maximum value (−2.0 ≤ c ₂ ≤ 4.0), suggesting possible overfitting of the theoretical profile of the proposed BibA_34–400 model in the experimental scattering profile.

Figure 5.

(a) Pair distance distribution [P(r)] function of intact BibA_34–400 protein. (b) Comparison of the experimental scattering profile (in blue) for BibA_34–400 with profiles from a theoretical model (FoXS; green) derived from the proposed BibA_34–400 model. (c) Fit of the crystal structure of BibA_126–398 (cyan) and the proposed BibA_34–400 (magenta) into the ab initio model of BibA_34–400 calculated with DAMMIF.

The ab initio envelope calculated for the full-length protein (BibA_34–400) using DAMMIF was an elongated and distorted rod-like shape of ∼200 Å in length. This ab initio envelope fits well with the experimental data, with a χ² of 1.084. The crystal structure of BibA_126–398 was fitted using UCSF Chimera and, as expected, it sits well in the ab initio envelope with an empty pocket on the N-terminal side of the pocket (Fig. 5 ▸ c). By comparison, the predicted BibA_34–400 hybrid model fits very well in the SAXS-derived envelope, with 2756 of 2879 atoms fittied inside the contour, agreeing reasonably with the proposed model (Fig. 5 ▸ c).

3.4. Binding-study results  

The binding of plasma-purified C4BP (Fig. 6 ▸ a) and rC4BP (Fig. 6 ▸ b) to immobilized BibA was measured, and the K _d was estimated to be 28.99 nM, suggesting moderate to tight binding of the C4BP α-chain and immobilized BibA_34–400. However, the binding-affinity estimation for C4BP and BibA (Fig. 6 ▸ a) is not accurate compared with that estimated for BibA and rC4BP, and we present this data only as an estimation that is useful for understanding the binding mode. Both proteins bind BibA_34–400 in a relatively similar range, indicating that the binding site is probably in the α-chain and not the β-chain. The interaction affinity between rC4BP immobilized in dextran matrix and BibA in solution analyzed by surface plasmon resonance (Fig. 6 ▸ c), in a similar range of direct binding, results in a K _d of 1.35 nM (k _a = 2.6 × 10⁵  M ⁻¹ s⁻¹ and k _d = 3.5 × 10⁻⁴ s⁻¹).

Figure 6.

Microtiter plates were coated with BibA (5 µg ml⁻¹) and increasing amounts of plasma-purified C4BP (a) or rC4BP (b) were added. Binding was detected using specific polyclonal anti-C4BP antibodies. BSA was used as a negative control. The mean ± SD of three independent experiments performed in duplicate is presented. Statistical significance was calculated using a two-way Anova test. ***, p < 0.001. (c) Surface plasmon resonance analysis to measure the binding of BibA to immobilized C4BP using Biacore. (d) The C4BP variants were allowed to bind to BibA immobilized on a plate. Bound C4BP was detected with polyclonal Abs. The graph represents data from three independent experiments performed in duplicate ± SD. Statistical significance was calculated using a one-way Anova test. *, p < 0.05; ***, p <0.001.

We have localized the binding of BibA_34–400 to the C4BP α-chain, and further observed that the ΔCCP1 mutant exhibited decreased binding (Fig. 6 ▸ d). There appears to be a second BibA binding site in CCP7–8 indicated by the decrease in binding shown by the ΔCCP7 and ΔCCP8 mutants. The CCP7–8 domains, however, may be less accessible in vivo considering that they are close to the crowded C-terminal oligomerization region of C4BP, which is also in proximity to Protein S bound to the α-chain of C4BP. This hypothesis is supported by the finding that rC4BP lacking the β-chain and Protein S binds more strongly to BibA than plasma-purified protein carrying these additional components. Therefore, we focused our structural effort on the CCP1–2 region and its interactions with BibA_34–399. However, we were not able to purify the BibA_34–399–CCP1–2 complex despite trying multiple buffers and salt concentrations.

4. Conclusions  

The crystal structure of BibA_126–398 was solved by the MAD phasing method. We could not determine the full-length BibA_34–400 crystal structure owing to protein degradation during purification and crystallization. Suspecting the N-terminal BibA_34–125 to be flexible or disordered, we introduced mutations into the BibA_34–400 constructs, which slowed down the degradation during purification (Fig. 3 ▸ b). However, the single crystals obtained were of BibA_126–398, and the crystal structure of this C-terminal segment is predominantly α-helical, with four sequentially connected three-helical bundles forming a slightly curved rod-shaped molecule.

We pursued circular dichroism spectrometry and small-angle X-ray scattering analysis to gain insights into the intact BibA_34–400 protein. The CD spectra of BibA_34–400 recorded at 20 and 40°C and the CD spectrum of denatured and refolded sample recorded at 20°C suggested the structure to be 80% α-helical, similar to the content observed in the crystal structure of BibA_126–398. We generated a model for the full-length BibA_34–400 structure by combining an N-terminal segment ab initio model, which was built assuming it to be similarly composed of two three-helical bundle motifs (MR-N1 and MR-N2), and the C-terminal BibA_126–398 crystal structure. This hybrid model fits well in the ab initio SAXS envelope. However, emphasizing that the N-terminal segment ab initio model generated by us may be biased, we could safely suggest that the full-length BibA_34–400 is a rod-shaped molecule and is predominantly α-helical in composition.

The overall structure of the N-terminal domain of BibA is novel and provides insight into the structure of a bacterial component involved in complement inhibition. The binding site of BibA to C4BP is located in the α-chain, with two potential interaction sites in CCP1 and CCP8. Currently, we have not produced a BibA–C4BP complex in solution, and we were not able to identify the precise region of BibA binding to C4BP. Moving forward, we will attempt to generate a crystal or solution structure of the BibA–C4BP complex, which may help in developing strategies for fighting GBS infections.

The amino-acid sequence of BibA is unique, and a non­redundant GenBank database BLAST search (Altschul et al., 1997 ▸) did not reveal any significant homologies. Nevertheless, on searching for three-helix-bundle proteins similar to the BibA_34–400 model, we found a host immune response-modulating surface protein from Staphylococcus aureus called protein A (Youn et al., 2017 ▸), which has 22% sequence identity to BibA. Protein A has antiparallel three-helix-bundle motif repeats and its secondary-structure content is 80% helical and turns, similar to BibA. However, protein A (PDB entry 5h7b) is a 237-residue polypeptide with five repeat modules, compared with the 273-residue BibA_126–398 protein, which is made of four repeat modules. Hence, the repeat modules of protein A are smaller in size and do not superimpose on the BibA structure and its structural motifs.

The GBS BibA functional homologs in group A streptococcus (GAS) are the surface-anchored virulence factor M proteins (Ghosh, 2011 ▸; McNamara et al., 2008 ▸). These proteins bind to C4BP (Persson et al., 2006 ▸) when recruited on the surface of GAS, and the binding has been localized to the N-terminal 50 residues of M proteins (hypervariable region, HVR; Lannergård et al., 2011 ▸; Penfound et al., 2010 ▸; Sandin et al., 2006 ▸). The crystal structure of C4BP α-chain CCP1 and CCP2 in complex with hypervariable regions of four different M proteins (M2, M22, M28 and M49) helped to identify conserved sequence patterns within their HVRs (Buffalo et al., 2016 ▸). However, we were unable to identify a similar pattern in BibA, suggesting that the interactions of C4BP with GBS surface proteins differ from those with GAS surface proteins.

Supplementary Material

PDB reference: BibA, 6poo

Funding Statement

This work was funded by National Institute of Allergy and Infectious Diseases grant R01-AI106808 to Sthanam Narayana. National Institute of Dental and Craniofacial Research grants R01-DE025015 and R01-DE017382.