Calcium Binds to Leptospiral Immunoglobulin-like Protein, LigB, and Modulates Fibronectin Binding*

Pathogenic Leptospira spp. express immunoglobulin-like proteins, LigA and LigB, which serve as adhesins to bind to extracellular matrices and mediate their attachment on host cells. However, nothing is known about the mechanism by which these proteins are involved in pathogenesis. We demonstrate that LigBCen2 binds Ca2+, as evidenced by inductively coupled plasma optical emission spectrometry, energy dispersive spectrometry, 45Ca overlay, and mass spectrometry, although there is no known motif for Ca2+ binding. LigBCen2 binds four Ca2+ as determined by matrix-assisted laser desorption ionization time-of-flight mass spectrometry. The dissociation constant, KD, for Ca2+ binding is 7 μm, as measured by isothermal titration calorimetry and calcium competition experiments. The nature of the Ca2+-binding site in LigB is possibly similar to that seen in the βγ-crystallin superfamily, since structurally, both families of proteins possess the Greek key type fold. The conformation of LigBCen2 was significantly influenced by Ca2+ binding as shown by far- and near-UV CD and by fluorescence spectroscopy. In the apo form, the protein appears to be partially unfolded, as seen in the far-UV CD spectrum, and upon Ca2+ binding, the protein acquires significant β-sheet conformation. Ca2+ binding stabilizes the protein as monitored by thermal unfolding by CD (50.7–54.8 °C) and by differential scanning calorimetry (50.0–55.7 °C). Ca2+ significantly assists the binding of LigBCen2 to the N-terminal domain of fibronectin and perturbs the secondary structure, suggesting the involvement of Ca2+ in adhesion. We demonstrate that LigB is a novel bacterial Ca2+-binding protein and suggest that Ca2+ binding plays a pivotal role in the pathogenesis of leptospirosis.

Lig proteins, which include LigA and LigB, possess bacterial immunoglobulin-like (BIg) 3 domains with 90-amino acid tandem repeats. Both proteins have identical N-terminal sequences of 630 amino acids, but their C termini are variable (11)(12)(13). LigB also encodes a C-terminal, nonrepeat domain with 771 amino acid residues (11,12). LigA and LigB may serve as microbial surface components recognizing adhesive matrix molecules that allow pathogenic Leptospira to bind to host extracellular matrix components, such as fibronectin (Fn), fibrinogen, laminin, and collagen (14 -16). Lig proteins may also serve as possible vaccine candidates and/or as diagnostic antigens (12,17,18), and their expression is regulated by osmolarity (19). A high affinity Fn binding region of LigB, designated LigBCen2, contains 152 amino acids that include part of an immunoglobulin-like domain and a nonrepeated region (15) (Fig. 1A).
Calcium plays a pivotal role in bacterial physiological activities, such as cell cycle, cell division (20), competence (21), pathogenesis (22), signal transduction (23), and motility and chemotaxis (24,25). Apart from these functions, it is also known that host-pathogen interactions of some bacteria are affected by calcium (26,27). Several types of Ca 2ϩ -binding motifs in bacterial proteins have been identified, which include EF-hand motif (28), leukotoxin (29) or hemolysin-type calcium-binding domain (30), and orphan motifs in which oxygen atoms provided by several charged glutamate or aspartate residues are used in ligation (28). It appears that Lig proteins do not have any of these known Ca 2ϩ -binding motifs. LigBCen2 shows sequence similarity to the c-type lectin domain of other adhesins, including invasin of Yersinia pseudotuberculosis and intimin of Escherichia coli (12) (Fig. 1B).
Although Lig proteins are likely to play a significant role in pathogenicity, little is known about the mechanisms of action of these proteins. We undertook this study to identify novel properties of Lig proteins. Based on the structural homology of immunoglobulin-like fold with lens ␤␥-crystallin type Greek key motif (31), we wondered if Lig proteins would bind Ca 2ϩ (32). We therefore performed these studies and report that Ca 2ϩ binds to the high affinity Fn binding region of LigB, LigBCen2. Ca 2ϩ binding increases the stability of LigBCen2 and significantly influences its conformation. Further, we demonstrate that Ca 2ϩ modulates the binding of LigB to N-terminal domain (NTD) of Fn, suggesting that it plays a major role in bacterial infection.

MATERIALS AND METHODS
Reagents and Antibodies-Calcium Green TM -1 was obtained from Molecular Probe (Eugene, OR). Fibronectin (human plasma fibronectin), NTD of Fn, EGTA, Stains-all, sodium chloride, Tris, and calcium chloride were from Sigma.
Plasmid Construction, Protein Purification, and Decalcification-The construct for the expression of histidine tag fused with a LigBCen2 DNA fragment (amino acids 1014 -1165) was generated using the vector pQE30 (Qiagen, Alencia, CA). Construction, expression, and purification procedures were as previously described (15,18). Protein was decalcified with 3 mM EDTA incubation for 45-60 min followed by buffer exchange with Chelex-100 resin-treated Tris buffer. All buffer solutions used for Ca 2ϩ -binding studies were passed through Chelex-100 resin and stored in plasticware.
Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES)-Standard analysis procedures were performed by the Soil and Plant Laboratory (Cornell University) to analyze the total mineral and heavy metal content by ICP-OES (Varian, Inc., Palo Alto, CA). Protein samples included 25 mM Tris buffer, pH 7.0, which contained 150 mM NaCl and 70 M of LigBCen2 either untreated or treated with 1 mM CaCl 2 or 5 mM EGTA, for 1 h at room temperature. Unbound calcium and EGTA were removed to trace levels (Ͻ0.1 g/ml) via extensive dialysis against Chelex-100-treated Tris buffer.
MALDI-TOF Mass Spectrometry-The molecular weight of purified LigBCen2 protein was analyzed using an Applied Biosystem 4700 mass spectrometer (Applied Biosystems, Foster City, CA). Protein samples (70 M LigBCen2) were incubated with 1 mM CaCl 2 , pH 7, for 1 h at room temperature. Unbound excess calcium was removed by dialysis against deionized water.
Energy-dispersive Spectrometry (EDS)-EDS analyses were performed with a JEOL 8900 electron probe microanalyzer (JEOL, Ltd. Tokyo, Japan). The operating conditions were 10-kV acceleration voltages. Sample preparation is described under "MALDI-TOF Mass Spectrometry." Aqueous solutions of proteins were lyophilized, and a powder form of each sample was analyzed by EDS.
Isothermal Titration Calorimetry (ITC)-The reaction enthalpy measurements were carried out with a VP-ITC calorimeter (MicroCal Inc.) at 30°C. Before the ITC experiment, residual Ca 2ϩ was removed from the protein by EDTA incubation, followed by taking EDTA off with Chelex-100 resintreated Tris pH 7 buffer containing 50 mM KCl. In a typical ITC experiment, the cell contained 1.39 ml of thoroughly degassed 75 M protein in 50 mM Tris (pH 7) containing 50 mM KCl, titrated against the same buffer containing 7 mM CaCl 2 . Following thermal equilibration, titrant was added to the 1.39-ml sample at an initial delay of 60 s, and sample in the cell was stirred at 300 rpm by a syringe. After 50 injections, protein was saturated with Ca 2ϩ until there was no further heat change. In another experiment with the same parameters, Mg 2ϩ versus LigBCen2 titration was performed, and after 50 injections the same protein was used for Ca 2ϩ titration. Titration was performed as follows: 50 injections of 3 l of ligand delivered over 4 s with a 220-s spacing between injections to allow complete equilibrium. Titration of Tris buffer with ligand was performed with the same parameters as mentioned above, and these reference data (heat of dilution) were subtracted with standard. Data were analyzed using MicroCal LLC ITC software (MicroCal), fitting them to an independent binding model.
Differential Scanning Calorimetry (DSC)-Excess heat capacity C p (T) of the apo and holo form of LigBCen2 was measured using a DSC Q1000 microcalorimeter (Waters, New Castle, DE). Degassed samples containing 3 M LigBCen2 with and without 1 mM CaCl 2 in Tris buffer (pH 7.0) were heated at a 10 K/h scan rate. C p (T) data were recorded, corrected for buffer base line, and normalized to the amount of the sample. The TA Universal Analysis software (Waters, New Castle, DE) was used for the data analysis and display.
CD Spectrometry-CD spectra were recorded on a Jasco J-815 spectropolarimeter under N 2 atmosphere at room temperature (25°C) in a 0.02-and 0.5-cm path length quartz cell for far-and near-UV CD spectra, respectively, with eight accumulations. Aliquots of calcium chloride solution were added to the protein solution and incubated for 5 min. All spectra were recorded in 50 mM Tris buffer, pH 7, containing 50 mM KCl. In thermal unfolding experiments, 2 M LigBCen2 in the absence and presence of 1 mM CaCl 2 was subjected to thermal unfolding, and data were collected at 1°C/min increments from 25 to 70°C recording the ellipticity at 215 nm, with 30-s temperature equilibrations, followed by 30-s data averaging. In order to measure the melting point, a first order derivative was applied to the results from the melting experiment. In all CD experiments, the background spectrum of Tris buffer (pH 7.0) without protein was subtracted from the protein spectra.
The Stains-all binding assay was performed essentially as described earlier (32,33). LigBCen2 (5 M) was mixed with the dye solution (60 -100 M) in 2 mM MOPS, pH 7.2, containing 30% ethylene glycol, and incubated for 5 min, and CD spectra were recorded from 400 -700 nm. For studying the interaction, aliquots of NTD of Fn were mixed to the LigBCen2-Stains-all complex in the presence or absence of Ca 2ϩ and CD spectra recorded.
Fluorescence Spectrometry-Fluorescence emission spectra were measured on a Hitachi F-4500 spectrofluorometer (Hitachi, Tokyo, Japan). All spectra were recorded in correct spectrum mode of the instrument with excitation and emission band passes of 5 nm each. The intrinsic Trp fluorescence of the protein was recorded by exciting the solution at 295 nm and measuring the emission in the 300 -400 nm regions. For calcium or magnesium titration, 0.1, 0.3, 0.5, 0.8, or 1.0 mM of calcium chloride was mixed with 10 M of LigBCen2 in 50 mM Tris buffer, pH 7.2, containing 50 mM KCl, and spectra were recorded after 3 min of incubation.
ANS fluorescence was measured by adding ANS to a final concentration of 100 M to the protein solution (18 M) and incubated for 5 min, and spectra were recorded between 400 and 600 nm at an excitation wavelength of 365 nm. Fibronectin binding to LigBCen2 was assayed by measuring the change in Trp fluorescence upon the addition of aliquots of NTD of Fn in the presence of 100 M CaCl 2 or in the absence of calcium chloride (in the presence of 100 M EGTA). The mixture was incubated for 5 min before recording the emission spectra at the excitation of 295 nm.
In a calcium competition experiment, 20 M Calcium Green TM -1 was mixed with or without 30 M LigBCen2, and 100 l/well was dispensed into a microtiter plate and incubated with 0.046, 0.093, 0.187, 0.375, 0.75, 1.5, or 3 M CaCl 2 in Tris buffer, pH 7.0, for 5 min. Enhanced fluorescence due to binding of free calcium and Calcium Green TM -1 was monitored at 528 nm with excitation at 485 nm using a Synergy TM HT multidetection microplate reader (Bio-TEK Instruments, Inc. Winooski, VT). The association constant, K a , was deduced using a Scatchard plot, and rearrangement of the Chang-Prusoff equation was used to calculate the dissociation constant (K D ) of LigBCen2 and calcium (34).
K app is the apparent dissociation constant measured in the presence of a concentration of LigBCen2 in Calcium Green TM -1 solution with calcium; K D(dye) or K D(LigBCen2) is the K D of calcium with Calcium Green TM -1 or LigBCen2, respectively. All spectra were recorded in the correct spectrum mode excitation and emission band passes of 5 nm each and corrected for volume changes before further analysis. All measurements were performed at 25°C. Fibronectin Binding Assay by ELISA-To determine the binding of LigBCen2 to NTD of Fn in the presence and absence of Ca 2ϩ , serial concentrations of NTD or bovine serum albumin (negative reference and data not shown) were coated on microtiter plate wells and blocked subsequently as previously described (15). 10 nM in the absence of or 1 mM in the presence of CaCl 2 , GST-LigBCen2, or GST in 100 l of Tris buffer (pH 7.0) was added onto microtiter plate wells and incubated for 1 h at 37°C.
To detect the binding of GST-LigBCen2 or GST, rabbit anti-GST (1:200) and horseradish peroxidase-conjugated goat anti-rabbit IgG (1:1,000) were used as primary and secondary antibody, respectively (15). After washing the plates three times with TBST, (0.05% Tween 20 in Tris buffer) 100 l of 3,3Ј,5,5Ј-tetramethylbenzidine (KPL, Gaithersburg, MD) was added to each well and incubated for 5 min. The reaction was stopped by adding 100 l of 0.5% hydrofluoric acid to each well. Each plate was read at 630 nm using an ELISA plate reader (Bioteck EL-312; Bio-TEK Instruments). Each value represents the mean Ϯ S.E. of three trials in triplicate samples. Statistically significant (p Ͻ 0.05) differences are indicated by an asterisk.
Statistical Analysis-Each data point represents the mean S.E. of sample tested in triplicate. Data were analyzed by Student's t test, and statistically significant differences were claimed at p values of Ͻ0.05.

The Sequence of LigBCen2 Is Similar to C-type Lectin Domains of Invasin and Intimin
Previously, Lig proteins, including LigA and LigB, have been shown to bind to extracellular matrix (14 -16). the central region of LigB, annotated as LigBCen2 (Fig. 1A) was selected for further functional studies (15). The sequence was selected from amino acids 1014 -1165 of LigB, as seen in Fig. 1B. Interestingly, this region contains the BIg domain and the nonrepeat region of 46 amino acids (15). Proteins that contain the BIg domain are found in a variety of bacterial and phage surface proteins, such as intimins or invasins, which are bacterial cell adhesion molecules that mediate intimate bacterial host-cell interaction (35,36). E. coli intimin contains three domains (two immunoglobulinlike domains and a C-type lectin-like module), implying that carbohydrate recognition may be important in intimin-mediated cell adhesion. However, the exact functions of these proteins, barring some preliminary studies on fibronectin interactions, are not known (15).

Rationale of Ca 2؉ Binding
As mentioned above, nothing is known about the function of Lig proteins except that they are thought to play a role in virulence or pathogenesis. We were interested in identifying the functions of these important proteins. Structurally, Lig proteins belong to the bacterial immunoglobulin fold or BIg fold. In one of the earlier studies, the structural and functional similarities between various proteins of the Greek key/immunoglobulin fold were assessed (31). Among the proteins selected for functional prediction, lens ␤␥-crystallins and immunoglobulin functions were chosen. Since both families of proteins possess the Greek key type fold (31), it prompted us to look for the function of Lig proteins in the context of lens ␤␥-crystallins. The exact function of lens ␤␥-crystallins is not known. However, we have shown earlier that ␤␥-crystallins belong to a different superfamily of low affinity Ca 2ϩ -binding proteins (37,38). Based on the fold similarities, we predicted that these Lig proteins might bind Ca 2ϩ . Therefore, we assessed Ca 2ϩ binding to LigBCen2 by a number of methods as described below.

Probing of Ca 2؉ Binding to LigBCen2 by ICP-OES, EDS, 45 Ca Overlay, and MALDI-TOF Mass Spectrometric Analysis
Since there is no known motif in LigBCen2 for Ca 2ϩ binding, it was necessary to probe Ca 2ϩ binding by a number of methods to examine the specificity of Ca 2ϩ binding. First, we probed Ca 2ϩ binding to LigBCen2 by ICP-OES. LigBCen2 in the presence or absence (in the presence of EGTA) of calcium chloride were applied to ICP-OES. As shown in Table 1, calcium was present only in calcium chloridetreated LigBCen2 and not in untreated or EGTA-treated LigB-Cen2. The results indicate that Ca 2ϩ binds to LigBCen2, since there was no Ca 2ϩ binding to EGTAtreated or untreated LigBCen2. To further confirm the Ca 2ϩ binding activity revealed by ICP-OES, Ca 2ϩ binding to LigBCen2 was assessed by EDS. As seen in Fig. 2, A and B, a prominent calcium signal was seen in Ca 2ϩ -bound LigBCen2 and not in the apo form of LigBCen2.
We also performed 45 Ca binding to LigBCen2 using a well known method of overlay (39). Seen as a dark spot on the membrane, radioactive calcium 45 Ca binds to LigB-Cen2, thereby confirming the specificity of Ca 2ϩ binding to the protein (data not shown). Further, the molecular mass of the holo form of LigBCen2 is higher (18,292 Da), as indicated by MALDI-TOF peaks, than that of the apo form of LigBCen2 (18,131 Da), further indicating that Ca 2ϩ binds to LigBCen2 (Fig. 2C). Since there was a 161-Da difference in the molecular mass between holo and apo forms, it is likely that at least four Ca 2ϩ molecules were bound to LigBCen2.

Stoichiometry of Ca 2؉ Binding to LigBCen2 by ITC
The above results of mass spectroscopy and 45 Ca binding experiments demonstrate that LigBCen2 is a Ca 2ϩ -binding protein. We next assessed the affinity and stoichiometry of the Ca 2ϩ binding to LigBCen2 by ITC. To quantitate the Ca 2ϩ binding affinity to LigBCen2, a titration of 7 mM calcium chloride with 75 M LigBCen2 was performed by ITC (Fig. 3A). The binding appeared to be an exothermic reaction with a favorable enthalpy (⌬H ϭ Ϫ9.9 kcal/mol) but is also associated with unfavorable entropy (⌬S ϭ Ϫ9.43 kcal/mol). As shown in Table 2, the K a for Ca 2ϩ binding to LigBCen2 was 1.32 ϫ 10 5 M Ϫ1 , and binding enthalpy was in the range of Ϫ10 kcal/mol. Fitting to one set of site macroscopic model data gives four binding sites for Ca 2ϩ with K D values of 7.5 M. Next, we examined if LigB-Cen2 also binds Mg 2ϩ . We did not see any change in molar heat in ITC (data not shown), suggesting that either Mg 2ϩ does not bind, or it binds very weakly to LigBCen2.

Ca 2؉ and Calcium Green Dye Competition Experiments
A calcium competition experiment using the dye Calcium Green TM -1 was performed to quantitatively confirm the binding of calcium to LigBCen2. When calcium chloride was added   -----LigBCen2 with EGTA  -------LigBCen2  -------a -, concentration below 0.1 g/ml. SEPTEMBER 12, 2008 • VOLUME 283 • NUMBER 37 either to the dye or dye mixed with LigBCen2, the fluorescence emission was significantly reduced, indicating that LigBCen2 was able to compete with the dye by binding calcium (Fig. 3B).

Calcium Binding to a Fibronectin-binding Domain of LigB
The binding data were plotted using Scatchard plot (Fig. 3C) and analyzed by rearrangement of the Chang-Prusoff equation.
The K D of calcium binding to LigBCen2 was calculated as 6 M.
The K D value thus obtained was close to that calculated from ITC (K D ϭ 7.5 M; Table 2).

Conformational Studies of Ca 2؉ Binding to LigBCen2
Ca 2ϩ Binding Influences the Secondary and Tertiary Structure Monitored by CD-In the far-UV CD spectrum of the apo form of LigBCen2, a broad negative peak at 215-208 nm is seen (Fig. 4A), suggesting that LigBCen2 is largely in a ␤-sheet conformation. However, the cross-over point is below 200 nm, suggesting that the protein is partially unfolded. The addition of Ca 2ϩ increases the negative value of ellipticity as well as red shifting the cross-over points, suggesting that upon binding Ca 2ϩ , the protein is folded and gains more ␤-sheet conformation (Fig. 4A).
Near-UV CD of LigBCen2 is shown in Fig. 4B. There are one Trp, five Tyr, and two Phe residues in this protein domain. In the near-UV CD, there are distinct bands (peaks) at 288 and 278 nm with a broad but strong peak at about 258 -266 nm. The strong ellipticity of this peak (258 -266 nm) suggests that Phe residues are either immobilized or interacting with neighboring residues. Two peaks at 288 and 292 nm represent the 1 L b bands of Trp, whereas the peak at 278 is due to Tyr. The addition of Ca 2ϩ brought significant changes in the near-UV CD spectrum (Fig. 4B). The ellipticity decreased over the entire range, a band for Trp transition appeared at about 302 nm, and some Trp bands (at 288 and 302) became negative in Ca 2ϩ -bound protein.
These data demonstrate that Ca 2ϩ binding imparts significant changes in the protein conformation.
Ca 2ϩ Binding Influences the Fluorescence Spectra of LigBCen2-Intrinsic Trp fluorescence of LigBCen2 is unusual in the sense that it shows a doublet or two peaks at 318 and 328 nm in its emission spectra (Fig. 4C). Maximum emission by Trp fluorescence is seen at 318 nm, suggesting that this single Trp is buried in a highly hydrophobic core of the protein. This unique phenomenon is rarely observed in proteins but has been documented in ␦-crystallin (40) and RNase P (41).
In the case of RNase P, it has been demonstrated that the emission doublet is due to the hydrophobic interaction of Trp with two Phe residues in the excited state (41). Therefore, in the case of LigBCen2, it appears that hydrophobic interaction among the lone Trp 60 and two Phe residues at positions 41 and 42 can form a complex when excited. When titrated with Ca 2ϩ , fluorescence intensity decreases significantly (10 -15% decrease at 318 nm), suggesting that upon binding Ca 2ϩ , there are conformational changes in the protein, and Trp moves toward a more nonpolar environment. The surface hydrophobicity of LigBCen2 was assessed by ANS binding. There was a weak fluorescence emission, suggesting that the protein is largely hydrophilic and binds ANS very weakly. In order to further characterize if there is any change in hydrophobicity upon Ca 2ϩ binding, the ANS-LigBCen2 complex was titrated by Ca 2ϩ .  Table 2.  There was an increase in ANS fluorescence intensity with a blue shift of about 4 -5 nm, suggesting that Ca 2ϩ binding changes the surface hydrophobicity of the protein (Fig. 4D).

Ca 2؉ Binding Increases the Conformational Stability of LigBCen2
In order to gain more insight into the possible role of Ca 2ϩ binding in protein stability, thermodynamic properties of thermal unfolding of LigBCen2, with or without calcium, were obtained by CD and DSC. The midpoint of unfolding was calculated by monitoring the change in ellipticity (in the CD) at 215 nm. As shown in Fig. 5A, the midpoint of LigBCen2 unfolding increased from 50.7 Ϯ 0.9 to 54.8 Ϯ 0.5°C when Ca 2ϩ was added. On the other hand, a shift in the midpoint of transition curves from 50.02 Ϯ 0.34°C for apo-LigBCen2 to 55.71 Ϯ 0.88°C for Ca 2ϩ -bound LigBCen2 was also calculated by DSC (Fig. 5B). Taken together, these data indicate that Ca 2ϩ binding stabilizes the overall structure of LigBCen2 significantly.

Implications of Ca 2؉ Binding for the Functions of LigBCen2
Our results confirm the hitherto unknown property of Ca 2ϩ binding by Lig proteins, whose functions are not yet fully understood. This is the first report confirming that a bacterial immunoglobulin domain is a Ca 2ϩ -binding domain. We have recently shown that LigBCen2 binds strongly to the NTD of Fn but very weakly to the gelatin binding domain (GBD) of fibronectin (15). We therefore studied if this binding is Ca 2ϩdependent or modulated by Ca 2ϩ .

Ca 2؉ Influences the Binding of LigBCen2 with Fibronectin as Monitored by Fluorescence
It is known that LigBCen2 strongly binds to Fn (15), and our data demonstrate that Ca 2ϩ binding stabilizes LigBCen2 conformation. In order to determine if Fn binding to LigBCen2 is Ca 2ϩ -dependent or -independent, we performed Trp fluorescence titration of LigBCen2 by Fn in the presence or absence of Ca 2ϩ . As seen in Fig. 6, A and B, there was a significant decrease in fluorescence (up to 20%), performed in the presence of Ca 2ϩ . There was an insignificant decrease when titration was performed in the absence of Ca 2ϩ (in the presence of EGTA), suggesting a weak interaction.

Fn Binding to LigBCen2 Monitored by CD
We monitored the binding of the NTD of Fn with LigBCen2 using far-UV CD. Varying concentrations of NTD were added to the LigBCen2 solution in the presence of Ca 2ϩ . The spectra were corrected for the NTD signal by subtracting the appropriate blank as mentioned under "Materials and Methods." The addition of NTD decreases the ellipticity to a great extent (Fig.  6C). The ellipticity in the region below 200 nm was more negative (as seen for a coiled protein) at a Lig/NTD ratio of about 1:2. When these experiments were performed in the absence of Ca 2ϩ (in the presence of EGTA), the influence was comparatively very weak (Fig. 6D). It is interesting to note that there is no significant change in the tertiary structure upon binding NTD when monitored by near-UV CD (Fig. 7A). Only minor changes were seen when the interaction was performed in the presence of Ca 2ϩ . The above results suggest that the NTD of Fn influ- ences the conformation of LigB-Cen2 significantly in the presence of Ca 2ϩ .

Ca 2؉ Modulates NTD of Fibronectin Binding to LigBCen2
We further assayed NTD binding to LigBCen2 by ELISA in the presence and absence of Ca 2ϩ . As shown in Fig. 7B, holo-LigBCen2 binds the NTD more strongly than apo-LigB-Cen2. Similarly, the binding affinity of the NTD to holo-LigBCen2 was increased (K D ϭ 93 nM) when compared with apoprotein (K D ϭ 272 nM) ( Table 3) (15). These results suggest that Ca 2ϩ binding to LigB-Cen2 assists the interaction with the NTD of Fn.

Binding of Fn to LigBCen2 Monitored by Stains-all Interaction
Stains-all, a carbocyanine dye, is a probe for Ca 2ϩ binding to proteins and has also been used to study protein-protein interactions (32,33,42). When a Ca 2ϩ -binding protein is mixed with Stains-all dye, CD bands at about 620 -650 nm (called J bands) and/or at 500 nm (called ␥-bands) of the dye are induced. LigBCen2 induces a strong J band of the dye at 650 nm, which indicates the Ca 2ϩ -binding nature of LigB-Cen2 as well as the similarity of its Ca 2ϩ -binding site to ␤-crystallin conformation (33,43). When NTD of Fn was added to the protein-dye complex in the presence of Ca 2ϩ , there was a reduction in the J band ellipticity (Fig. 7C), whereas in the absence of Ca 2ϩ (presence of EDTA), there was no such reduction in the ellipticity of the J band (Fig. 7D). These results of fluorescence, CD, ELISA, and Stains-all binding clearly demonstrate that the Ca 2ϩ aids LigBCen2 binding to the NTD of Fn.

DISCUSSION
A number of earlier studies on various Lig proteins have investigated their physiological roles in pathogenicity; still, their functions and structural properties are not yet known (12,13,17,18). This work was undertaken to understand the  conformational and unique features of these proteins by which they are supposed to be involved in pathogenesis. We, for the first time, demonstrate a novel feature of Lig proteins (i.e. Ca 2ϩ binding). Since there is no known Ca 2ϩ -binding motif in these proteins, this is a novel and significant finding. Furthermore, we demonstrate that Ca 2ϩ binding modulates the protein structures and the interaction of Lig with extracellular matrix proteins, such as Fn.
LigB Is a Ca 2ϩ -binding Protein-Our results confirm Lig proteins as Ca 2ϩ -binding proteins despite the lack of a known Ca 2ϩ -binding motif in these proteins, such as an EF-hand motif, C2 domain, or hemolysin-type motif (28). Besides these known motifs, some bacterial proteins also bind Ca 2ϩ through oxygen atoms provided by several charged glutamate or aspartate residues (28). Interestingly, there are five Asp and three Glu residues present in the tandem repeats of LigBCen2. However, often these motifs are less well defined and not easy to specify because of the absence of structural data for these proteins. It appears that Lig has a novel, orphan Ca 2ϩbinding motif that needs to be identified. Taking into account the similarities between various Greek key motifs (31), we suggest that Lig proteins should have a motif for Ca 2ϩ binding similar to that seen in lens ␤␥-crystallins (43). This motif is shown to bind Ca 2ϩ in many proteins, such as Protein S (44), spherulin 3a (45), caulollin (46), and yersinia crystallins (47), although this motif still needs to be more clearly defined by structural studies on various diverse proteins. We suggest that a motif similar to that of ␤␥-crystallins is present in LigB protein, although it needs to be verified by structural means. Extensive structural studies to identify this motif and its comparison with the crystallin-type Greek key motif are under way. There is a possibility that proteins of the BIg domain would bind Ca 2ϩ , thus forming a new family of Ca 2ϩ -binding proteins.
Significant changes in LigBCen2 conformation occur upon binding Ca 2ϩ . Similar results have been noted in some members of the lens ␤␥-crystallin superfamily, such as microbial crystallins (37,46,47). There is a significant tertiary fold in apo-LigBCen2, suggesting that this protein is well folded. However, upon binding Ca 2ϩ , there are significant changes in the tertiary structure of LigBCen2. These changes may differ from those seen in many calcium sensors, such as neuronal calcium sensor-1 (48), since these sensors, unlike LigBCen2, have poor tertiary structures in the apo form. Since Ca 2ϩ binds to LigBCen2 with moderate affinity and stabilizes the protein conformation, Ca 2ϩ binding appears to be structural rather than catalytic or regulatory.
The concentration of calcium in Leptospira spp. infected hosts (in vivo) is generally higher than in the environment, and LigB is an essential virulence factor required by Leptospira spp. to infect the host (13). Thus, a higher concentration of Ca 2ϩ inside the host tissues is generally available to bind to LigB, which would fold the protein and increase its structural stability and thus might help in the protection of the bacterium. This also explains the 50-year-old observations of Johnson and Gary (49) about the absolute requirement of calcium and magnesium for the growth and survival of this bacterium. This might explain how Leptospira spp. adapt and survive in vivo and highlights the importance of LigB for leptospiral infection.
LigB-mediated Binding of Leptospira to Fibronectin Is Enhanced by Ca 2ϩ -Using diverse methods of assays, we have shown that the interaction between LigB and Fn is   (15). SEPTEMBER 12, 2008 • VOLUME 283 • NUMBER 37 assisted by Ca 2ϩ . ClfA is a fibrinogen binding protein that serves as a clumping factor of Staphylococcus aureus. It has been reported that ClfA has a Ca 2ϩ -dependent inhibitory site for the interaction between ClfA and fibrinogen (50). However, our studies show that Ca 2ϩ enhances the binding affinity of LigBCen2 to the NTD of Fn significantly. LigB is an adhesin of Leptospira spp. that assists bacterial attachment to some ECM proteins, including Fn (14,16). Like other adhesions, LigBCen2 also requires Ca 2ϩ for its adhesive properties. Our finding of an increased binding affinity of LigBCen2 to the NTD of Fn in the presence of Ca 2ϩ in vitro indicates that the interaction would be fairly strong in vivo, since there is a continuing presence and availability of Ca 2ϩ in the host. Thus, we hypothesize that the level of Ca 2ϩ with host tissues can modulate the attachment of Leptospira spp. to the infected host. We further propose the possible use of Ca 2ϩ chelators in designing a potent therapeutic approach for the treatment of leptospirosis, which might minimize the interaction between LigB and extracellular matrix proteins.

Calcium Binding to a Fibronectin-binding Domain of LigB
In conclusion, we demonstrate that LigB is a novel Ca 2ϩbinding protein that binds Ca 2ϩ with high affinity. Ca 2ϩ binding stabilizes its structure and influences its local as well as global conformation. This work would add one more class of bacterial Ca 2ϩ -binding protein to the growing list of proteins in bacteria (28). The physiological impact of this study is on the identification of Ca 2ϩ as a factor for modulating Fn binding to LigB and possibly with other extracellular matrices and thus would have implications in pathogenesis. Our work presents a case of a novel bacterial Ca 2ϩ -binding protein and would help in designing better therapy using calcium chelators for bacterial infection.