Functional Manipulation of a Calcium-binding Protein from Entamoeba histolytica Guided by Paramagnetic NMR*

Background: EhCaBP1 is one of the EF-hand calcium-binding proteins (CaBP) involved in various Ca2+ signaling pathways. Results: Hydrophobic residues at the −4 position of the Ca2+-binding loop affects structure, Ca2+-binding properties, and cellular localization of EhCaBP1. Conclusion: The Y81F mutation in EhCaBP1 makes it more structured like CaM and TnC. Significance: Observed variations in the structures and cellular localization of the wt and mutant could have influence on their biological behavior. EhCaBP1, one of the calcium-binding proteins from Entamoeba histolytica, is a two-domain EF-hand protein. The two domains of EhCaBP1 are structurally and functionally different from each other. However, both domains are required for structural stability and a full range of functional diversity. Analysis of sequence and structure of EhCaBP1 and other CaBPs indicates that the C-terminal domain of EhCaBP1 possesses a unique structure compared with other family members. This had been attributed to the absence of a Phe-Phe interaction between highly conserved Phe residues at the −4 position in EF-hand III (F[-4]; Tyr81) and at the 13th position in EF-hand IV (F[+13]; Phe129) of the C-terminal domain. Against this backdrop, we mutated the Tyr residue at the −4th position of EF III to the Phe residue (Y81F), to bring in the Phe-Phe interaction and understand the nature of structural and functional changes in the protein by NMR spectroscopy, molecular dynamics (MD) simulation, isothermal titration calorimetry (ITC), and biological assays, such as imaging and actin binding. The Y81F mutation in EhCaBP1 resulted in a more compact structure for the C-terminal domain of the mutant as in the case of calmodulin and troponin C. The compact structure is favored by the presence of a π-π interaction between Phe81 and Phe129 along with several hydrophobic interactions of Phe81, which are not seen in the wild-type protein. Furthermore, the biological assays reveal preferential membrane localization of the mutant, loss of its colocalization with actin in the phagocytic cups, whereas retaining its ability to bind G- and F-actin.

ogy. The loop consists of a contiguous stretch of 12 amino acid residues and binds to Ca 2ϩ in a pentagonal-bipyramidal geometry (19). The residues in the loop at positions ϩ1, ϩ3, ϩ5, ϩ7, ϩ9, and ϩ12 coordinate with Ca 2ϩ . In addition, the Ca 2ϩ -binding affinity is controlled by other factors such as intrinsic binding affinity of each binding loop, conformational cost upon Ca 2ϩ binding (20,21), EF-␤-scaffold etc. Thus, despite the presence of highly conserved residues in the primary sequence of Ca 2ϩ -binding loops, a wide range of Ca 2ϩ -binding affinities were observed in different EF-CaBPs (22).
The inter-domain linker of EhCaBP1 is more flexible compared with that of CaM and TnC due to the presence of 3 Gly residues in the linker region and thus EhCaBP1 shows different target specificity compared with CaM and TnC (8). In addition, the hydrophobic packing, inside the core of the protein, may also influence the target specificity of EhCaBP1. A closer examination of the NMR structure revealed that the N-terminal domain is structurally more rigid compared with its C-terminal counterpart (8,15,16). Overlaying of individual Ca 2ϩ -binding loops of EhCaBP1 with CaM and TnC suggested that the 3rd (EF III) and 4th (EF IV) Ca 2ϩ -binding loops are more open compared with EF-hands I and II (8). Besides, it was reported that Ca 2ϩ gets displaced by lanthanides (Yb 3ϩ ) first from the 3rd Ca 2ϩ -binding loop in EhCaBP1, during the Ca 2ϩ displacement reaction (20,21). Such preferential Ca 2ϩ displacement from site III was attributed either due to the presence of four negatively charged Ca 2ϩ -coordinating residues (Asp 85 , Asp 87 , Asp 89 , and Glu 96 ) in the EF III binding loop as opposed to the presence of only three such residues in rest of the loops (I, II, and IV) or due to the structural differences between the N-and C-terminal domains. Furthermore, based on the amino acid sequence analysis of EhCaBP1, it was reported that the Tyr residue at the Ϫ4th position (Tyr 81 ) of the third Ca 2ϩ -binding loop (instead of a highly conserved Phe residue seen at this position in CaM and TnC) could be the cause of preferential Ca 2ϩ displacement from EF-hand III (20,21). Previously, it has been shown that, in a given domain containing a pair of EFhands, highly conserved Phe at the Ϫ4th position in the EF I (F[Ϫ4]) interacts with another highly conserved Phe at the 13th position in the EF II (Phe[ϩ13]) of the same domain such that their respective aromatic rings are oriented perpendicular to one another. Such an orientation results in a shorter C ␣ -C ␣ distance of 7-8 Å (8,20). Furthermore, proteins that lack such Phe(Ϫ4)-Phe(ϩ13) interactions are found to have a lower affinity for Ca 2ϩ (23). Such a Phe-Phe interaction, which is present in both CaM and TnC and also in the N-terminal domain of EhCaBP1, is absent in the C-terminal domain of EhCaBP1 due to the presence of Tyr at the Ϫ4th position (Tyr 81 ) instead of the highly conserved Phe residue. Furthermore, the aromatic rings of Tyr (Ϫ4) and Phe (ϩ13) of the C-terminal domain are oriented parallel to each other with an effective C ␣ -C ␣ distance of 10 -12 Å (20). This was attributed as one of the reasons for the structural variation of EhCaBP1 from CaM, TnC, and other CaBPs of its family. Against this backdrop, we set out to mutate, Tyr 81 to Phe 81 , in EhCaBP1 at the Ϫ4th position of EF III (Y81F). This mutant, (Y81F)-EhCaBP1, was overexpressed, purified, and the three-dimensional structure characterized both by multidimensional NMR and MD simulation to study the overall structural stability, dynamics, Ca 2ϩ binding properties, in vivo localization, and interaction with G-and F-actin, compared with the wild-type protein (wt-EhCaBP1).
Protein Expression and Purification-Overexpression and purification of the recombinant protein was carried out as described earlier (14). Apo-form of wt-and (Y81F)-EhCaBP1 were prepared by buffer exchange in a Centricon ultrafiltration system with a 3-kDa cut-off membrane.
Growth and Maintenance of Parasites-E. histolytica strain HM-1 IMSS and all transformed parasites were maintained and grown in TYI-S-33 medium containing 250 units ml Ϫ1 of benzyl penicillin and 0.25 mg ml Ϫ1 of streptomycin in 100 ml of medium. Cells carrying constructs with a constitutive expression system were maintained at 10 g ml Ϫ1 of G418. However, the in vivo experiments were carried out in the presence of 30 g ml Ϫ1 of G418.
Transfection and Selection of E. histolytica Trophozoites-Transfection was performed by electroporation. Briefly, trophozoites in log phase were harvested and washed with phosphate-buffered saline (PBS), followed by an incomplete cytomix buffer (10 mM K 2 HPO 4 /KH 2 PO 4 , pH 7.6, 120 mM KCl, 0.15 mM CaCl 2 , 25 mM HEPES, pH 7.4, 2 mM EGTA, 5 mM MgCl 2 ). The washed cells were then re-suspended in 0.8 ml of complete cytomix buffer (incomplete cytomix buffer containing 4 mM adenosine triphosphate, 10 mM glutathione). Plasmid DNA (200 mg) was added to the cell suspension and subjected to two consecutive pulses of 3000 V/cm (1.2 kV) at 25 mF (Bio-Rad, electroporator). The transfectants were initially allowed to grow without any selection. Drug selection was initiated after 2 days of transfection in the presence of 10 g ml Ϫ1 of G-418.
Immunofluorescence Staining-Immunofluorescence staining was carried out as described before (11,12). Briefly E. histolytica cells resuspended in TYI-33 medium were transferred onto (acetone-cleaned) coverslips placed in a Petri dish and allowed to adhere for 10 min at 35.5°C. The culture medium was removed and cells were fixed with 3.7% pre-warmed paraformaldehyde for 30 min. After fixation, the cells were permeabilized with 0.1% Triton X-100/PBS for 1 min. This step was omitted for non-permeabilized cells. The fixed cells were then washed with PBS and quenched for 30 min in PBS containing 50 mM NH 4 Cl. The coverslips with trophozoites were blocked with 1% BSA/PBS for 30 min followed by incubation with primary antibody. Rabbit polyclonal anti-EhCaBP1 at 1:100, anti-HA at 1:50, and TRITC-Phalloidin at 1:250 was at 37°C for 1 h. The coverslips were washed with PBS and 1% BSA/PBS before incubation with 1:300 diluted anti-rabbit Alexa 488 (Molecular Probes) for 30 min at 37°C. The coverslips with trophozoite cells were further washed with PBS and mounted on a glass slide using DABCO (1,4-diazbicyclo(2,2,2)octane (Sigma) 10 mg/ml in 80% glycerol). The edges of the coverslip were sealed with nail polish to avoid drying. Confocal images were obtained using an Olympus Fluoview FV1000 laser scanning microscope.
Preparation of Cytosolic and Membrane Fraction-To separate membrane proteins from the cytoplasmic fraction, the cell extract was prepared by resuspending the cell pellet (ϳ10 7 , washed with PBS) in 1 ml of 100 mM Na 2 HPO 4 buffer containing protease inhibitors (10 mM N-ethylmaleimide, 2 mM PMSF, 0.01 mM leupeptin, and 2 mM 4-(hydroxymercuri)benzoic acid). The suspension was then subjected to three cycles of freezethawing followed by centrifugation at 100,000 ϫ g for 30 min at 4°C. The resulting supernatant was labeled as the cytoplasmic fraction and the pellet as the membrane fraction. The pellet was washed twice with the above buffer and resuspended in the same buffer containing 1% Triton X-100 and re-centrifuged at 100,000 ϫ g for 20 min at 4°C, to separate the Triton-soluble fraction from Triton-insoluble fraction. The protein content of each fraction was estimated by BCA assay. The Triton-soluble fraction was further taken for analysis.
Circular Dichroism (CD)-CD spectra of wt-and (Y81F)-EhCaBP1 were recorded on a JASCO J-810 spectropolarimeter equipped with a Peltier temperature controller. A protein solution of 20 M was used for far-UV CD spectra. The temperature dependence of the CD spectra was carried out to determine the thermal stability of the mutant protein.
Dynamic Light Scattering-The hydrodynamic radius (R H ) was determined for (Y81F)-EhCaBP1 and wt-EhCaBP1 using the dynamic light scattering method on Dynapro-LS instrument at 830 nm. The protein samples were centrifuged at 13,000 ϫ g for 10 min and filtered through a 0.45-m syringe filter from Millipore into a quartz cuvette. A protein concentration of 8 mg/ml was used for each measurement. Of 50 to 60 measurements, only those showing a parabolic curve with a straight base line were considered for mean Ϯ S.D. calculation. The R H was thus determined from the regularization plot.
Isothermal Titration Calorimetry (ITC)-ITC experiments with (Y81F)-EhCaBP1 were performed with a Microcal VP-ITC titration calorimeter at 25°C. Apo-(Y81F)-EhCaBP1 was prepared by EGTA, EDTA, and Chelex-100 treatment, centrifuged, and degassed prior to the titration with Ca 2ϩ . Each titration consisted of injecting 3-l aliquots of 10 mM Ca 2ϩ solution (diluted from 1 M standard CaCl 2 solution purchased from Sigma) into 200 M protein solution (1.7 ml) at an interval of 3 min, to ensure that the titration peak returned to the base line prior to the following injection. The ITC data thus obtained was base line corrected and analyzed using the software ORIGIN, supplied with the instrument. The amount of heat released per addition of the titrant was fitted to different models to determine the number of binding sites and the metal binding affinities of the protein. All the ITC experiments were repeated three times with different protein concentrations (200, 150, and 100 M) to check the reproducibility of the data.
NMR Spectroscopy-For NMR studies, uniformly 15 N-labeled (U-15 N) and 13 C/ 15 N-doubly labeled (U-13 C/ 15 N) (Y81F)-EhCaBP1 protein samples were prepared in a mixed solvent of 90% H 2 O and 10% 2 H 2 O (50 mM Tris-HCl, pH 7.4, 100 mM NaCl, and 10 mM CaCl 2 ). NMR experiments were recorded at 25°C on a Bruker Avance 800 MHz NMR spectrometer equipped with a 5-mm cryogenically cooled triple-resonance probe and pulse field gradients. The spectra were processed using Felix 2002 (Accelrys Inc., San Diego, CA) and analyzed by Cara (24). The 1 H chemical shifts were referenced to an external standard 2,2-dimethyl-2-silapentene-5-sulfonates. The 13 C and 15 N chemical shifts were calibrated indirectly (25). Almost complete 1 H, 13 C, and 15 N resonance assignments of (Y81F)-EhCaBP1 in its Ca 2ϩ -bound state (holo-form) were carried out with a suite of three-dimensional double and triple resonance NMR experiments (26,28) and prior knowledge of the resonance assignments of wild-type protein (26), using the root mean square deviation-based algorithm discussed earlier (27). Furthermore, such assignments were verified by using other multidimensional NMR data (28 -32). The resonance assignments thus obtained were deposited separately for N-and C-terminal domains in BMRB under accession numbers 19196 and 19197, respectively. The three-dimensional [ 15 N-1 H]-NOESY-HSQC ( m ϭ 100 ms) and three-dimensional [ 13 C-1 H]-NOESY-HSQC ( m ϭ 120 ms) spectra were recorded to see additional or missing NOEs (if any) arising from point mutation Y81F in comparison with that of the wild-type protein, which in turn were used in the three-dimensional structure determination of (Y81F)-EhCaBP1.
Paramagnetic titration experiments with [Ca 2ϩ ] 4 -(Y81F)-EhCaBP1 were carried out with YbCl 3 (Sigma) to monitor Ca 2ϩ displacement by Yb 3ϩ . The protein concentration was 0.6 mM in a buffer containing 50 mM Tris-HCl (pH 6.3) and 100 mM NaCl. For each titration, an aliquot of 1 l of the stock solution (1 M) was added to the NMR tube containing the protein solution, mixed, equilibrated, and followed by recording two-dimensional [ 15 N, 1 H]-HSQC at 30°C. The HSQC spectra thus recorded were processed with identical processing parameters. Integral volumes for the individual 15 N-1 H cross-peaks in the resultant spectra were measured using Felix 2002 (Accelrys Inc.).
The 15 N relaxation data measurements were carried out on 800 MHz NMR spectrometer at 25°C with both wt-and (Y81F)-EhCaBP1 to see the effect of the point mutation (Y81F) on its dynamics. The 15 N spin-lattice relaxation rates (R 1 ϭ 1/T 1 ) were measured with inversion recovery delays of 50, 100, 200*, 300, 400, 500, 600*, and 800 ms. The 15 N spin-spin relaxation rates (R 2 ϭ 1/T 2 ) were measured with Carr-Purcell-Meiboom-Gill delays of 10*, 30, 50*, 70, 90, 110*, 130, 150, 170, and 190 ms. The uncertainty in the peak heights was determined in the duplicate spectra recorded with relaxation delays (marked with an asterisk) (33,34). The number of scans was set to 32. Phenomenological R 1 and R 2 values and corresponding uncertainties were determined by non-linear least-square fitting of the experimental data to single exponentials. The [ 15 N, 1 H]-NOE spectra were recorded with and without proton saturation during relaxation delay, with 64 scans for each complex t 1 point on both wt-and (Y81F)-EhCaBP1. A 2.5-s period of proton saturation was used in the NOE experiment. AUGUST 9, 2013 • VOLUME 288 • NUMBER 32

JOURNAL OF BIOLOGICAL CHEMISTRY 23475
Molecular Dynamics (MD) Simulation-MD simulations were performed on both wt-and (Y81F)-EhCaBP1 using GROMACS molecular modeling package (version 4.0.7) (35) and GROMOS96 (ffG43a1) force-field (36,37) with explicit simple point charge solvent (38). The starting structure for wild-type protein in its holo-form was the lowest-energy NMR structure of EhCaBP1 reported earlier (PDB code 1jfk) (8). As for the simulation of [Ca 2ϩ ] 4 -(Y81F)-EhCaBP1, the starting structure was constructed by mutating Tyr 81 to Phe 81 in the above mentioned wild-type protein structure using PyMOL (39). Simulations were performed with cubic periodic boundary under constant volume and temperature (NVT) conditions. Addition of two Na ϩ neutralized the total charge of the system. The electrostatic interactions were treated by the particle mesh Ewald method with Coulomb cut-off of 1 nm. The van der Waals interactions were treated using Lennard-Jones potential with a cut-off distance of 1 nm and a switching distance of 0.9 nm. Initially, the starting structures of (Y81F)-EhCaBP1 and wt-EhCaBP1 were subjected to energy minimization to remove van der Waals contacts of high potential energy using the steepest descent algorithm with a tolerance of 100 kJ mol Ϫ1 nm Ϫ1 and convergence was obtained in both cases. Following this, the position restrained MD were performed for 50 ps to equilibrate the solvent in and around the protein during which atomic positions of the proteins were restrained. Initial velocities were generated conforming to the Maxwell velocity distribution at 27°C. Subsequently, MD simulations were initiated by integrating the Newtonian equations of motion with a time step of 2 fs. All bonds were constrained using LINCS algorithm. Coordinates were saved every 1,000 steps (2 ps) for the analysis and corresponding velocities were saved every 10,000 steps (20 ps) for continuing the simulation. The temperature of the simulated system was maintained at 27°C by coupling to the Berendsen thermostat at a time constant of 0.1 ps, where the protein and solvent along with ions were coupled separately to the temperature bath. The MD simulations were unbiased and unconstrained. They were performed for a reasonable length of time (10 ns each). The simulation protocol was adopted following earlier work (40). Analyses were performed using the tools provided by the GROMACS software package, PyMOL (39), VMD (41), MATLAB (math-works.com), Xmgrace (plasmagate.weizmann.ac.il/Grace) etc. MD simulation runs were performed on an Intel Xeon quad core 2.4 GHz machine loaded with Linux 5.0 (RedHat Enterprise) (a facility provided by TIFR Computer Center, TIFR, Mumbai).
Surface Plasmon Resonance (SPR) Analysis-The SPR experiments were performed on a Twingle Surface Plasmon Resonance System (Kinetic Evaluation Instrument, Netherlands). The gold disk was incubated with 11-mercaptoundecanoic acid (10 mM in ethanol) solution for 36 h. Thereafter, the -COOH terminated surface was activated by the injection of 50 l of a 1:1 mixture of N-hydroxysuccinimide (25 mM) and N-ethyl-N-(3-dimethylaminopropyl)carbodimide (100 mM) at a flow rate of 15 l s Ϫ1 . The chemical modification on the surface of the chip enabled binding of the free amino groups of the G-actin. The G-actin (1 mg/ml) dissolved in G-buffer (2 mM Tris-Cl buffer, pH 8.0, 0.2 mM of CaCl 2 , 0.2 mM ATP, 0.5 mM DTT) was allowed to covalently link with the matrix for about 2 h at a flow rate of 15 l s Ϫ1 . The remaining activated groups on the chip, which did not react with the G-actin, were blocked by an injection of 50 l of 1 mol/liter of ethanolamine solution. The wt and mutant proteins (0.5 mg/ml) and their apo-forms were taken in corresponding buffers (20 mM Tris-Cl, 10 mM CaCl 2 , and 150 mM NaCl for holo-form and 20 mM Tris-Cl, 5 mM EGTA, and 150 mM NaCl for apo-form) and were allowed to bind to the immobilized G-actin for a time span of 1 h at a flow rate of 15 l s Ϫ1 . The binding was monitored from the angular shift.
Solid-phase Assay-The solid-phase assay experiments were performed to monitor the binding of wt-and (Y81F)-EhCaBP1 to G-actin. The wells of the ELISA plate were coated with 5 M G-actin in PBS buffer and incubated for 12 h at 4°C. The wells were washed with BPS-T buffer. The wt-and (Y81F)-EhCaBP1 protein of 5 M each was added to the wells in duplicates. Bound protein was detected with anti-EhCaBP1 antibody followed by horseradish peroxidase-linked anti-rabbit IgG using the colorimetric substrate 3,3Ј,5,5Ј-tetramethylbenzidine (Sigma). The reaction was stopped with 2 N H 2 SO 4 and absorbance was monitored at 405 nm with ELISA plate reader (Bio-Rad).
Co-sedimentation Assay-Co-sedimentation assay was carried out for monitoring binding of wt-and (Y81F)-EhCaBP1 with F-actin, following the protocol published earlier (11). Briefly, 5 M rabbit muscle G-actin (Sigma) was polymerized for 60 min in polymerization buffer containing 100 mM KCl and 2 mM MgCl 2 at room temperature. After polymerization, actin was mixed with 1 mM ATP and the appropriate target protein (5 M) in a total volume of 150 l of G-buffer (10 mM Tris-Cl, pH 7.5, 2 mM CaCl 2 , 2.5 mM ␤-mercaptoethanol, 0.5 M KCl, 10 mM MgCl 2 ) and incubated for 2 h at room temperature. The samples were centrifuged at 100,000 ϫ g for 45 min at 25°C, which favor F-actin formation. The supernatant (one-fourth of total) and pellet fractions (total) were analyzed by 12% SDS-PAGE, followed by Coomassie Blue staining. All target proteins were ultracentrifuged at 100,000 ϫ g for 1 h and the supernatant was used prior to the above assay to avoid detection of any spurious aggregate(s).

RESULTS AND DISCUSSION
(Y81F)-EhCaBP1 Is a Stable Monomer-The recombinant mutant protein was purified and analyzed by SDS-PAGE, mass spectrometry (MALDI), and dynamic light scattering. The protein essentially shows a single band in SDS-PAGE (supplemental Fig. S1A), indicative of its purity. The MALDI data clearly shows the purity and monomeric state of the protein, similar to that of EhCaBP1 (supplemental Fig. S1B). The molecular masses of (Y81F)-EhCaBP1 and the wild-type proteins are 14.806 and 14.822 kDa, respectively (supplemental Fig. S1B), as determined by MALDI. The dynamic light scattering measurements also reveal the monomeric form for both wt-and (Y81F)-EhCaBP1 (supplemental Fig. S2). The hydrodynamic radii determined by the dynamic light scatterings were 2.0 and 1.7 nm for the wt-and (Y81F)-EhCaBP1, respectively. The far-UV CD spectra recorded with both the wt-and (Y81F)-EhCaBP1 show a similar percentage of secondary structural components (supplemental Table S3 and Fig. S3). The temperature-dependent far-UV CD spectra show high stability for both the wt-and (Y81F)-EhCaBP1 (supplemental Fig. S3).
Ca 2ϩ -binding Affinity of the Mutant-Ca 2ϩ -binding affinity and associated thermodynamic parameters were reported earlier for wt-EhCaBP1 using both NMR and ITC (13). Although four metal ions have been shown to bind wt-EhCaBP1, detailed analysis revealed that three sites bind Ca 2ϩ in an exothermic manner, whereas one site in an endothermic manner. The latter is thought to be due to weak a Ca 2ϩ -binding affinity of the fourth Ca 2ϩ -binding site. The data could be fitted to only a sequential four-site model and the overall dissociation constant was reported to be 3.16 M (13). The presence of four bound Ca 2ϩ was further confirmed by the characteristic downfield shift of four backbone amide proton ( 1 H N ) resonances of homologous Gly residues at the 6th position (Gly 15 , Gly 51 , Gly 90 , and Gly 122 ) of the four functional Ca 2ϩ -binding loops, as described earlier (13). Each of these Gly( 1 H N ) is involved in hydrogen bonding with the side chain carboxyl oxygen atom (CЈO) of an invariant Asp at the first position of the respective Ca 2ϩ -binding loops (Scheme 1). Thus, the corresponding [ 15 N, 1 H] peaks of these Gly residues appear in the leastcrowded region of two-dimensional [ 15 N, 1 H]-HSQC (19) and act as good markers of Ca 2ϩ binding. As in the case of the wild-type protein, (Y81F)-EhCaBP1 also binds four Ca 2ϩ , as confirmed by the observation of four downfield shifted spectral signatures of the Gly residues ( Fig. 1A and supplemental Fig. S4) in the two-dimensional [ 15 N, 1 H]-HSQC. The Ca 2ϩ binding process is completely exothermic in nature for all the four sites of (Y81F)-EhCaBP1, unlike wild-type protein. The ITC data could be fitted to only two sets of sites model and overall dissociation constant was found to be 86 nM (Table 1). This shows a substantial enhancement in the Ca 2ϩ -binding affinity due to the point mutation, similar to the one observed in the case of a 36-residue long peptide representing the EF III of the protein with Tyr to Phe mutation (23). To explain this enhanced affinity, the factors that determine the intrinsic Ca 2ϩ -binding affinity need to be considered. The major controlling factors are the changes in enthalpy (⌬H) and entropy (⌬S). Among these two, the enthalpic contribution to the affinity in a favorable manner is in the formation of constructive electrostatic interactions between the ligands and the metal ion. In addition, loop-based helix-dipole and inter-helical interactions also play key roles in the binding enthalpy. The entropic contribution to the affinity in a favorable manner arises from the release of water molecules from the coordination sphere of the metal ion to the bulk solvent, when Ca 2ϩ coordinates to the protein, thus increasing the overall Ca 2ϩ -binding affinity. On the other hand, the factor that contributes to the entropy in an unfavorable manner arises from the loss of conformational entropy due to structural rigidity in the presence of Ca 2ϩ . The additional unfavorable entropy comes from the Ca 2ϩ -induced exposure of the hydrophobic surface, which also reduces Ca 2ϩ affinity.
Mutation-induced Chemical Shift Perturbations-Overlay of two-dimensional [ 15 N, 1 H]-HSQC spectra (supplemental Fig.  S4) of wt-and (Y81F)-EhCaBP1 revealed significant chemical shift perturbations for some residues, although the majority of residues did not show any noticeable change in their chemical shifts (Fig. 1B, supplemental Fig. S4). For a closer examination of the perturbations in the chemical shifts, we used the spectral signatures of the marker residues Gly 15 , Gly 51 , Gly 90 , and Gly 122 . Although Gly 15 and Gly 51 belonging to the N-terminal domain did not show any change in their chemical shifts, Gly 90 and Gly 122 belonging to the C-terminal domain expectedly showed significant perturbations in their chemical shifts (supplemental Fig. S4). For a greater insight into the chemical shift perturbations, all the 1 H N and 15 N chemical shifts were measured and residue-wise summed chemical shift changes (⌬␦) were calculated using Equation 1, where, ⌬H and ⌬N signify the observed changes in 1 H N and 15 N chemical shifts going from the wild-type protein spectrum to the mutant (Fig. 1B), respectively. As evident in Fig. 1B , and ␤4 (123-125) regions of the protein and can be expected due to the interaction of Phe 81 at the mutated site with the residues in the above specified stretches. Such chemical shift perturbations hint at conformational changes in the C-terminal domain, which may result in an altered hydrophobic surface. This is crucial for protein-protein interactions, which in turn controls the function of the protein. It was reported earlier that the N-and C-terminal domains show distinct unfolding behavior and hence are thought to function independently (15). The small chemical shift changes observed for the residues belonging to the N-terminal domain may be attributed to an altered domain orientation in the mutant protein.
Based on the well known empirical relation of 13 C chemical shifts (⌬C ␣ -⌬C ␤ ) (42,43), the secondary structure elements of the mutant protein were characterized. As evident in the Fig.  1C, (Y81F)-EhCaBP1 is predominantly ␣-helical in conformation with 8 ␣-helices and 4 short stretches of ␤-strands, similar to that of the wt-EhCaBP1, indicating that the Y81F point mutation does not alter the secondary structural elements in the protein. This is also evident from the far-UV CD spectra, which show 71 and 74% helical content for wt-and (Y81F)-EhCaBP1, respectively (supplemental Table S3 and Fig. S3).
Modulation of Conformational Dynamics-The 15 N relaxation study is one of the most established methods for probing millisecond to picosecond motions along the polypeptide backbone (33). In other words, NMR relaxation measurements provide information about the overall and internal motions in pro-SCHEME 1. Amino acid sequences of the four Ca 2؉ -binding loops in wt-EhCaBP1. AUGUST 9, 2013 • VOLUME 288 • NUMBER 32 teins, which is crucial for explaining its structural dynamics. The 15 N-longitudinal relaxation rates (R 1 ; Fig. 2A On the other hand, although the R 2 values for the (Y81F)-EhCaBP1 (Fig. 2B) also show variations all along the sequence, the variation was found to be more systematic and significant in the C-terminal domain as compared with that of the wild-type protein. As is evident from Fig. 2B, the R 2 values for the residues in the C-terminal domain of the mutant were found to be significantly higher compared with its counterpart in the wildtype protein, which could be attributed to the rigidity/compactness induced by the point mutation (Fig. 2B). The mean value of R 2 for N-and C-terminal domains of wt-EhCaBP1 were found to be 13.98 Ϯ 1.19 and 13.03 Ϯ 1.23 s Ϫ1 , whereas for (Y81F)-EhCaBP1 were 14.08 Ϯ 1.56 and 15.46 Ϯ 1.12 s Ϫ1 , respectively. Furthermore, to probe the differences in the backbone mobility and conformational dynamics, the difference in the residuewise R 2 values between wt-and (Y81F)-EhCaBP1 were calculated (Fig. 2C). As evident from this figure, there are both positive and negative variations in R 2 values occurring on different time scales in distinct parts of the protein representing concomitant conformational exchange. There is a continuous and large negative changes in R 2 values for the polypeptide stretches 79 -84 (␣ helix-5 with the mutation site), 88 -89 (Ca 2ϩ -binding site III), 99 -101 (␣ helix-6), 105-107 (linker between Ca 2ϩbinding sites III and IV), 109 -113 (␣ helix-7), 116 -118 and 122-126 (Ca 2ϩ -binding site IV), 129 (the Phe residue showing interaction with the mutated residue Phe 81), and 131-132 (terminal residues). Thus, theinteraction between the aromatic rings of Phe 81 and Phe 129 and the hydrophobic interaction discussed later could result in the C-terminal domain adopting a compact/rigid three-dimensional structure as in the case of CaM and TnC compared with the wt-EhCaBP1. Furthermore, the relative increase in the R 2 /R 1 values (Fig. 2D) of individual residues for the C-terminal domain in the case of the mutant compared with that of the wild-type protein indicates conformational exchanges induced due to the Y81F mutation. Mean values of R 2 /R 1 for wt-EhCaBP1 were found to be 16.5 Ϯ 0.53 and 13.89 Ϯ 0.46 for the N-and C-terminal domains, respectively, whereas for (Y81F)-EhCaBP1 these values were found to be 16.4 Ϯ 0.74 and 17.4 Ϯ 0.58 for the respective domains (Fig. 2D).

Functional Manipulation of a Calcium-binding Protein
To substantiate the observed changes in R 2 values upon point mutation, [ 15 N, 1 H]-NOEs were measured for both wt-and (Y81F)-EhCaBP1 (Fig. 2E) as mentioned under "Experimental Procedures." The mean value of NOEs was found to be 0.82 Ϯ 0.07 for the N-terminal domains of both wt-EhCaBP1 and (Y81F)-EhCaBP1, whereas the corresponding values were 0.82 Ϯ 0.02 and 0.89 Ϯ 0.05 for the C-terminal domains of wt-EhCaBP1 and (Y81F)-EhCaBP1, respectively. The increased mean value of NOEs for the C-terminal domain of (Y81F)-EhCaBP1 corroborates the compactness of its three-dimensional structure induced by the point mutation as compared with its wild-type protein. Thus, we propose that theinteraction of the Phe rings (Phe 81 and Phe 129 ) and hydrophobic interaction with other residues in (Y81F)-EhCaBP1 makes its three-dimensional structure more compact, similar to that of CaM and TnC. The relaxation data are further supported by the outcome of MD simulations (Fig. 3) and the NMR-derived three-dimensional structure of (Y81F)-EhCaBP1 (Fig. 4) discussed below.
MD Simulations on wt-and (Y81F)-EhCaBP1-MD simulations were performed on both wt-and (Y81F)-EhCaBP1 proteins as described under "Experimental Procedures" and various conformational parameters were monitored to assess the changes caused by the point mutation. To start with, the Phe 81 (C ␣ )-Phe 129 (C ␣ ) distance was monitored throughout the simulation (Fig. 3A). As evident from Fig. 3, the Phe 81 (C ␣ )-Phe 129 (C ␣ ) distance stabilized around 8 Å within 1 ns of simulation time. For a greater insight into the conformational changes that could have occurred at and around the mutated site, we carried out cluster analysis of trajectories obtained from MD simulations for both wild-type and mutant proteins. As mentioned earlier, Phe residues at the Ϫ4th and ϩ13th positions of the paired EF-hands in both the domains of the CaM/ TnC and in the N-terminal domain of EhCaBP1 are conserved and show edge to face interaction, with the corresponding Phe 81 (C ␣ )-Phe 129 (C ␣ ) distance around 7-8 Å. To visualize the existence of similar interaction in the C-terminal domain, and the relative orientation of these residues, cluster analysis were FIGURE 1. A, the two-dimensional [ 15 N, 1 H]HSQC spectrum of (Y81F)-EhCaBP1, recorded with 256 ϫ 1024 complex points along t 1 and t 2 dimensions, respectively, on a 800 MHz Bruker Avance NMR spectrometer at 25°C. The assignments are indicated by the one-letter amino acid code followed by the corresponding sequence number along the protein primary sequence. The peaks shown by solid circle (Arg 30 , Lys 33 , Glu 35 , and Gly 63 ) are below the counter level. The peaks connected with horizontal lines are correlations from side chain NH 2 spin pairs belonging to Asn and Gln residues. B, chemical shift perturbation (CSP) plot shown as residue-wise summed chemical shift changes (⌬␦) between wt-EhCaBP1 and (Y81F)-EhCaBP1 calculated from ⌬ 1 H N and ⌬ 15 N shifts, as explained in the text. C, chemical shift index (CSI) derived as residue-wise ⌬C ␣ -⌬C ␤ secondary chemical shifts of (Y81F)-EhCaBP1. ⌬C ␣ and ⌬C ␤ were obtained as the differences between the experimentally observed 13 C ␣ and 13 C ␤ chemical shifts and the corresponding random coil chemical shifts. The ⌬C ␣ -⌬C ␤ value for a particular residue i represents the average over three consecutive residues, iϪ1, i, and iϩ1. A stretch, having negative values of ⌬C ␣ -⌬C ␤ , indicates the presence of a ␤-strand, whereas a stretch of positive values indicates an ␣-helix. Overall dissociation constant for (Y81F)-EhCaBP1 ϭ 1/(K m1 K m2 ) 1/2 ϭ 86 nM. AUGUST (Fig. 3B). On the other hand, a single major cluster with 81% (Fig. 3B) of the total population was observed  for the mutant protein. A representative structure taken from the predominant cluster of (Y81F)-EhCaBP1 shows a Phe 81 (C ␣ )-Phe 129 (C ␣ ) distance of 8 Å, whereas representative structures belonging to clusters 1, 2, 3, and 4 of the wild-type protein show varying Phe 81 (C ␣ )-Phe 129 (C ␣ ) distances of 9.8, 9.0, 9.2, and 9.8 Å, respectively, which are ϳ1 to 2 Å larger than the distance observed in the mutant. Furthermore, it is interesting to note that Phe 81 and Phe 129 in (Y81F)-EhCaBP1 are oriented in an edge-to-face manner suggesting a stronginteraction as seen in CaM and TnC (Fig. 3B). In wild-type protein simulation, Tyr 81 and Phe 129 residues are far apart in clusters 1 and 4 suggesting the absence ofinteraction, whereas in clusters 2 and 3, they show an edge to face interaction with the Phe 81 (C ␣ )-Phe 129 (C ␣ ) distance larger than that seen in the predominant mutant cluster. We monitored the hydrophobic residues at and around the mutated site in both wild-type and mutant simulations. Although Leu 80 , Leu 92 , Phe 100 , and Phe 126 are in close proximity of Tyr 81 forming a hydrophobic core in wt-EhCaBP1, a total of seven residues (Leu 80 , Leu 92 , Val 97 , Phe 100 , Leu 126 , Ile 124 , and Phe 129 ) form part of the stronger hydrophobic core with Phe 81 in (Y81F)-EhCaBP1 (supplemental Fig. S5). Thus, the number of hydrophobic residues forming the hydrophobic core in the C-terminal domain of the protein is more in the mutant compared with the wild-type protein. Overall, the stronginteraction observed between the two aromatic residues (Phe 81 and Phe 129 ) and the large hydrophobic interaction (supplemental Fig. S5) present among nine residues facilitate the compactness of the C-terminal domain in the mutant protein similar to that reported for CaM and TnC (20).

Functional Manipulation of a Calcium-binding Protein
A closer examination of the simulated wt-EhCaBP1 structure reveals that Tyr 81 is not only involved in hydrophobic interaction with the aforementioned residues (Leu 80 , Leu 92 , Phe 100 , and Phe 126 ), but also involved in H-bonding. The OH group of Tyr 81 forms strong H-bonds with the 1 H N of Gly 90 , backbone carbonyl oxygen (CO) of Ile 124 and the side chain carboxylic groups (OD1 and OD2) of Asp 85 (Fig. 3C). This H-bond interaction is energetically more favorable than the expected hydrophobic interaction andinteraction between Tyr 81 and Phe 129 . As a result, the aromatic rings of Tyr 81 and Phe 129 in wt-EhCaBP1 remain away from each other, resulting in an average Tyr 81 (C ␣ )-Phe 129 (C ␣ ) distance of ϳ10 Å (Fig. 3, A and B). This is likely to be the reason for EF III and IV in the C-terminal domain of wt-EhCaBP1 to adopt open loop conformations compared with their counterparts (EF I and EF II) in the N-terminal domain and diverge from the corresponding loops in CaM and TnC. On the other hand, in the mutant, the mutated residue Phe 81 does not engage in any of the aforementioned H-bonding (Fig. 3C, Scheme 1). Instead, it involves a stronger hydrophobic interaction as described above. To understand the hydrophobic interaction at and around the mutation site, we monitored the hydrophobic solvent accessible surface area available for Phe 129 , which is a common interacting partner of Tyr 81 /Phe 81 . As expected, the hydrophobic solvent accessible surface areas for Phe 129 in (Y81F)-EhCaBP1 and wt-EhCaBP1 were 5 and 14 Å 2 , respectively (supplemental Fig. S6). As mentioned above, Phe 129 is involved in a stronger hydrophobic interaction with nine residues, and this large hydrophobic interaction facilitates the C-terminal domain to be relatively more closed in the mutant. Furthermore, the superposition of Ca 2ϩ -binding loops of EF III and IV of the mutant with wt-EhCaBP1, CaM, and TnC revealed that the mutant protein samples loop conformation (supplemental Fig. S7) similar to that of CaM and TnC, unlike the wt-EhCaBP1.
NMR Structure Calculation of (Y81F)-EhCaBP1-The threedimensional structure of (Y81F)-EhCaBP1 was determined using (i) dihedral angle constraints ( and ) derived from TALOS using 1 H, 15 N, 1 H ␣ , 13 CЈ, 13 C ␣ , and 13 C ␤ chemical shifts; (ii) distance constraints obtained from 15 N-and 13 Cedited NOESY after identifying and assigning all the NOESY cross-peaks (the lower bound and upper bound distance constraints were set to 2.4 and 6.0 Å, respectively); (iii) H-bond constraints for the residues, which are involved in ␣-helical and ␤-sheet conformation; and (iv) Ca 2ϩ -ligand coordination derived from pseudo-contact shifts. The structure was calculated using CYANA 3.0 (45,46), following the standard simulated annealing and torsion angle dynamics protocol. The structures of N-and C-terminal domains of (Y81F)-EhCaBP1 were determined separately (PDB codes 2m7m and 2m7n, respectively; Fig. 4, A and B). The details of structural statistics are provided in supplemental Tables S1 and S2. The three-dimensional structure of (Y81F)-EhCaBP1 thus derived is in good agreement with the outcome of MD simulations: (i) aromatic rings of Phe 81 (Ϫ4th position to EF-III) and Phe 129 (ϩ13 th position to EF-IV) are in close proximity as seen in the N-terminal domain of wt-EhCaBP1, CaM, and TnC, bringing the Phe 81 (C ␣ )-Phe 129 (C ␣ ) closer to each other (7-8 Å) (Fig. 5, C and D) unlike the C-terminal domain of wt-EhCaBP1 (Fig. 3B), (ii) Phe 81 shows NOEs to Leu 92 , Val 97 , Phe 100 , Ile 124 , Ile 126 , and Phe 129 (some of the resolved NOEs are shown in supplemental Fig. S8), which stabilizes theinteraction between aromatic rings of Phe 81 and Phe 129 , making the C-terminal domain more compact as compared with wt-EhCaBP1. The compact structure of the C-terminal domain of (Y81F)-EhCaBP1 is supported by the higher [ 15 N, 1 H]-NOE, R 2 , and R 2 /R 1 values as compared with the corresponding values of wt-EhCaBP1, as discussed above (Fig. 2, A-E).
Ca 2ϩ Displacement by Paramagnetic Trivalent Lanthanides (Ln 3ϩ )-The fact that the ionic radii of trivalent lanthanides (Ln 3ϩ ) are similar to that of Ca 2ϩ , one can use Ln 3ϩ for the displacement of Ca 2ϩ in diamagnetic proteins (like EhCaBP1). The pseudo-contact shifts arising from the paramagnetic Ln 3ϩ can then be used to determine and/or refine/validate the highresolution three-dimensional structure of any given protein under investigation (20,21). As mentioned under "Experimental Procedures," paramagnetic titration experiments were carried out with YbCl 3 using a U-15 N-[Ca 2ϩ ] 4 -(Y81F)-EhCaBP1 sample, to monitor Ca 2ϩ displacement by Yb 3ϩ . Yb 3ϩ was a preferred paramagnetic metal ion due to its favorable pseudocontact shift/line broadening ratio compared with other Ln 3ϩ (47). During the course of titration, a number of changes were noticed in the two-dimensional [ 15 N, 1 H]-HSQC spectrum. During the initial steps of titration, several of the original peaks broadened due to the enhancement of relaxation rates induced by the paramagnetic ion (Yb 3ϩ ), whereas several new peaks appeared all over the HSQC spectrum and several others remained unaffected in their positions and line widths. It is worth mentioning here that when Yb 3ϩ is added to a protein, the cross-peaks of residues undergo pseudo-contact shift if they lie within a distance range of 10 -25 Å of Yb 3ϩ , whereas it is expected to completely broaden out if they are located within 10 Å of Yb 3ϩ and mostly unaffected if they are beyond 25 Å. The pseudo-contact shifts thus arising from the paramagnetic ion provide long-range distance constraints with respect to NOEs FIGURE 5. Binding of (Y81F)-EhCaBP1 to G-actin. In the solid phase assay, the wt-EhCaBP1 and (Y81F)-EhCaBP1 proteins were incubated with G-actin-coated wells of a multiwell plate as described in the text. Binding was carried out either in the presence of Ca 2ϩ or EGTA, as indicated. EhCaBP1 binding was quantified by ELISA using an antibody against EhCaBP1. A, calcium dependence of binding of wt-EhCaBP1 and (Y81F)-EhCaBP1 to G-actin by indirect ELISA. CaBP2 that does not bind actin is used as a negative control (experiment repeated three times; n ϭ 3). B, the co-sedimentation assay was carried out for wt-EhCaBP1, (Y81F)-EhCaBP1, and CaBP2 with F-actin (n ϭ 3). The binding was monitored either in the presence of Ca 2ϩ or EGTA. Y81F mutant protein was found to bind actin filaments in Ca 2ϩ independent manner contrary to that observed for the wt-EhCaBP1 (n ϭ 3). EhCaBP2 was used as a negative control as it is a close homolog of EhCaBP1 (79% identity at amino acid level) and does not bind actin. and they can be used in calculation, refinement, and validation of high-resolution three-dimensional structures of proteins in solution. It also helps in fixing the position of the metal ions (Ca 2ϩ in CaBPs) and to study their binding motifs, which is important for metal-binding proteins.
During the initial stages of the titration, peaks due to Gly 90 belonging to EF III, broadened out first, whereas the peak due to Gly 122 of EF IV started showing pseudo-contact shift. Thereafter, Gly 51 of EF II broadened and Gly 15 of EF I showed pseudocontact shift. In the later stages of the titration Gly 15 and Gly 122 peaks started broadening out. Thus, Ca 2ϩ displacement by Yb 3ϩ follows a sequential path involving first, the site III followed by sites II, I, and IV, similar to that observed earlier (20,21) with wt-EhCaBP1 (supplemental Fig. S9). As mentioned earlier, preferential Ca 2ϩ displacement by Yb 3ϩ from EF III of wt-EhCaBP1 was attributed to charge-charge repulsion arising from the presence of four negatively charged Ca 2ϩ -coordinating residues (Asp 85 , Asp 87 , Asp 89 , and Glu 96 ) in the EF III binding loop or due to the open-loop conformation of the EF III. The lack of perturbation in the displacement pathway, despite the observed structural changes associated with the Y81F mutation, reveals that it is the higher density of acidic residues present in the EF III that preferentially dictates the Ca 2ϩ displacement from EhCaBP1 by Yb 3ϩ , rather than the open-loop conformation seen in the wild-type protein. Furthermore, it could be concluded that the observed structural changes associated with the point mutation substantially enhance the Ca 2ϩ -binding affinity as observed in our ITC measurements described above.
G-actin Binding to wt and Mutant Using SPR-The binding characteristics of G-actin with the wt and mutant proteins were studied in the presence and absence of Ca 2ϩ by SPR experiment. As described under "Experimental Procedures," the gold sensor chip was coated with G-actin and binding with wt and mutant proteins in their apo-and holo-forms were monitored by determining the change in the angular shift. For wt-EhCaBP1 protein the angular shifts in the presence and absence of Ca 2ϩ were 24 and 16 mDeg, respectively, whereas for (Y81F)-EhCaBP1 the corresponding values were 29 and 19 mDeg, respectively (supplemental Fig. S10). These findings clearly suggest that the wt and mutant proteins display almost similar binding affinity in the presence of Ca 2ϩ and both values decrease upon removal of Ca 2ϩ .
Functional Characterization of (Y81F)-EhCaBP1-It is likely that the mutation may disrupt some of the properties of wt-EhCaBP1. To check this, we have carried out a few assays based on the property of the wild-type protein. Functional differences such as binding to actin and subcellular localization between (Y81F)-EhCaBP1 and the wild-type were analyzed. In the first assay, G-actin binding was assessed using a solid-phase system described earlier (12,17). Both (Y81F)-EhCaBP1 and wt-EhCaBP1 bound G-actin almost to the same extent and the binding by both proteins was reduced by about 50% in the absence of Ca 2ϩ (Fig. 5A). This is in accordance with the SPR results using apo-and holo-forms, described above.
Further binding to F-actin was analyzed by a co-sedimentation assay. In this assay, F-actin was incubated with the indicated protein and sedimented by ultracentrifugation as described under "Experimental Procedures." The presence of the protein substantially more in the pellet as compared with that present in the supernatant suggested the binding to F-actin (Fig. 5B). Furthermore, (Y81F)-EhCaBP1 was found to bind actin filaments in a Ca 2ϩ -independent manner, unlike the wildtype protein, where the binding is Ca 2ϩ dependent. Our previous studies with CaBP1⌬EF, a Ca 2ϩ -binding defective mutant, where we have mutated all the Asp residues at the first position of each Ca 2ϩ -binding loop (Asp 10 , Asp 46 , Asp 85 , and Asp 117 ) present in EhCaBP1 to Ala residues (11) and the N-terminal domain of EhCaBP1 showed that these did not bind G-actin while retaining their ability to bind F-actin (17). In the absence of actin both the (Y81F)-EhCaBP1 and wild-type proteins were found only in the supernatant (supplemental Fig. S11) confirming that these proteins do not aggregate in the presence of Ca 2ϩ . Because the Y81F mutation has not affected calcium dependence of the (Y81F)-EhCaBP1 protein for binding G-actin, it is possible that mutation rendered interaction with F-actin to be Ca 2ϩ independent but this may be mediated by other mechanisms that led to defective dynamic behavior in vivo.
Subcellular localization of HA-tagged (Y81F)-EhCaBP1 and wt-EhCaBP1 was checked by immunofluorescence staining of transformed trophozoites. HA tagging was used to distinguish the mutant from endogenous proteins. The results are shown in Fig. 6, where E. histolytica trophozoites were shown to phagocytose RBCs. Actin polymerization and depolymerization is important for phagocytosis and Ca 2ϩ -binding proteins are involved in this process. The wt-EhCaBP1 was found in the cytosol and around the phagocytic cups as reported earlier (11) and also as shown in Fig. 6A. The (Y81F)-EhCaBP1 was found localized more in the plasma membrane instead of the cytosol in normal proliferating cells (Fig. 6B). Actin is observed around the phagocytic cups and pseudopods. Both actin and wt-EhCaBP1 co-localize at these sites. To delineate the localization pattern of (Y81F)-EhCaBP1, trophozoites were stained in the absence of erythrocytes where phagocytosis is not observed. Results (Fig. 6C) show that both actin and (Y81F)-EhCaBP1 proteins were found in the pseudopods. In contrast, only a small fraction of the mutant protein co-localized with actin at the phagocytic cups (Fig. 6, compare D with A). To validate these observations the levels of fluorescence in cytosol, membrane, and phagocytic cups were quantitated as described (11,48) and the results clearly showed that (Y81F)-EhCaBP1 is significantly more in the membrane region compared with the cytosol (Fig.  6E). This was further confirmed by fractionation of cellular compartments and checking the presence of both proteins by Western blot analysis (Fig. 6F). A significantly higher fraction of (Y81F)-EhCaBP1 was seen in the membrane fraction compared with that in the wild-type protein.
Although actin dynamics play critical roles in both phagocytosis and pseudopod formation, different actin modulating proteins may be required in the two processes (6,49). Variation in the localization of the mutant (Y81F)-EhCaBP1 and wild-type FIGURE 6. Localization of wt-EhCaBP1 and (Y81F)-EhCaBP1 in E. histolytica transforming trophozoites. A and D, trophozoits incubated with erythrocytes to observe phagocytosis, (phagocytic cups are shown by arrow); B and C, trophozoits in the absence of erythrocytes to visualize pseudopods (arrowheads); primary antibodies and actin staining are indicated for each panel. DIC, differential interference contrast. Scale bar, 10 m. E, quantification of the intensity of immunostain of HA tag (Y81F)-EhCaBP1 (black) and EhCaBP1 (gray) in trophozoites showing erythrophagocytosis. For E, five random regions were selected from membrane, cytosol, and phagocytic cups and average relative intensity was computed for each region by taking the signal from membrane as 100%. This was repeated for 10 such cells (n ϭ 10), bars represent standard error; scale bar, 10 m. TRITC-Phalloidin was used to mark actin filaments. EhCaBP1 was probed by specific antibody, HA tag (Y81F)-EhCaBP1 was probed by anti-HA antibody. Furthermore, both proteins were probed by Alexa 488-labeled secondary antibody. F, the distribution of HA tag (Y81F)-EhCaBP1 protein in membrane and cytosolic fractions as compared with EhCaBP1 as seen by Western blotting (Western blotting includes the antibody step, immunostaining is used for microscopy).