Targeting Plasmodium falciparum Invasion via Antibodies Against Merozoite Surface Enolase That Binds to Band 3 of Human Red Blood Cells

1. Introduction

Malaria is one of the most devastating parasitic infections, associated with a huge burden of morbidity and mortality [1]. P. falciparum, the causative agent for the deadly disease is an obligatory intracellular parasite. In its blood stage of infection, it invades erythrocytes making erythrocyte invasion an attractive target for malaria intervention. Invasion of host cells require interaction between parasite ligands and host cell receptors. In a bid to evade challenges, P. falciparum dabbles in extensive functional redundancy by deploying multiple invasion pathways that use different proteins as parasite ligands pairing with different receptors on erythrocytes. Parasite proteins which act as ligands come from two major families: Duffy binding like or erythrocyte binding like (DBL/EBL) and the reticulocyte binding like (RBL) proteins [2,3,4]. In P. falciparum, there are four and five known ligands of EBL class (EBL-175, EBL-1, EBA-140 and EBA-181) and RBL class (Rh1, Rh2a, Rh2b, Rh4 and Rh5) respectively [5]. The erythrocyte receptors that P. falciparum engages for invasion could be placed in two categories: the ones with sialic acid (SA) and the others without SA [6,7,8,9]. Accordingly, P. falciparum invasion pathways are either SA-dependent (SAD) or SA-independent (SAID). Interestingly, the malaria parasite has an uncanny ability to spontaneously switch from one invasion phenotype to another presumably to circumvent the host immune response [10]. Erythrocyte Glycophorins A-C are sialylated glycoproteins that serve as receptors for EBA-175, EBA-140 and EBL-1 respectively [7,11,12,13]. The other receptors that have been shown to be used by the parasite include complement receptor 1 (CR1) [14,15,16], basigin [17] and band 3 [18]. CR1 and basigin serve as receptors for Rh4 [4,15] and Rh5 [17] respectively. Apart from these, there are a host of other merozoite cell surface proteins that are believed to participate in the complex events of host cell invasion [5,19]. These include MSP1 that forms protein-protein complex with band 3 of erythrocyte [18], and PfMTRAP that binds to erythrocyte Semaphorin 7A [20]. However, there may be many more parasite ligands that are yet to be identified [21,22]. Thus, the parasite proteins expressed on the cell surface of invasive stages (i. e. sporozoites, merozoites and ookinetes) deserve to be examined for their functional involvement in host cell invasion. P. falciparum enolase (Pfeno), a moonlighting glycolytic enzyme has emerged as one such antigen. Surface expression of this protein has been observed in all three invasive forms of the parasite [23,24,25,26]. Exposure to antibodies against recombinant Pfeno (rPfeno) shows potent growth/invasion inhibitory effect on in vitro blood stage cultures presumably due to the blockade of merozoite surface enolase [26,27,28]. Likewise, antibody blockade of ookinete surface enolase is known to bring about inhibition of the ookinete traversal across the mosquito midgut epithelium [24]. Indeed, an enolase binding protein (EBP) has been isolated from mosquito gut epithelium that acts as host cell receptor for binding to ookinete cell surface enolase [29]. Molecular mechanisms of membrane bound enolase function in merozoite invasion of erythrocyte are yet to be explored. Recent observations about the ability of a monoclonal antibody directed against a unique pentapeptide (-¹⁰⁴EWGWS¹⁰⁸-) epitope of Pfeno to bring about potent invasion inhibition [28] indicate Pf enolase to serve as an invasion ligand. This would imply the presence of a receptor for merozoite enolase on host cell surface. In the present study, a series of experiments were undertaken in an attempt to identify and characterize the erythrocyte receptor for Pfeno. Multiple binding assays were performed to affinity pull down or co-immuno-precipitate the Pfeno binding human red blood cells (hRBC) membrane protein (s) and identify the pulled-out proteins by mass spectrometry and western blotting. Our results indicate that band 3 of hRBC is a Plasmodium merozoite cell surface enolase binding protein. In the present study we have also identified the peptide sequences in Pf enolase and band 3 that make the band 3 -Pf enolase interface which seems to facilitate invasion.

2. Materials and Methods

All the chemicals used were of analytical grade. The sources of different reagents were as given: Protein A-Sepharose (Pierce), Sepharose beads coupled to glutathione (GE Healthcare , Uppsala Sweden), Yeast enolase , anti-yeast enolase antibodies (α-Yeno), anti-band 3 antibodies (α-band 3), anti-spectrin antibody (α-spectrin), anti MSP1₁₉ antibody and Integrin alpha 7 antibodies (Sigma-Aldrich), Protease inhibitor cocktail (Roche Applied Science, Indianapolis, USA), HRP conjugated anti-mouse IgG secondary antibody (Molecular Probes, NJ, USA), Vectashield-mounting medium (Vector laboratories CA, USA), Biotin-labelled peptide library (15-mer peptides with five overlapping amino acids) covering the full sequence of Pfeno (Mimotopes, Australia) , synthetic band 3 peptides ( GenScript, USA Inc). Maxi-pro ELISA plates (Nunc), ready to use ABTS (, 2'-Azino-Bis-3-Ethylbenzothiazoline-6-Sulfonic Acid) (Sigma Aldrich, St. Louis, MO, USA), PVDF membrane (Millipore), Complete RPMI (cRPMI: RPMI with 0.5% Albumax II), Dibutyl phthalate, Iodoacetamide, Trypsin, Chymotrypsin, Neuraminidase and Coomassie stain (Sigma Aldrich). Sequencing grade Trypsin (Promega Madison, WI, USA).

2.1. Purification of recombinant P. falciparum enolase (rPfeno), preparation of Human RBC ghost, packed RBCs, and raising anti-rPfeno antibodies (α-rPfeno):

Recombinant P. falciparum enolase (rPfeno) [30,31], mouse anti-enolase antibodies (α-rPfeno) [31] and Human Red Blood Cell ghosts (hRBC^G) were prepared as described earlier [32]. Starting with human blood taken in heparinized tube, packed RBCs were prepared by centrifuging the sample at 14000 x g for 20 min. Cellular pellet was washed thrice with incomplete RPMI medium. The washed hRBCs were stored at 4^OC and used within 48 hr after collection. The α-rPfeno antibody was raised in Swiss mice in the animal house facility at TIFR and obtained as described earlier [31].

2.2. Erythrocytes binding Assays:

Packed hRBCs (100 µl) were incubated with varying concentrations of rPfeno or Yeast enolase (Yeno as control samples) in RBC binding buffer (50 mM Tris -HCl pH 8, 200 mM NaCl, 1mM EDTA, 2.5 mM MgCl₂, 2 mM dithiothreitol and 1% glycerol) for 2 hours (hrs) at 37^O C. Samples were centrifuged and hRBC pellet was washed twice with the binding buffer. This pellet was then layered on 100 µl of dibutyl phthalate and centrifuged (14000 x g , 5 min). The aqueous supernatant and the organic liquid layers were removed by aspiration. The RBC pellet was washed with phosphate buffer saline (PBS :10 mM Na-Phosphate, 137 mM NaCl, 2.7 mM KCl, pH 7.4). Recombinant Pfeno bound to the hRBCs was eluted using 25 µl of 2M-NaCl in PBS. The desalted eluate was run on 10% SDS-PAGE and the bound rPfeno / Yeno were visualized using western blot with α-rPfeno and α-Yeno antibodies respectively. In another experiment 100 µl of packed hRBCs in PBS were treated with trypsin, chymotrypsin (0.1 and 0.5 mg/ml) and neuraminidase (25 and 66 mU/ml) for 12 hours at 37^OC. Post treatment the enzymes were inactivated by incubating with 1mM PMSF for 1 hr. Samples were centrifuged (2100 x g, 10 min) and the pellet was washed five times with PBS containing 0.1 mM PMSF. The enzyme treated hRBCs were then used for binding assay [33] as described above. Protein samples were analyzed on 10% SDS-PAGE. Western blotting was performed to detect the bound enolase using appropriate antibodies (α-Pfeno / α-Yeno). For binding with human RBC ghosts (hRBC^G), ~100 µg protein equivalent ghost membranes were incubated with varying amounts of rPfeno (0, 5 and 25 µg) or Yeno (0, 5 and 100 µg) in PBS for 1 hr at 37^O C. Post incubation, hRBC^G membranes were collected by centrifugation and washed with PBS to remove unbound rPfeno. After each wash, absorbance at 280 nm was checked and wash step was repeated till the absorbance dropped to zero. Samples were analyzed on 10% SDS-PAGE and bound rPfeno was visualized by western blotting.

2.3. Blot Overlay assay:

The blot overlay assay or Far-western blot was done in accordance with the published protocol [34]. The blot bearing hRBC^G proteins was incubated with varying concentrations of rPfeno (or Yeno as negative control) in protein binding buffer (PBB) and washed with PBST. The presence of hRBC^G protein bound enolase was detected by probing the blot with α-rPfeno or α-Yeno antibodies.

2.4. Pull Down assay:

In a scouting experiment, purified hRBC^G were solubilized using different detergents (CHAPS, NP-40 and Triton-X100) to identify the detergent for maximal extraction of hRBC^G proteome [35]. In brief, the solubilized hRBC^G proteins were incubated with purified glutathione S transferase tagged rPfeno (GST-rPfeno) for overnight at 4^OC. For pulling down the GST-rPfeno bound with hRBC^G protein(s), GSH- Sepharose 4B resin was added and allowed to incubate for 1 hr. In another set of experiments (for detection of non-specific binding), a control sample consisting of solubilized hRBC^G were incubated with the Glutathione S-Sephadex-beads (GSH-B).The glutathione beads from the two sets of samples were separated by centrifugation (2000 rpm, 2 min), six washes of PBS were given, washed beads were suspended in SDS-PAGE sample buffer, boiled and proteins in the supernatant were resolved on 10% SDS PAGE, followed by silver staining. The bands (other than GST-rPfeno) were cut and processed for protein identification by mass spectroscopy.

2.5. Immuno-precipitation (IP) and reverse IP using Protein-A (PrA):

Protein A beads (PrA-B) (20 µl) were incubated with 2 µl α-Pfeno polyclonal antibody (1:200) for 6-8 hrs at room temperature (RT) and PrA-α-Pfeno beads were washed six times with PBS to remove unbound α-Pfeno. In a separate set, solubilized hRBC^G proteins were incubated with rPfeno (in PBS, pH 7.4) overnight at 4^OC. The solubilization was performed using solubilization buffer (20 mM Tris HCl pH-8, 137 mM NaCl, 10% glycerol and 1% NP-40 in presence of protease inhibitor cocktail). PrA-α-Pfeno beads prepared as above were added to the solubilized hRBC^G -rPfeno sample and incubated for 2 hrs at RT. After incubation, beads were washed with PBS and collected by centrifugation (2000 rpm, 2 minutes) and then analyzed for rPfeno-bound proteins on 10% SDS-PAGE. Western blotting was done using α-band 3, α-Spectrin and α-integrin alpha-7 antibodies. In reverse IP experiments, 20 µl of PrA-B were pre-incubated with antibodies against band 3/ Spectrin / Integrin alpha-7. These pre-incubated PrA-B were then mixed with solubilized hRBC^G proteins that were pre-incubated with rPfeno overnight at 4^O C and for 2 hr at RT. The PBS washed beads were collected by centrifugation, solubilized with Laemmli buffer, run on 10% SDS-PAGE and transferred to a blot which was then probed using α-Pfeno to detect rPfeno bound hRBC^G proteins. In an IP coupled with blot-overlay experiment, the band 3 protein from hRBC^G extract was pulled out using PrA-α-band 3 as antibody-antigen complex (PrA-α- band 3-band 3). The beads were washed and mixed with Laemmli buffer, run on 10% SDS-PAGE and proteins were transferred to a blot. This was followed by overlay with rPfeno. The washed blot was then probed with α-Pfeno to confirm direct binding of immuno purified and immobilized band 3 with purified rPfeno.

2.6. Two-dimensional Gel Electrophoresis (2DE) of hRBC ghost and blot overlay:

hRBC^G (120 µg protein) were solubilized in 100 µl of rehydration buffer (7M urea, 2M thiourea, 2% CHAPS, 50 mM DTT and 2% IPG buffer). The IPG strips (11 cm, 4 - 7 pH) were hydrated overnight and used for iso-electric focusing (IEF) with the following program, step 1: 250 V for 30 mins, step 2: 4000 V for 1hr and final step 3: 15000 V hrs. The strips were equilibrated with 2% DTT followed by 2.5% IAA (Iodoacetamide) (10 minutes each) before proceeding to second dimensional 10% PAGE. Three replicate gels were used for this sample. First gel was silver stained to profile the rehydration buffer solubilized proteome of hRBC^G. Second gel was used for a western blot using α- band 3 to visualize band 3 and its variants. Third gel was used to blot the proteins on to a PVDF membrane, followed by blocking and overlay with rPfeno (10 µg) in protein binding buffer (10 ml) overnight at 4^O C, washing and then probing with α-Pfeno. The method used for blot overlay with rPfeno followed by western blot was the same as described above.

2.7. In-gel Digestion of 2DE protein spots and Mass spectrometric analysis:

Spots corresponding to rPfeno binding proteins observed in gel overlay assay were excised from silver-stained gel and In-gel digestion was carried out using trypsin [36]. The gel pieces were treated with 25 mM ammonium bicarbonate (ABC): acetonitrile (ACN) 1:1 v/v). This was followed by reduction using 10 mM DTT at 56 ℃ for 1 hr and alkylation with 55 mM iodoacetamide at room temperature for 45 mins. Gel pieces were washed with 25 mM ABC and dried in vacuum evaporator. Finally, trypsin (400 ng) was added to gel pieces and the mixture was incubated for 16 hrs at 37^O C for digesting proteins to peptides. The peptides were eluted with 70% ACN containing 5% formic acid. Eluted solutions were dried in speed-vac evaporator at 30^O C. Peptides were dissolved in 0.1% formic acid (5 µl). Extracted peptides were analyzed on a Nano LC-ESI-MS/MS (Agilent 6250-Q-TOF) as described elsewhere [37]. Briefly, the extracted peptides dissolved in 0.1 % formic acid were applied to an Agilent HPLC chip (G4240-62002) which had a compact 40 nl C18 RP enrichment column and a 75 µm x 150 mm separation column that was packed with Zorbax 300 SB-C18 (5 µm) resin. Mobile phase solvents used were, solvent A: 0.1% formic acid in water and Solvent B: 0.1% formic acid in 90% ACN. After injection of the sample, column was washed with a 3 -12% gradient of solvent B for 3 min and the peptides were eluted with linear gradients of varying slopes, viz. 12-60% B for 3 to 23 min, and 60-95% B from 23 to 27 min. The mass spectrometric analysis was performed in extended dynamic mode in 50-1700 m/z range with active exclusion 0.5 min after each MS/MS spectrum. The scan rates (spectra/sec) for MS and MS/MS were 3.66 and 2.19 respectively. The acquired data were analyzed using Mass hunter software (Agilent technologies version B.07.00). For database search Mascot (Matrix sciences version 2.3.2) in-house server was used. The mascot generating files (mgf) were submitted to updated NCBI and Swiss-Prot database with parameter taxonomy set as Homo sapiens, fixed modification carbamidomethyl (cysteine), missed cleavage max 2, peptide mass tolerance 50 ppm and MS/MS mass tolerance of 0.6 Da.

2.8. P. falciparum and P. berghei parasite preparations:

P. falciparum 3D7 were maintained in culture. Briefly, human blood from healthy adults was collected in acid citrate dextrose (ACD) as the anticoagulant. After removing the leukocytes, the erythrocytes were washed and suspended in RPMI with 0.5% Albumax (cRPMI). Parasites were maintained at 5% haematocrit in cRPMI at 37^O C in a humidified chamber containing 5% CO₂. The ring stage synchronized P. falciparum culture was allowed to grow for 48 hrs at which time when the egressed merozoites would be maximum in the medium, the merozoites from the supernatant were harvested (at a parasitemia of ~ 5%) by centrifugation (1500xg, 10 min), and used for binding experiments. P. berghei ANKA infected RBCs (~10⁶ cells) were injected into Swiss mice intraperitoneally. Parasitemia was monitored every day and when it reached to ~40%, infected blood (1-2 ml) was collected in equal volume of ACD solution. For the preparation of merozoites, blood samples were centrifuged and cellular pellet was washed thrice with cRPMI. Cell pellet was suspended in 60% ice cold percoll and spun at 1500xg in a swing-out rotor at 4^O C for 15 min. Cellular layer at the interphase of aqueous and percoll media was collected. The infected RBCs in this layer were rich in schizonts and trophozoites. The interphase layer cells were repeatedly passed through a 25-gauge needle to release merozoites. Cellular debris from this suspension was removed by centrifugation at 600xg and the supernatant was centrifuged at 18,000xg for 15 min. to pellet down the merozoites [38]. Both the P. falciparum and the P. berghei merozoite pellets were washed with PBS and suspended in 100 μl of non-denaturing lysis buffer (20 mM Tris-HCl, pH 8 containing 137 mM NaCl, 10% glycerol, 1% NP-40 and protease inhibitor cocktail) and incubated on ice for 30 min. Suspension was centrifuged at 20,000xg (15 min, 4^OC) and supernatant was used as merozoite cellular extract. These native parasite preparations and hRBC ghost preparations (as the source for band 3 protein) were used for blot overlay, IP and reverse IP experiments.

2.9. ELISA for monitoring protein-protein interaction:

A library of biotin labelled fifteen-mer peptides (overlapping by five amino acids) covering the entire sequence of Pfeno {Accession No. Q27727, purchased from Mimotopes (Australia)} was immobilized on a streptavidin coated ELISA plate. The purity of the peptides was > 80% and the amino acid composition of each peptide was confirmed by mass spectrometric analysis. The custom ELISA plates consisted of 2 wells (wells 89 and 90) coated with control peptides. The methodology followed for ELISA was according to manufacturer’s protocol. In brief the Pfeno peptide coated wells were blocked with sodium caseinate (1%w/v) for 2 hrs at 22^O C, washed with PBST (phosphate buffer saline containing 0.05% Tween 20, pH 7.4) followed by incubation with solubilized hRBC^G proteins at 37^O C. Plates were developed with α-band 3 monoclonal antibody (1:5000). This was followed by addition of HRP conjugated mouse secondary antibody at a dilution of 1:1000 and incubation at room temperature for 45 min. The solution was discarded, and wells were washed thrice with PBS. Ready-to-use ABTS substrate (100 µl) was added. The color was developed for 10-15 min and the OD was measured at 405 nm on a Tican Plate Reader. ELISA values for control peptides were subtracted from value of each well before plotting the data.

For the converse experiment, i. e. to monitor the rPfeno - band 3 peptide interaction, the ELISA plate wells were coated with different synthetic peptides representing various regions of band 3. Binding of these peptides to rPfeno was probed by co-incubating with 1µg of purified rPfeno in PBS (100 µl) for 6 hrs at 37^O C. The wells were washed with PBST followed by incubation with α-Pfeno (1:2000) for 1 hour, washing three times with PBST, adding HRP conjugated mouse secondary antibody (1:1000) into the wells and incubating it at room temperature for 45 minutes. The supernatant was discarded, wells were washed with PBS and ready to use ABTS (200 µl) added into each well. The color was developed for 10-15 min. and absorbance was measured at 405 nm on a Tican Plate Reader.

2.10. Molecular Modelling:

The UniProt database was used to obtain the amino acid sequence of Pfeno (UniProt Accession no. Q27727; 446 residues). The structure was visualized using PyMOL 2.1.0. The 3D structure of Plasmodium falciparum enolase (Pfeno) was modelled using the crystal structure of Toxoplasma gondii enolase 1 (PDB 3OTR) [39] as a template, with the help of the homology modeller SWISS MODEL (http://swissmodel.expasy.org/). Toxoplasma gondii enolase 1 was used for modelling since it shows >90% sequence identity with Pfeno.

3. Results

3.1. Binding of rPfeno to hRBCs and hRBC ghosts:

To test for the binding of rPfeno to human red cells, rPfeno was incubated with hRBCs and the washed red cells were then run on SDS-PAGE. Western blot of such a gel was probed with α-rPfeno (Figure 1A). Lanes 2-4 showed a concentration dependent binding of rPfeno to hRBCs. The ligand concentration dependent binding isotherm for rPfeno showed a typical hyperbolic saturation response with increasing amounts of rPfeno (Figure 1B) suggesting that binding is most likely to a specific receptor.

Specific nature of this binding was further supported by the failure of yeast enolase to bind to the erythrocytes under similar experimental conditions (Figure 1C). Effect of prior treatment of hRBCs with trypsin and chymotrypsin on binding of rPfeno revealed that treatment with low concentrations of trypsin eliminated the rPfeno binding (Figure 1D, lanes 3 and 4) while it required high concentrations of chymotrypsin to observe a similar effect (Figure 1D, lanes 5 and 6). A negative control showing that hRBCs did not have any cross-reacting proteins to α-rPfeno is shown in lane 1 (Figure 1D). Untreated hRBCs incubated with rPfeno showed the normal binding as expected (positive control lane 2). The Western blots (Figure 2) show the specificity of direct binding of rPfeno to hRBC^G.

The concentration dependent interaction of parasite enolase with the ghost membranes was evident from the greater binding of rPfeno in 25 µg /ml sample as compared to the 5 µg/ml sample (Figure 2A). In a control sample where rPfeno was replaced with yeast enolase and the western blot was probed with α-yeast enolase antibody (α-Yeno), there was no detectable binding of Yeno to hRBC even at a concentration as high as 100 µg/ml (Figure 2B).

3.2. Pfeno interacting membrane proteins in hRBC:

Blot overlay assay (Method section - 2.3) was used to identify the likely interactors of rPfeno in hRBC^G proteins. hRBC^G proteins separated on a gel were allowed to interact with rPfeno and binding was detected by western blot analysis using α-rPfeno. Figure 3A (lane 2) shows the profile of proteins in hRBC^G proteome. Western blot of a replicate of lane 2 did not show any cross reactivity with α-rPfeno (Figure 3A, lane 4, a negative control) while it was able to detect purified rPfeno (Figure 3A, lane 3, a positive control). Results showed the binding of rPfeno to hRBC^G proteins at molecular masses ~91, 65, 48, 35 and 25 kDa (Figure 3B, lane 2; marked with arrowheads). Further, similar experiment with yeast enolase yielded negative (no binding) results (Figure 3C, lanes 2 and 3) indicating that binding of rPfeno to hRBC^G proteins was specific to parasite enolase.

3.3. Molecular Identification of rPfeno binding proteins in hRBC ghosts:

GST-tagged rPfeno was immobilized on GSH-B and used to pull-down rPfeno binding protein(s) from hRBC^G extracts. A silver-stained gel of the hRBC^G proteins captured using different detergents is shown in Figure 4A (lanes 2 - 4). Lane 5 shows a control sample where GSH-Beads (GSH-B) were incubated with hRBC^G extract to determine any non-specific adsorption of proteins by the beads. The bands marked with rectangular boxes (A-F) shown in various detergent solubilized hRBC^G pull-down samples represented the potential Pfeno binding proteins. The proteins identified in various bands using Mass spectrometry are listed in Table 1 along with their Protein scores. Two major proteins identified in these bands were spectrin and band 3. Interestingly, two other proteins ankyrin and integrin-alpha 7 were also detected. This was understood in the background of the fact that for the binding of merozoite cell surface enolase with a receptor on hRBC, it would be essential that binding domain of the receptor be exposed on the cell surface. Among the above mentioned four proteins, it is interesting to note that band 3 is the only one that has an extracellular domain [18,32]. Ankyrin and spectrin are peripheral membrane proteins associated with cytoskeletal structures and are likely to be captured in pull-down along with band 3.

In a separate experiment a pull-down assay was performed on a Triton-X100 solubilized sample of hRBC^G that were pre-treated with alkaline solution to eliminate peripheral proteins [33]. Lane 1 in Figure 4B shows the protein profile of alkali treated hRBC^G proteome. For the pull-down, immobilized GST-rPfeno was incubated with the solubilized extract of alkali treated hRBC^G. A 91 kDa protein bound to the parasite enolase was detected in this assay (Figure 4B, lane 3). Sepharose beads (without rPfeno) did not show any non-specific binding in hRBC^G extract (Figure 4B, lane 2; negative control). A western analysis of the three samples using α-band 3 to probe the blot showed the presence of band 3 in hRBC^G extract (Figure 4C, lane 3) as well as in the pull-down sample (lane 1). Thus, in this assay where peripheral membrane proteins were removed from hRBC^G, only band 3 was identified as an interactor of Pfeno (Figure 4C). The major protein band in the pull-down sample was of ~91 kDa besides which some low molecular weight proteins were also identified by band 3 antibody.

The pull-down protein profiles in samples solubilized using different detergents (CHAPS, NP-40 and Triton X100) showed variation in the isoforms of band 3 captured by rPfeno indicating differential solubilization properties of each detergent (Figure 4A, Table 1). CHAPS and NP-40 solubilized samples yielded lower molecular mass (25 and 48 kDa) variants of band 3 while Triton-X100 solubilized sample captured band 3 at ~91 kDa. Results presented in Figure 4B and 4C provide antibody-based identification of band 3 as the most likely candidate for Pfeno binding to hRBC.

3.4. The band 3 of human erythrocytes interacts with rPfeno:

The co-immuno-precipitate obtained using Protein A Beads (PrA-B) bound to α-Pfeno (PrA-B-αPfeno) along with the two control samples was analyzed on SDS-PAGE {Figure 5A(i)}. A western blot of the gel probed with α-band 3 showed three different forms of band 3 in hRBC^G detergent extract (Figure 5 A(ii) lane 1). A control sample where PrA-B were incubated with hRBC^G extract (lane 2) indicated that agarose beads did not bind any band 3 protein. In the sample where hRBC^G extract was incubated with rPfeno, prominent bands were observed at ~64 and 91 kDa {Figure 5A (ii) lane 3}. In the PrA-B-αPfeno system, the pull-down of band 3 with anti-rPfeno antibody implied the formation of a ‘band 3-rPfeno complex’. Similarly, α-band 3 antibody was immobilized on Protein A beads (PrA-B-α-band 3) and used to pull-down band 3 protein from a pre-incubated mixture of solubilized hRBC^G and rPfeno. Further the western blot analysis of co-immuno-precipitate (probed with α-Pfeno) showed the presence of rPfeno at ~50 kDa {Figure 5B (ii) lane 3}. As expected, there were no α-Pfeno cross-reacting proteins in hRBC^G (lane 1, control) and absence of any non-specific binding of rPfeno to PrA-B (lane 2, control). Also, the Western blot analysis of a pull-down sample obtained using α-rPfeno antibody bound Protein-A beads and probed with α-Spectrin antibody did not show spectrin in co-immunoprecipitate (data not shown). The pull-down sample was analyzed on SDS-PAGE. The Coomassie staining of the gel showed protein bands at molecular masses of 25, 64 and 91 kDa {Figure 5C (i) lane 2}. A blot of this immunoprecipitated sample was over laid with rPfeno and subsequently probed to detect the bound rPfeno using α-Pfeno antibody. The rPfeno showed specific binding to all three isoforms of the band 3 {Figure 5C (ii) lane 2 marked with arrowheads}. Protein bands at 91 and 64 kDa showed more pronounced binding of rPfeno as compared to the 25 kDa form of band 3 protein. Co-immunoprecipitation of rPfeno and band 3 in the reciprocal pull-down assays revealed direct interaction of rPfeno with band 3 providing unequivocal evidence for the direct interaction of parasite enolase to band 3 protein on human RBC plasma membrane.

3.5. Native enolases from P. falciparum and P. berghei merozoites bind to hRBC band 3:

Since all the earlier binding studies were carried out with recombinant Pfeno protein, we wanted to confirm whether native Plasmodium enolase also interacts with hRBC band 3 protein. Hence blot overlay and pull-down experiments were carried out using P. falciparum culture supernatant (as a source of P. falciparum merozoites) and P. falciparum infected RBCs (iRBC). Figure 6, Panel A, lane 2 shows a blot of hRBC ghost preparation, overlaid with merozoite extract, and then probed with α-Pfeno antibodies. The bands that light up are shown with arrows and some of them correspond to degradation products of band 3 protein. In lane 3 we show a pull-down experiment using Protein A beads coupled with α-band 3 antibodies, incubated with P. falciparum iRBC extract. The pulled-down proteins were run on a 10% SDS-PAGE and the Western blot was probed with α-band 3 antibodies. The band 3 protein was indeed seen to be pulled down from the P. falciparum iRBC extract. This pull down and the blot overlay showed band patterns similar to each other and was commensurate with canonical human band 3 protein fragments on SDS-PAGE. Lane 1 is a control lane to show that the α-Pfeno antibodies reacted with rPfeno protein specifically. Our P. falciparum cultures typically had a parasitemia of about 5-10%, and we realized that an iRBC preparation would contain uninfected RBCs as well. Therefore, the band 3 protein pull down in lane 3 could be due to pull down of the protein from uninfected RBCs. In order to ascertain whether band 3 protein did interact with native Pfeno present in iRBC, pull-down experiments were carried out with Protein beads coupled to α-band 3 antibodies, incubated with P. falciparum iRBCs, which were probed with α-Pfeno antibodies (Figure 6, Panel B). The α-band 3 antibodies did pull-down Pfeno from P. falciparum iRBCs as observed in lane 3. A similar result was observed when the rodent malaria parasite P. berghei infected RBCs were incubated with protein A beads coupled with α-band 3 antibodies (Panel B, lane 4). Lane 1 shows the molecular weight markers labelling the blot, while lanes 2 and 5 are controls to show that α-Pfeno antibodies are specific, since they react with rPfeno, but not with proteins from uninfected RBCs.

It is envisaged that the human RBC band 3 protein is a relevant binding partner for parasite enolase present on the merozoite surface. Since the P. falciparum culture did not provide us with adequate amounts of purified merozoites, we decided to use P. berghei merozoites purified as described in the Methods section. Protein profiles of the merozoite (M, lane 1) and hRBC^G extracts (G, lane 2) are shown in a Coomassie stained gel in Figure 6C. Western blots of M and G preparations were made and probed with α-MSP-1₁₉, α-Pfeno and α-band 3 (Figure 6C lanes 3, 5 and 7). Absence of band 3 (lane 5) and the presence of MSP-1₁₉ (lane 3) and enolase (lane 7) in these blots validated the merozoite preparation. The cross-reactivity of our α-Pfeno antibodies to murine Plasmodium parasites has been demonstrated by us earlier. Immunoblots of hRBC^G proteins showed no cross reactivity with α-MSP-1₁₉ (Figure 6C, lane 4) or α-Pfeno (Figure 6C, lane 8), but as expected, showed positive signals when probed with α-band 3 (Figure 6C, lane 6). In pull-down experiments, Pbeno was pulled out from Pb merozoite extract by incubation with PrA-B-α-Pfeno. The beads (PrA-B-α-Pfeno-Pbeno) were then incubated with hRBC^G extract. Presence of band 3 in this pull-down was detected in a Western blot probed with α-band 3 (Figure 6 Panel D, lane 2). Conversely, PrA-B-α-band 3 was used to pull down band 3 from hRBC^G extract as PrA-B-α-band 3-band 3 complex and then incubated with merozoite extract. Presence of band 3 protein in this pull-down was detected in a Western blot probed with α-Pfeno (Figure 6 Panel D, lane 4). The possibility of non-specific binding of band 3 or Pbeno proteins to PrA-B was ruled out by incubating PrA+ merozoite extract + hRBC^G extract and probing with α-band 3 antibodies (lane 1); PrA-B-αPfeno + hRBC^G ghost extract and probing with α-band 3 antibodies (lane 3) and PrA + merozoite extract and probing with α-Pfeno antibodies (lane 5). These controls showed an absence of nonspecific adsorption of band 3 or Pbeno on PrA beads under the experimental conditions used here. The sample where PrA-α-Pfeno-Pbeno incubated with hRBC^G extract was probed with α-band 3 showed two bands at ~25 and 65 kDa (Figure 6D, lane 2) indicating the binding of band 3 to immobilized Pbeno. The blot where PrA-B-α-band 3-band 3 was incubated with merozoite extract and probed with α-Pfeno, showed an enolase band at ~50 kDa and a proteolyzed product at a lower molecular weight (Figure 6D, lane 4). Presence of band 3 in lane 2 and Pbeno in lane 4 (marked by arrows) indicated the formation of PrA-B-α-band 3-band 3-Pbeno and PrA-B-α-Pfeno-Pbeno-band 3 complexes in the two samples. Observation of low molecular weight forms of Pbeno (Figure 6D lane 4) and band 3 (Figure 6A lane 3) are due to possible proteolysis of native forms. The experiments shown in the Figure 6 confirm the binding of native Plasmodium enolases to human band 3 protein.

3.6. Post translationally modified variants of band 3 exhibit differential binding to rPfeno:

In a two-dimensional gel electrophoresis (2DE) experiment (Method section-2.6), we attempted to separate the different variants of band 3. A silver stained 2DE of hRBC^G is shown in Figure 7A. Western blot of a replicate of this where band 3 was detected using α-band 3 showed two clusters of spots (Figure 7B). 25 kDa form of band 3 got resolved into at least five different spots with pIs ranging from 4.8-5.3 while the 64 kDa form showed two spots (pI~5.18). Intriguingly, the 91 kDa form was not observed in the western blot, presumably because it did not get solubilized in 2DE sample buffer. Using a replicate blot, the binding ability of band 3 variants to rPfeno was examined in an overlay assay. Interestingly, while rPfeno binding was observed with both forms at 64 kDa, the 25 kDa forms of band 3 exhibited binding to the three variants (pI~4.93, 5.03, 5.18). Band 3 variants with highest pI~5.31 and lowest pI~4.81 failed to bind Pf enolase (Figure 7C). The individual spots corresponding to band 3 (as seen in western blot Figure 7B) were excised from the silver-stained gel and subjected to Mass-spectrometry for the identification of the proteins. Two proteins that could be identified with high Prot Score were Spectrin and band 3. The protein spot at pI~5.18 and MW~64 kDa had both band 3 and spectrin while the ~25 kDa spot with pI~5.31 showed the presence of three proteins viz. diacylglycerol kinase, chromosome 20 ORF and cadherin 12 (Table 2). Although in Western blot using α-band 3, there was a positive identification of band 3 protein, mass spectrometry failed to confirm that. This can be due to missing the spot of interest on the silver-stained gel. Amounts of protein present in other spots with MW~25 kDa were too small to be identified by mass spectrometry.

As shown in Figure 8, band 3 has 911 residues (NCBI Reference Sequence Accession NP_000333.1). Table 3 lists the sequences of ten different band 3 peptides that were sequenced in mass spectrometry experiments covering ~15% (136 residues) of band 3. The residues that could be sequenced in peptides, led to identification of band 3.

3.7. Synthetic peptide towards mapping the binding sites of band 3 and Pfeno:

Band 3 is a trans membrane protein that has its N-terminal globular region on the cytosolic face while the membrane spanning and exocytic loop regions are at the C-terminal end [35,36]. Thus, the site for parasite enolase that binds to band 3 on extra cellular surface must map on the C-terminal region of band 3. For mapping the Pfeno binding region on band 3, a set of seven different peptides belonging to the exocytic domains of band 3 were custom synthesized (Table 4).

Four of these peptides (P^632-655, P^731–753, P^744–758 and P^852-879) localize in exocytic regions of band 3. Peptide P^813-826 is from a trans membrane region while the peptides P^220-246 and P^247-283 are from the endocytic domain of the band 3 [36]. Sequence positions of all the seven synthetic peptides are underlined (red) in Figure 8A. C-terminal region of band 3 (aa 361-911) forms the trans-membrane domains and exocytic regions of the protein (highlighted in turquoise blue in Figure 8A). Using these peptides as baits, their ability to capture rPfeno was further tested in an ELISA experiment. The results of the assay (Figure 8B) indicated that while the peptide P^813-826 that localizes in trans-membrane region had no interaction with rPfeno, the three exocytic peptides (P^731–753, P^744–758 and P^852-879) showed significant binding and are likely to form the binding region for the parasite enolase. The remaining three peptides (P^220-246, P^247-283 and P^632-655) showed intermediate level of reactivity (Figure 8B).

Interaction of rPfeno with the two N-terminal peptides (P^220-246 and P^247-283) is of unclear physiological relevance as these regions are located on the cytoplasmic face of the hRBC membrane. However, the possibility of post invasion interactions between enolase and the two N terminal peptides cannot be ruled out. Another peptide P^632-655 that showed an intermediate level of binding is also likely to be a part of enolase binding site. Thus, the data presented here imply that rPfeno binding region of band 3 includes sequences P^632-655, P^731–753, P^744–758 and P^852-879. It is important to realize that the method used here is largely limited to identifying sequentially continuous binding sites. Protein-protein interactions that are stabilized by secondary and/or tertiary structural features formed by non-contiguous sequence would not be identified by this approach.

For mapping the Pfeno sites that bind to band 3, a ‘Peptide Scanning’ strategy was used. The method involved synthesis of entire enolase protein sequence as short biotin tagged 15 mer overlapping peptides and testing them individually for binding to band 3 [37]. Binding of band 3 with the Pfeno peptide library was monitored using an ELISA format. As shown (Figure 9A), band 3 showed binding with peptides 16, 40, 49, 50 and 57-64. The sequences of these band 3 binding Pf enolase peptides as also their two neighboring peptides that failed to bind band 3 are shown in Figure 9B. Sequence alignment among peptides 15 - 17 indicated that the peptide 16 shares ten residues with these two non-binding peptides. Thus, if any stretch of five or ten residues in peptide 16 formed a binding site, one or the other neighboring peptides could also show some binding. Absence of such recognition with flanking peptides 15 and 17 was indicative of nonspecific binding of peptide 16. Similar is the case with peptide 40 which also showed no binding for the flanking peptides. Peptides 49 and 50 have ten residues in common (Panel B, shown in green) and both showed interaction with band 3 indicating that these residues are part of the binding domain. The N-terminal five residues of peptide 49 and the C-terminal five residues of peptide 50 are not involved in the binding to band 3. This may be the reason why the flanking peptides 48 and 51 that did not have adequate number of binding domain residues failed to interact with band 3. Thus, from these data we infer that Pf Enolase region ²⁴⁶GKVKIAMDAV²⁵⁵ is involved in binding interactions with band 3. Similar analysis of the other cluster of peptides (peptides 57-64) yielded a binding sequence of ²⁸⁶QLVDLYIDLVKKYPIVSIEDPFDQDDWENYAKLTAAIGKD³²⁵ as the band 3 binding domain in Pfeno. To visualize the location of this sequence in 3D structure of Pfeno, a model structure of Pfeno was generated [38] using the crystal structure of T. gondii enolase (pdb: 3OTR) [39] as template (Figure 9 C) as the two proteins have >90% sequence homology. In this structure, location of these sequences is shown in blue and green. Enolase is also known to bind plasminogen that facilitates invasion of mosquito mid gut epithelium [24]. The plasminogen interacting domain is shown in red. All apicomplexan enolases differ from host enolases in having an insert of five amino acids (¹⁰⁴EWGWS¹⁰⁸) (Figure 9 C, pink). This epitope has been shown to be a target for potent growth inhibiting antibodies [27,28]. It is interesting to note that the domains of Pfeno that are involved in executing one or the other moonlighting function are clustered on the same face of the molecule in the 3D model structure (Figure 9 C).

4. Discussion:

Plasmodium falciparum enolase (Pfeno) is a highly conserved glycolytic enzyme that is expressed on the surface of the merozoites. Involvement of enolase in merozoite invasion of erythrocytes was evident from the observed strong inhibition of the P. falciparum growth in blood cell cultures by α-Pfeno and reduction of parasitemia in P. yoelii infected mice immunized with α-Pfeno [26,27,28,30]. The target epitope for the inhibitory α-Pfeno antibodies was identified as the pentapeptide insert 104EWGWS108 present in all apicomplexan parasite enolases but absent in host enolases [46]. As the initial invasion event requires interaction between a relevant parasite ligand with host receptor, an attempt was made to identify and characterize the Pfeno binder/receptor on hRBC surface. Several experiments presented here indicated that rPfeno bound specifically to both hRBC and hRBC^G in a ligand concentration dependent manner. In the past, the effect of treatment of hRBC with proteases (trypsin and chymotrypsin) and neuraminidase on the binding of the parasite ligands have been used to define sialic acid dependent (SAD) or independent (SAID) pathways [6,7,8,9]. We found that the binding of rPfeno was sensitive to trypsin but resistant to neuraminidase suggesting that the rPfeno binding with the receptor is of the SAID kind. Results from a host of pull-down and overlay assays combined with western blotting and/or mass spectrometry (MS/MS) (Figs. 4-7) analysis led to identification of band 3 as the enolase binding protein in hRBC cell membrane. Band 3 is an anion exchanger that mediates the exchange of the cellular HCO3- with Cl- in plasma [42,47,48]. Structurally, it has 911 amino acid residues (Figure 8A) with a glycosylation site at N642. It is a trans-membrane protein that coimmunoprecipitates with RhopH3 and MSP1 [18,49]. Two extracellular domains of band 3, viz 5C (744ASTPGAAAQIGEVKE758) and 6A (807DRILLLFKPPKYHPDVPYVK826) form the binding site for MSP1. Interestingly, MSP1 mediated invasion pathway has also been shown to be of the SAID kind [18]. The enolase binding sequences identified here consisted of four exocytic peptides viz P632-655, P731–753, P744–758 and P852-879 (Figure 8A and 8B). Thus, there is quite a bit of overlap in the binding sites for MSP1 and Pfeno on band 3. Topographically, N-terminal part of band 3 is globular and located on cytosolic side while C terminal domain forms its transmembrane and exocytic regions [42,47]. As expected, the Pfeno binding site on band 3 mapped in C-terminal exocytic domain (Figure 8). Conversely, the binding site for band 3 on Pfeno was also mapped using a peptide scanning approach. The results showed that the Pfeno regions that bind band 3 are constituted by 246GKVKIAMDAV255and286QLVDLYIDLVKKYPIVSIEDPFDQDDWENYAKLTAAIGKD325 (Figure 9B). Pfeno appears to have multiple distinct sites that relate to its diverse physiological functions. In order to visualize the spatial disposition of various functional regions on the molecular surface of Pfeno, its 3D-structure was modelled using X-ray crystallographic structure of T. gondii enolase (Figure 9 C). Apart from the catalytic site (not shown), there is a plasminogen binding site constituted by 277DKSLVK282 residues [24], band 3 binding site formed by 246GKVKIAMDAV255and 286QLVDLYIDLVKKYPIVSIEDPFDQDDWENYAKLTAAIGKD325 and an invasion inhibitory antibody binding site 104EWGWS108 [27,28,30]. Presence of distinct Pfeno binding regions in band 3 and parasite growth inhibitory Pfeno antibodies indicates that band 3 enolase binding site may facilitate initial adhesion of merozoite to hRBC. The unique pentapeptide insert in Pfeno (104EWGWS108) is likely to serve as a docking port for some (yet to be identified) molecular factor of host origin whose interaction must be critical for merozoite invasion of hRBC in an enolase mediated pathway. In many pathogens as also in malignant cells, cell surface enolase is known to act as plasminogen binder which upon activation cleaves inter-cellular matrix protein facilitating the establishment of infection [25,50,51,52,53]. However, plasminogen binding function of Pfeno is unlikely to be of any relevance in blood stage merozoite invasion of hRBC as there is no barrier of matrix proteins between the host and the parasite.

Pfeno is an unusual surface antigen that expresses on cell surface of all the three invasive stages (sporozoites, merozoites and ookinetes) in Plasmodium’s life cycle [23,24]. In a series of elegant experiments, the physiological functions of ookinete cell surface enolase were demonstrated [23,24,25,29]. In mosquito mid gut, ookinete has to cross the peritrophic matrix barrier to reach to mid gut epithelium. This task is accomplished by binding of human plasminogen (derived from the blood meal of the mosquito) to ookinete cell surface enolase [24,54]. Disruption of this inter molecular interaction between Pfeno and plasminogen by anti-Pfeno antibodies or by feeding plasminogen depleted blood resulted in drastic reduction in traversal of ookinetes across the mid gut epithelium [24,29]. Most of the ookinete traversal inhibitory activity could be accounted for by the plasminogen receptor function of Pfeno. In addition, an enolase binding protein (EBP) that expresses on the luminal surface of the mosquito midgut epithelium was also identified [29]. Although physiological significance of Pfeno-EBP interaction is not well understood, it is likely to be important in mediating host cell recognition by the parasite (similar to merozoite cell surface enolase binding to band 3 of hRBC). Peritrophic matrix in mosquito mid gut is rich in proteins that contain PWWP, a protein-protein interaction motif [55]. Recently, this motif was shown to bind to 104EWGWS108 [56] raising an interesting possibility of aromatic interactions-based assembly of a ternary complex formed by Pfeno, plasminogen and PWWP containing matrix proteins, leading to high efficiency proteolysis of peritrophic matrix. Physiological functions of Pfeno on merozoite and ookinete cell surfaces are different and well adapted for host cell invasion needs of the parasite. Thus, it appears that enolase functions very much depend on when and where this versatile protein is present. This has been seen in several organisms where enolase has multiple moonlighting functions [57]. So far, the role of Pfeno on sporozoite cell surface has not been examined. It will be interesting to examine whether it has any functional role in sporozoite invasion of hepatocytes.

In conclusion, the identification of band 3 as a major erythrocyte receptor for the highly conserved glycolytic protein Pfeno should provide new insights in understanding the role of Pfeno on merozoite cell surface. More in-vivo and in-vitro studies such as RBC invasion inhibition assay using different laboratory strains and real time interaction using Surface Plasmon Resonance (SPR), are required involving both rPfeno and band 3 for better understanding of the host parasite interaction. It is hoped that the identification of erythrocyte band 3 as a receptor for Plasmodium enolase will provide a window of opportunity to develop new peptide/small molecule-based therapeutics and possible vaccine component(s) to combat the deadly disease of malaria.