Interaction of Plasmodium falciparum apicortin with α- and β-tubulin is critical for parasite growth and survival

Cytoskeletal structures of Apicomplexan parasites are important for parasite replication, motility, invasion to the host cell and survival. Apicortin, an Apicomplexan specific protein appears to be a crucial factor in maintaining stability of the parasite cytoskeletal assemblies. However, the function of apicortin, in terms of interaction with microtubules still remains elusive. Herein, we have attempted to elucidate the function of Plasmodium falciparum apicortin by monitoring its interaction with two main components of parasite microtubular structure, α-tubulin-I and β-tubulin through in silico and in vitro studies. Further, a p25 domain binding generic drug Tamoxifen (TMX), was used to disrupt PfApicortin-tubulin interactions which led to the inhibition in growth and progression of blood stage life cycle of P. falciparum.


Expression of Pfα-tubulin-I and Pfβ-tubulin in parasite and co-localization with PfApicortin.
Localization of PfApicortin with respect to Pfα-tubulin-I and Pfβ-tubulin was monitored in trophozoites and mature schizonts ( Fig. 2A,B). Localization of the tubulins was also observed in free merozoites. PfApicortin was found to localize on the surface of the parasite in subpellicular region in case of both the trophozoites and schizont ( Supplementary Fig. S8, Fig. 2A,B). In the merozoites, apicortin was observed to localize at the apical end of the parasite ( Fig. 2A,B). Both the tubulins were also observed to localize on parasite surface i.e. sub pellicular regions in the schizonts. In the merozoites, tubulins were also found to localize at the apical end as well as the surface ( Fig. 2A,B). Higher extent of colocalization was found for apicortin-α-tubulin-I (Pearson colocalization coefficient: 0.738 ± 0.0718, Fig. 2A) as compared to β-tubulin (Pearson colocalization coefficient: 0.707 ± 0.023, Fig. 2B). Scale bar represents the distance of 5 µm. The localization of apicortin was confirmed in the nuclear and cytoplasmic extracts of the parasite by western blotting. Bands of apicortin were observed in both the fractions with lighter bands in nuclear extract indicating major occurrence of apicortin in the cytoplasm (Fig. 2C, Supplementary Fig. S9A). The coommassie stained polyacrylamide gel after transfer of the proteins prior to immunoblotting showed equal loading in all the cases (Fig. 2C, Supplementary Fig. S9B). The nuclear and cytoplasmic fractions were also probed with anti H4(histone) and anti-NapL (localized in cytoplasm) antibodies as bonafide nuclear and cytoplasmic loading controls (Fig. 2C, Supplementary Fig. S9C,D).

Confirmation of the binding of PfApicortin with Pfα-tubulin-I and Pfβ-tubulin by immuno-precipitation, ELISA and SPR.
Binding of PfApicortin with Pfα-tubulin-I and Pfβ-tubulin was confirmed by immunoprecipitation assay. Bands of Pfα-tubulin-I and Pfβ-tubulin were observed when pulled down with recombinant PfApicortin and subsequently probed with anti-tubulin antibodies (Fig. 3A,B, Supplementary  Fig. S10A,B). Presence of PfApicortin in stripped blot after detection of the tubulins further confirmed the interaction between the proteins (Fig. 3C,D, Supplementary Fig. S10C,D). Presence of tubulins in input lysate was confirmed by western blotting which showed bands of α-tubulin-I and β-tubulin ( Supplementary Fig. S7D,E). Further confirmation of the interaction was done by performing indirect ELISA for determining protein-protein interaction where recombinant PfApicortin was coated in protein binding plates and increasing titer of tubulins were used. Increase in absorbance was observed with increasing titers of tubulins. For α-tubulin-I, the difference in absorbance among different titers was little (Fig. 3E) whereas for β-tubulin, the absorbance increased with increasing titer of protein (Fig. 3F). In order to determine the rate of binding reaction, SPR was performed by immobilizing PfApicortin and flowing α-and β-tubulins respectively at increasing concentrations. K d value of apicortin-α-tubulin-I binding was 1.2 × 10 -7 M (Fig. 3G) and that of apicortin-β-tubulin binding was 7.5 × 10 -7 M (Fig. 3H) indicating strong interaction between apicortin and Plasmodium tubulins.

Binding of TMX with PfApicortin.
In order to find out a small molecule candidate for binding studies with PfApicortin and obtaining proper knowledge about the function of the protein, literature was surveyed. TMX was selected to study as a candidate binder of apicortin as it binds with the p25 activator domain and disrupts the interaction of CDK5 (cyclin dependent kinase5) with p25 22 . It also possesses anti-malarial activity but www.nature.com/scientificreports/ the mechanism of action is not well defined 23 . Different conformations of TMX were docked on apicortin surface consisting of one chain containing multiple beta sheets and helices. The best conformations of the compound were selected based on their lowest free binding energy to the catalytic pocket (Fig. 4A). TMX showed its binding around p25α domain shown as surface (highlighted in blue) in Fig. 4C. Analysis of this pocket unravelled five residues including and near p25 domain (ILE1, LEU81, ASP140, VAL143 and LEU165) that are involved in interaction with the apicortin (Fig. 4B,D, Supplementary Table S1). TMX showed lower free binding energy of   (Fig. 4E). Paclitaxel and Nocodaole (tubulin binders) were used as controls for   www.nature.com/scientificreports/ binding experiment but did not show efficient binding ( Supplementary Fig. S12), with negative response units upon binding). A coumarin analogue which is a known tubulin binder 24 was used as a negative control in this study which showed poor/nil interaction with apicortin (data not shown).

Confirmation of TMX binding with apicortin within parasite. The binding of TMX with PfApicortin
was further confirmed by cellular thermal shift assay (CETSA) which is a method for examining ligand binding with the target protein in cell lysate or tissue extract 25 . Erythrocytes infected with P. falciparum 3D7 trophozoites were treated with 10 µM TMX for 12 h and parasite lysate was prepared subsequently. Thermal shift assay was performed at 45 °C and 65 °C. In western blot analysis, significant difference in band intensity (Fig. 5A, S11A 5C,*p < 0.05) of PfApicortin was observed between the control (without TMX) and TMX treated samples for both the temperatures. Higher intensity of PfApicortin band in case of TMX treated sample indicates binding of TMX with parasite apicortin leading to protection of the protein from precipitation. As a loading control, GAPDH was used which showed no difference between control and the drug treated sample (Fig. 5B, Supplementary Fig. S11B).

Disruption of apicortin-tubulin interaction in presence of TMX.
TMX was used to monitor the disruption of apicortin-tubulin interaction. ELISA was performed with apicortin (pre-incubated with 10 µM TMX) coated in the microtiter plate and subsequent incubation with varying titer of tubulins was done. Decrease in absorbance intensity at 450 nm was observed in case of both the tubulins as compared to the control. In Fig. 5D,E, differences in the absorbance intensities has been shown (grey and black bars indicating absorbances of samples without TMX as well as in presence of it respectively). Significant reduction in absorbance intensities was observed for all the concentrations of both the tubulins (Fig. 5D,E) indicating possible disruption of the apicortin-tubulin interaction.
Microtubule assembly in presence of apicortin. In order to monitor the effect of apicortin on microtubule assembly, tubulin polymerization assay was performed with equal amounts of recombinant α-tubulin-I and β-tubulin in presence of increasing concentrations of recombinant apicortin (1-10 µM). Extent of polymerization was observed by measuring the absorbance of the reaction mixture at 340 nm. With the rising concentrations of apicortin, absorbance was found to increase within a time span of 20 min (Fig. 6A). Similar experiment was performed with apicortin pre-incubated with 10 µM TMX and absorbance was monitored. Lesser extent of polymerization was observed in presence of TMX with approximately 50% reduction in absorbance in presence of 10 µM apicortin as compared to the control group (Fig. 6B). Paclitaxel (10 µM) was used as a positive control for tubulin polymerization in both the experiments. Increased level of tubulin polymerization in presence of apicortin indicates stabilization of the microtubule assembly in the parasite through apicortin-tubulin interaction. Apicortin mediated microtubular stabilization was further confirmed through downregulation of apicortin by microRNA. In our previous publication, we identified a human microRNA candidate miR-197 causing downregulation of parasite apicortin 26 . Parasites grown in miR-197 enriched erythrocytes were probed for apicortin and α-tubulin-I through immunofluorescence assay. Downregulation of apicortin was observed with decreased fluorescence intensity in green channel in the parasites infecting miRNA loaded erythrocytes. The staining of α-tubulin-I was found to be diffuse in the miRNA targeted parasites whereas clear microtubular surface structures were visible in case of the control parasite (Fig. 6C). The diffuse staining of α-tubulin-I indicates probable destabilization of the microtubule assembly as an effect of repression in apicortin expression.
Growth inhibition and cell death of P. falciparum by TMX. Growth inhibition assay was performed by applying different concentrations of TMX in two strains of P. falciparum: 3D7 and RKL9 (resistant to chloroquine). Growth inhibition was observed in case of both the strains with IC 50 value 8.3 µM (Fig. 7A,B). Progression of the parasite was also found defective in case of 10 µM TMX treated parasites as compared to the control as less number of mature schizonts were observed die to intraerythrocytic cell death at 44 h post invasion (Fig. 7D). The relative percentage of schizonts was less as compared to control at 44-46 h post invasion (Fig. 7E). In the second cycle of infection, less number of rings were formed in comparison to control and uninvaded merozoites were observed in drug treated culture (shown in pie charts, Fig. 7D). Chloroquine was used as a positive control for the growth inhibition assay with 72% growth inhibition at a concentration of 20 nM in case of P. falciparum 3D7 (Fig. 7C). Scale bar in Giemsa images represents the distance of 5 µm.

Discussion
Due to the rapid emergence of malaria parasite resistance to known anti-malarials, developing newer therapeutics is essential. For that, identification of more number of targets with diverse functions is necessary. Cytoskeletal proteins play important role in parasite growth and invasion due to the formation of specialized structures like apical complex, rhoptries, microneme etc. Targeting cytoskeletal proteins (tubulins) is a prevailing strategy for designing drugs against cancer. Application of cytoskeletal protein inhibitor or microtubule binder was also found to be effective in case of malaria parasite 27 . The presence of α-and β-tubulins among different Apicomplexan protists ( Supplementary Fig. S2B,C) makes them ideal component for targeting with inhibitors. The cytoskeletal assembly consists of cylindrical structures of 24 nm diameter called microtubules. Each microtubule is composed of 13 protofilaments which consists of equal amount of α-and β-tubulin. The polymerization process of microtubule is a GTP mediated process occurring at MTOC (microtubule organizing centre). Microtubule inhibitors work either by stabilizing the microtubule assembly like Taxol or disrupting the polymerization process like Nocodazole 28,29 . Therefore, tubulin polymerization and stability of microtubular structure are crucial factors for the survival of the parasite. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Several proteins are associated with the stabilization of microtubular structure. In vertebrates, proteins of MAP/Tau family are responsible for microtubule stabilization and regulation of microtubular network, recruitment of signaling proteins and execution of microtubule driven transport 30 . Similarly, Apicomplexan parasites also possess various proteins for microtubule stabilization and maintenance of specialized structures like conoid, rhoptries, microneme etc. used for invasion 17 . In Toxoplasma, two novel proteins SPM1 and SPM2 are present in the conoid to stabilize the subpellicular microtubular structure 31 . In Theileria, CLASP1 protein is responsible for the attachment of microtubules with the kinetochore and proper positioning of mitotic spindle 32 . Disruption of these proteins leads to impaired growth and invasion of the parasites. On the basis of these evidences, apicortin was selected as a target for studying further as it has microtubule binding domains, p25α and DCX with putative microtubule stabilization property. Moreover, its unique presence in Plasmodium and other Apicomplexan protists makes it a novel target to study ( Supplementary Fig. S2A, S3).
In silico and experimental data suggest a strong interaction of apicortin with α-and β-tubulin of P. falciparum. However, SPR and ELISA data indicate somewhat stronger interaction between apicortin and α-tubulin-I as compared to β-tubulin (lower K d value for Apicortin-α-tubulin-I binding in SPR). The reason behind this variation may be the refolded conformations of the proteins. In in silico experiments, an ensemble of different conformations of both the interacting proteins has been considered. In experimental conditions, fluctuation from the predicted conformations of binding may happen due to experimental conditions like temperature, pH, salt concentration etc 33 . Moreover, refolding of the proteins may give rise to a somewhat different conformation from that of the predicted binding mode. In that case, stronger interaction of apicortin and α-tubulin-I than that of apicortin and β-tubulin may result due to conformational fluctuations of apicortin as well as tubulins in course of the refolding process. Furthermore, the presence of disordered regions in proteins often influence the protein-protein interactions 33 . Plasmodium apicortin possesses a disordered region at its N terminal 8 . Presence of the disordered extension often set limitations in in silico prediction of binding of two proteins 33 . Refolding of the disordered extension might influence the apicortin-tubulin interaction which is evident as a varied binding www.nature.com/scientificreports/ affinity of apicortin with the tubulins in comparison to the in silico predictions. As α-tubulin-I is abundant in asexual stage of the parasite, disruption of its interaction with apicortin might affect the invasion and progression of the parasite in erythrocytes. The interaction of Plasmodium apicortin with parasite microtubule components α-and β-tubulins is an interesting area which needs to be explored further. Individual domains of apicortin are known to stabilize microtubules or cause polymerization of tubulins to different structures in brain cells 7 . Orthologue of apicortin in T. gondii known as TgDCX was found to stabilize the tubulin dimers of the microtubular structures present in the conoid of the parasite. Loss of TgDCX through knock out resulted in abnormal conoid structure along with loss of tubulin polymers from the conoid fibers. The growth and invasion of the parasite were also hindered 14 .
In the previous work, we identified human microRNA candidates causing reduction in apicortin expression 26 . The growth and invasion of the malaria parasite was stalled in response to the downregulation of apicortin. In this work, we investigated further checking for the interaction of apicortin with plasmodium tubulins present in erythrocytic stages. Interaction of apicortin with α-tubulin-I and β-tubulin was confirmed by in silico docking studies as well as experimental techniques such as immunoprecipitation, ELISA and surface plasmon resonance. In order to understand the effect of apicortin on microtubular assembly stabilization, polymerization assay was performed in presence of apicortin which showed higher extent of tubulin polymerization with the rising concentrations of apicortin (Fig. 6A). The polymerization process was also affected in presence of TMX, an apicortin binder used in this study (Fig. 6B). Furthermore, downregulation of apicortin by miR-197 resulted in loss of the organized structure of tubulin (Fig. 6C). On the basis of these results, it can be inferred that Plasmodium apicortin plays critical role in microtubular structure stabilization through the interaction with parasite tubulins.
In order to find out small molecule inhibitors, binders of individual domains of apicortin were searched. TMX was selected as it has p25 domain binding property as well as anti-malarial activity 22,34 . In earlier studies, the activity of TMX against P. falciparum and P. berghei had been shown [34][35][36][37][38] . Although the activity of TMX was reported at higher concentration 38 , other studies documented its inhibitory effect at 10 µM concentration 34,35,37 .
In an ex vivo cell line infection model with P. berghei infection 36 , TMX showed IC 50 value at 4 µM. In our work, the IC 50 of TMX was determined as 8.3 µM and the binding studies of apicortin with TMX showed effective result at a TMX concentration of 10 µM supporting the previous studies. Moreover, docking studies, cellular thermal shift assay and SPR experiment confirmed the binding of TMX with apicortin. From ELISA and polymerization assay data (Figs. 5D,E, 6B), it can be inferred that the binding of TMX with apicortin disrupted the interaction of apicortin and both the tubulins by hindering tubulin assembly which led to the death of the parasite (Fig. 7). The possible reason behind TMX mediated disruption of apicortin-tubulin interaction might be the conformational change of apicortin due to the binding of TMX at p25α domain. Similar molecules can be designed to develop novel anti-malarial therapeutics targeting an important and unique parasite protein. Also, any compound binding with the DCX domain might inhibit the apicortin-tubulin interaction in this way and serve as a probable anti-malarial therapeutic agent.

Summary
It can be concluded from the current study that apicortin, an Apicomplexan parasite specific protein interacts with P. falciparum α-tubulin-I and β-tubulin inducing tubulin polymerization and stabilization of microtubules. Disruption of this interaction hinders the growth and progression of the asexual blood stage parasites (Fig. 8). Antibody generation of the proteins. The purified proteins were injected into BALB/c mice (6 weeks, female) and they were bled twice followed by the first and second booster. Tubulins were also injected in white albino rabbits in order to raise anti-rabbit sera for tubulins. Antibodies of apicortin and tubulins were validated by western blot and immunofluorescence assay published in our previous articles 24,26 . All the experimental steps regarding immunization of mouse and rabbit were performed according to CPCSEA guidelines and approved by Institutional Animal Ethics Committee (IAEC), Jawaharlal Nehru University, New Delhi.

Cloning, expression and purification of PfApicortin, Pfα-tubulin-I and
Parasite culture and growth inhibition assay. Malaria  Preparation of cytoplasmic and nuclear extracts. Fractionation of P. falciparum cell lysate was performed following previously published protocols 40,41 . Briefly, sorbitol synchronized P. falciparum 3D7 culture of ring trophozoite and schizont stages were pelleted down and lysed with the help of lysis buffer composed of 20 mM HEPES, 10 mM KCl, 1 mM EDTA, 1 mM DTT (Sigma Aldrich, USA) along with protease inhibitor (Thermo, USA). The cells were incubated in the buffer for 5 min in ice followed by centrifugation at 2500g at 4 °C for 10 min. The cytoplasmic extract was present in the supernatant obtained which was separated and stored at − 20 °C. The pellets were further washed twice with the lysis buffer and incubated in that buffer for 30 min at rotating condition in 4 °C. The suspension was centrifuged at 12000g for 30 min and the nuclear extract was obtained in the supernatant. The extracts were used for western blotting after protein estimation by Bicinchoninic acid assay kit (G Biosciences,India) following the manufacturer's protocol.
Immunoprecipitation. Plasmodium tubulins (α and β) were detected by pulling down with purified recombinant apicortin from P. falciparum cell lysate. Briefly, recombinant PfApicortin (50 µg) was bound with Ni-NTA beads (50 µl packed) followed by washing of the beads with PBS to remove unbound protein. Parasite lysate (100 µg) was incubated with the beads at 4 °C overnight and beads were further washed to remove unbound proteins. The beads were finally boiled after adding 5 × loading dye and the supernatant was loaded in polyacrylamide gel (12%). Presence of α-and β-tubulin was detected through western blotting using anti-mouse polyclonal antibodies raised against the proteins. Ni-NTA beads incubated with PBS only (without recombinant apicortin) were used as control. The input lysate incubated with both types of beads was run in gel and immunoblotted to detect the presence of α-and β-tubulin as input controls ( Supplementary Fig. S7D,E). Estimation  Western blotting. Samples derived after immune-precipitation and CETSA experiments were run in 12% SDS-polyacrylamide gel and the proteins in gel were subsequently transferred to nitrocellulose membrane (Biorad, USA). Blocking was done in 5% skim milk (Himedia, India) for 2 h at room temperature and the blot was incubated with the primary antibody (anti-apicortin and anti tubulin, 1:5000) overnight at 4 °C. The blot was further incubated with HRP conjugated anti-mouse secondary antibody for 1 h at room temperature (1:5000, Sigma-Aldrich) after washing with PBST. The blot was washed further and antibody binding was detected using Enhanced Chemiluminescence Kit (Biorad, USA).
Immunofluorescence assay. Smears of infected erythrocytes were prepared and fixation was done by dipping the smears in chilled methanol for 20 min at − 20 °C. Smears were dried and blocked in 3% BSA (VWR, Pennsylvania, USA) solution in PBS overnight at 4 °C. Smears were incubated with anti-apicortin and antitubulin primary antibody for 2 h at room temperature. Further incubation with secondary antibody (conjugated with Alexa488/546, Invitrogen, USA) was done at room temperature for 1 h followed by washing with PBS. Slides containing smears were mounted using DAPI-antifade (Invitrogen, CA, USA). Images were captured in confocal microscope at 100 × magnification (Olympus Corporation, Tokyo, Japan) and analysed in CellSense Dimension 3 and ImageJ software.
Surface Plasmon Resonance (SPR) spectroscopy. Surface Plasmon resonance spectroscopy was performed to monitor the interaction of apicortin and tubulin as well as apicortin and TMX. Immobilization of Apicortin (12 µM) was done on the activated (through amine coupling) gold sensor chip. Different concentrations of α-tubulin-I (100 nM, 250 nM, 500 nM, 750 nM, 1 µM), β-tubulin (100 nM, 250 nM, 500 nM, 750 nM, 1 µM) and TMX (0.5 µM, 1 µM, 5 µM, 10 µM) were injected over the chip surface in order to monitor the interaction both for the association and dissociation throughout 500 s using the Auto LAB ESPRIT SPR instrument (Kinetic Evaluation Instruments BV, The Netherlands) with an open cuvette system along with electrochemical resonance facility. PBS was used for immobilization and binding solution. The surface was regenerated with 50 mM NaOH. Data were analyzed using Auto Lab ESPRIT Kinetic evaluation software.
Tubulin polymerization assay. Polymerization of recombinant tubulins were performed following the protocol published previously in studying microtubule associated proteins 43,44 . Polymerization reaction was set up with 10 µg of recombinant α-and β-tubulin and increasing concentrations of recombinant apicortin (1 µM, 2 µM, 4 µM, 6 µM, 8 µM and 10 µM) in presence of 1 mM GTP (Thermo, USA) and assembly buffer consisting of 100 mM PIPES, 1 mM EGTA and 1 mM MgCl 2 (Sigma-Aldrich, USA) with pH 6.8. The reaction mixture was incubated at 37 °C and absorbance was measured at different time points at 340 nm in UV-Vis spectrophotometer (Cary5000, Agilent technologies, USA). The assay was also set up in different concentrations of recombinant apicortin pre-incubated with 10 µM TMX (Cayman Chemical). One reaction was set up in presence of Paclitaxel (Sigma-Aldrich, USA) as a positive control of tubulin polymerization.
Loading of erythrocytes with miR-197 mimic. Packed erythrocytes were loaded with miR-197 mimic following the protocol mentioned in our previous publication 26 . Erythrocytes were lysed and filled with the cargo mimic followed by resealing and storage at 4 °C for growth inhibition assay. www.nature.com/scientificreports/ In silico experiments. Ab initio modelling of proteins was performed using I-TASSER server 45 . Protein sequences of Plasmodium falciparum apicortin (PF3D7_0517800), Plasmodium falciparum α-tubulin-I (PF3D7_0903700) and Plasmodium falciparum β-tubulin (PF3D7_1008700) were obtained from Uniprot 46 and also confirmed from the cloned sequence. Chemical structures of TMX was prepared through ChemSketch 47  Compliance with ethical standards. All the experiments were carried out in accordance with the guidelines and regulations of Jawaharlal Nehru University and approved by institutional IBSC committee. Animal handling and sera generation were performed as per CPCSEA guidelines and approved by the Institutional Animal Ethics Committee (IAEC), Jawaharlal Nehru University, New Delhi. All the steps of animal handling and sera generation were performed in compliance with the ARRIVE guidelines.