A screen of MMV box, pandemic box, and other synthetic compounds identifies eight anti-malarial candidates that target Pf Glideosome-associated protein, GAP50.
Since gliding machinery for different stages of Plasmodium spp possesses some conserved proteins such as GAP45, GAP50, and several IMC proteins, we aimed to identify a common inhibitor that can block gliding at all the life cycle stages of Plasmodium spp. Here we chose to target a conserved protein, PfGAP50, that anchors surface proteins to IMC. Screening of 951 compounds (Fig. 1a) from diverse chemical libraries like pathogen box, pandemic box, Azepino quinolines, quinolines with β benzomorphan framework (SI-1, table 1), and secondary metabolites purified from C. pareira34 for interactions with PfGAP50 via Schrödinger drug discovery suite (version 2019-3), differential scanning fluorimetry (DSF) for drug(s) interaction followed by cell-based in vitro screening assays led to the identification of eight promising molecules (Fig. 1b). In the absence of any known inhibitors and key binding sites on GAP50, permissive cut-off values for in silico screening were kept at < -20 kcal/mol (MMGBSA ΔG of binding), while a δ Tm cut-off kept at ± 2°C was chosen for DSF-based screening to ensure that nothing was missed. The independent use of both in silico docking and DSF (δ Tm) helped us to select 345 of the 951 compounds.
Since both in silico docking and DSF can occasionally give false positives35–37, we used Surface Plasmon Resonance38 to validate the binding of the shortlisted 345 molecules to recombinant PfGAP50 protein immobilized on the CM5 chip. GAP50 consists of an N-terminal signalling sequence, followed by a phosphatase domain and a C-terminal transmembrane domain (Fig. 1a). The phosphatase domain of GAP50 was expressed and purified to homogeneity in soluble form using two chromatography steps; Ni-NTA+ purification followed by hydrophobic chromatography (Figs. 2b, 2c & 2d).
Initial screening of 345 different molecules by SPR at 50 µM of each compound allowed the selection of 266 molecules that exhibited 1:1 binding with GAP50. Further, these 266 molecules 84 molecules were identified which were binding with GAP50 and were having activity against Pf (SI-2). We next screened all these 84 molecules for their antiplasmodial IC50s against different strains of the malaria parasite (chloroquine-sensitive Pf3D7, chloroquine-resistant PfINDO, and artemisinin-resistant PfCam 3.1R539T strains) (SI-2) leading to 39 compounds that showed IC50 against P. falciparum parasites of < 20 µM for further toxicity testing against human kidney and liver cell lines, using standard MTT assay39. Based on the selectivity index (> 100-fold) for P. falciparum parasites eight compounds; Hayatinine, Curine, MMV689758 (Bedaquiline), MMV1634402 (Brilacidin), and MMV688271, MMV782353, MMV642550, and USINB4-124-8 were selected for further analysis (Fig. 1b). These molecules exhibited PfGAP50 binding affinity of < 75 µM (Fig. 2e) with IC50 against Chloroquine sensitive and resistant strains < 3 µM (Table 1). It is noteworthy that among the shortlisted molecules (Table 1), Bedaquiline, Brilacidin, MMV688271, MMV642550 and Hayatinine showed IC50s ≤ 1.5 µM against all three strains of Plasmodium. Among our shortlisted compounds, MMV688271 have already been shown to possess antiplasmodial activity comparable to what we saw in our study40. USINB4-124-8 and Hayatinine were found to be more active against PfCam 3.1R539T (IC50 6.7 nM and 59 nM, respectively) than against PfINDO strain (IC50 2.42 µM and 0.41 µM, respectively). In comparison, MMV1782353 and Curine showed IC50s < 1 µM for PfINDO and Pf3D7 (IC50 of 0.64 µM and 0.34 µM, respectively) strains and exhibited IC50 of 21.25 µM against PfCam 3.1R539T strain. The differential activity seen for Hayatinine or Curine or MMV1782353 or USINB4-124-8 could be due to differential expression and modifications of protein(s) in different parasite strains resulting in alteration of parasite metabolic pathways.
Table-1: In vitro anti-plasmodial IC50, Mammalian cell CC50, Resistance Index (RI) and Selectivity index (SI) of top eight PfGAP50 binder
S.No.
|
Test molecule
|
IC50 (µM) against Pf strains
|
Resistance index*
|
CC50 (µM)
|
Selectivity index**
|
3D7
|
INDO
|
Cam3.1R539T
|
A
|
B
|
HEK293
|
HUH
7
|
A
|
B
|
1.
|
Hayatinine
|
0.5 ±
0.1
|
0.41 ± 0.06
|
0.059 ±
0.013
|
0.82
|
0.16
|
37.3 ± 9.42
|
42.71
±
4.22
|
90.84
|
104.00
|
2.
|
USINB4-124-8
|
0.27 ± 0.02
|
2.42 ± 0.26
|
0.0067 ±
0.0004
|
8.96
|
0.02
|
> 400
|
> 400
|
> 165.28
|
> 165.28
|
3.
|
Bedaquiline
|
1.22 ± 0.12
|
1.09 ± 0.03
|
0.72 ± 0.06
|
0.89
|
0.59
|
> 400
|
> 400
|
366.97
|
366.97
|
4.
|
MMV1782353
|
0.63 ± .05
|
0.49 ± 0.03
|
13.39 ±
2.70
|
0.64
|
21.25
|
> 200
|
> 200
|
500.00
|
500.00
|
5.
|
MMV642550
|
0.39 ± 0.04
|
1.2 ±
0.09
|
1.49 ± 0.06
|
3.076
|
3.82
|
> 200
|
> 200
|
166.67
|
166.67
|
6.
|
Brilacidin
|
0.45 ±
0.024
|
0.67 ± 0.03
|
0.865 ±
0.042
|
1.48
|
1.92
|
> 200
|
> 200
|
298.51
|
298.51
|
7.
|
MMV688271
|
0.33 ± 0.05
|
0.25 ± 0.04
|
0.337 ±
0.034
|
0.75
|
1.02
|
119.1 ±
33.72
|
29.65
±
3.76
|
478.31
|
119.08
|
8.
|
Curine
|
1.46 ± 0.10
|
0.51 ± 0.04
|
2.8 ±
0.11
|
0.34
|
1.91
|
104.1 ±
6.90
|
62.77
±
6.60
|
204.1
|
123.07
|
9.
|
Chloroquine#
|
0.014 ±
0.005
|
0.290 ±
0.02
|
0.055 ±
0.004
|
20.71
|
3.92
|
45.69 ±
3.12
|
58.54
±
4.73
|
157.55
|
201.86
|
10.
|
Artemisinin#
|
0.013 ±
0.004
|
0.010 ±
0.002
|
0.01 5 ±
0.002
|
0.77
|
1.15
|
> 100
|
> 100
|
> 6,666.66
|
> 6,666.66
|
*Resistance index (RI): A = IC50 PfINDO/IC50 Pf3D7; B = IC50 PfCam3.1R539T/IC50 Pf3D7 **Selectivity index (SI): A = CC50 HEK293/IC50 PfINDO; B = CC50 PfHUH7/IC50 PfINDO #Reference compounds |
Ring stage survival assay of lead molecules
Ring-stage Survival Assay41 (RSA) has been developed to analyse the response of Artemisinin on the survival of early rings (0 to 3 h p.i) on distinct parasite strains. To test the efficacy of selected compounds at P. falciparum ring stages, early rings (0 to 3 h p.i) of ART resistant (PfCam3.1R539T) were treated with molecules at their 10X IC50, with Dihydroartemisinin and Artemisinin at 700 nM serving as controls. Among the eight compounds, Bedaquiline was the only molecule that showed activity against early rings by RS assay (Fig. 3). Hayatinine and USINB4-124-8 were found to have no effect in RSA despite having nM activity against PfCam3.1R539T. Results of RS assays thus indicated that besides binding to PfGAP50, these compounds may be targeting other parasite proteins. It is also possible that PfGAP50 is playing another role(s) besides having a role in parasite gliding mobility.
Validation of selected drugs for their action against asexual blood stages of P. falciparum
To explore the effect of lead antiplasmodial molecules on other asexual blood stages of P. falciparum, we subjected synchronised different cell cycle stages to each lead molecule at different concentrations for 12 h followed by evaluation of % growth by SYBR green lysis as indicated in method. As shown (Fig. 4a), there were marked differences in the observed IC50s against different stages for these compounds. Taking early rings (E.R.) to late rings (L.R.), early trophozoites (E.T.) to late trophozoites (L.T.), early schizonts (E.S.) to late schizonts (L.S.), and late schizonts (L.S.) to early rings (E.R.) stages into consideration, USINB4-124-8 showed a much higher IC50 (µM) (38.12) against trophozoites and schizonts stages (54.32) than against rings (7.2). In contrast, MMV1782353 and MMV642550 showed significant activity (IC50 < 5 µM) against early and late schizonts while showing lesser potency (IC50 > 8 µM) against rings (Fig. 4a). It is noteworthy that despite the remarkable structural resemblance, Hayatinine and Curine showed marked differences in activities at different asexual blood stages (Fig. 1b). Thus, Hayatinine (608.7 Da) showed a marked preference for preventing L.S. to E.R. transition. In contrast, the Curine (594 Da) inhibited E.S. to L.S. along with E.T. to L.T. transitions. In summary, except for USINB4-124-8, all other shortlisted molecules showed potent activity against late trophozoites and schizonts stages of PfINDO parasites.
These molecules were studied further for their effects on parasite morphology/ development at different asexual stages of P. falciparum. Briefly, ring-stage parasites (6 to 12 h, p.i) were treated for a period of 57 h, and microscopic observation of Giemsa-stained slides was made at 12, 36, 48, and 57 h (Fig. 4b). Bedaquiline, and Hayatinine, allowed the proper development of parasites but prevented the invasion of merozoites. Thus, cultures treated with these two molecules showed RBC surface-attached merozoites even at 57 h. In the case of Curine and Brilacidin treatments, partial inhibition of invasion was seen as reflected in the observation of merozoites attached to RBC. MMV688271 prevented schizont egression, while MMV1782353 and MMV642550 allowed proper development of Pf at their respective IC50 concentration.
Together, these results indicated that despite being screened against PfGAP50 protein, eight lead compounds inhibited the development of different asexual blood stages of P. falciparum.
Differential action of lead compounds on P. falciparum invasion of human RBCs, gliding and egress
Since the invasion of merozoites is a crucial step preceding the exponential rise in % parasitemia, molecules with IC50 < 10 µM against schizonts (Fig. 4a) were studied to examine (a) inhibition of egress of merozoites from schizonts and (b) merozoite invasion into healthy RBCs. Microscopic interpretation and quantification by FACS showed that schizont rupture was stalled in the presence of MMV molecules 1782353 and 642550. However, upon three centrifugal washings to remove the test molecules, the schizonts were found to rupture and release merozoites. However, all the released merozoites could not invade, leading to a sharp decrease in % parasitemia akin to heparin, a known inhibitor of invasion42 (Fig. 5a and 5b). In the case of Hayatinine, 8 h treatment of schizonts led to inhibition of egress with microscopic images showing many schizonts and just a few merozoites stuck to the RBC membrane. Curine allowed schizont egress, but the level of invasion was significantly reduced, suggesting that the treatment rendered the merozoites non-invasive. With no egress and no invasion, both Bedaquiline and MMV688271 showed schizontocidal activity on mature (about to rupture) schizonts (Fig. 5a). These results were corroborated by co-staining the drug(s) treated parasites with anti-PhiL antibody and DAP1. As shown (Fig. 5b), Bedaquiline and Hayatinine prevented an active invasion by inhibiting gliding motility. We observed several unruptured schizonts upon Curine or MMV688271 treatment (Fig. 5b). In the case of Brilacidin, Curine, and MMV688271 treated parasites, even schizogony was hindered as pyknotic schizonts were seen in treated cultures. In summary, we identified five molecules, Hayatinine, Bedaquiline, Brilacidin, Curine, and MMV688271 that completely blocked invasion/schizont rupture at IC50, whereas MMV1782353 and MMV642550 were able to block parasite invasion partially.
Since PfGAP50 is a protein involved in forming IMC during schizogony and anchors several crucial surface proteins involved in invasion28, we next analysed our lead GAP50 binders (IC50 against schizonts < 10 µM) at their sublethal concentrations for their ability to perturb IMC formation. Recent publications17,43 have shown co-localization of the GAP50 complex with the Phil1 complex at the schizont stage. These complexes have been shown to work in tandem to facilitate the process of invasion. Hence, in control vs. test molecule treated schizonts, co-localization studies were performed for Phil1 and GAP50 proteins. As demonstrated (Fig. 5c), 0.4 % DMSOtreated control P. falciparum schizonts displayed proper development of schizonts and near complete co-localization of GAP50 and Phil1 (P-value 0.824). Similar co-localizations were observed also in the case of MMV688271 (P 0.886) and Curine (P 0.888) treated schizonts. In comparison, Hayatinine (P 0.617), Bedaquiline (P 0.641), and Brilacidin showed perturbations in IMC formation; hence, co-localization was also affected. The IMC disrupting activities of these compounds can thus be attributed to their ability to bind GAP50 or other proteins involved in proper IMC formation, thus inhibiting its activity or functions in gliding and invasion.
Based on compounds IMC disrupting and invasion inhibitory activity, Bedaquiline, Hayatinine, Curine, MMV688271, and Brilacidin were selected for further studies.
Interaction of lead molecules with GAP50 as seen through molecular dynamics (M.D.) simulation
To understand the interaction stability of lead invasion inhibitory compounds (Bedaquiline, Hayatinine, Curine, MMV688271, and Brilacidin) with PfGAP50 protein, blind docking of compounds was done followed by M.D. stimulation studies.
Bedaquiline was found to be interacting through hydrophobic interactions (W35, I69, H256, M278), hydrogen bonding (G65), and a water bridge (N221) (SI-1, Figs. 1 and 2). These interactions were maintained over 30% of 100 ns simulation time and appeared to be helping in stabilizing the binding of Bedaquiline with GAP50, resulting in -36.9 kcal/mol ΔG of binding energy. The sensorgram profile obtained using one-on-one interaction with PfGAP50 indicated a SPR based binding affinity (KD) of 9.27 µM (Fig. 2e). The binding of Brilacidin with PfGAP50 (SI-1, Figs. 3 and 4) was found to be majorly driven by water bridges mediated via the side chains and the backbone of N127, E129, N148, D201, I322, N349, E351, and L352, which were further strengthened by Hydrogen bond interactions involving E129, E351, and L352. Among these interactions, those with E129 and E351 were maintained for > 50% of 100 ns simulation time (SI-1, Fig. 4). As many as sixteen water bridges between Brilacidin and GAP50 may have contributed to the most promising MMGBSA ΔG binding of -64.82 kcal/mol (SI-2), with KD 2.81 µM (SI-2).
Interactions of MMV688271 (SI-1, Figs. 5 and 6) with PfGAP50 were found to be stabilized through Hydrogen bonds (E123, H152, F344, L347, P348), water bridges (E123, H152, F344, L347), and hydrophobic interactions (H152, F344, L347, V350). Varied interactions of MMV688271 with PfGAP50 might be why this compound showed the low MMGBSA ΔG binding of 41.86 kcal/mol with in vitro KD of 20 µM. Interestingly, even as the structurally close Curine and Hayatinine bind to PfGAP50, each bind to its unique site (SI-1, Figs. 7, 8, and 9). Further, SPR revealed that the KD of Curine was nearly three times higher than that of Hayatinine, suggesting Hayatinine interaction with PfGAP50 is more robust than the case with Curine (Fig. 2e). Hayatinine was found to interact with G119, Q120, M146, P147, and H152 via water bridges and E133 via hydrogen bonds (SI-1, Fig. 8). These bonds, which were found to persist for > 30% of the simulation time, might prevent Hayatinine from moving out of its surface binding site (SI-1, Fig. 7b). Residues interacting with Curine (SI-1, Fig. 7c) via hydrogen bonds, hydrophobic interactions, and water bridges were Y96, L134, D135, D137, A138, V350, and E351. Our in silico docking and simulation studies predicted the binding interaction of lead molecules to GAP50 which can help in designing more potent and GAP50-specific inhibitors. However, structural data is needed to confirm their binding modes and poses predicted by docking before going for designing new inhibitors.
Lead molecules create a prooxidant milieu in infected erythrocytes
Cells work best at optimum Redox conditions created by homeostasis mediated by the combined actions of pro- and anti-oxidant mechanisms. Any imbalance in this homeostasis is lethal for cell survival44. Although the parasite is adept at regulating its redox status via diverse mechanisms, the skewed prooxidant status stimulated by the lead molecules could overwhelm the parasite's redox homeostasis mechanisms and assist in killing the parasite. Given the robust heme detoxification mechanism that converts the toxic heme to hemozoin along with Thioredoxin and Glutathione dependent antioxidant systems in the malaria parasite's metabolically active trophozoite stage44, we tested whether these lead molecules (Bedaquiline, Hayatinine, Curine, MMV688271, and Brilacidin) were generating a prooxidant milieu. Briefly, trophozoites were pre-treated with a redox-sensitive H2DFCDA for 20 minutes and the dye was washed. Pre-treated trophozoites were then treated with each lead compound with 2x IC50 for 8 h and after the treatment dye fluorescence was measured using a fluorimeter. Except for Hayatinine, parasites treated with lead compounds showed a significant increase in fluorescence in comparison to untreated parasites (Fig. 6), indicating a change in redox potential in drug-treated parasites.
The combinational activity of lead molecules
To retard the development of resistance, reduce toxicity, and increase the efficacy of anti-malarial drugs, it is preferred to use drug combination(s) to treat malaria. Hence, we evaluated the effect of five lead compounds that block invasion in combination with three current antimalarials; Pyrimethamine (PYR), Mefloquine (MFQ), and Artemisinin (ART), on the PfCam3.1R539T strain, which is ART, Sulphadoxine and Pyrimethamine resistant. As shown in SI-1, Table 2, these drug combinations showed synergy (∑FIC < 1), additivity (FIC > 1 < 2), and antagonism (∑FIC < 2). When the ∑FIC values were averaged over varying ratios of drug combinations (Fig. 7), the combinations of Hayatinine and Curine with Artemisinin turned out to be synergistic (average ∑FIC ≈ 0.5) and additive (average ∑FIC ≈ 1) with PYR and MFQ. Likewise, the average ∑FIC values for Bedaquiline, Brilacidin, and MMV688271, in combination with PYR, MFQ, and ART, were additive. The additive/synergistic actions of these drugs in combination can be due to different targets in other metabolic pathways besides having PfGAP50 as a common target. This additive activity may allow us to use these compounds at their optimal concentrations with little risk of resistance development and fewer side effects/toxicity. Taken together, this study provides a lead for a new combination therapy, which may be highly effective against drug-resistant strains of Plasmodium falciparum.
Putative antiplasmodial targets of Bedaquiline, Hayatinine, Brilacidin, MMV688271, and Curine
The sub micromolar antiplasmodial IC50 of some of the compounds, together with µM K.D. against PfGAP50, and multi stage differential activity of invasion inhibitory molecules led us to surmise that the malaria parasite proteome could have some more targets in addition to PfGAP50 for them. Hence invasion inhibitory molecules from each library (Bedaquiline and MMV688271 (pathogen box), Brilacidin (pandemic box), Hayatinine, and Curine (phytometabolites from C. pareira) were taken for cross-docking against 68 different malaria parasite proteins. The co-crystallized ligand of each of these proteins was docked to its respective protein to generate docking scores that were used as cut-offs for our compounds. Each lead molecule was docked into the active sites of selected proteins, and the docking scores were calculated. Out of the 68 proteins studied (Supplementary Table 3), the number of different protein-ligand complexes with docking scores < that of co-crystallized ligand were 20 for MMV688271, 17 for Bedaquiline, 8 for Brilacidin, 7 for Hayatinine, and 10 for Curine (Supplementary Tables-4-8). Among the shortlisted proteins, three (Plasmepsins V, Cyclophilin, and Malaria Sporozoite Protein Uis3) were common for all five molecules (Fig. 8b). These three targets are known to be expressed only in trophozoites and schizont stages45 (Fig. 8a), and binding of the test molecules to them might explain their cidal activity towards mature stages of P. falciparum. We also identified peptide deformylase, Pyruvoyl Tetrahydropterin Synthase (PTPS), and glutaredoxin 1, uniquely expressed in trophozoites the putative targets of test molecules (Fig. 8a). To find the binding affinity of our lead molecules to their putative target proteins, MMGBSA ΔdG binding was computed. (Supplementary Tables 4–8). In all cases, dG binding was < − 30 kcal/mol (Fig. 8c). This low -ve score suggests that the identified proteins might be the targets for our lead molecules. However, extensive studies will be required to confirm the status of these proteins as drug targets for these molecules.
Bedaquiline treatment protects mice in the P. berghei-mouse model
Based on promising in vitro antiplasmodial potencies and SI, we next tested the protective potential of Bedaquiline and USINB4-124-8 in an in vivo P. berghei mouse model. Briefly, seven mice in four groups were intraperitoneally injected with 105 P. berghei parasites. Parasitemia was monitored daily, and when it reached 1%, drug treatment was started. Each drug was given orally for four consecutive days. Vehicle solution (2% (Hydroxypropyl) methylcellulose with 2% Tween 80 in Normal saline) was given without the drug in the negative control group. Parasitemia was monitored daily for forty days for each mouse. Treatment of P. berghei ANKA infected mice with vehicle solution (-ve control) led to increased fluctuations in body temperature and a decrease in mean body weight concomitant with the progressive increase in % Parasitemia with a median survival of 19 days and death of all 7 mice by day 22 (Fig. 9). Likewise, USINB4-124-8 fed mice also showed no difference from -ve control in terms of survival, % rise in parasitemia, fluctuation in body weight, and temperature (Fig. 9). In comparison, Bedaquiline (50 mg/kg b. wt.) treatment resulted in the complete suppression of % Parasitemia by day 19, reduced temperature and body weight fluctuations compared to -ve control, and survival of three out of seven mice till 34 days (Fig. 8). Together these results showed partial protection of mice treated with Bedaquiline at 50 mg/kg dose orally, thereby suggesting Bedaquiline can be carefully considered for combination therapy as well as for further chemical modifications. Bedaquiline is well tolerated at 700 mg dosage in humans (Cmax 6.47 mg/L) for MDR-TB treatment47,48 and is known to clear TB infection in mice model at 25 mg/Kg49. In the current study, Bedaquiline treatment at 50 mg/kg body weight for four days provided partial protection. It will be interesting to conduct future studies by increasing the dosage of Bedaquiline and tweaking its structure to see if this results in significant/complete protection in the P. berghei model of malaria.