Antimalarial cytoskeletal targeting with broad apicomplexan activity

Significance

Malaria parasites are becoming resistant to current medicines, creating an urgent need for drugs that work in new ways. This study shows that blocking parasite myosin A, a motor protein needed for movement and host-cell invasion, can block Plasmodium parasite development across several key stages of its life cycle. The compound KNX-115 acts potently and selectively on the parasite target, kills blood-stage parasites, blocks liver-stage infection and mosquito-stage transmission, and remains active against strains resistant to existing antimalarials, including patient isolates. Because related parasites use closely similar motors, KNX-115 also inhibits Cryptosporidium, Toxoplasma, and Eimeria. These results establish myosin A as a promising drug target and suggest a path toward broad therapies for apicomplexan parasitic diseases.

Abstract

Malaria is a devastating disease that resulted in an estimated 610,000 deaths in 2024, the majority being children under the age of five. Here, we use KNX-115 to illustrate multistage antiparasitic activity upon targeting the cytoskeletal enzyme Plasmodium falciparum myosin A (PfMyoA). KNX-115 inhibits purified actin-activated ATPase with a potency in the low nanomolar range and >50-fold selectivity against cardiac, skeletal, and smooth muscle myosins. KNX-115 traps PfMyoA in a state that binds weakly to actin. A 2.35 Å resolution structure of KNX-115 bound to PfMyoA reveals critical interactions contributing to its mechanism of action. Importantly, in vitro evolution data reveal that KNX-115 engages PfMyoA as a sole cellular target. Inhibiting PfMyoA blocks the development of the blood and liver stages of laboratory strains of P. falciparum, with no liver cell toxicity, sporozoite cell traversal and motility, and sporozoite development in the mosquito. Inhibiting PfMyoA completely kills parasites after 96 h of treatment. Furthermore, KNX-115 is equally effective at inhibiting a panel of Plasmodium strains resistant to experimental and marketed antimalarials and shows inhibitory activity against P. falciparum circulating isolates from the Brazilian Amazon. Inhibiting PfMyoA with KNX-115 also blocks the blood stage of a laboratory strain of Plasmodium vivax. In line with the evolutionary identity of MyoA among various apicomplexan parasites, KNX-115 also inhibits Cryptosporidium and Eimeria MyoA in vitro and is an effective inhibitor of Cryptosporidium, Toxoplasma, and Eimeria cellular growth, with EC₅₀s similar to those found for blood and liver stage Plasmodium.

Cells contain cytoskeletal networks of actin and microtubule tracks. These tracks are utilized by molecular motors like myosins, kinesins, and dyneins to perform motile and force-generating processes in the cell (1). This cytoskeletal system is vital to many essential cell functions in eukaryotes and is highly druggable, with multiple therapeutics currently in clinical trials (2–4), and two recently approved by the FDA for treatment of hypertrophic cardiomyopathy (5, 6). Kainomyx Inc. is a company devoted to using its extensive expertise in cytoskeletal biology for novel target-based drug discoveries to treat diseases, with a primary focus on malaria.

According to the 2024 WHO report, there were ~282 million malaria cases worldwide, resulting in 610,000 deaths, the majority of which occurred in children under the age of five (7). Resistance to drugs within current treatment regimens, such as artemisinin and its derivatives, is increasing in Southeast Asia and Africa, and there is an urgent need to develop new therapeutics (8). Many new drug discovery efforts have focused on phenotypic screens to find small molecules that inhibit the invasion/multiplication of Plasmodium parasites in the blood stages of the parasite (9). In general, however, knowing the target and understanding it well enables target-based screening and structure-based drug design, leading to a higher success rate in drug development.

Our approach is to develop small molecule inhibitors against mechanistically well-understood cytoskeletal targets that are vital components of Plasmodium parasite function and are likely to be essential at multiple stages of parasite infection. According to gene deletion studies, three Plasmodium falciparum myosins, MyoA, MyoF, and MyoK, have been shown to be independently essential for asexual blood stage parasite survival (10–12). P. falciparum MyoA (PfMyoA) was of primary interest for us to pursue because of its essential role in host cell invasion (13, 14). This single-headed myosin motor powers the sporozoite, merozoite, and ookinete stages of the parasite (13–16) through its ATP-dependent cyclical interactions with actin in the glideosome, a macromolecular complex necessary for the parasite’s motility and host cell invasion (17). It is biochemically well-characterized by in vitro studies using recombinantly expressed full-length PfMyoA (18). These studies revealed that motor function can be modulated by phosphorylation to optimize force and velocity parameters for the motile and invasion functions of the parasite during its human and mosquito life cycles (13). The high-resolution structure of PfMyoA combined with biochemical and parasitological studies (15, 16) has led to a detailed understanding of how this molecular motor works (13, 19–21). Importantly, none of the parasite myosins have more than 30% sequence identity to any of the human myosins (12), and even our early-stage molecules show high selectivity for the PfMyoA.

We have leveraged the amenability of PfMyoA to powerful molecular, structural, cellular, and genetic drug discovery tools to identify KNX-115, a 17-nM inhibitor of the actin-activated PfMyoA ATPase activity that serves as a potent and selective inhibitor useful for probing target biology and validating the consequences of PfMyoA inhibition; KNX-115 is not a compound that has been optimized or proposed as a clinical antimalarial drug candidate.

KNX-115 resulted from structure–activity relationship (SAR) studies on the original submicromolar-affinity Kainomyx molecule KNX-002 (22–24). Here, we demonstrate that by inhibiting PfMyoA, KNX-115 inhibits multiple stages of the Plasmodium life cycle, an attractive feature for a malaria therapeutic. In addition, we show that KNX-115 is a potent inhibitor of Cryptosporidium parvum (Cp), Toxoplasma gondii (Tg), and Eimeria tenella (Et) parasites, demonstrating the power of target-based drug discovery for apicomplexan parasitic diseases.

Results

In Vitro Molecular Assays Show KNX-115 Is a Potent Inhibitor of PfMyoA and Blocks Parasite Growth in Blood.

KNX-115 inhibits actin-activated PfMyoA ATPase activity with nanomolar potency.

KNX-002 inhibits full-length PfMyoA in an in vitro actin-activated myosin ATPase assay with an IC₅₀ of 0.4 ± 0.1 µM (Fig. 1A). SAR on KNX-002 led to KNX-115, which has KNX-002’s aminocyclopropyl group replaced by a sulfonamide (Fig. 1 A and B). KNX-115 shows a >20-fold gain in potency over KNX-002, with an IC₅₀ of <40 nM (Fig. 1A). Since this measured IC₅₀ is close to the PfMyoA concentration used in the assay (50 nM), we used the Morrison equation (25), which does not assume that the free concentration of the inhibitor equals the total concentration, to estimate the inhibition constant, K_i, to be 17 nM (Fig. 1C).

Fig. 1.

KNX-115 potently and selectively inhibits PfMyoA and blocks parasite growth in blood. (A) Actin-activated PfMyoA ATPase inhibition by parent compound KNX-002 (green) and its derivative KNX-115 (blue). 100% activity is the actin-activated ATPase rate in the DMSO control (ranges from 10 to 12 s⁻¹, at 15 µM actin). IC₅₀ is the concentration of KNX-115 needed to reduce the normalized ATPase rate to 50%. Data are expressed as mean ± SD. (B) Structure of KNX-115. (C) Morrison fit of a 22-point curve of PfMyoA inhibition. (D) KNX-115 mediated inhibition of rabbit skeletal myosin, bovine cardiac myosin, and chicken smooth muscle myosin actin-activated ATPase activity. (E) KNX-115 dose-dependently inhibits the growth of P. falciparum Dd2 strain in an asexual blood stage assay. Data are expressed as mean ± SEM, n = 52.

Dose–response curves of KNX-115 inhibition of bovine cardiac, rabbit skeletal, and chicken smooth muscle myosins showed IC₅₀s of 10 ± 6.4 µM, 1.0 ± 0.4 µM, and >50 µM, respectively (Fig. 1D). Considering a K_i for PfMyoA of 17 nM, selectivity is >50-fold for all three muscle myosins, suggesting high specificity for PfMyoA. Structure–activity relationship analysis yielded derivatives that did not inhibit cardiac, skeletal, or smooth muscle myosins at 50 µM, establishing that selectivity for PfMyoA over host myosins is synthetically tractable (SI Appendix, Table S1).

KNX-115 inhibits parasite growth in the blood stage of P. falciparum.

KNX-115 inhibition could potentially involve any of multiple steps in the parasite asexual reproductive cycle, including invasion of the red blood cells, development and schizogony of the parasites in the red blood cell after invasion, and egression from the red blood cell (13, 14, 16, 26). In asynchronous asexual blood stage cultures, KNX-115 inhibited the growth of P. falciparum Dd2 strain in red blood cell cultures with an EC₅₀ of 162 ± 14 nM (Fig. 1E). This apparent weakening of KNX-115 potency in parasite cellular assays versus the biochemical assays is likely in part explained by apparent weakening of KNX-115’s biochemical potency in the presence of Albumax (SI Appendix, Fig. S1), which binds small molecules leading to a decrease in their free fraction, thus leading to apparent weakening of potencies in engaging the target. Our measurements of KNX-115 binding to albumax showed 44% bound, while it was ~93% bound in a standard plasma protein binding assay using mouse plasma.

KNX-115 traps PfMyoA in a state that binds weakly to actin.

In vitro motility measurements with 0.15% methylcellulose, a condition where actin filaments remain close to the myosin-bound surface by interaction with only a small number of myosin molecules in the presence of ATP, were assessed in the absence or presence of KNX-115. Without inhibitor, all actin filaments showed directed motion with an average speed of 3.0 ± 0.4 μm/s (Movie S1), while there was no directed actin filament motion in the presence of KNX-115 (Movie S2). Furthermore, only a few actin filaments bound to surface-immobilized myosin in the presence of KNX-115 (8 ± 1 filaments/field) compared to robust binding in its absence (147 ± 5 filaments/field, Fig. 2A), suggesting that the inhibitor traps myosin in a state that binds weakly to actin. To define the nucleotide conditions under which KNX-115 acts, the number of actin filaments bound to surface-immobilized PfMyoA in the absence of methylcellulose was quantified (Fig. 2B). Without nucleotide, many actin filaments bound strongly to PfMyoA and KNX-115 had no effect, consistent with its binding site being inaccessible when the actin-binding cleft is closed in the high-affinity rigor state. In contrast, KNX-115 caused a marked reduction in actin binding in the presence of MgADP, reducing binding to levels comparable to those observed with MgATP. These data indicate that KNX-115 traps the motor in a low-affinity state for actin only when nucleotide is present, which is the physiological condition. KNX-115 also slowed the rate of Mg²⁺-mant-ADP and Mg²⁺-mant-ATP binding to PfMyoA by 6-fold and 14-fold, respectively (Fig. 2C), suggesting that although nucleotide and inhibitor can bind simultaneously, KNX-115 hinders the conformational rearrangements in the active site required for fast nucleotide binding.

Fig. 2.

KNX-115 traps nucleotide bound PfMyoA in a weak binding state for actin. (A) Representative still images from a gliding motility assay show that fewer filaments bind surface immobilized PfMyoA in the presence of KNX-115. Left panel, no inhibitor; Right panel, plus KNX-115. Data from two gliding motility assays each performed in triplicate (n = 6 fields). Values are means ± SD. (B) KNX-115 weakens the affinity of actin for ADP bound myosin but has no effect on binding in the absence of nucleotide. The normalized number of actin filaments bound ± SD to surface immobilized PfMyoA in the absence of methylcellulose is reported. The difference between ADP ± KNX-115 is significant (P < 0.0001) (one-sided ANOVA followed by Tukey’s post hoc test), but there were no significant differences between ATP ± KNX-115 (P > 0.9999), nor between no nucleotide ± KNX-115 (P = 0.9895). Data represent two experiments each performed in triplicate (open and filled symbols). Error bars are mean ± SD. (C) KNX-115 slows the binding of both mant-ATP (filled circles) and mant-ADP (open squares) to PfMyoA. Rates of the fast phase of a biphasic time course are plotted in the absence (blue) or presence of 50 µM KNX-115 (red). Association rate constants for mant-ATP are 5.3 ± 0.1 µM⁻¹ s⁻¹ in the absence of KNX-115 versus 0.37 ± 0.1 µM⁻¹ s⁻¹ in its presence. Association rate constants for mant-ADP are 4.9 ± 0.1 µM⁻¹ s⁻¹ in the absence of KNX-115 versus 0.79 ± 0.1 µM⁻¹ s⁻¹ in its presence. Data represent two experiments using different protein preparations.

A High-Resolution Structure of PfMyoA with KNX-115 Bound Explains Why KNX-115 Is More Potent Than KNX-002.

To investigate the inhibition mechanism of KNX-115, we cocrystallized PfMyoA full-length with the ATP analogue ATPγS and KNX-115 (PfMyoA/ATPγS/KNX-115) (SI Appendix, Table S2). The PfMyoA/ATPγS/KNX-115 full-length structure with the ELC1 and MTIP light chains was solved at 2.35 Å resolution (SI Appendix, Fig. S2A). The motor domain of PfMyoA is a typical myosin structure (Fig. 3A) with the active site found near the Transducer (deep teal), and the Relay (yellow) ending near the Converter (green), which is the beginning of the lever arm that rotates from a prestroke state to a poststroke state while bound to actin, generating force. The resulting electron density map allows the unambiguous placement of KNX-115, ATPγS, and the Mg²⁺ ion (green sphere) coordinated with water molecules (red spheres) (Fig. 3B). As for KNX-002 (23), PfMyoA with KNX-115 bound adopts a postrigor state (PR), which corresponds to the first state adopted by myosin motors upon ATP binding and their detachment from the actin track.

Fig. 3.

KNX-115 prevents PfMyoA inner cleft closure. (A) Crystal structure of PfMyoA full-length bound to ATPγS and KNX-115 (PfMyoA/ATPγS/KNX-115). Myosin domains are colored differently: N-terminal extension in purple, N-terminal (N-term) in gray, Relay in yellow, SH1-helix in red, Converter (Conv) in green. (B) The 2Fo-Fc electron density map allows the unambiguous placement of KNX-115, the nucleotide, the Mg²⁺ ion, and water molecules involved in the coordination. (C, Upper figure) Zoom in of the pocket occupied by KNX-115, and Lower figure, a diagram showing the details of the interactions: black dashed lines represent apolar interactions, green dashed lines represent π-stacking, and red dashed lines represent polar interactions. (D) Illustrates that both KNX-115 and KNX-002 interact with a water close to the Mg²⁺ coordination. (E) Shows that the only difference between KNX-002 and KNX-115 is the presence of a sulfonamide (KNX-115) in place of an aminocyclopropyl (KNX-002) group. The polar sulfonamide of KNX-115 allows polar interactions with R247. In (D and E) PfMyoA/ATPγS/KNX-115 and PfMyoA/ATPγS/KNX-002 [PDB code 8CDQ (23)] were superimposed on the entire structure.

KNX-115 occupies a pocket close to the active site, located in the cleft between the upper and the lower 50 kDa domains (U50 and L50, respectively), and interacts with both Switch-1 and Switch-2 (Fig. 3C). The structures of PfMyoA/ATPγS/KNX-115 and PfMyoA/ATPγS/KNX-002 (23) are highly similar and superimpose with a RMSD of 0.331 Å (SI Appendix, Fig. S2A). KNX-115 and KNX-002 occupy the same pocket (Fig. 3 C, Lower figure and SI Appendix, Fig. S2B and Table S3). In both molecules, the Mg²⁺ near the nucleotide is hexa-coordinated in a similar manner to what has been observed for the apo structure [PDB 8A12 (20, 23)]. The pyrazole group of KNX-115 is involved in γPi coordination as it binds a structural water that interacts directly with the γPi. This water molecule also interacts with another water molecule involved in Mg²⁺ and D469 coordination (Fig. 3C). The backdoor is open in this PR structure and no water is stabilized to promote ATP hydrolysis, unlike what is found for the postpowerstroke (PPS) structures, in which ATP is hydrolyzed and ADP.P_i gets trapped in the active site (27). Thus, KNX-115 inhibits ATP hydrolysis and prevents the recovery stroke through the same mechanism of action as previously described for KNX-002 (23): i) both compounds are sandwiched between Switches-1 and -2, thereby preventing inner cleft closure, and ii) they stabilize a conformation of the nucleotide pocket with an open backdoor, thus inhibiting ATP hydrolysis.

The difference between the two compounds is the presence of a polar sulfonamide group in KNX-115 replacing the apolar aminocyclopropyl group found in KNX-002 (Figs. 1 and 3D). This sulfonamide group establishes additional polar interactions with _Switch-1R247 (Fig. 3E and SI Appendix, Fig. S1B), a conserved residue that plays an important role to close the so-called phosphate backdoor (28, 29). We hypothesize that this difference is responsible for the increased potency of KNX-115, which fits tightly into the pocket through this additional polar interaction with Switch-1.

KNX-115 Kills P. falciparum Parasites.

Parasite reduction ratio assays show that KNX-115 kills blood stage P. falciparum parasites, and a stage-specificity assay shows that KNX-115 treated parasites cannot initiate a new round of parasite multiplication.

New antimalarials should ideally provide rapid relief of symptoms and have low rates of recrudescence after therapy. The parasite reduction ratio (PRR) assay, which measures speed of parasite killing (versus growth inhibition in standard EC₅₀ experiments with asexual stage parasites), showed that 5.4 µM KNX-115 killed P. falciparum 3D7 parasites after 96 h (Fig. 4A). Compared with other antimalarials, the profile closely resembled that of pyrimethamine. Both showed a 24-h lag before killing and both are faster than atovaquone but slower than chloroquine and the much more rapid artesunate, a member of the rapid-killing endoperoxide family of inhibitors (Fig. 4A). SI Appendix, Table S4 outlines the killing parameters of KNX-115, in which 99.9% of the parasite’s initial population was cleared in 58 h.

Fig. 4.

KNX-115 is parasiticidal and has a killing rate profile comparable to existing antimalarials. (A) P. falciparum viability time-course profiles demonstrate KNX-115 (blue) is parasiticidal and viable parasites are killed by 96 h. The assay was performed at 10X IC₅₀ of each compound. Data are expressed as mean ± SD. (B) Parasites of a NanoLuc reporter strain were synchronized at the ring stage, exposed to 5 µM KNX-115, and luciferase activity was compared over time to that of control (DMSO or 1 µM DHA-treated) cultures. (C) Giemsa staining and parasite counting of cultures treated with KNX-115, vehicle control (0.1% DMSO), or 1 µM DHA. The percentage values are total parasitemia and the pie charts indicate the proportion of ring/trophozoite versus schizont stages. Data for vehicle controls and DHA-treated cultures are from three independent flasks, and data for KNX-115 are from a single flask.

In stage-specificity assays, cultures from a P. falciparum NF54 strain expressing a Nanoluciferase reporter were synchronized at the ring/trophozoite stage, treated with 5 µM KNX-115, and followed over time. The control DHA led to rapid killing of the parasites. KNX-115 treated parasites replicated in the first 24 h and then showed a drop in parasitemia (Fig. 4B). Microscopy of the same cultures indicated a sharp increase in parasitemia in vehicle-treated cultures from 24 h onward (Fig. 4C). With KNX-115, the remaining parasites after 24 h were mostly in the schizont stage and did not develop beyond this stage. The combined data demonstrate that KNX-115 treated parasites are unable to initiate a new round of parasite multiplication.

Resistant Line Generation Identified PfMyoA as the Drug Target of KNX-115.

In vitro evolution experiments selected for mutations in PfMyoA that confer resistance to KNX-115.

To gain insight into potential mechanisms of resistance and to ensure on-target activity, in vitro evolution experiments were performed under drug pressure to favor selection of resistance mutations. The experiment started with bulk cultures to generate resistance which was followed by generation of clonal lines established from the drug-selected cultures. Treatment of 10⁸ NF54HGL P. falciparum parasites expressing a firefly luciferase reporter with 3× EC₅₀ of KNX-115 showed significant inhibition of parasite multiplication by 4 d, but recrudescent parasites were observed between 8 to 14 d into the experiment (Fig. 5A). When parasites from these cultures were analyzed by dose–response analyses using KNX-115, their EC₅₀s were ~threefold higher compared to wildtype parasites (Fig. 5B). Whole genome sequencing for the parent clone and resistant parasite populations F19, F20, and F21 identified ten different new missense mutations in the three genome sequences, relative to the parent clone, after removing variants in multigene families. Of these, nine were in PfMyoA with five different alleles detected, namely Q253P, A763T, K764E, E776K, and H107R (Fig. 5C and SI Appendix, Table S5). This is a very high level of enrichment that would not be expected by chance and shows that KNX-115 likely exerts its killing effect via PfMyoA.

Fig. 5.

In vitro evolution of drug resistance shows mildly resistant parasites that show no growth advantage over wildtype parasites. (A) Three flasks showing emergence of recrudescent P. falciparum parasites upon treatment with KNX-115 at 3× EC₅₀. Luciferase measurements were performed in triplicate, and the figure shows average values with error bars indicating SD. (B) KNX-115 sensitivity of drug-selected bulk cultures and mutant clones. The figure shows pIC₅₀ values for NF54 wildtype (WT) parasites, drug selected bulk cultures F19, F20 and F21 and mutant clonal lines rescued from the bulk cultures. Data are from three replicates, and the figure shows averages and SD. (C) Locations of PfMyoA mutations are not in the KNX-115 binding site. The eight mutations are located close to regions involved in the allosteric communication of the motor. Mutations shown as balls. The structure PfMyoA/ATPγS/KNX-115 is shown. Since the lever arm of the PR state of PfMyoA is kinked in this crystal packing (see ref. 20 for details), we used the structure of the PPS state (PDB code 6YCX) to model the straight lever arm for a better display. (D) Replication competition experiments between mutant clonal lines expressing a firefly luciferase reporter and wildtype parasites expressing a Nanoluciferase reporter. Cultures were inoculated at a 1:1 ratio and the ratio of the activities of the two reporters was followed in time. Experiments were repeated in three independent flasks. (E) Growth rates of wildtype and mutant parasites as determined by microscopic analyses of parasitemia. The figure shows data from three independent flasks. ** indicates significant (P < 0.01) difference as determined by one-way ANOVA.

To study the role of the individual PfMyoA mutations, clonal lines were established from the drug-selected cultures. Targeted sequencing of PfMyoA confirmed successful rescue of clonal lines for the individual A763T, E776K, and H107R mutations. In addition, we identified three new mutations, namely, S190Y, S742P, and M755R (Fig. 5 B and C) which were not seen in the whole genome sequencing of bulk culture. The Q253P and K764E mutations observed in the bulk culture sequencing were not rescued among the isolated clones, possibly due to insufficient sampling or fitness defects leading to these mutants becoming outcompeted over the course of limiting dilution. Phenotyping of the individual clones indicated a resistance profile aligned with that of the polyclonal cultures (Fig. 5B).

To study fitness effects of the mutations, we combined the mutant clones, which were made in a Firefly luciferase reporter background, with wildtype parasites expressing a Renilla luciferase reporter and followed the ratio of the two reporter activities over time. All clones replicated at rates comparable to wildtype, except for parasites with a H107R mutation that were rapidly outgrown by wildtype parasites (Fig. 5D). The slower replication rate for the H107R mutant clone was confirmed by microscopic analyses of parasitemia over time and calculation of replication rates (Fig. 5E). The growth rates of the other mutant clones did not significantly deviate from wildtype. These data indicate that the resistance mutations selected by KNX-115 are unlikely to have a fitness advantage in the human P. falciparum reservoir.

None of these PfMyoA mutations were in the KNX-115 binding pocket. They were all allosteric mutations in the myosin motor domain (Fig. 5C). Allosteric action is the likely explanation for the mild weakening of KNX-115 potency. For how these allosteric mutations might be affecting the myosin motor function, see the legend to SI Appendix, Fig. S3.

Minimum inoculum for resistance (MIR) analyses confirm PfMyoA as the target with an acceptable likelihood of resistance development.

The MIR assay using selection at high drug pressure is a valuable way to sense the likelihood of the development of resistance by Plasmodium parasites when treated with a drug (30). In MIR studies using the P. falciparum Dd2-B2 clone, EC₅₀ was 590 nM (N, n = 3, 2) and EC₉₀ was 1,072 nM (N, n = 3, 2). For KNX-115, two MIR single-step selections on the Dd2-B2 line were set up in parallel at a drug pressure of 3× EC_90-cellular (3,216 nM). The first selection was with 10⁶ parasites in 12 separate wells for a total of 1.2 × 10⁷ parasites. The second selection was with 3 × 10⁷ parasites in three separate wells for a total of 9 × 10⁷ parasites. Both selections had a consistent high drug pressure of (3× EC₉₀) over 60 d.

Parasite clearance was observed over the first 4 d. Cultures were screened three times weekly by flow cytometry and smearing, and no recrudescence was observed over the full 60 d in the 10⁶ parasites selection. Wells were counted as positive for recrudescence when the overall parasitemia reached 0.3%, and the parasites could be seen on a blood smear. Only one well of the 3 × 10⁷ parasites selection became positive on day 40 and was picked for further investigation. By day 60, the 3 × 10⁷ parasites selection yielded 1 recrudescent well out of three (MIR 1.02 × 10⁸, Log₁₀ MIR 8.01). The MIR value for KNX-115 is aligned with MMV’s criteria, which mandate that molecules must have Log₁₀ MIR > 7 to progress through the development pipeline. The bulk culture of the recrudescent well was then profiled against KNX-115 and showed a ~22× EC₅₀ shift and a ~31× EC₉₀ shift (Fig. 6A and Table 1).

Fig. 6.

Resistance to KNX-115 is observed in parasites with the G249V mutation in myosin A. (A) KNX-115 A1 bulk culture EC₅₀ and EC₉₀ shifts. Data shown are mean ± SEM (N, n = 3, 2). (B) The position of the G249V mutation relative to the KNX-115 binding site. The structure PfMyoA/ATPγS/KNX-115 is shown.

Table 1.

Bulk culture EC₅₀ and EC₉₀ shifts upon KNX-115 binding

	Dd2-B2	Dd2-KNX-115	Dd2-B2	Dd2-KNX-115
	EC50 (nM)		EC90 (nM)
Mean EC	714	15,618	1,146	35,470
SEM	41	1,139	7	2,112
Fold shift	N/A	22	N/A	31
N	3	3	3	3

The recrudescent well bulk culture was expanded and sequenced for further analysis. The selection drug pressure of 3× EC₉₀ was maintained throughout EC₅₀ and EC₉₀ phenotyping and expansion for gDNA extraction. Whole genome sequencing revealed one fixed SNP in PF3D7_1342600, the 3-exon gene encoding PfMyoA (SI Appendix, Table S6). The mutation was a missense mutation (G249V), abutting the KNX-115 binding site (Fig. 6B), which explains the larger EC₅₀ and EC₉₀ shifts compared to the allosteric mutants described above (Fig. 5C). No copy number variants (CNVs) were found in the recrudescent culture genome.

KNX-115 Inhibits Multiple Steps in the Parasite Life Cycle Using In Vitro Cellular Assays.

KNX-115 inhibits parasite growth in the liver stage of P. falciparum.

A highly desirable feature of antimalarials is activity against multiple stages of the parasite life cycle. Given the importance of the glideosome in sporozoite motility (15), we tested whether PfMyoA inhibition would affect the ability of the parasites to invade and multiply in liver cells, the first intracellular niche occupied by the parasite in a human host. When P. falciparum sporozoites were preincubated with KNX-115 before adding them to primary human hepatocytes, inhibition of parasite multiplication was observed with an EC₅₀ of 182 nM ± 64 nM (Fig. 7A, dark blue circles), which is comparable to the EC₅₀ seen in the asexual blood stage assay (Fig. 1E). Nearly identical results were obtained in experiments measuring inhibition of multiplication of P. berghei sporozoites in HepG2 cells when compounds were added preinvasion (EC₅₀ = 108 ± 21 nM) (Fig. 7B). The KNX-115 binding site (SI Appendix, Table S3) is 100% identical between P. falciparum and P. berghei MyoA. Interestingly, when P. falciparum sporozoite invasion was allowed to occur before adding KNX-115, no inhibition of parasite multiplication was seen (Fig. 7A, light blue triangles), suggesting that KNX-115 has its most potent activity early in infection, possibly by blocking invasion. Importantly, no evidence of toxicity was observed below 5 µM KNX-115 in the HepG2 cells (SI Appendix, Fig. S4A) and no liver cell toxicity was seen in the human hepatocyte experiments even at 10 µM KNX-115 (SI Appendix, Fig. S4B).

Fig. 7.

KNX-115 inhibits multiple stages of the Plasmodium parasite life cycle. (A) KNX-115 dose-dependently inhibits the growth of P. falciparum in a liver stage assay using human hepatocytes when sporozoites were preincubated with KNX-115 (dark blue circles). The inhibition is lost when KNX-115 is added postinvasion (light blue triangles). Data are from two biological replicates. Error bars indicate SDs. (B) KNX-115 dose-dependently inhibits the growth of P. berghei in a liver stage development assay using HepG2 cells (data expressed as mean ± SEM, n = 21) (31). Data are from two experiments performed on two different days. (C) KNX-115 inhibits sporozoite cell traversal through HC-04 cells. Data are from two biological and six technical replicates and are expressed as mean ± SEM. (D) The figure shows sporozoite gliding motility as measured by quantification of gliding trails relative to the assay controls (positive control: 10 µM cytochalasin D; negative control: 0.1% DMSO). The figure shows average values and SD for two independent measurements. (E) KNX-115 dose-dependently inhibits the conversion of gametocytes to oocysts in a standard membrane feeding assay performed with a P. falciparum NF54-HGL reporter strain. Each point is oocyst intensity in individual mosquitoes from two independent feeds per drug concentration tested (F) KNX-115 is not gametocytocidal in a gametocyte viability assay. Dihydroartemisinin (DHA) inhibition is a positive control. (G) KNX-115 does not affect exflagellation events in male gametocytes. Data are expressed as mean ± SD.

KNX-115 inhibits sporozoite traversal and sporozoite gliding.

An essential part of the parasite life cycle is the first step of cell traversal of sporozoites from the site of their injection in the dermis by a mosquito to the liver by gliding motility and transmigration through cells. It is rare to find an inhibitor of sporozoite cell traversal and gliding, but we hypothesized that we would observe such inhibition by explicitly targeting PfMyoA, which is a pivotal component of the glideosome used for sporozoite motility. Indeed, KNX-115 inhibited P. falciparum sporozoite cell traversal of hepatocytes with an EC₅₀ of 816 ± 55 nM (Fig. 7C) and sporozoite gliding with an EC₅₀ of 117 ± 20 nM (Fig. 7D).

KNX-115 inhibits development of sporozoites from gametocytes in the mosquito.

The mosquito stages of malaria parasites start with the activation of gametogenesis from the intraerythrocytic gametocytes immediately after a blood meal and end with the release of sporozoites. KNX-115 blocked the development of sporozoites from gametocytes in the mosquito host with an EC₅₀ of 230 nM (Fig. 7E). This could likely be due to blocking of the motile ookinete stage of the parasite. MyoA protein expression was not detected in the nonactivated gametocyte stages (11), and KNX-115 did not block the growth of stage III-IV gametocytes measured by a gametocyte viability assay (Fig. 7F) while the dihydroartemisinin (DHA) control showed inhibition. Furthermore, KNX-115 did not block exflagellation events of male gametocytes (Fig. 7G), in line with a lack of MyoA protein expression in the activated gametocyte stages of the parasite (11).

KNX-115 Effectively Inhibits P. falciparum Strains at the Blood Stage That Are Resistant to Approved Antimalarials and Advanced Portfolio Molecules.

Given the emergence of resistance to current antimalarials, a highly desirable profile of a new antimalarial is that it remains inhibitory against strains already resistant to approved antimalarials and advanced portfolio molecules. KNX-115 dose–response curves were comparable using Pf NF54 parasite strains showing multidrug resistance to chloroquine cycloguanil/pyrimethamine (strains 7G8 and K1), chloroquine cycloguanil/pyrimethamine, artemisinin and piperaquine (strain PH1263-C), and cycloguanil/pyrimethamine and atovaquone (strain TM90C2B), when compared to the parent Pf NF54 strain (Fig. 8A). We also tested five Pf Dd2 strains each resistant to a different advanced portfolio compound in addition to chloroquine cycloguanil/pyrimethamine. Again, KNX-115 dose–response curves were comparable using the five resistant Pf Dd2 strains compared to the parent Pf Dd2 strain (Fig. 8B), suggesting that, as expected, KNX-115 is not subject to resistance mechanisms mediated through the genetic alterations represented by these parasite lines.

Fig. 8.

KNX-115 inhibits asexual blood stage parasite replication in resistant strains of P. falciparum. (A, Top) Strains that are resistant to approved antimalarials. Five genetic backgrounds are represented in this subpanel (7G8 – South America; TM90C2B, K1 – South-East Asia; NF54 West Africa). NF54 is considered the reference pan-sensitive strain and TM90C2B, PH1263-C, 7G8, K1 are multidrug-resistant (MDR) strains, in which seven drug resistant genes are represented. The star (*) indicates a resistance mechanism that cannot be detected performing a standard ABS proliferation assay. (Bottom) KNX-115 is equipotent in inhibiting asexual blood stage parasitemia in resistant strains of P. falciparum compared to the parental NF54 strain. (B, Top) Dd2 strains that are resistant to advanced portfolio molecules. Five genetically engineered strains in a Dd2 genetic background are present in this subpanel, in which five drug resistance-correlated genes are represented. The parental Dd2 wild-type strain is used to determine a potential EC₅₀ shift. (Bottom) KNX-115 is equipotent in inhibiting asexual blood stage parasitemia in resistant strains of P. falciparum compared to the parental Dd2 strain. Data are expressed as mean of at least two independent biological replicates.

KNX-115 Effectively Inhibits Both P. falciparum and P. vivax in Ex Vivo Assays Using Patient Isolates.

KNX-115 effectively inhibits P. falciparum reinvasion of RBCs using patient isolates.

P. falciparum isolates were obtained from 15 monoinfected patients from the Brazilian Western Amazon. Parasitemia among these samples ranged from 2,100 to 17,773 parasites/µL. All isolates exhibited suitable in vitro growth conditions, enabling successful completion of the schizont maturation test (SMT). KNX-115 activity could be evaluated in 73% of the isolates, and its inhibitory effect was detectable during the second asexual life cycle of the parasites.

The EC₅₀ values determined for each clinical isolate are presented in Fig. 9A. The median EC₅₀ was 617 nM, which is consistent with the inhibitory values observed for the P. falciparum 3D7 reference strain (Fig. 9B). The clinical isolates were sensitive to artesunate, showing a median EC₅₀ of 3.5 nM, while chloroquine displayed a median EC₅₀ of 1,260 nM, confirming the expected resistance profile for Amazonian P. falciparum isolates (32).

Fig. 9.

KNX-115 inhibits ex vivo activity of P. falciparum and P. vivax patient isolates. (A) Ex vivo drug susceptibility (EC₅₀ values) of 11 field patient isolates for KNX-115, chloroquine, and artesunate inhibition. Values are presented as violin plots. (B) P. falciparum viability as a function of KNX-115 concentration used to determine 3× and 1× EC₅₀ values for use on field patient isolates of P. vivax, assuming similar dose–response for the two species. (C) KNX-115 was tested on 3 P. vivax patient isolates at two different concentrations (3× and 1× the EC₅₀ value). The invasion rate was calculated after 20 h incubation of purified schizonts with the compounds. The assays were performed in duplicate. *P < 0.05 in Mann–Whitney test.

KNX-115 effectively inhibits P. vivax reinvasion of RBCs using patient isolates.

P. vivax isolates were collected from three mono-infected patients in the Brazilian Western Amazon. The parasitemia ranged from 11,298 to 13,550 parasites/µL for the P. vivax isolates. Suitable growth for harvest was achieved in all of the P. vivax isolates in which KNX-115 was tested.

A P. falciparum 3D7 reference strain dose–response curve for viability (Fig. 9B) was used to determine the best concentration for the P. vivax reinvasion assays. Based on the P. falciparum EC₅₀ = 490 nM, test concentrations of 1x and 3x the EC₅₀ value were used for the P. vivax reinvasion into RBCs assay. Three human P. vivax samples were analyzed (Fig. 9C). For sample 1, the % parasitemia after 20 h incubation with or without KNX-115 was 0.8 ± 0.2 (untreated control), 0.8 ± 0.1 (1× EC₅₀), 0 ± 0 (3× EC₅₀); for sample 2, 0.7 ± 0.2, 0.5 ± 1.1, 0 ± 0; for sample 3, 2.3 ± 0.4, 1.2 ± 0.2, 0 ± 0. Thus, the 3× EC₅₀ KNX-115 treatment abolished the P. vivax invasion into RBCs in all 3 human samples.

KNX-115 Inhibits Other Members of the Apicomplexan Parasite Family.

In addition to malaria caused by species of Plasmodium, other Apicomplexa phyla members, Cryptosporidium, Toxoplasma, and Eimeria, cause significant disease in either humans, animals, or both. Cryptosporidiosis, a highly prevalent gastrointestinal parasitic disease in animals and humans that typically manifests as watery diarrhea, is increasingly identified as an important cause of morbidity and mortality worldwide, especially in low-resource areas (33). Coccidiosis caused by Eimeria species leads to massive losses in the poultry industry (34), and there has been an increase in drug-resistant forms of the parasite warranting development of novel therapeutics. Toxoplasmosis, caused by the intracellular parasite Toxoplasma gondii, is a globally prevalent zoonotic infection affecting roughly one-third of the world’s population. While often asymptomatic, it can cause severe disease in immunocompromised individuals and serious congenital complications when acquired during pregnancy (35).

Given the evolutionary conservation of MyoA among the various apicomplexans species, we examined the effects of KNX-115 on Cryptosporidium parvum MyoA (CpMyoA) motor domain in vitro using the actin-activated myosin ATPase assay. KNX-115 inhibited CpMyoA with an IC₅₀ < 100 nM (Fig. 10A). In keeping with its effects on CpMyoA in the in vitro molecular assay, KNX-115 was an effective inhibitor of C. parvum replication in an in vitro cellular growth assay using HCT-8 cells, with an EC₅₀ of 230 nM (Fig. 10B), like that found for Plasmodium (Fig. 1E).

Fig. 10.

KNX-115 inhibits Cryptosporidium parvum, Toxoplasma gondii, and Eimeria tenella. (A) KNX-115 inhibition of C. parvum MyoA actin-activated ATPase activity. For the in vitro biochemical assays, there is only an upper bound on the IC₅₀ values because accurate measurement of the inhibitory effect of KNX-115 was limited by the concentration of MyoA needed for the assay (40 nM). (B) KNX-115 inhibits proliferation of Cryptosporidium parvum (Cp) in HCT-8 cells. (C) Analysis of the effects of treatment during different time frames of C. parvum infection of HCT-8 cells: −2.5 to 0 refers to pretreatment of host cells, 0 to 2.5 refers to treatment only during the first 2.5 h post infection (hpi), 2.5 to 24 refers to treatment from 2.5 to 24 hpi, and 0 to 24 refers to continuous treatment for 24 h. Data are expressed as mean ± SD. For the C. parvum growth inhibition assay, data are derived from two biological replicates each with two technical replicates. Statistical test in C, two-way ANOVA with Tukey multiple comparison, ****P < 0.0001. (D) KNX-115 inhibits Toxoplasma gondii growth in human foreskin fibroblasts. The dose–response curve was conducted using a luciferase-expressing RH88 of T. gondii in 384 plates. KNX-115 was threefold serially diluted with a starting concentration of 10 µM. Points represent mean ± SEM of three independent biological replicates; technical triplicates were averaged prior to analysis. (E) KNX-115 inhibition of E. tenella MyoA actin-activated ATPase activity. For the in vitro biochemical assays, there is only an upper bound on the IC₅₀ values because accurate measurement of the inhibitory effect of KNX-115 was limited by the concentration of MyoA needed for the assay (25 nM). (F) KNX-115 inhibits E. tenella growth (dark blue) in MDBK cells. The dose–response curve was conducted using a dual antibody fluorophore (UGAMAB 1A3 primary). The graph is from four biological replicates, each with three technical replicates. KNX-115 had no effect on the host cell viability (light blue).

KNX-115 was relatively nontoxic for HCT-8 cells, with ~6% inhibition of host cell growth at 10 µM and ~15% inhibition at 100 µM. We also compared the inhibition of C. parvum growth by KNX-115 at the EC₅₀ level for different intervals followed by washout. Pretreatment of host cells had no effect on growth inhibition, while treatment during infection and continuing through the first 2.5 h post infection (hpi) was as effective as continuous treatment, with treatment beginning after 2.5 hpi being less effective (EC₅₀ = 730 nM) (Fig. 10C). These findings imply that, as for Plasmodium described above, KNX-115 has its most potent activity on Cryptosporidium early in infection, possibly by blocking invasion. Notably, there was also some effect when KNX-115 was added at EC₅₀ levels 2.5 h after infection, and this might be attributable to the fact that merozoites emerge from the first round of merogony at 12-15 h and then reinvade to initiate a second round of infection (Fig. 10C).

For Toxoplasma gondii, KNX-115 inhibition of cellular growth showed an EC₅₀ of 750 ± 280 nM (Fig. 10D). For Eimeria tenella, KNX-115 inhibited EtMyoA ATPase with an IC₅₀ < 100 nM (Fig. 10E) and E. tenella cellular growth with an EC₅₀ of 540 nM without impacting host cell viability (Fig. 10F).

Discussion

At Kainomyx, we developed KNX-115, which we first reported on in bioRxiv in 2022 (22). KNX-115 specifically targets MyoA in apicomplexan parasites. PfMyoA is an essential actomyosin motor required for motility and force production at multiple stages of the Plasmodium parasite life cycle (SI Appendix, Fig. S5). The breadth of activity observed with PfMyoA inhibition distinguishes this mechanism from those of the most currently used antimalarial therapeutics. Approved antimalarials such as chloroquine, amodiaquine, quinine, mefloquine, lumefantrine, piperaquine, and artemisinin-based combination therapies act primarily as blood-stage schizonticides and are not known to directly inhibit sporozoite motility, traversal, or host-cell invasion. Antifolates such as sulfadoxine–pyrimethamine also act mainly against replicating parasites, with clinical utility against blood-stage infection and chemoprophylaxis. Atovaquone–proguanil provides causal prophylactic activity against developing liver-stage parasites as well as blood-stage activity, but it is not known to block sporozoite motility or invasion directly. The 8-aminoquinolines primaquine and tafenoquine are notable for their activity against dormant liver-stage hypnozoites of P. vivax and P. ovale and for their transmission-reducing gametocytocidal effects, but they do not directly target the invasive motility machinery used throughout the parasite life cycle. Similarly, antibiotics such as doxycycline, clindamycin, and azithromycin have slow, delayed-death activity related to apicoplast function rather than direct effects on parasite motility or invasion.

By contrast, targeting PfMyoA interferes with an actomyosin motor required at multiple points in the parasite life cycle, including sporozoite motility and traversal, liver-cell invasion, red blood cell invasion, and mosquito-stage development leading to infectious sporozoites. Thus, PfMyoA inhibition offers a mechanistically distinct, stage-spanning strategy with potential therapeutic, prophylactic, and transmission-blocking applications. Inhibition of sporozoite motility and traversal suggests that PfMyoA inhibitors could have value as prophylactic agents for travelers, military personnel, and others at risk of malaria exposure. KNX-115 inhibits parasite development in the mosquito midgut, likely by targeting the motile ookinete. Because Anopheles populations are increasingly insecticide-resistant, transmission-blocking strategies that act on the parasite rather than the vector are attractive. Drug ingested in a blood meal can reach the mosquito hemolymph (36), and treated surfaces can deliver transmission-blocking antimalarials such as atovaquone to mosquitoes (37). With a suitable uptake profile, MyoA inhibitors might, in principle, block transmission by either route, though substantial further work would be required to establish their feasibility.

The findings that RBC invasion by P. vivax parasites was inhibited by KNX-115 and that liver-stage infection of P. berghei and P. falciparum were inhibited to similar extents by KNX-115 suggest that inhibition of PfMyoA will likely be effective across Plasmodium species. The liver-stage inhibition and inhibition of sporozoite motility also have potential applications of a MyoA inhibitor as a prophylactic for malaria. However, the dosing and safety parameters for such prophylactics need to be highly favorable and held to a much higher standard.

The data showing that KNX-115 is equipotent across strains of parasites resistant to other antimalarials and advanced portfolio compounds are very encouraging and indicate that the efficacy of a PfMyoA inhibitor is unlikely to be hampered by preexisting resistance mechanisms. Our experiments concerning the generation of resistant species upon KNX-115 treatment suggest that the selected parasites exhibit only weak resistance to KNX-115. There might be an evolutionary fitness cost for the parasite to develop resistance by mutating MyoA to escape MyoA inhibitors. An MIR score of 8 is considered a medium potential for resistance ranking.

While the slow onset of action of killing by KNX-115 has precedent in malaria (e.g., tetracyclines, pyrimethamine), this could pose issues if fever reduction is slow, which, for example, could lead to compliance issues. However, such issues could be mitigated by pairing a slow-acting drug with a fast-acting one in a typical antimalarial combination therapy regimen.

Importantly, we demonstrate that KNX-115 also has potent activity against other apicomplexan parasites from the Cryptosporidium, Toxoplasma, and Eimeria genera. Consistent with this, Snyder et al. have recently described a TgMyoA inhibitor with activity across other apicomplexan species and structural similarity to KNX-115 (38), further supporting MyoA as a tractable, conserved target across the phylum. In summary, our discovery and characterization of KNX-115 as a potent compound confirms MyoA as a promising target and lays the groundwork for developing MyoA inhibitors as part of the next generation of molecules for treating malaria, and possibly other apicomplexan parasitic diseases.

Methods

All Methods are described in detail in SI Appendix. Briefly,

KNX-115 origin.

Derivation of KNX-115 is described in SI Appendix.

Expression constructs.

Recombinant baculovirus expression vectors were generated by GenScript (Piscataway, NJ), and Sf9 codon-optimized sequences were cloned into individual pFastBac1 vectors. Sequences encoded for P. falciparum MyoA (PfMyoA) and other proteins, and details of the expression, were as described (18, 39).

Protein expression and purification.

PfMyoA, CpMyoA, and EtMyoA were expressed in Sf9 cells and purified as described (18). Skeletal, cardiac, and smooth muscle myosins were purified as described (40, 41). Bovine cardiac actin was purified as described (42).

Myosin ATPase assay.

The enzymatic activities of recombinantly expressed PfMyoA, CpMyoA, and EtMyoA were assessed by an actin-activated ATPase assay.

Protein expression and purification.

Sf9 cells were infected with relevant recombinant baculovirus, grown, harvested, and lysed by sonication, and proteins were purified. Skeletal muscle actin was purified from chicken skeletal muscle tissue as described (42).

In vitro motility assays.

PfMyoA driven movement of actin in the in vitro motility assay and an actin-binding assay were performed as detailed in SI Appendix.

Transient kinetics.

Stopped-flow experiments were performed on a KinTek stopped-flow apparatus (KinTek Corporation, model SF-2001).

Crystallization, data processing, structure determination, and refinement.

Full-length PfMyoA comprising PfELC and MTIP-ΔN (lacking residues 1 to 60) was cocrystallized complexed to KNX-115 and ATPγS. X-ray datasets were collected at the ESRF synchrotron, on the ID30b beamline. The atomic coordinates and structures factors have been deposited to the Protein Data Bank (43) with accession numbers 9T7I.

Analysis of the KNX-115 binding pocket.

The residues involved in KNX-115 binding were automatically determined by the software LigPlot+ (44).

Parasite reduction ratio (PRR) assay was performed as per the established protocol described by Sanz et al. (45). The parasite strain was cultured using standard procedures as described (46). The presence of parasite growth was detected using the P. falciparum LDH assay (47).

Asexual replication stage-specificity experiments.

P. falciparum NF54-ENL parasites expressing a nanoluciferase reporter gene were synchronized at the ring stage. Cultures were treated with vehicle (0.1% DMSO), 1 µM dihydroartemisinin, or 10 µM KNX-115, and nanoluciferase was followed in time. In independent experiments, asynchronous cultures were treated with the same compounds and stage progression of asexual blood stage parasites was monitored by microscopic analyses of Giemsa-stained thin smears.

P. falciparum in vitro evolution assays and whole genome sequencing.

P. falciparum NF54-HGL parasites were seeded at a density of 10⁸ parasites in 10 mL culture medium under pressure of KNX-115 at 3× IC₅₀. Parasite genomic DNA was extracted from the parent flask and the three resistant flasks, respectively, and sequenced on an Illumina NovaSeq 6000. Reads were aligned to P. falciparum reference genome 3D7 (PlasmoDB v13.0) using the pipeline described by Cowell et al. (48).

Mutant parasite cloning and replication competition assays.

Individual clones were rescued from drug-selected cultures by limited dilution cloning. Drug sensitivity was determined by parasite replication assays using the activity of the firefly luciferase reporter as a measure for parasite density (49). For replication competition experiments, mutant parasites expressing a firefly luciferase reporter enzyme were mixed with wildtype parasites expressing a Nanoluciferase reporter (50) at a 1:1 ratio and ratio of the activities of the two reporter enzymes was followed in time.

MIR assay.

P. falciparum asexual blood stage (ABS) parasites were cultured at 3% hematocrit. Dd2 parasites were obtained from T. Wellems (NIAID, NIH). Dd2-B2 is a genetically homogeneous line that was cloned from Dd2 by limiting dilution in the Fidock lab. To define the EC₅₀ of ABS parasites, growth inhibition assays were performed.

Whole-genome sequencing.

Whole-genome sequencing was performed using a Nextera Flex DNA library kit and multiplexed on a MiSeq flow cell to generate 300 bp paired-end reads. GATK HaplotypeCaller (GATK version 4.2.2) was used to identify all possible single nucleotide variants (SNVs) in test parasite lines filtered based on quality scores to obtain high quality single nucleotide polymorphisms (SNPs) that were annotated using SnpEff version 4.3t (51). BIC-Seq version 1.1.2 (52) was used to discover copy number variants (CNVs) against the parental strain using the Bayesian statistical model.

P. falciparum asexual blood stage assay.

A dose–response assay was performed to identify compounds that inhibit the growth of P. falciparum Dd2 strain parasites in human erythrocytes. To investigate whether KNX-115 yields higher potency with a culture starting with schizonts, the cultures were first synchronized.

Inhibition of P. falciparum liver stage development assay.

Assessment of compound effects on P. falciparum NF54 invasion of human primary hepatocytes and intrahepatocytic development was performed as described (53).

Inhibition of P. berghei liver stage assay.

For this assay, a human hepatoma cell line (HepG2-A16-CD81EGFP) was infected and parasite growth was assessed by bioluminescence measurements.

P. falciparum sporozoite cell traversal assay.

Traversal assays were performed essentially as described (54).

P. falciparum sporozoite gliding motility assay.

Sporozoite gliding motility was measured by staining gliding trails consisting of shed CSP protein.

P. falciparum transmission assays.

Transmission of malaria parasites was studied by Standard Membrane Feeding Assays using luminescent reporter parasite NF54-HGL. Following incubation of stage V gametocytes with compound for 48 h, a small aliquot was taken for analysis of exflagellation events by light microscopy. The remaining culture material was used to infect Anopheles stephensi mosquitoes. Eight days post infection, oocyst intensity was determined by luciferase assays on whole mosquito homogenates as described (49). Gametocyte viability was assessed as described (53).

P. falciparum asexual blood stage activity against Dd2 (MRA-156 from BEI-Resources, USA), a clone originally derived from the clinical isolate Indochina III/CDC, and a series of resistant laboratory strains derived from Dd2.

The activity was determined with the modified [³H]-hypoxanthine incorporation assay, as previously reported (55, 56).

P. falciparum schizont maturation assay.

P. falciparum isolates were obtained from monoinfected patients recruited at the Centre of Malaria Control (CEPEM) in Porto Velho, Rondônia, Brazil. Thick blood films were stained with 5% Giemsa and schizonts were counted. Data were normalized to drug-free controls and used to calculate EC₅₀ values.

P. vivax reinvasion of RBCs experiments.

The P. vivax isolates were collected from patients recruited at the Centre of Malaria Control (CEPEM) in the city of Porto Velho, state of Rondônia, in the Brazilian Western Amazon. Reinvasion assays were as described (57). KNX-115 was tested using the P. falciparum 3D7 lab strain (received as a donation from University of Campinas–UNICAMP).

Cryptosporidium parvum intracellular growth assays.

Human adenocarcinoma colon HCT-8 cells were cultured and plated in 96-well plates, then infected with 5 × 10⁴ bleached oocysts from the C. parvum AUCP-1 strain. KNX-115 was serially diluted to generate a dilution curve and compound dilutions were added to cells at the time of infection. For washout experiments, compounds were added as above and then incubated for intervals followed by washout and return to culture medium. All wells were fixed 24 hpi and labeled with rabbit anti-Cp antibody as described (58). Plates were imaged to quantify parasite growth and host cell counts.

A luciferase-expressing RH88 strain of Toxoplasma gondii was a gift from David Sibley at Washington University in St. Louis. The dose–response assay was performed as described (59).

Eimeria tenella intracellular growth assays.

MDBK cells (ATCC CCL-22) were cultured, and culture medium was removed and replaced with freshly excysted sporozoites of E. tenella (Penn ST strain). KNX-115 dilutions were added to each well of cultured, infected cells at the time of infection with sporozoites. Microscopy was with a Keyence BZX810 fluorescent microscopy system counting 20 random images per well (green and blue). Cytotoxicity was determined by live cell counts of nuclei (DAPI) in each well. Controls were infected/+monensin, infected/without KNX-115, and no infection.

Supplementary Material

Appendix 01 (PDF)

Movie S1.

Movement of actin filaments by PfMyoA in the absence of KNX-115. Representative movie at 4x normal speed showing directed motion of actin filaments powered by PfMyoA. Conditions: 25 mM imidazole pH 7.5, 150 mM KCl, 4 mM MgCl₂, 1 mM EGTA, 10 mM DTT, 0.15% methylcellulose, 1% DMSO, 30°C.

Movie S2.

No directed movement of actin filaments by PfMyoA is observed in the presence of KNX-115. Representative movie at 4x normal speed in the presence of 50 μM KNX-115. Few filaments are seen in the field of view, and none exhibit directed motion. Conditions: 25 mM imidazole pH 7.5, 150 mM KCl, 4 mM MgCl2, 1 mM EGTA, 10 mM DTT, 0.15% methylcellulose, 50 μM KNX-115, 1% DMSO, 30°C.

Data, Materials, and Software Availability

Study data are included in the article and/or supporting information.