Hemisynthetic alkaloids derived from trilobine are antimalarials with sustained activity in multidrug-resistant Plasmodium falciparum

Summary

Malaria eradication requires the development of new drugs to combat drug-resistant parasites. We identified bisbenzylisoquinoline alkaloids isolated from Cocculus hirsutus that are active against Plasmodium falciparum blood stages. Synthesis of a library of 94 hemi-synthetic derivatives allowed to identify compound 84 that kills multi-drug resistant clinical isolates in the nanomolar range (median IC₅₀ ranging from 35 to 88 nM). Chemical optimization led to compound 125 with significantly improved preclinical properties. 125 delays the onset of parasitemia in Plasmodium berghei infected mice and inhibits P. falciparum transmission stages in vitro (culture assays), and in vivo using membrane feeding assay in the Anopheles stephensi vector. Compound 125 also impairs P. falciparum development in sporozoite-infected hepatocytes, in the low micromolar range. Finally, by chemical pull-down strategy, we characterized the parasite interactome with trilobine derivatives, identifying protein partners belonging to metabolic pathways that are not targeted by the actual antimalarial drugs or implicated in drug-resistance mechanisms.

Graphical abstract

Highlights

• •First hemisynthetic trilobine derivatives for malaria treatment
• •Compounds active in multi-drug resistant isolates of Cambodia
• •Target multiple malaria life cycle stages
• •Chemical pull-down enriched proteins involved in translation and DNA replication

Multi-drug resistant organisms; Natural product chemistry; Applied microbiology; Microbiology parasite; Clinical microbiology

Introduction

Malaria, a vector-borne disease, continues to cause morbidity and mortality with more than 247 million cases and 619,000 annual deaths according to the World Health Organization (WHO).¹ This disease is caused by an apicomplexan protozoan parasite belonging to the genus Plasmodium. Among the five Plasmodium spp that infect humans, Plasmodium falciparum is, by far, responsible for the most severe disease and greatest number of deaths.

P. falciparum has a complex life cycle: the parasite is transmitted to humans by a mosquito bite that inoculates sporozoites. Sporozoites rapidly reach the liver where they replicate exponentially in a hepatocyte for approximately one week before the cells burst and release merozoites into the bloodstream. The merozoites invade red blood cells (RBC) and evolve in a 48h cycle of asexual replication through three stages called ring, trophozoite, and schizont, that will ultimately release new merozoites. Some parasites invading RBC differentiate into gametocytes, that will evolve in five stages. Through a blood meal, female Anopheles mosquitoes uptake parasites, and the mature stage V gametocytes will differentiate into gametes and fuse to form a zygote. The zygote differentiates into a motile ookinete that crosses the mosquito midgut and encysts in the midgut wall (oocyst). Exponential replication occurs again in the oocyst leading to sporozoites. Rupture of the oocyst releases the sporozoites that reach the salivary glands, where they can lead to a new infection of the human host. Symptoms appear during the exponential asexual replication in RBC. A malaria episode is characterized by fever, headache, and chills. If untreated, malaria can rapidly evolve into a life-threatening state with multivisceral failure and coma.

Treatment relies on artemisinin derivatives used alone (severe malaria) or in combination (uncomplicated malaria) with chloroquine derivatives (amodiaquine, piperaquine, pyronaridine-targeting heme crystallization), amino-alcohols (mefloquine, lumefantrine-targeting 80S ribosome²) or antifolates (sulfadoxine, pyrimetamine-targeting Plasmodium DHFR and DHODH). These treatments are highly efficient but resistances that arose in South-East Asia complicate disease management and threaten elimination.³ Most commonly used antimalarial medications are derived or inspired by natural products, such as quinine, artemisinin, atovaquone or doxycycline.⁴ Natural products have been an incredible source of inspiration in drug discovery and fills a chemical space that is not reached by chemical synthesis.⁵

Here, we screened an in-house chemical library of hemisynthetic derivatives of the natural product trilobine (bisbenzylisoquinoline alkaloid) obtained from the plant Cocculus hirsutus. C. hirsutus is of the Menispermaceae family, is rich in alkaloids and a perennial climber widespread in tropical and sub-tropical regions. Various parts of this plant are notably used in Asia for the treatment of fevers.⁶ Activity of the trilobine hemisynthetic library was tested against P. falciparum asexual blood stage parasites, the stage responsible for all clinical symptoms. Hits belonging to the natural product family showed killing activity in the nanomolar range against P. falciparum. Chemical modifications of the lead compound improved the activity against the disease-causing stage of malaria, but also in the stages that transmit the disease. This compound is fast-acting and has sustained activity in multi-drug-resistant clinical isolates originating from Cambodia, including artemisinin-resistant. Efforts were made to develop derivatives with an optimized in vivo efficacy in an experimental mouse malaria model and to decipher the protein target of these natural product derivatives.

Results

Trilobine hemisynthetic derivatives are active against P. falciparum asexual blood stage

Screening of an in-house chemical library that targets distinct human epigenetic factors^7,8 identified hits against asexual blood stages of P. falciparum, belonging to the double bridge bisbenzylisoquinoline family of alkaloids found in C. hirsutus. We measured the activity of 94 natural and hemisynthetic derivatives. We fixed the threshold for activity at IC₅₀ of 200 nM and identified compounds 31 and 84 (Figure 1A), the latter being the most active with a mean IC₅₀ of 130 nM. This compound was then chosen as the lead compound. Hemi-synthesis of 84 from the trilobine natural analog cocsuline (4) is described in Figure 1B.

Figure 1.

IC₅₀ of the most potent bisbenzylisoquinolines out of the 94 derivatives of the chemical series and hemi-synthesis of compound 84

(A) IC₅₀ was measured in triplicate using P. falciparum NF54 strain, in three independent experiments. Growth was determined using the SYBR Green I assay after 72h incubation with the compounds. Points represent individual IC₅₀ value (in nM) measured in each experiment and lines represent the mean. Compound 84 is the most active with a mean IC₅₀ of 130 nM and was selected as the lead compound. DHA = dihydroartemisinin.

(B) Hemi-synthesis of 84 starting from the natural compound cocsuline (4): a) i) NaH, DMF, 0°C then RT, 30 min; ii) 3-dimethylaminopropylchloride, KI, 80°C, 15h.

Compound 84 is a fast-acting compound and is active throughout the asexual blood cycle

Compound 84 was tested for its stage-specific activity in the asexual blood cycle (Figure 2A). Tightly synchronized P. falciparum (NF54) parasites were exposed to concentrations corresponding to 3x or 10x the IC₅₀ of 84. Parasites were treated for 6h windows covering the entire 48h cycle. Survival was assessed by counting parasites in a thin blood smear 72h post-invasion (hpi). Compound 84 is active throughout the entire asexual cell cycle, with the highest killing activity between 18 and 42hpi. Importantly, 84 is active in the early ring stage (0-6hpi), with less than 5% surviving parasites at 10x the IC₅₀ and about 20% survival at 3x the IC₅₀. As 6h of incubation were sufficient to obtain efficient parasite killing, we assessed the speed of action of 84 using the Parasite Reduction Rate assay (PRR) developed by Sanz et al.⁹ We found that 84 is acting as fast as dihydroartemisinin (DHA), the active metabolite of artemisinin derivatives (Figure 2B), and chloroquine, in comparison to the slow-acting compound pyrimethamine.

Figure 2.

Activity of compound 84 in P. falciparum asexual blood stages

(A) Stage-specific activity in P. falciparum NF54 asexual cell cycle. Tightly synchronized parasites (0-3h post-invasion) were incubated with 84 for 6h, for 8 consecutive, 6h periods, covering the 48h cell cycle, at two different concentrations (3x and 10x the IC₅₀). Survival rate was evaluated 72h after synchronization (n = 3 independent experiments).

(B) In vitro parasite reduction rate (PRR). P. falciparum culture containing mostly ring and mono-infected red blood cells (RBC) were incubated with 84 or reference drugs (dihydroartemisinin (DHA), chloroquine (CQ) and pyrimethamine (Pyr)) for different time windows (6h, 24h, 48h, 72h, and 96h), washed, serially diluted in fresh RBC and allowed to grow for 3 weeks, then growth was assessed using the SYBR Green I assay.⁹ N = 1 experiment run in duplicate.

(C) Activity against P. falciparum multi-drug resistant field isolates adapted to culture. Left graph: IC₅₀s (dots) of 84 and two reference drugs (mefloquine, MQ, and desethlyamodiaquine, AQ) obtained against Cambodian field isolates are plotted versus the type of resistance carried by the isolate. Bars represent the median value. Right graph: Percentage of survival using the Ring-stage Survival Assay (RSA). Tightly synchronized ring-stage parasites (0-3h post-invasion) were incubated with 700 nM dihydroartemisinin (DHA), 3x the IC₅₀ of 84 (equivalent to 400 nM), or the combination of both, for 6h. After the drugs are washed out and the parasites are allowed to grow until 72h post-synchronization. Parasite survival is assessed by counting parasitemia in 10,000 RBC in a thin blood smear, in comparison to a DMSO-treated control.¹⁰ Bars represent the median value. Statistics = multiple Mann-Whitney tests in comparison with the drug-sensitive Cambodia condition (if discovery: ∗∗∗p value < 0.001; ∗∗p value < 0.01; ∗p value < 0.05).

Compound 84 has a sustained activity in P. falciparum multi-drug resistant field isolates

Since compound 84 is active in early ring stages and is fast-acting, we assessed its activity against multi-drug resistant field isolates adapted to culture, notably isolates resistant to artemisinin (Figure 2C). Using IC₅₀ as a proxy, 84 has a similar IC₅₀ in the laboratory strain NF54 and drug-sensitive isolates originating from Cambodia (median of 147 nM). Interestingly, isolates resistant to artemisinin have an increased sensitivity to 84 (median IC₅₀ of 88 nM) especially when they carry additional resistance to piperaquine and amodiaquine, with median IC₅₀ values dropping to 46 and 36 nM respectively (Figure 2C, left panel). In the Ring-stage Survival Assay (RSA),¹⁰ DHA shows, as expected, a difference in survival between artemisinin-susceptible and artemisinin-resistant isolates (DHA concentration used corresponds to 200 times the IC₅₀). Compound 84 shows no difference in survival, whichever the resistance background of the isolates, when used at a concentration of 3 times the IC_50. The observed survival values correspond to the value obtained in the laboratory line NF54 (Figure 2A). Notably, 84 combined with DHA reduces almost to zero the survival of artemisinin-resistant strains (Figure 2C, right), indicating a synergistic effect.

Preclinical studies reveal that 84 binds to human plasma proteins

To assess compound 84 in vivo, preclinical studies were realized to test the cytotoxicity, microsomal stability, and plasma binding properties. Microsomal stability half-life of 84 was higher than 60min and the cytotoxicity in HepG2 human cells resulted in a selectivity index superior to 75 (data not shown). However, 84 was found highly bound to plasma protein (>99% in both humans and mice), thus only a very minor fraction would be able to reach the target in the infected host cell. To overcome this problem, the chemical modulation of 84 was carried out (Figure 3A).

Figure 3.

Compound 84 and its derivatives: toward the best candidate for in vivo studies

(A) General synthesis of compounds 125-129. a) i) NaH, EtSH, DMF, 0°C, 30 min; ii) isotrilobine 2, 150°C, 30 min; b) i) NaH, DMF, 0°C, 30 min; ii) desired chloro-derivative, 80°C, 1 ON, 25-45%.

(B) Table reporting the mean IC₅₀ measured against P. falciparum, the solubility of the compound, the plasma binding fraction in mouse and human plasma, microsomal stability measured in mouse and human microsomes and the cytotoxicity measured in HepG2 human cell line, for 84 and its derivatives.

(C) Pharmacokinetic profile of compound 125 after a single dose of compound, administered either intravenously (i.v., 5 mg/kg), per os (p.o., 10 mg/kg) or intraperitoneally (i.p., 5 mg/kg). Mean plasmatic concentration (+/−SD) is represented in ng/mL versus time (5 timepoints over 24h).

(D) In vivo activity of 125 in C57BL/6 mice infected with P. berghei ANKA. Mice were treated i.p., 2h after parasite i.v. inoculation, for 4 days, with a daily regimen of 20 mg/kg or 10 mg/kg of 125 or 25 mg/kg of chloroquine or vehicle control. Parasitemia (mean ± SD) and survival were followed over 11 days.

Chemical synthesis of novel derivatives and in vitro activity

Selective double demethylation of the two methoxy substituents of the natural product isotrilobine (2) was realized to obtain two phenol groups that were substituted with the corresponding alkylamines to afford compounds 125-129 (Figure 3A, Data S3 and Scheme S1 related to Figure 3A). Importantly, compound 125 showed significantly lower human plasma protein binding (90% in human plasma and 95% in mice) and a similar antimalarial activity (123 nM) as compound 84 (130 nM), a good microsomal stability (half-life >60min) and acceptable cytotoxicity. Moreover, the citrate salt of this compound is soluble in water since no aggregation was found by Dynamic Light Scattering at a concentration of 15 mM (Figure 3B).

Compound 125 increases survival in vivo

Thus compound 125 was chosen for in vivo studies. First, pharmacokinetic properties were assessed (Figure 3C): 125 has a long half-life (superior to 24h) with very slow elimination after a single dose administered either i.v., p.o. or i.p.. Plasma concentration reached up to 600 ng/mL, 30min after i.p. injection. Antimalarial in vivo activity was then conducted using the Peter’s 4-day suppressive test in C57BL/6 mice infected i.v. with Plasmodium berghei ANKA (Figure 3D). While the lowest dosage at 10 mg/kg i.p. increased the survival by 3.5 days compared to the control, it did not reduce parasitemia. Most interestingly, dosing at 20 mg/kg i.p. increased survival by 5 days and significantly lowered parasitemia compared to the control group (two-way ANOVA with Dunett’s correction, adjusted p value < 0.0001). However, complete cure was not achieved, and dose escalation was avoided due to potential toxicity in mice at higher doses.

125 is active in the transmission and liver stages

First, we tested compound 125 in early (stage IIb-III) and late (stage IV-V) gametocytes (Figure 4A). This compound has an IC₅₀ of around 1 μM in both stages in P. falciparum NF54 gametocytes and was slightly more active in the Cambodian strain 3601E1 resistant to artemisinin (Kelch13 R539T) with an IC₅₀ of 700 nM (Figure 4B). Interestingly, the IC_50s were identical in both early and late gametocyte developmental stages, while antimalarials are usually less efficient against stage V, which is in a temporary quiescent state. For example, puromycin, used here as a reference (Figure 4B), does not reach complete killing.

Figure 4.

Activity of compound 125 in transmission and liver stages

(A) Scheme showing the methodology used to assess anti-transmission activity. Early (IIb-III, blue) and late (IV-V, red) stage gametocytes were treated in vitro for 72h before survival was assessed using a flow cytometry method (SyBR Green I and Mitotracker deep red FM dyes). Bottom images represent the phenotype of early and late-stage gametocytes treated with 5 μM of compound 125. For in vivo mosquito studies (represented in orange), treatment was added to the infectious blood meal, containing mature stage V gametocytes, just before mosquito feeding. After 10 days of development, the number of oocysts was counted in each mosquito midgut. Pictures of gametocytes are taken and adapted from Josling and Llinás.¹¹

(B) Gametocytocidal activity of 125 against early (IIb-III) and late (IV-V) stage gametocytes in NF54 drug-susceptible strain and 3601E1, a cloned Cambodian isolate bearing R539T Kelch13 mutation that confers 21% survival after exposure to 700 nM dihydroartemisinin (DHA) in the RSA. Puromycin was used as a positive control. N = 3 independent experiments run in triplicates (mean ± SD).

(C) Impact of compound 125 on parasite development in the mosquito. Each dot represents the number of oocysts counted in each mosquito midgut and the line represents the median. MB = methylene blue. N = 4 independent experiments. Statistical analysis was done using the Mann-Whitney test. ∗∗∗∗p value < 0.0001; ∗∗∗p value = 0.0002.

(D) Activity of 125 in P. falciparum liver stages (mean with SD). Compound was added after sporozoites invasion of human primary hepatocytes. Media containing an increasing concentration of the compound was changed daily for 72h. Primaquine at 10 μM was used as a positive control. Exoerythrocytic forms (EEF) survival was assessed using immunostaining of parasites with PfHSP70 antibody. N = 1 experiment run in triplicate.

We then assessed the efficacy of parasite development in Anopheles mosquito stages by adding 125 to the infectious blood meal. The effect on parasite transmission was evaluated by counting the number of midgut oocysts, 10 days after the blood feeding (Figure 4C). Compound 125 significantly decreased the oocyst number at 10 μM without an increase in mosquito mortality compared to DMSO (data not shown).

Lastly, we tested compound 125 in primary human hepatocytes infected with P. falciparum sporozoites. Sporozoites invasion was performed for 3h, then cells were washed to remove remaining sporozoites and treated with 125 for 72h. Parasite survival was assessed by counting the number of intracellular parasite forms using immunofluorescence staining with anti-HSP70 antibody. We observed a dose-dependent inhibition of parasite development in hepatocytes (Figure 4D), indicating that 125 targets proteins expressed during multiple life cycle stages of P. falciparum.

Investigation of P. falciparum cellular targets

To decipher the mode of action of this family of bisbenzylisoquinoline alkaloids, we synthesized a chemical probe 132 (Data S3 related to Figure 5A) derived from the parent natural product trilobine (compound 1) and harboring a photo-activable benzophenone and a terminal alkyne group (Figure 5A). It was used to photo-crosslink directly in cells proteins bound to the probe or in its vicinity and, by 1,3-dipolar cycloaddition (CuAAC), to simultaneously label the cross-linked proteins with an azide derivative containing both a fluorescent reporter (TAMRA) and biotin for protein pull-down followed by mass spectrometry analysis and gel electrophoresis visualisation.¹⁴ The probe 132 has a similar IC₅₀ as trilobine (790 nM versus 410 nM, Figure S1 related to Figure 5), indicating that the derivatization was not impeding in-cellulo target engagement. It was then used to treat parasites either alone or in competition with increasing concentrations of the parent compound trilobine 1. Harvested parasites were saponin-lysed to remove hemoglobin contaminants and then parasite pellets were resuspended and UV-irradiated at 365 nm to crosslink the compound with the nearby target. Cells were then lysed and the click reaction was initiated in parasite extracts. Purified functionalized proteins were then enriched using neutravidin beads, digested, and finally analyzed via LC−MS/MS (Figure 5A). In gel fluorescence of the different conditions is shown in Figure 5B. Target deconvolution was done by plotting the competition curves i.e. protein abundance when probe 132 was used alone (10 μM) or in competition with 25 μM, 50 μM or 100 μM of parent compound 1. Only candidates producing curves with consistency between triplicates were selected. Among these, proteins that were not enriched compared to DMSO were discarded, as well as proteins with shared peptides (multigenic families). Clustering of the selected proteins was done using STRING software.¹² Two different pathways were revealed: polyadenylate-binding protein 1 complex (PABP1 and its partners), involved in protein translation, and proliferating cell nuclear antigen (PCNA1) involved in DNA replication and DNA methylation (Figures 5C and 5D). Noteworthy, none of those candidates has been previously associated with drug targeting by antimalarials or drug-resistance mechanisms.

Figure 5.

Chemical pull-down to decipher the protein targets

(A) Chemical pull-down strategy. Chemical probe 132, containing a photo-activable crosslinker moiety (in blue) and a reactive alkyne (in pink) to initiate the bioorthogonal 1,3-cycloaddition reaction, is incubated with parasites. After photo-crosslinking, click-labeling with biotin and a fluorescent reporter (TAMRA), and avidin-biotin affinity purification, enriched protein partners were analyzed by mass spectrometry.

(B) In-gel fluorescence of the pulled-down proteins labeled with TAMRA. Probe 132 was incubated with parasites at a concentration of 10 μM and competition with trilobine (1) was done at 25 μM (2.5x), 50 μM (5x), and 100 μM (10x), in triplicate. DMSO was used as a negative control. Colored arrows indicate the molecular weight of the proteins with matched colors in panel C.

(C) Clustering of enriched protein selected for their (i) decreasing quantity in competition with trilobine, (ii) enrichment compared to DMSO, and (iii) consistency between triplicates. Clustering was realized using the STRING¹² database.

(D) Table reporting the essentiality,¹³ molecular weight (MW) and GO term biological process of the chemical pull-down enriched proteins.

Discussion

We explored the activity of bisbenzylisoquinoline alkaloids, isolated from the plant C. hirsutus, in P. falciparum. The antimalarial activity of trilobine and its natural alkaloids analogs was significantly improved by synthesizing a chemical library of hemisynthetic compounds. Importantly, none of the multi-drug resistant clinical isolates from Cambodia showed cross-resistance to the lead compounds, indicating that this chemical scaffold targets parasite pathways distinct from commonly used antimalarial drugs. Furthermore, lead compounds are fast-acting, comparable to chloroquine and artemisinin. Another important activity is that the lead molecule kills the early-ring stage, which is able to resist high concentration of artemisinin by entering into dormancy.¹⁵ The lead compound 84 can restore antimalarial activity in artemisinin-resistant isolates when combined with DHA (Figure 2C), which highlights the potential of this chemical series to be used in combination with artemisinin. Natural bisbenzylisoquinoline alkaloids extracted from diverse plants have been already described for their antimalarial properties^{16,17,18,19,20,21,22} and previous re-sensitization has been described for bisbenzylisoquinolines in chloroquine-resistant P. falciparum strains, as well as in tumor cells resistant to vinblastin.^20,23 Here we report the first hemisynthetic study resulting in compounds with improved pharmacological properties of these natural products, active in resistant strains including artemisinin-resistant ones and all along the life cycle of the parasite.

As compound 84 was found to be highly bound to plasma protein (>99%), precluding its use for in vivo studies, new derivatives were synthesized resulting in compound 125 with a decreased plasma protein binding. Compound 125 showed sustained antimalarial activity (123 nM), water solubility, high microsomal stability, and low cytotoxicity (Figure 3B). Pharmacokinetic studies revealed rapid absorption and very slow elimination with an elimination half-life far superior to 24h (Figure 3C). In vivo antimalarial activity in mice infected with P. berghei shows that four doses at 20 mg/kg of 125 delay the onset of parasitemia and increase mice survival by 5 days (Figure 3D).

Compound 125 is also targeting transmission stages (in vitro and in vivo) and liver stages of P. falciparum, though at higher concentration than in the blood stage (Figure 4). This indicates that the protein target(s) are expressed in different stages of Plasmodium’s life cycle. To decipher the targeted proteins, we synthesized a chemical probe from the parent compound trilobine and used UV-affinity capture and chemical pull-down followed by mass spectrometry analysis. Target deconvolution of mass spectrometry hits revealed different protein interactions with the probe. The predicted targets have in common to have house-keeping functions in diverse vital pathways of P. falciparum such as the polyadenylate binding protein 1 (PABP1, Pf3D7_1224300) and the proliferating cell nuclear antigen 1 (PCNA1, Pf3D7_1361900) (for details, see Figure 5D). PABP1, its protein partner polyadenylate binding protein interacting protein 1 (PABP-IP1, Pf3D7_1107300) and MA3 domain-containing protein (MDCP, Pf3D7_1457300) are involved in protein translation.^24,25 They bind to the polyA tail of mRNA and are part of a complex with Eukaryotic translation Initiation Factor 4F (EIF4F) to mediate the recruitment of ribosomes to mRNA.²⁶ The PCNA1 protein is involved in DNA replication and DNA methylation,²⁷ and in DNA-damage response.²⁸ PCNA1 is clustering with the ribonucleoside-diphosphate reductase subunit m1 (RNRm1, Pf3D7_1437200). This protein has an important role in DNA synthesis since it catalyzes the rate-limiting step of de novo synthesis of the deoxyribonucleotides by reducing the corresponding ribonucleotides.²⁹ To our knowledge, the identified pathways have never been targeted by antimalarials or have been involved in a malaria parasite drug-resistance mechanism. Indeed, we were not able to obtain resistant parasites upon low-concentration continuous treatment with the compounds (data not shown). Interestingly, these pathways are vital for the other life cycle stages targeted by the compounds (transmission, liver stages). These proteins are well conserved in other malaria species (Plasmodium malariae, vivax, ovale, and knowlesi), and thus likely to kill these species as well, except for PABP-IP1, which is non-essential and not annotated in the genomes of Plasmodium vivax, ovale, and knowlesi. We show here for the first time that the bisbenzylisoquinoline trilobine could affect specific proteins involved in DNA replication and protein translation. This work lays out strategies for further studies to validate the protein target(s) by complementary methods such as the use of recombinant proteins. Defining the target will help the future preclinical drug-development process, by facilitating the improvement of selectivity compared to human homologs protein(s) in order to optimize the compound and deliver a new drug candidate.

In conclusion, we identified a novel chemical scaffold derived from a natural product with activities against all life cycle stages that interacts with diverse essential parasite pathways.

Limitations of the study

The main limitation is due to the supply of plant extracts. Importantly, the plant is abundant, and all parts contain isotrilobine, which is easily separated and is the precursor of lead 125. Additional studies are necessary to characterize the target and this will help optimize the best lead.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
PfHSP70	StressMarq Biosciences	Cat# SPC-186, RRID:AB_1608355)
Deposited data
Proteome data	PRIDE repository	Proteomics Identifications (PRIDE) (RRID:SCR_003411)
Experimental models: Cell lines
HEPG2 cell line	ATCC	RRID:CVCL_0027
3T3-J2	IZSLER	Cat# BS CL 153, RRID:CVCL_W667
Experimental models: Organisms/strains
C57BL/6 mice	MGI	Cat# 2159769, RRID:MGI:2159769

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Paola B. Arimondo (paola.arimondo@pasteur.fr).

Materials availability

There are restrictions to the availability of compounds due to necessity to resynthesise them ad hoc.

Experimental model and subject details

HepG2 cells

Human HepG2 (hepatocellular carcinoma, RRID:CVCL_0027) were grown according to the provider instructions and at 37°C and 5% CO₂.

Primary human hepatocytes

Cryopreserved primary human hepatocytes (purchased from BioIVT) were co-cultured with J2-3T3 murine embryonic fibroblasts (gift of Howard Green, Harvard Medical School) as described previously.³⁰ Cultures were maintained in Dulbecco’s Modified Eagle Medium (DMEM, Corning) with 10% fetal bovine serum (GIBCO), 1% ITS (insulin/transferrin/selenous acid and linoleic acid, BD Biosciences), 7 ng/mL glucagon (Sigma-Aldrich), 40 ng/mL dexamethasone (Sigma-Aldrich), 15 mM HEPES (GIBCO), and 100 mg/mL penicillin-streptomycin (Corning). Cells were kept at 37°C in a 5% CO₂ environment.

P. falciparum continuous culture

P. falciparum parasites were cultured using a standard protocol.³¹ The laboratory strains used were NF54 and 3D7; that were grown in RPMI1640 supplemented with HEPES (25 mM) and L-glutamine (2 mM) (Gibco), 0.5% (m/v) Albumax I, hypoxanthine (100 nM) and gentamycin (16 ng/mL), at 37°C, under reduced oxygen pressure (5% O₂, 5% CO₂). P. falciparum clinical isolates used in this study were collected in the framework of the therapeutic efficacy surveillance program in Cambodia between 2017 and 2019. Culture was performed in the same condition as for laboratory strains except that 5% (v/v) human serum and 0.25% (m/v) Albumax II were used for serum supplementation.

In vivo antimalarial activity

All animal experiments performed in the manuscript were conducted in compliance with French animal welfare laws and Institut Pasteur Animal Care and Use Committee guidelines. In vivo activity was determined as previously described³² following the Peters 4-day suppressive test³³ with slight modifications. 6-weeks old C57BL/6 female mice (Janvier Labs, MGI Cat# 2159769,RRID:MGI:2159769) were infected intravenously (i.v.) with 10⁵ iRBC of P. berghei ANKA GFP-expressing parasites.³⁴ 2h post-infection, mice were randomly assigned to experimental groups and treated intraperitoneally (i.p.) for 4 days with 125 following a daily regimen of 20 mg/kg or 10 mg/kg, or 25 mg/kg of chloroquine, or the equivalent vehicle control. Parasitemia was quantified from blood samples collected every day by flow cytometry of 200,000 RBCs and confirmed by Giemsa-stained thin blood smears. The experiment was performed once (n = 6 mice per group).

Method details

Chemical library synthesis

Purification of trilobine and its natural products and hemi-synthesis of the trilobine derivatives are described in.^7,35

Chemical synthesis

All chemicals were from Sigma-Aldrich, Alfa Aesar and FluoChem. NMR experiments were recorded on an Agilent DirectDrive 500 spectrometer (Agilent Technologies, Santa Clara) with a proton resonating frequency of 500 MHz. Spectra were recorded using VnmrJ 4.2A (Agilent Technologies). Chemical shifts are given in ppm. Coupling constants J are measured in Hz. Splitting patterns are designed as follows: s, singlet; bs broad singlet; d, doublet; bd broad doublet; t, triplet; brt, broad triplet; dd, doublet of a doublet; m, multiplet; ddd, doublet of a doublet of a doublet; q, quartet; p, quintet, sext, sextet. HRMS analysis were performed on a Q Exactive mass Spectrometer (Thermofisher) using direct injection. Samples were previously dissolved in a mix of water and acetonitrile (50/50) and 0.1% of formic acid. Full scans (150–2000 Da) were acquired in positive ion mode with a resolution of 70,000. The purity of all final compounds was verified to be higher than 95% using reversed-phase HPLC system (Agilent 1200 series) equipped with a diode-array detector on a C18 reverse phase column (Kromasil, 5 μm, 100 Å; 4.6 × 150 mm) at a flow rate of 1 mL min⁻¹ with a linear gradient of acetonitrile in 10 mM triethylammonium acetate buffer over 20 min (05–95% CH₃CN).

Synthesis and chemical analysis of compounds 84, 125, 126, 127, 128 and 129 is detailed in the Data S1 and Data S2. The chemical synthesis of probe 132 is reported in Data S3.

Compound solubility/aggregation

Compounds were mixed with citrate (1:1) to obtain the salt and then diluted in water at a final concentration of 15 mM 20 μL of each sample was then screened for aggregates by Dynamic Light Scattering on a DnyaPro plate reader III (Wyatt technology).

Microsomal stability and plasma binding

Rat and human microsomal stability and plasma binding studies were outsourced at Fidelta Ltd., Zagreb, Croatia.

Cytotoxicity in HepG2 cells

Cytotoxicity was evaluated after 72h of drug treatment in 96-well plates by ATPlite (PerkinElmer) according to the manufacturer’s protocol. Experiments were run in technical triplicates and at least in two biological replicates. GraphPad Prism 8 was used to interpolate IC₅₀. DMSO was used as control.

Inhibition activity in the asexual stages

IC₅₀ values were obtained using a range of 7-point and 2-step serial dilutions starting at 2 μM, 0.7% starting parasitemia and 2% final hematocrit. Growth was assessed after 72h using the SYBR Green I assay. GraphPad Prism 8 was used to interpolate IC₅₀ from three independent experiments run in triplicate. DHA and DMSO were used as positive and negative controls, respectively.

Stage-specificity during the asexual cell cycle

Stage-specificity was assessed using previously described methods.³⁶ Briefly, asexual NF54 parasites were tightly synchronized (0–3hpi) and incubated for 6h with compound 84 at 3× and 10× IC₅₀ value, at 0, 6, 12, 18, 24, 30, 36, or 42hpi. Following each treatment, cells were pelleted, washed with 10 mL of RPMI medium, and put back into culture in a new plate. Parasitemia was assessed at 72h post-synchronization using Giemsa-stained thin blood smears. The percentage of survival was compared to DMSO-treated parasites. Data were obtained from three independent experiments (one well per condition).

In vitro parasite reduction rate (PRR)

The in vitro parasite reduction rate was adapted from Sanz et al.⁹ 10⁶ infected RBC containing mostly rings were cultivated separately with 10× the IC₅₀ of 84 or reference drugs (DHA, Chloroquine (CQ) or Pyrimethamine (Pyr)) for 6h, 24h, 48h, 72h and 96h (0.5% starting parasitemia, 2% hematocrit of P. falciparum 3D7). After each incubation time, cells were washed and serially diluted (1/3) with fresh RBC. Growth was assessed after 3 weeks, using the SYBR green I assay. The number of wells showing parasite growth is correlated with the number of viable parasites at the different time points (as defined by Sanz et al.) and allows defining the growth parameters reported in Figure 2B. The experiment was performed once in duplicate.

Activity against P. falciparum Cambodian field isolates

P. falciparum isolates used in this study were collected in the framework of the therapeutic efficacy surveillance program in Cambodia between 2017 and 2019. PSA, RSA, AQSA and genotyping data have been performed in the Pasteur Institute in Cambodia. Parasites were chosen based on the pattern of multi-drug resistance they carry. IC₅₀ values were obtained as previously described³⁶ with slight modifications: ring-synchronized cultures (2% hematocrit, 1% starting parasitemia) were incubated for 96h before growth measurement using the SYBR green I assay. A range of 11-point and 2-step serial dilutions, starting at 1 μM were used to assess the activity of 84. Mefloquine (MQ) and monodesethylamodiaquine (dAQ) were used as positive controls. Experiments were performed in triplicate, with one biological replicate per isolate. Statistical analysis was performed using multiple Mann-Whitney tests in comparison with the drug-sensitive Cambodia condition.

Ring-stage survival assay (RSA)

The RSA¹⁰ was determined by treating 0-3h synchronized ring-stage parasites, for 6h, with either DMSO, 700 nM DHA, 3xIC₅₀ of 84 or the combination of both. The IC₅₀ used was the one obtained against the NF54 strain. After 6h, drugs were washed out and parasites were put back into culture. Parasitemia was assessed at 72h post-synchronization using Giemsa-stained thin blood smears. Percentage of survival was compared to DMSO-treated parasites and data were obtained from one independent experiment, in the different isolates. Statistical analysis was performed using multiple Mann-Whitney tests in comparison with the drug-sensitive Cambodia condition.

Pharmacokinetic studies

PK studies were outsourced at Fidelta Ltd., Zagreb, Croatia, according to standard procedures.

Activity against synchronous gametocytes

For this experiment, we used two different strains that retain their capacity to produce gametocytes: the drug-susceptible laboratory strain NF54 and an artemisinin-resistant (Kelch13 R539T) cloned Cambodian isolate 3601E1. Synchronous gametocytes were induced following previously described methods³⁷ with slight modifications: trophozoite purification steps were performed using gelatin flotation (Plasmion). Media containing 0.05% (w/v) Albumax I was used until induction (day 0, corresponding to gametocyte-committed ring stage parasites), followed by a media containing 5% (v/v) human serum and 0.025% (w/v) Albumax I. Asexual parasite populations was removed by N-acetylglucosamine treatment for 5 days. Parasites were treated with 125 and the positive control puromycin to evaluate efficacy against early (day 4, stage IIb-III) and late-stage gametocytes (day 9, stage IV-V). Parasites, at approximately 1% starting gametocytemia, were seeded in 96-well plates at 2% final hematocrit and treated for 72h with a 7-point and 2-step serial dilutions of compounds starting at 10 μM for 125 or 2.5 μM for puromycin. Readout of parasite survival was done by flow cytometry (Guava 6HT, Luminex corp) using a Mitotracker Deep Red^FM (molecular probes, 50 μg, reference M22426) and SYBR™ Green I (10,000x concentrate in DMSO, Lonza) staining. SYBR Green I was used at a final dilution of 0.5x and Mitotracker at a final concentration of 100 nM, in prewarmed PBS (37°C). Data were acquired on 10,000 red blood cells, in triplicate, in 3 independent experiments.

Standard membrane feeding assay

The culture of gametocytes is performed in a semi-automated system developed in Nijmegen, Netherlands³⁸ at an initial parasitemia of 0.5%. After 4–6 days, the parasitemia reaches approximately 5–10% and gametocytogenesis is induced. The cultures are checked twice a week to evaluate the presence of the different stages of gametocyte development. Infectious stage V gametocytes are observed after approximately two weeks of culture. To assess the maturity of the male gametocytes, an aliquot of the culture is taken and the number of exflagellation events are counted under optical microscope, using a 400x magnification, by diluting infected red blood cells in human serum. Cultures containing mature gametocytes are diluted to 1:6 in a mix of 50% fresh red blood cells and 50% human serum. Compound 125 (1 or 10 μM final concentration), methylene blue (MB) (10 μM) or DMSO are added to this mix and fed to 50 Anopheles stephensi females per condition, using a membrane feeder. The mosquitoes are maintained at 26°C and 70% humidity with 10% sucrose as food source. Ten days post-infection, the mosquito midguts are dissected and stained with 0.25% fluorescein and the number of oocysts per midgut are counted. The results have been obtained in 4 independent experiments. Statistical analysis was done using Mann-Whitney test.

Liver-stage activity

Micropatterned human primary hepatocytes surrounded by supportive 3T3-J2 stromal cells (IZSLER Cat# BS CL 153, RRID:CVCL_W667)³⁹ were seeded on glass-bottom 96-well plates and infected the following day with 70,000 sporozoites per well of P. falciparum NF54. Cells were centrifuged at 700g for 5min to settle down the parasites. Sporozoites were allowed to invade for 3h before washing the cells, then fresh media containing drugs was added and changed daily. Cells were incubated for 72 h at 37°C and 5% CO₂ before fixation and permeabilization with ice-cold methanol and staining with dapi and PfHSP70 antibody (StressMarq Biosciences Cat# SPC-186, RRID:AB_1608355). Readout was done by counting PfHSP70-stained exoerythrocytic forms (EEF) in fluorescence microscopy. 125 was tested at 4 different concentrations, ranging from 10 to 0.37 μM, using a 3-fold dilution. Primaquine diphosphate at 10 μM was used as a control. Experiment was done once in triplicate.

UV-affinity capture and protein pull-down

Asynchronous P. falciparum NF54 cultures (1.10⁸ parasites per condition) were treated for 1h with 10 μM of probe or DMSO. After 30min, the competitor was added in increasing concentration (DMSO, 25 μM, 50 μM or 100 μM of trilobine, 1) for 30min before infected red blood cells (iRBC) were harvested, saponin-lysed in ice (0.15% saponin in PBS) and washed thoroughly in ice-cold PBS to remove hemoglobin contaminants. Parasite pellets were resuspended in PBS and irradiated for 1h at 365 nm, on ice, to crosslink the probe to its protein targets. Parasites were centrifuged and pellets were lysed in a buffer containing 1% SDS, 1 mM dithiothreitol (DTT) and protease inhibitors (Roche) in PBS. Protein concentration was determined and extracts were diluted to a concentration of 1 mg/mL. Click-labelling was done following described protocol⁴⁰: 100 μL of extracts were treated with a 6 μL premix solution containing the following reagents to obtain a final concentration of: 60 μM of TAMRA-biotin-azide (Jena Bioscience, catalog number CLK-1048-5), 1 mM CuSO₄ in water (freshly prepared), 1 mM tris(2-carboxyethyl)phosphine (TCEP) and 100 μM tris[(1-benzyl-1H-1,2,3-triazol-4-yl)methyl]amine (TBTA). Tubes were shaken at room temperature for 1h after which proteins were precipitated with sequential addition and vortexing of methanol (200 μL), chloroform (50 μL) and ddH₂O (100 μL). Tubes were then centrifuged for 2 min at 14,000g and the protein layer was washed twice with 200 μL of methanol. Methanol was discarded and remaining solvent was let to evaporate at room temperature. Proteins were resuspended in SDS buffer (1% SDS, 10 mM DTT). Samples were then enriched on avidin-coupled agarose beads (PierceNeutrAvidin Agarose, ThermoFisher Scientific; 50 μL, pre-washed three times in 0.2% SDS in PBS) by incubation with gentle shaking for 2h at room temperature to bind and enrich biotin-labeled proteins thanks to the probe. The supernatant was then removed, and the beads were washed with 1% SDS in PBS (3 × 500 μL), 4M urea in 50 mM ammonium bicarbonate (AMBIC) (2 × 300 μL) and 50 mM AMBIC (5 × 300 μL). This experiment was done in three independent replicates.

Protein digestion

Reduction of disulfide bonds of enrich biotin-labeled proteins in 50 mM AMBIC was achieved by addition of DTT to final concentration of 5 mM and incubation for 30 min at 56°C with shaking. Then, iodoacetamide (IAA) was added at a final concentration of 20 mM for cysteine alkylation, and samples were incubated in dark for 30 min at 37°C. Samples were digested for 4h using LysC (Promega, MA, USA). Subsequently, 1 μg of trypsin (Promega, MA, USA) was added, and samples were digested overnight at 37°C. Digestion was quenched by addition of formic acid (FA). The digest was subsequently desalted using Sep-Pak C18 cartridges (Waters) and lyophilized in vacuum for further analysis.

LC-MS/MS analysis

All the samples were analyzed on a Q Exactive Plus Hybrid Quadrupole-Orbitrap Mass Spectrometer coupled with an Easy nLC 1200 ultra-high pressure liquid chromatography system (Thermo Fisher Scientific) with solvent A of 0.1% formic acid and solvent B of 0.1% formic acid/80% acetonitrile. 1 μg of each sample was injected on a 75-μm ID PicoFrit column packed in-house to approximately 30-cm length with ReproSil-Pur C18-AQ 1.9-μm beads (Dr. Maisch). Column equilibration and peptide loading were done at 900 bars in buffer A (0.1% FA). Samples were separated at a 250 nL/min flow rate with a gradient of 2–7% buffer B in 5min, 23% buffer B in 70min, 23 to 45% buffer B in 30min, 45 to 95% buffer B in 5min, followed by a hold at 95% for 7min and back to 2% for 15min. Column temperature was set to 60°C. Mass spectrometer was operated in data-dependent acquisition mode. The MS instrument parameters were set as follows: MS1, r = 70,000; MS2, r = 17,500; MS1 AGC target of 3e6; MS2 for the 10 most abundant ions using an AGC target of 1e6 and maximum injection time of 60ms; and a 45s dynamic exclusion. The isolation window was set to 1.6m/z and normalized collision energy fixed to 28 for HCD fragmentation. Unassigned precursor ion charge states as well as 1, 7, 8 and >8 charged states were rejected and peptide match was disable.

Protein identification and quantification

Acquired LC-MS/MS raw data were analyzed using MaxQuant 1.5.3.8 version⁴¹ using the Andromeda search engine⁴² against P. falciparum, strain NF54 proteome database (5,548 entries, download from PlasmoDB) and human reference proteome (89,260 entries, download Jan2018), concatenated with usual known mass spectrometry contaminants and reversed sequences of all entries. All searches were performed with oxidation of methionine and protein N-terminal acetylation as variable modifications and cysteine carbamidomethylation as fixed modification. Trypsin was selected as protease allowing for up to two missed cleavages. The minimum peptide length was set to 7 amino acids. The false discovery rate (FDR) for peptide and protein identification was set to 0.01. The main search peptide tolerance was set to 4.5ppm and to 20ppm for the MS/MS match tolerance. One unique peptide to the protein group was required for the protein identification. The match between run was selected. A false discovery rate cut-off of 1% was applied at the peptide and protein levels. All mass spectrometry proteomics data have been deposited at ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD036288 (RRID:SCR_003411).

Pull-down data analysis

MaxQuant LFQ values were filtered for contaminants and the dose-response curves of all proteins with a triplicate >0 LFQ value for the DMSO-dosed pulldown were fitted using a four-parameter log-logistic regression using an R script that utilizes the ‘drc’ package.⁴³ Manual inspection of the curves helped select the potential targets which were set in a functional context using STRING.¹²

Quantification and statistical analysis

Statistical analysis used is reported in each relevant section of the methodology and in the corresponding figure legend.

Published: February 17, 2023

Data and code availability

Data. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD036288.

Project Webpage:http://www.ebi.ac.uk/pride/archive/projects/PXD036288.

FTP Download:ftp://ftp.pride.ebi.ac.uk/pride/data/archive/2023/01/PXD036288.

Code. This paper does not report original code.

Additional information. Any additional information required to reanalyse the data reported in this paper is available from the lead contact upon request.