Identification of fungal natural products with potent inhibition in Toxoplasma gondii

ABSTRACT In an effort to identify novel compounds with potent inhibition against Toxoplasma gondii, a phenotypic screen was performed utilizing a library of 683 pure compounds derived primarily from terrestrial and marine fungi. An initial screen with a fixed concentration of 5 µM yielded 91 hits with inhibition comparable to an equal concentration of artemisinin. These compounds were then triaged based on known biological and chemical concerns and liabilities. From these, 49 prioritized compounds were tested in a dose response format with T. gondii and human foreskin fibroblasts (HFFs) for cytotoxicity. Ten compounds were identified with an IC50 less than 150 nM and a selectivity index (SI) greater than 100. An additional eight compounds demonstrated submicromolar IC50 and SI values equal to or greater than 35. While the majority of these scaffolds have been previously implicated against apicomplexan parasites, their activities in T. gondii were largely unknown. Herein, we report the T. gondii activity of these compounds with chemotypes including xanthoquinodins, peptaibols, heptelidic acid analogs, and fumagillin analogs, with multiple compounds demonstrating exceptional potency in T. gondii and limited toxicity to HFFs at the highest concentrations tested. IMPORTANCE Current therapeutics for treating toxoplasmosis remain insufficient, demonstrating high cytotoxicity, poor bioavailability, limited efficacy, and drug resistance. Additional research is needed to develop novel compounds with high efficacy and low cytotoxicity. The success of artemisinin and other natural products in treating malaria highlights the potential of natural products as anti-protozoan therapeutics. However, the exploration of natural products in T. gondii drug discovery has been less comprehensive, leaving untapped potential. By leveraging the resources available for the malaria drug discovery campaign, we conducted a phenotypic screen utilizing a set of natural products previously screened against Plasmodium falciparum. Our study revealed 18 compounds with high potency and low cytotoxicity in T. gondii, including four novel scaffolds with no previously reported activity in T. gondii. These new scaffolds may serve as starting points for the development of toxoplasmosis therapeutics but could also serve as tool compounds for target identification studies using chemogenomic approach.

individuals, newborns, and developing fetuses is often fatal.However, to date, no vaccine or fully efficacious drugs are available for prophylactic and therapeutic treatment.A combination of pyrimethamine and sulfadiazine is the current gold standard therapy for T. gondii.These two drugs act synergistically to disrupt DNA synthesis by interfering with folate production.During therapy, folinic acid (leucovorin) can be supplemented to restore the host's folate levels without impacting the parasite due to a lack of membrane transport mechanisms in T. gondii for exogenous folate.Unfortunately, sulfadiazine can induce side effects such as allergies and bone marrow suppression (2).As a result, sulfadiazine is sometimes replaced with a protein synthesis inhibitor such as azithro mycin, or an electron transport inhibitor such as atovaquone (3).Cerebral and ocular toxoplasmosis are the most challenging forms of infection to treat, as drugs must cross the blood brain barrier to reach the target sites at an effective concentration.For cerebral toxoplasmosis, in addition to pyrimethamine and sulfadiazine, trimethoprim combined with sulfamethoxazole (co-trimoxazole) or atovaquone can be used.Clindamycin or pyrimethamine-sulfadiazine in combination with corticosteroids are also given to treat ocular toxoplasmosis.However, resistance and treatment failures are well documented with pyrimethamine and sulfadiazine (2).Without long-term treatment, 80% of the cases will relapse and around 40% of the patients may discontinue treatment due to side effects (2).Drugs are also needed for treating bradyzoites, the chronic stage of T. gondii.Together, these challenges highlight the need to develop novel and efficacious therapeutics for the treatment of toxoplasmosis.
Natural products have historically been a major source of therapeutics for a variety of parasitic diseases (4).In addition to containing unexplored potential, natural products offer greater diversity, structural complexity, and molecular rigidity over synthetic molecules (4).Some of the most successful natural products to date include the antimalarials artemisinin and quinine.Additional screening of natural product libraries could provide exciting and unique starting scaffolds with the possibility for further optimization.To that end, we performed a phenotypic screen of a natural product library containing 683 fungal derived compounds against T. gondii tachyzoites in order to identify potent and selective inhibitors.This library had previously been cross-screened against the related apicomplexan parasite, Plasmodium falciparum (Dd2) by utilizing a SYBR Green I fluorescence-based assay (5).This work identified 78 unique compounds with greater than or equal to 50% inhibition at 1 µM, of which 57 demonstrated an EC 50 less than 1 µM.Additionally, the cytotoxicity of these compounds was determined in HepG2 cells, and scaffolds with high potency and low cytotoxicity were selected for further study (5).

RESULTS
This screen of 683 pure natural products at 5 µM yielded 91 initial hits with at least 94% inhibition, equal to or greater than the inhibition of the same concentration of artemisinin.Pyrimethamine or sulfadiazine was not used as positive controls due to the resistance of the strain to these drugs.This strain of RH88 was genetically engineered to express luciferase and had an insertion of DHFR as selection marker (6).The average Z′ score across all plates was 0.54 (Fig. 1).
Upon review of the initial set of hit molecules, 42 compounds were triaged out due to various biological and chemical concerns such as poor chemical stability, known human toxicity, compound availability, and target promiscuity.This resulted in a curated subset of 49 fungal metabolites with promising chemical properties and biological activities that warranted further investigation.These compounds were then subjected to IC 50 determination and cytotoxicity screening.To test the IC 50 assay, we used three positive controls that have IC 50 data published in T. gondii, namely, KAE609, artemisinin, and methylene blue.The IC 50 values of these compounds (Table 1) were found to be comparable to published values (7).With the method validated, we generated the IC 50 data of the 49 compounds as shown in Table 1.
Chemoinformatic analysis on the 49 compounds identified seven clusters with enrichment at rates greater than those expected by chance (P = 5.11 × 10 −4 -1.62 × 10 −2 ) (Fig. 2).Among these clusters, four scaffold groups stood out as containing compounds with high potency and low cytotoxicity.These included the peptaibol, xanthoquinodin, heptelidic acid, and cyclic tetrapeptide scaffold groups.In addition, fumagillin stood out as a single ungrouped scaffold.For further analysis, we focused our efforts on 18 compounds with submicromolar activity and favorable selectivity, of which 10 demon strated an IC 50 less than 150 nM (Table 1; Fig. 3).

DISCUSSION
As previously mentioned, this compound library had also been tested against the P. falciparum parasite line Dd2 (5).Many of the inhibitors demonstrated dual activity in both protozoan parasites (Table 1), suggesting similar possible drug targets in the two closely related species.Overall, higher potency and selectivity was observed in T. gondii compared to P. falciparum.The most significant difference was noted with the peptaibol compounds.For example, efrapeptin F (150003) and efrapeptin G (150004) were found to be 7 and 37 times more potent in T. gondii, respectively (Table 1).These two peptaibols also showed no toxicity to human foreskin fibroblasts (HFFs) at the highest concentra tions tested (20 µM).Peptaibols are membrane-active polypeptides characterized by the FIG 1 The workflow of phenotypic screen of the pure compound library against Toxoplasma gondii tachyzoites (RH88).One point screening at 5 µM yields 91 hits that have activity similar or better than artemisinin at the same concentration.Forty-nine compounds were selected for potency determination and cytotoxicity measurement.Four novel scaffolds were identified in addition to the reported cyclic tetrapeptide analogs.This figure was created with BioRender.com.f presence of a non-ribosomally synthesized amino acid, α-aminoisobutyric acid, and a Cterminal hydroxylated amino acid (Fig. 3).To date, no studies have reported the inhibi tory activity of peptaibols in T. gondii, although they have been implicated against the malaria parasites (8).These two 16-amino-acid peptaibols were procured from Tolypocla dium sp. and were originally shown to have insecticidal and antimicrobial activity (12).Whether fungus-derived peptaibols from other compound groups have activity in T. gondii is still unknown.These findings highlight the promise of this unique compound class for peptaibol-driven anti-T.gondii drug discovery.

Name
The second noteworthy scaffold group was the xanthoquinodin analogs.These compounds demonstrated broad spectrum activity against numerous human pathogens including Mycoplasma genitalium, P. falciparum, Cryptosporidium parvum, Trichomonas vaginalis, and others (13).As seen with the previous antimalarial screening, xanthoqui nodin A1 (130144) and A2 (130147) demonstrated the greatest activity with T. gondii IC 50 values of 111.7 and 56.4 nM, respectively.Dissimilar to the results in Plasmodium, xanthoquinodin A2 showed slightly higher potency and selectivity than the A1 analog.Additionally, both xanthoquinodin A1 and A2 were found to be 2 and 12 times more potent in T. gondii as compared to P. falciparum, respectively.This, again, highlights the potential of these compounds for use in T. gondii drug discovery.
A third scaffold group identified was the cyclic tetrapeptides, including 1-alaninech lamydocin (130134) that showed a T. gondii IC 50 of 28.7 nM and a SI of greater than 697.Other members of this group included apicidin B (120250) and C (120251) that demonstrated IC 50 values of 63.2 and 25.9 nM against T. gondii, respectively.Cyclic tetrapeptides such as apicidin have been shown to act as histone deacetylase (HDAC) inhibitors in Plasmodium (10).These compounds in particular are suspected to inhibit the plasmodial class I HDAC (PfHDAC1) (13), causing rapid hyperacetylation of histone H4 in vitro using P. falciparum.The anti-T.gondii activity of these apicidin analogs was also previously reported ( 14).One analog (FR235222) demonstrated success in inhibiting not only tachyzoites but also bradyzoites in cysts (15).Additional studies later identified the drug target of FR235222 as T. gondii HDAC3 (TgHDAC3) (16).While high cytotoxicity of these compounds was indicated in other cell lines (e.g., HepG2), we observed low cytotoxicity in HFFs.
Several verticillin analogs were also identified with T. gondii potency.Verticillins are a group of compounds made up of fungal 1,4-bridged epiplythiodioxopiperazines (ETPs).Their structures consist of two ETP monomers connected through a C-C bond joining two quaternary carbons.Although the members of this group had low IC 50 values against T. gondii, they were highly cytotoxic, and therefore possessed poor selectivity.Leptosin A is the exception with an IC 50 of 12.1 nM and SI of 45.Verticillins have demonstrated activity in a wide range of microbes including bacteria and nematodes and were found to be cytotoxic to many cancer cell lines including HeLa cells (17).The activity of this group is attributed to their sulfide bridge that targets proteins via a reaction with thiol groups, leading to toxicity through the production of reactive oxygen species (17).In previous work, verticillin A has shown to target histone methyltransfera ses, which are epigenetic regulators like HDACs (9).
Other notable hits include the heptelidic acid analog 130131, which demonstrated high potency against T. gondii (IC 50 of 17.6 nM) and good selectivity (SI greater than 1136).In other organisms, heptelidic acid was shown to inhibit GAPDH (glyceraldehyde 3-phosphate dehydrogenase) through the formation of a covalent bond between the thiol group of the catalytic cysteine and the epoxide ring of heptelidic acid (18).Heptelidic acid has also been found to have heme-dependent antimalarial activity and good murine bioavailability in vivo (19).There is limited research regarding the activity of these heptelidic acid analogs in T. gondii, although a large-scale screening at 100 µM found inhibition with the analog chlorohydrin, while the analog hydroheptelidic acid showed no activity (20).
Finally, we identified the fumagillin analog 150046 with an IC 50 of 58.8 nM against T. gondii.Zhang et al. (21) had previously shown that fumagillin and its synthetic analog TNP-470 were active in malaria parasites (21).Fumagillin is known to irreversibly inhibit the human type 2 methionine aminopeptidase (MetAP2) and was explored as an anticancer agent due to its inhibition of angiogenesis.Additionally, in a pull-down assay, fumagillin was found to bind to PfMetAP2, the human ortholog of MetAP2 (11).A clinical trial of TNP-470 as a cancer therapeutic was halted due to its neurotoxicity and short half-life in vivo (22).To reduce the neurotoxicity and pharmacological activities of fumagillin-based analogs, researchers created fumarranol that was found to have activity in P. falciparum Dd2 and 3D7 in vitro and to bind in vivo to PfMetAP2 (11).Beloranib, an analog of fumagillin, was advanced to phase II clinical trials due to its effect on weight loss (23).However, its phase III clinical trial was halted due to patient deaths.Given the high potency in T. gondii of fumagillin analog 150046, and its lack of toxicity at the highest concentration tested (60 µM), this compound group may warrant further study and development.

Conclusion
In conclusion, a screening of natural products primarily derived from fungi identified potent anti-T.gondii scaffolds, many of which show dual activity in P. falciparum.Overall, these inhibitors were more active in T. gondii, highlighting the potential advantages of reutilizing antimalarial screening libraries against other protozoan parasites.Additionally, having the established activity in a closely related species allows for the prioritization of compounds based on known activity and the potential of shared targets.While some of the inhibitors identified were previously known, many had not yet been implicated in T. gondii.This current discovery of potent inhibitors could spearhead T. gondii drug discovery featuring these scaffolds, opening up promising new research avenues in the search for anti-T.gondii therapeutics.

MATERIALS AND METHODS
For the initial fixed concentration screening and IC 50 determination in T. gondii, we utilized a luciferase expressing RH88 strain to quantify the number of live, actively replicating parasites.To construct this strain, the Fluc and DHFR selection markers had been integrated into the UPRT locus through CRISPR/Cas9, rendering the strain pyrimethamine-resistant (6).Screening was performed using a pure compound library provided by the University of Oklahoma Natural Products Discovery Group, containing 683 compounds derived primarily from fungal secondary metabolites.For purposes of testing, HFFs were grown in 96-well microtiter plates in preparation.The media we used to grow cells was referred to a D10 media that was composed of Dulbecco's Modified Eagle Medium (Thermo Fisher; Cat # 11995073), 10% Fetal Bovine Serum (Sigma-Aldrich; Cat # 12306C), 1 × non-essential amino acids (100 ×; Thermo Fisher; Cat # 11140050), HEPES buffer (1M; Fisher Scientific; Cat # SH3023701) at 0.4M, and 10 mg/mL of gentamicin (Fisher Scientific; Cat # 15710072).Compounds were then diluted to 10 µM in 100 µL of fresh D10 media.Each plate utilized artemisinin and DMSO vehicle controls.To each well, 2,500 parasites per 100 µL of D10 media were added.Three technical replicates were then performed.After 48 hours, 160 µL of the D10 media was removed, and plates were subjected to freezing then thawing.Following this, 10 µL of the passive lysis buffer (Promega; Cat# E1910) was added and incubated at room temperature for 15 min.Finally, 100 µL of luciferase assay reagent (Promega; Cat# E1501) was added to each well, and luminescence was measured using a PHERAstar FSX plate reader (BMG Labtech, Germany).Parasite inhibition was calculated based on the controls, with hits defined as those compounds with a mean potency equal to or greater than artemisinin (94% inhibition) in all three wells.
For IC 50 and cytotoxicity determination, we used 384 plates with confluent HFF cells.Compounds were serially diluted 3-fold with a starting concentration of 20 µM.Samples were further diluted by adding an equal volume of media containing 700 parasites.DMSO and 40 µL of KAE609 were used as negative and positive controls, respectively.Assays were performed with a total of three technical replicates and two biological replicates.After incubation for 48 hours, 50 µL of media was removed, and plates were subjected to freezing and thawing.Following this, 10 µL of passive lysis buffer was added and plates were incubated for 15 min at room temperature.Finally, 30 µL of luciferase reagent was added, and the luminescence was measured immediately using a PHERAstar FSX.In parallel, cytotoxicity dose response was performed in the same fashion without the addition of parasites.Cell counting kit-8 reagent (GlpBio; Cat # GK10001) was added at a 1:10 dilution and incubated at 37°C, 5% CO 2 for 1 hour.The absorbance was then measured at 450 nm using a PHERAstar FSX.The data management system CDD vault was used for curve fitting and IC 50 determination.In the same fashion, the IC 50 values of three positive controls (KAE609, artemisinin, and methylene blue) were measured (Table 1) and found to be comparable to published values (7).
To investigate inhibitory patterns within the structural diversity of these 49 hits, a chemoinformatic analysis was performed.The SMILES of each compound were converted to Morgan fingerprints using RDKIT (RDKit: Open-source cheminformatics.https://www.rdkit.org).Compounds were clustered together if they shared at least 80% Tanimoto similarity.Clusters and singletons were spatially arranged using fingerprint distances and visualized with Matplotlib (v3.7).
on next page)

FIG 2
FIG 2 Clustering analysis of selected 49 hit compounds.SMILES for 49 hit compounds were converted using Morgan fingerprint with radius 4 and 2048 bits as determined by RDKIT.Clusters were assigned using Tanimoto similarity threshold of 80%, indicated by light gray circles.T-distributed stochastic neighbor embedding (t-SNE) was used to visualize fingerprint similarities in a scatter plot.Each dot represents a unique hit compound and scaffold categories are indicated when known.IC 50 (nM) was represented by dot size with bigger dot size representing lower IC 50 and vice versa.The node's color intensity indicates the cytotoxicity IC 50 (nM) with the darker the color, the lower the cytotoxicity.Probability values were calculated using the hypergeometric mean function.Clusters all had enrichment values greater than expected by chance.Structures for the notably efficacious compounds were shown.

FIG 3
FIG3 The chemical structures and IC 50 values of the 18 potent compounds.

TABLE 1 T
. gondii and HFF activities of the 49 prioritized natural products (Continued)