Understanding the Antiplasmodial Action of Resistance-Refractory Xanthoquinodin A1

Abstract

Our previous work identified a series of 12 xanthoquinodin analogues and 2 emodin-dianthrones with broad-spectrum activities against Trichomonas vaginalis, Mycoplasma genitalium, Cryptosporidium parvum, and Plasmodium falciparum. Analyses conducted in this study revealed that the most active analogue, xanthoquinodin A1, also inhibits Toxoplasma gondii tachyzoites and the liver stage of Plasmodium berghei, with no cross-resistance to the known antimalarial targets PfACS, PfCARL, PfPI4K, or DHODH. In Plasmodium, inhibition occurs prior to multinucleation and induces parasite death following 12 h of compound exposure. This moderately fast activity has impeded resistance line generation, with xanthoquinodin A1 demonstrating an irresistible phenotype in both T. gondii and P. falciparum.

Graphical Abstract

Malaria caused by the Plasmodium parasite has long been a major global health concern, infecting more than 200 million each year.¹ Conventional antimalarial therapies are facing growing challenges due to the development of resistance, highlighting the urgent need for novel therapeutic approaches.^1–5 In this context, the search for effective, broad-spectrum antimicrobials has gained paramount importance in the battle against this deadly disease. These drugs have the benefit of targeting multiple aspects of the parasite’s biology, which enhances their effectiveness and reduces the risk of drug resistance compared to single target drugs.^6,7 Moreover, versatile inhibitors that have the potential to effectively treat various forms of malaria parasites, including the dormant liver stages of Plasmodium vivax, are desirable in order to simplify treatment regimens and improve patient compliance, particularly in regions with limited healthcare infrastructure.⁸ The cost-effectiveness of developing and distributing a single broad-spectrum drug can also help reduce the overall financial burden of malaria control and treatment programs.^6–8 These drugs have the potential to streamline research and development efforts, accelerate the approval processes, and bolster global health security by mitigating the risk of malaria outbreaks.

In our previous work, we identified a series of 12 xanthoquinodin analogues and 2 emodin-dianthrones with broad-spectrum activity against various pathogens, including Trichomonas vaginalis, Mycoplasma genitalium, Cryptosporidium parvum, and Plasmodium falciparum, with EC₅₀ values ranging from 0.12 to >10 μM, and selectivity indices ranging from 8.6 to >86.⁹ Our new efforts expand upon the most active analogue, xanthoquinodin A1, which demonstrated additional potential against the protozoa Toxoplasma gondii and Plasmodium berghei. We determined that this compound exhibits a high threshold of resistance in T. gondii and P. falciparum, with no cross-resistance observed for known antimalarial targets. As a starting scaffold, xanthoquinodin A1 holds the promise of being an excellent platform for hit-to-lead optimization in the global malaria elimination campaign.

RESULTS

Xanthoquinodin A1 Inhibits Plasmodium Blood and Liver Stages and T. gondii RH88 Tachyzoites with a High Threshold of Resistance.

To further define the range of xanthoquinodin A1’s activity in Apicomplexa, we performed an EC₅₀ determination in P. berghei liver stage parasites and the closely related parasite T. gondii. We found promising inhibition against both parasites, with a mean EC₅₀ of 1.27 μM against P. berghei and 0.12 μM against T. gondii RH88 tachyzoites.¹⁰ In comparison, xanthoquinodin A1 was roughly 3 times more active in T. gondii compared to the EC₅₀ previously determined in the P. falciparum blood stages (Pf Dd2 EC₅₀: 0.29 μM),^9,10 and approximately 4 times less active in the P. berghei liver stage.

To rule out known antiplasmodial drug targets, we next compared the EC₅₀ value of xanthoquinodin A1 between the chloroquine-resistant parental line Dd2 and five additional P. falciparum lines carrying resistance to antimalarial drug candidates. These included a chloroquine-susceptible strain (3D7), a transgenic parasite line with attb-inducible expression of Saccharomyces cerevisiae dihydroorotate dehydrogenase (Dd2 ScDHODH),¹¹ and lines with mutations in acyl-CoA synthetase (PfACS_A597),¹² cyclic amine resistance locus protein (PfCARL_I1139K),¹³ and phosphatidylinositol 4-kinase (PfPI4K_S1320L).¹⁴ No substantial change in parasite EC₅₀s was noted among these lines, with resistance indices (RIs) over the parental line at, or near 1 (Figure 1B), suggesting no overt cross-resistance with these leading drug candidate targets.

Figure 1.

Xanthoquinodin A1 apicomplexan activity and Plasmodium killing profile. (A) Structure of xanthoquinodin A1 and activity against P. falciparum, P. berghei liver stages, and T. gondii. P. falciparum Dd2 activity was previously reported in (PMID: 37276438). (B) EC₅₀ ratios obtained against parasite lines carrying resistance to known antiplasmodial targets: PfACS (control: MMV084978), PfCARL (control: GNF179), and PfPI4K (control: KDU691), along with Dd2/3D7 activity comparison (control: chloroquine). Activity in Pf Dd2 ScDHODH line with and without attb and with and without proguanil (PG). (C) The in vitro parasite killing rate obtained for xanthoquinodin A1 and controls atovaquone (slow acting) and dihydroartemisinin (DHA, fast acting). Lag phase: time needed to achieve the maximum rate of killing; log(PRR): log₁₀ of the parasite reduction ratio, corresponds to the decrease of parasite number over 48 h (excluding the lag phase); 99.9% PCT: time needed to reduce the initial parasite population by 99.9%. (D) Killing effect of xanthoquinodin A1 (10 × EC₅₀) following a 12 h incubation. Graph represents the percentage of viable parasites over time determined by SYBR⁺MTR⁺ gating. Data represent the mean and SEM of three biological replicates.

Multiple attempts were made to generate in vitro resistance to xanthoquinodin A1 in both T. gondii and P. falciparum in order to identify a potential target. Inocula consisting of 1.8 × 10⁸ in P. falciparum Dd2 or 2 × 10⁶ in T. gondii RH88 were used in each passage for continuous culturing in vitro for 8 and 3 months, respectively. Over 8 months, Dd2 parasites were exposed to a gradual increase in the concentration of xanthoquinodin A1 from 40 to 500 nM. The EC₅₀ of xanthoquinodin A1was then re-assessed three times throughout the selection process. However, no substantial EC₅₀ shift (2× or greater) was observed with the xanthoquinodin A1 exposed P. falciparum Dd2 culture at 46, 116, and 243 days following the start of the selection, indicating a lack of resistance development.

Similarly, the T. gondii in vitro evolution of resistance was attempted three times over 3 months, each time ending with 100% death of the parasites with no recrudescent parasites found in 2 weeks. The highest concentrations of xanthoquinodin A1 tolerated were 320, 340, and 350 nM for the three independent attempts. Resistance in P. falciparum was also attempted through a gatekeeper method to establish the minimum inoculum of resistance (MIR).¹⁵ After determining the EC₅₀ and EC₉₀ in a P. falciparum Dd2-B2 clone (181.6 and 429.4 nM, respectively), selection proceeded at 3 × EC₉₀ for over 60 days. During this time, no recrudescence was seen, indicating an MIR > 3 × 10⁷. This corroborates the “irresistible” property of xanthoquinodin A1 seen in P. falciparum, a desirable quality described by the medicines for malaria venture (MMV) to define inhibitors that failed to yield resistance.¹⁶

Xanthoquinodin A1 Shows a Fast-Killing Profile.

To examine the in vitro killing rate of xanthoquinodin A1, we used a previously established method¹⁵ to assess at the parasite reduction ratio (PRR). Synchronous ring-stage cultures containing 10⁶ 3D7 parasites/ml were treated with 10 × EC₅₀ of compound and incubated for 24, 48, 72, 96, or 120 h, prior to compound wash-out. Cultures were then serially diluted in uninfected RBCs. Parasites were allowed to grow for 21 days, after which growth was assessed using a SYBR green I assay. Dihydroartemisinin (DHA) was used as a fast-acting control and atovaquone as a slow-acting control (Figure 1C). Compared to these controls, the killing rate parameters determined for xanthoquinodin A1 rank it as a moderately fast-acting compound, with a lag phase similar to pyrimethamine.¹⁷ It exerts its killing effect after 24 h in the second part of the asexual cycle. After the lag phase, xanthoquinodin A1 eliminates 10^4.06 (11,481) parasites in 48 h, a value comparable to DHA, although the latter has no lag phase. Xanthoquinodin A1 also needed 47.2 h of incubation to eliminate 99.9% of the parasites, corresponding to one full asexual replication cycle.

Given the irresistible profile seen in the attempts to generate resistance in Dd2 lines, we performed additional profiling using Dd2 and a shorter incubation interval. Semisynchronous Dd2 parasites primarily at the ring stage of the asexual life cell cycle were incubated with 10 × EC₅₀ of xanthoquinodin A1, DHA, or a vehicle control for 12, 24, or 48 h. The parasitemia was then determined using flow cytometric staining with SYBR Green I, and the relative viability of parasites was assessed using Mitotracker Deep Red FM. After 12 h, xanthoquinodin A1-treated cultures showed a minor increase in parasitemia (<1% increase), and an immediate increase in the number of nonviable parasites (Figures 1D and S1A). Parasitemia began to decrease after the first collection time (48 h) and continued to decline throughout the remainder of the experiment. Similarly, the fraction of nonviable parasites remained high, suggesting that while parasites did not initially die when exposed to xanthoquinodin A1, a 12 h incubation was sufficient to cause a lasting effect. Comparable results were seen with the 24 and 48 h incubation times (Figure S1B,C), further supporting this finding.

Optimal Blood Stage Inhibition of Xanthoquinodin A1 Occurs Prior to 30 HPI.

To better define the stage-specific activity (SSA) of xanthoquinodin A1, its inhibitory effects were evaluated at several key stages of the intraerythrocytic life cycle. To that end, Dd2 culture was synchronized, and compound was added at 6, 18, 30, or 42 h postinvasion (HPI). Cultures were observed until the second life cycle with samples taken every 12 h following treatment for flow cytometric analysis and Giemsa-stained thin blood smears. Early stage compound addition (6 and 18 HPI) resulted in the complete inhibition of life cycle progression by 42 HPI (Figure 2A,B). Compared with the control, no initial distinction was noted at 18 HPI between xanthoquinodin A1 and vehicle-treated parasites. At 30 HPI, the 6 and 18 HPI-treated parasites appeared morphologically similar to the control, but interestingly showed a significant increase in the total fluorescence signal over the control (p = 0.0058 for the 6 HPI addition and p = 0.0001 for the 18 HPI addition) (Figure 2C). To explore this finding, we first confirmed that xanthoquinodin A1 by itself, and xanthoquinodin A1 alone with YOYO-1 had no effect on fluorescence signal. No changes to total fluorescence signal was seen in either condition. While YOYO-1 is a nucleic acid binding dye, the samples in this assay were treated with RNase to eliminate signal from RNA. Therefore, these results could be caused by other factors induced by xanthoquinodin A1 such as increased DNA synthesis, or increased YOYO-1 binding to DNA, possibly due to DNA conformational changes.

Figure 2.

The intraerythrocytic stage specificity of xanthoquinodin A1 shows that maximum inhibition occurs prior to 30 HPI. Samples were collected for thin blood smears and flow cytometry with YOYO-1 every 12 h after a 5 × EC₅₀ treatment at 6, 18, 30, and 42 HPI. (A) Histogram plots of the control and treated cultures. Results are representative of 3 biological replicates. (B) Giemsa staining of the control and treated cultures. (C) Total fluorescence signal across all 3 replicates. (D) Median fluorescence signal across all 3 replicates. Significance was determined using a two-way ANOVA against the vehicle-treated control per each collection point. For the final time point (treatment at 42 HPI), significance determined by paired t test.

By 42 HPI, parasites treated at 6 and 18 HPI showed lighter Giemsa staining overall, and little to no fluorescent shift associated with numerous multinucleated schizonts, as indicated by the significantly lower median fluorescence signal (p = 0.0017 and p = 0.0063, for 6 and 18 HPI treated parasites, respectively; Figure 2D). Compound addition at 30 HPI was only partially effective at blocking a shift to schizogony, with a significantly lower total fluorescence at 42 and 54 HPI (p = 0.0037 and p = 0.0046, respectively), and some reinvaded rings seen at 54 HPI. Compound addition after this time (42 HPI) appeared ineffective at blocking growth or reinvasion. Taken together, this suggests that addition prior to 30 HPI is essential for full life-cycle inhibition by xanthoquinodin A1, with growth arrest most apparent in late schizogony.

Xanthoquinodin A1 Affects mRNA Levels Related to Diverse Processes Including Meiosis, Transport, and Regulation of Phospholipids.

In our efforts to define the in vitro activity of xanthoquinodin A1 we utilized RNASeq of treated P. falciparum culture to explore the compound’s transcriptional impact. Synchronous 3D7 culture in the late trophozoite stage was initially treated for 2 and 4 h with an EC₅₀ concentration of xanthoquinodin A1 prior to collection. Exposure to xanthoquinodin A1 however, greatly reduced the total RNA yield compared to a DMSO vehicle control and additional test compounds. This necessitated performing RNA extraction after 1 h of incubation. It is worth noting that even with this shorter incubation time, roughly 58% less RNA was harvested from the xanthoquinodin A1-treated culture compared to the control (Figure S2A). There appeared to be no change to the RNA content or quality however, with a similar rRNA ratio (28s/18s) and RNA integrity number (RIN) seen between compound and vehicle-treated samples during the RNA quality check (Figure S2B,C).

Following sample preparation, differential expression was determined for 3 biological replicates. To increase fidelity, mRNA results were sorted at the gene (all isoforms grouped as one) and transcript levels (each isoform weighed separately) and the results compared. With the gene-level analysis, we identified 75 significant (p < 0.1) differentially expressed (log2(fc) > |1|) mRNA compared to a DMSO vehicle control (Figure 3A). Data were then grouped into six gene sets via hierarchical clustering. Cluster 1 with downregulated results showed GO-term enrichment in terms related to RNA export from the nucleus and the chromosomal region of the cell. Cluster 2, which also contained downregulated results, included the phosphatidylinositol phosphate (PIP) regulators phosphatidylinositol 3-kinase (PI3K) and phosphoinositide-binding protein 2 (PH2), and was enriched for terms related to protein folding, chromosome segregation, PIP biosynthesis, and endocytosis. The third cluster was enriched for kinetochore organization, mRNA metabolism, and chromosome location terms. The final cluster of downregulated genes, cluster 4, showed enrichment in the terms snoRNA localization, DNA endonuclease activity, and schizogony. Two small clusters of upregulated genes were also identified, clusters 5 and 6, which showed enrichment in the GO-terms meiotic spindle organization, NADPH activity, and euchromatin and nucleosome regions.

Figure 3.

RNASeq differential expression of mRNA with the incubation of xanthoquinodin A1. (A) Heatmap of mRNA transcripts (p-value < 0.1, log2 fc > |1|) differentially expressed at the gene level following 1 h exposure to xanthoquinodin A1 (XanQ A1) at EC₅₀ concentration. Results from three biological replicates. Transcripts validated via RT-qPCR are shown in bold and underlined. GO-term enrichment for individual gene clusters is shown to the right of the heatmap. (B) Heatmap of mRNA transcripts (p-value <0.1, log2 fc > |1|) differentially expressed at the transcript level. GO-term enrichment for individual gene clusters shown to the right of heatmap. (C) RT-qPCR validation of differentially expressed transcripts. Data represent the mean and SEM of three biological replicates using housekeeping genes serine tRNA ligase (PF3D7_0717700) and 60S ribosomal protein L18–2 (PF3D7_1341300).

At the transcript level, 40 significant (p < 0.1) differentially expressed (log2(fc) > |1|) mRNA were identified and grouped into four clusters (Figure 3B). Only two downregulated clusters were identified, with enrichment in GO-terms related to endosome to lysosome transport, phosphatidate cytidylyltransferase activity, protein folding, schizogony, and endocytosis. Cluster 3 showed upregulation and enrichment in terms realted to meiotic DNA double-strand break processing, lipoprotein metabolism, and mitochondrial translation termination. This cluster also had additional genes related to phosphatidylinositol phospholipids, including ras-related protein Rab-11B (RAB11b), and ORU-like cysteine protease (OTU), a negative regulator of protein lipidation. Finally, the fourth cluster showed enrichment in terms related to the euchromatin and nucleosome regions and chromosome organization. It is worth noting that 14 mRNA were identified in both the gene- and transcript-level analyses, the most notably being chaperone protein DnaJ (PF3D7_1143200), 60S ribosomal protein L37ae (RPL37A), DNA topoisomerase 6 subunit A (TOP6A), and ATP synthase-associated protein (ATPTG4).

Of the significant (p < 0.1) transcripts identified as differentially expressed (log2(fc) > |1|) via the RNASeq analysis, 9 were chosen for further validation via RT-qPCR. The acquired C_t values were normalized to two housekeeping genes, serine tRNA ligase (PF3D7_0717700) and 60S ribosomal protein L18–2 (PF3D7–1341300). All of the transcripts examined showed the same directional shift as the RNASeq results, and mean fold changes (FCs) ranging from −5.12 to 7.36 (Figure 3C). The greatest fold changes seen were the in the upregulation of RAB11b (mean FC: 7.17) and the downregulation of protein disulfide isomerase (PF3D7_1127500, mean FC: −5.17).

Xanthoquinodin A1 Impacts the Expression of lncRNA Predicted to Target Phospholipid Regulation, RNA, and Kinetochore Proteins.

To expand upon our transcriptomic analysis of xanthoquinodin A1 treated culture, we also examined the changes in expression among long noncoding RNAs (lncRNAs), which are known to be heavily involved in chromatin regulation.¹⁸ Twenty-eight lncRNA transcripts were identified as showing significant (p < 0.1) differential expression (log2(fc) > |1|), which were grouped into six subsets through hierarchical clustering. We then used Perl scripts to predict the cis targets 100,000 bp upstream and downstream of these lncRNAs. In total, 30 predicted cis targets were identified for 17 of the differentially expressed lncRNA (Figure 4A). GO-term analysis of these targets showed enrichment primarily in RNA regulation terms but also terms related to phospholipase binding, protein targeting, the kinetochore and chromosomal regions, and the cell cycle. Notable cis-targets include the RAC-beta serine/threonine protein kinase (PKB), kinetochore protein SPC25, and the signal recognition particle receptor subunit beta (SRPRB). While the majority of these targets showed some change in the transcript or gene results, only one of them was identified as significant (p < 0.1), likely due to the limited compound exposure time. This target was PF3D7_02188800, or ribonuclease P (RNase P), which was upregulated at the transcript level. Interestingly, RNase P has the distinction of being the only ribozyme RNase, and is ubiquitous across prokaryotes and eukaryotes.¹⁹ As with the mRNA results, we validated the differential expression seen in the RNASeq results via RT-qPCR of two of the lncRNA transcripts. Both the transcripts we assessed, MSTRG.3393.3 and MSTRG.2751.2, showed directional changes (up- or downregulation), comparable to the RNASeq results (Figure 4B).

Figure 4.

RNASeq differential expression of lncRNA with the incubation of xanthoquinodin A1. (A) Heatmap of lncRNA transcripts (p-value <0.1, log2 fc > |1|) differentially expressed following 1 h exposure to xanthoquinodin A1 (XanQ A1) at EC₅₀ concentration. Results from three biological replicates. Transcripts validated via RT-qPCR shown in bold and underlined. Predicted cis-targets of lncRNA per cluster are shown to the right of the heatmap, along with target-enriched GO-terms. (B) RT-qPCR validation of differentially expressed lncRNA. Data represent the mean and SEM of three biological replicates using housekeeping genes serine tRNA ligase (PF3D7_0717700) and 60S ribosomal protein L18–2 (PF3D7_1341300). Published on NCBI SRA Accession # PRJNA1073928.

STRING Analysis of Co-occurring Transcripts Identifies Key Regulators of Chromosome and Nucleotide Organization.

In order to refine the results of the RNASeq experiment, we chose to focus on transcripts or predicted cis targets that showed some degree of similarity in all of the organisms inhibited by xanthoquinodin A1. To validate the integrity of this approach, we looked at the activity of xanthoquinodin A1 and its respective analogues that had been tested in P. falciparum, T. vaginalis, M. genitalium, and C. parvum in our original communication.⁹ Using a Pearson r two-tailed correlation test with the pEC₅₀ values of these four organisms, we identified a consistent and statistically significant trend between pEC₅₀ values in P. falciparum and T. vaginalis (r = 0.8392, p = 0.0006), P. falciparum and M. genitalium (r = 0.8739, p = 0.0002), and P. falciparum and C. parvum (r = 0.6461, p = 0.0232) (Figure 5A). Given these results, we decided to reassess the RNASeq results obtained in P. falciparum for gene cooccurrence across these organisms, in addition to the newly confirmed species T. gondii and P. berghei. Looking at the differentially expressed transcripts (p < 0.05) identified at the gene and transcript level, in addition to the predicted cis targets of differentially expressed lncRNA (p < 0.05), we located 31 genes that showed similarity in at least five of the six organisms screened (Figure 5B). Of these, only five genes showed some similarity in all six organisms, including the putative AAA ATPase PF3D7_1412700, putative 40S ribosomal protein S5 (PF3D7_0721600), cytidine diphosphate-diacylglycerol synthase (CDS, PF3D7_1409900), GDP-L-fucose synthase (FS, PF3D7_1014000), and PKB (PF3D7_1246900). STRING analysis of these 31 genes showed a significant number of interactions (p = 0.00388) and several areas of functional enrichment. As highlighted in Figure 5C, the greatest enrichment was seen in subcellular localization to condensed nuclear chromosomes and GO-terms related to nucleotide processing functions, such as pyrophosphatase activity, ATPase activity, and nucleoside-triphosphatase activity. There was also an enrichment in protein domains that included a P-loop containing nucleoside triphosphate hydrolase. A full list of all functional enrichment results can be seen in Table S1.

Figure 5.

STRING analysis of the xanthoquinodin A1-shared transcriptome. (A) Correlation graph of the pEC₅₀ activity of xanthoquinodin A1 and its analogues across distinct phyla. Inhibition against these organisms originally communicated in the DOI: 10.1021/acs.jnatprod.3c00283. (B) Gene cooccurrence of differentially expressed transcripts (p < 0.05) or predicted cis targets of differentially expressed lncRNA (p < 0.05) identified following xanthoquinodin A1 incubation in P. falciparum. Transcripts shown are those with genome occurrence pattern similarity in five or more of the six organisms shown to be inhibited by xanthoquinodin A1, as determined by STRING. (C) Interaction network of transcripts with genome similarity. The top six functional enrichments based on STRING strength score are shown by varying node colors. Edge color is reflective of the interaction source.

DISCUSSION

Widespread increase in drug resistance in malaria and other protozoan infections underscores the need for novel drug leads. Our study of xanthoquinodin A1 and its characteristics may help to define this promising scaffold for further development. Attempts to generate in vitro resistance to xanthoquinodin A1 have proven to be unsuccessful, suggesting a potent and potentially irresistible mechanism of action, a desirable property for an antimalarial lead. Xanthoquinodin A1 displays a moderately fast-killing profile, eliminating a significant proportion of parasites within 24 h. While slower than rapid acting inhibitors such as DHA, this speed of action is still fast enough to position it as a promising candidate for malaria therapy.

Given the difficulty in generating resistance in vitro both in P. falciparum and T. gondii, we utilized transcriptomics to further explore antiplasmodial mode of action of xanthoquinodin A1. Transcriptomic analysis of xanthoquinodin A1 exposed cultures revealed significant changes in transcripts and GO-terms related to processes such as RNA trafficking, chromosome segregation, and schizogony. While these results do not point to a definitive target, they do appear to be unique for xanthoquinodin A1 compared to known inhibitor DHA and RNASeq datasets generated to identify potential drug targets.^20–22 RNASeq analysis of two other fungal-derived antiplasmodials tested alongside xanthoquinodin A1 showed comparatively different transcriptome changes. First, the peptaibol harzianin NPDG I, showed an enrichment in transcripts related to proton transport, translation, and miscellaneous metabolic processes.²³ While a second inhibitor (data not yet published) demonstrated transcriptome changes related to endocytosis, secretory vesicles, glycosylation, and others. During our search, novel lncRNAs were identified, with predicted cis targets related to phospholipid regulation, RNA regulation, and kinetochore proteins, results that were also unique to xanthoquinodin A1. These findings suggest that the transcriptional changes following exposure to xanthoqunodin A1 are not a general killing effect.

Despite these findings, identifying a clear mechanism of action remains challenging. However, we can draw conclusions based on the limited consensus about the pathways that may be involved. Future research may achieve greater success by utilizing additional approaches such as CETSA, IVIEWGA, and metabolic profiling.²⁴ In addition to identifying the potential target(s), xanthoquinodin A1’s activity in vivo also needed to be assessed. Nevertheless, this report may serve as a platform to explore the potential of xanthoquinodin A1 as a broad-spectrum antimicrobial and its potential promise in combating drug-resistant malaria.

METHODS

Plasmodium Culture.

Parasites grown in the Chakrabarti Lab were cultured as previously described²⁵ based on the protocol by Trager and Jensen²⁶ using human A+ blood. Culture was maintained at 37 °C with 5% CO₂, 5% O₂, 90% N₂, or 95% air. Parasites grown in the Winzeler Lab were cultured with RPMI 1640 medium (Gibco) supplemented with 0.25% Albumax I (Gibco), 26 mM Sodium bicarbonate (Sigma), 0.1 mM hypoxanthine (Sigma), and 50 μg/L gentamicin (Gibco) and Human O+ blood. Cultures were maintained at 37 °C with 5% CO₂, 3% O₂, and 92% N₂. Parasites grown in the Fidock Lab were grown in human O+ blood with RPMI 1640 with 25 mM HEPES, 50 mg/L hypoxanthine, 2 mM L-glutamine, 0.21% sodium bicarbonate, 0.5% Albumax II (Invitrogen), and 10 μg/mL gentamycin. Dishes were kept in modular chambers (Billups-Rothenberg) at 37 °C and 5% CO₂, 5% O₂, and 90% N₂.

Human Cell Culture.

Cytotoxicity screening was performed as previously described²⁵ using the MTS-cell viability assay. In brief, HepG2 cells were seeded 24 h prior to compound addition and then incubated with compound for 48 h. MTS solution was then added to assay wells and incubated at 37 °C for 3 h prior to an absorbance reading at 490 nm on a Synergy Neo2 multimode reader (BioTek, Winsooki, VT).

EC₅₀ Determination in T. gondii.

Dose response assays were carried out in 384-well plates as described recently.¹⁰ This assay utilized a luciferase expressing RH88 strain modified via the CRISPR/CAS9 system in which the luciferase expressing gene along with DHFR were inserted into the UPRT (uracil phosphoribosyl transferase) gene.²⁷ Due to the integration of the DHFR selection marker, this strain is resistant to pyrimethamine. The tachyzoite form of RH88 was grown in human foreskin fibroblasts (HFFs) in DMEM with sodium pyruvate and 10% FBS. In a 384-well plate, HFFs and RH88-luc were incubated with serially diluted xanthoquinodin A1 or controls (40 μM KAE609 and 0.2% DMSO) for 2 days. Host cells and parasites were incubated with 5× passive lysis buffer (Promega; Cat. No. E1910) at room temperature for 15 min, followed by the addition of a luciferase reagent (Promega, Cat. No. E1501) was added into each well, and luminescence was immediately read on PHERAstar FSX (BMG Labtech, Germany). The EC₅₀ curve was fitted in the CDD vault.

P. falciparum Blood Stage Screening.

Screening of xanthoquinodin A1 by the Chakrabarti Lab was conducted as previously described²⁵ based on protocols by Smilkstein et al.²⁸ In brief, an asynchronous parasite culture was incubated for 3 days with serial dilutions of compound or control (0.2% DMSO and 10 μM CQ). Plates were then frozen overnight, thawed, and incubated for 45 min −1 h with 1× SYBR Green I in lysis buffer (20 mM Tris–HCL, 0.08% saponin, 5 mM EDTA, and 0.8% Triton X-100). Fluorescence was read at 485 nm excitation and 530 nm emission on a Synergy Neo2 multimode reader (BioTek Winsooki, VT), and dose–response curves were generated with CDD Vault.

Screening in the Winzeler Lab involved incubation of the compound for 3 days with culture at 0.3% parasitemia, 2.5% hematocrit, dispensed into 1536-well black, clear bottom plates with prespotted compound using a MultiFloTM Microplate dispenser (BioTek). Plates were then incubated with 10× SYBR Green I (Invitrogen) in Lysis buffer (20 mM Tris/HCl, 5 mM EDTA, 0.16% (w/v) saponin, 1.6% (v/v) Triton X) for 24 h and read at 485 nm excitation and 530 nm emission on a EnVision Multilabel Reader (PerkinElmer). Dose response curves were also generated with the CDD Vault.

P. berghei Liver Stage Screening.

Liver stage screening was performed as previously described²⁹ using HepG2-A16-CD81 cells infected with P. berghei sporozoites (P. berghei ANKA GFP-Luc-SMcon) obtained by dissecting salivary glands of infected A. stephensi mosquitoes.

In Vitro Evolution of Resistance in P. falciparum and T. gondii.

T. gondii RH88-luc was cultured in HFFs in DMEM with 10% FBS in T25 flasks. In vitro evolution was carried out with an inoculum of 5 × 10⁶ T. gondii RH88 for each passage. Parasite growth was monitored daily. Once 80% of host cells were lysed, parasites were scratched off from the bottom of the T25, pressed through a 22 1/2 needle two to three times, and passed onto a fresh, confluent layer of HFFs. A stepwise increase in concentration of xanthoquinodin A1 was applied to induce resistance. Over the course of 1 month, the drug concentration increased from the initial 20 nM to approximately 320 nM. At the highest concentration, 100% death of the parasites was observed with no recrudescent parasites found 2 weeks later with drug pressure removed. Over the next two months, this process was repeated two more times, each time ending with complete death of the parasites.

For the in vitro evolution in P. falciparum, Dd2 parasites were cultured in human O-positive red blood cells at 1.5% hematocrit in T75 flasks with 40 mL of RPMI 1640 medium as described above. A control flask of parasites where no drug was added and three flasks of parasites with drug treatment as three biological replicates were set up. Parasites were passaged every 2 or 3 days. Each time, parasitemia was measured via microscopy and blood smear. Decisions regarding drug addition and concentration were made depending on the parasitemia. Over eight months, parasites were exposed to a gradual increase in the concentration of xanthoquinodin A1 from 40 to 500 nM over eight months. The IC₅₀s of xanthoquinodin A1 in the evolved parasites were measured three times (46, 116, and 243 days after the start of the selection) over the course of the selection to monitor resistance development. EC₅₀s were measured using flow cytometry at 1% parasitemia, 0.2% hematocrit in V-shaped 96-well plates as described.³⁰ Parasites were incubated with serially diluted xanthoquinodin A1 or control (10 μM artemisinin and 0.2% DMSO) for 3 days. After incubation, cells were stained with 20 μL of SYTO 61 (1:10,000 dilution; Thermo Fisher Scientific, Cat. No. S11343) in PBS for 15 min at room temperature, diluted with 180 μL of PBS and subjected to flow cytometry for parasitemia determination using a BD FACSCantoII (BD Biosciences, NJ). Flow cytometry data were analyzed using FlowJo Version 10.4.0. Dose–response curves were fitted with Prism 9 Version 9.1.1.

Minimum Inoculum of Resistance Assay.

To determine drug susceptibility, Dd2-B2 clonal line parasites were synchronized and exposed to xanthoquinodin A1 or a DSM265 control in the ring stage at 0.3% parasitemia and 1% hematocrit. Drug pressure was maintained consistently at 3 × IC₉₀, using starting inocula of 1.4 × 10⁷ and 3 × 10⁷. Cultures were assessed three times per week via blood smear and flow cytometry with SYBR Green and MitoTracker Deep Red FM (Life Technologies) using an iQue flow cytometer (Sartorius, Göttingen, Germany).

Synchronous Stage Specificity Assay.

Assay performed in biological triplicate as previously described³¹ in Dd2 culture. In brief, parasites were synchronized using MACS column³² and sorbitol³³ and plated at 1% parasitemia, 2% hematocrit. After reinvasion, xanthoquinodin A1 (or DMSO vehicle control) was added at 5 × EC₅₀ to different wells at 6, 18, 30, or 42 HPI. Samples were collected every 12 h starting at 6 HPI for flow staying with YOYO-1 and Giemsa thin smears. Gating of flow results was performed using vehicle, uninfected RBCs, and unstained controls using Flowjo version 10. Total and medium fluorescence in the FITC channel was then extracted from triplicate and graphed in GraphPad Prism version 10.

Parasite Reduction Ratio Assay.

This assay was adapted from Sanz et al.¹⁵ Briefly, 10⁶ monoinfected RBCs containing mostly early rings were cultivated separately with 10× the EC₅₀ of xanthoquinodin A1 or reference drugs (Dihydroartemisinin (DHA), Atovaquone (ATVQ)) for 24, 48, 72, 96, and 120 h (0.5% starting parasitemia, 2% hematocrit of P. falciparum 3D7). After each incubation time, cells were washed, and a volume corresponding to 10⁵ infected RBCs in the initial inoculum was serially diluted (1/3) with fresh RBCs. 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 and allows definition of the growth parameters reported in Figure 1C. The experiment was performed with two biological replicates with three technical replicates each.

Additional Killing Profiling.

The 12, 24, and 48 h killing profile assays were performed as previously described³¹ using Dd2 at 1% parasitemia and 4% hematocrit. Semisynchronous culture in ring (>80% of parasites at 2–10 HPI) was incubated with 10 × EC₅₀ of xanthoquinodin A1, DHA, or a DMSO vehicle control for 12, 24, or 48 h prior to compound wash-off. Samples were collected at the time of wash-off and every 24 h after for flow cytometric analysis with SYBR Green I and Mitotracker Deep Red FM. Gating of flow results utilized DHA, vehicle, no mitotracker, no staining, and uninfected RBC controls. Populations were identified as “SYBR positive, mitotracker positive” for viable parasites and “SYBR positive, mitotracker negative” for inviable parasites.

RNA Preparation.

For RNA work, assays were performed in biological triplicates using 3D7 culture synchronized using a combination of MACS column³² and sorbitol.³³ Xanthoquinodin A1 (or the DMSO vehicle control) was added at an EC₅₀ concentration to late trophozoite (~24 HPI) and incubated for 1 h. Culture was maintained ~10% parasitemia, 4% hematocrit prior to collection. RBCs lysis was achieved using a 5 min incubation with 0.1% saponin, followed by centrifugation at 900g and a 1 time wash in RPMI then 2 times wash in PBS. Parasites were then resuspended in a RNase free tube in TRIzol reagent, and RNA extraction was performed using Direct-zol RNA Miniprep Plus (Zymo Research Corporation, Irvine, CA) following the manufacturers protocol. RNA integrity was determined using nanodrop and RNA gel electrophoresis and then subsequently sent for RNASeq analysis or used for cDNA synthesis and RT-qPCR.

RNASeq.

For RNASeq, integrity was confirmed via an Agilent Technologies 2100 Bioanalyzer. Ribosomal RNA was then depleted, followed by RNA fragmentation and library preparation using the Illumina TruSeq-stranded-total-RNA-sample preparation protocol. Illumina’s NovaSeq 6000 sequencing system was used for pair-ended sequencing, and an Agilent Technologies 2100 Bioanalyzer High-Sensitivity DNA Chip was used for quality control. Transcripts were assembled using Cutadapt³⁴ with low-quality reads filtered out using LC Sciences in house perl scripts. FastQC³⁵ was then used to check sequence quality. Mapping to the P. falciparum genome was performed with Bowtie2³⁶ and HISAT2.³⁷ Genome assembly was performed using StringTie,³⁸ perl scripts, and gffcompare.³⁹ The transcript expression was determined in FPKM using StringTie³⁸ and edgeR⁴⁰ with the equation:

$$\mathrm{F}\mathrm{P}\mathrm{K}\mathrm{M}=\frac{\text{total}\text{exon}\text{fragments}}{\text{mapped}\text{reads}\text{(millions)}\times \text{exon}\text{length}\left(\mathrm{k}\mathrm{B}\right)}$$ (1)

EdgeR⁴⁰ was then used to perform a parametric F-test comparing nested linear models to determine p-value. Differential expression for mRNA was calculated at the gene level (all transcript isoforms grouped as 1) and the transcript level (all transcript isoforms separate). For RNASeq of lncRNA, transcripts <200 bp or matching known mRNAs were removed. CPC⁴¹ and CNCI⁴² were then used to assign coding potential. Cis targets 100,000 bp up- and downstream of the lncRNA were then predicted using Perl scripts.

Heatmaps for transcripts with p < 0.1 and log2(fc) > |1| were generated and clustered in Instant Clue⁴³ with generated Z-scores graphed in GraphPad Prism version 10. GO-enrichment was determined per each cluster using PlasmoDB and Revigo for semantic grouping and elimination of redundant/outdated terms.⁴⁴ Terms with the lowest p values were added to the figure. Published on NCBI SRA Accession # PRJNA1073928.

RT-qPCR.

For RT-qPCR, cDNA synthesis was performed on 500 ng of total RNA using SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen, Waltham, MA) per manufacturer’s protocol, with 100 ng of random hexamers. Select Master Mix (Life Technologies, Carlsbad, CA) was then used with 400 nM forward and reverse primers for RT-PCR. Primers used are listed in Table 1.

Table 1.

Primers Used in the Study

gene or transcript ID	forward primer	reverse primer
ALBA1	GGCCATTCAGAAGAGGTGG	TTCCACCATATCCGCTACCT
H3.3	AAATCCACAGGAGGAAAGGC	AGCAACAGTTCCTGGACGAT
PF3D7_1309400	GTCGGGCGCAAGATTAAGAA	TGTGTTGAACATTTCACCGT
PI3K	CGCATGATAAACCTGTGGCA	TCGACAGCTGGTAGATTTCGT
PF3D7_1127500	TATGGAGCCATGAGAGTACC	TCATCCTGCACACCAACAAA
PF3D7_0721600	GGTGGACCCAGAGAAGATTCA	CTGCATTTCTGGCACCTGTA
PF3D7_0926400	ACAGCTGTGCCGTATATTGT	GCTACATCACGGTAGGTCAC
RAB11b	GCACGGTCAAAGCTCAGATA	TGCGCTTCTCCTATAATGTGC
MSTRG.3393.3,	AGAATATCCACGCTTACGTT	CGCGTTCATGTAAACAGCCA
MSTRG.2751.2	ACATATACTGACCAGCATGGA	TCACTATCTATTATTAGCTTGCTCT
PF3D7_0717700 (HKG1)	AAGTAGCAGGTCATCGTGGTT	TTCGGCACATTCTTCCATAA
PF3D7_1341300 (HKG2)	ATTATCACATGGCCAATCACC	CAATCTCTTATCATCTGTTATT

For housekeeping genes (HKG), two transcripts were used: serine tRNA ligase (PF3D7_0717700) and 60S ribosomal protein L18–2 (PF3D7_1341300). All RT-qPCR assays were performed with 50 °C for 2 min for UDG attachment, 95 °C for 10 min for DNA polymerase dual lock, then 40× cycles of 95 °C denaturing for 15 s, 55 °C annealing for 30 s, and 60 °C extending for 30 s, followed by melt curve generation to check primer integrity. C_t values were obtained using a QuantStudio 7 Flex qPCR machine (Thermo Fisher, Waltham, MA) with ROX passive reference dye. Normalization to housekeeping genes and DMSO vehicle controls was performed using the 2^ΔΔCt method, with the fold change graphed using GraphPad Prism Version 10.

ABBREVIATIONS

P. falciparum

Plasmodium falciparum

P. berghei

Plasmodium berghei

T. gondii

Toxoplasma gondii

HepG2

human hepatocellular carcinoma cell line

Dd2

multidrug-resistant P. falciparum line

3D7

chloroquine-sensitive P. falciparum line

EC₉₀

50% effective concentration

MTS

3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium

μM

micromolar

SSA

stage-specific activity

HPI

hours post invasion

DMSO

dimethyl sulfoxide

DHA

dihydroartemisinin

CQ

chloroquine

PRR

parasite reduction ratio

PfACS

Plasmodium falciparum acyl-CoA synthetase

PfCARL

Plasmodium falciparum cyclic amine resistance locus protein

PfPI4K

Plasmodium falciparum phosphatidylinositol 4-kinase

ScDHODH

Saccharomyces cerevisiae dihydroorotate dehydrogenase

MMV

medicines for malaria venture

lncRNA

long noncoding RNA

RIN

RNA integrity number

mRNA

messenger RNA

PIP

phosphatidylinositol phosphate

PI3K

phosphatidylinositol 3-kinase

PH2

phosphoinositide-binding protein 2

snoRNA

small nucleolar RNA

RAB-11B

Ras-related protein 11

OTU

ORU-like cysteine protease

RT-qPCR

real-time quantitative polymerase chain reaction

SEM

standard error of means

tRNA

transfer RNA

FC

fold change

BP

base pairs

SRPRB

signal recognition particle receptor subunit beta

RNase P

ribonuclease P

HKG

housekeeping gene