Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Mar 21.
Published in final edited form as: Cell Chem Biol. 2018 Dec 27;26(3):411–419.e7. doi: 10.1016/j.chembiol.2018.11.003

Plasmodium PK9 Inhibitors Promote Growth of Liver Stage Parasites

Rene Raphemot 1,#, Amber Leigh Eubanks 1,#, Maria Toro-Moreno 1, Rechel Anne Geiger 1, Philip Floyd Hughes 3, Kuan-Yi Lu 2, Timothy Arthur James Haystead 3, Emily Rose Derbyshire 1,2,4,✉
PMCID: PMC6430656  NIHMSID: NIHMS1512045  PMID: 30595530

Summary

There is a scarcity of pharmacological tools to interrogate protein kinase function in Plasmodium parasites, the causative agent of malaria. Among Plasmodium’s protein kinases, those characterized as atypical represent attractive drug targets as they lack sequence similarity to human proteins. Here, we describe takinib as a small molecule to bind the atypical P. falciparum protein kinase 9 (PfPK9). PfPK9 phosphorylates the Plasmodium E2 ubiquitin-conjugating enzyme PfUBC13, which mediates K63-linkage specific polyubiquitination. Takinib is a potent human TAK1 inhibitor, thus we developed the Plasmodium-selective takinib analog HS220. We demonstrate that takinib and HS220 decrease K63-linked ubiquitination in P. falciparum, suggesting PfPK9 inhibition in cells. Takinib and HS220 induce a unique phenotype where parasite size in hepatocytes increases, yet high compound concentrations decrease the number of parasites. Our studies highlight the role of PK9 in regulating parasite development and the potential of targeting Plasmodium kinases for malaria control.

Graphical Abstract

graphic file with name nihms-1512045-f0001.jpg

eTOC Blurb

Malaria is caused by Plasmodium parasites and remains a global health burden. Few chemical probes exist for protein kinases in the parasite, although they are appealing drug targets. The authors have identified takinib and HS220 as small molecules that target PfPK9 to alter cell signaling pathways and parasite development.

Introduction

Intracellular parasites from the genus Plasmodium are the causal agents of malaria, a life-threatening disease that remains a major public health concern globally. Although many advancements have been made to reduce disease burden, over 445,000 deaths occur annually (WHO, 2017) and this number threatens to grow due to the emergence of drug-resistant parasites to first-line treatments (Ashley et al., 2014; Dondorp et al., 2011; Fairhurst and Dondorp, 2016). Despite the global impact of malaria, many aspects of Plasmodium biology remain elusive, which hinders the development of novel drugs. In other disease systems, chemical probes used to interrogate protein function have been critical for advancing our understanding of disease biology and facilitating drug development.

Protein kinases are important regulators of biological processes in eukaryotes, and thus have been exploited as chemotherapeutic targets in diverse conditions such as cancer, inflammatory diseases, and neurodegenerative disorders (Cohen, 2009; Zhang et al., 2009). The Plasmodium genome encodes 84–99 protein kinases (Anamika et al., 2005; Talevich et al., 2011; Ward et al., 2004) and several studies support their critical role for parasite survival at various stages throughout Plasmodium’s complex life cycle (Alam et al., 2015; Derbyshire et al., 2014; McNamara et al., 2013; Siden-Kiamos et al., 2006; Solyakov et al., 2011; Tewari et al., 2010). Furthermore, research in the last decade has highlighted the potential for kinases to be drug targets in Plasmodium (Doerig et al., 2005; Lucet et al., 2012). However, the paucity of small-molecules targeting Plasmodium protein kinases is a hindrance to probing their roles in cell signaling and regulation, especially for essential proteins that yield lethal phenotypes upon genetic disruption. In addition, Plasmodium is an obligate intracellular parasite, thus small molecules used for cellular studies require selectivity for the pathogen over host targets. Among the predicted Plasmodium kinases, a particularly intriguing group of atypical kinases are present that are phylogenetically divergent from mammalian kinases (Doerig et al., 2008; Talevich et al., 2012; Ward et al., 2004). This sequence divergence presents a unique opportunity to mitigate the challenges of species selectivity. Additionally, we predict that some of these atypical kinases are integral to Plasmodium’s life stages (Dorin-Semblat et al., 2013; Solyakov et al., 2011; Zhang et al., 2018), which includes an obligatory, asymptotic stage in the liver and a cyclical stage in red blood cells that causes disease symptoms.

Plasmodium falciparum protein kinase 9 (PfPK9) is an essential protein (Solyakov et al., 2011; Zhang et al., 2018) that does not cluster with established eukaryotic kinase groups. As a result, PfPK9 is referred to as an orphan kinase among Plasmodium protein kinase families (Solyakov et al., 2011; Ward et al., 2004) and the most similar proteins in P. falciparum have only ~35% sequence identity. To date, the only known target of PfPK9 is the E2 ubiquitin-conjugating enzyme 13 (PfUBC13) (Philip and Haystead, 2007), the homolog of human UBE2N/UBC13 (HsUBC13). HsUBC13 plays a role in DNA repair and immune response pathways by modifying target proteins with K63-linked polyubiquitin chains (Bothos et al., 2003; Brusky et al., 2000; Hodge et al., 2016; Hofmann and Pickart, 1999; Pickart, 2001), but details about PfPK9 mediation of PfUBC13 remain to be determined. The intriguing role of PfPK9 in Plasmodium biology, its low sequence homology to eukaryotic protein kinases, and compelling genetic evidence that it is essential, makes it an ideal candidate for functional studies.

In this study, we characterize the role of PfPK9 in Plasmodium with our discovery that the small molecule takinib binds to the parasite kinase with sub-micromolar affinity. Interestingly, takinib is a potent and selective inhibitor of human TAK1 (HsTAK1) (Totzke et al., 2017), which is regulated by HsUBC13 (Hodge et al., 2016). We demonstrate that takinib decreases K63-linked ubiquitination levels in Plasmodium, supporting the role of PfPK9 as a regulator of PfUBC13 in vivo. To achieve species selectivity, 15 takinib analogs were synthesized and characterized, leading to the identification of HS220 — a molecule that binds PfPK9 but does not inhibit HsTAK1. Takinib and HS220 inhibit Plasmodium parasite load in hepatoma cells through a unique mechanism that coincides with a disruption of controlled parasite growth. Our discovery and use of PfPK9-targeting molecules represent an important step in understanding biological pathways regulated by this kinase and supports future development of Plasmodium protein kinase inhibitors for malaria control.

Results

Target-Based Screen Identifies Takinib as PfPK9-Binding Compound

We established a high-throughput screen to discover small molecules that bind to PfPK9, as no inhibitors of the protein have been previously described. An ATP-competitive binding screen was optimized using GFP-PfPK9 overexpressed in HEK293 cells, which was bound to ATP-sepharose resin. Binding to the ATP-binding site was evaluated by measuring GFP fluorescence after protein elution in the presence of compound. With this approach, we first examined ATP binding to PfPK9 and determined an apparent Kd (Kd(app)) value for ATP of 0.38 mM (Figure S1A). A Z’ factor of 0.6 was achieved using DMSO as the negative control and 200 mM ATP as the positive control. After this validation, 3,218 compounds from a kinase-targeted library were screened at a single concentration of 500 μM to determine whether they could compete with ATP binding (Figure 1A). The initial screen was performed at 500 μM because the affinity resin has a high immobilized ATP concentration of ~10 mM. From this primary screen, 304 compounds that produced a ≥2-fold increase in fluorescence intensity relative to the DMSO control were prioritized. Western blot was used as a secondary assay to detect GFP-PfPK9, which confirmed 14 compounds (0.44% hit rate) able to bind to the kinase (Figure 1B).

Figure 1. Summary of PfPK9 screening results.

Figure 1.

Open in a new tab

(A) Relative PfPK9 binding of compounds screened (grey closed circles) assessed with an ATP-competitive binding assay. GFP-PfPK9 was quantified by relative fluorescence units and signal from compounds was normalized to the DMSO control to obtain relative binding (F/F0). False positives were discerned by Western blot and compounds confirmed via protein detection are shown (black open circles). Threshold for actives was ≥2-fold change relative to DMSO (dashed line).

(B) Schematic of screening strategy shows 3,218 compounds tested for binding to PfPK9. Actives from the fluorescence assay were confirmed by protein detection with Western blot and subsequently tested for activity against liver stage P. berghei parasites. Screening identified five hits that bind to PfPK9 and inhibit Plasmodium parasites where takinib (structure shown) was prioritized.

All 14 screening hits were tested for anti-Plasmodium activity and hepatocyte cytotoxicity. Plasmodium assays were completed utilizing the common liver stage malaria model system, which measures P. berghei parasite load in HepG2 cells via luciferase activity (Derbyshire et al., 2012). PfPK9 is 83% identical to PbPK9 (90% similarity), with particularly high conservation in the ATP-binding pocket, suggesting compounds would bind to both proteins (Figure S1C). When tested at 30 μM, five compounds (1–5) reduced P. berghei parasite load in liver cells by >50% (Figure 2A). HepG2 and HuH7 hepatoma cell viability was assessed to evaluate possible host toxicity and was examined in the same wells as the anti-parasite assays. Similar to other reports, compounds were eliminated for further study if they inhibited liver cell viability ≥50% relative to the DMSO control. When tested at 30 μM, takinib and 2–5 did not decrease HepG2 viability by >50% (Figure S1B). These results suggest that P. berghei inhibition by these compounds is not a consequence of host cytotoxicity. Interestingly, the five compounds with anti-Plasmodium activity represent two different scaffolds, benzimidazoles (takinib, 2–3) and quinolones (4–5). False positives are a common risk with high-throughput screening campaigns, but the presence of structural analogs among our prioritized hits increases our confidence in their selection for further study.

Figure 2. Plasmodium inhibition and PfPK9 binding affinity of screening actives.

Figure 2.

Open in a new tab

(A) Screening actives (1–14 at 30 μM) were tested for inhibition of P. berghei liver stage parasites. Compounds that reduced parasite load >50% (dashed line) were selected as hits. Data are shown as mean ± SEM (n = 6).

(B) Structures of compounds 1–5.

(C) Binding curves of takinib (1) and 2–5. Data are shown as mean ± SEM (n ≥ 2). Inset shows representative takinib curve, mean Kd(app) = 0.46 ± 0.10 μM.

To focus efforts on compounds with cell permeability and anti-Plasmodium activity, takinib and 2–5 were purchased for PfPK9 binding studies. All purchased hits reproduced in the PfPK9 binding assay and were subsequently used to measure binding affinity (Figure 2C). Takinib binding to PfPK9 was determined by converting an acquired EC50 value to a Kd(app) value with a previously reported equation (Cheng and Prusoff, 1973; Haystead, 2006). With this approach, the takinib Kd(app) value of 0.46 μM was determined. Compounds 2–5 did not saturate binding curves when tested up to 500 μM, suggesting Kd(app) values >0.5 μM. Due to low micromolar binding affinity, inhibition of Plasmodium liver-stage parasites, and lack of host cytotoxicity, takinib was prioritized for further functional studies.

Takinib Decreases K63-linked Protein Modification in Plasmodium

Previous work identified P. falciparum UBC13 (PfUBC13) as a native substrate of PfPK9 by tracking phosphate transfer in P. falciparum infected erythrocytes (Philip and Haystead, 2007). PfUBC13 is an E2 ubiquitin-conjugating enzyme and a homolog to Homo sapiens UBE2N/UBC13 (HsUBC13), which is the only E2 known to conjugate K63-linked ubiquitin to substrates. We confirmed that PfUBC13 is a PfPK9 substrate in vitro using a radiolabeled ATP kinase assay with purified proteins. In addition to phosphorylating PfUBC13, we found that PfPK9 is capable of phosphorylating HsUBC13 in vitro (Figure S2). To then explore the binding of takinib to PfPK9 in parasites, we monitored changes in K63-linked protein ubiquitination in P. falciparum infected erythrocytes. The blood-stage Plasmodium model system was selected for this study since there is currently no method to extract Plasmodium parasites from liver cells. Consequently, Plasmodium proteins cannot be deconvoluted from human proteins using commercially available antibodies that detect K63-linked ubiquitin. Blood stage P. falciparum parasites (ring stage) were treated with 30 μM takinib for 24 hrs and K63-linked ubiquitin levels were assessed via Western blot with a K63-linkage specific ubiquitin antibody (anti-K63Ub). We observed that takinib inhibits blood stage P. falciparum parasites at 100 μM (Figure S3C), therefore we incubated parasites with nonlethal concentrations (≤30 μM) for the ubiquitination study. Compound 5 was also evaluated to determine if a PfPK9-binding compound with a different scaffold similarly affects these post-translational modification (PTM) levels. After incubation with takinib or 5, K63-linked protein ubiquitination in P. falciparum decreased as indicated by reductions in several band intensities relative to the DMSO control (Figure 3A). To further validate downstream inhibition of PfUBC13 activity in parasites by takinib, a dose-response study was completed (0.001–100 μM). We observed that levels of K63-linked ubiquitin in P. falciparum decreased in a takinib-dependent manner (Figure 3B, bands A, C, and D). Interestingly, some protein targets (Figure 3B, band B) were not affected by takinib treatment. As a control, K48-linked ubiquitin levels were concurrently evaluated (Figure S3A) since K48-linked ubiquitin is involved in degradation pathways and is unrelated to the K63-linked ubiquitin signaling cascade. The total intensity of all bands was quantified after staining for K63-linked ubiquitin, K48-linked ubiquitin, and actin (Figure S3B). We observed that the K63-linked ubiquitin signal significantly decreased at both 30 and 100 μM, while no change in K48-linked ubiquitin or actin signals were detected.

Figure 3. K63-linked ubiquitination in Plasmodium blood stage parasites.

Figure 3.

Open in a new tab

(A) K63-linked ubiquitin was visualized via Western blot in 30 μM compound-treated and DMSO-treated P. falciparum 3D7-infected erythrocytes. Actin was detected as a loading control.

(B) Western blot of K63-linked ubiquitination levels in P. falciparum 3D7-infected erythrocytes with takinib (0.001–100 μM). Protein bands designated by A, C, and D exhibit dose-dependent reduction in K63-linked ubiquitin while B remains unaffected up to 100 μM based on densitometry. Data are representative of 2–3 independent experiments.

Takinib Alters Plasmodium Development in Liver Cells

To investigate the mode of takinib inhibition of Plasmodium liver stage parasites, dose-response experiments were completed in two hepatocyte cell lines, HepG2 and HuH7. Takinib inhibited P. berghei parasite load in HepG2 and HuH7 cells with EC50 values of 6.7 ± 1.3 and 7.3 ± 0.99 μM, respectively (Figure 4A). To evaluate potential compound cytotoxicity, hepatoma cell viability was assessed under the same experimental conditions as the anti-Plasmodium assays using a fluorescence-based protease activity assay and a luminescence-based assay that detects ATP. Based on these studies, host cell viability does not correlate with parasite load inhibition (Figure 4A). After infection, Plasmodium first migrates through several hepatocytes in a process termed traversal (0–3 hrs) before invasion (Mota et al., 2001). To establish if takinib inhibits early (traversal or invasion) or late (transformation and replication) stages of parasite development, we compared its potency when HepG2 cells were treated at the time of infection or after invasion. From these experiments, we observed no significant change in takinib potency (Figure 4B), suggesting it does not inhibit parasite invasion of host cells.

Figure 4. Takinib inhibition of Plasmodium liver stage parasites.

Figure 4.

Open in a new tab

(A) Takinib-dependent inhibition of P. berghei parasite load in HepG2 (black open circles) and HuH7 (black closed circles) hepatocytes, EC50 = 6.7 and 7.3 μM, respectively. Cell viability did not decrease by more than 40% when tested up to 200 μM, using both hepatocyte lines (red circles). Data are shown as mean ± SEM (n ≥ 4).

(B) Takinib-dependent inhibition of P. berghei parasite load in HepG2 cells following compound addition at 0 or 4 hpi, EC50 = 11 μM for both. Data are shown as mean ± SEM (n = 6).

(C) qRT-PCR analysis of PbPK9 mRNA transcripts throughout infection of HepG2 cells. Data are shown as mean ± SEM (n = 2). Data were normalized to the reference gene Pb18S rRNA at each timepoint and transcript levels shown relative to sporozoites (0 hpi).

To further probe the biological function of PK9 in Plasmodium, the gene expression profile of PbPK9 was measured throughout liver stage development. P. berghei-infected HepG2 and HuH7 cells were collected at various hours post-infection (hpi) and analyzed with qRT-PCR. Freshly dissected sporozoites were used to normalize samples collected at 4, 24, and 48 hpi and each time point was normalized to Pb18S to control for parasite numbers, since a single Plasmodium parasite develops into 10,000–30,000 over the course of liver stage infection (Prudencio et al., 2006). Pb18S exhibited a time dependent increase post-infection, indicating parasite maturation and replication (Figure S4A). We observed that PbPK9 is expressed in sporozoites and throughout infection of HepG2 cells (Figure S5A), with the highest levels observed at 4 hpi (Figure 4C). PbPK9 is also expressed throughout infection of HuH7 cells (Figure S4B).

To investigate the inhibition of liver stage Plasmodium parasites by takinib, immunofluorescence assays were performed. P. berghei-infected HuH7 cells were treated with 10 (~1× EC50) or 30 μM (~3× EC50) takinib for various periods of time, including 0–24 hpi, 24–48 hpi and 0–48 hpi. Infected cells were subsequently labeled with P. berghei HSP70 (PbHSP70) antibody at 48 hpi to visualize the parasites and DAPI to detect nuclei of P. berghei and HuH7 cells. Takinib treatment at 10 μM significantly increased exo-erythrocytic form (EEF) size when compared to DMSO-treated cells at all time periods tested (Figures 5A, B), but the number of EEFs (infected cells) was not affected (Figure 5C). Treatment of P. berghei-infected cells with 30 μM takinib also led to increased EEF size when added for either 0–24 hpi or 24–48 hpi (Figure 5D). Interestingly, the number of EEFs decreased when 30 μM takinib was added for 0–24 hpi or 0–48 hpi (Figure 5E), suggesting events after invasion are influencing the number of parasites.

Figure 5. Takinib causes significant increase in Plasmodium parasite size during liver-stage infection.

Figure 5.

Open in a new tab

(A) Representative immunofluorescence images of P. berghei EEFs in HuH7 cells treated with DMSO, 10 μM or 30 μM takinib at various times post-infection. Cells were stained for PbHsp70 (green) and DAPI (blue). Scale bars, 20 μm.

(B) P. berghei EEF size and (C) relative number of P. berghei-infected HuH7 cells after DMSO or 10 μM takinib treatment at various times post-infection.

(D) P. berghei EEF size and (E) relative number of P. berghei-infected HuH7 cells after DMSO or 30 μM takinib treatment at various times post-infection.

Image quantification completed with ImageJ. Data are shown as mean ± SEM (n = 2). **p<0.01, ***p<0.001, ****p<0.0001, one-way ANOVA with Dunnett’s multiple comparisons test for EEF size and Dunn’s multiple comparisons test for % EEFs.

Plasmodium Inhibition by Takinib Independent of TAK1 Activity

We have previously shown that takinib is a potent inhibitor of HsTAK1 (Totzke et al., 2017) and as a consequence it may bind to PfPK9 or HsTAK1 to inhibit Plasmodium. To discern between these two possibilities, we sought to develop compounds with selectivity for PfPK9 binding over HsTAK1. We synthesized and evaluated a series of 15 takinib analogs for PfPK9 binding using our ATP-competitive binding assay. When screened at 500 and 250 μM (Table S1), seven analogs were found to likely bind to PfPK9 based on relative fluorescence intensity. PfPK9 binding was further confirmed for six of these compounds with Western blot (Table S1). Of the validated PfPK9-binding analogs, HS220 and HS230 (Figure 6A) have been previously shown to have little or no effect on HsTAK1 activity (Totzke et al., 2017). Due to the poor solubility of HS230, HS220 (Figure 6B) was selected for further analysis.

Figure 6. PfPK9 selective Takinib analog HS220 inhibits Plasmodium liver stage parasites.

Figure 6.

Open in a new tab

(A) Relative PfPK9 binding of takinib analogs. From 15 analogs, HS220 and HS230 both bind PfPK9 >3-fold compared to the negative DMSO control, but displayed little or no inhibition of HsTAK1 (Totzke et al., 2017).

(B) Structure of HS220.

(C) HS220 inhibition of HsTAK1 (Totzke et al., 2017) and PfPK9 binding affinity relative to takinib. Fold change indicated above bars.

(D) HS220-dependent inhibition of P. berghei parasite load in HuH7 cells (black circles), EC50 = 43 ± 3.8 μM. Cell viability did not decrease more than 30% at any concentration tested (red circles). Representative curve shown, data shown as mean ± SEM (n = 3).

(E) P. berghei EEF size and (F) relative number of P. berghei-infected HuH7 cells after DMSO or HS220 treatment at 0 hpi. Analyses were conducted at 48 hpi. Data shown as mean ± SEM (n = 3). *p<0.05, **p<0.01, one-way ANOVA with Dunnett’s multiple comparisons test for EEF size and Dunn’s multiple comparisons test for % EEFs.

Using our competition binding assay, the HS220 Kd(app) of 4.1 ± 0.80 μM was determined for PfPK9 binding (Figure S5A). In previous activity assays, HsTAK1 inhibition was dramatically decreased (160,000-fold less potent) with the change from a primary benzamide (takinib) to a carboxylic acid (HS220) (Totzke et al., 2017). In contrast, only a minor change in PfPK9 binding affinity (8.9-fold less potent) is observed between HS220 and takinib. To probe takinib/HS220 selectivity in the context of the Plasmodium kinome, their affinity to two unrelated Plasmodium kinases, P. falciparum protein kinase 5 (PfPK5, a cyclin-dependent protein kinase) and P. falciparum calcium‐dependent protein kinase 1 (PfCDPK1), was measured. Takinib and HS220 had no affinity for either PfPK5 or PfCDPK1 when tested up to 30 μM (Kd>30 μM, KINOMEScan). This indicates that they do not generally bind to all Plasmodium kinases. The selectivity of HS220 to PfPK9 over HsTAK1 (Figure 6C) supports its use as a chemical probe in cell-based studies.

The anti-Plasmodium activity of HS220 was compared to that of takinib. Similar to takinib, HS220 inhibits P. berghei parasite load in HuH7 cells with an EC50 of 43 ± 3.8 μM (Figure 6D). Importantly, no significant change on HuH7 viability was observed upon HS220 treatment. Because of its reduced HsTAK1 affinity, the inhibitory activity of HS220 in parasites supports an essential role for PK9 in Plasmodium, in agreement with genetic studies demonstrating a lethal phenotype after gene disruption (Solyakov et al., 2011; Zhang et al., 2018). Immunofluorescence studies were also conducted to examine if PK9 inhibition influences parasite size. P. berghei-infected HuH7 cells were treated with 30 (~1× EC50) or 100 μM (~3× EC50) HS220 from 0–48 hpi and visualized at 48 hpi. HS220 treatment at 30 μM significantly increased P. berghei EEF size when compared to DMSO-treated cells (Figure 6E), but the number of EEFs did not change (Figure 6F). At 100 μM HS220 P. berghei EEF size was not affected (Figure 6E), but the number of EEFs decreased (Figure 6F). These observations are consistent with the effects of takinib when administered at concentrations above the EC50 value (Figures 5D, E). Thus, liver stage Plasmodium phenotypes are conserved between treatment with takinib and HS220, despite the fact that HS220 does not inhibit HsTAK1. Furthermore, we examined HS220 for the ability to modulate K63-linked ubiquitin levels in P. falciparum-infected erythrocytes. When parasites were incubated with 30 μM HS220, we observed a decrease in anti-K63Ub Western band intensities (Figure S5B) similar to the signal reduction seen with takinib and 5 treatment (Figure 3).

Discussion

Protein kinases in Plasmodium parasites likely play essential roles in cell signaling and homeostasis, yet the precise functions of many of these proteins remain unknown. Among the predicted eukaryotic-like P. falciparum protein kinases, those termed atypical are particularly intriguing due to their low sequence homology to known human protein kinases (Doerig et al., 2005; Solyakov et al., 2011; Ward et al., 2004). PfPK9 is one such kinase, and while many details of its function remain to be determined, it is known to be essential for parasite development during the blood stage (Solyakov et al., 2011; Zhang et al., 2018), and has one identified substrate, the E2 ubiquitin-conjugating enzyme PfUBC13 (Philip and Haystead, 2007). In humans, HsUBC13 is the only known E2 to conjugate K63-linkage specific polyubiquitin onto protein substrates and does so in complex with UBE2v1 or UBE2v2 (Hodge et al., 2016; Sato et al., 2012), which regulates signaling for DNA repair, JNK mediated stress responses, and NFκB/MAPK immune signaling. Despite the importance of PfPK9 for parasite survival and its potential role in ubiquitin signaling, no pharmacological modulators of this kinase were known to probe its physiological functions.

In this study, we employed a screening strategy to identify small molecules that compete with the PfPK9 ATP-binding site either by directly binding to this site or by allosterically influencing substrate affinity. Our library selection of ATP-like molecules and known kinase inhibitors was designed to increase the probability of obtaining hits. Interestingly, our modest hit rate (0.44%) is unusual for a biased library, suggesting unique structural features exist in PfPK9 when compared to other known ATP-binding proteins. Among the screening hits, takinib and compounds 2–5 were prioritized based on their binding affinity to PfPK9 and activity against liver stage P. berghei parasites.

The lack of structural information on PfPK9 represents a hurdle for studying molecular interactions involved in compound binding, but there is compelling evidence that PfPK9 is functionally similar to HsTAK1. We previously reported takinib as a potent and selective HsTAK1 inhibitor after profiling its activity against 140 human kinases (Totzke et al., 2017). Takinib also inhibits HsIRAK4 (IC50 120 nM) and HsIRAK1 (IC50 390 nM) although less potently than HsTAK1 (IC50 9.5 nM). HsTAK1, HsIRAK4 and HsIRAK1 are members of the Tyrosine Kinase-Like group and are all involved in protein polyubiquitination via HsUBC13 (Bhoj and Chen, 2009; Dunne et al., 2010; Wu et al., 2014). The inhibition of these kinases involved in HsUBC13 signaling and the decrease of PfUBC13 activity by takinib suggests structural homology may exist between HsTAK1 and PfPK9. This is further illustrated by our observation that PfPK9 can phosphorylate both PfUBC13 and HsUBC13. Taken together, these results allude to PfPK9 having a functional role similar to HsTAK1, but this proposal awaits experimental verification. This potential similarity is particularly intriguing in light of the low sequence identity (21%) between HsTAK1 and PfPK9. Plasmodium has an unusually AT-rich genome, which encodes proteins with higher Asn/Lys content compared to eukaryotic proteins. As a consequence, ~40% of Plasmodium genes have unknown functions (Swann et al., 2015). Our work with PfPK9 suggests some of Plasmodium’s unassigned proteins may have counterparts in other eukaryotes that cannot be predicted with current bioinformatic tools.

Importantly, takinib and other identified PfPK9-binding compounds reduced Plasmodium load in liver cells through a mechanism that does not involve host cell death, indicating their potential as probes to interrogate parasite biology. However, hepatocytes express HsTAK1 (Inokuchi et al., 2010), which complicates functional studies on PfPK9 using takinib. Therefore, we identified HS220 as a compound selective for PfPK9 (over HsTAK1) to study its cellular function. Both takinib and HS220 are unable to bind to PfPK5 and PfCDPK1, showing that they do not generally bind to Plasmodium kinases. A dramatic loss of HsTAK1 efficacy was observed upon replacing the primary benzamide (neutral species, takinib) with a carboxylic acid (negative charge, HS220), which could stem from electrostatic repulsion in the tight binding pocket (PDB: 5V5N). It is possible that PfPK9 has a larger binding pocket or a different electronic landscape to accommodate the additional charge on HS220. In support of this proposal, an alignment of PfPK9 with HsTAK1 suggests more residues may be present in the PfPK9 catalytic and hinge regions (Figure S6). Significantly, both takinib and HS220 reduce P. berghei load in hepatoma cells and produce the intriguing phenotype of increasing EEF size when administered at concentrations near their EC50. To the best of our knowledge, no small molecule inhibitor has been previously reported to enhance Plasmodium parasite size. Generally, liver stage Plasmodium inhibitors are characterized by a decrease in EEF size and/or EEF numbers. Thus, our observed increase in EEF size likely represents a unique mechanism of inhibition. This increase in size could indicate uncontrolled parasite growth due to the disruption of PK9-mediated cellular signaling or a role of PK9 in regulating nutrient acquisition (Portugal et al., 2011).

In contrast to the asexual blood stage, the host hepatocyte does not lyse during infection, but rather the parasites use the liver cell membrane to enter the bloodstream. Of note, we observed that EEF numbers decrease at high takinib concentrations. Since invasion is not inhibited by takinib, it suggests that liver cells encasing larger EEFs lyse or are cleared after invasion. During the course of liver stage infection, Plasmodium parasites manipulate signaling pathways within their host hepatocyte to prevent cell death (Kaushansky et al., 2013; van de Sand et al., 2005). Therefore, the significant increase in EEF size caused by takinib and HS220 treatments could disrupt this fine-tuning of host signaling pathways by the parasites, and potentially lead to apoptosis of infected hepatocytes. Additionally, it remains possible that the increased size induced by PK9 inhibitors results in the development of aberrant or immature parasites that are not viable. However, in vivo investigations are necessary to assess this potential phenotype. Currently, it is unclear what downstream PK9 substrate is responsible for the stimulation of parasite size, but we predict that this substrate and other members of the pathway could be drug targets.

We sought to evaluate if takinib can modulate PfPK9 activity in parasites, but few phospho-specific protein antibodies exist for Plasmodium. Since PfUBC13 is the only known PfPK9 substrate, we used its activity as a proxy for detecting kinase inhibition in cells. Takinib, HS220, and 5 each reduce K63-linked polyubiquitination in P. falciparum parasites. Furthermore, we observed that some proteins exhibit a dose-dependent reduction of K63-linked ubiquitin modification with increasing takinib treatment. We strategically conducted these studies using P. falciparum-infected erythrocytes since this host cell has a less complex protein composition compared to hepatocytes and methods exist to separate parasite proteins from erythrocyte proteins. Unfortunately, a similarly conducted experiment cannot currently be achieved in Plasmodium-infected liver cells due to the limitations of the hepatocyte model system. Our data support the proposal that PfPK9 targeting by takinib and HS220 in live blood-stage parasites modulates signaling involving K63-linked ubiquitin conjugation. While our studies do not eliminate the possibility of other binding partners in Plasmodium, the selectivity of takinib within the human kinome, the inability of both takinib and HS220 to bind to PfPK5 and PfCDPK1, and the unique nature of PfPK9 hints at a narrow binding profile. Future studies identifying the polyubiquitinated proteins influenced by takinib treatment will reveal downstream members of this signaling cascade in Plasmodium. However, the identification of these proteins by mass spectrometry is currently hindered by the low level of proteins with this PTM in Plasmodium-infected erythrocytes. Therefore, optimization of enrichment protocols for low cell numbers will be critical for mapping Plasmodium UBC13 substrates.

We observed that some proteins did not display a decrease in K63-linked ubiquitination with takinib treatment, suggesting a PfPK9 independent pathway exists to regulate PfUBC13. In mammalian cells, there are pathways involving HsUBC13 that are HsTAK1 independent as well as targets of HsUBC13 upstream of HsTAK1 (Hodge et al., 2016). Additionally, E3 ligases that act with the E2 conjugating enzyme determine the target specificity for polyubiquitination. In the human host, there are four E3’s that are known to pair with HsUBC13 to catalyze polyubiquitination activity (HLTF, SHPRH, RNF5, and TRIM5). Among these, only TRIM5 has a homolog in P. falciparum based on sequence analysis, but there are likely others that cannot be predicted based on currently available bioinformatic tools.

In summary, we discovered PfPK9 inhibitors and found that they induce an unusual phenotype – enhancement of parasite size in liver cells. While the mechanism of this growth stimulation remains to be elucidated, it signifies a potentially distinct means to target Plasmodium. Among our identified inhibitors, HS220 has the unique ability to bind to PfPK9 while not inhibiting HsTAK1. Thus, HS220 may be leveraged for future studies to elucidate the K63-linked ubiquitin signaling cascade in Plasmodium as well as uncover the role of PK9 in regulating Plasmodium parasite development.

Significance

Kinase inhibitors have the ability to change the current landscape for deciphering signaling pathways in Plasmodium parasites as well as enhance strategies for malaria control. The present work describes the identification of Plasmodium PK9 binding small molecules, including takinib, which modulate parasite development during the liver stage of infection. This study also discovered the PfPK9 selective compound, HS220, which unlike takinib does not have activity against host HsTAK1. HS220 and takinib have the unique ability to increase Plasmodium parasite size in liver cells, though eventually leading to a reduction in parasite numbers. We postulate these compounds act through regulation of a currently unknown parasite signaling pathway involving PfPK9. Our use of takinib and HS220 as an indirect inhibitor of PfUBC13, highlights potential Plasmodium proteins regulated by K63-linked ubiquitin signaling. Thus, these chemical probes may be useful for elucidating key components of Plasmodium PK9 and UBC13 signaling cascades.

STAR Methods

CONTACT FOR REAGENT AND RESOURCE SHARING

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Corresponding Author, Emily Derbyshire (emily.derbyshire@duke.edu).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell Lines

HEK293 (CRL-1573) and HepG2 (HB8065) cells were obtained from the Duke Cell Culture Facility. HuH7 cells were a kind gift from Dr. Peter Sorger (Harvard Medical School). The sex of the cell lines are male for HepG2 and HuH7, female for HEK293. Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Life Technologies) supplemented with 10% FBS (Sigma) and 1% antibiotic-antimycotic (Life Technologies) and maintained in a standard tissue culture incubator (37 °C, 5% CO2). For GFP-PfPK9-HEK293, P. falciparum PK9 (PF3D7_1315100) was cloned into the pEGFP-C1 mammalian expression vector prior to transfection of HEK293 cells and selection with G418 sulfate. Fluorescence activated cell sorting (FACS) was used to enrich for GFP-PfPK9 expressing cells and protein expression was verified by Western analysis with an eGFP antibody (ThermoFisher, CAB4211).

Human Blood, Mosquitoes, and Plasmodium

Whole human blood was obtained (Gulf Coast Regional Blood Center) from confidential donors. P. falciparum 3D7 parasites were obtained through BEI Resources Repository, NIAID, NIH, MRA-102, contributed by Daniel J. Carucci and were cultured in RPMI 1640 medium supplemented with 0.5% (m/v) AlbuMAX II, 25 mM HEPES, 25 ug/mL gentamycin, 24 mM sodium bicarbonate and 50 μg/mL hypoxanthine at a pH of 7.2 and maintained at 37°C under 92% N2, 5% CO2, and 3% O2. Synchronization was performed using 5% sorbitol as previously described (Radfar et al., 2009). P. berghei ANKA infected Anopheles stephensi mosquitoes were purchased from the New York University Langone Medical Center Insectary.

Bacterial Strains

Escherichia coli BL21(DE3) containing plasmids were cultured on Difco™ Luria-Bertani (Miller) agar plates or in Difco™ Luria-Bertani (Miller) broth, supplemented with Ampicillin (200 μg/mL) at 37 °C.

Ethics Statement

Experimental work involving human blood and Plasmodium parasites was approved by the Duke University Institutional Biosafety Committee, registration # 14-0026-01.

METHOD DETAILS

Anti-Plasmodium Liver-Stage Assays

Inhibition of P. berghei parasite load in hepatocytes was evaluated as previously described (Derbyshire et al., 2012). Briefly, 15,000 HepG2 or 8,000 HuH7 cells/well were seeded into 384-well plates in the presence or absence of compounds (0–100 μM) before infection with 3,000 P. berghei ANKA sporozoites. After 45 hrs, liver cell viability was assessed using CellTiter-Fluor (Promega) and parasite load was determined using Bright-Glo (Promega). The relative fluorescence and luminescence signal intensity of each well was normalized to the negative control (1% DMSO). Dose-response analysis was performed with GraphPad Prism.

Gene Expression Analysis

For gene expression of PbPK9 during the liver stage, HepG2 cells were seeded in 12-well plates and infected with P. berghei ANKA sporozoites at 0.3 multiplicity of infection. Total RNA was isolated at 4, 24 and 48 hpi and from 1×105 sporozoites. cDNA was synthesized from total RNA with GoScript™ Reverse Transcription System (Promega). PbPK9 levels were normalized to Pb18S and the relative expression of PbPK9 over time was compared to P. berghei ANKA sporozoites (0 hpi).

Immunofluorescence Analysis of EEFs

GFP-expressing P. berghei-infected HuH7 cells were stained with mouse anti-PbHSP70 (clone 2E6), goat anti-mouse AlexaFluor 488 (Life Technologies) and DAPI. Images were acquired on a Zeiss Axio Observer widefield fluorescence microscope and ImageJ was used to quantify EEF size. For immunofluorescence analyses with compound HS220, 10 fields in each well (triplicates) were randomly acquired using an Arrayscan VTI (Cellomics) with a 20× magnification.

Western Blot of K63-linked Ubiquitination in P. falciparum-Infected Erythrocytes

P. falciparum 3D7 isolate was maintained under standard conditions (Moll, 2013; Radfar et al., 2009). Synchronized ring-stage P. falciparum parasites (10 h post-reinvasion) at 10% parasitemia, 1% hematocrit were treated with either takinib (0.001–100 μM), compound 5 (30 μM), HS220 (30 μM), or ≤ 0.1% DMSO. After 24 hrs, cultures were harvested, treated with 0.03% saponin lysis buffer and then lysed by sonication. Lysates were analyzed by Western blot with K63-linkage specific anti-ubiquitin (Abcam ab179434) and Alexa Fluor 488 goat anti-rabbit (Life Technologies A11008). K48-linkage specific anti-ubiquitin (Abcam ab140601) was detected as a control using secondary Alexa Fluor 647 goat anti-rabbit IgG antibody (Life Technologies, A32733). Anti-actin (Abcam ab3280) was used as a loading control.

Recombinant Protein Expression and Purification

Codon optimized PfPK9-His (BlueHeron) in pET21a, PfUBC13-His6 (amplified from P. falciparum cDNA) in pET21a, and GST-HsUBC13 in pGEX6 (kind gift from Dr. Wei Xiao at the University of Saskatchewan)(McKenna et al., 2001) were expressed in E. coli BL21(DE3). Harvested cells were lysed by sonication, centrifuged and batch-bound to Ni-NTA agarose resin or glutathione sepharose 4B resin. GST-HsUBC13 was cleaved with PreScission Protease. Following size exclusion chromatography, PfPK9-His6 (>95%), PfUBC13-His6 (>98%), and HsUBC13 (>98%) purity was assessed by SDS-PAGE and was identity confirmed with MS/MS.

Takinib Analog Binding Screen

A small-molecule screen was performed similar to a previously report (Haystead, 2006). Briefly, HEK293-GFP-PfPK9 were lysed, bound to ATP-sepharose resin and dispensed into 96-well filter plates stacked on black 96-well catch plates. ATP (200 mM) and takinib (250 or 500 μM)(Totzke et al., 2017) were used as positive controls and DMSO (5%) was the negative control. Compounds were tested at 250 and 500 μM. Dose-response curves (0–500 μM) were generated for select compounds. Fluorescence intensity was measured on an EnVision and data analysis was completed in GraphPad Prism. Determined EC50s were used to calculate the observed PfPK9 Kd(app) as described previously (Haystead, 2006). Western blot with an eGFP antibody (ThermoFisher, CAB4211) was used to validate PfPK9 binding.

Synthesis of Takinib Analogs

The syntheses of analogs HS218 through HS225 have been previously described (Totzke et al., 2017). For the remaining analogs, reagents were obtained from commercial sources and used without further purification. Proton NMR spectra were obtained on Varian 400 and 500 MHz spectrometers. LC/MS were obtained on an Agilent ion-trap LC/MS system.

graphic file with name nihms-1512045-f0002.jpg

HS-230.

3-Carbamoylbenxoic acid (100 mg, 605 μmol), 2-amino-benzothiazole (91 mg, 605 μmol), EDC (174 mg, 908 μmol), HOBT (93 mg, 605 μmol) and DMAP (7 mg, 60 μmol) were combined in methylene chloride (2 mL) and treated with Hunig’s base (78 mg, 605 μmol) and stirred at RT. After a few minutes, DMF (1 mL) was added to aid dissolution. After 4 hr, the reaction was diluted with methanol and the solid precipitate filtered off and air dried to give amide 6 (148 mg, 82%) as a white powder. LC/MS showed a very broad peak with m/z = 298.0, [M+H]+ and 617.1, [2M + Na]+.

graphic file with name nihms-1512045-f0003.jpg

3-Cyanobenzoic acid (600 mg, 4.08 mmol) and 5-bromo-1H-benzo[d]imidazol-2-amine (865 mg, 4.08 mmol) were mixed with EDC (1.56 g, 8.16 mmol), HOBT (551 mg, 4.08 mmol) and DMAP (10 mg, 82 μmol), slurried in DMF (5 mL) and treated with Hunig’s base (527 mg, 4.08 mmol). The mixture was stirred for 16 h and then concentrated. The solid residue was slurried in hot ethanol, cooled and a white powder was filtered off, washed with ethanol and air-dried to give nitrile 7 (980 mg, 70%) as an off-white solid. LC/MS looks marvelous with a single peak with m/z = 341.0 and 343.0, [M+H]+.

graphic file with name nihms-1512045-f0004.jpg

Nitrile 7 (980 mg, 2.97 mmol) was dissolved in THF (18 mL) and DMSO (7.5 mL) and treated with potassium t-butoxide (5.7 mL of 1M THF solution) followed by 1-bromopropane (333 μL, 706 mg, 5.7 mmol) and stirred at RT. After 2 hr, the mixture was treated with a little acetic acid (200 μL) and stirred for an hour. The mixture was then poured into water (100 mL) giving rise to substantial precipitation. The slurry was stirred vigorously overnight then filtered and air-dried to give a mixture of 8a and 8b (838 mg, 76%) as a fluffy white powder. Some of the mixture (400 mg) was adsorbed onto silica and chromatographed (80 g Isco silica gel, 0 to 20% EtOAc in CH2Cl2) to give a partial separation of the two compounds. The clean fractions of the earlier eluting compound were combined to give 8a (called upper bromide, 184 mg). LC/MS shows a single tailing peak with m/z = 383.1 and 385.1, [M+H]+. 1H-NMR (dmso-d6) δ 8.52 (s, 1H), 8.51 (d, J = 8 Hz, 1H), 8.00 (d, J = 8 Hz, 1H), 7.71, (t, J = 8 Hz, 1H), 7.70 (s, 1H), 7.55 (d, J = 8 Hz, 1H), 7.43 (d, J = 8 Hz, 1H), 4.25 (t, J = 7 Hz, 2H), 1.83 (p, J = 7 Hz, 2H), 0.91 (t, J = 7 Hz, 3H). The clean fractions of the later eluting compound were combined to give 8b (called lower bromide, 145 mg). LC/MS shows a single tailing peak with m/z = 383.1 and 385.1, 1H-NMR (dmso-d6) δ 8.52 (s, 1H), 8.51 (d, J = 8 Hz, 1H), 7.99 (d, J = 8 Hz, 1H), 7.87 (s, 1H), 7.71, (t, J = 8 Hz, 1H), 7.47 (d, J = 8 Hz, 1H), 7.40 (d, J = 8 Hz, 1H), 4.25 (t, J = 7 Hz, 2H), 1.83 (p, J = 7 Hz, 2H), 0.93 (t, J = 7 Hz, 3H).

We were unable to obtain adequate NOe spectra to assign the structures. Identification of the isomeric structures was assigned later base on an unambiguous region-controlled synthesis of 8a (see below).

graphic file with name nihms-1512045-f0005.jpg

HS-231.

A mixture of nitriles 8a and 8b (22 mg, 57 μmol) was dissolved in DMSO (350 μL) and diluted with ethanol (500 μL). This was then treated with 50% NaOH (2 drops) followed by 30% hydrogen peroxide (3 drops). Lots of stuff fell out of solution. LC/MS of the crude material gave m/z = 401.1 and 403.1 for [M+H]+. After stirring for 1 hr, the mixture was transferred and concentrated onto silica gel (1 g) and chromatographed (12 g Isco silica gel, 0 to 10% MeOH in CH2Cl2) to give product which was slurried in MeOH to give a white crystalline solid. This was filtered off and air dried to give amides 9a and 9b (15.5 mg, 67%) as a white solid.

graphic file with name nihms-1512045-f0006.jpg

COMU (306 mg, 714 μmol) and 3-cyanobenzoic acid (100 mg, 680 μmol) were dissolved in DMF (1 mL) and treated with Hunig’s base (88 mg, 680 μmol). A lot of stuff fell out. After 5 min, 2-amino-5-bromobenzothiazole (156 mg, 680 μmol) was added along with more DMF (1 mL) and the mixture was stirred for 18 hr. The mixture was treated with a little ethanol (2 mL) and heated to clarity, then diluted with water (20 mL) which caused a lot of white solid to form. This was stirred vigorously for 2 hr then filtered and washed with water and air-dried to give nitrile 10 (243.9 mg, 100+%) as a white solid. LC/MS gave a major peak with m/z = 357.9 and 359.9 for [M+H]+. There were 2 additional minor peaks in the UV trace. The material was used for the next step.

graphic file with name nihms-1512045-f0007.jpg

HS-232.

Nitrile 10 (50 mg, 140 μmol) was dissolved in DMSO (100 μL) and diluted with ethanol (2 mL). This was then treated with 50% NaOH (5 drops). The mixture was stirred for 2 hr, then treated with acetic acid (50 μL), then ethanol (2 ml), then water (2 mL). The mixture was stirred overnight. The next day, the solids were filtered off and air-dried to give the amide product 11 (41 mg, 78%) as a white powder. LC/MS shows a small single peak and a trailing m/z = 376.0 and 378.0 for [M+H]+.

graphic file with name nihms-1512045-f0008.jpg

Nitrile 8a (upper bromide, 100 mg, 261 μmol) was mixed with tri(o-tolyl)phosphine (9.5 mg, 31 μmol) and palladium(II) acetate (6 mg, 26 μmol) in DMF (1 mL) and treated with triethylamine (53 mg, 522 μmol) and ethyl acrylate (52 mg, 522 μmol) and stirred at RT with nitrogen bubbling. The mixture was then heated to 125 °C under nitroge n for 20 hr and allowed to cool. The reaction mixture was diluted with DMSO (1 mL) and heated slightly to re-dissolved the product. The mixture was passed through a filter to remove palladium onto a column and chromatographed (43 g Isco C-18, 0 to 100% MeOH in 0.2% HCO2H) to give 12a (86 mg, 82%) as a white solid. LC/MS gave a single peak with m/z = 403.2 [M+H]+. 1H-NMR (dmso-d6) δ 8.54 (s 1H), 8.52 (d, J = 8 Hz, 1H), 8.00 (d, J = 8 Hz, 1H), 7.78 (s, 1H), 7.73 (d, J = 16 Hz, 1H), 7.72 (t, J = 8 Hz, 1H), 7.69 (d, J = 8 Hz, 1H), 7.62 (d, J = 8 Hz, 1H), 6.53 (d, J = 16 Hz, 1H), 4.28 (t, J = 7 Hz, 2H), 4.20 (q, J = 7 Hz, 2H), 1.84 (hex, J = 7 Hz, 2H), 1.27 (t, J = 7 Hz, 3H), 0.93 (t, J = 7 Hz, 3H).

graphic file with name nihms-1512045-f0009.jpg

HS-234.

Nitrile 12a (150 mg, 373 μmol) was slurried in ethanol (4 mL) and treated with 50% NaOH (12 drops, about 25 mg NaOH/drop, 300 mg) followed by 30% hydrogen peroxide (5 drops). The mixture was stirred at RT for 3 days. The mixture was then acidified by slow addition of 1N HCl to pH = 1 which led to precipitation of the product. The solid was then filtered off and washed with water and air dried overnight to give 13a (141 mg, 96%) as an off white solid. LC/MS gave a single peak with m/z = 393.2 [M+H]+. 1H-NMR (dmso-d6) δ 8.68 (s, 1H), 8.38 (d, J = 8 Hz, 1H), 8.11 (br s, 1H), 8.00 (d, J = 8 Hz, 1H), 7.77 (s, 1H), 7.66 (d, J = 16 Hz, 1H), 7.57–7.63 (m, 2H), 7.55 (t, J = 8 Hz, 1H), 7.42 (br s, 1H), 6.43 (d, J =16 Hz, 1H), 4.27 (t, J = 7 Hz, 2H), 1.85 (hx, J = 7 Hz, 2H), 0.93 (t, J = 7 Hz, 3H).

graphic file with name nihms-1512045-f0010.jpg

Nitrile 8b (100 mg, 261 μmol) was reacted as described above for 8a to give 12b (103 mg, 98%) as a white solid. LC/MS gave a single peak with m/z = 403.3 [M+H]+. 1H-NMR (dmso-d6) δ 8.53 (s 1H), 8.51 (d, J = 8 Hz, 1H), 8.06 (s, 1H), 8.00 (d, J = 8 Hz, 1H), 7.73 (d, J = 16 Hz, 1H), 7.71 (t, J = 8 Hz, 1H), 7.59 (d, J = 8 Hz, 1H), 7.54 (d, J = 8 Hz, 1H), 6.75 (d, J = 16 Hz, 1H), 4.28 (t, J = 7 Hz, 2H), 4.20 (q, J = 7 Hz, 2H), 1.86 (hex, J = 7 Hz, 2H), 1.27 (t, J = 7 Hz, 3H), 0.95 (t, J = 7 Hz, 3H).

graphic file with name nihms-1512045-f0011.jpg

HS-233.

Nitrile 12b (50 mg, 124 μmol) was reacted as described above for 12a to give 13b (103 mg, 98%) as a white solid. LC/MS gave a little peak which trailed out in the MS forever with m/z = 393.2, [M+H]+. NMR gave broad but consistent peaks in DMSO.]+. 1H-NMR (dmso-d6) δ 8.68 (br s, 1H), 8.37 (br d, 1H), 7.91 (br d, 1H), 7.74 (br s, 1H), 7.65 (br d, J = 16 Hz, 1H), 7.42–7.53 (m, 3H), 6.51 (br d, J = 16 Hz, 1H), 4.26 (br?, 2H), 1.86 (m, 2H), 0.94 (br?, 3H).

HS-235 was isolated as a by-product during the synthesis of affinity resins based on takinib

graphic file with name nihms-1512045-f0012.jpg

An earlier eluting peak was concentrated to give ester 14, HS-235 (2.8 mg). LC/MS gave a single peak with m/z = 407.2 for [M+H]+. 1H-NMR (dmso-d6) δ 8.68 (s, 1H), 8.38 (d, J = 8 Hz, 1H), 8.11 (br s, 1H), 8.00 (d, J = 8 Hz, 1H), 7.79 (s, 1H), 7.74 (d, J = 16 Hz, 1H), 7.68 (d, J = 8 Hz, 1H), 7.60 (d, J = 8 Hz, 1H), 7.55 (t, J = 8 Hz, 1H), 7.42 (br s, 1H), 6.53 (d, J = 16 Hz, 1H), 4.27 (t, J = 7 Hz, 2H), 3.74 (s, 3H), 1.85 (hex, J = 7 Hz, 2H), 0.93 (t, J = 7 Hz, 3H).

Independent synthesis of 8a

graphic file with name nihms-1512045-f0013.jpg

graphic file with name nihms-1512045-f0014.jpg

4-bromo-1-fluoro-2-nitrobenzene (1.5 g, 6.8 mmol) was dissolved in ethanol (10 mL) and treated with propyl amine (1.21 g, 1.68 mL, 20.4 mmol) and stirred at RT overnight. The next day, the reaction mixture was concentrated then re-dissolved in ethanol (20 mL) and acetic acid (5 mL). The mixture was treated with zinc powder (4 g) and heated to 80 °C with stirring for 1hr. The reaction was allowed to cool and was filtered through Celite® with an ethanol wash. The eluant was concentrated then chromatographed (Silica gel, 100% CH2Cl2) to give the aniline 15 (1.49 g, 95%) as a clear oil which solidified on standing. LC/MS gave a single peak with m/z = 229.0 for [M+H]+. 1H-NMR (dmso-d6) δ 6.65 (d, J = 2.3 Hz, 1H), 6.57 (dd, J = 2.3, 8 Hz, 1H), 6.29 (d, J = 8 Hz, 1H), 4.83 (br s, 2H), 4.49 (br t, J = 5 Hz, 1H), 2.94 (br q, J = 6 Hz, 2H), 1.58 (hex, J = 7 Hz, 2H), 0.94 (t, J = 7 Hz, 3H).

graphic file with name nihms-1512045-f0015.jpg

Analogous to J. Med. Chem. 2012, 55, 6523−6540. Bromoaniline 15 (440mg, 1.93 μmol) was dissolved in ethanol (4 mL) and treated with fresh cyanogen bromide (306 mg, 2.89 μmol) and stirred at RT. After 1 hr, the mixture was concentrated to give pink solid (728 mg). The pink solid was re-dissolved in ethanol, treated with xs 9/1 : MeOH/NH4OH (2 mL) and concentrated onto silica gel (4 g) and chromatographed (40 g Isco silica, 0 to 10% MeOH in CH2Cl2) to give 2-aminobenzimidazole 16 (421 mg, 86%) as a white solid. LC/MS gave a single peak with m/z = 254.0 for [M+H]+. 1H-NMR (dmso-d6) δ 7.23 (br s, 1H), 7.1 (d, J = 8 Hz, 1H), 6.29 (br d, J = 8 Hz, 1H), 6.59 (br s, 2H), 3.91 (t, J = 7 Hz, 2H), 1.62 (hex, J = 7 Hz, 2H), 0.84 (t, J = 7 Hz, 3H).

graphic file with name nihms-1512045-f0016.jpg

3-Cyanobenzoic acid (116 mg, 788 μmol) and 2-aminobenzimidazole 16 (200 mg, 788 μmol) were mixed with EDC (302 mg, 1.58 mmol), HOBT (106 mg, 788 μmol), DMAP (2 mg, 16 μmol) and slurried in DMF (2 mL) and treated with Hunig’s base (102 mg, 788 μmol). The mixture was gently heated to dissolve everything and stirred at RT overnight. The next day, the reaction mixture was diluted with methanol (20 mL) and stirred as solids precipitated out. The solid was filtered off, washed with methanol, and air-dried overnight to give nitrile 8a (229 mg, 76%) as a white powder. Compound 8a, prepared in this way, is identical to 8a and different from 8b, prepared by separation as described above.

graphic file with name nihms-1512045-f0017.jpg

HS-238.

Acid 13a (40 mg, 102 μmol) was slurried in ethanol (1 mL) and 10% Pd/C (5 mg) in EtOH (1 mL) and put under H2 atmosphere with 3 vacuum flushes. After a week, LC/MS showed some progress. Acetic acid (1 mL) and more catalyst were added and the reaction was mistakenly heated to reflux. LC/MS showed clean formation of product m/z = 395.1 and no sign of starting material. The reaction mixture, which contained substantial precipitate, was stirred under nitrogen overnight, then diluted with DMSO (to dissolve product, 2 mL) and filtered through Celite®, concentrated and chromatographed (50 g Isco C-18, 0.2% formic in water to 100% MeOH) to give product in approx. 4/1 : methanol/water. Crystals formed in the fractions and were filtered off to give the saturated acid 17a (18.2 mg, 45%) as a white solid. LC/MS gave a single peak with m/z = 395.2 for [M+H]+. 1H-NMR (dmso-d6) δ 8.67 (s, 1H), 8.37 (d, J = 8 Hz, 1H), 8.09 (br s, 1H), 7.98 (d, J = 8 Hz, 1H), 7.53 (t, J = 8 Hz, 1H), 7.44 (d, J = 8 Hz, 1H), 7.41 (br s, 1H), 7.40 (s, 1H), 7.14 (d, J = 8Hz, 1H), 4.24 (t, J = 7 Hz, 2H), 2.90 (t, J = 7 Hz, 2H), 2.55 (t, J = 7 Hz, 1H), 1.83 (hx, J = 7 Hz, 2H), 0.92 (t, J = 7 Hz, 3H).

QUANTIFICATION AND STATISTICAL ANALYSIS

GraphPad Prism 7 was used for statistical analysis of data from anti-Plasmodium assays, immunofluorescence, gene expression, blot quantification, and binding assays. Curves were fit using a standard non-linear equation with a variable four-parameter slope. For each experiment n and SEM or SD are reported in the figure legends. ImageJ was used for the quantification of EEF size in immunofluorescence assays.

Supplementary Material

1

Highlights.

  • PfPK9-binding compounds were discovered

  • PfPK9-binding compounds inhibit K63-linked ubiquitination in Plasmodium

  • Takinib and PfPK9-selective HS220 inhibit liver stage Plasmodium

  • Takinib and PfPK9-selective HS220 increase liver stage parasite size

Acknowledgments

We thank Dora Posfai and Dr. Brittany Speer for useful discussions. In addition, we thank Charlotte Farquhar and Kayla Sylvester for tissue culture expertise. We also thank the Derbyshire lab for critical reading of the manuscript. HsUBC13 was received as a kind gift from Dr. Wei Xiao at the University of Saskatchewan. This work was supported by the NIH (GM099796 to E.R.D.), the AAAS Marion Milligan Mason Award (to E.R.D.) and the Ralph E. Powe Junior Faculty Enhancement Award (to E.R.D.). We also thank the NIH (F32AI118294 to R.R.) and the U.S. Department of Education GAANN (P200A150114 to A.L.E.) for fellowship support.

Footnotes

Competing Financial Interests Statement

A patent disclosure describing Takinib and its analogs has been filed with Duke University by ERD and TAJH.

References

  1. Alam MM, Solyakov L, Bottrill AR, Flueck C, Siddiqui FA, Singh S, Mistry S, Viskaduraki M, Lee K, Hopp CS, et al. (2015). Phosphoproteomics reveals malaria parasite Protein Kinase G as a signalling hub regulating egress and invasion. Nature communications 6, 7285. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Srinivasan Anamika N., and Krupa A (2005). A genomic perspective of protein kinases in Plasmodium falciparum. Proteins 58, 180–189. [DOI] [PubMed] [Google Scholar]
  3. Ashley EA, Dhorda M, Fairhurst RM, Amaratunga C, Lim P, Suon S, Sreng S, Anderson JM, Mao S, Sam B, et al. (2014). Spread of Artemisinin Resistance in Plasmodium falciparum Malaria. New England Journal of Medicine 371, 411–423. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bhoj VG, and Chen ZJ (2009). Ubiquitylation in innate and adaptive immunity. Nature 458, 430–437. [DOI] [PubMed] [Google Scholar]
  5. Bothos J, Summers MK, Venere M, Scolnick DM, and Halazonetis TD (2003). The Chfr mitotic checkpoint protein functions with Ubc13-Mms2 to form Lys63-linked polyubiquitin chains. Oncogene 22, 7101–7107. [DOI] [PubMed] [Google Scholar]
  6. Brusky J, Zhu Y, and Xiao W (2000). UBC13, a DNA-damage-inducible gene, is a member of the error-free postreplication repair pathway in Saccharomyces cerevisiae. Current genetics 37, 168–174. [DOI] [PubMed] [Google Scholar]
  7. Cheng Y, and Prusoff WH (1973). Relationship between the inhibition constant (K1) and the concentration of inhibitor which causes 50 per cent inhibition (I50) of an enzymatic reaction. Biochemical pharmacology 22, 3099–3108. [DOI] [PubMed] [Google Scholar]
  8. Cohen P (2009). Targeting protein kinases for the development of anti-inflammatory drugs. Curr Opin Cell Biol 21, 317–324. [DOI] [PubMed] [Google Scholar]
  9. Derbyshire ER, Prudencio M, Mota MM, and Clardy J (2012). Liver-stage malaria parasites vulnerable to diverse chemical scaffolds. Proc Natl Acad Sci U S A 109, 8511–8516. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Derbyshire ER, Zuzarte-Luis V, Magalhaes AD, Kato N, Sanschagrin PC, Wang J, Zhou W, Miduturu CV, Mazitschek R, Sliz P, et al. (2014). Chemical interrogation of the malaria kinome. Chembiochem : a European journal of chemical biology 15, 1920–1930. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Doerig C, Billker O, Haystead T, Sharma P, Tobin AB, and Waters NC (2008). Protein kinases of malaria parasites: an update. Trends Parasitol 24, 570–577. [DOI] [PubMed] [Google Scholar]
  12. Doerig C, Billker O, Pratt D, and Endicott J (2005). Protein kinases as targets for antimalarial intervention: Kinomics, structure-based design, transmission-blockade, and targeting host cell enzymes. Biochim Biophys Acta 1754, 132–150. [DOI] [PubMed] [Google Scholar]
  13. Dondorp AM, Fairhurst RM, Slutsker L, Macarthur JR, Breman JG, Guerin PJ, Wellems TE, Ringwald P, Newman RD, and Plowe CV (2011). The threat of artemisinin-resistant malaria. The New England journal of medicine 365, 1073–1075. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Dorin-Semblat D, Bottrill AR, Solyakov L, Tobin A, and Doerig C (2013). Experimental tools for the study of protein phosphorylation in Plasmodium. Methods in molecular biology (Clifton, NJ) 923, 241–257. [DOI] [PubMed] [Google Scholar]
  15. Dunne A, Carpenter S, Brikos C, Gray P, Strelow A, Wesche H, Morrice N, and O’Neill LA (2010). IRAK1 and IRAK4 promote phosphorylation, ubiquitination, and degradation of MyD88 adaptor-like (Mal). The Journal of biological chemistry 285, 18276–18282. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fairhurst RM, and Dondorp AM (2016). Artemisinin-Resistant Plasmodium falciparum Malaria. Microbiology spectrum 4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Haystead TA (2006). The purinome, a complex mix of drug and toxicity targets. Current topics in medicinal chemistry 6, 1117–1127. [DOI] [PubMed] [Google Scholar]
  18. Hodge CD, Spyracopoulos L, and Glover JN (2016). Ubc13: the Lys63 ubiquitin chain building machine. Oncotarget 7, 64471–64504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hofmann RM, and Pickart CM (1999). Noncanonical MMS2-encoded ubiquitin-conjugating enzyme functions in assembly of novel polyubiquitin chains for DNA repair. Cell 96, 645–653. [DOI] [PubMed] [Google Scholar]
  20. Inokuchi S, Aoyama T, Miura K, Osterreicher CH, Kodama Y, Miyai K, Akira S, Brenner DA, and Seki E (2010). Disruption of TAK1 in hepatocytes causes hepatic injury, inflammation, fibrosis, and carcinogenesis. Proc Natl Acad Sci U S A 107, 844–849. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kaushansky A, Ye AS, Austin LS, Mikolajczak SA, Vaughan AM, Camargo N, Metzger PG, Douglass AN, MacBeath G, and Kappe SH (2013). Suppression of host p53 is critical for Plasmodium liver-stage infection. Cell reports 3, 630–637. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Lucet IS, Tobin A, Drewry D, Wilks AF, and Doerig C (2012). Plasmodium kinases as targets for new-generation antimalarials. Future medicinal chemistry 4, 2295–2310. [DOI] [PubMed] [Google Scholar]
  23. McKenna S, Spyracopoulos L, Moraes T, Pastushok L, Ptak C, Xiao W, and Ellison MJ (2001). Noncovalent interaction between ubiquitin and the human DNA repair protein Mms2 is required for Ubc13-mediated polyubiquitination. The Journal of biological chemistry 276, 40120–40126. [DOI] [PubMed] [Google Scholar]
  24. McNamara CW, Lee MC, Lim CS, Lim SH, Roland J, Nagle A, Simon O, Yeung BK, Chatterjee AK, McCormack SL, et al. (2013). Targeting Plasmodium PI(4)K to eliminate malaria. Nature 504, 248–253. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Moll K, Kaneko A, Scherf A and Wahlgren M (2013). Methods in malaria research, 6th edn (Manassas, Virginia: Glasgow E (ed).). [Google Scholar]
  26. Mota MM, Pradel G, Vanderberg JP, Hafalla JC, Frevert U, Nussenzweig RS, Nussenzweig V, and Rodriguez A (2001). Migration of Plasmodium sporozoites through cells before infection. Science 291, 141–144. [DOI] [PubMed] [Google Scholar]
  27. Philip N, and Haystead TA (2007). Characterization of a UBC13 kinase in Plasmodium falciparum. Proc Natl Acad Sci U S A 104, 7845–7850. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Pickart CM (2001). Ubiquitin enters the new millennium. Molecular cell 8, 499–504. [DOI] [PubMed] [Google Scholar]
  29. Portugal S, Carret C, Recker M, Armitage AE, Goncalves LA, Epiphanio S, Sullivan D, Roy C, Newbold CI, Drakesmith H, et al. (2011). Host-mediated regulation of superinfection in malaria. Nature medicine 17, 732–737. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Prudencio M, Rodriguez A, and Mota MM (2006). The silent path to thousands of merozoites: the Plasmodium liver stage. Nature reviews Microbiology 4, 849–856. [DOI] [PubMed] [Google Scholar]
  31. Radfar A, Mendez D, Moneriz C, Linares M, Marin-Garcia P, Puyet A, Diez A, and Bautista JM (2009). Synchronous culture of Plasmodium falciparum at high parasitemia levels. Nat Protoc 4, 1899–1915. [DOI] [PubMed] [Google Scholar]
  32. Sato Y, Yamagata A, Goto-Ito S, Kubota K, Miyamoto R, Nakada S, and Fukai S (2012). Molecular basis of Lys-63-linked polyubiquitination inhibition by the interaction between human deubiquitinating enzyme OTUB1 and ubiquitin-conjugating enzyme UBC13. The Journal of biological chemistry 287, 25860–25868. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Siden-Kiamos I, Ecker A, Nyback S, Louis C, Sinden RE, and Billker O (2006). Plasmodium berghei calcium-dependent protein kinase 3 is required for ookinete gliding motility and mosquito midgut invasion. Mol Microbiol 60, 1355–1363. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Solyakov L, Halbert J, Alam MM, Semblat JP, Dorin-Semblat D, Reininger L, Bottrill AR, Mistry S, Abdi A, Fennell C, et al. (2011). Global kinomic and phospho-proteomic analyses of the human malaria parasite Plasmodium falciparum. Nature communications 2, 565. [DOI] [PubMed] [Google Scholar]
  35. Swann J, Jamshidi N, Lewis NE, and Winzeler EA (2015). Systems analysis of host-parasite interactions. Wiley interdisciplinary reviews Systems biology and medicine 7, 381–400. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Talevich E, Mirza A, and Kannan N (2011). Structural and evolutionary divergence of eukaryotic protein kinases in Apicomplexa. BMC evolutionary biology 11, 321. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Talevich E, Tobin AB, Kannan N, and Doerig C (2012). An evolutionary perspective on the kinome of malaria parasites. Philosophical transactions of the Royal Society of London SeriesB, Biological sciences 367, 2607–2618. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Tewari R, Straschil U, Bateman A, Bohme U, Cherevach I, Gong P, Pain A, and Billker O (2010). The systematic functional analysis of Plasmodium protein kinases identifies essential regulators of mosquito transmission. Cell Host Microbe 8, 377–387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Totzke J, Gurbani D, Raphemot R, Hughes PF, Bodoor K, Carlson DA, Loiselle DR, Bera AK, Eibschutz LS, Perkins MM, et al. (2017). Takinib, a Selective TAK1 Inhibitor, Broadens the Therapeutic Efficacy of TNF-alpha Inhibition for Cancer and Autoimmune Disease. Cell chemical biology 24, 1029–1039.e1027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. van de Sand C, Horstmann S, Schmidt A, Sturm A, Bolte S, Krueger A, Lutgehetmann M, Pollok JM, Libert C, and Heussler VT (2005). The liver stage of Plasmodium berghei inhibits host cell apoptosis. Mol Microbiol 58, 731–742. [DOI] [PubMed] [Google Scholar]
  41. Ward P, Equinet L, Packer J, and Doerig C (2004). Protein kinases of the human malaria parasite Plasmodium falciparum: the kinome of a divergent eukaryote. BMC Genomics 5, 79. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. WHO (2017). World Malaria Report (World Health Organization, Geneva: ). [Google Scholar]
  43. Wu X, Zhang W, Font-Burgada J, Palmer T, Hamil AS, Biswas SK, Poidinger M, Borcherding N, Xie Q, Ellies LG, et al. (2014). Ubiquitin-conjugating enzyme Ubc13 controls breast cancer metastasis through a TAK1-p38 MAP kinase cascade. Proceedings of the National Academy of Sciences of the United States of America 111, 13870–13875. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Zhang J, Yang PL, and Gray NS (2009). Targeting cancer with small molecule kinase inhibitors. Nature reviews Cancer 9, 28–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Zhang M, Wang C, Otto TD, Oberstaller J, Liao X, Adapa SR, Udenze K, Bronner IF, Casandra D, Mayho M, et al. (2018). Uncovering the essential genes of the human malaria parasite Plasmodium falciparum by saturation mutagenesis. Science 360. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1512045-supplement-1.pdf (922.1KB, pdf)

RESOURCES