Ethics statement

All procedures performed on mice adhered to the United Kingdom Home Office Animals (Scientific Procedures) Act of 1986 (project license PPL 40/3700) as well as the European Union Animals Directive 2010/63/EU and were approved by Aberystwyth University’s Animal Welfare and Ethical Review Body.

Compound acquisition, storage and handling

All epigenetic probes, epigenetic inhibitors and chemotype matched controls (where available) were received from the SGC and solubilised in DMSO (Fisher Scientific, Loughborough, UK) at stock aliquot concentrations of 1.6 mM and 10 mM for schistosomula and adult worm screening respectively. Auranofin (AUR) and praziquantel (PZQ) were purchased from Sigma-Aldrich and reconstituted similarly to the SGC compounds. All reconsituted compounds were stored at -80°C.

Parasite material

A Puerto Rican strain (NMRI) of S. mansoni was used throughout the study and passaged between Mus musculus (Tuck Ordinary; TO) and Biomphalaria glabrata (NMRI albino and pigmented hybrid [40]) hosts. Cercariae were shed from both B. glabrata strains by exposure to light in an artificially heated room (26°C) for 1 h and used to percutaneously infect M. musculus (200 cercariae/mouse) [41] for generation of adult schistosomes or to mechanically transform into schistosomula [42] for in vitro compound screening. Adult schistosomes were obtained from M. musculus at 7 wks post-infection and used for in vitro compound screening and RNA interference (RNAi).

Bioinformatics

The specific Homo sapiens target (histone reader, writer or eraser) of the SGC-provided epigenetic probes/inhibitors initially was derived from the SGC website and corresponding reference literature [30]. Uniprot IDs of the representative H. sapiens epigenetic target were obtained from Uniprot [43] and their downloaded protein sequences were used as queries for protein BLAST (BLASTp) searches against the predicted protein database derived from the S. mansoni genome hosted in NCBI [44] and Wormbase-Parasite [45] using default settings. For the BLASTp searches, both full-length and catalytic domain peptide sequences were used as queries against the S. mansoni genome (v7.0). The most closely related S. mansoni protein (as compared by E values) and their sequence similarity were reported. Multiple sequence alignments (MSAs) of the amino acid sequence of the catalytic domain amino acid sequences within the identified SmSMYD and human SMYD proteins were performed with MUSCLE v3.8 (Multiple Sequence Comparison by Log Expectation) using the default parameters [46] and selecting ClustalW as the output format [47]. The alignments were then analysed in Jalview v2.9 [48] and visually inspected to check for ambiguities and sequences not aligning correctly. Phylogenetic trees were constructed by MEGA7 using the neighbour-joining method based on the JTT matrix-based model with default settings [49]. A total of 1000 bootstrap replications were run to estimate the confidence of each node.

In vitro schistosomula screening

Mechanically transformed schistosomula phenotype and motility metrics were assessed at 72 h post compound incubation as previously described [50], with minor modifications. Briefly, 384-well tissue culture plates (Perkin Elmer, cat 6007460), containing 20 μl of Basch medium [51], were wet-stamped using a Biomek NX^P liquid handling platform (Beckman Coulter, UK) with negative (0.625% dimethyl sulfoxide, DMSO) and positive controls (AUR and PZQ at 10 μM final concentration in 0.625% DMSO) as well as 37 SGC compounds (at 10 μM final concentration in 0.625% DMSO). Two-fold titrations of 14 SGC compounds, which were consistent hits at 10 μM, were also conducted at 10 μM, 5 μM, 2.5 μM, 1.25 μM and 0.625 μM to generate approximate EC₅₀s using GraphPad Prism. To each pre-loaded well (single concentration or titration screens), a total of 100–120 mechanically transformed schistosomula (in 60 μl) were deposited via a WellMate (Thermo Scientific, UK). Schistosomula/compound co-cultures (80 μl in total) were then incubated at 37°C for 72 h in a humidified atmosphere containing 5% CO₂. At 72 h, tissue culture plates (containing schistosomula/compound co-cultures) were imaged under the same conditions (37°C in a humidified atmosphere containing 5% CO₂) using an ImageXpressXL high content imager (Molecular Devices, UK) with subsequent images processed for phenotype and motility as previously reported [50] and successfully implemented by us at Aberystwyth University [7, 32–34, 52]. Single point schistosomula screens (10 μM) were repeated two or three times whereas dose response titrations were performed once. Zˊ values obtained from all schistosomula screens ranged from 0.27–0.48 (mean = 0.40) for phenotype and 0.41–0.56 (mean = 0.48) for motility.

In vitro adult schistosome screening

The anthelmintic effect that selected SGC compounds (i.e. those 14/37 compounds that demonstrated anti-schistosomula activity at 10 μM) had on adult male and female schistosome pairs as well as egg production was assessed according to the methodology described by Edwards et al. [31]. In vitro screening was replicated three times on different dates to account for any biological variation. Briefly three adult worm pairs/well (48-well tissue culture plate format) were placed into DMEM (Gibco, Paisley, UK) supplemented with 10% (v/v) FCS (Gibco, Paisley, UK), 1% (v/v) L-glutamine (Gibco, Paisley, UK) and an antibiotic mixture (150 Units/ml penicillin and 150 μg/ml streptomycin; Gibco, UK). DMSO (0.5% negative control), AUR (10 μM concentration in 0.5% DMSO; positive control) and two-fold titrations of SGC compounds (50 μM– 6.25 μM or 50 μM– 0.05 μM; in 0.5% DMSO max) were added, and together, these adult worm/compound co-cultures were incubated at 37°C for 72 h in a humidified atmosphere containing 5% CO₂. Following compound incubation, WHO/TDR readouts of adult worm motility [53], abundance of H3K36me2 (LLY-507 and BAY-598 treated worms only) in histone extracts and egg counts/well were recorded at 72 h.

RNA interference (RNAi)

Following the perfusion of 7 wks infected mice, adult worms were recovered and RNAi performed as previously described [20, 21]. Briefly, smp_000700 and non-specific luciferase siRNA duplexes were purchased from Sigma (siSmp_000700 = sense: GGUAAUCGGUCAUGUGUAU[dT][dT] and anti-sense: AUACACAUGACCGAUUACC[dT][dT]; siLuc = sense CUUACGCUGAGUACUUCGA[dT][dT] and anti-sense UCGAAGUACUCAGCGUAAG[dT][dT]) and used at a final concentration of 50 ng/μl. Mixed sex adult worm pairs (for knockdown assessment by quantitative reverse transcription PCR, qRT-PCR) were cultured at 37°C in DMEM supplemented with 10% FCS, 2 mM L-glutamine, 10% v/v HEPES (Sigma-Aldrich, UK), 100 Units/ml penicillin and 100 μg/ml streptomycin in an atmosphere of 5% CO₂ with a 70% media exchange performed every 48 h. The experiment was replicated three times (21 worm pairs/replicate). Quantitative reverse transcription PCR (qRT-PCR) of smp_000700 abundance was performed at 48 h and adult worm motility [53] as well as egg counts were quantified at 168 h. Levels of H3K36me2 detected in schistosome histone extracts were assessed at 72 h.

Quantitative reverse transcription PCR (qRT-PCR)

Following RNAi with siSmp_000700 and siLuc, mixed-sex adult worms were incubated for a total of 48 h before processing them for RNA isolation. Briefly, worms were homogenised using a TissueLyser LT (Qiagen, UK) in TRIzol Reagent (Invitrogen, UK) before isolation of total RNA using the Direct-zol RNA Kit (Zymo, UK). cDNA was then generated using the SensiFAST cDNA synthesis kit (Bioline), qRT-PCR performed and data analysed as previously described [54]. qRT-PCR primers for amplifying smp_000700 (Forward primer 5’-GTCTTGCATGTATAGAGGATTGGTC-3’, Reverse 5’-GCAGTCAACCGATTCAATTAAAGT-3’) and internal standard alpha tubulin (SmAT1; Forward primer 5’-CGAAGCTTGGGCGCGTCTAGAT -3’, Reverse 5’-CTAATACTCTTCACCTTCCCCT -3’) were purchased from Sigma.

Homology modelling and epi-drug docking

Homology modelling of Smp_000700 was prepared as previously described [34] with some minor modifications. The homology model of full length Smp_000700 was constructed using the crystal structure of the human SMYD3 (PDB ID: 5EX3 [55]) as a template containing a substrate peptide and S-adenosyl homocysteine (SAH), the demethylated metabolite of the cofactor S-adenosyl methionine (SAM). Although the sequence similarity between the parasite target and the homologous human protein is not very high (32%), it’s still above the critical level (30%) in producing homology models [56].

The FASTA amino acid sequences for the protein to be modelled was obtained from Uniprot [57] and then used to perform a protein BLAST (BLASTp) search in NCBI [44] to obtain a homologous sequence of a protein to be used as a template. The crystal structure of the selected protein (H. sapiens SMYD3, PDB ID: 5EX3) suitable as a template was downloaded from PDB. The homology model was built using the homology modelling tool and a single template approach with Amber99 force field in MOE2018.10 [58]. Briefly, the primary sequence of Smp_000700 to be modelled was loaded into MOE together with the 3D structure of HsSMYD3, both sequences were aligned and the final 3D model of Smp_000700 was produced as a single output structure. The model was subsequently refined by energy minimization with RMSD of 0.1. MOE-Homology (developed by Chemical Computing Group, Inc.) combines the methods of segment-matching procedure and the approach to the modelling of insertion/deletion regions [59]. Using default parameters, four different softwares (Ramachandran plot analyses, ProSA-web, ERRAT and Verify3D) were used to validate the robustness of the homology model [60–63].

Docking simulations of BAY-598 and LLY-507 were performed using the Glide docking software within Maestro (Schrödinger Release 2017 [64]) as previously discussed [34]. The model was pre-processed using the Schrödinger Protein Preparation Wizard by assigning bond orders, adding hydrogens and performing a restrained energy minimisation of the added hydrogens using the OPLS_2005 force field. The docking site was identified over the substrate binding pocket of each homology model and a 12 Å docking grid (inner-box 10 Å and outer-box 22 Å) was prepared using, as a centroid, the substrate peptide. Glide SP precision was used keeping the default parameters and setting 5 as number of output poses per input ligand to include in the solution. The output poses were saved as a mol2 file. The docking results were visually inspected for their ability to bind the active site.

H3K36me2 detection

LLY-507 (6.25 μM), BAY-598 (25 μM) or siSmp_000700 treated male and female worms (alongside DMSO or siLuc controls; 21 individuals per condition) were homogenized with a TissueLyser (Qiagen) and total histones extracted using the EpiQuik^TM Total Histone Extraction Kit (Epigentek). Levels of H3K36me2 in adult histone extracts were measured using the EpiQuik^TM Global Di-Methyl Histone H3-K36 Quantification Kit (Fluorimetric, Epigentek). Technical duplicates of three biological replicates of each treatment were analysed according to the manufacturer’s instructions. Fluorescent readings (530_EX/590_EM nm) were obtained using a POLARstar Omega (BMG Labtech, UK) microtiter plate reader. Fluorescent values of the samples were corrected by subtracting the fluorescent readings of the blank (buffer only, provided in the kit). The mean of the adjusted control values (DMSO for LLY-507/BAY-598 treated worms and siLuc for siSmp_000700 treated worms) was set at 100% H3K36me2 and the standard deviation (SD) was calculated from the normalised values.

Vitellocyte and egg volume quantification

Prior to laser scanning confocal microscopy (LSCM) visualisation, the total number of eggs produced by GSK-J4 (0.2 μM), GSK-J1 (6.25 μM) or DMSO treated adult worm pairs were enumerated and subsequently immersed in PBS supplemented with DAPI (4',6-diamidino-2-phenylindole, 2 μg/ml). Fluorescent microscopic images (10 eggs per treatment) were captured on a Leica TCS SP8 super resolution laser confocal microscope fitted with a 63 X (water immersion) objective using the Leica Application Suite X. Green (egg autofluorescence) fluorescence was visualised with an argon or diode-pumped, solid state (DPSS) laser at 488 nm. DAPI was visualised using a 405 nm blue diode laser. The number of vitellocytes (DAPI⁺ cells) and overall volume (mapped by the green autofluorescence) for individual eggs were calculated using IMARIS 7.3 software (Bitplane).

HepG2 cell culture and MTT assays

Human Caucasian Hepatocyte Carcinoma (HepG2) cells (Sigma Aldrich, UK) were utilized to assess SGC epigenetic probe cell cytotoxicity in the application of the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay [52]. Briefly, cells were passaged at 70–80% confluence and seeded (20,000 cells/ 50 μL per well) in a 96-well, black-sided, clear-bottomed falcon plate (Fisher Scientific, UK) with the final plate column (8 wells) being treated as a blank (media only). Following a 24 h incubation (5% CO₂, 37°C, humidified), HepG2 cells were then treated with 50 μL of pre-warmed media (37°C) containing SGC EPs/EIs at 200 μM to 0.02 μM (final concentration 100 μM, 10 μM, 1 μM, 0.1 μM and 0.01 μM). Three positive (1% Triton X-100) and negative (1% DMSO) control wells per plate were additionally included.

After addition of compounds, each plate was then incubated for a further 20 h before application of MTT for assessment of overt compound cytotoxicity using the MTT assay [52, 65]. The MTT assay was read using the POLARstar Omega (BMG LabTech, UK) plate reader at an absorbance of 570 nm. CC₅₀s were calculated in GraphPad Prism.

Statistical analysis

All statistical analyses were performed using a Student’s t-test (two samples) or a two-way ANOVA followed by Least Significant Difference post-hoc correction (more than two samples). A p value less than 0.05 was considered statistically significant.

Compounds obtained from the Structural Genomics Consortium (SGC) display anti-schistosomal activity

Continuing our search for novel anti-schistosomal drug targets from within the parasite’s epigenetic machinery has led to the acquisition of 34 EPs and 3 EIs targeting histone modifying enzyme (HME) readers, writers and erasers from the SGC (S1 Table). The difference between EPs and EIs is based on the compound’s in vitro activity (EPs display in vitro potency > 100 nM; EIs do not), selectivity (EPs display > 30-fold selectivity vs other subfamilies; EIs do not) and on-target cell activity (EPs display activity > 1 μM; EIs do not) [30]. These EPs and EIs have initially been developed for oncology research and inflammatory disorders in humans or animal models [30]. Because of this, the human target is well characterised allowing us to identify the most likely S. mansoni target of each of these EPs and EIs by BLASTp analysis of the parasite’s genome (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. The Structural Genomics Consortium (SGC) epigenetic probes (EPs) and epigenetic inhibitors (EIs) used in this study.

https://doi.org/10.1371/journal.pntd.0007693.t001

In most cases, a clear one to one S. mansoni homolog (E values less than 1E^-10) could be identified for all EP and EI targets. However, no support for a schistosome PAD-4 (Uniprot Q9UM07) homolog was found in the S. mansoni genome (v7.0) and only weak support for a SMYD2 (Uniprot Q9NRG4) protein lysine methyltransferase (PKMT) homolog could be identified. In this later case, both Smp_342100 and Smp_000700 contained moderate sequence similarity (over the full-length or main catalytic domain sequences) to HsSMYD2 suggesting that this particular SMYD homolog may present unique (non-human) features useful from a drug development standpoint. In cases where EPs or EIs are unable to distinguish between family members such as PFI-1, JQ1 and BSP antagonists targeting the BET (Bromo and ExtraTerminal domain protein) family, IOX1 targeting the JMJD (JuMonJi C Domain-containing) protein demethylase family or LAQ824 and CI-994 targeting the class I HDACs (Histone DeACetylases), homologous schistosome families were the presumed targets. Therefore, a single Smp could not be identified for these compounds. Following on from this comparative sequence analysis, a single-point (10 μM) schistosomula screen of these 37 compounds was implemented to assess their anthelmintic properties (Table 1 and Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Epigenetic probes/inhibitors targeting histone modifying enzymes negatively modify schistosomula phenotype and motility metrics.

A collection of 37 EPs/EIs targeting histone modifying enzymes (readers, writers and erasers) as well as controls (PZQ and AUR) were screened against S. mansoni schistosomula at 10 μM for 72 h using the Roboworm platform as described in the Materials and Methods. (A) Parasite phenotype and (B) Parasite motility were both negatively affected by fourteen (38%) of these compounds upon repeat screening (each filled circle represents average phenotype or motility scores derived from ~ 80–120 schistosomula; n = 2–3; horizontal bars represent average scores derived from replicates). (C) Representative images of schistosomula phenotypes affected by the fourteen active EPs/EIs, compared to controls (Media only, DMSO, Auranofin and Praziquantel). Zˊ values obtained from these screens ranged from 0.27–0.48 (mean = 0.40) for phenotype and 0.41–0.56 (mean = 0.48) for motility. Table 1 (closest S. mansoni homolog) and S1 Table (structures, SMILES, MW, etc.) provide further information related to EPs/EIs screened and their putative S. mansoni targets.

https://doi.org/10.1371/journal.pntd.0007693.g001

Using the Roboworm platform [52], 13 EPs and 1 EI (LAQ824) were classified as hits (14/37; 38% hit rate) affecting both schistosomula phenotype (Fig 1A) and motility (Fig 1B) at 10 μM. These hits consisted of 5 compounds targeting epigenetic readers, 6 targeting epigenetic writers and 3 targeting epigenetic erasers. Reassuringly, GSK343 was identified as a hit in these screens, supporting previous anti-schistosomal investigations [66] and providing further confidence in the results obtained for the additional EPs and EIs. In one (out of three) JQ1 replicate screen, the motility result was not considered a hit as the value fell above the ‘hit’ cut-off (-0.35). However, as the other two replicates were identified as motility hits and all three replicates were within the phenotype hit threshold (below -0.15), JQ1 was included as an anti-schistosomal EP. Roboworm ‘hit’ cut-off boundaries are fully defined by Paveley et al [50] and have been successfully applied in our laboratory during other anthelmintic projects [7, 32–34, 52]. Furthermore, not all EPs targeting the same protein (or family) were equally active in these studies (Table 1 and Fig 1) [30]. For example, while JQ1 was a hit, PFI-1 and BSP were not; these compounds are all promiscuous BET bromodomain inhibitors [67–69]. Similarly, PFI-4 was a hit, but the promiscuous EPs NI-57 and OF-1 were not; these compounds all target human BRPF1/2/3 (BRomodomain and PHD Finger containing) members [70]. Likewise, SGC-CBP30 demonstrated activity against schistosomula, but I-CBP112 did not; both compounds were developed against the bromodomains within human CREBBP (cAMP Responsive Element Binding protein Binding Protein) and EP300 transcriptional co-activators [71, 72]. Finally, the BRomoDomain-9 epigenetic probe I-BRD9 [73], but not LP99 [74] and BI-9564 [75] was active against schistosomula. These results, and others including the pan-acting HDAC inhibitors LAQ824 (hit) vs CI-994 (non-hit), would suggest that not all EPs designed against the same human targets (or families) are equally effective against S. mansoni homologues. Two exceptions to this interpretation exist as both GSK343 [76] and UNC1999 [77] (EZH1/H2 EPs) as well as BAY-598 [38] and LLY-507 [37] (SMYD2 EPs) are hits against schistosomula. Therefore, in these cases, the use of complementary EPs may provide stronger support for schistosome target validation and progression during further investigations.

Microscopic assessment of affected schistosomula demonstrated a range of abnormal phenotypes including granulation, swelling, elongation and other irregular shape modifications (Fig 1C). Some compounds (NVS-CECR2-1, LLY-507 and GSK-J4) induced phenotypes comparable to auranofin. Where chemotype-matched negative (or less active) control compounds for hits were available (A-197 for A-196, BAY-369 for BAY-598, UNC2400 for UNC1999, GSK-J1/GSKJ5 for GSK-J4), they were also tested against schistosomula at 10 μM (S1 Fig). None of these chemotype controls demonstrated anti-schistosomula activity suggesting that specificity of the hits for the S. mansoni homologue is similar to that found for the original human target. Further SAR of the chemotype-matched controls and their related hit EPs/EIs with the schistosome target or whole organism could help provide evidence to support this contention. Nevertheless, encouraged by these single-point schistosomula screens (and the results of the chemotype controls), follow-on dose-response titrations of the 14 hits against schistosomula, adult schistosome pairs and a surrogate human cell line (HepG2) were next pursued (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Dose response titrations of the 14 active SGC compounds against schistosomula and adult schistosome pairs.

https://doi.org/10.1371/journal.pntd.0007693.t002

Upon dose-response titrations of these 13 EP and one EI (LAQ824) anti-schistosomula hits, several of the results warranted further comment. Firstly, several compounds had some degree of general cytotoxicity against HepG2 cells. While HepG2 cytotoxicity is a flag, it is not considered a no-go output in our compound triaging scheme. Nevertheless, the cytotoxicity results are unsurprising as these EPs were primarily developed for use in humans as anti-cancer agents [30], and, as HepG2 cells are derived from a hepatocellular carcinoma, varying sensitivity to these EPs was expected (and, as an example, confirmed for LLY-507, [37]). However, PFI-4, SGC-CBP30, A-196, BY-598 and GSK-J4 had minimal activity on this cell line’s viability (CC₅₀s > 50 μM) in these studies. Secondly, low μM anti-schistosomula EC₅₀s were generally a good predictor for low μM adult worm EC₅₀s (e.g. NVS-CECR2-1, LLY-507, GSK-J4). However, this was not always the case (e.g. I-BRD9, SGC-CBP30, MS023) and indicated that stage specific expression of the target may be an important consideration for compound progression. Thirdly, while LLY-507 may contain some off target activity in schistosomes (similar to what has been reported for human cells [30]), the anti-schistosomal potency and selectivity of BAY-598 were similar. These results suggested that both LLY-507 and BAY-598 (likely targeting a schistosome SMYD protein lysine methyltransferase homologue, Table 1) could be exchangeable in further progression of these anti-schistosomal compounds and their predicted target (s). Finally, these dose response titrations revealed that the best anti-schistosomal hit, when considering schistosomulum and adult worm potencies alongside HepG2 cytotoxicity, was GSK-J4. Because our previous investigations have independently identified crucial developmental roles for the schistosome protein methylation machinery [34, 36], we decided to further our investigations of LLY-507/BAY-598 and GSK-J4.

Protein lysine methyltransferase (LLY-507 and BAY-598) and demethylase (GSK-J4) modulators are amongst the most selectively-potent, anti-schistosomal epigenetic probes within the SGC collection

To provide a mechanistic context associated with the anti-schistosomal activity of LLY-507 and BAY-598, a combination of homology modelling, in silico molecular docking and detection of H3K36 dimethylation (me2) was performed (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The HMT inhibitors LLY-507 and BAY-598 significantly reduce adult worm H3K36me2 levels.

Homology modeling of Smp_000700, using H. sapiens SMYD3 (PDB 5EX3) as the template, was performed according to the Materials and Methods. (A) Domain architecture and homology model of Smp_000700 showing the tetratricopeptide repeat (TPR, gold, AA 76–195), the SET N-terminal domain (red, AA 279–312), the myeloid, nervy and DEAF-1 domain (MYND, blue, AA 313–370) and the SET C-terminal domain (green, AA 371–642). SAH (S-adenosyl homocysteine) and histone H3 are indicated. (B) Predicted binding of LLY-507 to Smp_000700 substrate binding pocket. (C) Predicted binding of BAY-598 to Smp_000700 substrate binding pocket. (D) Adult schistosome pairs (21 pairs/biological replicate; n = 3; 63 worm pairs in total) were co-cultured for 72 h in a sub-lethal concentration of LLY-507 (6.25 μM) or BAY-598 (25 μM) in 0.625% DMSO. After co-cultivation, schistosomes were separated by sex (males and females), histones extracted and total levels of H3K36me2 quantified by ELISA according to Methods.

https://doi.org/10.1371/journal.pntd.0007693.g002

Phylogenetic analyses of the three putative S. mansoni SMYD family members (Smp_121610, Smp_000700 and Smp_342100 [34]) alongside the five characterised H. sapiens SMYD (SMYD 1–5) proteins (using the SET domain only) indicated that S. mansoni does not contain a close HsSMYD2 homolog (supporting the findings in Table 1) (S2 Fig). However, as the substrate competitive inhibitors of HsSMYD2 (BAY-598 and LLY-507) are both active against schistosomula (but BAY-369, the negative control of BAY-598, is not, S1 Fig), these compounds are likely acting on one of these SmSMYDs (best estimate is Smp_000700 or Smp_342100, Table 1). As smp_000700 is more consistently expressed in 0–72 h schistosomula compared to smp_342100 [34], we, therefore, chose to specifically examine the putative binding of LLY-507/BAY-598 to a homology model of Smp_000700 (built from HsSMYD3, PDB 5EX3) (Fig 2A). Initial assessment of the homology model by Ramachandran plot analysis, ProSA-web, ERRAT and Verify3D indicated a high-quality structure (S3 Fig) surpassing agreeable standards defined in the literature [60–63]. Together, these tools verified that Smp_000700 adopted a similar three-dimensional arrangement to that of the structurally-solved human template and was suitable for predicting the binding organization of LLY-507 and BAY-598.

Smp_000700’s SET domain contains a series of β strands, which fold into three discrete sheets around a unique knot-like structure. In this structure, the C-terminal SET domain threads though a loop region, which is formed by a hydrogen bond between two segments of the protein chain [78]. The MYND domain is a putative zinc-finger motif, mainly involved in proline-rich protein interactions, and defines a distinctive class of histone writer called the SMYD proteins [79, 80]. The N-terminal SET domain is flanked by an additional domain known as tetracopeptide repeat domain (TPR). Smp_000700 contains two distinct binding sites within its SET domain: one for the substrate (histone protein, H3) and one for the demethylated cofactor (SAH). These two pockets are located on opposite faces of the SET domain and are connected by a deep channel running though the SET domain core. This architecture allows the side chain of the substrate (H3K36) to be close to the cofactor, which facilitates the transfer of a methyl group.

From the analysis of LLY-507 docking to Smp_000700, this compound adopted a similar binding confirmation to what was previously reported for HsSMYD2 bound to LLY-507 (deposited under the code 4WUY [37]). Specifically, LLY-507 is predicted to bind the substrate binding pocket with the pyrrolidine group (highlighted in yellow, Fig 2B) extending into the lysine channel (Phe184, Tyr240 and Tyr258 of HsSMYD2; His554, Tyr642 and Tyr662 of Smp_000700; S4 Fig, panels A and B) and connecting to the cofactor (represented here as SAM for HsSMYD2 and the product SAH, after donating a methyl group from SAM for Smp_000700) binding pocket. Furthermore, the nitrogen atom of LLY-507’s pyrrolidine ring is predicted to form a hydrogen bond with Smp_000700 Ser555 (S4 Fig, panel B). The benzonitrile group (as well as the piperazine; both highlighted in brown, Fig 2B) binds to the periphery of the substrate binding pocket (Thr185 for HsSMYD2 and Gly643 for Smp_000700; S4 Fig, panels A and B). This interaction putatively prevents the natural substrate (histone or other target protein) from binding to Smp_000700 and supports LLY-507’s mechanism of action as a substrate competitive inhibitor. Finally, this model suggests that the indole group of LLY-507 stacks into an accessory pocket of Smp_000700, just flanking the substrate binding pocket (area in blue, Fig 2B; S4 Fig, panel B). A similar indole confirmation is also observed for HsSMYD2/LLY-507 interactions (S4 Fig, panel A)

Regarding BAY-598, the docking simulation revealed a good occupation of Smp_000700’s substrate binding site (Fig 2C). However, in contrast to the binding mode of this compound with the human target (HsSMYD2) where the lysine binding channel is occupied by the 3, 4-chloro phenyl residue involved in π-stacking interactions with two aromatic residues (Phe184 and Tyr240, S4 Fig, panel C) [38], this particular BAY-598 feature is predicted to bind to an adjacent hydrophobic pocket of Smp_000700 (S4 Fig, panel D). Furthermore, the 3 - (difluoromethoxy) phenyl ring of BAY-598 (in yellow, Fig 2C) is predicted to insert into the channel interconnecting the two binding pockets (substrate and cofactor) of Smp_000700 with the difluoromethoxy group pointing toward the cofactor (in this case the SAM demethylated derivative, SAH). Two Smp_000700 aromatic amino acid residues (His554 and Tyr642, S4 Fig, panel D) are likely essential for positioning BAY-598’s hydroxyacetamide substituent (highlighted in green, Fig 2C) in the lysine binding channel with support from two hydrogen bonds donated by Asn641 (S4 Fig, panel D). These hydrogen bonds are donated by Thr185 in HsSMYD3 (S4 Fig, panel C) as previously shown [38]. The variation in predicted binding modes of LLY-507 and BAY-598 to Smp_000700 (i.e. LLY-507 adopting a confirmation analogous to the human homolog; BAY-598 binding to an adjacent hydrophobic pocket) may explain the difference in anti-schistosomal activity observed (LLY-507 > BAY-598). Further SAR could be helpful in providing answers to the differential activity of these two putative Smp_000700 interactors.

The homology model and in silico docking analyses suggested that both LLY-507 and BAY-598 could inhibit the protein lysine methyltransferase activity of Smp_000700. Therefore, as H3K36 is a specific target of SMYD2 dimethylation (me2) [81], we specifically examined this epitope (H3K36me2) in histone preparations derived from adult schistosome pairs incubated with sub-lethal concentrations of LLY-507 (6.25 μM) or BAY-598 (25 μM) (Fig 2D). Here, a significant reduction in the amount of H3K36me2 was detected in drug-treated parasites compared to controls (DMSO) with LLY-507 slightly more effective than BAY-598 (25% vs 23%) in mediating this activity. LLY-507 and BAY-598 mediated reductions in H3K36me2 were equivalent between the sexes. Together, these data provided evidence that both SMYD2 inhibitors affected H3K36me2 in schistosomes presumably via their inhibitory activity on Smp_000700. To verify this contention, RNAi of Smp_000700 was subsequently performed on adult schistosome pairs (Fig 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Schistosome motility, egg production and H3K36me2 are regulated by Smp_000700.

RNAi of adult schistosome pairs (21 worm pairs/biological replicate; n = 3 replicates) using siRNAs directed against smp_000700 and luc was performed as described in the Materials and Methods. (A) qRT-PCR analysis of smp_000700 RNA levels in siLuc vs siSmp_000700 treated worms at 48 h. (B) Quantification of adult schistosome worm motility at 168 h. (C) Enumeration of in vitro laid egg (IVLE) production at 168 h. (D) Detection of H3K36me2 in adult schistosome nuclear extracts at 168 h. Statistical significance is indicated (Student’s t test, two tailed, unequal variance). *** represents p < 0.001.

https://doi.org/10.1371/journal.pntd.0007693.g003

Here, short interfering RNAs (siRNAs) targeting smp_007000 (siSmp_000700) significantly reduced mRNA abundance by 66% when compared to siLuc controls (Fig 3A). This knockdown was also associated with a significant reduction in adult worm motility as well as egg production defects at 7 days post-treatment (Fig 3B and 3C). When H3K36me2 was examined, knockdown of smp_000700 led to a significant reduction in this epitope in both male and female schistosomes when compared to the controls (Fig 3D). This functional genomics data broadly recapitulates the drug studies using LLY-507/BAY-598 and provides supporting evidence that Smp_000700 is an active SMYD protein lysine methyltransferase required for schistosome motility, egg production and H3K36me2. Adult worm [34] and gonad (ovaries > testes) [82] expression of smp_000700 further provides a spatial temporal context as to how inhibition (chemical or functional genomics) of this gene product’s activity may negatively affect egg-production.

Amongst the EPs/EIs studied in this investigation, GSK-J4 demonstrated the most selective activity on schistosomes (schistosomula and adults, Table 2). This EP is a cell-permeable pro-drug (likely processed into GSK-J1 by intracellular schistosome esterases [39]) that is predicted to target a schistosome JMJD3 homolog (Smp_034000, Table 1) responsible for Fe (II) and 2-oxoglutarate-dependent demethylase activities. As IVLE production defects were consistently observed in adult worm screens (even at drug concentrations where worm motility was unaffected), we initiated an extensive titration (50 μM—0.05 μM) of GSK-J4 on adult schistosome pairs (Fig 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. The cell-permeable JMJD3 inhibitor GSK-J4, but not cell impermeable GSK-J1, significantly affects IVLE production and vitellocyte packaging.

Adult schistosome pairs (3 pairs/well; n = 3 or 6) were co-cultured for 72 h in GSK-J4 (50 μM– 0.05 μM), GSK-J1 (6.25 μM or 3.13 μM) or DMSO (0.625%) as described in the Materials and Methods. (A) Adult worm motility scores (red circles–males; blue circles–females). (B) Number of IVLEs produced. (C) Representative IVLE phenotypes (autofluorescence; Ex = 488 nm, Em = 519 nm and DAPI; Ex = 405 nm, Em = 458 nm) from schistosome pairs co-cultivated in GSK-J4 (0.2 μM), GSK-J1 (6.25 μM) or 0.625% DMSO for 72 h. (D) Quantification of egg volumes between treatment groups (n = 10 per group; GSK-J4–0.2 μM, GSK-J1–6.25 μM). (E) Number of vitellocytes per egg (n = 10 per group) between treatment groups (GSK-J4–0.2 μM, GSK-J1–6.25 μM). *, p < 0.05; ***, p < 0.001.

https://doi.org/10.1371/journal.pntd.0007693.g004

A direct comparison of GSK-J4 vs GSK-J1 (the cell impermeable parent of GSK-J4 [39]) on adult schistosome motility and egg production revealed the importance of an ethyl ester modification (found in GSK-J4, but not GSK-J1) in these in vitro phenotypic metrics (Fig 4A and 4B). For example, whereas GSK-J4 at 3.13 μM and 6.25 μM led to adult worm motility and IVLE deficiencies, GSK-J1 at these concentrations did not substantially alter these phenotypes when compared to DMSO controls. LSCM of IVLEs derived from schistosome cultures co-incubated with GSK-J4 (0.2 μM) vs GSK-J1 (6.25 μM) revealed further details about these differentially membrane permeable EPs (Fig 4C, 4D, 4E and S5 Fig). Of those phenotypically normal IVLEs found in GSK-J4 and GSK-J1 cultures at these concentrations, there was no difference in surface autofluorescence (Fig 4C and S5 Fig) or overall egg volume (Fig 4D and S5 Fig) measurements. However, when the number of vitellocytes was quantified, GSK-J4 significantly inhibited the packaging of this critical cell population into IVLEs whereas GSK-J1 did not (Fig 4C, Fig 4E and S5 Fig). Together, these data provided evidence that the specific anthelmintic activity of a JMJD3 EP is directly related to cellular (or in this case, schistosome) permeability. Alongside the results obtained with both LLY-507 and BAY-598, these findings additionally supported the important role of histone (protein) methylation/demethylation processes in schistosome development (Fig 1, Table 2) and adult worm phenotypes (Table 2, Figs 3 and 4).

Conclusions

The active compounds identified here from within the SGC epigenetic probe/epigenetic inhibitor collection represent exciting new starting points for the pursuit of next-generation anti-schistosomals. As many of these compounds are currently being explored for the treatment of non-communicable diseases in humans [83], their direct repositioning as schistosomiasis control agents could be rapid (depending upon medicinal chemistry improvements in selectivity). Specifically, our collective results suggest that compounds (LLY-507, BAY-598, GSK-J4) affecting protein methylation homeostasis in schistosomes are amongst some of the most potent agents from within the tested SGC collection. Therefore, and in agreement with other studies [17, 34, 36, 66], members of the schistosome protein methyltransferase/demethylase families should be considered validated anthelmintic targets for progression.