Discovery of a Potent and Selective Coactivator Associated Arginine

Article pubs.acs.org/jmc

Discovery of a Potent and Selective Coactivator Associated Arginine Methyltransferase 1 (CARM1) Inhibitor by Virtual Screening Renato Ferreira de Freitas,*,†,∥ Mohammad S. Eram,†,∥ David Smil,† Magdalena M. Szewczyk,† Steven Kennedy,† Peter J. Brown,† Vijayaratnam Santhakumar,† Dalia Barsyte-Lovejoy,† Cheryl H. Arrowsmith,†,§ Masoud Vedadi,*,†,‡ and Matthieu Schapira*,†,‡ †

Structural Genomics Consortium, University of Toronto, Toronto, ON M5G 1L7, Canada Princess Margaret Cancer Centre and Department of Medical Biophysics, University of Toronto, Toronto, ON M5G 1L7, Canada ‡ Department of Pharmacology and Toxicology, University of Toronto, Toronto, ON M5S 1A8, Canada §

S Supporting Information *

ABSTRACT: Protein arginine methyltransferases (PRMTs) represent an emerging target class in oncology and other disease areas. So far, the most successful strategy to identify PRMT inhibitors has been to screen large to medium-size chemical libraries. Attempts to develop PRMT inhibitors using receptor-based computational methods have met limited success. Here, using virtual screening approaches, we identify 11 CARM1 (PRMT4) inhibitors with ligand efficiencies ranging from 0.28 to 0.84. CARM1 selective hits were further validated by orthogonal methods. Two structure-based rounds of optimization produced 27 (SGC2085), a CARM1 inhibitor with an IC50 of 50 nM and more than hundred-fold selectivity over other PRMTs. These results indicate that virtual screening strategies can be successfully applied to Rossmann-fold protein methyltransferases.

■

The first crystal structure of a PRMT enzyme (PRMT3) was deposited in the PDB in 2001.14 Since then, many groups have been trying to identify PRMT inhibitors using a variety of strategies such as high throughput screening (HTS),15−17 adhoc synthesis of analogues of previous inhibitors,18−22 and, more recently, a fragment-based approach.23 A few PRMT1 inhibitors were also discovered by structure-based virtual screening.24−29 However, the development of PRMT inhibitors using computational methods has met with limited success, as compounds identified so far are relatively weak. In this work, we built a PRMT-focused virtual library and used structurebased in silico screening methods to identify potent and selective inhibitors of CARM1. To the best of our knowledge, this is the first report on a virtual screening campaign producing potent CARM1 inhibitors.

INTRODUCTION

Protein arginine methyltransferases (PRMTs) transfer a methyl group from S-adenosyl-L-methionine (SAM) to the terminal guanidino nitrogens of arginine on substrate proteins.1 PRMTs can be divided into three types (I, II, and III) according to the degree and position of methylation.2 Type I (PRMT1, 3, 4, 6, and 8) and II (PRMT5 and 9) enzymes convert arginine into monomethylarginine (MMA) and further into asymmetric and symmetric dimethylarginine, respectively, while PRMT7 (the only known type III PRMT) can only generate MMA. PRMTs have a wide range of protein substrates3,4 that are implicated in a variety of cellular functions including RNA processing, transcriptional regulation, DNA damage repair, and apoptosis.5−8 Deregulation of type-I PRMTs has been associated with diverse disease areas, including cancer.1 CARM1 (PRMT4), an enzyme that is both nuclear and cytoplasmic, methylates histone 3 on arginines 2, 17, and 26, as well as numerous nonhistone proteins involved in transcriptional regulation and mRNA processing.9,10 CARM1 is an important positive modulator of Wnt/β-catenin transcription and neoplastic transformation in colorectal cancer11 as well as a critical factor in estrogen-stimulated breast cancer growth,12 and its depletion results in decreased proliferation of myeloid leukemia cells in vivo.13 Selective inhibitors of CARM1 would be useful tools to validate this PRMT as a therapeutic target. © XXXX American Chemical Society

■

DESIGN RATIONALE

At the time we started our work, compounds 1 and 2 were the most potent PRMT inhibitors known, inhibiting CARM1 with IC50 values of 27 and 30 nM, respectively (Figure 1A).30 These two compounds were selective against PRMT1 and PRMT3, although no selectivity data for other PRMTs has been reported. Profiling of a close analogue of 2 where the fluorine is replaced with a chlorine revealed equipotency against Received: April 29, 2016

A

DOI: 10.1021/acs.jmedchem.6b00668 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 1. (A) Known type I PRMT inhibitors. The tails of the inhibitors are highlighted in blue. (B) Structural alignment of type I PRMT inhibitors in complex with CARM1 (1, 2: PDB codes 2Y1X and 2Y1W) and PRMT6 (3−5: PDB codes 4Y30, 5E8R, and 5EGS): 1 (green), 2 (yellow), 3 (orange), 4 (magenta), and 5 (cyan). The tails are anchored in the arginine binding channel.

CARM1 (70 nM) and PRMT6 (60 nM), but greater than 24fold selectivity over PRMT1, PRMT3, PRMT5, and PRMT8.23 Recently, EPZ020411 (3) was disclosed as the first potent small molecule PRMT6 inhibitor,31 and MS023 (4) as a potent type I PRMT inhibitor.32 We also identified 5 as a fragment inhibitor of PRMT6 (IC50 = 300 nM).23 A common feature of these five inhibitors is that they are anchored in the PRMT substrate arginine-binding channel through a basic amine tail (Figure 1B). Deconstructing these potent inhibitors shows that fragments that preserve the basic tail bind with exceptionally high ligand efficiency, probably through interaction with a conserved catalytic glutamate sidechain at the bottom of the substrate arginine-binding pocket.23 We therefore hypothesized that these two tails could serve as chemical warheads targeting type-I PRMTs. Based on these observations, we built a PRMT-focused virtual library composed of commercially available compounds where diverse scaffolds were appended to the two basic amine tails featured in pioneer type-I PRMT inhibitors.

CARM1 (PDB code 2Y1X), but the rest of the molecule adopted a very different, suboptimal conformation (Figure S2), suggesting that this compound is probably not active on PRMT6. Once docking of the PRMT warheads was validated in our system, the PRMT-focused library was docked to PRMT6 using the core docking option of Glide35 where the tail structures of 1 and 2, observed both in the CARM1 crystal structures and PRMT6 docking models, were used as the reference “core”. Out of the 17,000 docked compounds, the top 2300 were subjected to molecular mechanics Poisson−Boltzmann surface area (MM/PBSA) energy minimization and rescoring with AMBER 12.37 The top-ranked 1084 compounds were subjected to a 1 ns molecular dynamics (MD) simulation in explicit water with AMBER 12, and trajectories were postprocessed through the MM/PBSA approach. The binding poses of the best scoring compounds were then visually inspected, ligands with dubious binding poses rejected, and 51 compounds purchased from commercial vendors (Figure 2).

VIRTUAL SCREENING Substructure searches focused on the two basic amine tails (Figure 1A) were performed using the software FILTER33 against ∼22 million commercial compounds downloaded from the ZINC database,34 resulting in ∼132,000 compounds. Next, these were docked with Glide35 to the crystal structure of mouse PRMT6 (mPRMT6: PDB 4C03),36 the only available PRMT6 structure with a complete substrate-binding pocket at the time (see Experimental Section for details). As a preliminary validation step, compounds 1 and 2 were docked into the binding pocket of PRMT6 with hydrogen bonding constraints at Glu158, His320, Met160 (only for 2), and Glu167 (only for 1). Only compound 2 recapitulated the binding pose previously observed with CARM1 (PDB code 2Y1W), though the phenyl ring of 2 was rotated by 180°, pointing the methoxy group toward Cys53 (Figure S1). This was in agreement with the fact that a close analogue of 2 where the fluorine is replaced with a chlorine binds both CARM1 and PRMT6 with virtually the same potency.23 The tail of 1 also recapitulated the binding pose experimentally observed with

BIOCHEMICAL SCREENING The initial screening of the 51 compounds was done at two compound concentrations (10 and 50 μM), and 16 of them showed inhibition at both concentrations. However, when we performed full range titration (24 points) to determine IC50

■

■

Figure 2. Virtual screening workflow. B



Article

Table 1. Structure, logD, and IC50 Values for the 11 Type-I PRMT Inhibitors

Ligand efficiency was calculated using the equation LE = (1.37 × pIC50)/HA; LE is expressed as kcal/mol/atom. bLogD was calculated using ChemAxon’s MarvinSketch, version 15.8.31. cNI: no inhibition. IC50 determination experiments were performed in triplicate, and the values are presented as mean ± SD. a

well buried binding sites.38 Indeed, the arginine pocket of type-I PRMTs contains a catalytic glutamic acid (E258 in CARM1, E155 in PRMT6) buried at the bottom of the pocket that makes an electrostatic interaction with the terminal amino group of inhibitors 1−5 in complex with CARM1 and PRMT6. In addition to 7, six other fragments (8, 10, 11−13, and 15) were found to inhibit at least one PRMT with ligand efficiencies ranging from 0.34 to 0.64. With the goal of developing chemical probes selective for a single PRMT isoform, we pursued three CARM1-selective hits (compounds 6, 10, and 11) for further characterization. Direct binding of the compounds to CARM1 was tested by differential scanning fluorimetry (DSF) and differential static light scattering (DSLS). Compounds 6 and 11 induced a positive Tm shift of 0.45−2.1 °C by both DSF and DSLS, while 10 destabilized the protein (Table 2). Running the experiments in the presence of an excess of the cofactor S-adenosyl-Lmethionine (SAM) did not increase stabilization (data not shown). Binding of 6 was further confirmed by SPR, with KD values in good agreement with enzymatic IC50 values (Table 2).

values, only 11 compounds showed dose−response inhibition (Table 1). The alanine-amide tail of 1 was found in five inhibitors (compounds 6, 11−14), while the ethylenediamine tail of 2−5 was found in six inhibitors (compounds 7−10, 15, and 16). Surprisingly, the majority of hits were either pan-type-I PRMT inhibitors or selective for CARM1. The hit-rate for CARM1 was particularly high (11 inhibitors out of 51 tested), followed by PRMT6 (five inhibitors) and PRMT8 (three inhibitors). None of the compounds was able to inhibit PRMT1 or PRMT3 when tested up to 50 μM, even though the arginine pocket is highly conserved across all type-I PRMTs. Compound 6 was the most potent CARM1 inhibitor (IC50 = 1.9 μM; LE = 0.36) and was about 7-fold selective over PRMT6. Fragment 7 inhibited CARM1 (IC50 = 3.1 μM; LE = 0.84) and PRMT6 (IC50 = 1.9 μM; LE = 0.87) at low micromolar concentrations and showed exceptionally high ligand efficiencies. Fragments derived from the tail of inhibitors 1−5 also had very high ligand efficiencies,23 which can be explained by the fact that these fragments are small, charged ligands and bind in highly charged, C



Article

pyridin-3-yl group of the inhibitor adopts different conformations in CARM1 and PRMT6. In particular, predicted πstacking interactions between CARM1’s F153 and the pyridine and phenyl rings of 6 are lost in PRMT6, which could account for the reduced inhibitory activity.

Table 2. Biophysical Validation of CARM1 Hits Compound CARM1 IC50 (μM) 6 10 11 a

1.9 ± 0.3 16.6 ± 1.7 20.5 ± 4.5

ΔTm (°C)

ΔTagg (°C)

KD SPR (μM)

2.1 −2.4 0.45

1.9 −2.0 0.5

9 NBa NB

■

No binding.

SAR BY CATALOG

Following the biochemical validation of 6 and encouraging structural insight, this compound was selected as an attractive candidate for further optimization. A substructure search for commercially available compounds sharing the core scaffold of 6 was performed, following the widely used “SAR by catalog” approach39 resulting in nine close analogues (17−25, Table 3). Compound 17, which is the (R)-enantiomer of 6, was about 21-fold less active. Removing the methyl group on the α-carbon (18) resulted in 7-fold loss of potency. Additionally, in agreement with previous work by Purandare et al.,19 neither the methylation of the terminal amino group (19 and 20) nor bulky substituents on the α-carbon (21 and 22) were tolerated. Next, we kept the alanine-amide tail of 6 unchanged and evaluated the structure−activity relationship on the left side of

Compound 6 inhibited CARM1 selectively, with an IC50 of 1.9 μM and a ligand efficiency of 0.36, and was validated by orthogonal biophysical methods. The >7-fold selectivity of 6 over PRMT6 was not expected, given the high sequence similarity of the arginine binding pocket. To better understand this selectivity profile, we docked 6 to the structure of CARM1 in complex with 1 (PDB code 2Y1X, Figure 3A). The binding pose of the alanine-amide tail of 6 in CARM1’s argininebinding channel recapitulated that observed with PRMT6 (Figure 3B). At the mouth of the arginine-binding pocket, three residues are not conserved between the two enzymes: F153, Q159, and N266 in CARM1 and C53, V59, and H166 in PRMT6, respectively). As a result the 2-(2,4-difluorophenoxy)-

Figure 3. (A) Top: predicted binding mode of 6 (orange sticks) against CARM1. Bottom: view of the mouth of the arginine-pocket in CARM1. (B) Top: predicted binding mode of 6 against PRMT6. Bottom: view of the mouth of the arginine-pocket in PRMT6. Residues in the active site are shown as cyan sticks. Dashed lines represent intermolecular hydrogen bonds. Surface is colored in light blue. D



Article

Table 3. Structure−Activity Relationships of the Analogues of Compound 6

a Ligand efficiency was calculated using the equation LE = (1.37 × pIC50)/HA; LE is expressed as kcal/mol/atom. bLogD was calculated using ChemAxon’s MarvinSketch, version 15.8.31. cNI: no inhibition. IC50 determination experiments were performed in triplicate, and the values are presented as mean ± SD.

substituent at R1 is essential for potent and selective inhibition of CARM1.

the molecule. Compound 23, with a 3,4-difluorophenoxy group, was 3-fold more potent than 6. Replacement of the 2(2,4-difluorophenoxy)pyridin-3-yl group of 6 with a 4-(3,5dimethyl-1H-pyrazol-1-yl)phenyl moiety (24) was not tolerated, but the introduction of a 3-fluoro-4-(2-fluorophenoxy)phenyl group (25) improved potency significantly (IC50 = 0.36 μM). Encouraged by this result, we synthesized the corresponding (S)-enantiomer (26), which resulted in an IC50 of 0.24 μM. Following the identification of compound 26, we synthesized six analogues (27-32) selected in silico to briefly explore the structure−activity relationship around this scaffold (Table 4). Compound 27 (SGC2085), which features a methyl at position R1 and a 3,5-dimethylphenoxy at R2 had an IC50 of 50 nM for CARM1 and was over 100-fold selective for CARM1 over PRMT6. Compound 28, which also has a methyl group at R1, was equipotent to 26 and showed approximately ∼32-fold preference for CARM1 inhibition versus PRMT6 inhibition. Other compounds where R1 is hydrogen and R2 contains hydrophobic groups were either less potent or less selective over PRMT6. These results indicate that the presence of a

■

BINDING MODE OF 27

To investigate the structural basis for the observed selectivity of 27 for CARM1 over PRMT6, a docking model of this compound in complex with CARM1 and PRMT6 was built with Glide,35 using the crystal structure of CARM1 in complex with 1 (PDB code 2Y1X) and the crystal structure of PRMT6 in complex with 4 (PDB code 5E8R). Compound 27 had a better docking score when in complex with CARM1 than with PRMT6 (−9.15 vs −6.32 kcal/mol, respectively). The protein− ligand complexes obtained from the docking study were then subjected to a 10 ns molecular dynamics simulation and MM/ PBSA binding free energy calculations. In agreement with the docking results, compound 27 also showed a stronger binding energy with CARM1 (−33.2 kcal/mol) than with PRMT6 (−26.2 kcal/mol). In addition, only the CARM1−27 complex was stable over the 10 ns simulation (Figure S3). In the predicted binding mode, not only does the alanine-amide tail of 27 recapitulate interactions observed between CARM1 and 1, E



Article

Table 4. Structure−Activity Relationships of the Analogues of Compound 26.

Ligand efficiency was calculated using the equation LE = (1.37 × pIC50)/HA; LE is expressed as kcal/mol/atom. bLogD was calculated using ChemAxon’s MarvinSketch, version 15.8.31. IC50 determination experiments were performed in triplicate, and the values are presented as mean ± SD. a

Figure 4. (A) Predicted binding mode of 27 (orange sticks) derived from the molecular dynamics simulation. Residues in the active site are shown as cyan sticks. Dashed lines represent intermolecular hydrogen bonds. (B) Rotated view of 27 showing the 3-methyl group binding into the pocket formed by F153, Q159, and M163. Surface is colored in light blue.

Table 5. IC50 Determination for 27 on PRMTs Compound PRMT1 IC50 (μM) 27

>100

PRMT3 IC50 (μM)

CARM1 IC50 (μM)

PRMT5 IC50 (μM)

PRMT6 IC50 (μM)

PRMT7 IC50 (μM)

PRMT8 IC50 (μM)

>100

0.05 ± 0.02

>100

5.2 ± 0.9

>100

>50

but the compound also forms stabilizing π-stacking interaction with the side-chain of H415 and favorable hydrophobic interactions with the aromatic residues Y62, F153, F154, and F475 (Figure 4A). The 3-methyl group is occupying a pocket formed by F153, Q159, and M163 (Figure 4B). This pocket is absent in PRMT6 since F153 and Q159 are replaced with C50 and V56, respectively (Figure S4). While a crystal structure would be necessary to unequivocally confirm the predicted

binding mode of 27, we believe that the docking model is sufficiently robust to guide follow-up chemistry. With the exception of PRMT6 (IC50 = 5.2 μM), 27 did not inhibit other PRMTs (Table 5). Considering its small size (MW = 312.4 Da), 27 has an excellent selectivity profile, which can probably be further improved by exploiting differences in the binding sites of the two enzymes outside the argininebinding pocket. Compound 27 also showed complete selectivity against a panel of 21 human protein methyltransF



Article

reasonable physicochemical properties (logD = 2.7; HBA = 2; HBD = 2). Considering that alanine-amide-based PRMT inhibitors can cross cell membranes,42 we believe that a structure-based approach may be applied to improve both biochemical and cellular activities of this scaffold.

ferases (Figure S5) tested at three different concentrations (1, 10, and 50 μM). To characterize the mechanism of action of 27 in solution, IC50 values were determined at various concentrations of SAM and peptide substrate (Figure S6). Increasing concentration of substrate peptide or cofactor did not affect IC50 values, indicative of a noncompetitive mechanism of inhibition, which has been previously shown for other protein methyltransferase inhibitors binding at the substrate pocket.23,40 One hypothesis for this mechanism is that the substrate binds outside the catalytic pocket of CARM1 with significant affinity, and it is not completely displaced by the inhibitor.

■

CONCLUSION Analysis of structural information combined with virtual screening yielded the ligand-efficient CARM1 inhibitor 6. Two rounds of optimization guided by docking studies produced 27, a potent and selective CARM1 inhibitor. Based on its high ligand efficiency, 27 is a good starting point for further optimization. However, this and previous work suggest that overcoming the cellular liability associated with its alanineamide tail may be challenging. While SET-domain methyltransferases have so far resisted virtual screening approaches, these results establish virtual screening as a valid tool toward the discovery of chemical probes targeting type-I PRMTs.

■

CELLULAR ACTIVITY OF 27 CARM1 was previously shown to asymmetrically dimethylate BAF155 at R1064.41 No cellular activity was observed for 27 when tested up to 10 μM (48 h exposure in HEK293 cells), while methylation of BAF155 was abrogated by 10 μM of the dual CARM1/PRMT6 inhibitor MS049 (http://www.thesgc. org/chemical-probes/MS049) (Figure 5).

■

EXPERIMENTAL SECTION

Docking. The X-ray crystal structure of mouse PRMT635 (mPRMT6: PDB 4C03) was used for docking studies. The preparation of the protein for docking was performed with PrepWizard using the standard protocol, including the addition of hydrogens, the assignment of bond order, assessment of the correct protonation states, and a restrained minimization using the OPLS-AA 2005 force field. The CARM1 structure in complex with 1 (2Y1X) was aligned with PRMT6 (4C03); inhibitor 1 was then merged with the PRMT6 structure. The PRMT6 structure was merged with inhibitor 2 using the same procedure. The two PRMT6 complexes generated in the previous steps were used to define the binding site where the compounds were docked. Receptor grids were calculated at the centroid of the inhibitors with the option to dock ligands of similar size. Hydrogen bonding constraints at Glu158, His320, Met160 (only for 2), and Glu167 (only for 1) were defined. All docking calculations were performed using Glide SP35 using default settings. Also, the core docking option was turned on to allow only ligand poses that have their core aligned within 1.0 Å of the reference core (the tails of inhibitors 1 and 2). Preparing Ligands for Docking. Using the tails of inhibitors 1 and 2 as reference, a substructure search against the ZINC database34 resulted in ∼132,000 compounds. LigPrep44 was used to prepare the ligands using default settings. The total number of ligands resulting from this procedure was 365,745. CARM1 Expression and Purification. Transient expression of full-length (residues 1−608) human CARM1-HaloTag in pFN21A vector was achieved by transfection of suspension culture of FreeStyle 293-F cells (at the exponential growth phase) grown in FreeStyle 293 Expression Medium (Invitrogen) with FectoPRO transfection reagent (Polyplus-transfection, New York, NY). The transfected cells were harvested 76 h post-transfection. The harvested cells were resuspended in PBS plus protease inhibitor (cOmpleteTM EDTA free protease inhibitor cocktail tablets, Roche). The cells were then lysed chemically after adding 10% glycerol and 2 mM 2-mercaptoethanol by rotating 30 min with NP40 (final concentration of 0.6%) and 10 U/L Benzonase nuclease (Sigma) followed by sonication at frequency of 7 Hz (10S on/10S off) for 2 min (Sonicator 3000, Misoni). The crude extract was clarified by high-speed centrifugation (60 min at 36,000 × g at 4 °C) by Beckman Coulter centrifuge. The recombinant protein was purified using the HaloTag Protein Purification System (Promega, Madison, WI), and the tag was removed using HaloTEV Protease as described previously.45 Enzymatic Assays. A radiometric assay was used to study the in vitro inhibition of PRMTs as described previously.32,46 In principle, radiolabeled S-adenosylmethionine (3H-SAM, PerkinElmer Life Sciences, specific activity range 12−18 Ci/mmol) served as methyl donor, for methylation of the biotinylated histone peptides.

Figure 5. Effect of 27 on the methylation of BAF155 by CARM1 in cells.

We assume that the absence of cellular activity for 27 is due to poor permeability. Indeed, Huynh et al. found that potent meta-substituted (S)-alanine benzyl amide CARM1 inhibitors only had modest permeability (PAMPA assay).42 They were able to optimize the permeability of the inhibitors by increasing their hydrophobicity, while still keeping the alanine-amide moiety. However, they did not report if the improved permeability translated into cellular activity. In another report, Therrien et al. replaced the alanine-amide moiety with an ethylene diamine tail.43 Despite the observed CARM1 enzymatic activity (IC50 = 0.20 μM), the diamine compound failed to show any cellular activity. To test whether the alanine-amide could be favorably replaced with an ethylene diamine, a PRMT-anchoring tail found in cell-active compounds, in the context of the scaffold of 26, we synthesized compound 33 (Table 4). In contrast to results observed by Therrien et al.,43 this compound was 48fold less active than 26 against CARM1 but was equipotent against PRMT6. Probably due to its mediocre potency, compound 33 was inactive in cells (data not shown). These results clearly illustrate that the two basic amine tails present in all type-I PRMT inhibitors are not uniformly interchangeable. We speculate that poor membrane permeability resulting from the large polar surface area of the alanine-amide tail of 27 is underlying the observed lack of cellular activity. Other PRMT inhibitors featuring the same alanine-amide moiety have been reported with high biochemical potency but no cellular activity,42,43 suggesting that the alternate ethylene diamine warhead featured in recent type-I PRMT chemical probes (Figure 1) is more likely to yield compounds active in cells. However, unlike previous compounds, 27 is relatively small (MW = 312.4 Da) with high ligand efficiency (LE = 0.43) and G



Article

Incorporation of the tritiated methyl into the arginine residues of the substrate histone peptides then was measured in scintillation proximity FlashPlates Plus (PerkinElmer, Waltham, MA). The amount of methylated peptides was then quantified by tracing the radioactivity (counts per minute) using TopCount NXT plate reader (PerkinElmer, Waltham, MA). The C-terminally biotinylated histone H3 peptide composed of the first 25 amino acids residues (H3 1−25) was used as substrate for CARM1. The typical assay mixture (10 μL volume) contained 25 nM CARM1, 0.7 μM H3 1−25, and 1.9 μM SAM in 20 mM bicine (pH 8.5). The IC50 values were determined under balanced conditions at apparent KM concentrations of both substrates by titration of the compound in the reaction mixture in a range between 100 and 0.006 μM. DSF. Differential scanning fluorimetry (DSF) measurements were performed with a Light Cycler 480 II instrument from Roche Applied Science. The protein was assayed at 0.2 mg/mL in 100 mM HEPES (pH 7.5), 150 mM NaCl, 2% DMSO final, and 5× Sypro Orange (Sypro Orange, purchased from Invitrogen as a 5000× stock solution was diluted 1:1000 to yield a 5× working concentration). The compounds were titrated up to 600 μM to assess their stabilizing effect. DSF was carried out by increasing the temperature by 4 °C/min from 30 to 95 °C, and data points were collected at 0.4 °C intervals. The temperature scan curves were fitted to a Boltzmann sigmoid function, and the Tm values were obtained from the midpoint of the transition as described previously.47 DSLS. Differential static light scattering (DSLS) experiments were performed as previously described.46 Briefly, CARM1 at 0.2 mg/mL in 100 mM HEPES pH 7.5 and 150 mM NaCl was incubated with the titrated compound (2% DMSO final). Forty microliters of the protein/ compound mixture was heated from 30 to 80 °C at a rate of 1 °C per min in a clear-bottom 384-well plate (Nunc, Rochester, NY). Protein aggregation was monitored by measuring the intensity of the scattered light every 30 s with a CCD camera. These total intensities were then plotted against temperature and fitted to the Boltzman equation by nonlinear regression. SPR. Surface plasmon resonance (SPR) experiments were performed using a Biacore T200 (GE Health Sciences Inc.) at 20 °C. Approximately 4500 RU of CARM1 was amino coupled to a CM5 Chip (according to the manufacturer’s protocol), and another cell being left blank for reference subtraction. Compounds were serially diluted in DMSO and transferred to the buffer (HBS-EP) giving 5% DMSO final. Compounds were tested with 30 s contact time at 75 μL/ min. KD values were determined using Steady State Affinity Fitting and the Biacore T200 Evaluation Software (GE Health Sciences Inc.). SAM binding to CARM1 showed the protein to be approximately 90% functional on the chip. Cellular Assay. HEK293 cells were grown in 12-well plates in DMEM supplemented with 10% FBS, penicillin (100 U/mL), and streptomycin (100 μg/mL). Thirty percent confluent cells were treated with inhibitors or DMSO. After 48 h, media were removed and cells were lysed in 100 μL of total lysis buffer (20 mM Tris-HCl pH 8.0, 150 mM NaCl, 1 mM EDTA, 10 mM MgCl2, 0.5% Triton X-100, 12.5 U/mL benzonase (Sigma)), complete EDTA-free protease inhibitor cocktail (Roche). After 3 min incubation at room temperature, SDS was added to 1% final concentration. Lysates were run on SDS-PAGE, and immunoblotting was done as outlined below to determine the levels of unmethylated and methylated BAF155. Western Blot. Total cell lysates were resolved in 4−12% Bis-Tris Protein Gels (Invitrogen) with MOPS buffer (Invitrogen) and transferred in for 1.5 h (80 V) onto PVDF membrane (Millipore) in Tris-Glycine transfer buffer containing 20% MeOH and 0.05% SDS. Blots were blocked for 1 h in blocking buffer (5% milk in 0.1% Tween 20-PBS) and incubated with primary antibodies mouse anti-BAF155 (1:500, SantaCruz, #sc-48350) and rabbit antidimethyl BAF155 (R1064 asymmetrically dimethylated) (1:1000, Millipore, #ABE1339) in blocking buffer overnight at 4 °C. After five washes with 0.1% Tween 20 PBS, the blots were incubated with goat-anti rabbit (IR800 conjugated, LiCor, #926-32211) and donkey antimouse (IR 680, LiCor, #926-68072) antibodies (1:5000) in Odyssey Blocking Buffer (LiCor) for 1 h at RT and washed five times with 0.1% Tween

20 PBS. The signal was read on an Odyssey scanner (LiCor) at 800 and 700 nm. Chemistry and Compound Purity. Compounds 26−33 were synthesized according to the procedures described in the Supporting Information. All commercial and synthesized compounds tested in vitro were >95% pure. Purity determination was conducted by UV absorbance at 254 nm during tandem liquid chromatography/mass spectrometry (LC−MS) using a Waters Acquity separations module. Identity was determined via low resolution mass spectrometry (LRMS) conducted in positive ion mode using a Waters Acquity SQD mass spectrometer (electrospray ionization source) fitted with a PDA detector. Mobile phase A consisted of 0.1% formic acid in water, while mobile phase B consisted of 0.1% formic acid in acetonitrile. The gradient ran from 5% to 95% mobile phase B over 3 min at 0.5 mL/ min. An Acquity CSH C18 (2.1 mm × 50 mm, 130 Å, 1.7 μm) column (part no. 186005296) was used with column temperature maintained at 25 °C. The sample solution injection volume was 5 μL. The chlorine analogue of compound 2 was synthesized as previously described.48

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jmedchem.6b00668. Experimental procedure for the molecular dynamics simulation, proposed binding models of 1 and 2 in complex with PRMT6, RMSD plot for the MD simulation of 27 in complex with CARM1, proposed binding mode of 27 in complex with PRMT6, and the general procedure for synthesis of 26−33 (PDF) Molecular strings for all compounds that were tested in vitro (CSV) Molecular strings for all compounds that were tested in vitro (CSV)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Phone: 416-946-0132. *E-mail: [email protected]. Phone: 416-976-0897. *E-mail: [email protected]. Phone: 416-9783092. Author Contributions ∥

These authors contributed equally to this work.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The SGC is a registered charity (No. 1097737) that receives funds from AbbVie, Bayer Pharma AG, Boehringer Ingelheim, Canada Foundation for Innovation, Eshelman Institute for Innovation, Genome Canada, Innovative Medicines Initiative (EU/EFPIA) [ULTRA-DD Grant 115766], Janssen, Merck & Co., Novartis Pharma AG, Ontario Ministry of Economic Development and Innovation, Pfizer, São Paulo Research Foundation-FAPESP, Takeda, and the Wellcome Trust. We thank OpenEye Scientific Software, Inc., for the free academic license of the FILTER software.

■

ABBREVIATIONS USED PMT, protein methyltransferase; PRMT, protein arginine methyltransferase; CARM1, coactivator associated arginine methyltransferase 1; SAM, S-adenosyl-L-methionine; SAH, H



Article

Sadenosyl-homocysteine; DSF, differential scanning fluorimetry; DSLS, differential static light scattering; SPR, surface plasmon resonance

■

Biological Evaluation of Carboxy Analogues of Arginine Methyltransferase Inhibitor 1 (AMI-1). ChemMedChem 2010, 5, 398−414. (19) Purandare, A. V.; Chen, Z.; Huynh, T.; Pang, S.; Geng, J.; Vaccaro, W.; Poss, M. A.; Oconnell, J.; Nowak, K.; Jayaraman, L. Pyrazole Inhibitors of Coactivator Associated Arginine Methyltransferase 1 (CARM1). Bioorg. Med. Chem. Lett. 2008, 18, 4438−4441. (20) Wan, H.; Huynh, T.; Pang, S.; Geng, J.; Vaccaro, W.; Poss, M. A.; Trainor, G. L.; Lorenzi, M. V.; Gottardis, M.; Jayaraman, L.; Purandare, A. V. Benzo[d]imidazole Inhibitors of Coactivator Associated Arginine Methyltransferase 1 (CARM1)-Hit to Lead Studies. Bioorg. Med. Chem. Lett. 2009, 19, 5063−5066. (21) t Hart, P.; Lakowski, T. M.; Thomas, D.; Frankel, A.; Martin, N. I. Peptidic partial bisubstrates as inhibitors of the protein arginine Nmethyltransferases. ChemBioChem 2011, 12, 1427−1432. (22) Bissinger, E. M.; Heinke, R.; Spannhoff, A.; Eberlin, A.; Metzger, E.; Cura, V.; Hassenboehler, P.; Cavarelli, J.; Schüle, R.; Bedford, M. T.; Sippl, W.; Jung, M. Acyl Derivatives of P-Aminosulfonamides and Dapsone as New Inhibitors of the Arginine Methyltransferase hPRMT1. Bioorg. Med. Chem. 2011, 19, 3717−3731. (23) Ferreira de Freitas, R.; Eram, M. S.; Szewczyk, M. M.; Steuber, H.; Smil, D.; Wu, H.; Li, F.; Senisterra, G.; Dong, A.; Brown, P. J.; Hitchcock, M.; Moosmayer, D.; Stegmann, C. M.; Egner, U.; Arrowsmith, C.; Barsyte-Lovejoy, D.; Vedadi, M.; Schapira, M. Discovery of a Potent Class I Protein Arginine Methyltransferase Fragment Inhibitor. J. Med. Chem. 2016, 59, 1176−1183. (24) Xie, Y.; Zhou, R.; Lian, F.; Liu, Y.; Chen, L.; Shi, Z.; Zhang, N.; Zheng, M.; Shen, B.; Jiang, H.; Liang, Z.; Luo, C. Virtual Screening and Biological Evaluation of Novel Small Molecular Inhibitors against Protein Arginine Methyltransferase 1 (PRMT1). Org. Biomol. Chem. 2014, 12, 9665−9673. (25) Spannhoff, A.; Heinke, R.; Bauer, I.; Trojer, P.; Metzger, E.; Gust, R.; Schüle, R.; Brosch, G.; Sippl, W.; Jung, M. Target-Based Approach to Inhibitors of Histone Arginine Methyltransferases. J. Med. Chem. 2007, 50, 2319−2325. (26) Spannhoff, A.; Machmur, R.; Heinke, R.; Trojer, P.; Bauer, I.; Brosch, G.; Schüle, R.; Hanefeld, W.; Sippl, W.; Jung, M. A Novel Arginine Methyltransferase Inhibitor with Cellular Activity. Bioorg. Med. Chem. Lett. 2007, 17, 4150−4153. (27) Heinke, R.; Spannhoff, a; Meier, R.; Trojer, P.; Bauer, I.; Jung, M.; Sippl, W. Virtual Screening and Biological Characterization of Novel Histone Arginine Methyltransferase PRMT1 Inhibitors. ChemMedChem 2009, 4, 69−77. (28) Feng, Y.; Li, M.; Wang, B.; Zheng, Y. G. Discovery and Mechanistic Study of a Class of Protein Arginine Methylation Inhibitors. J. Med. Chem. 2010, 53, 6028−6039. (29) Wang, J.; Chen, L.; Sinha, S. H.; Liang, Z.; Cai, H.; Muniyan, S.; Chou, Y.; Yang, C.; Yan, L.; Feng, Y.; Li, K.; Lin, M.; Jiang, H.; Zheng, Y.; Luo, C. Pharmacophore-Based Virtual Screening and Biological Evaluation of Small Molecule Inhibitors for Protein Arginine Methylation. J. Med. Chem. 2012, 55, 7978−7987. (30) Sack, J. S.; Thieffine, S.; Bandiera, T.; Fasolini, M.; Duke, G. J.; Jayaraman, L.; Kish, K. F.; Klei, H. E.; Purandare, A. V.; Rosettani, P.; Troiani, S.; Xie, D.; Bertrand, J. a. Structural Basis for CARM1 Inhibition by Indole and Pyrazole Inhibitors. Biochem. J. 2011, 436, 331−339. (31) Mitchell, L. H.; Drew, A. E.; Ribich, S. A.; Rioux, N.; Swinger, K. K.; Jacques, S. L.; Lingaraj, T.; Boriack-Sjodin, P. A.; Waters, N. J.; Wigle, T. J.; Moradei, O.; Jin, L.; Riera, T.; Porter-Scott, M.; Moyer, M. P.; Smith, J. J.; Chesworth, R.; Copeland, R. A. Aryl Pyrazoles as Potent Inhibitors of Arginine Methyltransferases: Identification of the First PRMT6 Tool Compound. ACS Med. Chem. Lett. 2015, 6, 655− 659. (32) Eram, M. S.; Shen, Y.; Szewczyk, M. M.; Wu, H.; Senisterra, G.; Li, F.; Butler, K. V.; Kaniskan, H. Ü .; Speed, B. A.; dela Seña, C.; Dong, A.; Zeng, H.; Schapira, M.; Brown, P. J.; Arrowsmith, C. H.; Barsyte-Lovejoy, D.; Liu, J.; Vedadi, M.; Jin, J. A Potent, Selective, and Cell-Active Inhibitor of Human Type I Protein Arginine Methyltransferases. ACS Chem. Biol. 2016, 11, 772−781. (33) FILTER; OpenEye Scientific Software: Santa Fe, NM, 2014.

REFERENCES

(1) Yang, Y.; Bedford, M. T. Protein Arginine Methyltransferases and Cancer. Nat. Rev. Cancer 2013, 13, 37−50. (2) Bedford, M. T.; Richard, S. Arginine Methylation: An Emerging Regulator of Protein Function. Mol. Cell 2005, 18, 263−272. (3) Boisvert, F.-M.; Chénard, C. A.; Richard, S. Protein Interfaces in Signaling Regulated by Arginine Methylation. Sci. Signaling 2005, 2005, re2. (4) Guo, A.; Gu, H.; Zhou, J.; Mulhern, D.; Wang, Y.; Lee, K. a; Yang, V.; Aguiar, M.; Kornhauser, J.; Jia, X.; Ren, J.; Beausoleil, S. a; Silva, J. C.; Vemulapalli, V.; Bedford, M. T.; Comb, M. J. Immunoaffinity Enrichment and Mass Spectrometry Analysis of Protein Methylation. Mol. Cell. Proteomics 2014, 13, 372−387. (5) Jenuwein, T.; Allis, C. D. Translating the Histone Code. Science 2001, 293, 1074−1080. (6) Boisvert, F.-M.; Côté, J.; Boulanger, M.-C.; Richard, S. A Proteomic Analysis of Arginine-Methylated Protein Complexes. Mol. Cell. Proteomics 2003, 2, 1319−1330. (7) Bedford, M. T.; Clarke, S. G. Protein Arginine Methylation in Mammals: Who, What, and Why. Mol. Cell 2009, 33, 1−13. (8) Lee, D. Y.; Teyssier, C.; Strahl, B. D.; Stallcup, M. R. Role of Protein Methylation in Regulation of Transcription. Endocr. Rev. 2005, 26, 147−170. (9) Lee, Y.-H.; Coonrod, S. A.; Kraus, W. L.; Jelinek, M. A.; Stallcup, M. R. Regulation of Coactivator Complex Assembly and Function by Protein Arginine Methylation and Demethylimination. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 3611−3616. (10) Cheng, D.; Côté, J.; Shaaban, S.; Bedford, M. T. The Arginine Methyltransferase CARM1 Regulates the Coupling of Transcription and mRNA Processing. Mol. Cell 2007, 25, 71−83. (11) Ou, C. Y.; LaBonte, M. J.; Manegold, P. C.; So, A. Y.; Ianculescu, I.; Gerke, D. S.; Yamamoto, K. R.; Ladner, R. D.; Kahn, M.; Kim, J. H.; Stallcup, M. R. A Coactivator Role of CARM1 in the Dysregulation of Beta-Catenin Activity in Colorectal Cancer Cell Growth and Gene Expression. Mol. Cancer Res. 2011, 9, 660−670. (12) Frietze, S.; Lupien, M.; Silver, P. A.; Brown, M. CARM1 Regulates Estrogen-Stimulated Breast Cancer Growth through upRegulation of E2F1. Cancer Res. 2008, 68, 301−306. (13) Zhang, X.; Zhou, L.; Cheng, X. Crystal Structure of the Conserved Core of Protein Arginine Methyltransferase PRMT3. EMBO J. 2000, 19, 3509−3519. (14) Vu, L. P.; Perna, F.; Wang, L.; Voza, F.; Figueroa, M. E.; Tempst, P.; Erdjument-Bromage, H.; Gao, R.; Chen, S.; Paietta, E.; Deblasio, T.; Melnick, A.; Liu, Y.; Zhao, X.; Nimer, S. D. PRMT4 Blocks Myeloid Differentiation by Assembling a Methyl-RUNX1Dependent Repressor Complex. Cell Rep. 2013, 5, 1625−1638. (15) Cheng, D.; Yadav, N.; King, R. W.; Swanson, M. S.; Weinstein, E. J.; Bedford, M. T. Small Molecule Regulators of Protein Arginine Methyltransferases. J. Biol. Chem. 2004, 279, 23892−23899. (16) Dillon, M. B. C.; Bachovchin, D. A.; Brown, S. J.; Finn, M. G.; Rosen, H.; Cravatt, B. F.; Mowen, K. A. Novel Inhibitors for PRMT1 Discovered by High-Throughput Screening Using Activity-Based Fluorescence Polarization. ACS Chem. Biol. 2012, 7, 1198−1204. (17) Siarheyeva, A.; Senisterra, G.; Allali-Hassani, A.; Dong, A.; Dobrovetsky, E.; Wasney, G. A.; Chau, I.; Marcellus, R.; Hajian, T.; Liu, F.; Korboukh, I.; Smil, D.; Bolshan, Y.; Min, J.; Wu, H.; Zeng, H.; Loppnau, P.; Poda, G.; Griffin, C.; Aman, A.; Brown, P. J.; Jin, J.; AlAwar, R.; Arrowsmith, C. H.; Schapira, M.; Vedadi, M. An Allosteric Inhibitor of Protein Arginine Methyltransferase 3. Structure 2012, 20, 1425−1435. (18) Castellano, S.; Milite, C.; Ragno, R.; Simeoni, S.; Mai, A.; Limongelli, V.; Novellino, E.; Bauer, I.; Brosch, G.; Spannhoff, A.; Cheng, D.; Bedford, M. T.; Sbardella, G. Design, Synthesis and I



Article

(34) Irwin, J. J.; Shoichet, B. K. ZINC - A Free Database of Commercially Available Compounds for Virtual Screening. J. Chem. Inf. Model. 2005, 45, 177−182. (35) Glide, version 6.2; Schrödinger, LLC: New York, 2013. (36) Bonnefond, L.; Stojko, J.; Mailliot, J.; Troffer-Charlier, N.; Cura, V.; Wurtz, J. M.; Cianférani, S.; Cavarelli, J. Functional Insights from High Resolution Structures of Mouse Protein Arginine Methyltransferase 6. J. Struct. Biol. 2015, 191, 175−183. (37) Case, D. A.; Darden, T. A.; Cheatham, T. E., III; Simmerling, C. L.; Wang, J.; Duke, R. E.; Luo, R.; Walker, R. C.; Zhang, W.; Merz, K. M.; Robertson, B.; Wang, B.; Hayik, S.; Roitberg, A.; Seabra, G.; Kolossvary, I.; Wong, K. F.; Paesani, F.; Vanicek, J.; Liu, J.; Wu, X.; Brozell, S.; Steinbrecher, T.; Gohlke, H.; Cai, Q.; Ye, X.; Wang, J.; Hsieh, M.-J.; Cui, G.; Roe, D. R.; Mathews, D. H.; Seetin, M. G.; Sangui, C.; Babin, V.; Luchko, T.; Gusarov, S.; Kovalenko, A.; Kollman, P. A. AMBER 12; University of California: San Francisco, 2010. (38) Smith, R. D.; Engdahl, A. L.; Dunbar, J. B.; Carlson, H. A. Biophysical Limits of Protein-Ligand Binding. J. Chem. Inf. Model. 2012, 52, 2098−2106. (39) Schulz, M. N.; Landström, J.; Bright, K.; Hubbard, R. E. Design of a Fragment Library That Maximally Represents Available Chemical Space. J. Comput.-Aided Mol. Des. 2011, 25, 611−620. (40) Mitchell, L. H.; Boriack-Sjodin, P. A.; Smith, S.; Thomenius, M.; Rioux, N.; Munchhof, M.; Mills, J. E.; Klaus, C.; Totman, J.; Riera, T. V.; Raimondi, A.; Jacques, S. L.; West, K.; Foley, M.; Waters, N. J.; Kuntz, K. W.; Wigle, T. J.; Scott, M. P.; Copeland, R. A.; Smith, J. J.; Chesworth, R. Novel Oxindole Sulfonamides and Sulfamides: EPZ031686, the First Orally Bioavailable Small Molecule SMYD3 Inhibitor. ACS Med. Chem. Lett. 2016, 7, 134−138. (41) Wang, L.; Zhao, Z.; Meyer, M.; Saha, S.; Yu, M.; Guo, A.; Wisinski, K.; Huang, W.; Cai, W.; Pike, J. W.; Yuan, M.; Ahlquist, P.; Xu, W. CARM1Methylates Chromatin Remodeling Factor BAF155 to Enhance Tumor Progression and Metastasis. Cancer Cell 2014, 25, 21−36. (42) Huynh, T.; Chen, Z.; Pang, S.; Geng, J.; Bandiera, T.; Bindi, S.; Vianello, P.; Roletto, F.; Thieffine, S.; Galvani, A.; Vaccaro, W.; Poss, M. A.; Trainor, G. L.; Lorenzi, M. V.; Gottardis, M.; Jayaraman, L.; Purandare, A. V. Optimization of Pyrazole Inhibitors of Coactivator Associated Arginine Methyltransferase 1 (CARM1). Bioorg. Med. Chem. Lett. 2009, 19, 2924−2927. (43) Therrien, E.; Larouche, G.; Manku, S.; Allan, M.; Nguyen, N.; Styhler, S.; Robert, M. F.; Goulet, A. C.; Besterman, J. M.; Nguyen, H.; Wahhab, A. 1,2-Diamines as Inhibitors of Co-Activator Associated Arginine Methyltransferase 1 (CARM1). Bioorg. Med. Chem. Lett. 2009, 19, 6725−6732. (44) LigPrep, version 2.9; Schrödinger, LLC: New York, 2014. (45) Chumanov, R. S.; Kuhn, P. A.; Xu, W.; Burgess, R. R. Expression and Purification of Full-Length Mouse CARM1 from Transiently Transfected HEK293T Cells Using HaloTag Technology. Protein Expression Purif. 2011, 76, 145−153. (46) Kaniskan, H. Ü ; Szewczyk, M. M.; Yu, Z.; Eram, M. S.; Yang, X.; Schmidt, K.; Luo, X.; Dai, M.; He, F.; Zang, I.; Lin, Y.; Kennedy, S.; Li, F.; Dobrovetsky, E.; Dong, A.; Smil, D.; Min, S. J.; Landon, M.; LinJones, J.; Huang, X. P.; Roth, B. L.; Schapira, M.; Atadja, P.; BarsyteLovejoy, D.; Arrowsmith, C. H.; Brown, P. J.; Zhao, K.; Jin, J.; Vedadi, M. A Potent, Selective and Cell-Active Allosteric Inhibitor of Protein Arginine Methyltransferase 3 (PRMT3). Angew. Chem., Int. Ed. 2015, 54, 5166−5170. (47) Vedadi, M.; Niesen, F. H.; Allali-Hassani, A.; Fedorov, O. Y.; Finerty, P. J.; Wasney, G. A.; Yeung, R.; Arrowsmith, C.; Ball, L. J.; Berglund, H.; Hui, R.; Marsden, B. D.; Nordlund, P.; Sundstrom, M.; Weigelt, J.; Edwards, A. M. Chemical Screening Methods to Identify Ligands That Promote Protein Stability, Protein Crystallization, and Structure Determination. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 15835−15840. (48) Purandare, A. V.; Wan, H.; Huynh, T. N. Heterocyclic inhibitors of protein arginine methyl transferases. U.S. Patent 2006/0235037, 2006. J


Discovery of a Potent and Selective Coactivator Associated Arginine

Recommend Documents