A First-in-Class, Highly Selective and Cell-Active Allosteric Inhibitor of Protein Arginine Methyltransferase 6

Abstract

Protein arginine methyltransferase 6 (PRMT6) catalyzes monomethylation and asymmetric dimethylation of arginine residues in various proteins, plays important roles in biological processes, and is associated with multiple cancers. To date, a highly selective PRMT6 inhibitor has not been reported. Here we report the discovery and characterization of a first-in-class, highly selective allosteric inhibitor of PRMT6, (R)-2 (SGC6870). (R)-2 is a potent PRMT6 inhibitor (IC₅₀ = 77 ± 6 nM) with outstanding selectivity for PRMT6 over a broad panel of other methyltransferases and nonepigenetic targets. Notably, the crystal structure of the PRMT6−(R)-2 complex and kinetic studies revealed (R)-2 binds a unique, induced allosteric pocket. Additionally, (R)-2 engages PRMT6 and potently inhibits its methyltransferase activity in cells. Moreover, (R)-2’s enantiomer, (S)-2 (SGC6870N), is inactive against PRMT6 and can be utilized as a negative control. Collectively, (R)-2 is a well-characterized PRMT6 chemical probe and a valuable tool for further investigating PRMT6 functions in health and disease.

Graphical Abstract

INTRODUCTION

Protein arginine methyltransferases (PRMTs) catalyze monomethylation and symmetric or asymmetric dimethylation of arginine residues at histone and/or nonhistone substrates.^1,2 PRMTs are involved in multiple cellular processes, including regulation of cell cycle, cell pluripotency, estrogen-stimulated transcription, DNA repair, and modulation of pre-mRNA splicing.^3,4 Aberrant expression of PRMTs are associated with various types of cancer, such as leukemiaand breast, lung, prostate, and bladder cancers.⁵ PRMTs are commonly classified into three types.^1,2 Type I PRMTs, including PRMT1, PRMT3, PRMT4 (CARM1), PRMT6, and PRMT8, catalyze monomethylation and asymmetric dimethylation of arginine residues. Type II PRMTs, including PRMT5 and PRMT9, catalyze monomethylation and symmetric dimethylation of arginine residues. PRMT7 as the only type III PRMT catalyzes monomethylation of arginine residues. All PRMTs can methylate multiple protein substrates and some PRMTs share the same substrates. For example, H4R3 can be methylated by PRMT1, PRMT5, PRMT6, and PRMT7;^6–9 TP53 can be methylated by PRMT3 and PRMT5;^10,11 and H3R2 can be methylated by PRMT5, PRMT6, and PRMT7.^5,12 To further understand the roles of each PRMT in cell biology, highly selective and cell-active inhibitors of each PRMT are needed. In addition to being invaluable chemical tools, selective PRMT inhibitors could also serve as leads for drug discovery programs. To date, highly selective inhibitors for PRMT3,¹³ CARM1,^14,15 PRMT5,^16–20 and PRMT7²¹ have been reported. However, it remains challenging to develop highly selective inhibitors for other PRMTs, including PRMT6, due to high homology in the conserved catalytic core shared among type I PRMTs.²² PRMT6 is associated with proliferative phenotypes,²³ and it contributes to negative regulation of global DNA methylation in cancer cells.²⁴ However, its function is still poorly understood. While several PRMT6 inhibitors have been reported to date, a highly selective and cell-active PRMT6 inhibitor that can be utilized as a PRMT6 chemical probe has not yet been reported. Previously, we reported the type I PRMT pan inhibitor MS023,²⁵ PRMT6 covalent inhibitor MS117 with modest selectivity,²⁶ and PRMT4/6 dual inhibitor MS049 (Figure 1).²⁵ In addition, EPZ020411²⁷ was reported as a PRMT6 inhibitor with limited selectivity (Figure 1). Here, we report the first highly selective and cell-active allosteric inhibitor of PRMT6, (R)-2, which is one of the only few PRMT allosteric inhibitors, including our previously reported PRMT3 allosteric inhibitor, SGC707,¹³ and a recently reported PRMT5 allosteric inhibitor.²⁸ We also developed (S)-2, which is the enantiomer of (R)-2 but is inactive for PRMT6, as a negative control of (R)-2 for chemical biology studies.

Figure 1.

The structures and potencies of reported PRMT6 inhibitors

RESULTS AND DISCUSSION

Discoveries of (R)-2 and (S)-2.

Our quest for the discovery of a selective PRMT6 inhibitor began with the screening of a diverse library of 5000 compounds, from which a submicromolar inhibitor (±)-1 (IC₅₀ = 755 ± 87 nM) was identified (Figure 2A). Optically pure (R)-1 enantiomer and (S)-1 enantiomer were further chirally separated. The (R)-1 enantiomer was identified as the active isomer (IC₅₀ = 388 ± 77 nM), while the (S)-1 enantiomer was inactive (IC₅₀ > 100 μM) (Figure 2A). To discover the first highly selective chemical probe of PRMT6, we conducted structure-activity relationship (SAR) studies and successively optimized the amide group, benzodiazepine ring substituent, and pendant ring moiety of compound (±)-1. Design, synthesis, and testing of more than 60 derivatives of (±)-1 led to the identification of compound (±)-2, which contains 3,5-dimethyl phenyl and thiophene rings, with the most potent inhibitory effect of PRMT6 (IC₅₀ = 214 ± 37 nM). The synthesis of (R)-2 and (S)-2 was outlined in Scheme 1. Intermediate 3 was prepared through the addition reaction between 2,6-dimethyl-4H-benzo[d][1,3]oxazin-4-one and (3,5-dimethylphenyl)-magnesium bromide, deprotection of acetal group in the presence of hydrochloric acid, and cyclization reaction with Oxazolidine-2,5-dione, successively. Intermediate 3 was further reduced to afford diazepine in the presence of sodium cyanoborohydride, followed by amide coupling reaction with 5-bromothiophene-2-carboxylic acid to afford racemic compound (±)-2. Optically pure enantiomer (R)-2 and enantiomer (S)-2 were then obtained by chiral separation. Similar to compound (±)-1, the (R)-enantiomer of (±)-2, (R)-2, showed potent PRMT6 inhibition (IC₅₀ = 77 ± 6 nM) while the (S)-enantiomer, (S)-2, was inactive (IC₅₀ > 50 μM) (Figure 2A,B).

Figure 2.

Discovery of the PRMT6 chemical probe (R)-2. (A) Structure-activity relationship studies and chiral separation led to the discovery of (R)-2 as a potent PRMT6 inhibitor and its enantiomer, (S)-2, as a negative control. (B) IC₅₀ determination of (R)-2 and (S)-2, n = 3.

Scheme 1.

Synthesis of Compound (R)-2 and Compound (S)-2^a

^a(a) dichloromethane, −78 °C to rt; (b) HCl, MeOH/H₂O, reflux; (c) oxazolidine-2,5-dione, CF₃COOH, triethylamine, toluene, 60–85 °C; (d) NaBH₃CN, AcOH, MeOH, rt; (e) 5-bromothiophene-2-carboxylic acid, EDCI, HOAt, NMM, DMSO, rt; (f) chiral separation.

Selectivity Assessment.

Next, we assessed the selectivity of (R)-2 and (S)-2 against a total of 33 methyltransferases, including 8 PRMTs, 21 protein lysine methyltransferases (PKMTs), 3 DNA methyltransferases (DNMTs), and 1 RNA methyltransferase. (R)-2 at both 1 and 10 μM potently inhibited PRMT6, but did not significantly inhibit other 32 methyltransferases (Figure 3, Table S1). As expected, negative control (S)-2 did not significantly inhibit any of the 33 methyltransferases, including PRMT6, at 1 or 10 μM (Figure 3, Table S1). Furthermore, (R)-2 was remarkably selective for PRMT6 over a broad range of nonepigenetic targets, including G protein-coupled receptors (GPCRs), ion channels, transporters, and other enzymes at 1 μM (Table S2).

Figure 3.

Selectivity assessment of (R)-2 (SGC6870) at 1 μM (gray) and 10 μM (black) and (S)-2 (SGC6870N) at 1 μM (green) and 10 μM (blue) against 8 PRMTs, 21 protein lysine methyltransferases (PKMTs), 3 DNA methyltransferases (DNMTs), and 1 RNA methyltransferase, n = 3.

Mechanism of Action.

Typically, when the addition of an inhibitor leads to quick equilibrium, the IC₅₀ values can be determined very reproducibly regardless of small variations in the preparation of reaction mixtures, and time of inhibitor incubation with protein. However, as we further assessed the inhibitory effect of (R)-2 in different conditions, we observed significant differences in IC₅₀ values. We then investigated a possible time-dependent inhibition of PRMT6 by (R)-2 (Table S3). Inhibitory effect of (R)-2 was indeed time-dependent with more than 100-fold decrease in the IC₅₀ value after 2 h of compound preincubation with PRMT6 (Table S3). This indicates that (R)-2 binds slowly and requires significant time to reach equilibrium before steady state turnover initiated by the addition of the substrate (Figure 4A,B and Table S3). To confirm that the time-dependent inhibition of PRMT6 by (R)-2 is not due to covalent binding of the compound, we tested samples of PRMT6 incubated with (R)-2 for 2 h by mass spectrometry. No covalent modification of PRMT6 by (R)-2 was observed (Figure S1). Using 2 h preincubation, we investigated the mechanism of action of (R)-2 by determining the IC₅₀ values at various concentrations of the cofactor (Figure 4C) or substrate (Figure 4D). Our data supported a noncompetitive pattern of inhibition with respect to both the S-adenosyl methionine (SAM) cofactor and peptide substrate as no changes in IC₅₀ values were observed as the cofactor or substrate concentrations were increased (Figure 4C,D). These mechanism of action results suggest that (R)-2 may be an allosteric inhibitor binding away from the catalytic active site.

Figure 4.

Mechanism of action. (A) IC₅₀ values were determined at various compound-protein incubation times ranging from 15 to 120 min. The value obtained with no incubation time was used as a control (values are presented in Table S3). (B) Part of the plot in Figure A is magnified for better view of the decrease in the IC₅₀ value at longer incubation times. (C) IC₅₀ values of (R)-2 at increasing concentrations of SAM (up to 30 μM) at fixed concentration of peptide substrate (3 μM of H4 1–24 peptide). (D) IC₅₀ values of (R)-2 at increasing concentration of peptide (up to 8 μM) at fixed concentration of SAM (12 μM).

Cocrystal Structures.

To further investigate the mechanism of inhibition, we solved the crystal structures of PRMT6 in complex with (R)-2 (PDB: 6W6D) (Figure 5A, Table S4) and (R)-1 (PDB: 5WCF) (Figure S2A and Table S4), respectively. Structural alignment of PRMT6-(R)-2-SAM complex, and previously reported PRMT6-MS023-SAH²⁵ and PRMT6-SAH²⁹ complexes revealed (R)-2 binds in a unique pocket distinct from either the substrate or SAM binding pocket (Figure 5A,B) consistent with our kinetic data. A nine-amino acid loop consisting of residues Gly158 through Met166 flipped toward and downsized the substrate-binding pocket, creating a new pocket for accommodating (R)-2 (Figure 5B). Notably, this loop movement was highly localized, and no other large-scale changes were observed in the complex. Several key interactions were observed between (R)-2 and residues in this new site (Figure 5A). The endo-amide on the diazepine ring formed a hydrogen bond with the backbone nitrogen of Gly158, and the exo-amide formed a hydrogen bond with the backbone nitrogen of Gly160. In addition, the thiophene group formed a T-shaped π-π stacking interaction with Tyr159, and another T-shaped π-π stacking interaction was also observed between the dimethylphenyl group and Trp156. Furthermore, the methylphenyl portion of the benzodiazepine moiety points to a hydrophobic pocket consisting of Leu267, Val271, Leu343, and Pro345 residues. Residue Ala321 is one of the key components creating the newly formed binding pocket for the bromothiophene group of (R)-2 and bromofuran group of (R)-1 (Figure S2B). We mutated it to Ile, Gln, and Met bearing bulkier side chains, respectively. While these directed mutations did not significantly impact the substrate or SAM binding affinity to PRMT6 or PRMT6 catalytic activity itself (Table 1), they did impair PRMT6 inhibitory potency of (R)-1 by 4–12-fold (Table 2). The structural alignment of PRMT6-(R)-2 and PMRT6-(R)-1 complexes revealed that two inhibitors possess the same binding mode (Figure S2B). Taken together, these results show both (R)-2 and (R)-1 bind to the induced PRMT6 allosteric pocket. As previously reported, electron donating or withdrawing substituents of aromatic rings impact the T-shaped π-π stacking interactions.³⁰ Compared with (R)-1, the enhanced inhibitory potency of (R)-2 could be partially due to higher electron density of thiophene and dimethylphenyl groups of (R)-2 than furan and fluorophenyl of (R)-1, as a result of stronger T-shaped π-π stacking interactions between (R)-2 and PRMT6 residues Tyr159 and Trp156. Collectively, both structural and kinetic data consistently revealed that (R)-2 is an allosteric PRMT6 inhibitor.

Figure 5.

Cocrystal structure of PRMT6 in complex with (R)-2. (A) Cocrystal structure of PRMT6 (tinted blue) in complex with (R)-2 (green), and SAM (orange), (PDB: 6W6D). Dashed yellow lines indicate key hydrogen bonds. (B) Structural alignments of complexes PRMT6-(R)-2-SAM (cyan), PRMT6-MS023-SAH (blue) (PDB: 5E8R) and PRMT6-SAH (magenta) (PDB: 4C05). Movement of loop consisting of nine residues, EWMGYGLLH, indicates that (R)-2 binds to a novel induced allosteric pocket of PRMT6.

Table 1.

Activity Assessment on PRMT6-Wild Type (1–375) and Its Mutants (17, 19, 20)

parameters	wild type	PRMT6−17	PRMT6−19	PRMT6−20
mutation		A321I	A321Q	A321M
Km B-H4(1–24), μM	0.17	0.4	0.18	0.17
kcat, h−1	42	85	20	29
Km SAM, μM	2	4	3	4

Table 2.

Inhibitory Potency of (R)-1 against PRMT6-Wild Type (1–375) and Its Mutants (17, 19, 20)^a

protein	IC50 (μM)	hill slope
PRMT6-Wt	4	1.3
PRMT6–17	46	1
PRMT6–19	17	1.5
PRMT6–20	41	0.8

^aThe IC₅₀ values in Table 2 were generated using 20 min preincubation.

Evaluation in Cellular Assays.

Next, given that (R)-2 is a highly selective PRMT6 inhibitor in biochemical assays, we did not assess its cellular selectivity for PRMT6, but evaluated its cellular potency on ectopically expressed PRMT6 in HEK293T cells and used a catalytically inactive PRMT6 mutant (V86K/D88A) as a positive control. HEK293T cells were treated with (R)-2 (Figure 6A) or (S)-2 (Figure 6B) at titrated concentrations for 20 h. (R)-2 potently and concentration-dependently reduced cellular levels of H3R2me2a (IC₅₀ = 0.9 ± 0.1 μM) (Figure 6C) and H4R3me2a (IC₅₀ = 0.6 ± 0.1 μM) (Figure 6D), both of which are known substrates of PRMT6.⁵ Furthermore, (R)-2 did not show significant toxicity to HEK293T cells at up to 10 μM (Figure 6A). As expected, ectopical expression of the PRMT6 catalytically inactive V86K/D88A mutant led to near complete reduction of H3R2me2a and H4R3me2a markers (Figure 6A,B). Consistent with its poor potency against PRMT6 in the biochemical assay, (S)-2 did not significantly reduce cellular levels of H4R3me2a or H3R2me2a at up to 10 μM (Figure 6B). However, at 30 μM, significant toxicity to HEK293T cells was observed (Figure 6B).

Figure 6.

Inhibition of PRMT6-dependent H3R2 and H4R3 asymmetric di-methylation in cells. HEK293T cells were transfected with Flag-tagged PRMT6 and treated with indicated compounds for 20 h. The flag-tagged PRMT6 catalytically inactive mutant V86K/D88A (mut) was used as a positive control. (A) Western blot representation of the effect of (R)-2 on PRMT6 activity. (B) Western blot representation of the effect of (S)-2 on PRMT6 activity. (C) IC₅₀ determination of (R)-2 and (S)-2 at reducing H3R2me2a. The graph represents nonlinear fit of H3R2me2a fluorescence intensities normalized to intensities of H3, n = 4. (D) IC₅₀ determination of (R)-2 and (S)-2 at reducing H4R3me2a. The graph represents nonlinear fit of H4R3me2a fluorescence intensities normalized to intensities of H4, n = 3.

Finally, we further investigated the effect of (R)-2 and (S)-2 on cell growth in three different cell lines, HEK293T (embryonic kidney), PNT2 (prostate), and MCF-7 (breast cancer). Neither (R)-2 nor (S)-2 showed any significant toxicity to these three cell lines at concentrations up to 10 μM (Figure 7A,B) after 3 days treatment. Therefore, (R)-2 and (S)-2 can be utilized in cellular studies at up to 10 μM concentration.

Figure 7.

Inhibitory effect of (R)-2 and (S)-2 on cell viability (A) Effect of (R)-2 on MCF-7, PNT2, and HEK293T cell viability, n = 3. (B) Inhibitory effect of (S)-2 on MCF-7, PNT2, and HEK293T cell viability, n = 3.

CONCLUSIONS

In conclusion, we discovered the first highly selective, potent and cell-active PRMT6 allosteric inhibitor, (R)-2. The crystal structure of the PRMT6-(R)-2 complex and kinetic studies revealed that (R)-2 binds to a unique, induced allosteric pocket, rendering this inhibitor highly selective for PRMT6 over 32 other methyltransferases, as well as a broad range of common drug targets. In addition, (R)-2 can significantly inhibit PRMT6 activity in cells with submicromolar potency, but did not display significant cellular toxicity at concentrations up to 10 μM. Furthermore, we demonstrated that the enantiomer of (R)-2, (S)-2, showed very poor inhibitory effect on PRMT6 activity in both biochemical and cell-based assays and can be used as a negative control. Collectively, (R)-2 is a valuable chemical probe of PRMT6 and together with (S)-2 is an excellent tool compound to further investigate physiological and pathophysiological functions of PRMT6.

EXPERIMENTAL SECTION

Chemistry General Procedure.

All commercially available chemical reagents were directly used without further purification. A Teledyne ISCO CombiFlash Rf⁺ instrument equipped with a variable wavelength UV detector and a fraction collector was used to conduct flash column chromatography. RediSepRf HP C18 and flash silica columns were used for purification. The purity of each tested compound was >95%, determined based on the HPLC spectrum, which was generated on an Agilent 1200 series system with a DAD detector and a 2.1 mm × 150 mm Zorbax 300SB-C18 5 μm column. 0.8 μL of the sample (methanol solution) was injected onto a C18 column at room temperature, and the flow rate of eluents was 0.4 mL/min. The eluents were as follows: water containing 0.1% formic acid was designated as solvent A, while acetonitrile containing 0.1% formic acid was designated as solvent B. The linear gradient was set such that 1% B was used from 0 to 1 min, 1–99% B from 1 to 4 min, and 99% B from 4 to 8 min. High-resolution mass spectra (HRMS) data were acquired in positive ion mode using an Agilent G1969A API-TOF with an electrospray ionization (ESI) source. Nuclear magnetic resonance (NMR) spectra were acquired on Bruker Avance-III 800 MHz spectrometer (800 MHz ¹H NMR, 201 MHz ¹³C NMR). Chemical shifts are reported in ppm (δ). The enantiomers were separated by prep SFC using an AS-H column (2 × 25 cm) and mobile phase of 25% MeOH/CO₂ with 0.1% DEA at 70 mL·min⁻¹ (100 bar).

5-(3,5-dimethylphenyl)-7-methyl-1,3-dihydro-2H-benzo-[1,4]diazepin-2-one (3).

To the solution of 2,6-dimethyl-4H-benzo[d][1,3]oxazin-4-one (3.5 g, 20 mmol) in 60 mL of dichloromethane at −78 °C, was added dropwise 3,5-dimethylphenylmagnesium bromide (0.5 M in THF, 48 mL, 24 mmol). The resulting mixture was slowly warmed to room temperature and stirred overnight. Then, 30 mL of saturated aqueous ammonium chloride was carefully added, and the mixture was extracted with dichloromethane (60 mL × 3). The organic phase was dried over anhydrous sodium sulfate and concentrated under reduced pressure. The resulting residue was dissolved in 60 mL of methanol and 10 mL of hydrochloric acid (36.5–38.0%). The solution was heated to reflux for 6 h. The cooled solution was then carefully basified to pH = 8 by saturated aqueous sodium bicarbonate and concentrated under reduced pressure. The resulting mixture was extracted with ethyl acetate (50 mL × 3). The organic phase was dried over anhydrous sodium sulfate and concentrated under reduced pressure. The resulting residue was dissolved in 80 mL of toluene. To this solution, was added trifluoroacetic acid (1.7 mL, 22 mmol) and added portion wise oxazolidine-2,5-dione (2.6 g, 26 mmol). The reaction was heated to 60 °C for 30 min. Then triethylamine (3.1 mL, 22 mmol) was added, and the mixture was heated to 80 °C for another 30 min. To the cooled mixture was carefully added 20 mL of saturated aqueous sodium bicarbonate and the mixture was extracted with ethyl acetate (50 mL × 3). The organic phase was dried over anhydrous sodium sulfate and concentrated under reduced pressure. The resulting residue was purified by flash chromatography on silica gel column with eluent (EtOAc/hexane: 0–50%) to give white solid (2.7 g, yield 49%). ¹H NMR (800 MHz, DMSO-d₆) δ 10.43 (s, 1H), 7.36 (d, J = 8.4 Hz, 1H), 7.15 (d, J = 8.3 Hz, 1H), 7.11 (s, 1H), 7.06 (s, 2H), 7.01 (s, 1H), 4.46–3.66 (m, 2H), 2.27 (s, 6H), 2.25 (s, 3H). ¹³C NMR (201 MHz, DMSO-d₆) δ 170.9, 170.3, 139.8, 137.7, 137.7, 132.8, 132.2, 131.9, 130.8, 127.5, 126.9, 121.5, 57.3, 21.3, 20.7. MS (ESI) m/z 279.2 [M + H]⁺. HRMS (TOF) m/z [M + H]⁺ calcd for C₁₈H₁₉N₂O⁺ 279.1492, found 279.1509.

4-(5-bromothiophene-2-carbonyl)-5-(3,5-dimethylphenyl)-7-methyl-1,3,4,5-tetrahydro-2H-benzo[e][1,4]diazepin-2-one ((±)-2).

To the solution of 5-(3,5-dimethylphenyl)-7-methyl-1,3-dihydro-2H-benzo[e][1,4]diazepin-2-one (2.7 g, 9.7 mmol) in 40 mL of methanol was added portion wise sodium cyanoborohydride (1.8 g, 28 mmol) and acetic acid (5.5 mL, 97 mmol). The mixture was stirred at room temperature overnight, and the volatile was removed under reduced pressure. The residue was carefully treated by 20 mL of saturated aqueous sodium bicarbonate and extracted with ethyl acetate (50 mL × 3). The organic phase was dried over anhydrous sodium sulfate and concentrated under reduced pressure. To the solution of residue above in 50 mL of dimethyl sulfoxide was added 5-bromothiophene-2-carboxylic acid (2 g, 10 mmol), EDCI (2.4 g, 13 mmol), HOAt (1.7 g, 13 mmol), and N-methylmorpholine (2.8 mL, 26 mmol). The resulting solution was stirred at room temperature for 8 h and extracted with ethyl acetate (150 mL) and water (100 mL). The organic phase was washed by another 100 mL of water for twice and concentrated. The residue was purified by flash chromatography on silica gel column with eluent (EtOAc/hexane: 0–50%) to give a pale yellow solid (2.7 g, yield 59%).

The enantiomers were separated by prep SFC using an AS-H column (2 × 25 cm) and mobile phase of 25% MeOH/CO₂ with 0.1% DEA at 70 mL.min⁻¹ (100 bar). Then, 15 mg aliquots in EtOH:DCM (1:1; 1 mL) were injected and the absorbance was monitored at 220 nm. Analytical SFC was used to determine the purity of combined separated fractions using an AS-H column (0.46 × 25 cm) and mobile phase of 40% MeOH/CO₂ with 0.1% DEA at 3 mL·min⁻¹ (120 bar). Peak 1 eluting at 2.44 min using the analytical conditions corresponded to (S)-2, whereas Peak 2 eluting at 3.34 min corresponded to (R)-2. Lastly, 1.0 g of each enantiomer was obtained from chiral separation of 2.2 g of racemic material.

(R)-2: ¹H NMR (800 MHz, DMSO-d₆) δ 10.19–9.82 (m, 1H), 7.73–7.49 (m, 1H), 7.41–7.33 (m, 2H), 7.19 (d, J = 8.1 Hz, 1H), 7.00–6.81 (m, 2H), 6.67 (s, 2H), 6.59–6.31 (m, 1H), 4.48–3.82 (m, 2H), 2.32 (s, 3H), 2.19 (s, 6H). ¹³C NMR (201 MHz, DMSO-d₆) δ 168.1, 161.5, 140.2, 139.8, 137.8, 134.9, 134.4, 132.5, 131.6, 130.3, 129.8, 129.1, 124.42, 124.41, 117.6, 62.6, 51.4, 21.5, 20.7. HPLC >95% pure, t_R = 5.39 min; MS (ESI) m/z 469.1 [M + H]⁺. HRMS m/z [M + H]⁺ calcd for C₂₃H₂₂BrN₂O₂S⁺ 469.0580, found 469.0574. [α]_D²⁰ +275.2° (c 4.55, MeOH).

(S)-2: ¹H NMR (800 MHz, DMSO-d₆) δ 10.22–9.85 (m, 1H), 7.60 (s, 1H), 7.49–7.10 (m, 3H), 7.01–6.80 (m, 2H), 6.67 (s, 2H), 6.60–6.27 (m, 1H), 4.55–3.75 (m, 2H), 2.32 (s, 3H), 2.18 (s, 6H). ¹³C NMR (201 MHz, DMSO-d₆) δ 168.1, 161.5, 140.2, 139.8, 137.8, 134.9, 134.4, 132.5, 131.6, 130.4, 129.8, 129.1, 124.42, 122.40, 117.6, 62.6, 51.3, 21.5, 20.7. HPLC >95% pure, t_R = 5.37 min; MS (ESI) m/z 469.1 [M + H]⁺. HRMS m/z [M + H]⁺ calcd for C₂₃H₂₂BrN₂O₂S⁺ 469.0580, found 469.0590. [α]_D²⁰ −268.2° (c 4.77, MeOH).

Due to the amide rotamers of (R)-2 and (S)-2, which can interconvert at ambient temperature, broadening of NMR peaks was observed.

Crystallization and Structure Determination.

Crystallization.

Human PRMT6 protein was expressed and purified according to the previously published protocol.³¹ PRMT6 at 5.6 mg/mL was mixed with fivefold molar excess of S-adenosyl-L-homocysteine (SAH) or S-Adenosyl methionine (SAM) and 3-fold molar excess of (R)-1 or (R)-2 (dissolved from a previously prepared 100 mM DMSO stock solution). Diffraction quality crystals were obtained by setting a 96-well vapor-diffusion sitting drops at room temperature, in a precipitant solution containing 0.1 M sodium malonate pH 7.0, 12% (w/v) PEG 3350. The PRMT6-(R)-2/(R)-1 crystal was cryo-protected by immersing it into a precipitant solution supplemented with 10% glycerol and then into paratone, and cryo-cooled in liquid nitrogen.

Structure Determination.

X-ray diffraction data for PRMT6 + (R)-2/(R)-1 were collected at 100 K at beam line 24IDE of Advanced Photon Source (APS), Argonne National Laboratory and CMCF 08ID-1 of Canadian Light Source. Both data sets were processed using the HKL-3000 suite.³² The structure of PRMT6 + (R)-2 was directly refined against PDB entry 6W6D as template. The structure of PRMT6 + (R)-1 was directly refined against PDB entry 5WCF as template. REFMAC was used for structure refinement.³³ GRADE was used to generate all restrains for compound refinement.³⁴ Graphics program COOT was used for model building and visualization.³⁵ Molprobity was used for structure validation.³⁶

PRMT6 Mutation.

Three PRMT6 mutants were generated by mutating the alanine residue at position 321 of the wild type PRMT6 (PRMT6-WT) (residues 1–375) to isoleucine, glutamine and methionine of mutants 17, 19, and 20, respectively.

Kinetic Characterization of PRMT6-WT and PRMT6-Mutants.

Using the optimized assay conditions (25 nM of PRMT6-wt or PRMT6-mutants, 20 mM Tris-HCl pH 7.5, 0.01% Triton X-100, and 10 mM DTT), the kinetic parameters were determined for biotinylated H4 (B-H4) (residues 1–24) by varying the peptide concentration (0–1.75 μM) and keeping the AdoMet at a saturation concentration of 10 μM. Apparent Km values were determined for AdoMet in reactions with 2 μM of B-H4 by varying the AdoMet concentration (up to 12 μM). The reactions were started by the addition of a mixture of ³H-AdoMet and unlabeled AdoMet. The reaction mixtures were incubated for 20 min at 23 °C and quenched by adding 10 μL of 7.5 M guanidinium hydrochloride. Ten microliters of the quenched reaction mixture was spotted onto streptavidin-coated membrane squares (SAM2® Biotin capture membrane, Promega). The membrane was washed three times in 2 M NaCl for 2 min each time, and then in water three times for 30 s each. The membrane was dried at 50 °C for 1 h, and then each spotted square was cut from the membrane and placed in a scintillation vial. The amount of methylated peptide was quantified by tracing the radioactivity (cpm) as counted by a TriCarb liquid scintillation counter (PerkinElmer Life Sciences).

IC₅₀ Determination of (R)-1 against PRMT6-WT and PRMT6-Mutants.

IC₅₀ values of (R)-1 (data summarized in Table 2) were determined using the radiometric method.³⁷ The preincubation time was 20 min.

Biochemical Assays.

IC₅₀ values inhibiting PRMT6 activity were also determined using the radiometric method.³⁸ The preincubation time was 2 h for all biochemical assays except the ones for Table 2 as noted above.

Mass Spectrometry Analysis for Assessing the Covalent Binding.

PRMT6 was incubated with 20 molar excess of the compound (R)-2 for 1 h at RT, followed by adding 0.1% trifluoroacetic acid (aq.). The samples were separated over a HPLC column over a 5–95% acetonitrile/water gradient and analyzed using an Agilent LC/MSD Time-of-Flight Mass Spectrometer equipped with an electrospray ionization source.

Selectivity Assays.

The effect of compounds on 33 methyltransferase (MT) activities was assessed as previously described.³⁹ Selectivity assay of 44 nonepigenetic targets (kinases, GPCRs, ion channels, and transporters) were conducted by Eurofins (https://www.eurofinsdiscoveryservices.com/cms/cms-content/services/adme-tox).

Cellular PRMT6 Assay.

HEK293T cells were grown in 12-well plates (2e5cells/well) in DMEM supplemented with 10% FBS, penicillin (100 units/mL), and streptomycin (100 μg/mL). Next day cells were transfected with FLAG-tagged PRMT6/mutant V86K, D88A PRMT6 (1 μg of DNA per well) using jetPRIME transfection reagent (Polyplus-Transfection), following manufacturer instructions. After 4 h media were removed, and cells were treated with compounds. After 20 h, the media was removed, and cells were lysed in 100 μL of lysis buffer (in mM: 20 Tris-HCl pH = 8, 150 NaCl, 1 EDTA, 10 MgCl₂, 0.5% Triton-X100, 12.5 U/mL benzonase (Sigma), complete EDTA-free protease inhibitor cocktail (Roche). After 1 min incubation at RT, SDS was added to the final 1% concentration. 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-H4 (1:1000, Abcam #174628), rabbit anti-H4R3me2a (1:1000 Active Motif #39705), mouse anti-FLAG (1:5000, Sigma #F1804), mouse anti-H3 (1:1000, Abcam #174628), and rabbit anti-H3R2me2a (1:1000, Millipore #04–808) in blocking buffer o/n at 4 °C. After five washes with 0.1% Tween 20 PBS, the blots were incubated with goat-anti rabbit (IR800 conjugated, LiCor #926–32,211) and donkey anti-mouse (IR 680, LiCor #926–68,072) 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.

Cell Viability Assay.

1 × 10⁴ cells (HEK293T, PNT2 and MCF7) were seeded on 96-well plates. After 24 h, the serial diluted compounds ((R)-2 and (S)-2) from 10 μM were treated for 3 days. Cell viability was evaluated by CCK-8 (Cell counting kit-8, WST-8). Briefly, 10 μL/well of CCK-8 was treated and then incubated for 4 h at 37 °C. The absorbance was recorded by Infinite F PLEX plate reader (TECAN, Morrisville, NC, USA) at 450 nm. GraphPad Prism 8 was used for data analysis, and the experimental results are shown as the average ± SD from three replicate experiments.

ABBREVIATIONS USED

PRMT

protein arginine methyltransferase

SAM

S-adenosylmethionine

TFA

trifluoroacetic acid

EDCI

1-ethyl-3-(3-dimethylaminopropyl)carbodiimide

HOAt

1-hydroxy-7-azabenzo-triazole

NMM

N-methylmorpholine

NMP

N-Methyl-2-Pyrrolidone

DMSO

dimethyl sulfoxide

ESI

electrospray ionization

H3R2me2a

H3R2 asymmetric dimethylation

H4R3me2a

H4R3 asymmetric dimethylation