- Email: [email protected]
Identification of LEM-14 inhibitor of the oncoprotein NSD2
Biochemical and Biophysical Research Communications xxx (xxxx) xxx
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Identification of LEM-14 inhibitor of the oncoprotein NSD2 Yunpeng Shen a, e, 1, Masayo Morishita a, b, 1, Doohyun Lee c, d, Shinae Kim c, d, Taeho Lee c, Damiaan E.H.F. Mevius a, e, Yeonjeong Roh a, Eric di Luccio a, * a
Department of Genetic Engineering, School of Life Sciences, College of Natural Sciences, Kyungpook National University, Daegu, 41566, Republic of Korea Institute of Agricultural Science and Technology, Kyungpook National University, Daegu, 41566, Republic of Korea c College of Pharmacy, Research Institute of Pharmaceutical Sciences, Kyungpook National University, Daegu, 41566, Republic of Korea d New Drug Development Center, Daegu-Gyeongbuk Medical Innovation Foundation, Daegu, 41061, Republic of Korea e Department of Food Biomaterials, College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, Republic of Korea b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 29 October 2018 Accepted 6 November 2018 Available online xxx
The NSD family (NSD1, NSD2/MMSET/WHSC1, and NSD3/WHSC1L1) are histone lysine methyltransferases (HMTases) essential for chromatin regulation. The NSDs are oncoproteins, drivers of a number of tumors and are considered important drug-targets but the lack of potent and selective inhibitors hampers further therapeutic development and limits exploration of their biology. In particular, MMSET/NSD2 selective inhibition is being pursued for therapeutic interventions against multiple myeloma (MM) cases, especially in multiple myeloma t(4;14)(p16.3;q32) translocation that is associated with a significantly worse prognosis than other MM subgroups. Multiple myeloma is the second most common hematological malignancy, after non-Hodgkin lymphoma and remains an incurable malignancy. Here we report the discovery of LEM-14, an NSD2 specific inhibitor with an in vitro IC50 of 132 mM and that is inactive against the closely related NSD1 and NSD3. LEM-14-1189, a LEM-14 derivative, differentially inhibits the NSDs with in vitro IC50 of 418 mM (NSD1), IC50 of 111 mM (NSD2) and IC50 of 60 mM (NSD3). We propose LEM-14 and derivative LEM-14-1189 as tools for studying the biology of the NSDs and constitute meaningful steps toward potent NSDs therapeutic inhibitors. © 2018 Elsevier Inc. All rights reserved.
Keywords: Epigenetic therapy of cancer Multiple myeloma Histone-lysine methyltransferase NSD2 Inhibitors Drug-design
1. Introduction The nuclear receptor-binding SET domain (NSD) family consists of NSD1, NSD2/MMSET/WHSC1 and NSD3/WHSC1L1 histone methyltransferases (HMTases). The NSD family are important in regulating the chromatin, evidenced by their implications in developmental syndromes and numerous cancers [1e5]. Chromatin regulation by NSD2 is mediated through H3K36 and H4K20 methylations, although in vitro its substrate specificity is broaden to H3K4, H3K9 H3K27, H3K36, H3K79, and H4K20, underlying an interesting regulatory mechanism that is still unknown [6e8]. Overexpression of NSD2 increases cellular proliferation,
Abbreviations: NSD, nuclear receptor binding SET domain; MMSET, Multiple Myeloma SET domain; HMTase, Histone lysine Methyltransferase; AdoMET, Sadenosylmethionine. * Corresponding author. E-mail addresses: [email protected], [email protected] (E. di Luccio). 1 Both authors contributed equally to this study.
chromosomal accessibility, and alteration of regulation of gene expression levels [7]. Whereas, knock-down of NSD2 affects cell adhesion, inhibits proliferation, and induces apoptosis [9]. Furthermore, NSD2 has roles in DNA damage repair being recruited to double strand-breaks (DSBs) with an increase of H4K20 methylation level at DSBs [8]. Additionally, expression of REIIBP, a short NSD2 isoform, leads to elevated H3K27 methylation and recruits histone deacetylases (HDACs) [10]. REIIBP might also be able to methylate H3K79, a mark previously thought to be methylated exclusively by DOT1L [11]. NSD2 haploinsufficiency results in the developmental WolfHirschhorn syndrome [12]. NSD2 abnormal expression level is associated with tumor aggressiveness and is implicated in more than 20 cancers [12,13]. Specifically, abnormal overexpression of NSD2 is a hallmark of the t(4;14)(p16.3;q32) translocation in multiple myeloma that is associated with a poor prognosis. Multiple Myeloma, accounts for 20% of hematological malignancy deaths and there is no cure. Inhibition of the NSDs, especially NSD2 represents a new avenue for cancer therapies, especially in multiple
https://doi.org/10.1016/j.bbrc.2018.11.037 0006-291X/© 2018 Elsevier Inc. All rights reserved.
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
2
Y. Shen et al. / Biochemical and Biophysical Research Communications xxx (xxxx) xxx
myeloma cases [5,7,13,14]. NSDs inhibitors are needed, but HMTase inhibitors remain scarce [15,16]. MCTP39, Sinefungin and an N-Alkyl Sinefungin derivative have been identified as NSD2 inhibitors with in vitro IC50 of 3 mM, 26 mM and 3.3 mM, respectively but inhibition data on NSD1 and NSD3 is unknown [17,18]. PTD2, a norleucine-containing peptide derived from the histone H4 sequence inhibits both NSD2 and NSD3 with IC50 of 22 mM and 3.2 mM, respectively [14]. PTD2 inhibition on NSD1 is unknown. BIX-01294 and BIX-01338 are inhibitors of G9a/G9a-like protein (GLP) H3K9 HMTase. BIX-01294 inhibits GLP and G9a with IC50 values of 0.7 and 1.9 mM, respectively and also inhibits NSD1, NSD2, and NSD3 with IC50 values of 112, 41, and 95 mM, respectively [19,20]. Taken together, the selective inhibition of each NSD and especially NSD2 remains to be achieved. The NSDs operate in distinct pathways but share a highly conserved SET domain, which is challenging for drug-design (Fig. S1) [2,4]. The catalytic SET domain (Su(var)3-9, Enhancer-ofzeste and Trithorax (SET)) is composed of two binding pockets, the histone-tail binding/active site and the coenzyme S-adenosyl methionine (SAM/AdoMet) site (Fig.1 A&D). The histone lysine substrate extends into a narrow channel linking the two pockets. Inhibition of the methyltransferase activity can primarily be achieved by targeting either binding pocket. Previously, we showed that pan inhibitors such as BIX-01294 could bind to the active site and the coenzyme site of all NSDs with however a slightly higher affinity toward the active site than the coenzyme site [20]. The AdoMet binding site is also highly conserved across
methyltransferases. However, the histone-tail binding area is heterogeneous at the sequence level while having a scaffold structurally conserved, which may offer opportunities for achieving selective inhibition (Fig. S1&S2) [2,4]. In this study, we report the discovery by virtual ligand screening of LEM-14, the first NSD2 specific inhibitor amongst the NSDs (Fig. 1& Table 1). LEM-14 is a hit inhibitor with an in vitro IC50 of 132 mM against H3K36 methylation by NSD2 and has very weak activity against NSD1 and has no activity against NSD3 (Table 1). We further designed, produced and investigated 14 derivatives of LEM-14 (Fig. S3 & Table 1). Derivative LEM-14-1189 differentially inhibits the NSDs with an in vitro IC50 of 418 mM (NSD1), IC50 of 111 mM (NSD2) and IC50 of 60 mM (NSD3) (Table 1).
2. Materials and methods 2.1. Cloning, protein expression, and protein purification The cloning of the SET domains of human NSD1, NSD2, and NSD3 genes and the expression and purification of NSD proteins were described previously [20]. The resultant NSD1-SET (309 a.a., 35.1 kDa), NSD2-SET (299 a.a., 34.0 kDa), and NSD3-SET (309 a.a., 35.0 kDa) were proven to be stable and retained catalytic properties for an extended period of time when stored at 80 C and confirmed by SDS-PAGE and Coomassie brilliant blue G-250 staining (Fig. 1 B).
Fig. 1. Schematic NSD2 and structure of LEM-14, LEM-14e1189 and LEM-06. (A) Schematic of the primary structure of NSD2: PWWP domain; PHD zinc fingers domain; SET histone methyl transferase (HMTase) with the AWS and postSET domains. The regulatory loop closing onto the histone-binding site is indicated in red. (B) Bacterially expressed and purified NSD1, NSD2, and NSD3 SET domain used for in vitro assays were detected by SDS-PAGE and Coomassie blue staining. (C) Structure of LEM-14. (D) Structure of NSD2 catalytic domain bound with LEM-14 into the active site. The overall structure and molecular surface of the loop-opened NSD2 SET domain model bound with the computed most favorable docked conformation of LEM-14 (carbon atoms colored in yellow) in the active site. AdoHcy (carbon atoms colored in green) is shown in the coenzyme site. The negative-charged surface is indicated as red and positive charged as blue. LEM-14 docking conformations are focused and magnified in boxes. (E) Structure of LEM-14-1189. (F) Structure of LEM-06. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
Y. Shen et al. / Biochemical and Biophysical Research Communications xxx (xxxx) xxx
3
Table 1 Biological in vitro IC50 (mM) for H3K36 mono-methylation with histone H3 as substrate; n. a. indicates no inhibitory activity; >1000 indicates a weak inhibition with an IC50 above 1000 mM. MW: molecular weight (Da); cLogP: octanol/water partition coefficient; TPSA: total polar surface area; Rot. B.: number of rotatable bonds; Rule of 5: The number of Lipinski's rule of 5 violations for each molecule. Molecule
NSD1 IC50 (mM)
NSD2 IC50 (mM)
NSD3 IC50 (mM)
MW (Da)
TPSA (Å2)
cLogP
Rot. B.
Rule of 5
Ref.
BIX-01294 LEM-06 LEM-14 LEM-14-1185 LEM-14-1187 LEM-14e1189 LEM-14-1190 LEM-14-1191 LEM-14-1192 LEM-14-1193 LEM-14-1195 LEM-14-1196 LEM-14-1199 LEM-14-1200 LEM-14-1202 LEM-14-1203 LEM-14-1204
112 ± 57 >1000 >1000 n. a. n. a. 418 ± 111 >1000 >1000 >1000 >1000 >1000 >1000 >1000 >1000 >1000 n. a. n. a.
41 ± 2 890 ± 250 132 ± 50 n. a. n. a. 111 ± 11 >1000 >1000 >1000 >1000 >1000 >1000 >1000 >1000 >1000 n. a. n. a.
95 ± 53 >1000 n. a. n. a. n. a. 60 ± 11 >1000 >1000 >1000 >1000 >1000 >1000 >1000 >1000 >1000 n. a. n. a.
492.6 492.7 478.6 334.4 476.5 683.8 465.5 465.6 493.6 479.6 465.6 493.6 478.6 492.5 492.5 406.5 307.4
68.7 58.7 93.2 99.7 91.3 141.5 113.5 75.4 75.4 92.5 75.4 75.4 91.3 108.4 108.4 108.9 98.2
3,9 2.4 1.7 0.7 2.8 1.15 2.7 2.7 3.3 2.3 2.7 3.3 2.2 1.9 1.9 2.4 0.9
7 4 3 3.9 4.3 9.7 7.1 5.3 6.3 5.3 5.3 6.3 3.8 3.8 3.8 5.4 3.9
0 0 0 0 0 2 0 0 0 1 0 0 0 0 0 0 0
[20] [1]
2.2. Compounds Compounds were purchased from MolPort (http://www. molport.com/) and prepared in dimethyl sulfoxide (DMSO) > 99.9% pure (D8418 Sigma). 2.3. Histone methyltransferase inhibitor assay Histone methyltransferase activity of NSD2-SET on monomethylation of H3K36 for the dose-response curve was measured by colorimetric quantification kits (P-3046, Epigentek, USA) following the manufacturer's protocol as previously described. Methyltransferase activity of NSD1-SET, NSD2-SET, and NSD3-SET treated with or without LEM-14 and its derivatives were measured with Methyltransferase Activity Assay kits (ADI-907-025, Enzo Life Sciences, USA) following the manufacturer's instruction. Assays were done in triplicate individual experiments. The results were normalized against the control that does not contain any enzymes. GraphPad Prism 7.0 was used to process the data, plot dose-response curves and calculate IC50 values. 2.4. Molecular modeling of the closed and open conformation of NSD2-SET Homology models of NSD1 (UniProt: Q96L73), NSD2 (UniProt: Q9BZ95) and NSD3 (UniProt: Q9BZ95) were built, refined and validated as we previously described [2]. Initial NSDs-SET build models are in a closed-conformation with the flexible regulatory loop blocking the substrate from entering the active site. The openconformation model of the catalytic domain was used for the docking of small molecules following the methods we previously established [4]. 2.5. Docking model of LEM-14 and derivatives to the NSDs The docking of LEM-14 into the active site and the coenzyme site of the open-conformation models of NSD1, NSD2 and NSD3 where the regulatory loop is in an active state. S-Adenosyl-L-homocysteine (AdoHcy) was manually placed at the coenzyme site openconformation models docked with LEM-14 at the active site, which was guided by superposition of the crystal structures of NSD1 (PDB:3OOI) and NSD3 (PDB:4YZ8) as well as NSD SET models. The NSD-LEM-14 model complexes were prepared with Autodock
Tools 1.5.6 with default settings. All bonds of LEM-14 were treated as rotatable bonds except for amine bonds. Autodock Vina was used for docking. Results from the rigid docking were utilized to assess the interacting amino acid residues that were selected for a following flexible docking. Due to the absence of NSDs structure data, the interaction prediction between NSDs and histone tail peptide was achieved by using GalaxyPepDock. The top ten predicted Set7-histone tail peptide complex structures were analyzed. The predicted structure (the histone tail peptide GGVKKPHR as input) used for further experiment was selected in terms of the highest TM-score and RMSD. 2.6. Virtual ligand screening and selection of the compounds AutoDock Vina was used for virtual ligand screening with the “Clean Drug-Like compounds” downloaded from the ZINC docking database version 11. Hydrogen atoms on the target were added using PDB2PQR program in conjunction with the CHARMM force field. All the ligands were prepared with the AutoDock tools (ADT) v1.5.4 r29. A subset of 40,325 molecules (190 Da < MW < 500 Da) was docked and the top 3000 results were analyzed and filtered with FieldView from Cresset (Herts, AL7 3AX, United Kingdom). For each molecule, the top 9 binding conformations were retained. The top docking solutions were manually inspected. LEM-14 (ZINC ID: 23116671) was purchased from Molport (cat.: 005-735-962). Selected molecules for in vitro inhibition are named LEM. 2.7. Synthesis of LEM-14 derivatives LEM-14 derivatives (Fig. 3) were synthesized following schemes described in supplementary materials (Scheme 1-5). All chemicals were reagent grade and used as purchased. Flash column chromatography was carried out on Merck silica gel 60 (230e400 mesh). 1H NMR spectra were recorded in d units relative to deuterated solvent as internal reference by a Bruker 500 MHz NMR instrument. 3. Results and discussion The rationale of this study was to discover small molecules to selectively inhibit the NSDs especially NSD2 in order to have tools to further investigate the biology of the NSDs and for drug-design purposes.
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
4
Y. Shen et al. / Biochemical and Biophysical Research Communications xxx (xxxx) xxx
Inhibition by small molecules of histone methyltransferases is achieved primarily by targeting the catalytic SET domain on either the coenzyme AdoMet binding pocket or the histone-tail binding area. Alternatively, the modulation of an HMTase activity can also be achieved by targeting an allosteric site or disrupting the interaction with proteins partner. The catalytic SET domain of the NSDs is subdivided into the AWS, SET, and post-SET sub-domains and is conserved across HMTases (Fig. 1A, S1&S2). At the sequence level, the catalytic SET domain of the NSDs is highly conserved (Fig. S1). NSD2-SET shares 75.9% identity and 90.1% similarity with the NSD1-SET, whereas the NSD3-SET shares 72.9% identity and 85% similarity with the NSD1-SET (Fig. S1). Structure-based drug-design on the NSDs, especially NSD2, is hampered by difficulties in crystallizing the catalytic domain. The structure of the catalytic SET domain of NSD1 in a closed conformation has been solved and revealed a flexible regulatory loop located at the interface of the SET and post-SET regions and is a conserved feature amongst the NSDs [21]. This regulatory loop acts as an open/close gate for the substrate access to the active site and is thought to be important in drug-design as it sits on top of the substrate in the opened conformation of the SET domain [4]. The inherent flexibility of this loop is one of the challenges for protein crystallization of the NSDs (Fig. 1D) [2,20]. Our previous study showed that the regulatory loop of NSD2-SET is displaced upon substrate binding with a ~45 rotation and a ~6 Å translation at the tip and it opens the negatively-charged binding groove for the docking of positivelycharged histones H3 or H4 [4]. The scaffold of the SET domain of NSD1, NSD2 and NSD3 is nonetheless conserved [14,22]. Taken
together, the high level of sequence conservation and scaffold similarity of the catalytic SET domain across the NSDs give a valid ground for relying on molecular modeling for a structure-based drug-design on NSD2-SET [2]. The NSDs are challenging drug-targets and only a few nonspecific inhibitors have been identified. MCTP39, Sinefungin and an N-Alkyl Sinefungin derivative inhibits NSD2 in vitro with IC50 of 3 mM, 26 mM and 3.3 mM, respectively [17,18]. PTD2, a norleucinecontaining peptide derived from histone H4 inhibits both NSD2 and NSD3 with IC50 of 22 mM and 3.2 mM, respectively [14]. Previously, we identified the hit inhibitor LEM-06 against NSD2 with an in vitro IC50 of 890 mM [1]. However, LEM-06 has a high IC50 value and is non NSD2 specific. Our previous study showed that BIX01294, an inhibitor for GLP/G9a H3K9 HMTases with IC50 values of 0.7 and 1.9 mM, respectively [19], is also an inhibitor against NSD1, NSD2, and NSD3 with IC50 values of 112, 41, and 95 mM, respectively [20]. Based upon our virtual ligand screening strategy we established to isolate LEM-06 [1], here we identified LEM-14 with an improved in vitro IC50 of 132 mM against NSD2 and, most importantly, has no activity against NSD3 and has a high IC50 > 1000 mM against NSD1 (Table 1). LEM-14 discriminates NSD2 over NSD1 and NSD3, which represents an interesting improvement over existing inhibitors. LEM14 is a small compound with a molecular weight of 478.6 Da, which is lighter than BIX-01294 (492.6 Da) but has a total polar surface area of 93.2 Å2 larger that BIX-01294 (68.4 Å2) (Table 1, Fig. 1 & S3). Amongst any other LEMs, LEM-14 displays the strongest positive electrostatic field akin to BIX-01294, which may be
Fig. 2. LEM-14 binding detail in the catalytic SET domain of the NSDs. The key residues (carbon atoms colored in black) that stabilize LEM-14 are displayed in the active and the coenzyme sites of NSD2 (light gray). Hydrogen bond (blue), aromatic hydrogen bond (light blue), pi-pi stacking (yellow), pi-cation interaction (green) and salt bridge (pink) are depicted with dashes. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
Y. Shen et al. / Biochemical and Biophysical Research Communications xxx (xxxx) xxx
5
Fig. 3. Chemical structures of the LEM-14 derivatives synthesized and tested for H3K36 inhibition in vitro against the NSDs-SET.
related to the strength of the interaction with the negatively charged surface of the histone-tail binding domain (Fig. 2). To understand the binding specificity of LEM-14 against NSD2, we used molecular modeling and flexible dockings. Small molecules of less than 500 Da can possibly bind either on the coenzyme or histone-tail site of SET containing enzymes. Therefore, to assess the likelihood of binding to either site, we docked LEM-14 into both the active and coenzyme pockets of NSD2-SET (Tables 1e3 and Fig. 2). LEM-14 appears to bind preferably to the histone tail site of all three NSDs (Table 2 and Fig. 2). A common binding scheme is seen with the flexible regulatory loop anchoring LEM-14 into the histone tail site (Fig. 2). Notably, LEM-14 doesn't bind into the narrow hydrophobic channel, where the lysine substrate extends, that links the coenzyme site to the histone tail site. This narrow hydrophobic channel seems to be important to target for better inhibition of the NSDs [1,4,20]. LEM-14 binds tightly in the active binding pocket of NSD2 with two hydrogen bonds, one aromatic hydrogen bond, and three pication interactions. In contrast, LEM-14 is stabilized in the coenzyme pocket with two hydrogen bonds and one salt bridge. LEM-14 binding to the coenzyme pocket of NSD2 appears weaker than to the active pocket (Fig. 2). LEM-14 binds to the coenzyme pocket with a docking energy of 11.2 kcal/mol, which is less favorable than binding into the active pocket ( 12.0 kcal/mol) (Table 2). In addition, LEM-14 is bulkier than the inhibitor MCTP39 or the coenzyme AdoMet. The binding of LEM-14 into the sterically constrained coenzyme pocket of the NSDs may be at the expense of a
higher energy cost, thus reducing the likelihood of occurrence. LEM-14 is weakly stabilized in the active site of NSD1 by a hydrogen bond and pi-pi stacking with Phe1996 as well as pication interaction with Tyr2058 (Fig. 2). In NSD3, LEM-14 bind also weakly to the active site through a single hydrogen bond with Thr1203 and a pi-pi stacking (Fig. 2). However, in the active site of NSD2, LEM-14 docks with a tighter network of interactions compared to NSD1 and NSD3 (Fig. 2). In NSD2, LEM-14 is anchored by hydrogen bonds with Met1119, Leu1120 and Gly1185 as well as an aromatic hydrogen bond with Phe1117 (Fig. 2). Moreover, a pipi stacking with Tyr1179 and a pi-cation interaction with Hsd1110 stabilizes the LEM-14 from two opposite sites (Fig. 2). Taken together, LEM-14 binds preferably to the histone-tail site of the NSDs but is better stabilized in NSD2, compared to NSD1 and NSD3, which may explain the selective NSD2 inhibition over NSD1 and NSD3 (Tables 1 and 2, Fig. 2). Next, in order to improve the biochemical inhibition of the NSDs, we investigated LEM-14 derivatives (Tables 1 and 3 and Fig. 3). The goal was to modify the structure of LEM-14 to maximize NSD2 inhibition and also explore possible inhibition of NSD1 and NSD3 as well. LEM-14 derivatives were designed by modification of oxoisochromane and piperazine parts of LEM-14 to confirm the changes of NSDs inhibitory activities. Oxoisochromane part of LEM14 was changed to isochromane, oxoisochromene and its chiral center as well as the piperazine part of LEM-14 was changed to the length of amine linker and the introduction of amide linker. The synthetic routes (see supplementary materials) were begun with
Table 2 Docking of LEM-14 into the catalytic SET domain of NSD1, NSD2 and NSD3.
Table 3 Docking of LEM-14-1189 into the catalytic SET domain of NSD1, NSD2 and NSD3.
Affinity (kcal/mol) NSD1 NSD2 NSD3
Histone-tail site 11.3 12.0 13.1
AdoMet site 10.6 11.2 11.2
Affinity (kcal/mol) NSD1 NSD2 NSD3
Histone-tail site 13.2 10.9 12.8
AdoMet site 10.1 11.7 15.0
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
6
Y. Shen et al. / Biochemical and Biophysical Research Communications xxx (xxxx) xxx
Fig. 4. LEM-14e1189 binding detail in the catalytic SET domain of the NSDs. The key residues (carbon atoms colored in black) that stabilize LEM-14-1189 are displayed in the active and coenzyme sites of NSD2 (light gray). Hydrogen bond (blue), aromatic hydrogen bond (light blue), pi-pi stacking (yellow), pi-cation interaction (green) and salt bridge (pink) are depicted with dashes.
three types of thieno[2,3-d]pyrimidine and the variable oxoisochromane derivatives, which relies on the amide formation and nucleophilic substitution for LEM-14 derivatives (Fig. 3 and Scheme 1-5 in supplementary materials). Interestingly, derivative LEM-141189 showed a notable improvement of inhibition compared to LEM-14 with in vitro IC50 of 418 mM (NSD1), IC50 of 111 mM (NSD2) and IC50 of 60 mM (NSD3) (Table 1), whereas LEM-14 is only active on NSD2 with an IC50 of 132 mM (Table 1). LEM-14-1189 inhibits NSD3, unlike LEM-14 (Table 1). LEM14-1189 bound to the coenzyme pocket of NSD3 showed the lowest docking energies of 15.0 kcal/mol, amongst the NSDs (Table 3), which is consistent with experimentally measured in vitro inhibitions. LEM-14-1189 is firmly bound in the coenzyme site of NSD3, in contrast to LEM-14 (Figs. 2 and 4). LEM-14 is anchored by a single hydrogen bond with Thr1197 in the coenzyme site of NSD3, whereas LEM-14-1189 is anchored by two hydrogen bonds with Asn1198 and Arg220 and pi-pi stackings with Tyr1261 and Hsd1274 (Fig. 4). When bound to the active site, LEM-14-1189 extends in the narrow lysine channel of NSD3 unlike LEM-14. LEM-14-1189 inhibits NSD1 in contrast to LEM-14 (Table 1). LEM-14-1189 has increased stability in both the active site and the coenzyme site of NSD1 with hydrogen bonds with Arg1982, Thr2000 and Lys2031 as well as a pi-pi stacking with Tyr 2058 in the active site and a single hydrogen bond with Lys2071 in the
coenzyme site, contrasting to the weaker binding network observed with LEM-14 (Figs. 2 and 4). LEM-14-1189 improved NSD2 inhibition compared to LEM-14 (Table 1). LEM-14-1189 appears to preferably bind into the coenzyme site of NSD2 and seems to be anchored tighter than LEM-14 (Figs. 2 and 4). In NSD2, LEM-14-1189 is stabilized by hydrogen bonds with Lys1069, Trp1075, Thr1115 and Arg1192. Taken together, LEM-14-1189 has a tighter interaction network with the NSDs compared to LEM-14 and is better stabilized in NSD3, which explains the selective NSD3 inhibition over NSD1 and NSD2. Small molecules inhibitors of the NSDs are needed [13,14,16]. Thus, we propose LEM14 and LEM-14-1189 as tools for investigating the biology of the NSDs and for further development of specific NSDs inhibitors suitable against malignancies including multiple myeloma. Acknowledgments This work was supported by the Basic Science Research Program of the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology of Korea: NRF2018R1D1A1B07049298 to EDL, NRF-2017R1A1A1A05001129 to TL and NRF-2016R1D1A1B01014286 to MM. The Kyungpook National University research funds contributed to support the research presented in this study.
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
Y. Shen et al. / Biochemical and Biophysical Research Communications xxx (xxxx) xxx
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.bbrc.2018.11.037. References [1] E. di Luccio, Inhibition of nuclear receptor binding SET domain 2/multiple myeloma SET domain by LEM-06 implication for epigenetic cancer therapies, J. Canc. Prev. 20 (2015) 113e120. [2] M. Morishita, D. Mevius, E. di Luccio, In vitro histone lysine methylation by NSD1, NSD2/MMSET/WHSC1, and NSD3/WHSC1L, BMC Struct. Biol. 14 (2014) 25. [3] E. di Luccio, P. Koehl, The H-factor as a novel quality metric for homology modeling, J. Clin. Bioinf. 2 (2012) 18. [4] M. Morishita, E. di Luccio, Structural insights into the regulation and the recognition of histone marks by the SET domain of NSD1, Biochem. Biophys. Res. Commun. 412 (2011) 214e219. [5] M. Morishita, E. di Luccio, Cancers and the NSD family of histone lysine methyltransferases, Biochim. Biophys. Acta 1816 (2011) 158e163. [6] M. Morishita, D. Mevius, E. di Luccio, In vitro histone lysine methylation by NSD1, NSD2/MMSET/WHSC1 and NSD3/WHSC1L, BMC Struct. Biol. 14 (2014) 25. [7] E. Martinez-Garcia, R. Popovic, D.J. Min, S.M. Sweet, P.M. Thomas, L. Zamdborg, A. Heffner, C. Will, L. Lamy, L.M. Staudt, D.L. Levens, N.L. Kelleher, J.D. Licht, The MMSET histone methyl transferase switches global histone methylation and alters gene expression in t(4;14) multiple myeloma cells, Blood 117 (2011) 211e220. [8] H. Pei, L. Zhang, K. Luo, Y. Qin, M. Chesi, F. Fei, P.L. Bergsagel, L. Wang, Z. You, Z. Lou, MMSET regulates histone H4K20 methylation and 53BP1 accumulation at DNA damage sites, Nature 470 (2011) 124e128. [9] F. Mirabella, P. Wu, C.P. Wardell, M.F. Kaiser, B.A. Walker, D.C. Johnson, G.J. Morgan, MMSET is the key molecular target in t(4;14) myeloma, Blood Canc. J. 3 (2013) e114. [10] J.Y. Kim, H.J. Kee, N.W. Choe, S.M. Kim, G.H. Eom, H.J. Baek, H. Kook, H. Kook, S.B. Seo, Multiple-myeloma-related WHSC1/MMSET isoform RE-IIBP is a histone methyltransferase with transcriptional repression activity, Mol. Cell Biol. 28 (2008) 2023e2034.
7
[11] J. Woo Park, K.B. Kim, J.Y. Kim, Y.C. Chae, O.S. Jeong, S.B. Seo, RE-IIBP methylates H3K79 and induces MEIS1-mediated apoptosis via H2BK120 ubiquitination by RNF20, Sci. Rep. 5 (2015) 12485. [12] A. Kassambara, B. Klein, J. Moreaux, MMSET is overexpressed in cancers: link with tumor aggressiveness, Biochem. Biophys. Res. Commun. 379 (2009) 840e845. [13] T. Vougiouklakis, R. Hamamoto, Y. Nakamura, V. Saloura, The NSD family of protein methyltransferases in human cancer, Epigenomics 7 (2015) 863e874. [14] M.J. Morrison, P.A. Boriack-Sjodin, K.K. Swinger, T.J. Wigle, D. Sadalge, K.W. Kuntz, M.P. Scott, W.P. Janzen, R. Chesworth, K.W. Duncan, D.M. Harvey, J.W. Lampe, L.H. Mitchell, R.A. Copeland, Identification of a peptide inhibitor for the histone methyltransferase WHSC1, PLoS One 13 (2018), e0197082. [15] P.L. Bergsagel, W.M. Kuehl, Molecular pathogenesis and a consequent classification of multiple myeloma, J. Clin. Oncol. 23 (2005) 6333e6338. [16] A. Swaroop, J.A. Oyer, C.M. Will, X. Huang, W. Yu, C. Troche, M. Bulic, B.H. Durham, Q.J. Wen, J.D. Crispino, A.D. MacKerell Jr., R.L. Bennett, N.L. Kelleher, J.D. Licht, An activating mutation of the NSD2 histone methyltransferase drives oncogenic reprogramming in acute lymphocytic leukemia, Oncogene (2018 Aug 31) [Epub ahead of print]. [17] A.M. Chinnaiyan, S. Lnu, Q. Cao, COMPOSITIONS AND METHODS FOR INHIBITING MMSET, EP Patent, 2012. [18] D. Tisi, E. Chiarparin, E. Tamanini, P. Pathuri, J.E. Coyle, A. Hold, F.P. Holding, N. Amin, A.C. Martin, S.J. Rich, V. Berdini, J. Yon, P. Acklam, R. Burke, L. Drouin, J.E. Harmer, F. Jeganathan, R.L. van Montfort, Y. Newbatt, M. Tortorici, M. Westlake, A. Wood, S. Hoelder, T.D. Heightman, Structure of the epigenetic oncogene MMSET and inhibition by N-Alkyl Sinefungin derivatives, ACS Chem. Biol. 11 (2016) 3093e3105. [19] Y. Chang, X. Zhang, J.R. Horton, A.K. Upadhyay, A. Spannhoff, J. Liu, J.P. Snyder, M.T. Bedford, X. Cheng, Structural basis for G9a-like protein lysine methyltransferase inhibition by BIX-01294, Nat. Struct. Mol. Biol. 16 (2009) 312e317. [20] M. Morishita, D.E.H.F. Mevius, Y. Shen, S. Zhao, E. di Luccio, BIX-01294 inhibits oncoproteins NSD1, NSD2 and NSD3, Med. Chem. Res. (2017) 1e10. [21] Q. Qiao, Y. Li, Z. Chen, M. Wang, D. Reinberg, R.M. Xu, The structure of NSD1 reveals an autoregulatory mechanism underlying histone H3K36 methylation, J. Biol. Chem. 286 (2011) 8361e8368. [22] N. Amin, D. Nietlispach, S. Qamar, J. Coyle, E. Chiarparin, G. Williams, NMR backbone resonance assignment and solution secondary structure determination of human NSD1 and NSD2, Biomol. NMR Assign. 10 (2016) 315e320.
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Identification of LEM-14 inhibitor of the oncoprotein NSD2 Yunpeng Shen a, e, 1, Masayo Morishita a, b, 1, Doohyun Lee c, d, Shinae Kim c, d, Taeho Lee c, Damiaan E.H.F. Mevius a, e, Yeonjeong Roh a, Eric di Luccio a, * a
Department of Genetic Engineering, School of Life Sciences, College of Natural Sciences, Kyungpook National University, Daegu, 41566, Republic of Korea Institute of Agricultural Science and Technology, Kyungpook National University, Daegu, 41566, Republic of Korea c College of Pharmacy, Research Institute of Pharmaceutical Sciences, Kyungpook National University, Daegu, 41566, Republic of Korea d New Drug Development Center, Daegu-Gyeongbuk Medical Innovation Foundation, Daegu, 41061, Republic of Korea e Department of Food Biomaterials, College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, Republic of Korea b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 29 October 2018 Accepted 6 November 2018 Available online xxx
The NSD family (NSD1, NSD2/MMSET/WHSC1, and NSD3/WHSC1L1) are histone lysine methyltransferases (HMTases) essential for chromatin regulation. The NSDs are oncoproteins, drivers of a number of tumors and are considered important drug-targets but the lack of potent and selective inhibitors hampers further therapeutic development and limits exploration of their biology. In particular, MMSET/NSD2 selective inhibition is being pursued for therapeutic interventions against multiple myeloma (MM) cases, especially in multiple myeloma t(4;14)(p16.3;q32) translocation that is associated with a significantly worse prognosis than other MM subgroups. Multiple myeloma is the second most common hematological malignancy, after non-Hodgkin lymphoma and remains an incurable malignancy. Here we report the discovery of LEM-14, an NSD2 specific inhibitor with an in vitro IC50 of 132 mM and that is inactive against the closely related NSD1 and NSD3. LEM-14-1189, a LEM-14 derivative, differentially inhibits the NSDs with in vitro IC50 of 418 mM (NSD1), IC50 of 111 mM (NSD2) and IC50 of 60 mM (NSD3). We propose LEM-14 and derivative LEM-14-1189 as tools for studying the biology of the NSDs and constitute meaningful steps toward potent NSDs therapeutic inhibitors. © 2018 Elsevier Inc. All rights reserved.
Keywords: Epigenetic therapy of cancer Multiple myeloma Histone-lysine methyltransferase NSD2 Inhibitors Drug-design
1. Introduction The nuclear receptor-binding SET domain (NSD) family consists of NSD1, NSD2/MMSET/WHSC1 and NSD3/WHSC1L1 histone methyltransferases (HMTases). The NSD family are important in regulating the chromatin, evidenced by their implications in developmental syndromes and numerous cancers [1e5]. Chromatin regulation by NSD2 is mediated through H3K36 and H4K20 methylations, although in vitro its substrate specificity is broaden to H3K4, H3K9 H3K27, H3K36, H3K79, and H4K20, underlying an interesting regulatory mechanism that is still unknown [6e8]. Overexpression of NSD2 increases cellular proliferation,
Abbreviations: NSD, nuclear receptor binding SET domain; MMSET, Multiple Myeloma SET domain; HMTase, Histone lysine Methyltransferase; AdoMET, Sadenosylmethionine. * Corresponding author. E-mail addresses: [email protected], [email protected] (E. di Luccio). 1 Both authors contributed equally to this study.
chromosomal accessibility, and alteration of regulation of gene expression levels [7]. Whereas, knock-down of NSD2 affects cell adhesion, inhibits proliferation, and induces apoptosis [9]. Furthermore, NSD2 has roles in DNA damage repair being recruited to double strand-breaks (DSBs) with an increase of H4K20 methylation level at DSBs [8]. Additionally, expression of REIIBP, a short NSD2 isoform, leads to elevated H3K27 methylation and recruits histone deacetylases (HDACs) [10]. REIIBP might also be able to methylate H3K79, a mark previously thought to be methylated exclusively by DOT1L [11]. NSD2 haploinsufficiency results in the developmental WolfHirschhorn syndrome [12]. NSD2 abnormal expression level is associated with tumor aggressiveness and is implicated in more than 20 cancers [12,13]. Specifically, abnormal overexpression of NSD2 is a hallmark of the t(4;14)(p16.3;q32) translocation in multiple myeloma that is associated with a poor prognosis. Multiple Myeloma, accounts for 20% of hematological malignancy deaths and there is no cure. Inhibition of the NSDs, especially NSD2 represents a new avenue for cancer therapies, especially in multiple
https://doi.org/10.1016/j.bbrc.2018.11.037 0006-291X/© 2018 Elsevier Inc. All rights reserved.
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
2
Y. Shen et al. / Biochemical and Biophysical Research Communications xxx (xxxx) xxx
myeloma cases [5,7,13,14]. NSDs inhibitors are needed, but HMTase inhibitors remain scarce [15,16]. MCTP39, Sinefungin and an N-Alkyl Sinefungin derivative have been identified as NSD2 inhibitors with in vitro IC50 of 3 mM, 26 mM and 3.3 mM, respectively but inhibition data on NSD1 and NSD3 is unknown [17,18]. PTD2, a norleucine-containing peptide derived from the histone H4 sequence inhibits both NSD2 and NSD3 with IC50 of 22 mM and 3.2 mM, respectively [14]. PTD2 inhibition on NSD1 is unknown. BIX-01294 and BIX-01338 are inhibitors of G9a/G9a-like protein (GLP) H3K9 HMTase. BIX-01294 inhibits GLP and G9a with IC50 values of 0.7 and 1.9 mM, respectively and also inhibits NSD1, NSD2, and NSD3 with IC50 values of 112, 41, and 95 mM, respectively [19,20]. Taken together, the selective inhibition of each NSD and especially NSD2 remains to be achieved. The NSDs operate in distinct pathways but share a highly conserved SET domain, which is challenging for drug-design (Fig. S1) [2,4]. The catalytic SET domain (Su(var)3-9, Enhancer-ofzeste and Trithorax (SET)) is composed of two binding pockets, the histone-tail binding/active site and the coenzyme S-adenosyl methionine (SAM/AdoMet) site (Fig.1 A&D). The histone lysine substrate extends into a narrow channel linking the two pockets. Inhibition of the methyltransferase activity can primarily be achieved by targeting either binding pocket. Previously, we showed that pan inhibitors such as BIX-01294 could bind to the active site and the coenzyme site of all NSDs with however a slightly higher affinity toward the active site than the coenzyme site [20]. The AdoMet binding site is also highly conserved across
methyltransferases. However, the histone-tail binding area is heterogeneous at the sequence level while having a scaffold structurally conserved, which may offer opportunities for achieving selective inhibition (Fig. S1&S2) [2,4]. In this study, we report the discovery by virtual ligand screening of LEM-14, the first NSD2 specific inhibitor amongst the NSDs (Fig. 1& Table 1). LEM-14 is a hit inhibitor with an in vitro IC50 of 132 mM against H3K36 methylation by NSD2 and has very weak activity against NSD1 and has no activity against NSD3 (Table 1). We further designed, produced and investigated 14 derivatives of LEM-14 (Fig. S3 & Table 1). Derivative LEM-14-1189 differentially inhibits the NSDs with an in vitro IC50 of 418 mM (NSD1), IC50 of 111 mM (NSD2) and IC50 of 60 mM (NSD3) (Table 1).
2. Materials and methods 2.1. Cloning, protein expression, and protein purification The cloning of the SET domains of human NSD1, NSD2, and NSD3 genes and the expression and purification of NSD proteins were described previously [20]. The resultant NSD1-SET (309 a.a., 35.1 kDa), NSD2-SET (299 a.a., 34.0 kDa), and NSD3-SET (309 a.a., 35.0 kDa) were proven to be stable and retained catalytic properties for an extended period of time when stored at 80 C and confirmed by SDS-PAGE and Coomassie brilliant blue G-250 staining (Fig. 1 B).
Fig. 1. Schematic NSD2 and structure of LEM-14, LEM-14e1189 and LEM-06. (A) Schematic of the primary structure of NSD2: PWWP domain; PHD zinc fingers domain; SET histone methyl transferase (HMTase) with the AWS and postSET domains. The regulatory loop closing onto the histone-binding site is indicated in red. (B) Bacterially expressed and purified NSD1, NSD2, and NSD3 SET domain used for in vitro assays were detected by SDS-PAGE and Coomassie blue staining. (C) Structure of LEM-14. (D) Structure of NSD2 catalytic domain bound with LEM-14 into the active site. The overall structure and molecular surface of the loop-opened NSD2 SET domain model bound with the computed most favorable docked conformation of LEM-14 (carbon atoms colored in yellow) in the active site. AdoHcy (carbon atoms colored in green) is shown in the coenzyme site. The negative-charged surface is indicated as red and positive charged as blue. LEM-14 docking conformations are focused and magnified in boxes. (E) Structure of LEM-14-1189. (F) Structure of LEM-06. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
Y. Shen et al. / Biochemical and Biophysical Research Communications xxx (xxxx) xxx
3
Table 1 Biological in vitro IC50 (mM) for H3K36 mono-methylation with histone H3 as substrate; n. a. indicates no inhibitory activity; >1000 indicates a weak inhibition with an IC50 above 1000 mM. MW: molecular weight (Da); cLogP: octanol/water partition coefficient; TPSA: total polar surface area; Rot. B.: number of rotatable bonds; Rule of 5: The number of Lipinski's rule of 5 violations for each molecule. Molecule
NSD1 IC50 (mM)
NSD2 IC50 (mM)
NSD3 IC50 (mM)
MW (Da)
TPSA (Å2)
cLogP
Rot. B.
Rule of 5
Ref.
BIX-01294 LEM-06 LEM-14 LEM-14-1185 LEM-14-1187 LEM-14e1189 LEM-14-1190 LEM-14-1191 LEM-14-1192 LEM-14-1193 LEM-14-1195 LEM-14-1196 LEM-14-1199 LEM-14-1200 LEM-14-1202 LEM-14-1203 LEM-14-1204
112 ± 57 >1000 >1000 n. a. n. a. 418 ± 111 >1000 >1000 >1000 >1000 >1000 >1000 >1000 >1000 >1000 n. a. n. a.
41 ± 2 890 ± 250 132 ± 50 n. a. n. a. 111 ± 11 >1000 >1000 >1000 >1000 >1000 >1000 >1000 >1000 >1000 n. a. n. a.
95 ± 53 >1000 n. a. n. a. n. a. 60 ± 11 >1000 >1000 >1000 >1000 >1000 >1000 >1000 >1000 >1000 n. a. n. a.
492.6 492.7 478.6 334.4 476.5 683.8 465.5 465.6 493.6 479.6 465.6 493.6 478.6 492.5 492.5 406.5 307.4
68.7 58.7 93.2 99.7 91.3 141.5 113.5 75.4 75.4 92.5 75.4 75.4 91.3 108.4 108.4 108.9 98.2
3,9 2.4 1.7 0.7 2.8 1.15 2.7 2.7 3.3 2.3 2.7 3.3 2.2 1.9 1.9 2.4 0.9
7 4 3 3.9 4.3 9.7 7.1 5.3 6.3 5.3 5.3 6.3 3.8 3.8 3.8 5.4 3.9
0 0 0 0 0 2 0 0 0 1 0 0 0 0 0 0 0
[20] [1]
2.2. Compounds Compounds were purchased from MolPort (http://www. molport.com/) and prepared in dimethyl sulfoxide (DMSO) > 99.9% pure (D8418 Sigma). 2.3. Histone methyltransferase inhibitor assay Histone methyltransferase activity of NSD2-SET on monomethylation of H3K36 for the dose-response curve was measured by colorimetric quantification kits (P-3046, Epigentek, USA) following the manufacturer's protocol as previously described. Methyltransferase activity of NSD1-SET, NSD2-SET, and NSD3-SET treated with or without LEM-14 and its derivatives were measured with Methyltransferase Activity Assay kits (ADI-907-025, Enzo Life Sciences, USA) following the manufacturer's instruction. Assays were done in triplicate individual experiments. The results were normalized against the control that does not contain any enzymes. GraphPad Prism 7.0 was used to process the data, plot dose-response curves and calculate IC50 values. 2.4. Molecular modeling of the closed and open conformation of NSD2-SET Homology models of NSD1 (UniProt: Q96L73), NSD2 (UniProt: Q9BZ95) and NSD3 (UniProt: Q9BZ95) were built, refined and validated as we previously described [2]. Initial NSDs-SET build models are in a closed-conformation with the flexible regulatory loop blocking the substrate from entering the active site. The openconformation model of the catalytic domain was used for the docking of small molecules following the methods we previously established [4]. 2.5. Docking model of LEM-14 and derivatives to the NSDs The docking of LEM-14 into the active site and the coenzyme site of the open-conformation models of NSD1, NSD2 and NSD3 where the regulatory loop is in an active state. S-Adenosyl-L-homocysteine (AdoHcy) was manually placed at the coenzyme site openconformation models docked with LEM-14 at the active site, which was guided by superposition of the crystal structures of NSD1 (PDB:3OOI) and NSD3 (PDB:4YZ8) as well as NSD SET models. The NSD-LEM-14 model complexes were prepared with Autodock
Tools 1.5.6 with default settings. All bonds of LEM-14 were treated as rotatable bonds except for amine bonds. Autodock Vina was used for docking. Results from the rigid docking were utilized to assess the interacting amino acid residues that were selected for a following flexible docking. Due to the absence of NSDs structure data, the interaction prediction between NSDs and histone tail peptide was achieved by using GalaxyPepDock. The top ten predicted Set7-histone tail peptide complex structures were analyzed. The predicted structure (the histone tail peptide GGVKKPHR as input) used for further experiment was selected in terms of the highest TM-score and RMSD. 2.6. Virtual ligand screening and selection of the compounds AutoDock Vina was used for virtual ligand screening with the “Clean Drug-Like compounds” downloaded from the ZINC docking database version 11. Hydrogen atoms on the target were added using PDB2PQR program in conjunction with the CHARMM force field. All the ligands were prepared with the AutoDock tools (ADT) v1.5.4 r29. A subset of 40,325 molecules (190 Da < MW < 500 Da) was docked and the top 3000 results were analyzed and filtered with FieldView from Cresset (Herts, AL7 3AX, United Kingdom). For each molecule, the top 9 binding conformations were retained. The top docking solutions were manually inspected. LEM-14 (ZINC ID: 23116671) was purchased from Molport (cat.: 005-735-962). Selected molecules for in vitro inhibition are named LEM. 2.7. Synthesis of LEM-14 derivatives LEM-14 derivatives (Fig. 3) were synthesized following schemes described in supplementary materials (Scheme 1-5). All chemicals were reagent grade and used as purchased. Flash column chromatography was carried out on Merck silica gel 60 (230e400 mesh). 1H NMR spectra were recorded in d units relative to deuterated solvent as internal reference by a Bruker 500 MHz NMR instrument. 3. Results and discussion The rationale of this study was to discover small molecules to selectively inhibit the NSDs especially NSD2 in order to have tools to further investigate the biology of the NSDs and for drug-design purposes.
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
4
Y. Shen et al. / Biochemical and Biophysical Research Communications xxx (xxxx) xxx
Inhibition by small molecules of histone methyltransferases is achieved primarily by targeting the catalytic SET domain on either the coenzyme AdoMet binding pocket or the histone-tail binding area. Alternatively, the modulation of an HMTase activity can also be achieved by targeting an allosteric site or disrupting the interaction with proteins partner. The catalytic SET domain of the NSDs is subdivided into the AWS, SET, and post-SET sub-domains and is conserved across HMTases (Fig. 1A, S1&S2). At the sequence level, the catalytic SET domain of the NSDs is highly conserved (Fig. S1). NSD2-SET shares 75.9% identity and 90.1% similarity with the NSD1-SET, whereas the NSD3-SET shares 72.9% identity and 85% similarity with the NSD1-SET (Fig. S1). Structure-based drug-design on the NSDs, especially NSD2, is hampered by difficulties in crystallizing the catalytic domain. The structure of the catalytic SET domain of NSD1 in a closed conformation has been solved and revealed a flexible regulatory loop located at the interface of the SET and post-SET regions and is a conserved feature amongst the NSDs [21]. This regulatory loop acts as an open/close gate for the substrate access to the active site and is thought to be important in drug-design as it sits on top of the substrate in the opened conformation of the SET domain [4]. The inherent flexibility of this loop is one of the challenges for protein crystallization of the NSDs (Fig. 1D) [2,20]. Our previous study showed that the regulatory loop of NSD2-SET is displaced upon substrate binding with a ~45 rotation and a ~6 Å translation at the tip and it opens the negatively-charged binding groove for the docking of positivelycharged histones H3 or H4 [4]. The scaffold of the SET domain of NSD1, NSD2 and NSD3 is nonetheless conserved [14,22]. Taken
together, the high level of sequence conservation and scaffold similarity of the catalytic SET domain across the NSDs give a valid ground for relying on molecular modeling for a structure-based drug-design on NSD2-SET [2]. The NSDs are challenging drug-targets and only a few nonspecific inhibitors have been identified. MCTP39, Sinefungin and an N-Alkyl Sinefungin derivative inhibits NSD2 in vitro with IC50 of 3 mM, 26 mM and 3.3 mM, respectively [17,18]. PTD2, a norleucinecontaining peptide derived from histone H4 inhibits both NSD2 and NSD3 with IC50 of 22 mM and 3.2 mM, respectively [14]. Previously, we identified the hit inhibitor LEM-06 against NSD2 with an in vitro IC50 of 890 mM [1]. However, LEM-06 has a high IC50 value and is non NSD2 specific. Our previous study showed that BIX01294, an inhibitor for GLP/G9a H3K9 HMTases with IC50 values of 0.7 and 1.9 mM, respectively [19], is also an inhibitor against NSD1, NSD2, and NSD3 with IC50 values of 112, 41, and 95 mM, respectively [20]. Based upon our virtual ligand screening strategy we established to isolate LEM-06 [1], here we identified LEM-14 with an improved in vitro IC50 of 132 mM against NSD2 and, most importantly, has no activity against NSD3 and has a high IC50 > 1000 mM against NSD1 (Table 1). LEM-14 discriminates NSD2 over NSD1 and NSD3, which represents an interesting improvement over existing inhibitors. LEM14 is a small compound with a molecular weight of 478.6 Da, which is lighter than BIX-01294 (492.6 Da) but has a total polar surface area of 93.2 Å2 larger that BIX-01294 (68.4 Å2) (Table 1, Fig. 1 & S3). Amongst any other LEMs, LEM-14 displays the strongest positive electrostatic field akin to BIX-01294, which may be
Fig. 2. LEM-14 binding detail in the catalytic SET domain of the NSDs. The key residues (carbon atoms colored in black) that stabilize LEM-14 are displayed in the active and the coenzyme sites of NSD2 (light gray). Hydrogen bond (blue), aromatic hydrogen bond (light blue), pi-pi stacking (yellow), pi-cation interaction (green) and salt bridge (pink) are depicted with dashes. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
Y. Shen et al. / Biochemical and Biophysical Research Communications xxx (xxxx) xxx
5
Fig. 3. Chemical structures of the LEM-14 derivatives synthesized and tested for H3K36 inhibition in vitro against the NSDs-SET.
related to the strength of the interaction with the negatively charged surface of the histone-tail binding domain (Fig. 2). To understand the binding specificity of LEM-14 against NSD2, we used molecular modeling and flexible dockings. Small molecules of less than 500 Da can possibly bind either on the coenzyme or histone-tail site of SET containing enzymes. Therefore, to assess the likelihood of binding to either site, we docked LEM-14 into both the active and coenzyme pockets of NSD2-SET (Tables 1e3 and Fig. 2). LEM-14 appears to bind preferably to the histone tail site of all three NSDs (Table 2 and Fig. 2). A common binding scheme is seen with the flexible regulatory loop anchoring LEM-14 into the histone tail site (Fig. 2). Notably, LEM-14 doesn't bind into the narrow hydrophobic channel, where the lysine substrate extends, that links the coenzyme site to the histone tail site. This narrow hydrophobic channel seems to be important to target for better inhibition of the NSDs [1,4,20]. LEM-14 binds tightly in the active binding pocket of NSD2 with two hydrogen bonds, one aromatic hydrogen bond, and three pication interactions. In contrast, LEM-14 is stabilized in the coenzyme pocket with two hydrogen bonds and one salt bridge. LEM-14 binding to the coenzyme pocket of NSD2 appears weaker than to the active pocket (Fig. 2). LEM-14 binds to the coenzyme pocket with a docking energy of 11.2 kcal/mol, which is less favorable than binding into the active pocket ( 12.0 kcal/mol) (Table 2). In addition, LEM-14 is bulkier than the inhibitor MCTP39 or the coenzyme AdoMet. The binding of LEM-14 into the sterically constrained coenzyme pocket of the NSDs may be at the expense of a
higher energy cost, thus reducing the likelihood of occurrence. LEM-14 is weakly stabilized in the active site of NSD1 by a hydrogen bond and pi-pi stacking with Phe1996 as well as pication interaction with Tyr2058 (Fig. 2). In NSD3, LEM-14 bind also weakly to the active site through a single hydrogen bond with Thr1203 and a pi-pi stacking (Fig. 2). However, in the active site of NSD2, LEM-14 docks with a tighter network of interactions compared to NSD1 and NSD3 (Fig. 2). In NSD2, LEM-14 is anchored by hydrogen bonds with Met1119, Leu1120 and Gly1185 as well as an aromatic hydrogen bond with Phe1117 (Fig. 2). Moreover, a pipi stacking with Tyr1179 and a pi-cation interaction with Hsd1110 stabilizes the LEM-14 from two opposite sites (Fig. 2). Taken together, LEM-14 binds preferably to the histone-tail site of the NSDs but is better stabilized in NSD2, compared to NSD1 and NSD3, which may explain the selective NSD2 inhibition over NSD1 and NSD3 (Tables 1 and 2, Fig. 2). Next, in order to improve the biochemical inhibition of the NSDs, we investigated LEM-14 derivatives (Tables 1 and 3 and Fig. 3). The goal was to modify the structure of LEM-14 to maximize NSD2 inhibition and also explore possible inhibition of NSD1 and NSD3 as well. LEM-14 derivatives were designed by modification of oxoisochromane and piperazine parts of LEM-14 to confirm the changes of NSDs inhibitory activities. Oxoisochromane part of LEM14 was changed to isochromane, oxoisochromene and its chiral center as well as the piperazine part of LEM-14 was changed to the length of amine linker and the introduction of amide linker. The synthetic routes (see supplementary materials) were begun with
Table 2 Docking of LEM-14 into the catalytic SET domain of NSD1, NSD2 and NSD3.
Table 3 Docking of LEM-14-1189 into the catalytic SET domain of NSD1, NSD2 and NSD3.
Affinity (kcal/mol) NSD1 NSD2 NSD3
Histone-tail site 11.3 12.0 13.1
AdoMet site 10.6 11.2 11.2
Affinity (kcal/mol) NSD1 NSD2 NSD3
Histone-tail site 13.2 10.9 12.8
AdoMet site 10.1 11.7 15.0
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
6
Y. Shen et al. / Biochemical and Biophysical Research Communications xxx (xxxx) xxx
Fig. 4. LEM-14e1189 binding detail in the catalytic SET domain of the NSDs. The key residues (carbon atoms colored in black) that stabilize LEM-14-1189 are displayed in the active and coenzyme sites of NSD2 (light gray). Hydrogen bond (blue), aromatic hydrogen bond (light blue), pi-pi stacking (yellow), pi-cation interaction (green) and salt bridge (pink) are depicted with dashes.
three types of thieno[2,3-d]pyrimidine and the variable oxoisochromane derivatives, which relies on the amide formation and nucleophilic substitution for LEM-14 derivatives (Fig. 3 and Scheme 1-5 in supplementary materials). Interestingly, derivative LEM-141189 showed a notable improvement of inhibition compared to LEM-14 with in vitro IC50 of 418 mM (NSD1), IC50 of 111 mM (NSD2) and IC50 of 60 mM (NSD3) (Table 1), whereas LEM-14 is only active on NSD2 with an IC50 of 132 mM (Table 1). LEM-14-1189 inhibits NSD3, unlike LEM-14 (Table 1). LEM14-1189 bound to the coenzyme pocket of NSD3 showed the lowest docking energies of 15.0 kcal/mol, amongst the NSDs (Table 3), which is consistent with experimentally measured in vitro inhibitions. LEM-14-1189 is firmly bound in the coenzyme site of NSD3, in contrast to LEM-14 (Figs. 2 and 4). LEM-14 is anchored by a single hydrogen bond with Thr1197 in the coenzyme site of NSD3, whereas LEM-14-1189 is anchored by two hydrogen bonds with Asn1198 and Arg220 and pi-pi stackings with Tyr1261 and Hsd1274 (Fig. 4). When bound to the active site, LEM-14-1189 extends in the narrow lysine channel of NSD3 unlike LEM-14. LEM-14-1189 inhibits NSD1 in contrast to LEM-14 (Table 1). LEM-14-1189 has increased stability in both the active site and the coenzyme site of NSD1 with hydrogen bonds with Arg1982, Thr2000 and Lys2031 as well as a pi-pi stacking with Tyr 2058 in the active site and a single hydrogen bond with Lys2071 in the
coenzyme site, contrasting to the weaker binding network observed with LEM-14 (Figs. 2 and 4). LEM-14-1189 improved NSD2 inhibition compared to LEM-14 (Table 1). LEM-14-1189 appears to preferably bind into the coenzyme site of NSD2 and seems to be anchored tighter than LEM-14 (Figs. 2 and 4). In NSD2, LEM-14-1189 is stabilized by hydrogen bonds with Lys1069, Trp1075, Thr1115 and Arg1192. Taken together, LEM-14-1189 has a tighter interaction network with the NSDs compared to LEM-14 and is better stabilized in NSD3, which explains the selective NSD3 inhibition over NSD1 and NSD2. Small molecules inhibitors of the NSDs are needed [13,14,16]. Thus, we propose LEM14 and LEM-14-1189 as tools for investigating the biology of the NSDs and for further development of specific NSDs inhibitors suitable against malignancies including multiple myeloma. Acknowledgments This work was supported by the Basic Science Research Program of the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology of Korea: NRF2018R1D1A1B07049298 to EDL, NRF-2017R1A1A1A05001129 to TL and NRF-2016R1D1A1B01014286 to MM. The Kyungpook National University research funds contributed to support the research presented in this study.
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037
Y. Shen et al. / Biochemical and Biophysical Research Communications xxx (xxxx) xxx
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.bbrc.2018.11.037. References [1] E. di Luccio, Inhibition of nuclear receptor binding SET domain 2/multiple myeloma SET domain by LEM-06 implication for epigenetic cancer therapies, J. Canc. Prev. 20 (2015) 113e120. [2] M. Morishita, D. Mevius, E. di Luccio, In vitro histone lysine methylation by NSD1, NSD2/MMSET/WHSC1, and NSD3/WHSC1L, BMC Struct. Biol. 14 (2014) 25. [3] E. di Luccio, P. Koehl, The H-factor as a novel quality metric for homology modeling, J. Clin. Bioinf. 2 (2012) 18. [4] M. Morishita, E. di Luccio, Structural insights into the regulation and the recognition of histone marks by the SET domain of NSD1, Biochem. Biophys. Res. Commun. 412 (2011) 214e219. [5] M. Morishita, E. di Luccio, Cancers and the NSD family of histone lysine methyltransferases, Biochim. Biophys. Acta 1816 (2011) 158e163. [6] M. Morishita, D. Mevius, E. di Luccio, In vitro histone lysine methylation by NSD1, NSD2/MMSET/WHSC1 and NSD3/WHSC1L, BMC Struct. Biol. 14 (2014) 25. [7] E. Martinez-Garcia, R. Popovic, D.J. Min, S.M. Sweet, P.M. Thomas, L. Zamdborg, A. Heffner, C. Will, L. Lamy, L.M. Staudt, D.L. Levens, N.L. Kelleher, J.D. Licht, The MMSET histone methyl transferase switches global histone methylation and alters gene expression in t(4;14) multiple myeloma cells, Blood 117 (2011) 211e220. [8] H. Pei, L. Zhang, K. Luo, Y. Qin, M. Chesi, F. Fei, P.L. Bergsagel, L. Wang, Z. You, Z. Lou, MMSET regulates histone H4K20 methylation and 53BP1 accumulation at DNA damage sites, Nature 470 (2011) 124e128. [9] F. Mirabella, P. Wu, C.P. Wardell, M.F. Kaiser, B.A. Walker, D.C. Johnson, G.J. Morgan, MMSET is the key molecular target in t(4;14) myeloma, Blood Canc. J. 3 (2013) e114. [10] J.Y. Kim, H.J. Kee, N.W. Choe, S.M. Kim, G.H. Eom, H.J. Baek, H. Kook, H. Kook, S.B. Seo, Multiple-myeloma-related WHSC1/MMSET isoform RE-IIBP is a histone methyltransferase with transcriptional repression activity, Mol. Cell Biol. 28 (2008) 2023e2034.
7
[11] J. Woo Park, K.B. Kim, J.Y. Kim, Y.C. Chae, O.S. Jeong, S.B. Seo, RE-IIBP methylates H3K79 and induces MEIS1-mediated apoptosis via H2BK120 ubiquitination by RNF20, Sci. Rep. 5 (2015) 12485. [12] A. Kassambara, B. Klein, J. Moreaux, MMSET is overexpressed in cancers: link with tumor aggressiveness, Biochem. Biophys. Res. Commun. 379 (2009) 840e845. [13] T. Vougiouklakis, R. Hamamoto, Y. Nakamura, V. Saloura, The NSD family of protein methyltransferases in human cancer, Epigenomics 7 (2015) 863e874. [14] M.J. Morrison, P.A. Boriack-Sjodin, K.K. Swinger, T.J. Wigle, D. Sadalge, K.W. Kuntz, M.P. Scott, W.P. Janzen, R. Chesworth, K.W. Duncan, D.M. Harvey, J.W. Lampe, L.H. Mitchell, R.A. Copeland, Identification of a peptide inhibitor for the histone methyltransferase WHSC1, PLoS One 13 (2018), e0197082. [15] P.L. Bergsagel, W.M. Kuehl, Molecular pathogenesis and a consequent classification of multiple myeloma, J. Clin. Oncol. 23 (2005) 6333e6338. [16] A. Swaroop, J.A. Oyer, C.M. Will, X. Huang, W. Yu, C. Troche, M. Bulic, B.H. Durham, Q.J. Wen, J.D. Crispino, A.D. MacKerell Jr., R.L. Bennett, N.L. Kelleher, J.D. Licht, An activating mutation of the NSD2 histone methyltransferase drives oncogenic reprogramming in acute lymphocytic leukemia, Oncogene (2018 Aug 31) [Epub ahead of print]. [17] A.M. Chinnaiyan, S. Lnu, Q. Cao, COMPOSITIONS AND METHODS FOR INHIBITING MMSET, EP Patent, 2012. [18] D. Tisi, E. Chiarparin, E. Tamanini, P. Pathuri, J.E. Coyle, A. Hold, F.P. Holding, N. Amin, A.C. Martin, S.J. Rich, V. Berdini, J. Yon, P. Acklam, R. Burke, L. Drouin, J.E. Harmer, F. Jeganathan, R.L. van Montfort, Y. Newbatt, M. Tortorici, M. Westlake, A. Wood, S. Hoelder, T.D. Heightman, Structure of the epigenetic oncogene MMSET and inhibition by N-Alkyl Sinefungin derivatives, ACS Chem. Biol. 11 (2016) 3093e3105. [19] Y. Chang, X. Zhang, J.R. Horton, A.K. Upadhyay, A. Spannhoff, J. Liu, J.P. Snyder, M.T. Bedford, X. Cheng, Structural basis for G9a-like protein lysine methyltransferase inhibition by BIX-01294, Nat. Struct. Mol. Biol. 16 (2009) 312e317. [20] M. Morishita, D.E.H.F. Mevius, Y. Shen, S. Zhao, E. di Luccio, BIX-01294 inhibits oncoproteins NSD1, NSD2 and NSD3, Med. Chem. Res. (2017) 1e10. [21] Q. Qiao, Y. Li, Z. Chen, M. Wang, D. Reinberg, R.M. Xu, The structure of NSD1 reveals an autoregulatory mechanism underlying histone H3K36 methylation, J. Biol. Chem. 286 (2011) 8361e8368. [22] N. Amin, D. Nietlispach, S. Qamar, J. Coyle, E. Chiarparin, G. Williams, NMR backbone resonance assignment and solution secondary structure determination of human NSD1 and NSD2, Biomol. NMR Assign. 10 (2016) 315e320.
Please cite this article as: Y. Shen et al., Identification of LEM-14 inhibitor of the oncoprotein NSD2, Biochemical and Biophysical Research Communications, https://doi.org/10.1016/j.bbrc.2018.11.037