and Y365) translocates side chains Y365 and W364 to create a deeper
pocket allowing for insertion of EED226 further into the pocket. EED226 did not
alter the conformation of the EBD-binding site. In vivo, EED226 induced the
regression of tumor xenografts in mouse models. Finally, EED226 was capable of
inhibiting cancer cell lines with acquired resistance to S-adenosylmethionine EZH2
inhibitors and also displayed a synergistic effect to inhibit cancer cell growth when
combined with EZH2 inhibitors [245].
Another high-throughput screening effort against EED, followed by structurebased medicinal chemistry efforts, led to the identification of A-395 (K i ¼ 0.4 nM) as
an allosteric inhibitor of PRC2 (Fig. 17) [246]. The compound selectively and
potently binds EED and inhibits the catalytic activity of the trimeric PRC2 complex
(EZH2-EED-SUZ12) in an in vitro radioactivity-based assay with an IC 50 of 18 nM.
Crystallographic studies and an AlphaLISA assay demonstrated the capability of
A-395 to compete with H3K27me3 for binding to the aromatic cage formed by the
WD40 repeat domain of EED (IC 50 ¼ 7 nM). Additionally, A-395 potently reduced
H3K27 methylation in rhabdoid tumor cells and inhibited proliferation of human
cancer cells, specifically DLBCL Pfeiffer and Karpas422, which are sensitive to
EZH2 inhibition. In DLBCL Pfeiffer cell xenograft models, A-395 displayed significant inhibition of tumor growth (84%). Finally, A-395 retained antitumor effects
against cell lines resistant to SET domain targeted EZH2 inhibitors [246].
A different approach to discover ligands of EED started from the identification of
JARID2 trimethylated at lysine 116 (JARID2-K116me3) as an EED-binding partner
and allosteric activator of PRC2 methyltransferase activity at low micromolar
potency (K D ¼ 3 μM), which is tenfold higher than the affinity of EED for
H3Kme7 in vitro [247]. The co-crystal structure of EED in complex with
JARID2-K116me3 (PDB: 4X3E) provided insight to the structural features necessary for binding, specifically at residues 114–118 [248]. Using a split-and-pool
synthesis method, a first-generation library of 1,029 peptides based on a truncated
JARID2 114–118 K116me3 scaffold (Fig. 18) was synthesized, and an on-bead magnetic enrichment screening approach for EED was used to identify lead compounds
[248]. Compound #1 (Fig. 18, compound numbering as in the original paper) was
identified and when evaluated by ITC had improved potency compared to
JARID2 114–118 K116me3 (K D ¼ 4.8 μM and 8.8 μM, respectively). A subsequent
second-generation library consisting of 4,410 compounds was designed to enhance
ligand affinity and compound #3 (Fig. 18, compound numbering as in the original
paper), which demonstrated similar binding values in both ITC and an EED FP assay
(K D ¼ 1.09 μM and IC 50 ¼ 1.65 μM, respectively) was identified. Crystallographic
studies and further modifications to the N- and C-terminus as well as replacement of
the Kme3 resulted in the more potent UNC5114 (ITC, K D ¼ 0.68 μM and FP assay
IC 50 ¼ 1.74 μM) and UNC5115 (ITC, K D ¼ 1.14 μM and FP Assay, IC 50 ¼ 3.87 μM).
A biotinylated UNC5114 chemiprecipitated PRC2 components, EED, EZH2, and
SUZ12, confirming the endogenous interaction of UNC5114 with the Kme reader
pocket of EED without disrupting the PRC2 complex. Finally, both UNC5114 and
UNC5115 were able to competitively inhibit the PRC2 complex in the presence of
H3K27me3 [248].
Methyl-Readers and Inhibitors
381
pocket allowing for insertion of EED226 further into the pocket. EED226 did not
alter the conformation of the EBD-binding site. In vivo, EED226 induced the
regression of tumor xenografts in mouse models. Finally, EED226 was capable of
inhibiting cancer cell lines with acquired resistance to S-adenosylmethionine EZH2
inhibitors and also displayed a synergistic effect to inhibit cancer cell growth when
combined with EZH2 inhibitors [245].
Another high-throughput screening effort against EED, followed by structurebased medicinal chemistry efforts, led to the identification of A-395 (K i ¼ 0.4 nM) as
an allosteric inhibitor of PRC2 (Fig. 17) [246]. The compound selectively and
potently binds EED and inhibits the catalytic activity of the trimeric PRC2 complex
(EZH2-EED-SUZ12) in an in vitro radioactivity-based assay with an IC 50 of 18 nM.
Crystallographic studies and an AlphaLISA assay demonstrated the capability of
A-395 to compete with H3K27me3 for binding to the aromatic cage formed by the
WD40 repeat domain of EED (IC 50 ¼ 7 nM). Additionally, A-395 potently reduced
H3K27 methylation in rhabdoid tumor cells and inhibited proliferation of human
cancer cells, specifically DLBCL Pfeiffer and Karpas422, which are sensitive to
EZH2 inhibition. In DLBCL Pfeiffer cell xenograft models, A-395 displayed significant inhibition of tumor growth (84%). Finally, A-395 retained antitumor effects
against cell lines resistant to SET domain targeted EZH2 inhibitors [246].
A different approach to discover ligands of EED started from the identification of
JARID2 trimethylated at lysine 116 (JARID2-K116me3) as an EED-binding partner
and allosteric activator of PRC2 methyltransferase activity at low micromolar
potency (K D ¼ 3 μM), which is tenfold higher than the affinity of EED for
H3Kme7 in vitro [247]. The co-crystal structure of EED in complex with
JARID2-K116me3 (PDB: 4X3E) provided insight to the structural features necessary for binding, specifically at residues 114–118 [248]. Using a split-and-pool
synthesis method, a first-generation library of 1,029 peptides based on a truncated
JARID2 114–118 K116me3 scaffold (Fig. 18) was synthesized, and an on-bead magnetic enrichment screening approach for EED was used to identify lead compounds
[248]. Compound #1 (Fig. 18, compound numbering as in the original paper) was
identified and when evaluated by ITC had improved potency compared to
JARID2 114–118 K116me3 (K D ¼ 4.8 μM and 8.8 μM, respectively). A subsequent
second-generation library consisting of 4,410 compounds was designed to enhance
ligand affinity and compound #3 (Fig. 18, compound numbering as in the original
paper), which demonstrated similar binding values in both ITC and an EED FP assay
(K D ¼ 1.09 μM and IC 50 ¼ 1.65 μM, respectively) was identified. Crystallographic
studies and further modifications to the N- and C-terminus as well as replacement of
the Kme3 resulted in the more potent UNC5114 (ITC, K D ¼ 0.68 μM and FP assay
IC 50 ¼ 1.74 μM) and UNC5115 (ITC, K D ¼ 1.14 μM and FP Assay, IC 50 ¼ 3.87 μM).
A biotinylated UNC5114 chemiprecipitated PRC2 components, EED, EZH2, and
SUZ12, confirming the endogenous interaction of UNC5114 with the Kme reader
pocket of EED without disrupting the PRC2 complex. Finally, both UNC5114 and
UNC5115 were able to competitively inhibit the PRC2 complex in the presence of
H3K27me3 [248].
Methyl-Readers and Inhibitors
381