Complexity of Blocking Bivalent Protein-Protein Interactions: Development of a Highly Potent Inhibitor of the Menin-Mixed-Lineage Leukemia Interaction

Article abstract of DOI:10.1021/acs.jmedchem.8b00071

The protein-protein interaction between menin and mixed-lineage leukemia 1 (MLL1) plays an important role in development of acute leukemia with translocations of the MLL1 gene and in solid tumors. Here, we report the development of a new generation of menin-MLL1 inhibitors identified by structure-based optimization of the thienopyrimidine class of compounds. This work resulted in compound 28 (MI-1481), which showed very potent inhibition of the menin-MLL1 interaction (IC₅₀ = 3.6 nM), representing the most potent reversible menin-MLL1 inhibitor reported to date. The crystal structure of the menin-28 complex revealed a hydrogen bond with Glu366 and hydrophobic interactions, which contributed to strong inhibitory activity of 28. Compound 28 also demonstrates pronounced activity in MLL leukemia cells and in vivo in MLL leukemia models. Thus, 28 is a valuable menin-MLL1 inhibitor that can be used for potential therapeutic applications and in further studies regarding the role of menin in cancer.

Full text of DOI:10.1021/acs.jmedchem.8b00071

Subscriber access provided by Kaohsiung Medical University
Article
The complexity of blocking bivalent protein-protein
interactions: development of a highly potent inhibitor
of the menin-Mixed Lineage Leukemia (MLL) interaction
Dmitry Borkin, Szymon Klossowski, Jonathan Pollock, Hongzhi Miao, Brian Linhares, Katarzyna Kempinska,
Zhuang Jin, Trupta Purohit, Bo Wen, Miao He, Duxin Sun, Tomasz Cierpicki, and Jolanta Grembecka
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00071 • Publication Date (Web): 08 May 2018
Downloaded from http://pubs.acs.org on May 8, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The complexity of blocking bivalent proteinꢀprotein interactions: development of a highly
potent inhibitor of the meninꢀMixed Lineage Leukemia (MLL) interaction
a,#
a,#
a
a
a
Dmitry Borkin , Szymon Klossowski , Jonathan Pollock , Hongzhi Miao , Brian M. Linhares ,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
a
a
b
b
b
Katarzyna Kempinska , Zhuang Jin , Trupta Purohit , Bo Wen , Miao He , Duxin Sun , Tomasz
a
a,*
Cierpicki , Jolanta Grembecka
a
Department of Pathology, University of Michigan, Ann Arbor, MI, 48109, USA
College of Pharmacy, University of Michigan, Ann Arbor, MI, 48109, USA
D.B. and S. K. contributed equally to this work
b
#
*
Corresponding author;
Jolanta Grembecka, PhD
Associate Professor,
Department of Pathology,
University of Michigan
1
150 West Medical Center Dr, MSRB I, Room 4510D
Ann Arbor, MI, 48108
eꢀmail: jolantag@umich.edu
Tel. 734ꢀ615ꢀ9319
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract
The proteinꢀprotein interaction between menin and Mixed Lineage Leukemia 1 (MLL1) plays an
important role in development of acute leukemia with translocations of the MLL1 gene and in
solid tumors. Here, we report the development of a new generation of meninꢀMLL1 inhibitors
identified by structureꢀbased optimization of the thienopyrimidine class of compounds. This
work resulted in compound 28 (MIꢀ1481), which showed very potent inhibition of the meninꢀ
MLL1 interaction (IC50 = 3.6 nM), representing the most potent reversible meninꢀMLL1
inhibitor reported to date. The crystal structure of the meninꢀ28 complex revealed a hydrogen
bond with Glu366 and hydrophobic interactions, which contributed to strong inhibitory activity
of 28. Compound 28 also demonstrates pronounced activity in MLL leukemia cells and in vivo
in MLL leukemia models. Thus, 28 is a valuable meninꢀMLL1 inhibitor that can be used for
potential therapeutic applications and in further studies regarding the role of menin in cancer.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
ACS Paragon Plus Environment

Page 3 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Introduction
Menin is a scaffold protein that is involved in a network of proteinꢀprotein interactions,
among which the interaction with the Mixed Lineage Leukemia 1 (MLL1; hereafter referred as
MLL) protein or MLL fusion proteins have been recognized as highly relevant to human
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
diseases, specifically to cancer. The MLL1 gene is translocated in 5ꢀ10% of acute leukemia
cases, and this translocation leads to the expression of MLL fusion proteins, which are critical
2
ꢀ3
drivers of leukemogenesis. The MLL leukemia patients have an approximately 35% fiveꢀyear
4
ꢀ5
survival rate, which reflects poor response to currently available treatments and supports the
need for the development of new therapies. The interaction of menin with MLL fusion proteins
2
, 6ꢀ
plays a key role in leukemogenesis, which was demonstrated before using genetic approaches.
7
The importance of this interaction in MLL leukemia has been further validated by small
8ꢀ12
molecule inhibitors of the meninꢀMLL interaction that we have reported in earlier studies.
Furthermore, the interaction of menin with the wild type MLL plays an important role in solid
1
3
14
tumors, including castration resistant prostate cancer, Ewing sarcoma, hepatocellular
15ꢀ16
17
carcinoma (HCC)
and pediatric gliomas. Therefore, the development of small molecule
inhibitors of the meninꢀMLL interaction with optimized drugꢀlike properties may lead to new
therapeutics for acute leukemia and a subset of solid tumors.
The development of potent inhibitors with favorable drugꢀlike properties that target proteinꢀ
1
8ꢀ19
protein interactions (PPI) is still a challenge.
The specific architecture of proteinꢀprotein
interfaces (e.g. the lack of wellꢀdefined pockets and large contact areas) frequently implies the
need for increasing the molecular weight of PPI inhibitors, which makes it difficult to achieve
20
strong potency and favorable drugꢀlike properties. Nevertheless, successful examples of PPI
2
1ꢀ24
inhibitors exist,
including compounds advanced to clinical trials (e.g. Bclꢀ2 family inhibitors,
2
5ꢀ26
p53 activators
), thus supporting efforts in this direction. Furthermore, the recent FDA
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
approval of the BCLꢀ2 inhibitor Venetoclax demonstrates that PPI inhibitors can be valuable
therapeutics, despite challenges it obtaining favorable drugꢀlike properties for such compounds
as a result of inherently higher than average molecular weight (M > 800 Da for Venetoclax).
w
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The proteinꢀprotein interaction between menin and MLL or MLL fusion proteins involves
2, 6
the Nꢀterminal fragment of MLL (MLL4ꢀ43
)
, which binds to menin with a strong binding
affinity (K = 6.8 nM) by utilizing the bivalent binding mode, with MBM1 (meninꢀbinding motif
d
1
, MLL4ꢀ15) and MBM2 (meninꢀbinding motif 2, MLL23ꢀ43) motifs involved in the interactions
27
with menin. Structural studies have revealed that MLL binds to a large central cavity on
9
, 28
menin,
thus supporting the existence of pockets to bind small molecules. Indeed, our previous
studies have validated that the meninꢀMLL interaction can be effectively blocked with small
8ꢀ12
molecules that bind to the MLL binding site on menin.
However, due to the extended meninꢀ
MLL interface resulting from the bivalent binding mode of MLL to menin, optimization of the
1
2
potency and drugꢀlike properties of meninꢀMLL inhibitors remains challenging.
Our previous efforts resulted in the development of the thienopyrimidine class of meninꢀ
MLL inhibitors, with MIꢀ463, MIꢀ503 and MIꢀ538 demonstrating the most potent inhibitory
1
0, 12
activity (Figure 1a).
These compounds selectively block proliferation of human MLL
leukemia cells (GI50 = 200ꢀ500 nM), demonstrate specific mechanism of action and delay
10, 12
leukemia progression in mice models of MLL leukemia.
MIꢀ503 also demonstrated activity
1
3
14
16
in prostate cancer, Ewing sarcoma and hepatocellular carcinoma models. However, meninꢀ
MLL inhibitors with improved therapeutic potential over MIꢀ503, namely with increased
potency (e.g. IC < 10 nM to make them comparable or more potent than MLL , GI < 50
5
0
4ꢀ43
50
nM and >100ꢀfold selectivity to MLL leukemia cells, strong reduction in leukemia burden and
>50% decrease in target gene expression in vivo) are still required. In this study we report
4
ACS Paragon Plus Environment

Page 5 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
structureꢀbased optimization of the MIꢀ463 and MIꢀ503 meninꢀMLL inhibitors by exploring a
collection of saturated sixꢀmembered rings introduced at the indole nitrogen of these compounds
to further improve their activity and drugꢀlike properties. These efforts resulted in the
identification of compound 28 (MIꢀ1481), which demonstrated low nanomolar activity (IC =
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
3
.6 nM) for the inhibition of the interaction between menin and the bivalent MLL4ꢀ43
encompassing the entire menin binding motif on MLL. Thus, 28 demonstrates >10ꢀfold
improved inhibitory activity over MIꢀ503, representing the most potent reversible small
molecule meninꢀMLL inhibitor reported to date. Compound 28 also demonstrated substantially
increased activity in MLL leukemia cells (~10ꢀfold) over the previous generation of reversible
1
0, 12
meninꢀMLL inhibitors,
activity in a mice model of MLL leukemia. We also report the coꢀcrystal structure of the meninꢀ
8 complex, which revealed that both polar and hydrophobic interactions likely contribute to the
high selectivity towards MLL leukemia cells and pronounced in vivo
2
strong inhibitory activity of this compound. This compound could serve as a valuable chemical
probe to further study the role of menin in pathological and physiological states and could lead to
the development of a new therapeutic agent for tumors that rely on meninꢀMLL interaction.
Results and Discussion
Biochemical assay to assess potent meninꢀMLL inhibitors
We have recently reported the thienopyrimidine class of meninꢀMLL interaction, with the most
10, 12
potent MIꢀ463 (compound 1) and MIꢀ503 (compound 2), Figure 1a, Table 1.
These
compounds bind to menin with K ~ 10 nM and block the interaction of menin with a short
d
MLLꢀderived peptide MBM1 (MLL4ꢀ15) with nanomolar inhibitory activities (IC ~15 nM for
50
both compounds) as measured in the fluorescence polarization (FP) assay, Figure 1a. However,
in the FP competition assay the low end of inhibitor K values that can be accurately measured is
i
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the K value of the fluorescent ligand, and for inhibitors more potent than the fluorescent probe
d
the IC values are independent of inhibitor potency but linearly proportional to the K value of
5
0
d
2
9
the fluorescence ligand. Based on the modest binding affinity of the fluorescein labeled (FLSN)
2
7
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
MBM1 to menin (K = 49 nM) we concluded that meninꢀMLL inhibitors with low nanomolar
d
binding affinity or better would reach the detection limit of the competition FP assay that uses
FLSN_MBM1. To avoid this issue, we utilized the bivalent fragment of MLL (MLL4ꢀ43) in the
competition FP assay for testing of the new generation of meninꢀMLL inhibitors. The MLL4ꢀ43
6, 27
includes the entire menin binding motif (MBM1 and MBM2)
and binds to menin with K =
d
6
.8 nM as determined by the Isothermal Titration Calorimetry (ITC). First, we determined the
binding affinity of the fluorescein labeled MLL4ꢀ43 peptide (FLSNꢀMLL4ꢀ43) to menin in this new
FP assay, resulting in K = 1.0 nM Supplementary Figure 1, which remains in a good
d
27
agreement with the results from the ITC experiment. We then applied the FP competition assay
with FLSNꢀMLL4ꢀ43 to test the inhibitory activity of MIꢀ463 and MIꢀ503 meninꢀML inhibitors,
resulting in IC values of 32 and 33 nM, respectively, Figure 1a, Table 1. Because MLL
5
0
4ꢀ43
encompasses the entire menin binding motif, these data support that the thienopyrimidine
compounds (e.g. MIꢀ463 and MIꢀ503) can effectively displace the bivalent MLL fragment from
menin, which is an important feature in the context of functional studies. Overall, based on these
results we concluded that the FP assay utilizing the biꢀvalent MLL4ꢀ43 peptide represents an
appropriate approach to provide accurate activity measurements for low nanomolar inhibitors of
the meninꢀMLL interactions. Thus, we used this assay to assess the in vitro activity of all
compounds reported in this study.
Development of meninꢀMLL inhibitors with saturated rings substituting indole nitrogen
6
ACS Paragon Plus Environment

Page 7 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In our previous studies, we developed MIꢀ463 and its analogues with different substitutions at
the indole ring, including aliphatic and aromatic substituents on indole nitrogen (e.g. MIꢀ503),
Figure 1a. However, these modifications did not result in increased inhibitory activity as
1
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
compared to MIꢀ463. To further optimize these compounds we performed detailed analysis of
the coꢀcrystal structure of menin in complex with MIꢀ503, which revealed three hydrogen bonds
1
0
(
with Tyr276, Trp341 and Glu366) and hydrophobic interactions with menin, Figure 1b.
Furthermore, we found that the methylꢀpyrazole ring substituting indole nitrogen in MIꢀ503 is
partly solvent exposed, with two negatively charged residues on menin, Glu363 and Glu366,
approaching this ring, Figure 1b. Based on this analysis we designed various saturated rings
harboring positively charged and/or uncharged polar groups to be introduced at the indole
nitrogen of MIꢀ463 to form electrostatic interactions or hydrogen bonds with this area of the
binding site on menin.
First, we explored several saturated rings with short aliphatic linkers substituting indole
nitrogen of MIꢀ463, Table 1. Introduction of the piperazine or piperidine rings, which resulted in
compounds 3 and 4, Table 1, led to a ~2ꢀfold improved inhibitory activity over MIꢀ463 (IC =
5
0
1
5 and 14 nM for 3 and 4, respectively). On the other hand, incorporation of the morpholine
(compound 5) or morpholinone (compound 6) rings connected to indole through an ethyl linker
resulted in less pronounced activity improvement (<1.5ꢀfold) over MIꢀ463, Table 1.
Interestingly, the introduction of the methylꢀmorpholinone substituent at indole nitrogen
(compound 7, MIꢀ568) resulted in an over 4ꢀfold improvement in the IC value (IC = 7.5 nM
5
0
50
for the racemic mixture of 7) as compared to MIꢀ463, Table 1. Thus, although 7 represents the
racemic mixture of two enantiomers, this compound demonstrates low nanomolar inhibition of
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the meninꢀMLL interaction, outperforming other compounds with various saturated rings
introduced to the scaffold, Table 1.
We then assessed the cellular activity in MLL leukemia cells and the microsomal stability
of compounds 3ꢀ7, Table 1. Notably, compound 7 demonstrated the most pronounced effect on
proliferation of the MLLꢀAF9 transformed leukemia cells (GI = 50 nM), representing ~5ꢀfold
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
stronger cellular activity than MIꢀ463, Table 1. The cellular activities of compounds 3ꢀ6 are less
pronounced and range from 0.15ꢀ0.81 µM, Table 1. Interestingly, despite improved inhibitory
activity, compounds 3 and 4 showed weaker cell growth inhibition as compared to MIꢀ463,
which may have resulted from the positive charge in the piperazine or piperidine rings of these
compounds, possibly affecting their cellular permeability. In addition, compounds 3, 4 and 7
showed the best microsomal stability measured in mouse liver microsomes (T1/2 > 60 min),
representing substantial improvement over MIꢀ463 or MIꢀ503, Table 1. Furthermore, compound
7 has also the lowest clogP value (Table 1), thus supporting favorable polarity and solubility.
Overall, these results support that the analogues of MIꢀ463 with saturated rings introduced at
indole nitrogen represent valuable candidates for further optimization.
Optimization of analogues with piperazine ring
Since compound 3 showed 2ꢀfold improvement in inhibitory activity and substantially increased
metabolic stability over MIꢀ463 (Table 1), we pursued further optimization of 3 by introducing
substituents to the piperazine ring. First, we introduced aliphatic groups, such as methyl
(compound 8) and nꢀpropyl (compound 9), but they did not improve substantially the in vitro
inhibitory activity or cell growth inhibition in MLLꢀAF9 cells over 3, Table 2. Analysis of the
1
0
crystal structure of the meninꢀMIꢀ503 complex revealed the presence of Arg330, which could
8
ACS Paragon Plus Environment

Page 9 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
potentially be involved in hydrogen bond/salt bridge with ligands that harbor the appropriate
substituent at the indole nitrogen of MIꢀ463. Therefore, we introduced the formamide group to
the piperazine ring (compound 10), but this modification did not improve inhibitory activity
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
IC = 13 nM for 10), although activity in the MLLꢀAF9 transformed cells demonstrated a 2ꢀ
50
fold improvement for 10 over 3, Table 2. This was likely due to the neutralization of the positive
charge in 3, which could have positively affected the cellular permeability. Further substitution
of the amide with methyl (compound 11) did not significantly change the inhibitory activity,
while ethyl analog (compound 12) showed an over 6ꢀfold reduction in inhibitory activity (IC =
5
0
8
3 nM) as compared to 10. Interestingly, the analogue with the monofluoromethyl group
(compound 13, MIꢀ853) demonstrated slightly improved inhibitory activity (IC = 10 nM) over
5
0
1
0, while the diꢀ and trifluoromethyl analogues (14 and 15) were 2ꢀ5ꢀfold less potent than 10,
Table 2. Improved in vitro activity of 13 resulted in more potent cell growth inhibition of the
MLLꢀAF9 transformed cells observed for this compound (GI = 75 nM), Table 2. Furthermore,
5
0
introduction of a larger hydrophobic group substituting amide in 10, which resulted in
compounds 16 and 17, led to a 3ꢀ4ꢀfold reduced inhibitory activity, Table 2. Finally, we
introduced the 2ꢀpropanamide group (compound 18), but did not observe increased inhibitory
activity for the racemic mixture of this compound over 10, Table 2.
To further explore substitutions at the piperazine ring of 3, we substituted this ring with
sulfonyl or sulfonamide groups (compounds 19ꢀ23), Table 2. Among these compounds the
analogue with the methyl sulfonamide (compound 19), demonstrated the most pronounced
inhibitory activity for blocking the meninꢀMLL interaction (IC50 = 13 nM), Table 2. The
structureꢀactivity relationship for the sulfonyl analogues (compounds 19ꢀ23) is similar to the
corresponding amide derivatives, Table 2, with reduced inhibitory activity upon increasing the
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
size of the aliphatic substituent (e.g. compounds 20 and 21, Table 2). The sulfamide derivatives
2
2 and 23 showed about 2ꢀfold reduction in inhibitory activity as compared to the methylsulfone
analogue 19, Table 2.
Despite comparable or weaker inhibitory activity in in vitro biochemical assay, the
majority of compounds with amides, sulfones or sulfonamides substituting the piperazine ring in
demonstrated stronger cell growth inhibition of the MLLꢀAF9 transformed cells (GI values
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
50
below 200 nM) as compared to compound 3, Table 2. This may have resulted from the reduced
basicity of these compounds over 3, which could have affected their cellular permeability.
Importantly, the piperazine analogues demonstrated good correlation of in vitro inhibitory
activity with cell growth inhibition, with compound 13 showing the strongest inhibitory activity
(
IC = 10 nM) for blocking the meninꢀMLL interaction and inhibition of cell proliferation in
50
MLLꢀAF9 leukemia cells (GI = 75 nM), Table 2. On the other hand, the microsomal stability
5
0
of the amide and sulfonamide analogues is reduced as compared to compound 3 (Table 2),
although compounds 19 and 22 still retain relatively good microsomal stability (T 1/2 > 25 min).
Development of analogues with substituted piperidine ring
Because unsubstituted piperazine (3) and piperidine (4) derivatives have similar inhibitory
activity in blocking the meninꢀMLL interaction, we also explored several analogues derived
from 4, Table 3. In contrast to the corresponding piperazine analogues, the introduction of the
amide group (compound 24) and its further substitution with methyl (compound 25) resulted in a
5
ꢀ10 fold reduction in inhibitory activity as compared to 4, Table 3. Similarly, the incorporation
of the sulfonamide (compound 26) or sulfamide (compound 27) groups to this scaffold resulted
in >2ꢀfold reduced inhibitory activity over 4, Table 3. Based on these results we concluded that
10
ACS Paragon Plus Environment

Page 11 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the reduced basicity of the piperidine series (Table 3) makes these compound less potent meninꢀ
MLL inhibitors than the corresponding piperazine derivatives (Table 2) possibly due to the
absence of longꢀrange electrostatic interactions with Glu363 and/or Glu366 side chains, Figure
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
b. Interestingly, despite weaker in vitro inhibitory activity, all piperidine analogues
demonstrated comparable or increased activity in MLL leukemia cells over 4 (e.g. compound 26
has over 3ꢀfold improved GI50 as compared to 4, Table 3). Furthermore, compounds 24ꢀ27
demonstrate good microsomal stability, with the sulfonyl (26) or sulfonamide (27) analogues
exhibiting the longest halfꢀlife in mouse liver microsomes (T1/2 > 60 min).
Structural studies reveal the binding mode of piperazine derivatives to menin
To understand the molecular basis of the interactions of the piperazine analogues with menin, we
coꢀcrystallized menin in complex with 13, which represents the most potent compound from this
series (Table 2) and solved a highꢀresolution crystal structure of the complex, Figure 2a,
Supplementary Table 1. The crystal structure revealed that the binding mode of 13 to menin is
similar as that observed for MIꢀ503, with two hydrogen bonds (with Tyr276 and Trp341) and
hydrophobic interactions retained in 13 within the core structure preserved in both compounds
(Figure 1b, 2a). The differences in the binding mode of 13 versus MIꢀ503 were only observed
for binding of the substituent introduced at indole nitrogen, which varies between these two
compounds, Table 1, 2. The ethyl linker connecting indole with the piperazine ring in 13 is
involved in hydrophobic interactions with the side chains of Glu363, Met322, Val367 and
Glu366 on menin, Figure 2a. One of the nitrogen atoms in the piperazine ring of 13 points
towards the carboxyl group of Glu366 (distance ~4.5Å), supporting the longꢀrange electrostatic
interactions, Figure 2a. Furthermore, one side of the piperidine ring in 13 is involved in the
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
hydrophobic contacts with Val367 and Val371 on menin, while the other side of this ring in
exposed to the solvent, Figure 2a. In addition, the oxygen atom from the fluoroacetamide group
of 13 forms a hydrogen bond with the side chain of Arg330 (3.3 Å), while the fluoromethyl
group is involved in the hydrophobic interactions with the side chains of Cys329 and Arg330
with the fluorine atom approaching Cys329 and Gly326, Figure 2a. Overall, favorable
interactions of the fluoroacetamide group in 13 with menin result in the most pronounced activity
of this compound among all amide or sulfonamide derivatives, Table 2, 3. The crystal structure
also demonstrates limited space in the binding site to accommodate hydrophobic groups larger
than acetamide, reflected by the reduced activity of such analogs (e.g. compounds 12 and 14ꢀ17,
Table 2). Overall, the crystal structure of menin in complex with 13 supports that both
hydrophobic contacts and hydrogen bond with Arg330 on menin contribute to the pronounced
inhibitory activity of this compound.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Development of compound 28 as a highly potent meninꢀMLL inhibitor
Exploring different saturated rings introduced at indole nitrogen led to the identification of 7,
which demonstrated the most potent inhibitory activity for blocking the meninꢀMLL interaction
(
IC = 7.5 nM for the racemic mixture), Table 1. This compound harbors a morpholinone ring
50
introduced to the indole nitrogen through a methyl linker, Table 1. In addition to exhibiting the
most potent in vitro activity, compound 7 also demonstrated the most pronounced cell growth
inhibition in the MLLꢀAF9 leukemia cells (GI = 50 nM), Table 1. To assess the activity of
5
0
individual enantiomers of this compound we synthesized both of them, resulting in 28 (S
enantiomer, MIꢀ1481) and 29 (R enantiomer, MIꢀ1482), Table 4. The S enantiomers (28)
demonstrated a ~2ꢀfold improvement in inhibitory activity for blocking the meninꢀMLL
12
ACS Paragon Plus Environment

Page 13 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
interaction (IC = 3.6 nM) over the racemic mixture (compound 7), Figure 3a, Table 4. Thus,
5
0
2
8 demonstrates about 10ꢀfold increase in inhibitory activity over MIꢀ463 and MIꢀ503 (Table1),
representing the most potent reversible meninꢀMLL inhibitor reported to date. In contrast, the R
enantiomer (compound 29) demonstrated a ~35ꢀfold reduction in inhibitory activity in blocking
the meninꢀMLL interaction (IC = 123 nM for 29) as compared to 28, Figure 3a, Table 4. Thus,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
2
8 and 29 provide a valuable pair of enantiomers with distinct inhibitory activity to study meninꢀ
MLL inhibition in biological systems.
Molecular basis of meninꢀMLL inhibition by compound 28
To understand the molecular basis of the strong inhibition of the meninꢀMLL interaction
by 28, we coꢀcrystallized menin in complex with this compound, resulting in a highꢀresolution
crystal structure, Supplementary Table 1, Figure 2b. We found a similar interaction pattern of
28 with menin to the one observed in the meninꢀMIꢀ503 complex for the structural fragment
retained in both compounds (e.g. hydrogen bonds with Tyr276 and Trp341 and similar pattern of
hydrophobic interactions), Figure 1b, 2b. Again, the differences were found in the binding of
substituents introduced at indole nitrogen. Specifically, the methyl group linking indole with the
morpholinone ring in 28 is involved in the hydrophobic interactions with Glu363, Met322 and
Val 367, Figure 2b. Furthermore, one side of the morpholinone ring in 28 is involved in
hydrophobic contacts with the side chains of Glu363, Glu366 and Val367 on menin, while the
amino portion of the amide forms a short hydrogen bond (2.6 Å) with the carboxyl group of
Glu366 (Figure 2b). This hydrogen bond, which was not observed for the piperazine analogues
(e.g. in the meninꢀ13 complex), most likely contributes to the pronounced inhibitory activity of
28. The carbonyl carbon and the adjacent methylene group in 28 are engaged in the hydrophobic
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
contacts with Val371, while the ether oxygen is solvent exposed and does not form major
contacts with menin, Figure 2b. In contrast, the pyrazole ring introduced at the indole nitrogen
of MIꢀ503 is mostly solvent exposed, with one hydrogen bond formed with Glu366, Figure 1b.
To understand better the improvement in inhibitory activity for 28 over MIꢀ503 we
performed the bioꢀlayer interferometry (BLI) studies and compared binding affinities of both
compounds to menin. This study demonstrated that the association rates to menin are similar for
both compounds, however, the dissociation rate is significantly slower for 28 over MIꢀ503,
resulting in the K value of ~ 3ꢀfold improved for 28 (K = 9 nM) as compared to MIꢀ503 (K =
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
d
d
d
2
4 nM), Supplementary Figure 2. Furthermore, we have also performed structural analysis of
menin interactions with ligands by overlapped the crystal structures of menin in complex with:
the biꢀvalent MLL peptide encompassing both MBM1 and MBM2 motifs, MIꢀ503 and 28,
Figure 2c. This analysis have revealed that both compounds occupy the same region of the
binding site on menin where MBM1 motif binds, however 28 approaches more closely the
region occupied by MBM2, particularly Arg24 from MLL, as compared to MIꢀ503, Figure 2c.
Therefore, 28 might be able to more effectively displace the biꢀvalent MLL peptide from its
binding to menin likely due to stronger interference with the MBM2 portion of MLL, as reflected
by ~10ꢀfold improved inhibitory activity of 28 over MIꢀ503. Overall, both structural data (e.g.
extensive interactions with menin and interference with the MBM2 motif) as well as binding
kinetics support 28 as a more effective inhibitor of the meninꢀMLL interaction over the previous
generation of meninꢀMLL inhibitors (e.g. MIꢀ503 or MIꢀ463), Tables 1, 4. To better understand
the molecular basis of the differential inhibitory activity between both enantiomers of 7, we also
solved the coꢀcrystal structure of menin in complex with 29 (R enantiomer of 7), Figure 2d,
Supplementary Table 1. The core structure of 29 binds to menin in the same way as observed
14
ACS Paragon Plus Environment

Page 15 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
for 28, but the orientation of the morpholinone ring in 29 is different from that of 28, Figure 2d,
which results in significantly reduced contacts of this ring with menin. The most striking
difference in the binding mode of 29 and 28 is the lack of a hydrogen bond between the amide
nitrogen in 29 and the carboxyl group of Glu366 in menin, which is present when 28 binds to
menin, Figure 2b,d. Compound 29 also has limited hydrophobic contacts with menin as
compared to 28, as 29 engages only a small portion of the morpholinone ring in the interactions
with Glu366, Val367 and Val371 on menin, Figure 2d. Overall, we concluded that limited
interactions of the morpholinone ring in 29 with menin, in particular the lack of a hydrogen bond
with Glu366, result in a 35ꢀfold reduction in inhibitory activity of this compound to menin as
compared to 28.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compound 28 demonstrates potent on target activity in MLL leukemia cells
To assess whether potent inhibition of the meninꢀMLL interaction by 28 results in strong activity
of this compound in functional assays, we tested 28 in several leukemia cell lines. Importantly,
2
8 markedly reduced cell growth of murine bone marrow cells transformed with MLLꢀAF9, with
a GI value of 34 nM (Table 4, Figure 3b). In contrast, limited activity of this compound was
50
observed in murine bone marrow cells transformed with the control oncogenes Hoxa9/Meis1
(
HMꢀ2, GI > 6 µM), Figure 3b. Thus, 28 demonstrated over 200ꢀfold selectivity towards MLL
50
fusion protein transformed cells, Table 4, demonstrating substantial improvement in selectivity
over the previous generation of meninꢀMLL inhibitors (e.g. selectivity index, SI ~ 27 for MIꢀ503
12
in MLL leukemia cells). When tested in human MLL leukemia cell lines, 28 also showed
pronounced cell growth inhibition, with GI values of 36 nM in MV4;11 (harboring MLLꢀAF4
5
0
fusion protein) and 61 nM in MOLM13 (harboring MLLꢀAF9), Figure 3c. Importantly, no
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
substantial cell growth inhibition was observed in the control human leukemia cell lines K562
and U937, which do not harbor MLL translocations (GI > 6 µM), Figure 3c. Therefore, 28
5
0
shows the most pronounced activity in MLL leukemia cells among all reversible meninꢀMLL
1
0, 12
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
inhibitors reported to date, Tables 1ꢀ4,
demonstrating ~7ꢀ10ꢀfold improvement in the
10
activity in MLL leukemia cells over MIꢀ463 or MIꢀ503 (Table 1). As pointed out above, the
increased cellular activity of 28 is associated with ~10ꢀfold improved inhibitory activity of this
compound in blocking the meninꢀMLL interaction as compared to MIꢀ463 or MIꢀ503, Table 1,
4
. In contrast, compound 29, which represents the R enantiomer of 7, showed limited activity in
MLL leukemia cells (over 16ꢀfold weaker than 28), Figure 3d, in agreement with the
substantially reduced inhibitory activity of 29, Table 4.
To validate the on target activity of 28 in MLL leukemia cells, we assessed the effect of
this compound on the expression level of the MLL fusion target genes: HOXA9, MEIS1 and
DLX2, all of which are required for the MLLꢀfusion protein mediated leukemogenesis. Six days
of treatment of the MV4;11 human MLL leukemia cells with 28 resulted in a substantial
downregulation of the MLL fusion protein target genes, with a strong reduction in the expression
level of HOXA9 and MEIS1 at 100 nM of 28, Figure 3e. The effect of 28 on the expression level
of DLX2 was even more pronounced, Figure 3e. Furthermore, marked upregulation of the
ITGAM differentiation marker was observed upon treatment of MV4;11 cells with 28 (Figure
3
e), supporting reprograming of these cells from primitive blasts to more mature blood cells.
Overall, 28 demonstrates a pronounced effect on the expression level of MLL fusion target
genes, which strongly supports the on target activity of this compound in MLL leukemia cells.
Compound 28 strongly inhibits leukemia progression in vivo
16
ACS Paragon Plus Environment

Page 17 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
We then performed pharmacokinetic (PK) studies in mice with 28 and found relatively high
exposure of this compound in blood plasma using i.v. (intravenous) and i.p. (intraperitoneal)
administration (Supplementary Figure 3), however, oral bioavailability of this compound was
low (<5%). Based on these results we selected i.p. as a way of administration of 28 for in vivo
efficacy studies. To establish a potential therapeutic value of 28, we tested the effect of this
compound in a mice model of MLL leukemia. For this purpose, we utilized the
xenotransplantation model of the MLL leukemia with MV4;11ꢀluc cells (expressing luciferase)
injected into the tail vein of NSG mice. Treatment with 28 (80 mg/kg, i.p., n = 8) or vehicle (n =
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
8
) was initiated 13 days after transplantation of mice with the MV4;11 cells to account for a
compound effect in an advanced leukemia model, and was continued for 21 consecutive days.
No reduction in the body weight of mice was observed during the entire treatment period with
2
8, Supplementary Figure 4. To assess treatment efficacy, the bioluminescence level in mice
was measured at days 0, 7, 14 and 21 after initiating the treatment. Very pronounced (~75%) and
consistent reduction in the luciferase signal was observed over the course of the experiment upon
treatment of mice with 28 as compared to the vehicle treated mice, Figure 4a,b. These results
demonstrate the pronounced effect 28 on leukemia progression in an advanced model of MLL
leukemia.
To validate the onꢀtarget in vivo activity of 28 we performed gene expression studies in
bone marrow and spleen samples isolated from mice transplanted with MV4;11 cells and treated
with 28 or vehicle. Treatment with 28 was initiated 21 days after transplantation of mice with the
leukemic cells to account for the effect of 28 in advanced leukemia model. Compound 28 (or
vehicle) was administered for six consecutive days, followed by collection of the bone marrow
and spleen samples. Importantly, substantial reduction in the spleen weight and size was
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
observed for mice treated with 28 compared to the vehicle control, Figure 4c. Furthermore, the
level of leukemic blasts (human CD45+ cells) was very high (~75%) in the bone marrow cells
isolated from vehicle treated mice, which was indicative of advanced leukemia, while the level
of leukemic blasts was substantially reduced (to ~35%) in mice treated with 28, Figure 4d. We
then performed gene expression studies in bone marrow and spleen samples isolated from mice
treated with 28 or vehicle using qRTꢀPCR and found that treatment with 28 resulted in a strong
reduction (2ꢀ5 fold) in the expression level of MLL fusion target genes MEIS1 and HOXA9,
Figure 4e. Furthermore, an over 4ꢀ6ꢀfold increase in the expression of ITGAM differentiation
marker was observed in the bone marrow and spleen samples upon treatment with 28, Figure 4e.
Overall, the results from gene expression studies demonstrate pronounced, onꢀtarget activity of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
8 in the advanced mouse model of MLL leukemia, thus supporting the potential therapeutic
value of this compound.
Chemistry
Synthesis of compounds 3 and 4 was performed using the synthetic route shown in
Scheme 1. Two major intermediates: 5ꢀformylꢀ4ꢀmethylꢀ1Hꢀindoleꢀ2ꢀcarbonitrile (30) and Nꢀ
(piperidinꢀ4ꢀyl)ꢀ6ꢀ(2,2,2ꢀtrifluoroethyl)thieno[2,3ꢀd]pyrimidinꢀ4ꢀamine (33) were obtained
10
according to the procedure described before. The alcohols 31a,b were converted into mesylates
followed by the substitution with the aldehyde 30 in the presence of cesium carbonate, which
afforded aldehydes 32a,b. (Scheme 1). Compounds 32a,b were then used in the reductive
amination reaction with the secondary amine 33 utilizing sodium triacetoxyborohydride as the
reducing agent to afford compounds 34a,b. Removal of the tertꢀbutoxycarbonyl (Boc) group in
the presence of tin (IV) chloride resulted in compounds 3 and 4. Compound 5 bearing the
18
ACS Paragon Plus Environment

Page 19 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
morpholinone ring was obtained using the strategy depicted in Scheme 2. The 4ꢀ(2ꢀ
chloroethyl)morpholine (35) was used for the indole alkylation of aldehyde 30, resulting in
compound 36, Scheme 2. The reductive amination of aldehyde 36 with the amine 33 resulted in
compound 5, Scheme 2.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The racemic mixtures of compounds 6 and 7 were prepared according to the synthetic
route presented in Scheme 3. The intermediates 39a and 39b were obtained using the methods
3
0ꢀ31
described before.
Briefly, the commercially available 3ꢀaminoꢀ1,2ꢀpropanediol (37a) and
methyl 4ꢀaminoꢀ3ꢀhydroxybutanoate (37b) were subjected to acylation with the chloroacetic
acid, resulting in compounds 38a and 38b, respectively. Direct cyclisation of 38a with potassium
tꢀbutoxide afforded alcohol 39a. Homologous alcohol 39b was achieved by reduction of the ester
group in 38b followed by the analogous cyclisation with potassium tꢀbutoxide as a base (Scheme
3
). The alcohols 39a,b were then converted to the corresponding mesylates and used for
alkylation of the indol 30 to afford aldehydes 40a and 40b. Reductive amination reactions of the
intermediate 33 with aldehydes 40b or 40a led to the target compounds 6 and 7, respectively
(
Scheme 3).
The synthetic route for compounds 8 and 9 harboring the piperazine ring in the structure
is shown in Scheme 4. Briefly, both compounds were prepared from 3 using either formaline
for 8) or propionaldehyde (for 9) in the reductive amination reaction, Scheme 4. The piperazine
(
and piperidine analogues 10ꢀ27 were obtained according to the synthetic strategy depicted in
Scheme 5. In general, the amide series (compounds 10ꢀ17 and 24ꢀ25) was obtained by acylation
of the amines 3 or 4. The formamides 10 and 24 were prepared by formylation of 3 or 4,
respectively, with the methyl formate. The corresponding acyl chlorides were used to convert the
amine 3 into compounds 11, 12 or 17 with DIPEA serving as a base, Scheme 5. The analogous
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
reaction was performed with the amine 4 to afford compound 25. Synthesis of the amides
1
3,14,16 was performed by utilizing the coupling reaction between the amine 3 and the
corresponding carboxylic acids in the presence of 2ꢀ(1HꢀBenzotriazoleꢀ1ꢀyl)ꢀ1,1,3,3ꢀ
tetramethylaminium tetrafluoroborate (TBTU) serving as a condensing reagent, Scheme 5.
Compound 15 with the trifluoroacetyl moiety was prepared using the triflate anhydride as an
acylating agent. Alkylation of the nitrogen in the piperazine ring of 3 by the 2ꢀ
bromopropionamide led to the amine 18. Scheme 5.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The sulfonamide series (compounds 19ꢀ21, 26) was obtained using the sulfonylation
reactions with compounds 3 or 4 and the appropriate sulfonyl chlorides, Scheme 5. The twoꢀstep
synthetic approach was applied to afford compounds 22 and 27. The benzyl
(chlorosulfonyl)carbamate was used for sulfonylation of the amines 3 or 4, followed by the
cleavage of the Cbzꢀprotected amino group performed with Pd/C and ammonium formate, which
led to sulfamides 22 or 27, Scheme 5. Conversion of the amine 3 to compound 23 was achieved
using the methylsulfamoyl chloride in the presence of DIPEA.
Synthesis of compounds 28 and 29, which represent single enantiomers of 7, was
performed by applying the same methodology as for compound 7, Scheme 3 through utilizing
the nonꢀracemic 3ꢀaminopropaneꢀ1,2ꢀdiols: 37a(S) or 37a(R).
Conclusions
Identification of potent inhibitors targeting proteinꢀprotein interactions and optimization of their
drugꢀlike properties represents a challenge as such compounds frequently have large molecular
1
8ꢀ20
weight, which is required for effective blocking of PPI interfaces.
This is particularly true for
compounds that target extended proteinꢀprotein interfaces, including inhibitors of the BCLꢀ2
20
ACS Paragon Plus Environment

Page 21 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
26
1
family of proteins and others. On the other hand, recent FDA approval of the BCLꢀ2 inhibitor
Venetoclax demonstrates that PPI inhibitors with large molecular weight (>800 Da) can serve as
valuable therapeutics. In the current study, we report optimization of the small molecule
inhibitors of the meninꢀMLL interaction, with the major goal of improving the potency and drugꢀ
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3 9
like properties of these compounds. The MLL binding site on menin is large (over 5,000 Å ) as
menin binds long, bivalent MLL peptide (MLL4ꢀ43). This suggests that compounds with large
molecular weight may be required for effective inhibition of the meninꢀMLL interaction.
Using the crystal structure of menin in complex with MIꢀ503 we designed, synthesized
and evaluated a series of new analogues by exploring various saturated rings that substitute
indole nitrogen in MIꢀ503, Tables 1ꢀ4. The goal of this work was to increase the inhibitory
activity and metabolic stability of meninꢀMLL inhibitors by introducing substituents with
different physicochemical properties (e.g. charged versus uncharged substituents, hydrogen bond
donor or acceptor groups, etc.). These efforts resulted in compound 28, which demonstrates very
potent inhibition of the meninꢀMLL interaction (IC = 3.6 nM with MLL4ꢀ43). Therefore, 28
5
0
represents the most potent reversible small molecule inhibitor of the meninꢀMLL interaction
1
0, 12
reported to date
. About 10ꢀfold improved inhibitory activity for 28 over previous generation
of reversible meninꢀMLL inhibitors (e.g. MIꢀ503, MIꢀ463) results from stronger binding affinity
of this compound to menin, but also likely from more pronounced overlap with MBM2 binding
to menin, as supported by the structural data, Figure 2c. Furthermore, 28 demonstrates stronger
effect (>8ꢀfold) and better selectivity towards MLL leukemia cells than MIꢀ503, and it has also
more pronounced effect on the target gene expression in mouse models of MLL leukemia,
1
0
Figure 5E. On the other hand, increased polarity of 28 (clogP value reduced) and increased
molecular weight (by ~30 Da) over MIꢀ503 (M = 597 Da for 28) negatively affected oral
w
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
bioavailability of this compound. These results demonstrate the complexity of blocking biꢀvalent
proteinꢀprotein interactions with large molecular interfaces, where achievement of potent
inhibition requires effective interference with both motifs occupying different regions of the
binding site. The increase in the molecular weight of inhibitors targeting biꢀvalent PPIs (e.g. >
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
00 Da) can negatively affect their drugꢀlike properties, making the lead optimization process
extensive and complex.
Importantly, strong in vitro inhibition induced by 28 correlates well with pronounced
activity of this compound in MLL leukemia cells (GI = 30ꢀ60 nM), which represents almost
5
0
1
0, 12
1
0ꢀfold improvement over MIꢀ503 or any other reversible meninꢀMLL inhibitor
. In
addition, compound 28 demonstrates high selectivity (>100ꢀfold) and specific mechanism of
action in MLL leukemia cells. Furthermore, the pronounced effect of 28 in inhibiting leukemia
progression in the disseminated model of MLL leukemia combined with the onꢀtarget
mechanism of action in vivo make this compound a very attractive candidate compound for
further studies. During the preparation of this article, irreversible meninꢀMLL inhibitors with
32
distinct molecular scaffold were reported, but these compounds did not show stronger in vitro
inhibitory activity over 28 despite covalent binding to menin. Furthermore, biological activity of
these irreversible meninꢀMLL inhibitors, especially in the in vivo models of MLL leukemia, and
their potential offꢀtarget effects remain to be assessed.
Overall, compound 28 identified in this study represents a valuable molecule to further
investigate the role of meninꢀMLL interaction in normal and pathological states. This compound
may also lead to the development of new therapeutic agents for MLL leukemia and possibly
1
3
14
16
solid tumors, including prostate cancer Ewing sarcoma, HCC, and other tumors that may
rely on the intact meninꢀMLL interaction. Our work demonstrates that, despite challenges
22
ACS Paragon Plus Environment

Page 23 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
associated with the development of potent PPI inhibitors targeting biꢀvalent proteinꢀprotein
interfaces, successful identification of such compounds is possible and can lead to valuable
chemical tools for the scientific community and potentially to new therapeutic agents.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Experimental Section
Expression and purification of menin
8
ꢀ9
The expression and purification of menin have been described previously. In order to generate
33
biotinylated menin, we introduced Aviꢀtag sequence (AAALEGLNDIFEAQKIEWHE) to the Cꢀ
terminus of menin. Protein was coꢀexpressed in E. Coli BL21 (DE3) cells with the construct encoding
3
3
BirA enzyme and purified using the same protocol as for the intact menin.
Biochemical characterization of menin-MLL inhibitors
Inhibition of the meninꢀMLL interaction by small molecules was assessed by fluorescence
polarization (FP) assay. The fluoresceinꢀlabeled MLL4ꢀ43 peptide (FLSNꢀMLL4ꢀ43) at 4 nM,
menin at 4 nM and varying concentrations of compounds were used for IC determination in the
5
0
FP buffer (50 mM Tris, pH 7.5, 50 mM NaCl, 1 mM TCEP). The complex of menin with the
FLSNꢀMLL 4ꢀ43 peptide was incubated for 1h. Compounds (5% final DMSO concentration) were
then added to the meninꢀFLSN_MLL4ꢀ43 complex and incubated for 3h before changes in
fluorescence polarization (mP values) were measured using the PHERAstar microplate reader
(
BMG). The IC values were calculated in Origin 7.0 (OriginLab) by plotting the mP values
50
measured for each compound as a function of compound concentration.
Bio-Layer Interferometry (BLI) binding assay
BLI assay was performed in 96ꢀwell microplates at room temperature with continues shaking at
1
000 rpm using the Octet Red 96 system (Fortebio, Menlo Park, CA, USA). Biotinylated menin
was bound on Super Streptavidin (SSA) biosensors (ForteBio) by dipping sensors into 200 nM
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
biotinylated menin in the assay buffer composed of 50 mM Tris, pH 7.5, 50 mM NaCl, 1 mM
TCEP, 0.05% BSA, 0.01% Tweenꢀ20 and 2% DMSO. Menin binding was performed for 600 s
followed by equilibration in the assay buffer for 1200 s. Compounds (28 and MIꢀ503) were used
at 150 and 200 nM concentrations and binding was performed for 600 s followed by 1200 s
dissociation. Data were fit globally and generated automatically by Octet User Software
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(ForteBio, Inc.). Analysis of binding kinetics was performed in Prism (GraphPad Software, Inc.).
Kobs values were extracted using one phase association equation, and Koff values were extracted
using one phase decay equation. Subsequently, K was derived using K = K × [Ligand] +
on
obs
on
K , where [Ligand] is a free ligand concentration in solution. K was calculated as K /K .
off
D
off on
Crystallization of menin complexes with small molecule inhibitors
For coꢀcrystallization experiments 2.5 mg/mL menin was incubated with small molecule
inhibitors at 1:3 molar ratio. Crystals were obtained using the sitting drop technique at 10 ºC by
9
applying the procedure described previously. Prior to data collection, crystals were transferred
into a cryoꢀsolution containing 20% PEG550 MME and flashꢀfrozen in liquid nitrogen.
Crystallographic data collection and structure determination
Diffraction data for menin and menin complexes were collected at the 21ꢀIDꢀD and 21ꢀIDꢀF
beamlines at the Life Sciences Collaborative Access Team at the Advanced Photon Source. Data
34
were processed with HKLꢀ2000. Structures of the complexes were determined by molecular
replacement using MOLREP with the apoꢀstructure of human menin (PDB code: 4GPQ) as a
3
5
36
search model in molecular replacement. The model was refined using REFMAC, COOT ,
37
38
CCP4 package
and PHENIX .Validation of the structures was performed using
3
9
40
MOLPROBITY and ADIT. Details of data processing and refinement are summarized in
Supplementary Table 1. Coordinates and structure factors for meninꢀinhibitor complexes have
24
ACS Paragon Plus Environment

Page 25 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
been deposited in the Protein Data Bank under the following codes: 6BXH (meninꢀcompound
1
3), 6BXY (meninꢀcompound 28), 6BY8 (meninꢀcompound 29).
Viability assays
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The MLLꢀAF9, HMꢀ2 and E2AꢀHLF transformed murine bone marrow cells (BMCs) were
41
prepared as described previously. MV4;11, MOLMꢀ13, K562 and U937 cells were cultured in
RPMIꢀ1640 medium with 10% FBS, 1% penicillin/streptomycin and NEAA (Nonꢀessential
5
5
amino acid) solution. For viability assay, MOLMꢀ13 (1x10 /mL), MV4;11 (1x10 /mL), K562
5
5
(
1x10 /mL) and U937 (1x10 /mL) human leukemia cells as well as MLLꢀAF9, HMꢀ2 and E2Aꢀ
4
HLF murine bone marrow cells (2.5x10 /mL) were plated (1 mL/well), treated with compounds
or 0.25% DMSO and cultured at 37 ᵒC for 7 days. Media were changed at day four with viable
cell number restored to the original concentration and compounds were reꢀsupplied. 100 µL of
cell suspension were transferred to 96 well plates for each sample in quadruplicates. A Vybrant
MTT cell proliferation assay kit (Molecular Probes) was employed. Plates were read for
absorbance at 570 nm using a PHERAstar BMG microplate reader. The experiments were
performed 2ꢀ3 times in quadruplicated with calculation of mean and standard deviation for each
condition.
Real-Time PCR
Effect of 28 on the expression level of MLL fusion target genes was assessed by Realꢀtime
1
0
quantitative PCR (qRTꢀPCR) using the protocol described previously.
Microsomal stability studies
The metabolic stability was assessed using CDꢀ1 mouse liver microsomes. 1 µM compounds
were incubated with 1.3 mg/mL microsomes and 1.7 mM cofactor NADPH in 0.1 M phosphate
buffer (pH =7.4) containing 3.3 mM MgCl at 37 °C. The DMSO concentration was less than
2
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
.1% in the final incubation system. At 0, 3, 5, 10, 15, 30, 45 and 60 min of incubation, the
reactions were stopped by adding 3ꢀfold excess of acetonitrile containing 100 ng/mL of internal
standard for quantification. The collected fractions were centrifuged at 3000 g for 20 min to
collect the supernatant for LCꢀMS/MS analysis, from which the amount of compound remaining
was determined. The natural log of the amount of compound remaining was plotted against time
to determine the disappearance rate and the halfꢀlife of tested compounds.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Pharmacokinetic studies in mice
All animal experiments in this study were approved by the University of Michigan Committee on
Use and Care of Animals and Unit for Laboratory Animal Medicine (ULAM).
The pharmacokinetics of 28 was determined in female C57BL/6 mice following intravenous (iv)
dosing at 15 mg/kg and 30 mg/kg for oral (po) and intraperitoneal (ip) dosing. For iv and po
administration, 28 was dissolved in the vehicle containing 25% (v/v) DMSO, 25% (v/v) PEG400
and 50% (v/v) PBS. For ip administration, 28 was dissolved in the vehicle containing 20% HPꢀβꢀ
cyclodextrin and 5% Cremophor. Serial blood samples were collected over 7h, centrifuged at
1
5,000 rpm for 10 min and saved for analysis. Plasma concentrations of 28 were determined by
the LCꢀMS/MS method developed and validated for this study. The LCꢀMS/MS method
consisted of an Agilent 1200 HPLC system and chromatographic separation of 28 was achieved
using an Agilent Zorbax ExtendꢀC18 column (5 cm x 2.1 mm, 3.5 µm; Waters). An AB Sciex
QTrap 3200 mass spectrometer equipped with an electrospray ionization source (ABIꢀSciex,
Toronto, Canada) in the positiveꢀion multiple reaction monitoring (MRM) mode was used for
detection. Pharmacokinetic parameters were calculated by noncompartmental methods using
WinNonlin® version 3.2 (Pharsight Corporation, Mountain View, CA, USA).
In vivo efficacy studies
26
ACS Paragon Plus Environment

Page 27 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
All animal experiments in this study were approved by the University of Michigan Committee on
Use and Care of Animals and Unit for Laboratory Animal Medicine (ULAM).
For in vivo efficacy studies, 6 to 8ꢀweekꢀold female NSG mice (Jackson laboratory) were
7
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
intravenously injected with 1x10 luciferaseꢀexpressing MV4;11 cells. At day 13 postꢀ
transplantaion, mice were split intro vehicle and the treatment groups of 8 mice per group and
were treated for 21 consecutive days with once daily intraperitoneal injections of vehicle (20%
HPꢀβꢀcyclodextrin and 5% Cremophor) or 80 mg/kg of 28 dissolved in vehicle. Imaging of mice
was performed at days 0, 7, 14 and 21 of treatment with 28. For imaging, mice were kept
anesthetized with Isoflurane, and luciferase at 50 mg/kg was administered by intraperitoneal
injection. Photonic emission was imaged using the in Vivo Imaging System BLI (IVIS, 200)
with total imaging time of 10 seconds for each mouse. Survival of mice was monitored to assess
treatment efficacy.
For the in vivo gene expression studies, 21 days after transplantation of mice with MV4;11 cells
mice were grouped into vehicle or compound 28 treatment groups (n = 3 / group), and the
treatment was initiated and continued for 6 consecutive days (80 mg/kg, i.p., once daily). The
next day after completing the treatment mice were sacrificed and bone marrow samples were
isolated for gene expression studies by qRTꢀPCR.
Flow cytometric analysis of mice samples
To assess levels of leukemia burden in the mice transplanted with leukemic cells, cells from bone
marrow, peripheral blood, and spleen were buffered with PBS and 1% FBS for flow cytometery
analysis. Red blood cells were lysed with ACK (Lonza). All flow cytometry experiments were
performed on a LSR II FACSCanto or FACSAria (BD Biosciences), and all data were analyzed
with FlowJo software (Tree Star, Inc.).
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 70
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
General Chemistry
All commercially available solvents and reagents were used without further preparation unless
1
13
otherwise indicated. The H and C NMR data were taken on Bruker Avance III 600 MHz or
Varian MR400. Chemical shifts are reported in ppm relative to tetramethylsilane or residual
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
13
solvent signal. Some signals that overlapped with solvents were identified using Hꢀ C HSQC
spectrum (multiplicity not known). The mass measurements were determined on a Micromass
LCT timeꢀofꢀflight mass spectrometer using positive mode and electrospray ionization. The
exact mass measurements were determined on Agilent QꢀTOF timeꢀofꢀflight mass spectrometer
using positive ion mode and electrospray ionization. Analytical TLC was performed on Merck
TLC aluminum plates precoated with F254 silica gel 60 (UV, 254 nm, and iodine). The purity
analysis of final compounds was determined on Shimadzu Prominence HPLC system (20 series:
binary pump, UV/vis at 254 nm, heated column compartment 28 °C), using Restek Ultra C18 (5
ꢁm) 150 mm × 4.6 mm column. LRꢀMS was recorder on Shimadzu LCꢀ2020 system (DUISꢀ
ESI). The solvents were programmed to run at gradient starting 20% CH CN in water to 80%
3
during 8 min run. If not indicated, the purity of all final compounds was >95% as determined by
HPLC via integration of UV spectra at 254 nm. Optical rotation was measured using Jasco
P2000 polarimeter (path lenght = 100 mm). All final (unless indicated otherwise) compounds 1ꢀ
2
9 were tested as hydrochloride salts. The general procedure for HCl salt formation: free base
was dissolved in minimum volume of methanol followed by addition of 1 equiv of 1 M HCl in
water and thorough drying. During NMR experiment, most of the hydrochloride salts of these
compounds exist as a mixture of two forms in solution of either DMSO or MeOH with the
approximate ratio of 10:1; NMR is reported for the major rotamer or a free base as indicated.
28
ACS Paragon Plus Environment

Page 29 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
4
4
ꢀmethylꢀ1ꢀ(2ꢀ(piperazinꢀ1ꢀyl)ethyl)ꢀ5ꢀ((4ꢀ((6ꢀ(2,2,2ꢀtrifluoroethyl)thieno[2,3ꢀd]pyrimidinꢀ
ꢀyl)amino)piperidinꢀ1ꢀyl)methyl)ꢀ1Hꢀindoleꢀ2ꢀcarbonitrile (3). To a stirred solution of tertꢀ
butyl 4ꢀ(2ꢀhydroxyethyl)piperazineꢀ1ꢀcarboxylate (31a, 1 g 4.3 mmol) and trimethylamine (1.5
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
eq, 898 µL, 6.45 mmol) in DCM (14 ml) methanosulfonyl chloride (1.2 eq, 399 µL, 5.16 mmol)
o
was added dropwise at 0 C. Mixture was allowed to warm to the room temperature and stirred
for 3 h. The organic phase was diluted with DCM (40 mL) and washed with water, brine and
dried over sodium sulfate. Solvent was evaporated (partial decomposition was observed when
o
heated over 35 C during evaporation) and crude product (1.25 g) was used in the next step. The
5
ꢀformylꢀ4ꢀmethylꢀ1Hꢀindoleꢀ2ꢀcarbonitrile (30, 300 mg, 1.63 mmol) was suspended in DMF
(3.2 mL) and cesium carbonate was added (1.63 g, 4.68 mmol, 3 eq). The tertꢀbutyl 4ꢀ(2ꢀ
(methylsulfonyl)oxy)ethyl)piperazineꢀ1ꢀcarboxylate, from the previous step (753 mg, 2.45
(
mmol, 1.5 eq) was then added and resulting suspension was stirred at room temperature for 22 h.
Reaction was quenched with water and extracted with DCM (3x40 mL). Organic phase was
evaporated with silica gel and separated using flash chromatography (silica gel, hexanes:ethyl
acetate 3:1 to 1:1, v:v). 316 mg of compound tertꢀbutyl 4ꢀ(2ꢀ(2ꢀcyanoꢀ5ꢀformylꢀ4ꢀmethylꢀ1Hꢀ
indolꢀ1ꢀyl)ethyl)piperazineꢀ1ꢀcarboxylate (32b) was isolated (48% yield). MS (m/z) found for
+
[
M+H ]: 397 Da. Compound 32b (880 mg, 2.79 mmol, 1 eq) and amine 33 (1.01 g, 3.06 mmol,
1
.1 eq) were then dissolved in DCM (28 mL). Triethylamine was added (1.6 mL, 4 eq) followed
by sodium triacetoxyborohydride (1.2 g, 5.6 mmol, 2 eq). Reaction was stirred overnight at room
temperature. Mixture was diluted with DCM (50 mL) and washed with 1M NaOH . Organic
aq
phase was dried over sodium sulfate and evaporated with silica gel. Purified using flash
chromatography (silica gel, dichloromethane:methanol with 5% ammonia, 0% to 5%, v:v). 1.2 g
of product 34b (tertꢀbutyl 4ꢀ(2ꢀ(2ꢀcyanoꢀ4ꢀmethylꢀ5ꢀ((4ꢀ((6ꢀ(2,2,2ꢀtrifluoroethyl)thieno[2,3ꢀ
29
ACS Paragon Plus Environment