Selective Inhibition of Histone Deacetylase 10: Hydrogen Bonding to the Gatekeeper Residue is Implicated

Article abstract of DOI:10.1021/acs.jmedchem.8b01936

The discovery of isozyme-selective histone deacetylase (HDAC) inhibitors is critical for understanding the biological functions of individual HDACs and for validating HDACs as drug targets. The isozyme HDAC10 contributes to chemotherapy resistance and has

Full text of DOI:10.1021/acs.jmedchem.8b01936

Subscriber access provided by IDAHO STATE UNIV
Article
Selective Inhibition of Histone Deacetylase 10: Hydrogen
Bonding to the Gatekeeper Residue is Implicated
Magalie Géraldy, Michael Morgen, Peter Sehr, Raphael Steimbach, Davide Moi, Johannes Ridinger,
Ina Oehme, Olaf Witt, Mona Malz, Mauro S. Nogueira, Oliver Koch, Nikolas Gunkel, and Aubry K. Miller
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b01936 • Publication Date (Web): 09 Apr 2019
Downloaded from http://pubs.acs.org on April 9, 2019
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 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Selective Inhibition of Histone Deacetylase 10:
Hydrogen Bonding to the Gatekeeper Residue is
Implicated
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,†
1
2
1,3
Magalie Géraldy , Michael Morgen , Peter Sehr , Raphael R. Steimbach , Davide
Moi , Johannes Ridinger3,4,5,6, Ina Oehme4,5,7, Olaf Witt4,5,6,7, Mona Malz , Mauro S.
1
1
8
8,†
1,7
1,7,
Nogueira , Oliver Koch , Nikolas Gunkel , Aubry K. Miller *
1
Cancer Drug Development Group, German Cancer Research Center (DKFZ), 69120
Heidelberg, Germany
2
Chemical Biology Core Facility, European Molecular Biology Laboratory, 69117
Heidelberg, Germany
3
Biosciences Faculty, University of Heidelberg, 69120 Heidelberg, Germany
4
Hopp Children’s Cancer Center Heidelberg (KiTZ), 69120 Heidelberg, Germany
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 90
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
5
Clinical Cooperation Unit Pediatric Oncology, German Cancer Research Center
(DKFZ), 69120 Heidelberg, Germany
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
6
Department of Pediatric Oncology, Hematology and Immunology, University Hospital
Heidelberg, 69120 Heidelberg, Germany.
7
German Cancer Consortium (DKTK), 69120 Heidelberg, Germany
8
Faculty of Chemistry and Chemical Biology, TU Dortmund University, 44227
Dortmund, Germany
ABSTRACT
The discovery of isozyme-selective histone deacetylase (HDAC) inhibitors is critical for
understanding the biological functions of individual HDACs and for validating HDACs as
drug targets. The isozyme HDAC10 contributes to chemotherapy resistance and has
recently been described to be a polyamine deacetylase, but no studies toward selective
HDAC10 inhibitors have been published. Using two complementary assays, we found
tubastatin A, an HDAC6 inhibitor, to potently bind HDAC10. We synthesized tubastatin A
2
ACS Paragon Plus Environment

Page 3 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
derivatives and found that a basic amine in the cap group was required for strong
HDAC10 binding. HDAC10 inhibitors mimicked knockdown by causing dose-dependent
accumulation of acidic vesicles in a neuroblastoma cell line. Furthermore, docking into
human HDAC10 homology models indicated that a hydrogen-bond between a cap group
nitrogen and the gatekeeper residue Glu272 was responsible for potent HDAC10 binding.
Taken together, our data provide an optimal platform for the development of HDAC10-
selective inhibitors, as exemplified with the tubastatin A scaffold.
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
INTRODUCTION
The discovery of histone deacetylase 1 (HDAC1) by Taunton and Schreiber in 19961
provided the sought-after enzyme target for substances like trichostatin A (TSA (1), Figure
1) and SAHA (2). At the time, 1 and 2 were reported to increase histone lysine acetylation
levels, thereby inducing cellular differentiation, but their mechanism(s) of action were
2
unknown. In the intervening time, it has become clear that HDACs, a family of 18
functionally related isozymes, have a broader role than catalyzing the hydrolysis of
acetylated histone lysines: not only do they act in both the nucleus and the cytoplasm,
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 90
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
but they catalyze the removal of acyl groups from a variety of different proteins, and
therefore regulate numerous biological processes.3-7 Misregulation of acetylation
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
mechanisms has been linked to multiple human diseases and there are four HDAC
inhibitor drugs approved by the U.S. FDA (2–5) and one in China (6) (Figure 1).8
O
O
O
O
OH
H
N
OH
N
OH
H
H
N
H
N
N
H
Me Me
O
Me N
2
HN
SAHA (2)
Me
TSA (1)
Panobinostat (3)
O
O
HN
O
N
H
O O
S
O
N
NH2
F
O
O
H
H
OH
N
S
NH
N
N
S
N
HN
H
H
O
O
O
Chidamide (6)
Belinostat (4)
Ph Ph
Romidepsin (5)
Cap group
O
N
O
Me
OH
S
N
N
H
N
N
O
N
O
O
H
H
HO
N
N
ACY-738 (9)
N
H
OH
OH
Linker
O
O
Zinc Binding
Group
Tubacin (7)
Tubastatin A (8)
Figure 1: Well-known HDAC inhibitors
The approved drugs are known to cause severe side effects, which is attributed in part
to their lack of isozyme selectivity.9-10 More selective inhibitors are expected to overcome
4
ACS Paragon Plus Environment

Page 5 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
these liabilities and are likely to improve the clinical value of this target class. Moreover,
the development of isozyme-selective chemical probes will be critical to further
disentangle the biological role(s) of individual HDAC members.12
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 number of selective HDAC6 (Class IIB) inhibitors have been described, and tubacin
(
7), tubastatin A (8), and ACY-738 (9) are all designated as HDAC6 chemical probes in
13
the Chemical Probes Portal (Figure 1). Most HDAC6-selective inhibitors, like 8 and 9,
achieve selectivity over Class I enzymes by incorporating a relatively bulky phenyl
hydroxamate “linker” moiety in addition to a “cap group” that can make specific
interactions with the HDAC6 protein surface (see 8, Figure 1).
HDAC10,14-16 the only other HDAC Class IIB member, has received little attention from
the medicinal chemistry community,17-19 and to the best of our knowledge no study
dedicated to the development of HDAC10-selective inhibitors has been published. On the
other hand, a number of studies have highlighted HDAC10 as a potential cancer drug
target.20-23 In one study, high HDAC10 expression levels were found to correlate with poor
clinical outcome for advanced stage 4 neuroblastoma patients who received
chemotherapy.23 Consistent with these findings, HDAC10 depletion in neuroblastoma
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 90
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
cells interrupts autophagic flux and sensitizes cells for chemotherapy, and enforced
HDAC10 expression protects neuroblastoma cells against doxorubicin treatment.18
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
Recently, Christianson and co-workers solved the x-ray crystal structure of zebrafish
HDAC10 (zHDAC10), and elegantly demonstrated that both zHDAC10 and human
HDAC10 (hHDAC10) are highly active polyamine deacetylases (PDACs), while being
4
poor lysine deacetylases. Two specific structural features near the active site of the
enzyme(s) were identified as being responsible for PDAC activity. First, the negatively
charged Glu272 (hHDAC10 numbering) amino acid was demonstrated to be a unique
gatekeeper residue, which establishes specificity for cationic polyamine substrates over
acetylated lysines. Second, the L1 loop of HDAC10 contains a four residue 3 -helix
10
(Pro21-Glu22-Cys23-Glu24, hHDAC10 numbering) that constricts the active site, causing
an acetylated lysine side chain to be too short to reach the active site zinc atom.
In the present study, we report our initial studies to discover selective HDAC10
inhibitors. Using two ligand displacement assays, we demonstrate that tubastatin A (8)
binds to HDAC10 even more potently than to HDAC6. Structure–activity relationship
(SAR) studies on 8, supported with selectivity testing against Class I enzymes, functional
6
ACS Paragon Plus Environment

Page 7 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
cell-based experiments, and docking into human HDAC10 homology models, suggest
that binding to the Glu272 gatekeeper residue is critical for potent HDAC10 inhibition.
Collectively, our results present the first systematic medicinal chemistry study of HDAC10
inhibition, and lay the foundation for the development of isozyme-selective HDAC10
inhibitors.
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
RESULTS AND DISCUSSION
Testing of known substances in two displacement assays reveals unexpectedly potent
HDAC10 binding. Although it is known that HDAC10 enzymatic activity is challenging to
assay with acetylated lysine substrates,12 some contract research organizations offer
enzymatic profiling of all HDACs, and IC values measured in this way against HDAC10
50
are sometimes reported in the literature. It has also been claimed that residual co-purified
Class I HDACs from HDAC preparations of Class IIA HDACs are largely responsible for
the activity observed when testing these proteins in acetyl-lysine-based enzymatic
24
assays. In light of this claim, and the recent finding that HDAC10 has significant PDAC
4
activity, we analyzed a selection of literature data, where enzymatic activities across
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 90
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
multiple HDAC isozymes were reported.25-33 Interestingly, strong correlations between
HDAC10 and Class I HDAC inhibition values were found, while HDAC6 gave a completely
distinct activity profile, despite its high sequence homology to HDAC10 as a Class IIB
enzyme (Figure 2, Supplementary Figures 1 and 2).
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
Figure 2. Hierarchical clustering reveals a correlation between HDAC Class I and
HDAC10 activities. A selectivity profile of each compound was calculated relative to
HDAC2. After log2 transformation, the values were subjected to hierarchical clustering
using the complete-linkage agglomeration rule. Red and green colors indicate higher and
lower reported IC values, respectively, of a given drug on an HDAC isozyme relative to
50
HDAC2. Superscripts indicate the reference from where the data was taken.
8
ACS Paragon Plus Environment

Page 9 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
On the basis of this analysis, and due to the low-throughput format of Christianson’s
PDAC assay, we decided to assay HDAC10 in a non-enzymatic format. To eliminate any
complications that could arise from co-purified HDAC impurities, two ligand displacement
assays that employ genetically tagged HDAC proteins for signal generation were utilized,
thereby ensuring reliable identification of HDAC10 binders. For screening against pure
recombinant HDAC10, a time-resolved fluorescence resonance energy transfer
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
(TR-FRET) assay that measures displacement of a fluorescently-labeled ligand was used
34
(
Figure 3A). A conceptually similar bioluminescence resonance energy transfer (BRET)
assay was also utilized with cells that overexpress nano-luciferase-tagged HDAC10
35
(
Figure 3B, Supplementary Figure 3). The BRET assay can independently confirm the
TR-FRET assay and provides target engagement values in living cells.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 90
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
A
B
3
40 nm
340 nm
615 nm
FRET
665 nm
6
15 nm
Eu
Eu
HDAC10
binder
tag
tag
tracer
tracer
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
HDAC10
HDAC10
610 nm
450 nm
450 nm
BRET
HDAC10
binder
nano
luc
nano
luc
tracer
tracer
HDAC10
HDAC10
Figure 3: HDAC10 ligand displacement assays. A) TR-FRET assay: A Europium labelled
anti-GST tag antibody is coupled to GST-tagged HDAC10 protein. When HDAC10 is
bound by a dye-labelled tracer molecule, irradiation with 340 nm light produces productive
FRET between the Europium (615 nM emission) and the dye (665 nM emission). In the
presence of an HDAC10 binder, which competes with the tracer ligand, FRET is disrupted
and a change in the signal is observed. B) BRET assay: When cells expressing nano-
luciferase-tagged HDAC10 are treated with a dye-labelled tracer and nano-luciferase
substrate, productive BRET is observed. In the presence of an HDAC10 binder, BRET is
disrupted.
A series of commercially available HDAC inhibitors, which have been reported to have
high HDAC6 activity, were tested in both HDAC10 assays (Table 1, Supplementary Table
1, Supplementary Figure 1). Ricolinostat (10) showed almost a 10-fold difference in
activity between the two assays, but the majority of compounds had pIC50 values that
10
ACS Paragon Plus Environment

Page 11 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
were within 0.5 log units of each other. The BRET assay typically gave lower potencies,
which can be expected when comparing cellular and biochemical assay formats. It was
not surprising that some of the pan-HDAC inhibitor compounds (2, 10–15) showed quite
high activity in the HDAC10 displacement assays. Indeed, all of these compounds, except
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
pracinostat (13), had sub-micromolar IC50 values (pIC ≥ 6.0) in at least one of the
50
33
assays. Most interesting were the results with tubastatin A (8) HPOB (16) , and
nexturastat (17),36 compounds that have been described as highly selective HDAC6
inhibitors with poor HDAC10 inhibitory activities. While both HPOB and nexturastat were
found to be moderately potent HDAC10 binders, Tubastatin A was active in the low
nanomolar range (pIC = 7.9) in both the FRET and BRET assays. A control experiment
50
with PCI-34051 (18), a hydroxamate-containing HDAC8 selective inhibitor, gave low
binding values in both assays.
Table 1. TR-FRET and BRET HDAC10 binding data with commercial inhibitors.
Cmpd.
Substance
HDAC10 HDAC10
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 90
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
FRET
pIC50a
BRET
pIC50b
5.9
SAHAc
Ricolinostatc
CAY-10603c
CUDC-101c
Pracinostatc
Quisinostatc
Abexinostatc
Tubastatin Ad
2
6.7
6.9
6.5
6.9
5.8
8.0
8.4
7.9
7.1
7.0
4.4
10
5.8
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
11
12
13
6.1
6.4
5.5
14
7.4
15
8.1
8
7.9
16
HPOBd
Nexturastatd
PCI-34051e
6.5
17
6.3
18
<5.0
a
b
FRET pIC values are the mean of at least two independent experiments. BRET pIC
50
50
c
values are calculated from one experiment performed in triplicate. pan-HDAC inhibitor.
d
e
HDAC6 “selective” inhibitor. HDAC8 selective inhibitor. All pIC values with errors are
5
0
reported in Supplementary Table 1.
Tubastatin A analogs lacking a basic amine in the cap group have diminished HDAC10
binding. Intrigued by the high potency of the HDAC6-selective inhibitor tubastatin A
against HDAC10, we next examined the structure–activity relationship (SAR) of the “cap
group” with respect to HDAC10 and HDAC6 binding. A targeted selection of tricyclic
tubastatin A cap-group analogs were synthesized and tested in both the TR-FRET and
the BRET HDAC10 assays, as well as against HDAC6 in the BRET assay format. As was
12
ACS Paragon Plus Environment

Page 13 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
found with the commercial inhibitors, pIC values of the TR-FRET and BRET HDAC10
50
assays were quite consistent for all compounds assayed (Table 2, Supplementary Table
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
2
). More illuminating was the comparison of the HDAC10 and HDAC6 BRET pIC values
50
for the various derivatives. Tubastatin A (8) gave a pIC of 7.9 against HDAC10 and of
50
25
7
.0 against HDAC6.18, 23 Tetrahydro--carboline 19, a regioisomer of tubastatin A that
differs only in the location of the carboline nitrogen, was found to be roughly an order of
magnitude less active than 8 in the two HDAC10 assays, but essentially equipotent in the
HDAC6 assay. An even greater drop in HDAC10 activities, relative to 8, was observed
37
for compounds 20 and 21, where the -carboline nitrogen of 8 is replaced with sulfur
and oxygen atoms, respectively. As with 19, however, compounds 20 and 21 have only
slightly diminished (20) or improved (21) HDAC6 values in comparison to 8.
Tetrahydrocarbazole 22 and carbazole 2325 were over 2 orders of magnitude weaker
HDAC10 binders than 8, but again retained their HDAC6 binding capabilities. The
acetylated tubastatin A derivative 24 was also found to have significantly diminished
HDAC10 binding, but to be equipotent with tubastatin A (8) against HDAC6.
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 90
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
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
Table 2: HDAC6 and HDAC10 binding values of tricyclic tubastatin A analogs
N
H
N
OH
O
HDAC10
HDAC6
Selectivity
Cmpd.
FRETa BRETb BRETb BRET HDAC10/
pIC50 pIC50 pIC50
HDAC6 (fold)
NMe
8
7.9
6.7
6.5
6.2
6.0
5.7
6.3
7.9
6.8
6.4
6.0
5.6
5.4
5.7
7.0
7.2
6.9c
7.7
7.2
7.4
7.1
6.9
19
20
21
22
0.41
NMe
S
0.47
O
0.018
0.025
0.011
0.043
23
O
N
2
4
a
b
pIC values are the mean of at least 2 independent measurements. pIC values are
50
50
calculated from pooled data of two biological replicates which were each performed in
c
triplicate. pIC value is from one experiment performed in triplicate. All pIC values with
50
50
errors are reported in Supplementary Table 2.
It is apparent that while HDAC6 activity is influenced by changes to the indole-fused
ring, most modifications result in substances that are similarly potent, or even more
14
ACS Paragon Plus Environment

Page 15 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
potent, than tubastatin A (8). This is consistent with published data for 19, 20, and 23,
which are all reported to be potent HDAC6 inhibitors.25, 37 In contrast, HDAC10 activity
appears to be quite sensitive to modifications of the indole-fused ring. Comparing BRET
data for the two enzymes shows that while tubastatin A (8) binds HDAC10 better than
HDAC6 by almost a factor of 10, 19-24 have moderate (19; 2.5 fold) to high (23; 100 fold)
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
38
preferences for HDAC6 over HDAC10. This is visually represented with the BRET dose
response curves of 8, 19, and 21 (Figure 4). A basic nitrogen atom embedded in a
tetrahydro--carboline structure, therefore, appears to be a critical feature for potent
HDAC10 binding, suggestive of a specific polar interaction in the HDAC10 binding site.
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 90
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
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
Figure 4: Dose-response curves of compounds 8, 19, and 21 in the HDAC6 and HDAC10
BRET assay. Data are from one representative biological replicate that was performed in
triplicate. Error bars are standard deviations.
SAR of bicyclic Tubastatin A analogs. In order to investigate this hypothesis further, we
prepared a series of “ring-opened” bicyclic tubastatin A analogs, whose structures and
binding activities are shown in Table 3 (Supplementary Table 3). Consistent with our
expectations, indole 25 and 2,3-dimethylindole 26 showed diminished activity in both
HDAC10 assays, in comparison to 8. The 2-tert-butylindole derivative 27 showed lower
16
ACS Paragon Plus Environment

Page 17 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
activity against both HDAC6 and HDAC10, indicating that the t-butyl group is too bulky
for either of the two binding pockets. Interestingly, the “re”-introduction of a basic amine
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
39
into the cap group, as in gramine 28 and pyrrolidine 30, produced very potent HDAC10
binders, both with BRET pIC50 values of 8.4. In comparison, their HDAC6 binding is
moderate (pIC = 6.8). Indazole 29, which is expected to be more pharmacologically
50
stable than 28, was found to have similar HDAC10 potency to 8, while BOC-protected
amine 31 lost significant HDAC10 binding capacity. Tryptamine 3240 showed good
potency against HDAC10 (BRET pIC = 7.6), despite the presence of two methylenes
50
between the indole C3 and the nitrogen atom, which was detrimental in the case of 19.
The increased conformational flexibility of ring-opened 32 may allow it to still make a key
interaction with the enzyme that 19 cannot. Both the C3 hydroxyethyl (33) and
methoxyethyl (34) derivatives lose almost an order of magnitude binding to HDAC10 in
comparison to 32, further highlighting the importance of a basic nitrogen in the cap group
scaffold. Compound 35, which incorporates a gramine type structure into a cyclic side
chain is also an excellent HDAC10 binder (BRET pIC = 8.3), while a structurally similar
50
aromatic imidazole analog (36) is 10-fold less potent.
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 90
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
Table 3: HDAC6 and HDAC10 binding values of bicyclic Tubastatin A analogs
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
R2
X
R1
N
H
N
OH
O
HDAC10
HDAC6
Selectivity
R2
X
R1
FRET BRET BRET BRET HDAC10/
Cmpd.
a
b
b
pIC50 pIC50 pIC50
HDAC6 (fold)
NMe
8
C
7.9
7.9
7.0
7.9
25
C
C
C
C
N
C
C
H
Me
t-Bu
H
H
Me
H
7.2
6.5
5.3
8.3
7.7
8.4
6.8
6.8
6.2
<4.4
8.4
7.8
8.4
6.6
7.2
7.5
5.7
6.8
6.9
6.8
7.1
0.37
0.049
<0.05
43
26
27
28
29
30
NMe2
–
NMe2
N
7.4
H
37
31
H
NHBoc
NMe2
0.30
32
33
34
C
C
C
H
H
H
7.7
7.1
6.9
7.6
6.7
6.7
6.9
7.5
7.2
4.7
OH
0.17
0.32
OMe
3
5
6
C
C
H
H
N
8.3
7.6
8.3
7.3
6.9
7.1
29
Me
N
3
1.8
N
Me
a
b
pIC values are the mean of at least 2 independent measurements. pIC values are
50
50
calculated from pooled data of two biological replicates which were each performed in
triplicate. All pIC values with errors are reported in Supplementary Table 3.
50
As found with the “fused” analogs in Table 1, BRET HDAC6 activity was relatively
insensitive to structural changes, with all compounds except t-Bu substituted 27 having
activities within a single order of magnitude range (6.8 < pIC < 7.5). Amine analogs 28–
50
18
ACS Paragon Plus Environment

Page 19 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
0, 32, and 35 all show activities similar or better than tubastatin A against HDAC10, with
moderate selectivity (≥0.7 log units) over HDAC6. Compounds 28, 30, and 35 appear to
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
have the best profiles in the BRET assays, with pIC values greater than 8.3 (IC = 5.0
50
50
nM) against HDAC10 and selectivities greater than 25 fold against HDAC6.
Selectivity testing against Class I enzymes. Selected compounds were also tested in
enzymatic assays against HDAC1, HDAC2, HDAC3/NcoR2, and HDAC8 to gain insight
into their activities against Class I enzymes (Table 4, Supplementary Table 4). SAHA (2)
was used as a positive control for HDAC1, HDAC2, and HDAC3/NcoR2, and PCI-34051
(18) was used as a positive control for HDAC8. Tubastatin A (8) was found to have pIC50
values in accordance with the literature.30, 41 Tubastatin A regioisomer 19 gave lower
inhibition values than 8 against all the Class I enzymes. This, coupled with the fact that
19 is more potent against HDAC6 than 8, suggests that it may be a better HDAC6 probe
than tubastatin A (8) itself. Compound 21 had similar Class I pIC50 values to 8.
Interestingly, dimethylindole 26 saw no increase in potency against Class I enzymes
despite its smaller bulk, indicating that a tricyclic cap group is not strictly required for Class
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 90
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
IIB selectivity. Compounds 28, 29, 30, 35, and 36 all saw increased potency against Class
I enzymes in comparison to 8. While 29, 35, and 36 had overall decreased HDAC10
selectivity relative to 8, 28 and 30 retained the HDAC10 selectivity of 8 due to their
concomitant increase in potency against HDAC10. Taking all the tested Class I and Class
IIB enzymes into account, compounds 28 and 30 represent the best combination of
potency (IC < 10 nM) against HDAC10 and selectivity over other HDACs.
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
50
Table 4. HDAC2 pIC values of selected inhibitors.
50
pIC50
pIC50
pIC50
pIC50
pIC50
Compound
(SAHA)
HDAC10
(FRET)b
HDAC1a HDAC2a HDAC3a HDAC8a
2
7.3
6.4
4.9
7.4
5.8
6.1
6.0
6.7
7.9
8
(Tubastatin 5.9
A)
1
1
2
2
8 (PCI-34051) 4.3
N.M.
4.6
4.2
5.4
5.9
5.5
7.5
5.4
6.1
5.8
4.4
6.7
6.2
6.5
9
1
6
5.5
5.8
5.6
5.1
4.9
20
ACS Paragon Plus Environment

Page 21 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
8
9
6.0
5.9
6.1
6.5
6.2
5.2
5.1
5.2
5.7
5.5
6.2
6.1
6.4
6.8
6.3
6.4
6.1
6.5
6.4
6.4
8.3
7.7
8.4
8.3
7.6
2
30
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
35
36
a
b
pIC values are calculated from one experiment performed in triplicate. Values are
50
taken from HDAC10 FRET data in Tables 2 and 3. All pIC values with errors are reported
50
in Supplementary Table 4.
Compound HDAC6/10 Inhibition Profiles are Reflected in On-target Activity in Cells.
Some selective HDAC10 inhibitors (8, 29, 30, 35), some selective HDAC6 inhibitors (20
and 21), and one inactive substance (27) were also tested in the neuroblastoma BE(2)-C
cell line to confirm on-target activity.18-19, 23, 42 To test for HDAC6 inhibition, tubulin
acetylation was measured by Western blot, and to test for HDAC10 inhibition, expansion
of the lysosomal compartment was measured using the pH-dependent LysoTracker DND-
99 fluorescent probe. As expected, tubastatin A (8) showed a dose-dependent increase
in tubulin acetylation (Figure 5A). Substances 20, 21, 29, 30, and 35 had profiles similar
to 8, consistent with the fact that they are all similarly potent HDAC6 inhibitors. Only
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 90
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
compound 27, which showed reduced HDAC6 activity in the BRET assay, required higher
concentrations to increase tubulin acetylation. In the LysoTracker assay, only those
compounds which showed highly potent HDAC10 binding in the BRET and FRET assays
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, 29, 30, and 35) strongly induced fluorescence, while 20, 21, and 27 gave no to weak
increases (Figure 5B).
Figure 5. A) Western blot analysis of acetylated tubulin versus total tubulin, as induced
by HDAC6 inhibition from tubastatin A (8) and analogs. B) Expansion of the lysosomal
compartment as a measure of HDAC10 inhibition, shown by increased fluorescence from
the LysoTracker DND-99 probe. Data is normalized to the fluorescence measured with
only DMSO and represents Mean ± SD from at least 3 experiments. A normalized value
of 1 means that fluorescence of the treated sample is equal to the DMSO control.
22
ACS Paragon Plus Environment

Page 23 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Homology Modeling Provides a Rationale for Selective HDAC10 Binding. The Glu272
gatekeeper residue is the single determining factor which gives HDAC10 PDAC over
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
4
HDAC activity. This selectivity is the result of a polar interaction between the negatively
charged glutamic acid and a positively charged polyamine. We wondered whether
inhibitors with a basic amine in the cap group were such strong HDAC10 binders due to
a similar polar interaction with Glu272. Moreover, highly potent binders such as 8, 30,
and 35 indicate that bulky inhibitors are able to fit into the constricted binding pocket of
HDAC10, presumably requiring significant movement of the L1 loop. To investigate, we
built several different homology models of hHDAC10 and performed docking studies.
Initially, two hHDAC10 homology models were prepared, one based on a crystal structure
4
43
of zHDAC10 (PDB ID 5td7) and one based on hHDAC6 (PDB ID 5edu), hereafter
referred to as Model I and Model II, respectively. Both Model I and Model II were used for
docking studies of tubastatin A (8). Model I has an L1 loop that closely resembles that of
the parent zHDAC10 structure, and failed to show any reliable docking poses. Model II
has an L1 loop more similar to its parent hHDAC6 structure, and showed high-scoring
docking poses of 8 with a hydrogen bond to Glu272 (Figure 6A). A superposition of the
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 90
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
two models clearly indicated that the L1 loop in Model I prevents binding, as 8 would clash
with the L1 loop (Figure 6B). To investigate the loop flexibility that seems to be needed
for docking of 8, two additional models were built. For one of these models (Model III),
the ligand of PDB ID 5edu (trichostatin A (1)) was transferred into PDB ID 5td7 to
introduce an induced fit during model building. For the other model (Model IV), residues
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
14–28 from the L1 loop were removed from PDB ID 5td7 and rebuilt using MOE’s “build
loop” module. Both Model III and Model IV resulted in new L1 loop conformations.
Interestingly, docking of 8 into these models succeeded, leading to hydrogen-bonded
docking poses similar to what was found in Model II (Figures 6C and 6D). To summarize,
in the four models reliable docking poses were only identified when the L1 loop adopted
a conformation different from what is found in the zHDAC10 structure, suggesting that
flexibility of the L1 loop is needed for tubastatin A binding.
24
ACS Paragon Plus Environment

Page 25 of 90
Journal of Medicinal Chemistry
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
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
Figure 6: Docking poses of Tubastatin A in hHDAC10 models (red: oxygen of Glu272 as
hydrogen bond partner; magenta: L1 loop). A) Model II (green). B) Superposition of 8
docking pose in Model II (green) with Model I (yellow). C) Induced fit Model III. D)
Alternative L1 loop Model IV.
Compounds 20 and 21, as tricyclic tubastatin A analogues, and the bicyclic derivatives
28, 29, 30 and 35 were also docked in the human HDAC10 homology models. The
docking poses of 20 and 21 show only hydrophobic interactions in the cap group region,
with no polar interactions to Glu272 (Figure 7A), while adopting similar conformations to
8 (Figure 7B). For 20 and 21, the flexible Glu272 moved beneath the ring system during
docking because no hydrogen bonding could occur. This difference in binding between 8
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 90
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
and 20/21 could explain their disparate activities against HDAC10. Tubastatin A (8), 20
and 21 were also docked into hHDAC6 (PDB ID 5edu), giving one main binding pose for
all three substances (data not shown). These docking poses show largely hydrophobic
interactions of the cap groups with HDAC6, consistent with the fact that the structural
differences between the compounds do not have dramatic effects on HDAC6 binding in
the BRET assay.
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
For the bicyclic derivatives 28, 29, 30 and 35, the results are more ambiguous since
many of the calculated poses only show monodendate binding of the hydroxamic acid to
zinc. This monodentate binding was different to what has been found with some selective
44
HDAC6 inhibitors. Bidendate binding poses were only found in Model II. In these poses,
hydrogen bonds were found to Glu272 for 28, 29, and 35 (Figure 7C–E), while 30 showed
a hydrogen bond to Glu22 (Figure 7F), a glutamic acid found in the L1 loop close to the
binding pocket. Interestingly, 29 also gave a reasonable pose to Glu22 (Figure 7G). Here,
the homology modelling clearly reaches its limit to provide one reliable binding pose. It is
possible that a different conformation of the loop region is required to provide a reliable
interaction pattern for binding. However, these results indicate that the bicyclic derivatives
26
ACS Paragon Plus Environment

Page 27 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
8, 29, 30 and 35 can address at least one of the two glutamic acids in the binding region
of HDAC10 (Glu22, Glu272) which are not present in HDAC6.
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
Figure 7: Docking poses of 8, 20, 21, 28, 29, 30, and 35 in Model II (tan); pink dotted line
represents a hydrogen bond. A) 20 (orange) and 21 (lime green). B–E) 8 (cyan), 28
(yellow), 29 (coral), and 35 (orchid) with hydrogen bonds to Glu272. F–G) 30 (olive) and
29 (coral) bound to Glu22.
CHEMISTRY
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 90
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 compounds used in this study were synthesized in a straightforward manner.
Commercially available or synthesized indoles (37–50) were N-alkylated with methyl 4-
carbomethoxybenzyl bromide (51) to give esters 52–57, 59–63, 65, 67, and 68, followed
by conversion to the corresponding hydroxamic acids 20–27, 31–36, and 70 (Scheme 1).
In some instances, intermediate functional group manipulations were carried out at the
methyl ester stage (e.g. 57 to 58, 63 to 64, and 68 to 69) or the hydroxamic acid stage
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
(e.g. 70 to 19, and 25 to 28 and 30). Compound 29 was synthesized starting with
alkylation of 1H-indazole-3-carbaldehyde (71) with 51 to give ester 72, followed by
reductive amination with dimethylamine and then hydroxamic acid formation (Scheme 2).
Scheme 1: Synthesis of indole-based inhibitorsa
28
ACS Paragon Plus Environment

Page 29 of 90
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
R2
R2
R2
Br
R1
R1
(
a)
(b)
R1
N
N
+
H
OMe
N
OMe
N
OH
H
O
51
O
O
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
R1
R2
R1
R2
R1
R2
3
3
3
7:
8:
9:
52:
53:
54:
55:
70:
19:
NBoc
S
NBoc
S
NBoc
c,h
NMe
S
O
O
20:
21:
22:
O
40:
41:
56:
57:
NBoc
NBoc
O
23:
42:
O
N
c,d
43:
H
H
Me
N
Me
24:
25:
4
4
4
4: Me
Me
58:
5: t-Bu
H
H
H
59:
H
H
2
6: Me
Me
6:
H
NHBoc
NHBoc
60: Me
Me
H
i
2
7: t-Bu
H
j
4
7:
8:
H
H
61: t-Bu
28:
0:
H
H
NMe2
N
OH
62:
H
4
NHBoc
3
NHBoc
NMe2
6
3:
4:
H
H
4
9:
0:
H
H
N
e,c,f
e
31:
H
Me
NHBoc
6
5
NMe2
OH
O
32:
33:
34:
H
H
H
OH
65:
6:
H
H
OMe
6
OMe
67:
68:
H
N
35:
H
H
Me
N
Me
N
H
H
O
g
N
36:
69:
N
N
Me
Me
a
Reagents and conditions: (a) NaH or K CO , DMF, 20–74%; (b) NH OH (aq), KOH or
2
3
2
KCN, dioxane 1–93%; (c) TFA/CH Cl or HCl/Et O; (d) AcCl, Et N, CH Cl , 54%; (e) NaH,
2
2
2
3
2
2
MeI, DMF, 39–53%; (f) CH O (37% in H O), NaCNBH , MeCN; (g) MeNH ,
2
2
3
2
CH Cl /MeOH, 40 °C, then TOSMIC, K CO , DME/MeOH, rt, 41%; (h) MeI, Et N, MeCN,
2
2
2
3
3
60 °C; (i) MeNH , AcOH, (CH O) , MeOH/H O 5:1, 70 °C, 46%; (j) pyrrolidine, AcOH,
2
2
n
2
(
CH O) , MeOH/H O 5:1, 70 °C, 90%.
2
n
2
29
ACS Paragon Plus Environment