Largazole Analogues Embodying Radical Changes in the Depsipeptide Ring: Development of a More Selective and Highly Potent Analogue

Article abstract of DOI:10.1021/acs.jmedchem.6b01271

A number of analogues of the marine-derived histone deacetylase inhibitor largazole incorporating major structural changes in the depsipeptide ring were synthesized. Replacing the thiazole-thiazoline fragment of largazole with a bipyridine group gave analogue 7 with potent cell growth inhibitory activity and an activity profile similar to that of largazole, suggesting that conformational change accompanying switching hybridization from sp³ to sp² at C-7 is well tolerated. Analogue 7 was more class I selective compared to largazole, with at least 464-fold selectivity for class I HDAC proteins over class II HDAC6 compared to a 22-fold selectivity observed with largazole. To our knowledge 7 represents the first example of a potent and highly cytotoxic largazole analogue not containing a thiazoline ring. The elimination of a chiral center derived from the unnatural amino acid R-a-methylcysteine makes the molecule more amenable to chemical synthesis, and coupled with its increased class I selectivity, 7 could serve as a new lead compound for developing selective largazole analogues.

Full text of DOI:10.1021/acs.jmedchem.6b01271

Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)
Article
Largazole Analogues Embodying Radical Changes in the Depsipeptide
Ring: Development of a More Selective and Highly Potent Analogue
Jehad Almaliti, Ayad A Al-Hamashi, Ahmed T Negmeldin, Christin l Hanigan, Lalith
Perera, Mary Kay H. Pflum, Robert A. Casero, and L. M. Viranga Tillekeratne
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01271 • Publication Date (Web): 03 Nov 2016
Downloaded from http://pubs.acs.org on November 3, 2016
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 free 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 accessible to all
readers and 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.
Journal of Medicinal Chemistry 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
Largazole Analogues Embodying Radical Changes in the Depsipeptide Ring:
Development of a More Selective and Highly Potent Analogue
ɸ§
ɸ
±
┴
Jehad Almaliti , Ayad A. AlꢀHamashi , Ahmed T. Negmeldin , Christin L. Hanigan , Lalith
£
±
┴
ɸ
Perera , Mary Kay H Pflum , Robert A. Casero, Jr. , and L. M. Viranga Tillekeratne *
ɸ
Department of Medicinal and Biological Chemistry, College of Pharmacy and Pharmaceutical Sciences, The
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
University of Toledo, 2801, W. Bancroft Street, Toledo, OHꢀ43606 USA
§
Department of Pharmaceutical Sciences, Faculty of Pharmacy, The University of Jordan, Amman, 11942, Jordan.
±
Department of Chemistry, Wayne State University, 5101 Cass Avenue, Detroit, Michigan 48202
┴
The Sidney Kimmel Comprehensive Cancer Center, The Johns Hopkins University School of Medicine,
Bunting/Blaustein Cancer Research Building 1 1650 Orleans Street ꢀ Room 551, Baltimore, MD 21231
£
Laboratory of Structural Biology, National Institute of Environmental Health Sciences, National Institutes of
Health, Department of Health and Human Services, Research Triangle Park, North Carolina 27709
(
A.T.N and C.L.H contributed equally to biological studies)
ABSTRACT
A number of analogues of the marineꢀderived histone deacetylase inhibitor largazole
incorporating major structural changes in the depsipeptide ring were synthesized. Replacing the
thiazoleꢀthiazoline fragment of largazole with a bipyridine group gave analogue 7 with potent
cell growth inhibitory activity and an activity profile similar to that of largazole, suggesting that
3
2
conformational change accompanying switching hybridization from sp to sp at Cꢀ7 is well
tolerated. Analogue 7 was more class I selective compared to largazole, with at least 464ꢀfold
selectivity for class I HDAC proteins over class II HDAC6 compared to a 22ꢀfold selectivity
observed with largazole. To our knowledge 7 represents the first example of a potent and highly
cytotoxic largazole analogue not containing a thiazoline ring. The elimination of a chiral center
derived from the unnatural amino acid Rꢀαꢀmethylcysteine makes the molecule more amenable
to chemical synthesis and, coupled with its increased class I selectivity, 7 could serve as a new
lead compound for developing selective largazole analogues.
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
INTRODUCTION
Epigenetics is defined as heritable changes in gene expression that do not involve
1
changes in the DNA sequence itself. Epigenetic gene regulation includes DNA methylation and
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
histone modifications, and nonꢀcoding RNA. This method of gene regulation is a vital part of
normal development and cellular and tissue differentiation. Like many normal cellular
processes, epigenetic regulation can become dysregulated in disease states. In cancer, aberrant
2
silencing of tumor suppressor genes is a hallmark of tumorigenesis. Since epigenetic regulation
of genes does not change the DNA sequence, these aberrant changes can be targeted
pharmacologically and reversed.
Histone modification as mentioned above is one mode of epigenetic regulation.
Modification of histones’ lysine residues include: methylation, acetylation and phosphorylation,
3
which contribute to regulation of gene transcription. Histone acetylation/deacetylation is one
method of epigenetic regulation, which has been of interest to target pharmacologically.
Acetylation and deacetylation of histones are controlled by two enzyme families; the histone
acetyl transferases (HAT), which transfer acetyl groups from coenzyme A to lysine residues on
histones and the histone deacetylases (HDAC), which catalyze the removal of acetyl groups from
histone lysine residues. In general, acetylation of histones is associated with active
transcriptionally competent euchromatin regions, while hypoacetylated histones mark
4
transcriptionally inactive heterochromatic regions. In cancer, aberrant gene silencing is
associated with histone hypoacetylation and in some cases overꢀexpression of HDAC proteins.
Inhibition of HDAC proteins in some cancers could reactivate the silenced tumor suppresser
5
genes and inhibit the growth of tumors. Therefore, HDAC inhibitors have become a promising
area of research in the development of selective drugs for the treatment of cancer. HDAC
ACS Paragon Plus Environment

Page 3 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
inhibitors can block tumor cell proliferation by inducing cell differentiation, cell cycle arrest,
6
and/or apoptosis. Many nonhistone proteins are also substrates for HDAC thus underscoring
their role in cellular process besides transcriptional regulation and their therapeutic potential in
other nonꢀcancer pathological conditions such as neurodegenerative diseases and
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7
–10
immune/inflammatory disorders.
There are currently four HDAC inhibitors approved by
the FDA for clinical use; suberoylanilide hydroxamic acid (SAHA) 1 (Figure 1) approved by the
1
1
FDA in 2006 for the treatment of cutaneous manifestations of Tꢀcell lymphoma, FK228 2 ,
1
2
approved in 2009 for the same disorder, belinostat 3 approved in July 2014 for the treatment of
13
peripheral Tꢀcell lymphoma, and very recently approved panobinostat 4 for the treatment of
1
4
multiple myeloma. There are at least 18 different isoforms of HDAC proteins belonging to four
classes found in humans; the inhibitors FDA approved for cancer treatment to date all target
15
either class 1 and/or class 2 HDAC enzymes. SAHA, belinostat and panobinostat target class I
and II HDACS, whereas FK228 is a classꢀI selective HDAC inhibitor. Selective HDAC
inhibitors are warranted primarily to reduce side effects, target other disorders, and to probe the
16
still not well understood physiological and pathological roles of different HDAC isoforms. The
structureꢀactivity relationship (SAR) studies of HDAC inhibitors have revealed three essential
domains required for a small molecule to act as an HDAC inhibitor; a zinc binding group which
interacts with the zinc ion in the active site and precludes it from catalyzing the deacetylation
process, a linker group, and a surface recognition cap group, which interacts with the
17
hydrophobic residues on the rim of the active site.
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
O
NH
HN
3
O
O
H
N
O
O
O
NH
HO
N
O
H
S
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
O
N
SAHA (1)
H
S
FK228 (2)
H
N
H
N
H
N
HO
S
N
O
O
O
H
H
N
OH
Belinostat (3)
O
Panobinostat (4)
O
S
10
7
3
6
N
1
2
NH
2
N
S
13
O
O
O
O
H C(H C) S
6
22
17
15 HN
3
2
Largazole (5)
Figure 1. HDAC inhibitors and their pharmacophore.
Largazole 5 is a natural product isolated in 2008 from a marine cyanobacterium of the
genus Symploca collected from Key Largo in Florida. Biological assays showed that largazole
possesses potent growth inhibitory activity on several cancer cell lines and its effect on other
1
8–21
disorders such as liver fibrosis, inflammation and rheumatoid arthritis is being evaluated
.
Largazole is a classꢀI selective HDAC inhibitor and is a promising lead molecule in the
ACS Paragon Plus Environment

Page 5 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
22
development of class/isoform selective HDAC inhibitors. A number of largazole analogues has
been synthesized and their biological activities evaluated delineating some of the structureꢀ
2
3–29
activity relationship (SAR) information of the molecule.
Modifications of the linker and the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
27,30–32
zinc binding group are not well tolerated and have mostly yielded less potent analogues.
In contrast, potent analogues have been generated by structural modification of the depsipeptide
2
8,30,31,33–35
ring (surface recognition cap group).
The lower sequence homology between different
HDAC isoforms at the surface near the opening to the binding pocket may be exploited to design
1
7
more potent HDAC inhibitors by targeting the depsipeptide ring.
Here in, we describe the synthesis and biological activity of novel largazole analogues
with major modifications in the depsipeptide ring, while maintaining its 16ꢀmembered ring
character (Figure 2). In analogue 6, the thiazoleꢀthiazoline moiety was replaced with a simple
polyꢀether linkage. Williams and coworkers reported that a pyridyl analogue in which the
thiazole ring of largazole was replaced with a pyridine ring had enhanced HDAC inhibitory
3
6
activity. We synthesized analogue 7 in which a biꢀpyridine moiety replaced the entire thiazoleꢀ
thiazoline fragment of largazole. The biꢀpyridine moiety introduces additional lipophilicity to the
molecule, which may enhance interaction with the lipophilic rim of the active site. However,
3
2
change in hybridization from sp to sp at Cꢀ7 position of largazole will impose significant
conformational change in the depsipeptide ring of the molecule. In analogue 8, the thiazole ring
was replaced with a pyridine but the thiazoline moiety was replaced with a methylamine group.
2
3
While the sp nature of the thiazoline nitrogen is lost, the sp configuration at Cꢀ7 was retained.
Although there is a decrease in hydrophilicity, the Nꢀmethyl group provides a handle for
structural manipulation to modulate hydrophilicity/lipophilicity in this part of the molecule. The
analogue 9 which contains a benzylꢀamine at this position was designed to determine the effect
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
of increasing bulk and lipophilicity at this position. As a further extension of this study, various
electron donating and withdrawing groups can be introduced in the benzyl moiety in analogue 9.
With this in mind, analogues 10ꢀ13 with an electron withdrawing trifluoromethyl, an electron
donating methoxy, an electron withdrawing and hydrophobic pꢀchloro, and a bulky 1ꢀnaphthyl
group, respectively, were designed.
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. Largazole analogues synthesized
ACS Paragon Plus Environment

Page 7 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Chemistry:
From a synthetic view point, two major improvements were made in the convergent synthetic
27
approach that we previously reported for largazole and analogues (Scheme 1). First, an
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
improved procedure for the synthesis of the thiolꢀcontaining linker moiety was developed to
27,28
remove some of the limitations of our previous method.
Second, a new protectionꢀ
deprotection protocol using mild conditions was developed to obtain the acyclic intermediate
required for macrolactamization in pure form and in high yields. The closure of the 16ꢀ
23
membered depsipeptide ring in largazole synthesis is challenging. The macrocyclization yield
is relatively low, and depends on the purity of the starting material and the reaction
37
conditions. . In the synthesis of analogues 12-13, we used a new protection/deprotection
strategy in which the carboxylic end was protected as a trimethylsilylethyl ester and the amino
end was protected with Fmoc and both protecting groups were removed simultaneously under
mild conditions using TBAF to obtain the deprotected intermediate in a high degree of purity and
in high yields (vide infra). This simple protectionꢀdeprotection strategy may be employed
successfully in other macrolactamization reactions as well.
Retrosynthetically, all analogues could be approached by acyl transfer from fragment 30 to
amines 22-29, followed by Yamaguchi esterification with Fmocꢀvaline to form acyclic
precursors 14ꢀ21. Macrolactamization and thioꢀesterification would give the target largazole
analogues. The synthesis of the amine moieties 22-29 is described for each specific analogue.
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
O
O
pg
O
X
Y
X
NH
FmocHN
O
O
Y
O
O
O
O
O
S
N
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
N
H
H
6-13
1
4-21
TrtS
O
FmocHN
O
pg
O
X
OH
Fmocꢀvaline 31
OH
O
S
S
O
O
Y
S
TrtS
N
H2N
TrtS
H
N
S
Surface recognition cap
groups 22-29
3
3
3
2
Tail (linker) moiety 30
Scheme 1. Retrosynthetic analysis of largazole analogues 6-13.
Michael addition of triphenylmethanethiol 34 to acrolein yielded aldehyde 36, which was
converted to the key aldehyde 33 in larger quantities and in higher yields using a Wittig reagent
with a nitrile function as a masked aldehyde group (Scheme 2), instead of one containing an
2
7
unmasked aldehyde group reported earlier.
Reaction of aldehyde 36 with
(
cyanomethyl)triphenylphosphonium chloride yielded an E:Zꢀmixture of nitrile 38. Upon
investigating varying experimental conditions, the highest ratio of the required Eꢀisomer (5:1,
E:Z) was obtained when the aldehyde 36 was reacted with (cyanomethyl)triphenylphosphonium
o
chloride in benzene under reflux at 90 C. Attempts to isomerize this mixture to the desired E
ACS Paragon Plus Environment

Page 9 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
isomer in high yields under acidic and basic conditions were not successful. Therefore, the nitrile
mixture was taken to the next step and reduced with DIBALꢀH to an E/Z mixture (E:Z, 5:1) of
3
8
the corresponding aldehydes. Gratifyingly, the slow passage of the mixture through a silica
column ensured complete isomerization to the desired Eꢀisomer and the aldehyde 33 was
obtained in 85% overall yield without the need for tedious separation procedures.
Diastereoselective aldol reaction of the aldehyde 33 with acetyl Nagao auxiliary 32 gave alcohol
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
30, the Sꢀstereochemistry of the new chiral center of which was confirmed by Mosher ester
3
9
analysis.
Scheme 2. Synthesis of intermediate 30, the tail (linker) moiety.
The synthesis of the surface recognition cap moiety of analogue 6 is shown in scheme 3. As the
alkylation of the NꢀBoc protected aminoꢀalcohol 40 with methyl bromoacetate was sluggish,
40
Finkelstein conditions were used to prepare the polyether 43 in 78% yields. Removal of the
Boc protecting group gave the amine 22 as its TFA salt, which was used in the next step without
further purification.
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
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
Scheme 3. Synthesis of amine 22 required for depsipeptide ring of analogue 5.
For the synthesis of the surface recognition moiety of analogue 7, 2ꢀbromoꢀ6ꢀ
methylpyridine 44 was converted to the tin derivative 45 by reaction with nꢀBuꢀLi followed by
trimethyltin chloride (Scheme 4). Stille coupling of the tin derivative 45 with 2,6ꢀ
dibromopyridine 46 in the presence of Pd(PPh ) catalyst gave 6ꢀbromoꢀ6’ꢀmethylꢀ2,2’ꢀ
3
4
bipyridine 47. A modified Jones oxidation procedure using chromium oxide in concentrated
sulfuric acid was used to obtain the carboxylic acid 48 in high yields after working up the
4
1
reaction mixture with cold ice water to precipitate out the pure carboxylic acid. Esterification,
followed by reaction with potassium cyanide in the presence of CuCN under microwave heating
4
1
gave the corresponding nitrile 50 in high yields. Nitrile 50 was reduced to methyl amine 23 by
catalytic hydrogenation.
ACS Paragon Plus Environment

Page 11 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 4. Synthesis of amine 23 required for depsipeptide ring of analogue 7.
To synthesize the surface recognition cap moiety for analogue 8-13, 2,6ꢀdihydroxymethyl
pyridine 51 was treated with silver oxide and potassium iodide followed by tosyl chloride at ꢀ20
o
C (Scheme 5) to obtain predominantly the monoꢀtosylated product 52, with a trace amount of
4
2
diꢀtosylated byꢀproduct. After purification by passing through a pad of silica gel, compound 52
was converted to azide 53 by heating with sodium azide in DMF. Compound 53 was reacted
with tosyl chloride and sodium hydroxide to give intermediate 54, which was used to alkylate the
secondary amines 55ꢀ60 yielding the tertiary amines 61-66. They were reduced with H /Pd to
2
give amines 24ꢀ29 in almost quantitative yield.
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
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
Scheme 5. Synthesis of amines 24-29 required for depsipeptide ring of analogues 8-13.
With the required building blocks in hand, we proceeded to synthesize each analogue 6-
13 by the general approach as shown in Scheme 6. Acyl transfer from 30 to amines 22-29 was
carried out in the presence of DMAP to obtain alcohols 67-74. They were esterified with Fmocꢀ
Lꢀvaline using Yamaguchi esterification conditions to afford esters 14-21. Esters 14ꢀ19 were
saponified with aqueous LiOH and Fmoc group was removed with diethylamine, while in esters,
20ꢀ21 both trimethylsilylethyl ester and Fmoc groups were simultaneously removed with TBAF.
The deprotected acyclic intermediates were macrolactamized using HOAt, HATU, and Hunig’s
3
7
base to afford the macrocyclic depsipeptide intermediates 75-82. The trityl group was removed
ACS Paragon Plus Environment

Page 13 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
with TFA and the resulting thiols were dried azeotropically with toluene and esterified with
octanoyl chloride. The products were purified by reversed phase chromatography on C18 in
acetonitrile/water to obtain analogues 6-13.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 6. Synthesis of largazole analogues 6-13.
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
BIOLOGICAL STUDIES
Analogues 6-13 were subjected to preliminary screening of antiproliferative activity in a
colon adenocarcinoma cell line (HCT116) by a standard 3ꢀ(4,5ꢀdimethylthiazolꢀ2ꢀyl)ꢀ5ꢀ(3ꢀ
carboxymethoxyphenyl)ꢀ2ꢀ(4ꢀsulfophenyl)ꢀ2Hꢀtetrazolium (MTS) reduction assay (Promega)
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
7,43
using largazole as the control, as reported earlier.
The cells were incubated for 96 h with the
analogues at 0.1 nM, 1 nM, 10 nM, 100 nM, 1 µM, and 5 µM concentrations. The results of the
MTS assay for largazole and analogues 6 – 8 are shown in Figure 3. Largazole inhibited the
growth of HCT116 cells with a GI of 65 nM, while the biꢀpyridine analogue 7 with a GI of 75
5
0
50
nM showed an activity profile similar to that of largazole, and was the most active of the three
analogues (see Figure 5a below for a comprehensive list of GI values on NCI 60 cell lines).
5
0
The polyꢀether analogue 6 with major changes in the surface recognition cap group was less
active than analogue 7 and largazole and but had moderate antiproliferative activity with a GI50
of 0.75 ꢀM. on HCT 116 cells. Analogue 8 too showed moderate antiproliferative activity, with a
GI value of 0.8 ꢀM on HCT 116 cells.
5
0
ACS Paragon Plus Environment

Page 15 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 3. Cell survival MTS assay on HCT116 colorectal cancer cell line, treated for 96 hours at
0.1 nM, 1 nM, 10 nM, 100 nM, 1 µM and 5 µM. Each point represents the mean of three
experiments; each with 4 replicates with the error bars indicating the SEM
Replacing the thiazoline ring of largazole with a methylamine group in analogue 8, in addition to
change of hybridization of the nitrogen atom, involved loss of hydrophobicity in this region of
the molecule. Therefore, the benzylamine analogue 9 was designed to determine the effect of
increasing bulk and lipophilicity at this position. As a further extension of this study, analogues
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
10 – 13 with electron withdrawing trifluoromethyl, electron donating methoxy, a hydrophobic pꢀ
chloro, and hydrophobic and bulky 1ꢀnaphthyl group, respectively, were designed to investigate
the effect of altering electronic and steric properties at this position. However, all these structural
modifications resulted in analogues with reduced levels of activity (Table 1 and in supporting
informaiton, see figure S1 for growth inhibitory effect of analogues 9ꢀ13 on HCT 116 cells).
Table 1. GI values of MTS assay on survival of HCT116 colorectal cancer cell line treated with
5
0
largazole and analogues 6ꢀ13 for 96 hours at 0.1 nM, 1 nM, 10 nM, 100 nM, 1 µM and 5 µM
See Figure 3 and Figure S1 in supporting information).
(
Compound
GI values (ꢀΜ)
5
0
Largazole
0.065 ± 0.003
0.75 ± 0.017
0.075 ± 0.0002
0.8 ± 0.048
>5
6
7
8
9
1
1
1
1
0
1
2
3
3.2 ± 0.125
4.25 ± 0.063
2.67 ± 0.065
>5
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
All the analogues were tested in the National Cancer Institute (NCI) sixty human tumor
cell line assay (NCIꢀ60), where the antiproliferative activities of the compounds were screened
4
4
initially at a single dose of 10 µM (Figure 4 and Figures S2 – S9 in supporting information).
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
Analogue 7 was the most active of the three compounds 6ꢀ8. The mean percentage growth for
analogues 6 and 8 were 38.58 and 40.77, respectively, which amounted to a 60% inhibition of
the growth of all the cancer cell lines on average at 10 ꢀM. In contrast, the mean percentage
growth for analogue 7 was ꢀ16.18, suggesting strong antiproliferative and cytotoxic activity.
However, analogues 6 and 8 exhibited moderate inhibitory activity on some of the melanoma,
renal and leukemia cancer cell lines, but were less active on other cancer cell lines. Analogues 9
– 13 had mean percentage growths of ꢀ4.39, 0.43, 5.41, 1.42 and 34.86, respectively (Figures S5ꢀ
S9 in supporting information). In the dose response assay against the 60 NCI cell lines,
analogues 6 and 8 and 9ꢀ13 had moderate activity (micromolar GI s) on most of the cancer cell
5
0
lines (Figure S10a in supporting information), and did not proceed to in vivo testing at the NCI.
However, the benzyl and substituted benzyl analogues 9-12 were found to be several folds more
active than the methylamine analogue 8 on most of the cell lines and the electron donating or
withdrawing nature of the substituents did not seem to have an effect on activity. However, the
presence of a bulky and hydrophobic chlorine atom in the para position in analogue 12 caused a
significant reduction in activity, which was further compromised by the replacement of the
phenyl group with an even bulkier and more hydrophobic naphthyl group leading to almost
complete loss of activity.
In contrast, analogue 7 inhibited the growth of most of the cancer cell lines at
concentrations <10 nM, (Figure 4 and Figure S10). The cytotoxicity profile of 7 is similar to that
ACS Paragon Plus Environment

Page 17 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of largazole and it was differentially more toxic to some of the ovarian, colon, and melanoma
4
5
cancer cell lines at concentrations < 10 nM.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Heat map showing the antiproliferative activity of 6ꢀ8 in the NCI 60 cell line assay at
0 ꢀΜ
1
As the antiproliferative activity of the synthesized analogues is most likely to be
mediated through HDAC inhibition as in largazole, global histone H3 lysine 9 (H3K9)
acetylation was evaluated by Western blot analysis of colorectal cancer cell line HCT116 after
24 h of cellular exposure at 10 nM, 100 nM, and 1 µM concentrations for each compound
2
7
(
Figure 5) as reported earlier. Significant induction of global acetylation was observed by
exposure to analogue 7 and largazole at 10 nM, and analogues 6 and 8 at 100 nM. These results
are consistent with the results of the antiproliferative assay that 7 is more active than 6 and 8.
Figure 5. All compounds induced global changes of histone acetylation at varying
concentrations. 7 showed the most comparable induction of global acetylation to largazole.
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
Western blot analysis of global histone H3 acetylation following the treatment of HCT116 cells
with largazole and its analogues 6ꢀ8 (concentrations are 10, 100 nM and 1 µM). The Western
blot is a representative figure from an experiment done 3 times.
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
Consistent with the results of the antiproliferative assay, analogues 9 – 12 induced less
global histone acetylation than analogue 7 and largazole at 10 nM, and hardly any global
acetylation was observed with the naphthyl analogue 13 even at 100 nM (Figure S13 in
supporting information).
The HDAC inhibitory activity of the synthesized analogues was determined using an
ELISAꢀbased HDAC activity assay and mammalian cell derived HDAC1, HDAC2, HDAC3 and
4
6
HDAC6 (Figure 6). These assays were conducted after incubation of the test compound with
each isoform for 15 minutes before the addition of the HDACꢀGlo™ substrate. In a single
concentration screen, analogues 6 (at 2 ꢀM) and 7 (at 250 nM) inhibited Class I recombinant
enzymes HDAC1, HDAC2 and HDAC3 significantly, while the Class II enzyme HDAC6 was
relatively unaffected. The selectivity profile was similar to that of largazole tested at 1 ꢀM.
However, analogue 8 (1 ꢀM) showed slightly greater inhibition of HDAC6 compared to 6 and 7,
suggesting that analogue 8 is less class I selective. Consistent with the antiproliferative activity
results, analogues 9 – 13 showed lower inhibitory activity and lower selectivity on the enzymes,
compared to 6ꢀ8.
ACS Paragon Plus Environment

Page 19 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Isoform selectivity screen of analogues 6ꢀ13 and largazole against HDAC1, HDAC2,
HDAC3, and HDAC6 at a single concentration of 2 µM for 6 and 13, 250 nM for 7, 9, 11 and
12, 1 µM for 8, 10 and largazole. Mean percent of deacetylase activity remaining from at least
two independent trials with standard errors were plotted (Table S14 in supporting information).
To further assess selectivity, IC values were determined for both largazole and the most
5
0
potent analog 7 (Table 2). Largazole showed at least 22ꢀfold selectivity for class I HDAC
proteins over Class II HDAC6. Both the IC values and the fold selectivity of largazole are
5
0
3
6,37
consistent with prior literature reports.
Interestingly, analogue 7 showed higher potency and
selectivity for class I HDACs than largazole, with at least 464ꢀfold selectivity for HDAC1,
HDAC2, and HDAC3 over HDAC6 (Table 2).
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
a
Table 2. IC values of largazole and 7 with HDAC1, HDAC2, HDAC3, and HDAC6.
5
0
IC values (nM)
5
0
Compound
HDAC1
61 ± 7
HDAC2
64 ± 5
HDAC3
50 ± 6
HDAC6
1400 ± 100
13000 ± 2000
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
Largazole
7
21 ± 1
28 ± 1
27 ± 2
a
Mean IC value and standard error of at least two independent trials are shown (Figures S15
5
0
and S17 and Tables S16 and S18 in supporting information).
Analogue 7 was selected for in vivo studies in the Developmental Therapeutics Program
of the NCI. In the nontumored animal toxicity assay for acute toxicity, no signs of toxicity was
4
4
observed with 7 at the highest dose of 50 mg/kg tested. Further in vivo testing is in progress at
the NCI.
Molecular Modeling
28
In our previous study we obtained a reasonable solution structure for largazole bound to
HDACꢀ8 using a 30 ns MD trajectory with explicit water molecules . That study was based on
6
the Xꢀray crystal structure (pdb ID:3RQD), in which two largazole thiol molecules were bound
to two HDACꢀ8 molecules. In the binding conformation, not only the two largazole thiol
molecules were in contact with each other, but also both HDACꢀ8 peptides were in contact with
each largazole thiol molecule, possibly making this binding pose somewhat adulterated. During
molecular dynamics, we observed that the position of the macrocyclic ring of one largazole thiol
molecule was shifted towards the area primarily occupied by the second largazole thiol molecule.
47
In the present study, we used available Xꢀray crystal coordinates for HDACꢀ1 {pdb ID: 4BKX}
ACS Paragon Plus Environment

Page 21 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
48
and HDACꢀ2 {pdb ID: 4LY1} as starting structures to study largazole and analogue 7 binding
on these HDAC proteins tested in the study. However, only a partial Xꢀray crystal structure for
HDACꢀ6 is currently available in the pdb database and therefore, within the context of the
present computational study we did not pursue on looking at the binding of largazole and the
analogue 7 in the ligand binding site of HDACꢀ6. As has been the case for their amino acid
sequences (in supporting information Figure S19), overall structural features of HDACꢀ1 and
HDACꢀ2 compared with HDACꢀ8 are quite similar to each other as displayed in Figure 7. After
introducing largazole thiol onto HDACꢀ1 and HDACꢀ2 ligand binding sites using the largazole
thiol bound HDACꢀ8 solution structure, each resultant system was subjected to a lengthy
solution MD simulation to evaluate its binding behavior. In addition, a parallel set of simulations
was performed after appropriately modifying the structure of largazole thiol in each of the above
systems to yield the thiol of analogue 7 to elucidate its binding behavior in HDACꢀ1 and HDACꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.
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
Figure 7. (a) Structural alignments of the Xꢀray crystal structures of HDACꢀ1 (pdb ID: 4BKX;
in red) and HDACꢀ2 (pdb ID: 4lY1; in blue) with HDACꢀ8 (pdb ID: 3RQD; in cyan). The
largazole thiol molecule from largazole thiolꢀbound HDACꢀ8 solution structure is also shown (in
ball and stick representation) (b) The area of the binding site is zoomed in for a closer
examination. All segments that can be in contact with the ligand are remarkably similar in
structures. The loop directly above largazole thiol is seen to be different from that of the HDACꢀ
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
and HDACꢀ2 and became involved in ligand interactions during the MD simulation of
largazole thiol/HDACꢀ8.
To evaluate the properties of isolated ligands used in the molecular modeling, we first
geometry optimized the thiols of largazole and analogue 7 (7ꢀthiol) and calculated the point
charge distribution using the CM5 charge derivation scheme introduced by Truhlar and
49
coworkers which works relatively accurately for systems with buried atoms. Two orthogonal
views of optimized structures of largazole thiol and 7-thiol are shown in Figure 8 with the color
code using partial atomic charges. As can be seen from this figure, the part of the analogue
molecule embedded into HDACs shows remarkable similarities in both structure and charge to
that of largazole thiol. Stability of the proteinꢀligand complexes during molecular dynamics
simulations is confirmed from the root mean square deviations (RMSD) calculated using the last
40 ns segments of each MD trajectory (see Figure 9).
ACS Paragon Plus Environment

Page 23 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. Optimized structures of (a) largazole thiol and (b) 7ꢀthiol. The values of the point
charges at the atomic positions are used in the color coding. Two orthogonal orientations are
given for each compound.
From the plot of RMSD values calculated using backbone heavy atoms after aligning
them with the Xꢀray crystal structure versus the time (Figure 9), one can observe the type of
fluctuations expected from a stable solution structure of a receptor protein in its ligand bound
conformations. Average fluctuations are reasonable ( < 2 Å) for peptides of this size during the
solution dynamics and the analogue bound conformation shows slightly more stability compared
to the largazole thiolꢀbound HDACs. Since RMSD provides information on the global stability
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
of the solvated complex, we have further evaluated the position fluctuations of each residue that
were estimated as the Bꢀfactors from the last 30 ns of the MD trajectory for each HDAC system.
The Bꢀfactors, representing average positional fluctuations plotted here against the residue ID (in
supporting information, Figure S20), are the values calculated for all the back bone heavy atoms
in each residue. Not many specific variations in Bꢀfactors among ligand/HDAC complexes are
observed. When there are variations, they need to be looked at with respect to residues that are in
contact with each ligand (largazole thiol or the analogue). From the previous MD simulation of
the largazole thiol/HDACꢀ8 complex, we found K33, Y100, D101, H143, G151, F152, H180,
F208, D267, M274, G302, and Y306 residues of HDACꢀ8 are in contact with largazole thiol. We
emphasize here that K33 is not in contact with HDACꢀ8 in the Xꢀray crystal structure, but during
dynamics it came in contact with largazole thiol. All the corresponding residues of the two
HDACs can be found in the contact residue list of largazole thiol and 7-thiol bound to HDACꢀ1
and HDACꢀ2 (see Table 3). Note here that residues P29 and H28 (in most cases) in HDACꢀ1 and
HDACꢀ2 are in the loop corresponding to the one that contains K33 of HDACꢀ8.
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
In Figure 10, we display typical binding configurations for largazole thiol and 7-thiol in
HDACꢀ1 and HDACꢀ2. Except for the residues H28 and P29 in the Nꢀterminal region, other
residues listed in Table 3 participate in the common stem region and the macrocycle segment of
largazole thiol and 7-thiol. Largazole thiol in both HDACs seems to be more upright at the top
part of the macrocycle and corresponding macrocyclic ring segment of 7-thiol adopts a slightly
tilted conformation (note that the planes of the two rings are also at an angle to each other). Also,
we note that the atomic fluctuations of heavy atoms of largazole thiol are quite similar to one
another in the two HDAC proteins. Though the values are slightly larger, especially for the
atoms in the macrocyclic ring, than for largazole thiol, 7ꢀthiol (See supporting information,
ACS Paragon Plus Environment

Page 25 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure S21) also has quite similar atomic fluctuations in the two HDACs. This information
points to rather similar binding interactions for each compound in the two slightly different
protein environments. However, most significant information results from the enthalpy
calculations carried out for the four simulations described in the methods. From the results
summarized in the Table 4, largazole thiol is about 3ꢀ4 kcal/mol more stable in the protein
environment as opposed to it being solvated in water. The difference in interaction strength is
only slightly over one kcal/mol for this ligand. For 7-thiol, this difference becomes only half a
kcal/mol for the two proteins, but in the protein environment the analogue is more stable by
about 6 kcal/mol. Also note that with this force field, for 7-thiol, one can estimate approximately
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
kcal/mol more favorable interaction energy between water molecules when compared to
largazole thiol.
Figure 9. The root mean square deviations (A) calculated from the last 40 ns of each MD
trajectory (HDAC1/largazole thiol – black; HDAC2/largazole thiol – green; HDAC1/ 7-thiol –
red; HDAC2/ 7-thiol ꢀ blue).
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
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 10. Snapshots representing typical configurations of (a) and (c) largazole thiol and (b)
and (d) 7-thiol bound to HDACꢀ1 and HDACꢀ2, respectively from molecular dynamics
simulations. Initial structures of HDAC1 and HDAC2 used in MD were from pdb IDs 4BKX
and 4IY1, respectively.
ACS Paragon Plus Environment

Page 27 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. HDAC residues that are in contacts with the ligands.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Protein
Ligand
Protein residues that are in
contact with the ligand
HDACꢀ1
largazole
H28 P29 E98 D99 H141
G149 F150 H178 F205 D264
L271 G301 Y303
HDACꢀ1
Analogue 7
largazole
H28 P29 E98 D99 H141
G149 F150 H178 F205 D264
L271 G301 Y303
HDACꢀ2
H33 P34 D104 H146 G154
F155 H183 F210 D269 L276
G306 Y308
HDACꢀ2
Analogue 7
largazole
H33 P34 D104 H146 G154
F155 H183 F210 D269 L276
G306 Y308
HDACꢀ8*
K33 Y100 D101 H143 G151
F152 H180 F208 D267 M274
G304 Y306
*from reference 28.
Table 4. Enthalpies of binding (in kcal/mol) of largazole thiol and 7-thiol bound to HDACꢀ1 and
HDACꢀ2 calculated from the four simulations described in the method.
Ligand
In HDAC-1
In HDAC-2
In solution (interaction
energy with water)
Largazole thiol
251.1 ± 7.8
256.2 ± 13.2
252.7 ± 8.1
255.7 ± 10.2
248.1 ± 13.0
250.1 ± 10.7
7
-thiol
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
DISCUSSION
Targeting aberrant epigenetic gene silencing has gained broad recognition as a promising
5
approach to cancer treatment. The approval of four HDAC inhibitors as anticancer drugs within
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 last decade, the last two of which were approved within the last two years, underscores the
great potential that HDAC inhibitors hold as new cancer therapeutics. Eighteen different
15
isoforms of HDAC have been identified. Of the four HDAC inhibitors approved as anticancer
drugs by the FDA, SAHA, belinostat, and panobinostat target HDACs in class 1 and 2, while
FK228 is a class Iꢀselective HDAC inhibitor. HDAC inhibitors targeted to specific HDAC
isoforms that are overexpressed in different forms of cancer could be useful as effective
anticancer agents with minimal of undesirable side effects compared their nonꢀselective
2+
counterparts. Although the high sequence homology present in the vicinity of the active site Zn
ion in different HDAC isoforms may not augur well for altering the zincꢀbinding domain of
6
,23,50
HDAC inhibitors to achieve isoform selectivity,
the 14 Å internal cavity near the active site
51,52
constituting the acetate release channel has been explored for this purpose.
In contrast, the
differences in sequence homology in the hydrophobic rim at the entrance to the active site may
offer a site for exploitation to develop isoformꢀselective HDAC inhibitors. In this context, the
class I selectivity of romidepsin may be due to the ability of its large depsipeptide ring to reach
and interact with less conserved regions located further away than those accessible to smaller
head groups of panꢀHDAC inhibitors like vorinostat and belinostat. The depsipeptide ring of
largazole may therefore be considered a viable target for chemical modification to develop
isoformꢀselective HDAC inhibitors as well as to develop analogues with enhanced
23
30,32,45
27,28
pharmacological effects. Several groups,
including ours,
have previously reported
analogues of varying levels of activity generated by altering the western half of the depsipeptide
ACS Paragon Plus Environment

Page 29 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ring. Alterations in the thiazoleꢀthiazoline moiety constituting the eastern half of the ring have
resulted in HDAC inhibitors with promising antiproliferative activity. In fact, the analogue in
which the thiazole ring of largazole was replaced with a pyridine ring is a potent HDAC
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
32
inhibitor. In this study we made radical changes in this part of the depsipeptide ring, while
maintaining the 16ꢀmembered ring size. Analogue 6 incorporated a major structural change in
the depsipeptide ring with the rigid thiazoleꢀthiazoline moiety being replaced with a more
flexible and more hydrophilic polyether chain. In 7, the thiazoleꢀthiazoline moiety was replaced
3
with a biꢀpyridine unit, which was accompanied by a change in the hybridization at Cꢀ7 from sp
2
to sp . Likewise, the thiazole ring of 8 was replaced with a pyridine ring, but the thiazoline ring
3
was replaced with an Nꢀmethyl group. The sp character of the Cꢀ7 center is retained in this
2
3
molecule, but the sp nitrogen of the thiazoline is replaced with an sp nitrogen of the Nꢀmethyl
group. The Nꢀmethyl group was introduced as a handle for structural modification to modulate
hydrophilicity/lipophilicity in this part of the molecule (vide infra).
The three analogues 6 ꢀ 8 showed moderate to high antiproliferative activity on HCT 116
cell line and in the NCI 60 cell line assay. Analogue 7 is the most active of the three analogues,
with an activity profile similar to that of largazole. It showed highly potent growth inhibition on
most cell lines tested in the NCI 60 cell line assay. In the HCT 116 cell growth inhibition assay,
it had a GI of 75 nM compared to a GI of 65 nM observed for largazole The slightly higher
5
0
50
3
1
GI value observed for largazole compared to GI of 44 nM previously reported may be
5
0
50
attributed to differences in the assay conditions.. Analogues 6 and 8 were less active than
largazole and analogue 7, both on HCT 116 cells (GI of 0.75 ꢀM and 0.8 ꢀM, respectively) and
5
0
in the NCI 60 cell line assay. They exhibited moderate inhibitory activity on some of the
melanoma, renal and leukemia cancer cell lines in the NCI assay, but were less active on other
ACS Paragon Plus Environment