Structure-Activity Relationships of Potent, Targeted Covalent Inhibitors That Abolish Both the Transamidation and GTP Binding Activities of Human Tissue Transglutaminase

Article abstract of DOI:10.1021/acs.jmedchem.7b01070

Human tissue transglutaminase (hTG2) is a multifunctional enzyme. It is primarily known for its calcium-dependent transamidation activity that leads to formation of an isopeptide bond between glutamine and lysine residues found on the surface of proteins, but it is also a GTP binding protein. Overexpression and unregulated hTG2 activity have been associated with numerous human diseases, including cancer stem cell survival and metastatic phenotype. Herein, we present a series of targeted covalent inhibitors (TCIs) based on our previously reported Cbz-Lys scaffold. From this structure-activity relationship (SAR) study, novel irreversible inhibitors were identified that block the transamidation activity of hTG2 and allosterically abolish its GTP binding ability with a high degree of selectivity and efficiency (k_inact/K_I > 10⁵ M^-1 min^-1). One optimized inhibitor (VA4) was also shown to inhibit epidermal cancer stem cell invasion with an EC₅₀ of 3.9 μM, representing a significant improvement over our previously reported "hit" NC9.

Full text of DOI:10.1021/acs.jmedchem.7b01070

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES
Article
Structure-Activity Relationships of potent, targeted covalent
inhibitors that abolish both the transamidation and GTP
binding activities of human tissue transglutaminase
Abdullah Akbar, Nicole M.R. McNeil, Marie R. Albert, Viviane Ta, Gautam
Adhikary, Karine Bourgeois, Richard L. Eckert, and Jeffrey W. Keillor
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01070 • Publication Date (Web): 31 Aug 2017
Downloaded from http://pubs.acs.org on September 1, 2017
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
StructureꢀActivity Relationships of potent, targeted covalent
inhibitors that abolish both the transamidation and GTP binding
activities of human tissue transglutaminase
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
‡1
‡1
1
1
Abdullah Akbar, Nicole M. R. McNeil, Marie R. Albert, Viviane Ta,
2
1
2
1
Gautam Adhikary, Karine Bourgeois, Richard L. Eckert and Jeffrey W. Keillor *
1
Department of Chemistry and Biomolecular Sciences, University of Ottawa
1
0 MarieꢀCurie, Ottawa, Ontario, K1N 6N5 Canada
2
Department of Biochemistry and Molecular Biology,
University of Maryland School of Medicine, Baltimore, Maryland, USA
ACS Paragon Plus Environment
1

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
Human tissue transglutaminase (hTG2) is a multifunctional enzyme. It is primarily known for its
calciumꢀdependent transamidation activity that leads to formation of an isopeptide bond between
glutamine and lysine residues found on the surface of proteins, but it is also a GTP binding
protein. Overexpression and unregulated hTG2 activity has been associated with numerous
human diseases, including cancer stem cell survival and metastatic phenotype. Herein, we
present a series of targeted covalent inhibitors (TCIs) based on our previously reported CbzꢀLys
scaffold. From this structureꢀactivity relationship (SAR) study, novel irreversible inhibitors were
identified that block the transamidation activity of hTG2 and allosterically abolish its GTP
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
ꢀ1
ꢀ1
binding ability with a high degree of selectivity and efficiency (kinact/K > 10 M min ). One
I
optimized inhibitor (VA4) was also shown to inhibit epidermal cancer stem cell invasion with an
EC of 3.9 µM, representing a significant improvement over our previously reported ‘hit’ NC9.
50
ACS Paragon Plus Environment
2

Page 3 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
Human tissue transglutaminase (hTG2) is a ubiquitously expressed enzyme and is the most
1
ꢀ2
studied member of the transglutaminase (TGase) family. It is a complex, multifunctional
enzyme and the only member of the TGase family that catalyses protein crossꢀlinking and also
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
plays a role in GTP binding and Gꢀprotein signalling. Its conformation and consequently its
3
ꢀ5
5ꢀ6
function are tightly regulated by the presence of specific allosteric and redox regulators. In
the extracellular matrix (ECM) and plasma membrane, hTG2 is allosterically modulated by
7
calcium ions and redox proteins to primarily exist in an open or extended conformation that is
8
catalytically active when specific disulfide bonds are reduced. In this conformation the catalytic
core, which contains the catalytic triad Cys277, His335 and Asp358, is accessible and capable of
catalysing the formation of isopeptide bonds (crossꢀlinks) between peptide bound Gln and Lys
9
residues. This crossꢀlinking activity is also referred to as transamidation and occurs via a pingꢀ
1
0
pong mechanism. First, the nucleophilic active site thiolate (Cys277) attacks an acyl donor
substrate (ie. a peptideꢀbound Gln residue) resulting in the release of one equivalent of ammonia
and the formation of the intermediate thioester. Subsequently, the thiolate is regenerated by
nucleophilic attack of an acyl acceptor substrate (ie. a peptideꢀbound Lys residue) to afford the
isopeptide product, or the thioester is cleaved by water to afford the deamidated (or hydrolysis)
1
1
product. Intracellularly, hTG2 is regulated by guanidineꢀcontaining nucleotides, such as GTP
9
that bind hTG2 at a site remote from the catalytic active site. Upon GTP binding, hTG2
primarily adopts a closed or compact conformation where the catalytic Cys277 becomes
inaccessible. In this closed conformation, it is believed that little or no transamidation activity
1
2
occurs while several cellular signaling pathways are impacted.
ACS Paragon Plus Environment
3

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
1
Due to the multifunctional nature and ubiquitous expression of hTG2, it is involved in a
2
, 13ꢀ14
variety of normal physiological processes that have been reviewed elsewhere;
however, its
1
5
unregulated activity has also been implicated in several disease processes. The unregulated
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
16
transamidation/deamidation activity of hTG2 has been implicated in atheroscelorsis, pulmonary
1
7ꢀ18
19ꢀ20
21
22
fibrosis,
liver fibrosis,
chronic kidney disease, Alzheimer’s disease
In addition, recent studies have shown that hTG2 is also an oncogenic protein, where
its GTPꢀbinding function has been associated with cancer cell proliferation, metastasis, and
and celiac
2
3ꢀ26
disease.
27ꢀ31
cancer stem cell survival.
Interestingly, hTG2 knockout mice appear developmentally and
32
reproductively normal, mainly displaying delayed wound healing and poor response to stress.
Due to the lack of serious and fatal deficiencies observed in hTG2 knockout mice with respect to
normal biological functions, as well as its involvement in a wide range of diseases, hTG2 has
32ꢀ34
been proposed as a safe therapeutic target.
Several academic and industrial research groups have contributed to the design and evaluation
of peptidomimetic and small molecule reversible and covalent hTG2 inhibitors, as summarized
35ꢀ37
in several recent reviews.
While some studies have focussed on reversible inhibitors of
3
8
hTG2, the majority have focussed on the development of potent and selective covalent
inhibitors that target the nucleophilic activeꢀsite Cys277 residue. Some representative examples
of peptidic and peptidomimetic irreversible inhibitors of hTG2 are shown in Figure 1. These
inhibitors show some structural similarity with respect to the scaffold: a hydrophobic group (Cbz
or AcꢀPro) located on the Nꢀterminal side of the electrophilic warhead, followed by a small
hydrophobic residue Cꢀterminal side (Phe or Tyr) or a more extended hydrophobic sequence
(
e.g. ProꢀPhe or ProꢀLeu). These structures also display a broad variety of electrophiles used as
39
40
a warhead, including a sulfonium methyl ketone, a halodihydroisoxazole, a diazomethyl
ACS Paragon Plus Environment
4

Page 5 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
8
41
42ꢀ47
ketone, an α,βꢀunsaturated ester and an acrylamide group,
all of which have been used in
3
5, 37, 48
the design of hTG2 inhibitors.
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 1. Representative irreversible inhibitors of hTG2 from the recent literature.
The development and use of targeted covalent inhibitors (TCIs) as potential therapeutics has
49
also been reviewed recently. Some of the advantageous features of this class of inhibitors are
their high efficiency and extended duration of action. The current consensus of the hTG2
inhibitor field suggests that continued efforts in the design and evaluation of hTG2 inhibitors is
3
6
needed due to the limited amount of available structural data of the inhibitorꢀbound protein. In
ACS Paragon Plus Environment
5

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
this work, we present the synthesis and in vitro kinetic evaluation of a series of novel irreversible
inhibitors that block both the transamidation and GTPꢀbinding activities of hTG2, as well as
biological data illustrating cellular efficacy.
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
ACS Paragon Plus Environment
6

Page 7 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
RESULTS AND DISCUSSION
Inhibitor Design
8
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
Khosla and coꢀworkers have reported a crystal structure of the open form of hTG2 with an
irreversible inhibitor bound in the catalytic active site – namely, AcꢀP(DON)LPFꢀNH , shown in
2
Figure 1. Based on this crystal structure (PDB code 2Q3Z), the catalytic active site has been
described as a relatively hydrophobic binding groove flanking a hydrophobic tunnel that contains
1
1, 40, 50
the catalytic Cys277 at its base. Previous inhibition studies based on the CbzꢀPhe
and
4
6, 51
CbzꢀLys
scaffolds have shown that the Nꢀterminal Cbz protecting group confers affinity to
these irreversible inhibitors (see Figure 1). As such, the Cbz group was maintained throughout
this SAR study.
As mentioned above, a variety of electrophilic warheads have also been used in irreversible
14, 35, 37
inhibitors of TGases.
We, and others, have found that the acrylamide warhead offers an
excellent balance between stability and reactivity for use in targetedꢀcovalent inhibitors of
4
2ꢀ47, 50, 52
hTG2.
This balance may also explain why it has seen broad application in TCI’s
4
9, 53
recently developed against other target proteins, as well.
Historically, some of the most potent acrylamide inhibitors that we prepared in previous
studies featured a CbzꢀLys scaffold and an acrylamide warhead. For example, we found that
compounds 1 and 2 (Figure 2) were remarkably efficient inhibitors of guinea pig liver tissue
transglutaminase (gplTG2). Since then, CHDI (Cure for Huntington’s Disease Initiative) has
46
reported that our compound 1 (Figure 2) served as a starting point for their SAR study that
resulted in Compound 4l, shown in Figure 1. In our group, compound 3, also known as NC9
5
1
(
Figure 2) was subsequently designed to be a fluorescent TGase probe.
ACS Paragon Plus Environment
7

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
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. Lysineꢀbased irreversible inhibitors and their kinetic parameters of inhibition against
4
2, 51
gplTG2.
In this work, we evaluated some of our past compounds such as 1 and 3 against human TG2.
We also envisioned a series of modifications that may provide insight into the structural features
necessary for potent and selective targetedꢀcovalent inhibition of hTG2. While maintaining the
N-terminal Cbz protecting group for affinity and the ‘select’ acrylamide warhead (see above), we
systematically varied the length of the side chain, the peptide backbone spacer, and the Cꢀ
terminal R group (Figure 3), seeking to optimize not only inhibition efficiency, but also drugꢀlike
properties such as calculated LogP, PSA, and the number of rotatable bonds.
ACS Paragon Plus Environment
8

Page 9 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 3. Proposed structural modifications for this SAR study with hTG2.
Synthesis and StructureꢀActivity Relationship studies
4
2, 46
Starting from compound 1
the carboxylic acid group was activated using peptide coupling
5
4
reagents such as EDC/NHS or EDC/HOBt. From this activated intermediate, a variety of
nitrogen nucleophiles were added, giving access to a series of inhibitors in just a few synthetic
steps. In the case of compounds 4 and 5 (Scheme 1), ammonium chloride or glycine methyl ester
were used as the nucleophile, respectively. From compound 5, saponification of the methyl ester
using LiOH afforded the free carboxylic acid derivative 2. Collectively, these compounds are
abbreviated as CbzꢀLys(Acr)ꢀR deriatives. Irreversible inhibitor 3 was synthesized according to
51
our previously published procedure.
ACS Paragon Plus Environment
9

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
Scheme 1. Synthesis of CbzꢀLys(Acr)ꢀR derivatives of 1. (a) EDCꢀHCl, NHS, acetonitrile, rt, 16
h; (b) NH Cl, DABCO, acetonitrile, rt, 16 h; (c) glycine methylester, triethylamine, acetonitrile,
4
rt, 4 h; (d) LiOH, THF:H O, rt, 1 h.
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
When the carboxylic acid derivatives 1 and 2 were previously synthesized and evaluated as
irreversible inhibitors of gplTG2, they displayed excellent kinetic inhibition parameters with KI
6
ꢀ1
ꢀ1
values in the high nanomolar range and efficiency constants of 10 M min , whereas 3 was two
4
2, 51
orders of magnitude less efficient.
To begin the present SAR study, we reꢀevaluated
ACS Paragon Plus Environment
10

Page 11 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
compounds 1ꢀ2 and 3 as irreversible inhibitors of hTG2. A representative example of this kinetic
evaluation is shown in Figure 4.
A
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
B
Figure 4. Determination of irreversible inhibition parameters for compound 2 using a previously
5
5
reported continuous colorimetric assay. (A) Timeꢀdependent inactivation curves of enzymatic
reaction in the presence of various concentrations of 2. (B) Nonlinear regression to a hyperbolic
equation of kobs versus varying concentrations of 2.
5
5
56
Using a colorimetric assay under Kitz & Wilson conditions, rate constants were measured
for the timeꢀdependent loss of acylꢀtransferase activity (Figure 4A). Fitting of the hyperbolic
concentration dependence of the observed rate constants of inactivation (Figure 4B) provided the
kinetic parameters shown in Table 1 (see Experimental for details).
ACS Paragon Plus Environment
11

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
Table 1. Kinetic parameters of substituted CbzꢀLys(Acr)ꢀR.
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
kinact/KI
ꢀ1
Inhibitor
R group
K (µM)
I
kinact (min )
3
ꢀ1
ꢀ1
(
10 M min )
1
2
60.3 ± 18.6 0.796 ± 0.130
13.2 ± 4.6
48.1 ± 10.3
35.1 ± 15.1
40.0 ± 5.9
1.26 ± 0.17
1.08 ± 0.46
1.45 ± 0.12
26.1 ± 6.6
30.7 ± 2.5
36.2 ± 6.1
a
4
5
3
33.9 ± 3.4
2.60 ± 0.17
76.4 ± 9.1
a) Kinetic parameters for compound 4 were determined using a double reciprocal method of
data analysis (see Experimental).
ACS Paragon Plus Environment
12

Page 13 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
While compounds 1 and 2 were among the best irreversible inhibitors of gplTG2, their affinity
for hTG2 and their reaction efficiency are shown here to be two orders of magnitude worse.
Inhibitor 2, which contains a glycine linker between the CbzꢀLys(Acr) and the negatively
charged carboxylate, shows slightly increased affinity and efficiency compared to 1. This gain in
inhibition efficiency may be due to distancing the negative charge from possible unfavorable
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
8
interactions in the hydrophobic binding pocket. Replacing the carboxylic acid groups of 1 and 2
with a neutral terminal amide in 4 or a methyl ester in 5 results in minor gains in affinity and
efficiency, consistent with the elimination of negative charge and increase in LogP values (3.42
for 4, 3.09 for 5).
Extending the CbzꢀLys(Acr) scaffold with the addition of a
polyethyleneglycol linker and dansyl group in 3 resulted in a comparable inhibition constant to
that determined for gplTG2 (K = ~30 µM). However, 3 appears to inactivate the human enzyme
I
ꢀ1
with a much higher rate constant based on the elevated kinact value of 2.60 min compared to
ꢀ1
0
.483 min , for hTG2 and gplTG2 respectively. The observed increase in k
with respect to
inact
hTG2 may be due to a slightly different binding mode that places the acrylamide warhead closer
to the active site Cys277 residue. This increased rate constant results in a 5ꢀfold increase in the
overall efficiency (kinact/K ) of 3 as an inhibitor of hTG2.
I
The results from Table 1 also suggest that the aromatic and hydrophobic dansyl group may
provide beneficial contacts with a distal binding pocket of hTG2 that the truncated compounds 1ꢀ
5 cannot access. More specifically, we speculated that the dansyl group of 3 may be discretely
bound by hTG2, rather than extended away from it by the PEG spacer, as was our intent in the
5
1
design of this fluorescent probe. Docking 3 into the open conformation of hTG2 (PDB code
8
2
Q3Z) suggested that the dansyl group may be bound in the hydrophobic pocket on the surface
of hTG2 in which the Phe side chain of AcꢀP(DON)LPFꢀNH was bound, in the coꢀcrystal
2
ACS Paragon Plus Environment
13

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
structure reported by Khosla. In this putative binding model (Figure S1), the PEG spacer forms a
long, flexible loop, and few productive interactions with the enzyme.
On the other hand, 3 was tested in biological assays and found to be surprisingly effective
against human squamous cell carcinoma. Human TG2 has been shown to be highly elevated in
epidermal cancer stem (ECS) cells and hTG2 knockdown or inactivation with 20 µM 3 reduced
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
9
ECS cell proliferation and survival, the expression of epithelialꢀmesenchymal transition (EMT)
2
8
28
transcription factors and markers, and metastatic phenotype. In TG2 knockout ECS cells,
treatment with 3 had no effect, suggesting that its biological activity is not related to offꢀtarget
30
interaction. Furthermore, administration of 3 by intraperitoneal injection at 20 mg/kg reduced
3
1
the growth of mouse xenograft tumors initiated by injection of cancer stem cells. In light of
this biological activity, we considered 3 to be an interesting hit, and sought to optimize it further.
For an inhibitor to become a drug candidate, its potency, efficiency and bioavailability must all
be optimized. With the ultimate goal of improving all of these features, we considered
5
7
58
Lipinksi’s and Veber’s rules and turned our attention to reducing the number of rotatable
bonds and the polar surface area (PSA) in our inhibitors. To this end, we first sought to
systematically decrease the length of the diamine spacer between the CbzꢀLys(Acr) and the
dansyl functional group of 3. The monoꢀdansylated alkyl amines 6ꢀ8 in Scheme 2 were
synthesized and coupled to the activated carboxylic acid of 1 to give inhibitors 10ꢀ12, and a
piperazine linker (9) was also coupled to 1 to produce the more rigid inhibitor 13 (also known as
VA4).
ACS Paragon Plus Environment
14

Page 15 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 2. Synthesis of hTG2 inhibitors with varying spacer length. (a) EDCꢀHCl, HOBt,
triethylamine, acetonitrile, rt, 16 h.
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
ACS Paragon Plus Environment
15

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
Table 2. Kinetic parameters of irreversible inhibitors with varying backbone spacers.
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
kinact/KI
PSA
ꢀ1
a
Inhibitor
X group
K (µM)
I
kinact (min )
nRB
3
ꢀ1
ꢀ1
2
(
10 M min )
(Å )
NHꢀ
3
(CH
2
CH
2
O)
2
ꢀ
33.9 ± 3.4
2.60 ± 0.17
76.4 ± 9.1
164
23
CH CH NH
2
2
b
1
1
1
0
1
2
HNꢀ(CH ) ꢀNH 27.1 ± 11.9
2.19 ± 0.94
0.85 ± 0.25
1.45 ± 0.12
1.40 ± 0.13
80.8 ± 4.7
27.6 ± 1.7
61.5 ± 9.6
107 ± 23.8
146
146
146
128
19
18
17
14
2
4
b
b
HNꢀ(CH ) ꢀNH
30.5 ± 9.4
23.5 ± 3.1
12.9 ± 2.6
2
3
HNꢀ(CH ) ꢀNH
2
2
13
piperazine
a) Polar Surface Area and the number of Rotatable Bonds were calculated using the
Molinspiration molecular property calculation engine (http://www.molinspiration.com).
b) Kinetic parameters for compounds 10ꢀ12 were determined using a double reciprocal
method of data analysis (see Experimental).
As shown in Table 2, reducing the spacer length from eight atoms in 3 to four atoms in
compound 10 resulted in a dramatic reduction in the number of rotatable bonds and PSA, but
little change with respect to the kinetic parameters measured. Interestingly, reducing the spacer
ACS Paragon Plus Environment
16

Page 17 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
to three methylene groups (11) resulted in a marked reduction in efficiency, mainly due to a
significant drop in the inactivation rate constant. Decreasing the spacer length to two methylene
units (12) restored inhibitor efficiency to the values seen for the hit compound 3. Maintaining the
distance associated with a twoꢀmethylene spacer but in the form of the more rigid piperazine
group gave inhibitor 13, displaying the best inhibition parameters in the series. In addition to
exceeding the overall inhibition efficiency of 3, inhibitor 13 also showed a greater than twoꢀfold
reduction in the inhibition binding constant, suggestive of more favorable interactions with the
binding pocket of hTG2 compared to inhibitors with more flexible spacers. It is interesting to
note that CHDI also discovered the suitability of the piperazine group at this position in
analogous inhibitors, but from an expansion strategy among a series of Cꢀterminal tertiary
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
4
6
2
amides. With a PSA below 140 Å and nearly half as many rotatable bonds as 3, we were also
hopeful that the bioavailability of 13 would be substantially improved (see below).
Using 13 as a new parent compound, additional inhibitors were then prepared that contained
the dansyl piperazine but whose side chains bearing the acrylamide warhead were systematically
varied with respect to their length. Compounds 14–16 were synthesized from commercially
4
3, 46
available starting materials using modified literature procedures,
and were coupled to the
monoꢀdansyl piperazine to afford irreversible inhibitors 17ꢀ19 (Scheme 3).
ACS Paragon Plus Environment
17

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
Scheme 3. Synthesis of irreversible hTG2 inhibitors 17ꢀ19. (a) EDCꢀHCl, HOBt, triethylamine,
acetonitrile, rt, 16 h.
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 3. Kinetic parameters of derivatives having varied side chain length.
kinact/KI
ꢀ1
a
Compound (CH2)n
K (µM)
I
kinact (min )
nRB
3
ꢀ1
ꢀ1
(
10 M min )
17
n = 1
n.d.
n.d.
n.d.
11
1
8
n = 2
n = 3
n = 4
n.d.
n.d.
n.d.
12
13
14
b
19
47.7 ± 37.8
12.9 ± 2.6
1.18 ± 0.93
1.40 ± 0.13
24.7 ± 2.2
107 ± 23.8
13
a) The number of Rotatable Bonds was calculated using the Molinspiration molecular
property calculation engine (http://www.molinspiration.com).
b) Kinetic parameters for compound 19 were determined using a double reciprocal
method of data analysis (see Experimental). n.d = not detectable
ACS Paragon Plus Environment
18

Page 19 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
As shown in Table 3, although reducing the length of the side chain reduced the number of
rotatable bonds in inhibitors 17ꢀ19, it also had a drastic effect on the efficacy of the inhibitors.
Timeꢀdependent inhibition of hTG2 within a reasonable inhibitor concentration range was only
observed for compound 19 and consequently we were unable to determine the kinetic parameters
for 17 and 18. As another means of comparison, we tested high concentrations of inhibitors 17ꢀ
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
1
9 in our colorimetric assay and measured relative observed rate constants (k ) from the timeꢀ
obs
dependent inactivation curves, as shown in Figure 5.
Figure 5. Observed rate constants (k ) of inactivation of hTG2 with 18 µM of inhibitors 17ꢀ19.
obs
From both Table 3 and Figure 5, it is obvious that reducing the number of methylene units of
the side chain below 4 results in a dramatic loss of activity. These results are consistent with
42
previous studies performed on the CbzꢀPheꢀX and CbzꢀXꢀGly scaffolds. We hypothesize that
both the Cbz and dansyl functional groups present favorable contacts with the extended
8
hydrophobic binding pocket on the surface of hTG2 and thus help to position the acrylamide
warhead for attack by Cys277 at the base of the active site tunnel. A sidechain of one or two
ACS Paragon Plus Environment
19

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
methylene units is apparently not long enough to position the acrylamide group for nucleophilic
attack, and therefore no significant enzyme inactivation is observed. Extending the side chain to
three methylene units results in the reappearance of activity for inhibitor 19, but only modestly in
comparison to 3 and 13. It is evident that the side chain length is a critical component of this
inhibitor scaffold for successful inhibition of hTG2.
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
Having established the advantages of a piperazine linker and a fourꢀmethylene side chain, we
moved our attention to modification of the dansyl group. We began by synthesizing derivatives
that maintained the sulfonamide connection, while replacing the dansyl group with various
aromatic and aliphatic functional groups. Our goal was to probe the interaction with the putative
hydrophobic binding pocket, while optimizing for the number of rotatable bonds and
lipophilicity. First, the piperazine sulfonamide intermediates 21 – 29 were synthesized (see
supporting information), followed by peptide coupling with carboxylic acid 1 to give final
compounds 30 – 38 (Scheme 4).
ACS Paragon Plus Environment
20

Page 21 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 4. Synthesis of sulfonamideꢀcontaining hTG2 inhibitors. (a) triethylamine, DCM, rt, 16
h; (b) EDCꢀHCl, HOBt, triethylamine, acetonitrile, 16 h.
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
Compound 30 differs from 13 only by the lack of the dimethyl amino group on the
naphthalene ring, so these inhibitors are very similar with respect to calculated LogP and
rotatable bonds. Interestingly, the inhibition constants (K ) for both 30 and 13 are also very
I
comparable (Table 4); however, the rate constant of inactivation (kinact) of 30 is more than threeꢀ
fold less than that of 13, resulting in a lower overall inhibition efficiency for 30. Changing the
orientation of the naphthyl moiety in compound 31 gave a modest gain in inhibition efficiency.
Reducing the size of the aromatic ring from naphthyl to phenyl in compound 32 resulted in a
slight decrease in the calculated LogP, to below 5, although its inhibition kinetics were no better
than those of 31. The introduction of the more flexible phenylacetyl group in compound 33
resulted in a reduction of inhibition efficiency to values comparable with 30. Comparison of the
kinetic data for compounds 30 – 33 suggests that the binding pocket is relatively indiscriminate
and will accept moderate to large hydrophobic groups, and that there is little difference in the
calculated LogP values, as expected.
The importance of the aromaticity of these substituents was then investigated by synthesizing
and evaluating nonꢀaromatic compounds 34 – 37. Based on the kinetic data shown in Table 4,
aromaticity is not necessary to achieve modest inhibition efficiency; the cyclohexyl (34),
isopropyl (35), ethyl (36) and methyl (37) sulfonamides all gave comparable inhibition values to
ACS Paragon Plus Environment
21

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
those of the aromatic derivatives. Interestingly, the electronꢀrich thiophene sulfonamide
derivative (38) showed poor inhibition, while the methyl sulfonamide (37) had the best inhibition
kinetics of the series. Furthermore, inhibitor 37 has a significantly lower calculated LogP value,
fewer rotatable bonds and a molecular weight below of only 480.6 g/mol, all of which may
slightly favor its cellular permeability.
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
ACS Paragon Plus Environment
22

Page 23 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 4. Kinetic parameters of piperazine sulfonamide derivatives.
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
a
kinact/KI
PSA
ꢀ
1
a
a
Inhibitor
R group
K (µM)
kinact (min )
LogP
nRB
I
3
ꢀ1
ꢀ1
2
(10 M min )
(Å )
13
12.9 ± 2.6
1.40 ± 0.13
107 ± 23.8
5.49
128
14
30
10.4 ± 1.8
6.7 ± 1.1
0.39 ± 0.03
0.38 ± 0.03
37.2 ± 7.1
5.43
5.46
125
125
13
13
31
57.3 ± 10.6
ACS Paragon Plus Environment
2
3

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
3
2
3
8.8 ± 1.5
14.7 ± 3.2
11.0 ± 1.1
9.1 ± 1.6
36.0 ± 3.6
6.4 ± 1.8
40.1 ± 6.0
0.47 ± 0.04
53.3 ± 9.8
34.9 ± 8.7
62.9 ± 7.1
54.7 ± 10.4
39.4 ± 4.5
78.4 ± 23.7
36.8 ± 6.4
4.28
125
125
125
125
125
125
125
13
14
13
13
13
12
13
3
0.47 ± 0.06
0.69 ± 0.03
0.50 ± 0.03
1.42 ± 0.09
0.50 ± 0.05
1.48 ± 0.13
4.34
4.65
3.48
3.12
2.74
4.22
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
34
3
5
6
3
37
3
8
a) LogP and the number of Rotatable Bonds were calculated using the Molinspiration molecular property
calculation engine (http://www.molinspiration.com).
ACS Paragon Plus Environment
2
4

Page 25 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The final series of derivatives that were investigated was also constructed from a Cbzꢀ
Lys(Acr)ꢀpipꢀX scaffold, but the Cꢀterminal sulfonamide group was replaced with a carboxamide
functional group. This was expected to decrease the polar surface area of the inhibitors, and
possibly to orient the Cꢀterminal hydrophobic group slightly differently, due to the
conformational difference between a carboxamide and a sulfonamide. Piperazine amide
intermediates 40 – 48 were prepared from the Bocꢀprotected piperazine (39). Various acid
chlorides or activated carboxylic acids were also coupled to 39 to afford the desired Bocꢀ
piperazine intermediates. Removal of the Boc group, followed by peptide coupling with
acrylamide 1 gave the desired final compounds 49 – 57 (Scheme 5).
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 5. Synthesis of carboxamide hTG2 inhibitors. (a) EDCꢀHCl, NHS, acetonitrile, rt, 16 h;
(b) triethylamine, acetonitrile, rt, 4 h; (c) triethylamine, DCM, rt, 4 h.
ACS Paragon Plus Environment
25

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
Table 5. Kinetic parameters of piperazine amide derivatives.
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
kinact/KI
ꢀ1
a
a
2
a
Inhibitor
R group
K (µM)
I
kinact (min )
LogP
PSA (Å ) nRB
3
ꢀ1
ꢀ1
(
10 M min )
49 (AA9)
8.9 ± 1.0
0.90 ± 0.05
101 ± 12.8
5.09
108
12
5
0 (AA10)
51
13.4 ± 2.3 1.21 ± 0.13
36.6 ± 4.8 4.77 ± 0.46
20.0 ± 5.8 2.26 ± 0.37
90.3 ± 18.0
130 ± 21.2
113 ± 37.5
5.11
3.93
4.49
108
108
108
12
12
13
52
ACS Paragon Plus Environment
2
6

Page 27 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5
3
52.1 ± 6.7 2.06 ± 0.18
36.9 ± 6.1
91.3 ± 17.9
61.4 ± 6.4
37.0 ± 6.4
2.73
3.23
2.69
2.64
108
121
121
121
11
12
12
12
54
32.1 ± 5.4 2.93 ± 0.29
23.4 ± 2.2 1.44 ± 0.07
38.7 ± 5.7 1.43 ± 0.13
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
5
5
6
5
5
7 (VA5)
11.2 ± 2.6 0.69 ± 0.09
62.1 ± 16.4
3.43
158
12
a) LogP, Polar Surface Area and the number of Rotatable Bonds were calculated using the Molinspiration
molecular property calculation engine (http://www.molinspiration.com).
ACS Paragon Plus Environment
2
7

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
59
Direct comparison of the 1ꢀnaphthyl (49), 2ꢀnaphthyl (50), phenyl (51) and benzyl (52)
amide derivatives in Table 5 to their corresponding sulfonamide derivatives in Table 4 suggests
that the presence of an amide functional group in lieu of the sulfonamide is beneficial for the
inhibition of hTG2. All four of these aromatic derivatives (49 – 52) show at least a twoꢀfold
increase in overall inhibition efficiency compared to their sulfonamide counterparts (30 – 33),
and a significant reduction in their polar surface areas. Surprisingly, this trend does not hold for
the acetyl derivative (53). Although the polar surface area is less than that of the methyl
sulfonamide derivative (37), as expected, its overall inhibition efficiency almost three times
lower.
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 introduction of an Nꢀheterocycle in the form of three different pyridine isomers showed an
interesting trend in inhibition efficiencies as the nitrogen was moved around the aromatic ring
relative to the amide bond (Table 5). The picolinamide derivative (54) showed the best overall
efficiency, followed by the nicotinamide derivative (55) and then the isonicotinamide derivative
(
56). While all three of the Nꢀheterocycle inhibitors had lower calculated LogP values compared
to the nonꢀpolar phenyl compound (51), they were also significantly less efficient inhibitors. This
may suggest that a heterocycle substituent is not necessary but can be tolerated by hTG2;
perhaps conformational rotation of the pyridine ring allows the polar nitrogen of the
picolinamide derivative to be oriented towards solvent, while the rest of the ring occupies a
3
0, 59ꢀ60
hydrophobic cavity. Finally, the coumarin derivative (57, aka VA5)
was intended for use
as a cellular fluorescent probe that would be visible in the blue region. We were gratified to note
that the inhibition kinetic parameters of this compound were comparable to those of 13, although
its polar surface area was considerably larger, raising concerns over its permeability.
ACS Paragon Plus Environment
28

Page 29 of 70
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Overall, upon consideration of affinity, efficiency, and physicochemical properties, we judged
1
3 and compound 49, also known as AA9, to be the best lead compounds from the sulfonamide
and amide piperazinyl linked derivatives. Both compounds are highly efficient irreversible
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
5
ꢀ1
ꢀ1
2
inhibitors (kinact/K > 10 M min ) and have polar surface areas smaller than 140 Å ; however,
I
both have calculated LogP values just over 5, and more than 10 rotatable bonds.
Inhibition of GTPꢀbinding
As mentioned above, unregulated transglutaminase activity has been implicated in the
pathogenesis of several diseases. In most of these diseases, it is the enzymatic transamidase
function of hTG2 that is believed to contribute to the pathology; however, more recently it has
been shown that the GTPꢀbinding ability of the enzyme is involved in certain types of
2, 27ꢀ28
cancer.
Consequently, targeting hTG2 to abolish its enzymatic activity and/or GTP binding
offers a potential therapeutic approach towards treatment of associated diseases.
A crystal structure of hTG2 in complex with GDP illustrated how guanidine nucleotide
9
binding causes the two Cꢀterminal βꢀbarrel domains to be folded over the catalytic core,
showing how acylꢀtransfer activity and GTPꢀbinding are conformationally mutually exclusive.
Our irreversible inhibitors are competitive with respect to the acylꢀdonor substrate of hTG2, and
we have observed by mass spectrometry that one equivalent of 3, 50, or 57 are covalently
5
9
incorporated into irreversibly inhibited hTG2, presumably through reaction with the active site
cysteine. Therefore, our inhibitors should not compete directly with GTP binding. However, we
1
2, 30
59, 61
have also shown, by both FRETꢀFLIM
and by CEꢀMS,
that micromolar concentrations
of these acrylamideꢀbased irreversible inhibitors can lock hTG2 in its open/extended
conformation, rendering hTG2 incapable of binding GTP. To that end, two of our lead
ACS Paragon Plus Environment
29