Imidazolium-based warheads strongly influence activity of water-soluble peptidic transglutaminase inhibitors

Article abstract of DOI:10.1016/j.ejmech.2013.05.018

New peptidic water-soluble inhibitors are reported. In addition to the carboxylate moiety, a new polar warhead was explored. Depending on the size of its substituents, the newly appended imidazolium scaffold designed to enhance the hydrophilic character o

Full text of DOI:10.1016/j.ejmech.2013.05.018

European Journal of Medicinal Chemistry 66 (2013) 526e530
Contents lists available at SciVerse ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Short communication
Imidazolium-based warheads strongly influence activity
of water-soluble peptidic transglutaminase inhibitors
*
Eduard Badarau, Alexandre Mongeot, Russell Collighan, Dan Rathbone, Martin Griffin
School of Life and Health Sciences, Aston University, Aston Triangle, Birmingham B4 7ET, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
New peptidic water-soluble inhibitors are reported. In addition to the carboxylate moiety, a new polar
warhead was explored. Depending on the size of its substituents, the newly appended imidazolium
scaffold designed to enhance the hydrophilic character of the inhibitors could induce a good inhibition
for tissue transglutaminase (TG2) and blood coagulation factor XIIIa (FXIIIa). Correlated with the narrow
tunnel that hosts the target catalytic cysteine residue, the various modulations suggest a bent confor-
mation of the ligands as the binding pattern mode. Analogues in the dialkylsulfonium series were also
tested and showed specificity for TG2 over FXIIIa.
Received 8 September 2012
Received in revised form
10 May 2013
Accepted 16 May 2013
Available online 7 June 2013
Keywords:
Imidazolium warheads
Dialkylsulfonium
Ó 2013 Elsevier Masson SAS. All rights reserved.
Peptidic inhibitors
Tissue transglutaminase (TG2)
Coagulation factor XIIIa (FXIIIa)
1. Introduction
concept, such polar molecules were successfully used in preclinical
studies of diabetic nephropathy and kidney scarring where they
Transglutaminases (TGs) are a group of enzymes that catalyse
the formation of protein crosslinks by forming isopeptide bonds
between peptide bound glutamine and lysine residues, either in
an inter- or intramolecular manner [1]. Although under normal
physiological conditions they are responsible for the stabilisation of
high molecular weight protein structures e.g. in skin and in hair,
abnormal levels of transglutaminases in particular the ubiquitous
TG2 are associated with many disease states. These include
neurodegenerative diseases, celiac sprue, cataract and fibrosis [2].
Starting from the hypothesis that in many disease states such as
fibrosis, celiac sprue and cancer, transglutaminases such as TG2 act
extracellularly, we initiated a medicinal chemistry programme
aimed at developing tools for such diseases. The design of highly
hydrophilic inhibitors that could control the over-expressed levels
of extracellular TG2 associated with the above-mentioned pathol-
ogies would also have the advantage of a desirable cellular toxicity
profile for such molecules, as their intracellular access would be
restrained. These were, in fact, the fundamentals when we devel-
oped a new series of TG2 inhibitors incorporating, in addition to the
polar dimethylsulfonium warhead pioneered by Pliaura et al. in
early 90s [3], a new carboxylate moiety [4]. As a proof of this
induced no animal toxicity up to 120 days and successfully reduced
kidney fibrosis and scarring by up to 85% with a significant increase
in kidney function [5e7].
As a continuation of our research programme on TG2, we syn-
thesized new analogues by replacing the previous warhead (dime-
thylsulfonium) with a new polar moiety, namely imidazolium-
based warheads. The discovery that imidazolium derivatives inhi
bit transglutaminases was previously disclosed by Merck in 1992 [8]
(A, Fig. 1). Subsequently, the same group showed that the key
structural feature for the irreversible reaction with the enzyme is in
fact the acetonyl-thioimidazolium. Acetonylation of the catalytic
CYS residue would thus render the enzyme inactive while releasing
the imidazolium thione [9].
Since such small compounds patented by Merck have many
drawbacks for future pharmacological use in terms of TG specificity
and target attainability, we incorporated this imidazole feature into
our previously developed inhibitors. The hydrophilic character of
our new compounds would be preserved due to the presence of
both carboxylate and imidazolium salts in their structure (Fig. 1). As
the imidazolium-based derivatives were patented by Merck for
their ability to inhibit the FXIIIa, we evaluated our new series on
both TG2 and FXIIIa.
Several crystal structures for TG2 have been published.
They revealed a huge conformational change for this enzyme be-
tween the nucleotide-bound conformation [11,12] (“closed”
* Corresponding author.
E-mail address: martin.griffin2404@gmail.com (M. Griffin).
0223-5234/$ e see front matter Ó 2013 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.ejmech.2013.05.018

E. Badarau et al. / European Journal of Medicinal Chemistry 66 (2013) 526e530
527
Fig. 1. Influence of polar groups on the hydrophilicity of newly-designed compounds. Log D values at pH ¼ 7.4 in an octanol/water mixture were predicted using SPARC
Software [10].
conformation) and an irreversible inhibitor-bound conformation
[13,14] (“open” conformation). As the key catalytic CYS residue was
revealed to be hosted in a narrow tunnel, several research groups
assumed that the electrophilic warhead should be size-limited, in
order to reach and react irreversibly with the nucleophilic CYS
acids as the central substituents (H for GLY vs Bn for PHE) could
induce different conformational change for the final derivatives in
their binding to the enzymes and whether they had any influence on
the potency of the compound for TG2 compared with FXIIIa.
residue. Examples of such small-size warheads are a,b-unsaturated
Table 1
amides [15e17], chloroacetamides [17], epoxides and vinyl acet-
amides [15,16] and maleimides [17,18]. We were thus intrigued to
find out if appending a bulky imidazolium moiety to our previous
peptidic inhibitors would interfere with the binding and/or
bonding with the enzyme. Although the reaction with the
CYS277 residue is mechanistically different than in the case of
thioimidazolium warheads, we were encouraged by the finding
that other bulky group such as thiadiazoles could also be accom-
modated by the binding site [19].
Inhibition results in case of imidazolium-based derivatives.
O
O
H
N
Aa
S
O
N
H
R'
O
COOH
S-R0
Aa ¼ PHE
IC50 (TG2) IC50 (FXIII)
Aa ¼ GLY
IC50 (TG2) IC50 (FXIII)
S
2. Results and discussion
1a
3
5
1b
1
2
N
N
S
Our studies started with the synthesis of the imidazolium-based
derivative 1a (Table 1). This compound inhibited TG2 with an IC50
of 3 ꢀ 1
dimethylsulfonium series [4]. When tested in Swiss 3T3 cell lines,
this inhibitor did not show any toxicity up to 250 M for 72 h,
indirectly confirming that it would successfully target only the
extracellular TG2. Encouraged by these preliminary results, we
subsequently decorated the imidazolium scaffold with various
substituents in order to evaluate the influence of these structural
modifications on TG2 inhibition.
mM, comparable with the most potent derivatives in the
2a >200
50
2b >200
>200
N
S
N
m
3a >200
4a >200
>200
>200
3b >200
4b >200
30
N
N
S
>200
N
N
2.1. Chemistry
S
As the first step in the synthesis, a set of imidazolium-thiones was
obtained by condensation of different thioureas and acetoins in
refluxing hexan-1-ol. Because the catalytic CYS277 residue was
revealed by crystallography to be situated in a small cavity (PDB
codes: 1KV3, 2Q3Z), different sizes and positions of the substituents
(Me, Et, Pr, Ph, Bn) on the imidazolium moiety were tested. The ob-
tained thiones were reacted with the bromo-methylketones syn-
thesized as previously described by our group [4] (Scheme 1). All
modifications were conducted in both the glycine and the phenyl-
alanine series. We were interested in testing if these different amino
N
N
5a >200
>200
5b >200
>200
S
N
N
6a
5
10
6b
10
20
IC50 values are reported in
ments run in triplicate for which SD ꢁ20%.
mM and represent the average of 2e3 different experi-

528
E. Badarau et al. / European Journal of Medicinal Chemistry 66 (2013) 526e530
Scheme 1. Synthesis of target inhibitors.
In order to correlate better the new imidazole-based derivatives
inhibition activity is lost. The same dramatic decrease of activity is
observed even when only one nitrogen substituent is increased in
size (1a,b vs 3a,b, 4a,b and 5a,b). The fact that the activity is
gradually lost with the increase of the size of the substituent was
confirmed for a monoethyl substituent (data not presented). The
hypothesis that extra aromatic substituents like phenyl or benzyl
(3a,b or 4a,b) could perhaps stabilise the initial ligand binding by
interaction with aromatic residues around the catalytic CYS277
residues (e.g. TRP241, TRP323 or PHE325) proved unsuccessful.
Interestingly, modifications of substituents on C4 and C5 positions
of imidazolium scaffold were tolerated, as compounds 6a,b
retained a good inhibition range.
with our previous inhibitors, analogues in the dialkylsulfonium
series were also synthesized and biologically tested. The alkyl
substituents on sulfonium warhead were varied in methyl and ethyl
groups. The newly synthesized derivatives were characterised by
infra-red spectroscopy (IR), high resolution mass spectrometry
(HRMS), proton and carbon-13 NMR spectroscopy (1H and 13C
NMR), melting points, and had properties consistent with the
proposed structures (see Supporting Information).
2.2. Inhibition activity
In the alkylsulfonium series, a similar activity trend was
observed. Small substituents favoured a better binding profile,
whilst increasing their size induced a fall of activity (7a,b vs 8a,b).
Concerning the specificity profile of these water-soluble in-
hibitors, it was discovered that the dialkylsulfonium warheads have
privileged interactions with TG2 residues, while these are not
present in the case of FXIIIa. The imidazolium warheads could not
discriminate between the two extracellular transglutaminases,
even if a certain trend could be identified in the case of 2a and 3b.
Thus, this type of compound is either equipotent for both TG2 and
FXIIIa (1a,b and 6a,b) or devoid of any inhibition trend (2a,be5a,b).
The final molecules were evaluated for their inhibitory activities
for TG2 and FXIIIa (Tables 1 and 2). Briefly, as previously described,
the enzyme assay in which the inhibitors were evaluated for their
IC50 was based on the Ca2þ-mediated incorporation of N-(5-
aminopentyl)biotinamide into N,N0-dimethylcasein by recombi-
nant human TG2 as previously described by Griffin et al. [4] (see
Supporting Information).
The first important remark concerning the IC50 results for TG2 is
that the overall inhibition trends are the same in the PHE series
(1ae6a) compared to the GLY series (1be6b). Thus, despite the
increased conformational freedom of GLY-based molecules, as
there is practically no discrimination between the activities of the
compounds with a GLY versus PHE core, we can conclude that the
binding pattern in both series is most probably the same.
Concerning the conducted structural modifications (Table 1),
one can observe that the activity is extremely sensitive to modifi-
cations of the substituents on imidazolium nitrogens. In fact, when
passing from dimethyl to diethyl substituents (1a,b vs 2a,b) the
2.3. Molecular modelling
The fact that small structural modifications induce a huge ac-
tivity change for TG2 (1a,b vs 2a,b) seems to be in accordance with
the fact that the catalytic CYS residue is situated in a narrow cavity
which has a very minor conformational change in both steps of the
inhibition process: the formation of the reversible Michaelise
Menten complex and the subsequent covalent bonding of the in-
hibitor. In fact, when the closed and the extended conformations of
TG2 are superimposed in the core region (PDB codes: 1KV3 vs
2Q3Z), little change is observed around the key CYS277 residue. The
inhibitors should thus have a strong complementarity with the
enzyme in the warhead region, as the dynamics of the induced fit
process is limited in CYS277 area, but more extensive in the adja-
cent hydrophobic and loop regions.
Table 2
Inhibition results for dialkylsulfonium derivatives.
O
O
R"
S
H
N
Aa
O
N
H
R"
O
COOH
Br
Preliminary docking studies showed that, in order to form viable
complexes with the enzyme, the inhibitors from the newly syn-
thesized imidazolium family should adopt a bent conformation. As
the irreversibility of the reaction was verified in the case of our
potent inhibitors by dilution experiments (as previously described
[4]), a distance constraint between the electrophilic carbon of the
inhibitor and the nucleophilic sulphur atom of the CYS277 residue
was imposed during the docking process as implemented in the
Gold software [20]. Compound 6a was selected as reference for our
Aa ¼ PHE
Aa ¼ GLY
IC50 (TG2) IC50 (FXIII)
S
IC50 (TG2) IC50 (FXIII)
R"
R"
7a 10
105
7b
8
>200
S
8a 20
>200
8b 20
>200
S
IC50 values are reported in
mM and represent the average of 2e3 different experi-
ments run in triplicate for which SD ꢁ20%.

E. Badarau et al. / European Journal of Medicinal Chemistry 66 (2013) 526e530
529
Fig. 2. Selected TG2 complexes with 6a (A) and 7a (B) derived from docking. The solvent accessible surface around the ligand is also represented (right). The continuous red line
represents the distance between the electrophilic carbon of the ligand and the sulphur atom from residue CYS277, while the red dotted line represents the hydrogen bond with the
residue TRP241. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
modelling study, since this derivative incorporates in its structure
more simplified analogues either in the core (cpd 6b) or in the
imidazolium region (cpds 1a and 1b). Among the different clusters
generated by docking from different binding patterns, two were
retained as they could explain the activity trends of the other
imidazolium-based compounds. Interestingly, the representative
poses for each cluster differed only in the orientation of the imi-
dazolium warheads, the rest of the peptidic backbone adopting a
similar conformation. Subsequently, both complexes were relaxed
for 4 ns by molecular dynamics simulation in explicit water at 300 K
[21,22]. The same distance between electrophilic carbon (inhibitor)
and the nucleophilic sulphur atom (CYS277) was monitored (see
Supporting Information). In one case, the distance remained
bellow an average of 3.5 Å (Fig. 3 e Complex 1), while in the other
case the ligand seems to move away from the CYS277 region (Fig. 3
e Complex 2). The first complex represents thus a plausible binding
pattern for this type of ligands (Fig. 2A). The ligand adopts a bent
conformation in order to present its electrophilic site closer to the
catalytic CYS residue. This pose can also explain why substituents
on imidazolium nitrogens would interfere with the binding, while
the substituents in C4 and C5 positions would not have any strong
impact as they point in the outer space of the enzyme.
When analyzing the molecular dynamics trajectory, we
observed that the imidazolium warhead has privileged interactions
with the TRP241 and TRP323 residues. It is important to emphasize
that these TRP residues are also conserved in the case of FXIIIa, a
fact that could explain the potency of this inhibitor also for this
transglutaminase.
The same modelling protocol was used in the case of the
dimethylsulfonium warheads (7a,b). Despite the fact that these
Fig. 3. Distances recorded during the molecular dynamics trajectory in case of 6a, for selected complexes resulting from docking. In red/upper: distance between the electrophilic
carbon (ligand) and the sulphur atom (CYS277); in blue/bottom: H-bond distance with TRP241. (For interpretation of the references to colour in this figure legend, the reader is
referred to the web version of this article.)

530
E. Badarau et al. / European Journal of Medicinal Chemistry 66 (2013) 526e530
[2] C.M. Bergamini, R.J. Collighan, Z. Wang, M. Griffin, Adv. Enzymol, John Wiley &
Sons, Inc., 2011. pp. 1e46.
[3] D.H. Pliura, B.J. Bonaventura, H.W. Pauls, J.F. Killackey, A. Krantz, Irreversible
inhibition of transglutaminases by sulfonium methylketones: optimization of
specificity and potency with omega-aminoacyl spacers, J. Enzym. Inhib. 6
(1992) 181e194.
[4] M. Griffin, A. Mongeot, R. Collighan, R.E. Saint, R.A. Jones, I.G. Coutts,
D.L. Rathbone, Synthesis of potent water-soluble tissue transglutaminase in-
hibitors, Bioorg. Med. Chem. Lett. 18 (2008) 5559e5562.
[5] T.S. Johnson, M. Fisher, J.L. Haylor, Z. Hau, N.J. Skill, R. Jones, R. Saint, I. Coutts,
M.E. Vickers, A.M. El Nahas, M. Griffin, Transglutaminase inhibition reduces
fibrosis and preserves function in experimental chronic kidney disease, J. Am.
Soc. Nephrol. 18 (2007) 3078e3088.
inhibitors also adopt a bent conformation in the warhead region
(Fig. 2B), the orientation is different and seems to favour a catione
pi interaction with the PHE325 residue. While the corresponding
residue in FXIII seems similar (TYR372 in the case of FXIIIa vs
PHE325 in the case of TG2), with the assumption that the region
around the catalytic CYS residue is also conformationally highly
preserved (as in TG2 case), the crystal structure of FXIII in its closed
form (PDB code: 1EVU) shows that the position of this key TYR372
residue is different than that of the PHE325 in case of TG2, thus
rendering difficult a similar pose of the inhibitor in the catalytic
site. This could explain the specificity of inhibitors containing the
dimethylsulfonium moieties for TG2.
[6] T. Johnson, M. Fisher, J. Haylor, Z. Hau, N. Skill, R. Jones, R. Saint, I. Coutts,
A. El Nahas, M. Griffin, Transglutaminase inhibition ameliorates tissue
scarring and fibrosis: experience in a kidney model, J. Am. Soc. 14 (2008)
2052.
[7] L. Huang, J.L. Haylor, Z. Hau, R.A. Jones, M.E. Vickers, B. Wagner, M. Griffin,
R.E. Saint, I.G. Coutts, A.M. El Nahas, T.S. Johnson, Transglutaminase inhibi-
tion ameliorates experimental diabetic nephropathy, Kidney Int. 76 (2009)
383e394.
[8] J.J. Baldwin, G. Valley, D.C. Remy, N. Wales, D.A. Claremon, Merck & Co, Inc.,
Rahway, N. J., U.S Patent 386642, 1990.
[9] K.F. Freund, K.P. Doshi, S.L. Gaul, D.A. Claremon, D.C. Remy, J.J. Baldwin,
S.M. Pitzenberger, A.M. Stern, Transglutaminase inhibition by 2-[(2-
oxopropyl)thio]imidazolium derivatives: mechanism of factor XIIIa inactiva-
tion, Biochemistry 33 (1994) 10109e10119.
[10] SPARC_v4.6, Sparc Performs automated Reasoning in Chemistry http://
archemcalc.com/sparc/.
[11] B.-G. Han, J.-W. Cho, Y.D. Cho, K.-C. Jeong, S.-Y. Kim, B.I. Lee, Crystal structure
of human transglutaminase 2 in complex with adenosine triphosphate, Int. J.
Biol. Macromol. 47 (2010) 190e195.
[12] S. Liu, R.A. Cerione, J. Clardy, Structural basis for the guanine nucleotide-
binding activity of tissue transglutaminase and its regulation of trans-
amidation activity, Proc. Natl. Acad. Sci. U S A 99 (2002) 2743e2747.
[13] D.M. Pinkas, P. Strop, A.T. Brunger, C. Khosla, Transglutaminase 2 undergoes a
large conformational change upon activation, PLoS Biol. 5 (2007) 2788e2795.
[14] I. Lindemann, A. Heine, G. Klebe, Transglutaminase 2 in complex with a novel
inhibitor (2012). PDB codes: 3S3P, 3S3S, 3S3J.
3. Conclusion
New peptidic inhibitors incorporating a new polar warhead like
imidazolium are reported. Inhibition activity of both TG2 and FXIIIa
proved extremely sensitive to modification of the substituents on
the imidazolium nitrogens. For example, passing from methyl to an
ethyl substituent induced a complete loss of activity. When corre-
lated with the crystal structure of TG2, these results suggested a
bent conformation for the inhibitors in the warhead region. A
similar bent conformation was predicted by molecular modelling in
the case of another polar warhead: dialkylsulfonium. The synthe-
sized inhibitors were tested for two members of the trans-
glutaminase family, TG2 and FXIIIa, the dimethylsulfonium
warhead rendering the inhibitors specific for TG2. Several other
inhibitors from this water-soluble peptidic series are currently
synthesized in our laboratory. A more complete set of ligands
would help finding a general binding pattern, with the final goal of
designing more potent selective TG2 inhibitors.
[15] P. de Macedo, C. Marrano, J.W. Keillor, Synthesis of dipeptide-bound epoxides
and alpha, beta-unsaturated amides as potential irreversible transglutaminase
inhibitors, Bioorg. Med. Chem. 10 (2002) 355e360.
[16] C. Marrano, P. de Macedo, J.W. Keillor, Evaluation of novel dipeptide-bound
alpha, beta-unsaturated amides and epoxides as irreversible inhibitors
of guinea pig liver transglutaminase, Bioorg. Med. Chem. 9 (2001) 1923e
1928.
[17] C. Pardin, S.M. Gillet, J.W. Keillor, Synthesis and evaluation of peptidic irre-
versible inhibitors of tissue transglutaminase, Bioorg. Med. Chem. 14 (2006)
8379e8385.
[18] D. Halim, K. Caron, J.W. Keillor, Synthesis and evaluation of peptidic mal-
eimides as transglutaminase inhibitors, Bioorg. Med. Chem. Lett. 17 (2007)
305e308.
Acknowledgements
The authors acknowledge EPSRC National Mass Spectrometry
Centre, Swansea (UK) for recording the HRMS spectra. E.B. and A.M.
are Marie Curie research fellows with the support of the following
grants (E.B.: “TRANSCOM”, FP7 No: 251506, A.M.: “TRACKS”, MRTN-
CT-2006-03032).
[19] C. Marrano, P. de Macedo, P. Gagnon, D. Lapierre, C. Gravel, J.W. Keillor,
Synthesis and evaluation of novel dipeptide-bound 1,2,4-thiadiazoles as
irreversible inhibitors of guinea pig liver transglutaminase, Bioorg. Med.
Chem. 9 (2001) 3231e3241.
[20] GOLD, version 5.1; software available from Cambridge Crystallographic Data
Centre Cambridge Crystallographic Data Centre (12 Union Road Cambridge
CB2 1EZ UK). http://www.ccdc.cam.ac.uk.
Appendix A. Supplementary data
Supplementary data related to this article can be found at
http://dx.doi.org/10.1016/j.ejmech.2013.05.018.
[21] Amber, version 11; software developed at the University of California, San
Francisco, by Case D. A. et coll. http://ambermd.org/.
References
[22] D.A. Case, T.E. Cheatham 3rd, T. Darden, H. Gohlke, R. Luo, K.M. Merz Jr.,
A. Onufriev, C. Simmerling, B. Wang, R.J. Woods, The Amber biomolecular
simulation programs, J. Comput. Chem. 26 (2005) 1668e1688.
[1] M. Griffin, R. Casadio, C.M. Bergamini, Transglutaminases: nature’s biological
glues, Biochem. J. 368 (2002) 377e396.