Interrogation of the active sites of protein arginine deiminases (PAD1, -2, and -4) using designer probes

Article abstract of DOI:10.1021/ml300377d

Protein arginine deiminases (PADs) are involved in a number of cellular pathways, and they catalyze the transformation of peptidyl arginine residue into a citrulline as part of post-translational modifications. To understand ligand preferences, a group of probe molecules were investigated against PAD1, PAD2, and PAD4. These probe molecules carried a well-known covalent modifier of the catalytic cysteine residue, 2-chloroacetamidine moiety, which was tethered to an α-amino acid via a carbon linker. The chain length for the linker varied from 0 to 4. Time-dependent assays indicated that 2-chloroacetamidine (2CA) with no linker inhibited all PAD enzymes with a similar trend in the second-order rate constants, although with poor affinity. Among the other three probe molecules, compound 3 with a three-carbon linker exhibited the best second-order rate constants for optimal ligand reactivity with the binding site. These analyses provide insights into the relative patterns of covalent inactivation of PAD isozymes and the design of novel inhibitors targeting PAD enzymes as potential therapeutic targets.

Full text of DOI:10.1021/ml300377d

Letter
pubs.acs.org/acsmedchemlett
Interrogation of the Active Sites of Protein Arginine Deiminases
(PAD1, -2, and -4) Using Designer Probes
Angelica M. Bello,†,‡,§ Ewa Wasilewski,†,‡ Lianhu Wei,†,‡,§ Mario A. Moscarello,∥
and Lakshmi P. Kotra*,†,‡,§
‡
†Center for Molecular Design and Preformulations and Toronto General Research Institute, University Health Network, Toronto,
Ontario, M5G 1L7, Canada
§Department of Pharmaceutical Sciences, Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, Ontario, M5S 3M2,
Canada
∥Department of Structural Biology and Biochemistry, The Hospital for Sick Children, Toronto, Ontario, M5G 1X8, Canada
S
* Supporting Information
ABSTRACT: Protein arginine deiminases (PADs) are involved in a number of cellular pathways, and they catalyze the
transformation of peptidyl arginine residue into a citrulline as part of post-translational modifications. To understand ligand
preferences, a group of probe molecules were investigated against PAD1, PAD2, and PAD4. These probe molecules carried a
well-known covalent modifier of the catalytic cysteine residue, 2-chloroacetamidine moiety, which was tethered to an α-amino
acid via a carbon linker. The chain length for the linker varied from 0 to 4. Time-dependent assays indicated that 2-
chloroacetamidine (2CA) with no linker inhibited all PAD enzymes with a similar trend in the second-order rate constants,
although with poor affinity. Among the other three probe molecules, compound 3 with a three-carbon linker exhibited the best
second-order rate constants for optimal ligand reactivity with the binding site. These analyses provide insights into the relative
patterns of covalent inactivation of PAD isozymes and the design of novel inhibitors targeting PAD enzymes as potential
therapeutic targets.
KEYWORDS: PAD, protein arginine deiminase, chemical probes, inhibitor design
rotein arginine deiminases (PAD) are a group of five
Penzymes, PAD1−PAD4 and PAD6, involved in the post-
translational modification of peptidyl arginine residues and are
distributed in various tissues. PAD enzymes belong to the
superfamily of amidinotransferases and are essential for a
variety of cellular processes. In recent years, arginine-related
biochemical pathways have been reported to have an important
role in several major human diseases including rheumatoid
arthritis, Alzheimer’s disease, scrapie, Creutzfeld−Jacob disease,
and multiple sclerosis.1−5 Recently, it was also shown that
PAD4 may be involved in cancer via the tumor suppressor
genes; thus, PAD inhibitors may also have potential therapeutic
benefits as anticancer agents.6 Modulation of the activities of
these enzymes, which are up-regulated in the above-mentioned
diseases, may lead to novel therapeutic strategies against these
diseases.
residues, forming a closely interacting tetrad.7 Although the
mechanism is not fully established, there is sufficient evidence
that PAD4, one of the five known human PAD isozymes,
utilizes the active site cysteine, Cys645, to catalyze the
hydrolytic deimination of arginine (Arg) residues involving a
covalent mechanism.7−11 Through a series of PAD4 mutant
studies, it was shown that the active site Cys645 is indeed
deprotonated and exists in the thiolate form prior to the
substrate binding in the catalytic site.7−11 Furthermore, it was
suggested that a properly oriented His471 side chain is required
for carrying out the acid−base chemistry for the deimination
process.
A series of PAD4 inhibitors that utilize a covalent inhibition
mechanism have been reported over the past 7 years,12,13
including N-α-benzoyl-N5-(2-chloro-1-iminoethyl)-L-ornithine
PAD hydrolyses peptidyl arginine residues to peptidyl
citrulline and ammonia (Figure 1A). In the catalytic site of
PAD, four residues appear to be critical for the deimination
reaction: one cysteine, one histidine, and two aspartate
Received: November 5, 2012
Accepted: January 15, 2013
Published: January 15, 2013
© 2013 American Chemical Society
249
dx.doi.org/10.1021/ml300377d | ACS Med. Chem. Lett. 2013, 4, 249−253

ACS Medicinal Chemistry Letters
Letter
Figure 1. (A) Hydrolysis of peptidyl Arg residue into peptidyl
citrulline by PAD enzyme. (B) Structures of haloamidines.
or Cl-amidine, with an IC50 of 5.9 0.3 μM (Figure 1B).6,14−16
Similarly, various other haloamidines (Figure 1B) were
reported as potent inhibitors and as probe molecules to
investigate PAD enzymes.13 In the above studies, PAD4
enzyme was investigated using amidine molecular probes.12
For comparative analysis with PAD1, -3, and -4, only one probe
molecule with two different halogens (X-amidine, X = Cl, F;
Figure 1B) was used to compare the relative inactivations of the
PAD enzymes, primarily through inhibitor concentrations to
inhibit 50% of enzyme activity (IC50).9
As a part our medicinal chemistry program, we were
interested in comparing the effectiveness of various inhibitors
against PAD1, PAD2, and PAD4 enzymes, due to their
therapeutic importance in neurological diseases. Here, we
report comparative reactivities of PAD1, -2, and -4 isozymes
using probe molecules designed with a comprehensive
structural analysis and provide a strong link toward medicinal
chemistry for selective inhibitors of PAD isozymes. Addition-
ally, enzyme kinetics analyses with second-order rate constants
provide a comparative tool across various isozymes and
inhibitors classes for the study of enzymes and inhibitors.
Here, we present a systematic approach, using strategically
designed chloroamidine derivatives carrying between zero and
four side chain C−C bonds, to examine the depth of the
catalytic site pocket and the varying reactivities with the PAD
enzymes. Probe molecules in the current study include 2-
chloroacetamidine (2CA) (1, Figure 2A) through chloroami-
dine structures 2−4. The shortest of these, 2CA, is a known
inhibitor of PAD enzymes, with very interesting pharmaco-
logical activities in multiple sclerosis.17,18 The chloroamidine
group is used as the reactive center because the simplest
compound 1 has already been shown to be an active covalent
inhibitor of PAD enzyme, and other derivatives with varied
chain lengths could be synthetically accessed. Thus, compounds
1−4 were designed with varied lengths of the side chains
connected to the chloroamidine moiety (Figure 2A). Three of
the four compounds carry an acyl moiety on the amino
terminus and a methyl ester on the carboxyl terminus. Because
of the differences in the length of the side chains, these
compounds are expected to enter the catalytic site of PAD with
varying degrees of access to the reactive center for covalent
bond formation. A detailed investigation in this direction would
Figure 2. (A) Chemical probe molecules targeting PAD enzymes. (B)
peptidyl Arg binding site in a tunnel shape in PAD4 (1WDA) and the
entry of the substrate indicated by the arrow in white. (C) Connolly
surface depicting the water accessible surface on 1 (2CA). (D)
Compounds 1−4 docked into the active site of PAD4 (1WDA).
Inhibitor molecules are shown as a ball-and-stick model with the
following: compound 1, red; 2, purple; 3, orange; and 4, green. PAD4
residues are shown as capped-stick models, color-coded according to
atom type.
help design inhibitors with improved precision to modulate the
activities of PAD enzymes.
Compound 1 is a simple acetamidine derivative carrying the
chloromethyl group as the reactive center but does not carry
any other structural elements such as a linker moiety or the
amino acidlike structural components (Figure 2A). Compound
1 is approximately 5.5−7.0 Å in length when one measures on
the Connolly accessible surface on this molecule (Figure 2C).
When one compares the size of 1 and the opening of the
binding site in PAD4, they are very close to conclude that 1
could easily enter the binding site channel (Figures 2B). The
remaining three probe molecules carry 2, 3, or 4 −C−C− bond
spacers between the amidine moiety and the Cα of the amino
acid. The amino acid backbone, that is, the −NH−CH−CO−
moiety, keeps the side chain anchored at the opening of the
binding pocket, in contrast to 2CA (1), which does not carry
such an anchoring moiety. Because of this anchor and their
250
dx.doi.org/10.1021/ml300377d | ACS Med. Chem. Lett. 2013, 4, 249−253

ACS Medicinal Chemistry Letters
Letter
varying side chain lengths, compounds 2−4 reach into different
depths of the binding pocket (Figure 2D). Hence, the reactive
carbon for each compound is located at different depths in the
binding pocket.
Compound 1 was used as a hydrochloride salt in enzymatic
assays. Compounds 2−4 were synthesized according to Scheme
1.15,19 Briefly, acylation of the commercially available starting
We evaluated compounds 1−4 against PAD1, PAD2, and
PAD4 probing the ligand interactions in the context of binding
site of PAD enzymes (Table 1). Compound 1 with no specific
interactions at the binding site, other than carrying the reactive
chloroimidine moiety, inhibited all three PAD enzymes in a
time-dependent and concentration-dependent manner, with
almost similar second-order rate constants with kinact/KI in the
range of 9−60 M−1 min−1.
a
There is a clear loss of PAD activity upon incubation with 1
over time, with a complete loss of activity at 5 mM
concentration against PAD4 within 30 min. Poor affinity of 1
against PAD enzymes as demonstrated by the high
concentrations required to achieve any inhibition, as well as
the lowest second-order rate constants in comparison to those
with inhibitors 2−4. This is not surprising due to the absence of
any structural features attributed to high affinity. Because of the
small size of compound 1, and possibly for its entry into the
catalytic sites of PAD enzymes with little barrier, compound 1
exhibited similar affinities and kinetics of inactivation of the
PAD enzymes (Table 1 and Figure 1B,C). Compound 2, with a
2-carbon linker, exhibited a little higher affinity than that seen
with 1 against PAD4 but failed to provide complete loss of
activity in a time-dependent fashion against PAD4. Compound
3, on the other hand, showed higher potency as well as faster
inactivation of PAD4 than 1 and 2. In fact, PAD4 lost all its
activity within 3 min when incubated with 3 at 2 mM
concentration and within 30 min at 100 μM concentration of 3
(kinact/KI = 865 55 M−1 min−1). As expected, compound 3
exhibited higher second-order rates of inactivation against
PAD1 and PAD2 as well (kinact/KI = 497 21 and 167 μM−1
min−1, respectively).
Scheme 1. Synthesis of Compounds 2−4
Compound 4, with the longest linkera four-carbon linker
between the chloroamidine moiety and the amino acid
backboneexhibited a decreased rate of inactivation of
PAD1 and PAD4 (kinact/KI = 62 2 and 50 3 M−1 min−1,
respectively) when compared to those with 3. Interestingly,
when the inhibition profiles by 2 and 3 against PAD enzymes
are compared, the general trend indicated that a 3-carbon chain
as in 3 is optimal for the chloroamidine moiety to inactivate the
enzyme in a time-dependent fashion, presumably through the
participation of the catalytic Cys moiety. Compounds 2−4
carry simple groups on the carboxyl and amino ends, viz. a
methyl ester and an acyl moiety, respectively. These
substitutions do not present strong interactions at the binding
site but do anchor the ligands to the opening of the active site
channel. When compounds 2−4 were docked into the binding
site of PAD4, compound 3 showed the optimal orientation and
positioning for a potential reaction with the Cys645 residue, in
comparison to that with 2 or 4 (indicated by an arrow, Figure
2D).21
a
Reagents and conditions: (i) 2 M HCl/ether, EtOH. (ii) AcCl, Et3N,
DCM, rt, 5 h. (iii) TFA, DCM, rt. (iv) Compound 6, Et3N, EtOH, rt,
16 h.
materials 7 and 8 afforded the derivatives 9 and 10, respectively
(Scheme 1B). These two served as precursors for the final
target compounds carrying the side chains with n = 1 and n = 2,
respectively. The carbamate moiety was removed by treatment
with TFA, yielding 11 and 12, respectively. Compound 13, a
lysine analogue purchased from commercial sources, along with
11 and 12 were transformed into corresponding chloroamidine
derivatives 2−4 by treatment with the imidoate derivative 6.
Compound 6 was synthesized from chloroacetonitrile 5
(Scheme 1A).15,20 Compounds 2−4 were obtained as the
corresponding TFA salts, and their purity was confirmed using
rigorous analytical protocols (see the Supporting Information).
Altogether, compounds 1−4 provided the probe molecules to
interrogate the PAD enzymes’ activities.
This trend is also reflected when the efficiencies of
inactivation are compared for each inhibitor, against each
a
Table 1. Enzyme Inhibition Kinetics for Compounds 1−4 Using Human PAD1, Mouse PAD2, and Human PAD4
kinact/KI (M−1 min−1
)
1
2
3
4
Hs PAD1
Mm PAD2
Hs PAD4
32 1 (1)
111 9 (3.5)
50 3 (5.5)
183 23 (3.6)
497 21 (15.5)
167 17 (18.5)
865 55 (17.3)
62 2 (1.9)
120 14 (13.3)
50 3 (1)
9
0.2 (1)
60 4 (1.2)
a
Hs, Homo sapiens; Mm, Mus musculus. Numbers in parentheses indicate relative rates of efficiency of inhibition of each PAD isozyme due to
inhibitors 1−4. Compound 3 with an optimal side chain length exhibited the best second-order rate constant for all three PAD enzymes.
251
dx.doi.org/10.1021/ml300377d | ACS Med. Chem. Lett. 2013, 4, 249−253

ACS Medicinal Chemistry Letters
Letter
in Pathogenesis of Sporadic Creutzfeld-Jacob Disease. Acta Neuro-
pathol. 2009, 119, 199−210.
isozyme (Table 1, numbers in parentheses). Compound 3
inactivates each isozyme, specifically PAD4, 15−18 times faster
than the corresponding poorest inhibitor, 1. In fact, compound
1 is almost the least reactive compound against all three
isozymes tested, and there appears to be no selectivity either.
The Connolly surface for compound 1 (Figure 2C) indicates
that this compound is small enough to enter the active sites of
the PAD enzymes (Figure 2B) and equally be accessible to the
thiolate in the active site for covalent bond formation. If there
were no reactive moieties such as a chloroamidine group
present on the inhibitors, high-affinity inhibitors could be
designed to take advantage of the interactions inside the PAD
active site, and the linker may not be a concern for optimal
reactivity.
In summary, this report highlights the architecture of the
active site of PAD enzyme through the reactive ligands and
their inactivation rates of these enzymes for direct comparison
with other inhibitors and enzymes families. Further systematic
interrogation will provide tools to design nonreactive chemical
probes as potential inhibitors of PAD enzymes with therapeutic
potential.
(5) Moscarello, M. A.; Wood, D. D.; Ackerley, C.; Boulias, C. Meylin
in Multiple Sclerosis is Developmentally Immature. J. Clin. Invest.
1994, 94, 146−154.
(6) Wang, Y.; Li, P.; Wang, S.; Hu, J.; Chen, X. A.; Wu, J.; Fisher, M.;
Oshaben, K.; Zhao, N.; Gu, Y.; Wang, D.; Chen, G.; Wang, Y.
Anticancer Peptidylarginine Deiminase (PAD) Inhibitors Regulate the
Autophagy Flux and the Mammalian Target of Rapamycin Complex 1
Activity. J. Biol. Chem. 2012, 287, 25941−25953.
(7) Arita, K.; Hashimoto, H.; Shimizu, T.; Nakashima, K.; Yamada,
M.; Sato, M. Structural basis for Ca(2+)-induced Activation of Human
PAD4. Nature Struct. Mol. Biol. 2004, 11, 777−783.
(8) Kearney, P. L.; Bathia, M.; Luo, Y.; Glascock, M. C.; Catchings,
K. L.; Yamada, M.; Thompson, P. R. Kinetic Characterization of
Protein Arginine Deiminase 4: A Transcriptional Corepressor
Implicated in the Onset and Progression of Rheumatoid Arthritis.
Biochemistry 2005, 44, 10570−10582.
(9) Knuckley, B.; Bhatia, M.; Thompson, P. R. Protein Arginine
Deiminase 4: Evidence for a Reverse Protonation Mechanism.
Biochemistry 2007, 46, 6578−6587.
(10) Ke, Z.; Zhou, Y.; Hu, P.; Wang, S.; Xie, D.; Zhang, Y. Active site
cysteine is protonated in the PAD4Michaelis complex: evidence from
Born-Oppenheimer ab initio QM/MM molecular dynamics simu-
lations. J. Phys. Chem. B 2009, 113, 12750−12758.
(11) Ke, Z.; Wang, S.; Xie, D.; Zhang, Y. Born-Oppenheimer ab initio
AM/MM molecular dynamics simulations of the hydrolysis reaction
catalyzed by protein arginine deiminase 4. J. Phys. Chem. B 2009, 113,
16705−16710.
(12) Luo, Y.; Knuckley, B.; Lee, Y. H.; Stallcuo, M. R.; Thompson, P.
R. A Fluoroacetamidine-Based Inactivator of Protein Arginine
Deiminase 4: Design, Synthesis, and in Vitro and in Vivo Evaluation.
J. Am. Chem. Soc. 2006, 128, 1092−1093.
(13) Luo, Y.; Arita, K.; Bhatia, M.; Knucley, B.; Lee, Y. H.; Stallcup,
M. R.; Sato, M.; Thompson, P. Inhibitors and Inactivators of Protein
Arginine Deiminase 4: Functional and Structural Characterization.
Biochemistry 2006, 45, 11727−11736.
(14) Knuckley, B.; Causey, C. P.; Jones, J. E.; Bhatia, M.; Dreyton, C.
J.; Osborne, T. C.; Takahara, H.; Thompson, P. R. Substrate
Specificity and Kinetic Studies of PADs 1, 3 and 4 Identify Potent
and Selective Inhibitors of Protein Arginine Deiminiase 3. Biochemistry
2010, 49, 4852−4863.
(15) Causey, C. P.; Jones, J. E.; Slack, J. L.; Kamei, D.; Jones, L. E.,
Jr.; Subramanian, V.; Knucley, B.; Ebrahimi, P.; Chumanevic, A. A.;
Luo, Y.; Hashimoto, H.; Sato, M.; Hofseth, L. J.; Thompson, P. R. The
Development of N-α-(2-Carboxyl)benzoyl-N5-(2-fluoro-1-iminoeth-
yl)-L-ornithine Amide (o-F-amidine) and N-α-(2-Carboxyl)benzoyl-
N5-(2-chloro-1-iminoethyl)-L-ornithine Amide (o-Cl-amidine) As
Second Generation Protein Arginine Deiminase (PAD) Inhibitors. J.
Med. Chem. 2011, 54, 6919−6935.
ASSOCIATED CONTENT
■
S
* Supporting Information
Compounds purity data, mass spectral data, and enzyme
kinetics. This material is available free of charge via the Internet
at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Tel: 416-581-7601. Fax: 416-581-7621. E-mail: lkotra@
uhnres.utoronto.ca.
Author Contributions
The manuscript was written through contributions of all
authors.
Funding
We acknowledge the financial support of MaRS Innovations,
University Health Network, and Hospital for Sick Children.
Notes
The authors declare no competing financial interest.
ABBREVIATIONS
■
PAD, protein arginine deiminase; 2CA, 2-chloroacetamidine;
Arg, arginine
(16) Jones, J. E.; Slack, J. L.; Fang, P.; Zhang, X.; Subramanian, V.;
Causey, C. P.; Coonrod, S. A.; Guo, M.; Thompson, P. Synthesis and
Screening of a Haloacetamidine Containing Library to Identify PAD4
Selective Inhibitors. ACS Chem. Biol. 2012, 7, 160−165.
(17) Stone, E. M.; Schaller, T. H.; Bianchi, H.; Person, M. D.; Fast,
W. Inactivation of two diverse enzymes in the amidinotransferase
superfamily by 2-chloroacetamidine: dimethylarginase and peptidyl
arginine deiminase. Biochemistry 2005, 44, 13744−13752.
(18) Moscarello, M. A.; Lei, H.; Mastronardi, F. G.; Winer, S.; Tsui,
H.; Li, Z.; Ackerley, C.; Zhang, L.; Raijmakers, R.; Wood, D. D.
Inhibition of peptidyl-arginine deiminase reverses protein-hyper-
citrullination and disease in mouse models of multiple sclerosis. Dis.
Models Mech. 2013, in press.
(19) Causey, C. P.; Thompson, P. R. An improved synthesis of
haloacetamidine-based inactivators of protein-arginine deiminase-4
(PAD4). Tetrahedron Lett. 2008, 49, 4383−4385.
(20) Stillings, M. R.; et al. Substituted 1,3,4-thiadiazoles with
anticonvulsant activity. 2. Aminoalkyl derivatives. J. Med. Chem. 1986,
29, 2280−2284.
REFERENCES
■
(1) Vossenaur, E. R.; Zendman, A. J.; van Venrooji, W. J.; Pruijin, G.
J.; PAD, A. Growing Family of Citrullinating Enzymes: Genes,
Features and Involvement in Disease. Bioessays 2003, 25, 1106−1115.
(2) Ishigami, A.; Oshawa, T.; Hiratsuka, M.; Taguchi, H.; Kobayashi,
S.; Saito, Y.; Murayama, S.; Asaga, H.; Kimura, N.; Maruyama, N.
Abnormal Accumulation of Citrullinated Proteins Catalyzed by
Peptidylarginine Deiminase in Hippocampal Extracts from Patients
with Alzheimer’s Disease. J. Neurosci. Res. 2005, 80, 120−128.
(3) Jang, B.; Kim, E.; Choi, J. K.; Jin, J. K.; Kim, J. I.; Ishigami, A.;
Maruyama, N.; Carp, R. I.; Kim, Y. S.; Choi, E. K. Accumulation of
Citrullinated Proteins by Up-Regulated Peptidylarginine Deiminase 2
in Brains of Scrapie Infected Mice. A Possible Role in Pathogenesis.
Am. J. Pathol. 2008, 173, 1129−1142.
(4) Jang, B.; Jin, J. K.; Jeong, Y. C.; Cho, H. J.; Ishigami, A.; Choi, K.
C.; Carp, R. I.; Maruyama, N.; Kim, Y. S.; Choi, E. K. Involvement of
Peptidylarginine Deiminase-Mediated Post-translational Citrullination
252
dx.doi.org/10.1021/ml300377d | ACS Med. Chem. Lett. 2013, 4, 249−253

ACS Medicinal Chemistry Letters
Letter
(21) Models were generated from the X-ray crystal structure of
PAD4 complexed with benzoyl-L-arginine amide, and the inhibitors 2−
4 were overlapped with the ligand. The active site Cys645 was
regenerated from Ala645 PAD4 inactive mutant used for crystal-
lization.
253
dx.doi.org/10.1021/ml300377d | ACS Med. Chem. Lett. 2013, 4, 249−253