Rational Design of Calpain Inhibitors Based on Calpastatin Peptidomimetics

Article abstract of DOI:10.1021/acs.jmedchem.6b00267

Our previously reported structures of calpain bound to its endogenous inhibitor calpastatin have motivated the use of aziridine aldehyde-mediated peptide macrocyclization toward the design of cyclic peptides and peptidomimetics as calpain inhibitors. Inspired by nature's hint that a β-turn loop within calpastatin forms a broad interaction around calpain's active site cysteine, we have constructed and tested a library of 45 peptidic compounds based on this loop sequence. Four molecules have shown reproducibly low micromolar inhibition of calpain-2. Further systematic sequence changes led to the development of probes that displayed increased potency and specificity of inhibition against calpain over other cysteine proteases. Calculated K_i values were in the low micromolar range, rivaling other peptidomimetic calpain inhibitors and presenting an improved selectivity profile against other therapeutically relevant proteases. Competitive and mixed inhibition against calpain-2 was observed, and an allosteric inhibition site on the enzyme was identified for a noncompetitive inhibitor.

Full text of DOI:10.1021/acs.jmedchem.6b00267

Subscriber access provided by Nanyang Technological Univ
Article
Rational design of calpain inhibitors based on calpastatin peptidomimetics
Kristin E. Low, Spencer Ler, Kevin Chen, Robert Laurence Campbell, Jennifer L. Hickey,
Joanne Tan, Conor C. G. Scully, Peter L. Davies, Andrei K Yudin, and Serge Zaretsky
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00267 • Publication Date (Web): 05 May 2016
Downloaded from http://pubs.acs.org on May 7, 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 49
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
Rational design of calpain inhibitors based on
calpastatin peptidomimetics
Kristin E. Low,1 Spencer Ler,2 Kevin Chen,1 Robert L. Campbell,1 Jennifer L. Hickey,3 Joanne
Tan,2 Conor C. G. Scully,2 Peter L. Davies,1* Andrei K. Yudin,2* and Serge Zaretsky2*
1Department of Biomedical and Molecular Sciences, Queen's University, Kingston, ON, K7L
3N6, Canada
2Department of Chemistry, University of Toronto, Toronto, ON, M5S 3H6, Canada
3Encycle Therapeutics Inc., 101 College Street, Suite 314, Toronto, Ontario M5G 1L7, Canada
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 49
1
2
3
4
Abstract
5
6
7
8
9
Our previously reported structures of calpain bound to its endogenous inhibitor calpastatin have
motivated the use of aziridine aldehyde-mediated peptide macrocyclization towards the design of
cyclic peptides and peptidomimetics as calpain inhibitors. Inspired by nature‘s hint that a ꢀ-turn
loop within calpastatin forms a broad interaction around calpain‘s active site cysteine, we have
constructed and tested a library of 45 peptidic compounds based on this loop sequence. Four
molecules have shown reproducibly low micromolar inhibition of calpain-2. Further systematic
sequence changes led to the development of probes that displayed increased potency and
specificity of inhibition against calpain over other cysteine proteases. Calculated Ki values were
in the low ꢁM range, rivaling other peptidomimetic calpain inhibitors and presenting an
improved selectivity profile against other therapeutically relevant proteases. Competitive and
mixed inhibition against calpain-2 was observed and an allosteric inhibition site on the enzyme
was identified for a non-competitive inhibitor.
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
Keywords
cyclic peptide; calpain; calcium-activated protease; inhibitor; enzyme kinetics; cysteine protease;
macrocycles
Introduction
Calpains are a family of Ca2+-activated cytosolic multi-domain cysteine proteases that are
responsible for specific, limited cleavage of a wide variety of proteolytic targets.1 They act on
2
ACS Paragon Plus Environment

Page 3 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Ca2+ signals to bring about cellular responses such as reorganization of the cytoskeleton, cell
cycle regulation, and apoptosis.2-4 Over-activation, either by specific mutations or upstream
effects, cause numerous pathological conditions in heart attack, stroke, Alzheimer‘s disease,
cancer, neurodegeneration, muscular dystrophy, and traumatic brain injury.4-8
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
The most extensively studied isoforms, calpains-1 and -2, are heterodimers composed of
an isoform-specific 80-kDa large subunit and a common 28-kDa small subunit.1,3,5,9-13 Calpain-1
requires multi-µM concentrations of Ca2+, while calpain-2 requires an order of magnitude higher
level in the sub-mM range. 14-20 Activation by Ca2+ induces conformational changes in the
enzyme, aligning the domains of the protease core for active site cleft formation and resulting in
an overall more compact enzyme.
Calpains-1 and -2 have a highly specific natural inhibitor, calpastatin, which is thought to
limit their proteolytic activity within the cell. Calpastatin binds the enzyme only when the latter
is in the Ca2+-activated form and is released from the enzyme at low concentrations of Ca2+.21,22
Calpastatin is an intrinsically unstructured 70-kDa protein comprised of five domains, four of
which (CAST 1 through 4) are homologous and independently inhibitory and each contain
subdomains A, B, and C.2,23 Crystal structures of Ca2+-bound calpain inhibited by calpastatin
have revealed the mode of inhibition.15,17,24 Subdomains A and C, which help anchor the inhibitor
to the enzyme (Figure 1), form amphipathic .-helices upon binding to hydrophobic clefts in the
two penta-EF hand (PEF) domains. Subdomain B is responsible for inhibition of calpain by
binding and occluding the primed and unprimed sides of the active site cleft formed by the
protease core domains. Subdomain B avoids cleavage by looping out and away from the active
site cysteine with a ꢀ-turn at the conserved Lys611-Leu612-Gly613-Glu614. It also forms a two-turn
amphipathic .-helix that binds on the primed side of the cleft.
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 49
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
ACS Paragon Plus Environment

Page 5 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
highly potent with moderate selectivity, but lack therapeutic potential due to their low aqueous
solubility.28 In their work on ketoamide inhibitors of calpain, AbbVie noted pharmacokinetic
issues in development due to poor cellular stability.29,30 Moreover, as the physiological targets of
calpain cleavage remain largely undefined, calpain-selective inhibitors are sought for their use as
in vitro tools for target identification in addition to use against aforementioned calpain-associated
pathologies. Calpain isoform specificity is also desired in cases of isoform-specific pathologies.
Currently, most compounds show marginal selectivity between calpains-1 and -2, which can
prevent progression into clinic trials of such compounds, such as seen with N-(4-
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
fluorophenylsulfonyl)-l-valyl-l-leucinal, SJA-6017.31,32
The recent crystal structures of Ca2+-activated calpain have allowed for the structure-
based design of inhibitors along the entire active site cleft, including those regions of potential
interaction with the calpain-specific C2-like (C2L) domain.15,17 This had not been possible
before, as the only structures available of calpain in the active conformation were those of the
uninhibited and inhibited protease core (PC) (domains PC1 + PC2).14,16,18-20
Proteases universally bind to the extended ꢀ-strand conformation of their substrates.33
This insight provides a possible explanation for the superior inhibition achieved by many
peptidomimetic molecules with conformational restrictions that favour a ꢀ-strand-like
formation.24,34 Similar trends have been noted for calpain inhibition.26 Since recognition
sequences are often 4-5 residues in length, one successful approach to inhibitor synthesis is to
lock the sequence into a specific conformation within a cyclic peptide. Cyclization can result in a
reduced entropic penalty for binding to a protein host, thereby making it a favourable
interaction.35,36 In addition, cyclic peptides exhibit greater stability to proteolytic degradation by
peptidases in the cell.37 Cyclic peptides have previously been used to successfully inhibit HIV
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 49
1
2
3
4
5
6
7
8
9
protease,21 HCV NS3 protease,23 and even calpain (Figure 2).24 To further capitalize on the
cyclization ꢀ-strand locking mechanism, we noted that calpastatin has a conserved ꢀ-turn where
it binds to the active site of calpain.
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
We hypothesized that by using the calpain-specific inhibition characteristics of
calpastatin, we would be able to design novel, potent calpain inhibitors. Specifically, inhibitor
design based on the calpastatin ꢀ-turn residues KLGE would propagate calpastatin-like
specificity and potency to such inhibitors. Designing inhibitors in this manner without a chemical
warhead would allow to avoid the high potential for toxicity found with compounds containing
reactive centers, thus potentially improving clinical use.26 Based on prior work with cyclic
peptides derived from aziridine aldehyde-mediated peptide macrocyclization, we expected 18-
membered rings to adopt rigid ꢀ -turn structures. Additionally, we were interested in comparing
the structure-activity effects of constraining the KLGE-type sequences within cyclic peptides to
amidine-based cycle-tail motif peptidomimetics that are also derived from the aziridine
aldehyde-mediated macrocyclization chemistry (Figure 2). Here, we show that cyclic peptide and
amidine-based cycle-tail motif peptidomimetics inhibit cysteine proteases, including calpain, and
that both show various modes of inhibition of calpain-2 with Ki values in the low micromolar
range. In addition, photo-reactive cross-linking has been used to identify inhibitor binding sites
for two of the library compounds.
6
ACS Paragon Plus Environment

Page 7 of 49
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 2. Comparison of Abell‘s cyclic peptide inhibitors of calpain24 to compounds from
this study, which were derived from the multicomponent synthesis of peptides, aziridine
aldehydes, and isocyanides
Experimental Section
Materials. Active wild-type and inactive C105S mutant rat recombinant calpain-2,38 and rat
calpain-1 protease core14 were expressed and purified as previously described. Rat calpain
enzymes were used over human-sourced calpains based on the availabilities of recombinant
expression techniques for high quantities of protein, and the crystal structures of rat full-length
enzyme and protease core construct. Papain, human cathepsin L, the cysteine protease substrate
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 49
1
2
3
4
5
6
Z-FR-AMC, and leupeptin were purchased from Calbiochem. The calpain substrate (EDANS)-
EPLFAERK-(DABCYL) was synthesized by Biomer Technology.
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
Cyclic peptide and peptidomimetic synthesis. Peptides were synthesized using the telescopic
cyclization/aziridine ring-opening synthesis.39 Unless otherwise noted, a Pro-terminated peptide
was cyclized at a 0.10 mmol scale in a 2-dram vial. To it, was added aziridine aldehyde dimer
(0.10 mmol, 2.0 eq. as monomer), TFE (1.0 mL, 0.1 M), and tert-butyl isocyanide (0.2 mmol, 2.0
eq.). The reaction was left to stir at room temperature for 4 h and monitored throughout by RP-
HPLC/MS. Ring-opening of the aziridine moiety was initiated by addition of thiobenzoic acid
(0.4 mmol, 4.0 eq) and left to stir for 2 more hours. Afterwards, Raney®-Nickel slurry (approx.
1.5 mL) was added, reaction capped, and allowed to stir for a final 16 h. The mixture was filtered
through Celite with methanol, and concentrated under reduced pressure. TFA/DCM (1:1, 4 mL)
was then added to the mixture to deprotect the side chains. The cleavage cocktail was evaporated
with a stream of nitrogen after 2 h of stirring at room temperature. The TFA salt was then
purified by semi-preparative RP-HPLC/MS or reversed-phase flash chromatography. The pure
fractions were pooled together and lyophilized to afford the final products as white or off-white
powders. Purity was confirmed to be ≥ 95% by HPLC/MS and products were identified by NMR
and HPLC/MS (see the Supporting Information).
RP-HPLC/MS. Low resolution mass spectra (ESI) were collected on an Agilent Technologies
1200 series HPLC paired to a 6130 Mass Spectrometer. Compounds were resolved on an Agilent
Poroshell 120 EC-C18, 2.7 µm, 4.6 x 50 mm2 column at room temperature with a flow of 1
mL/min. The gradient consisted of eluents A (0.1% formic acid in double-distilled water) and B
8
ACS Paragon Plus Environment

Page 9 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(0.1% formic acid in HPLC-grade acetonitrile). The gradient method started with a step of 5% of
B for the first 0.99 min, followed by a linear gradient from 5% to 95% B in 8.0 min. The column
was then washed with 95 % B for 1.0 min and equilibrated at 5% B for 1.5 min.
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
Peptide NMR. 1H and 13C NMR spectra were recorded on Varian Mercury 400 and Agilent 500
MHz or 600 MHz spectrometers. 1H NMR spectra were referenced to CDCl3 (7.26 ppm),
CD3OD (3.30 ppm), DMSO-d6 (2.50 ppm). 13C NMR spectra were referenced to CDCl3 (77.2
ppm), CD3OD (49.0 ppm), and DMSO-d6 (39.52 ppm). NMR spectra were recorded at 25 °C
unless otherwise specified. Peak multiplicities are designated by the following abbreviations: s,
singlet; bs, broad singlet; d, doublet; t, triplet; q, quartet; m, multiplet; ds, doublet of singlets; dd,
doublet of doublets; ddd, doublet of doublet of doublets; bt, broad triplet; td, triplet of doublets;
tdd, triplet of doublets of doublets.
Modelling, docking simulation, and molecular dynamics simulation studies. Initial designs for
the cyclic peptides were modelled using PyMOL based on the crystallographic structure of
calpastatin-bound calpain-2 (PDB: 3BOW).15 SBGrid Consortium-supported software
applications were used.40 Molecular dynamics (MD) calculations were done with the
GROMACS package41 on calpain-2. The protein was solvated in a box using 30365 water
molecules. The net charge of the system was neutralized with 18 sodium ions. The system was
then subjected to energy minimization, position-restrained molecular dynamics to settle the water
molecules, and unrestrained molecular dynamics simulations. The simulations were run at a
temperature of 273 K using a time step of 2 fs for a total duration of 20 ns. Long-range
electrostatics were treated with the particle mesh Ewald method. The simulations were conducted
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 49
1
2
3
4
5
6
7
8
9
under isothermal conditions using V-rescale temperature coupling. MD studies with calpain-2
used the CHARMM42 and OPLS43 force fields for protein and the SPCE water model. MD
calculations were also done on select cyclic peptides. Topology files and parameters for the
cyclic peptides were provided by SwissParam.44 MD studies with cyclic peptides were run in a
similar method to those of calpain-2 and used the CHARMM force field.42 Results were
analyzed using VMD.45 Docking simulations were done using AutoDock4 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
AutoDockTools4,46 and Glide47-49 in the Maestro Schrödinger Software Suite for docking the
cyclic peptides to the active site cleft of calpain-2, calpain-1 protease core, and papain. The
protein receptors were used in rigid docking mode. For the cyclic ligands, residue side chains
maintained flexibility while the peptide backbone was torsionally constrained based on NMR
structures. The AutoDock4 dockings were performed using the Lamarckian genetic algorithm
with parameters as follows: 100 dockings, population size of 150, random starting position and
conformation, translation steps of 2.0 @, rotation steps of 50°, mutation rate of 0.02, crossover
rate of 0.8, local search frequency of 0.06, and 2.5 million energy evaluations. Final docked
conformations were clustered using a root-mean-squared deviation (RMSD) of 2.0 @. Glide
dockings were performed using default settings with some exceptions: Glide-XP (extra
precision) option was selected, torsional constraints were set for cycle backbone and peptide
bonds starting from the compound‘s NMR structure, and at most 100 poses per ligand were
produced. Glide docking results were analyzed by Prime MM-GB/SA50,51 to determine the free
energies of binding.
Peptide inhibitor library screening and calpain inhibition kinetics using a fluorescence-based
hydrolysis assay. The rate of cleavage of the (EDANS)-EPLFAERK-(DABCYL) fluorescent
10
ACS Paragon Plus Environment

Page 11 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
substrate was monitored in a 0.5-mL cuvette using a Perkin-Elmer LS55 fluorescence
spectrometer with ꢂex = 335 nm, ꢂem = 500 nm and slit widths of 10 nm. Triplicate readings were
obtained at 0.1 s intervals. The cuvette contained 100 nM rat calpain-1 protease core or 10 nM
rat calpain-2, 5 µM substrate, and 1-100 µM peptide inhibitor (dissolved in DMSO at stock
concentrations of 50 mM) in 10 mM HEPES (pH 7.4) and 10 mM DTT. Control assays were
performed with an equivalent amount of DMSO. The total concentration of DMSO in each
reaction was 0.4%. The reaction was initiated with the addition of CaCl2 to 4 mM in a final
volume of 0.5 mL.
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
Inhibition of papain and cathepsin L in a fluorescence-based hydrolysis assay. For papain
activity assays, the rate of cleavage of the (EDANS)-EPLFAERK-(DABCYL) fluorescent
substrate was measured in a similar manner as the calpain activity assays. For human cathepsin L
activity assays, the rate of cleavage of the Z-FR-AMC fluorescent substrate was quantified again
in a similar manner but with ꢂex = 360 nm, ꢂem = 460 nm and slit widths of 8 nm. Triplicate
readings were obtained at 0.1 s intervals. The cuvette contained 5 µM (EDANS)-EPLFAERK-
(DABCYL) or 5 µM Z-FR-AMC, and 100 µM peptide inhibitor (dissolved in DMSO at stock
concentrations of 50 mM) in 10 mM HEPES (pH 7.4) and 10 mM DTT (papain) or 0.1 M
NaOAc-HCl (pH 5.5), 1 mM EDTA, and 1 mM DTT (cathepsin L). Control assays were
performed with an equivalent amount of DMSO. The total concentration of DMSO in each
reaction was 0.4%. The reactions were initiated by the addition of 30 nM papain or 1 nM
cathepsin L, respectively. The initial rates were determined according to the procedure used for
the calpain activity assay. All reactions were performed at 298 K.
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 49
1
2
3
4
5
6
7
8
9
Analysis of enzyme kinetics with lead compounds. Fluorogenic hydrolysis assays were performed
with concentrations of substrate ranging from 1-100 µM and inhibitor concentrations from 0-50
µM. The initial rates were determined by fitting a straight line to the linear portion of the
progress curves (first 10 s after Ca2+ addition) (Figure S1). The resulting rate data were fitted to
Michaelis-Menten curves for the visualization of enzyme inhibition. Ki, Km, and Vmax values for
the reactions were obtained from simultaneously fitting the uninhibited and inhibited data to the
full mixed enzyme inhibition equation below,
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
àÔë
> ?
5
> ?
+
s E
s E
s E
-
Ü6
R L
> ?
+
-
Ü5
> ?
E 5
-
à
> ?
+
-
Ü6
where Ki1 is the dissociation constant for inhibitor from free enzyme, and Ki2 is the dissociation
constant for inhibitor from the ES complex.
For those inhibitors that appeared to follow competitive inhibition kinetics, the data was fit to the
competitive inhibition equation below,
> ?
8
5
àÔë
R L
> ?
+
> ?
p E 5
-à ls E
-
Ü5
Simultaneous non-linear regression fitting was done by a similar method as Kakkar et al.52 using
the non-linear least squares fitting function of gnuplot.53 All reactions were corrected for the
inner filter effect as previously described54 by measuring the fluorescence of increasing amounts
12
ACS Paragon Plus Environment

Page 13 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
of substrate with a fixed concentration of free EDANS fluorophore and thereby calculating a
correction factor.
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
Identifying the inhibitor binding site by photo-reactive cross-linking. Peptides were synthesized
using the thioester deacylation methodology to establish a thiol group on the peptide for
downstream conjugation.55 The reactions proceeded by first exposing the peptides to cyclization
conditions and then ring-opening with a thioacid. DTT and K2CO3 were used for deacylation of
the resulting thioester. The thiol-containing peptide was conjugated to benzophenone-4-
maleimide in DMF, deprotected under acidic conditions, and purified by reversed-phase flash
chromatography. For more details, see Supporting Information. A control inhibition study was
performed to ensure the photoreactive probes remained active. The photo-reactive probes were
cross-linked to calpain using similar methods to those of O‘Neil, et al.56,57 Benzophenone-
containing compounds am*[PGLo] and c*[PGALK] were each incubated with calpain at 5 and
10 molar excess (25 and 50 ꢁM, respectively). A control reaction was carried out with equivalent
amounts of DMSO. Inactive C105S mutant calpain-2 was present at 0.5 mg/mL (5 ꢁM). The
reaction took place in a solution of 10 mM HEPES pH 7.4, 10 mM DTT, and 4 mM CaCl2,
placed in 96-well Corning polystyrene plates, and allowed to proceed under long-wave UV light
for 2.5 h. The photo-adduct products were then submitted to the Protein Function Discovery
Facility at Queen‘s University for tryptic digestion and MALDI-TOF mass fingerprinting (see
Supporting Information).
Results and Discussion
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 49
1
2
3
4
5
6
Peptide macrocyclization results in the formation of calpastatin-like ꢀ-turns
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
Building on the initial work of peptide macrocyclization with aziridine aldehyde
dimers,58 we recently observed a propensity for the cyclic products to form ꢀ-turns.39 In the
cyclization process, three additional atoms are introduced into the peptide backbone. This
cyclization linker is comprised of an exocyclic amide and an aziridine moiety, the latter of which
can react with a nucleophile (Figure 3A). The presence of the carbonyl group in close proximity
to the peptide backbone allows the peptide to adopt a multitude of conformations with
transannular hydrogen bonds. Peptide backbones of pentamers cyclized with the aziridine
aldehyde methodology have been found to adopt ꢀ-turn structures in both solution phase (NMR)
and solid phase (X-ray).39 In the solid phase structure of c*[PGLGF], a type II ꢀ-turn centered on
the Gly-Leu residues was observed, while in the solution phase, the turn was indexed by one
residue, such that the turn was centered on the Pro-Gly residues and a type I ꢀ-turn was formed
(Figure 3B).
14
ACS Paragon Plus Environment

Page 15 of 49
Journal of Medicinal Chemistry
1
2
3
A
4
5
6
7
PR1R2R3R4
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
c*[ PR1R2R3R4]
c[PR1R2R3R4]
B
c*[PGLGF] X-
ray crystal
c*[PGLGF] NMR
structure (DMSO)
Figure 3. Macrocyclization of peptides to mimic natural calpastatin ꢀ-turn motifs. (A)
Aziridine aldehyde-based macrocyclization provides access to ꢀ-turn motifs in peptides. The
macrocycles are denoted with c*[PR1R2R3R4], where PR1R2R3R4 corresponds to the linear
peptide pentamer precursor. Amino acids of the linear peptide precursor are shown in black,
while the ring closure-associated chemical groups are highlighted in red. (B) Cyclic peptide
c*[PGLGF], derived from linear peptide PGLGF, arranges into a distinct ꢀ-turn when observed
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 49
1
2
3
4
by X-ray in the solid state versus the solution phase structure in DMSO. c*[PGLGF] is shown in
5
6
stick representation with oxygen atoms coloured red and nitrogen atoms coloured blue.
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
Aziridine aldehyde-mediated peptide cyclization was performed at 0.05 œ 0.2 M
concentration without appreciable amounts of cyclodimerization and polymerization by-products
formed, which is a distinct improvement over conventional peptide cyclization methods.55 A
competing intramolecular reaction pathway led to peptidomimetics bearing an amidine
functionality (Figure 4).59 The amidine products are a novel structural class bearing a cycle-tail
motif with a piperazinone-like core and linear peptide chain. The cyclization reaction favours
cyclic peptide (Figure 3) products when pentamers formed from all L-amino acids are cyclized,
and forms the amidine products predominantly with smaller linear peptides substrates or
pentamers with varied side chain stereochemistry. Armed with the methodology to make two
types of peptidic structures, ꢀ-turns with cyclic peptides and cycle-tail motifs with the amidine
peptidomimetics, we turned to the crystal structures of calpastatin-inhibited Ca2+-activated full-
length calpain as a model for inhibitor development.
16
ACS Paragon Plus Environment

Page 17 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
PR1R2R3R4
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
am*[PR1R2R3R4]
Figure 4. Cycle-tail motif peptides derived from aziridine aldehydes and isocyanide. The
amidine (functionality highlighted in red) peptidomimetics are denoted with am*[PR1R2R3R4],
where PR1R2R3R4 corresponds to the linear peptide pentamer precursor.
Modelling potential inhibitors on the ꢀ-turn structure of calpastatin in the calpain active site
cleft.
Previously published methods for making cyclic peptides39,58 provided an option for the
design and modelling of peptides to mimic the conformation of calpastatin in the calpain active
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 49
1
2
3
4
5
6
7
8
9
site (Figure 5). Some of these initial models were analyzed using the molecular dynamics
software package GROMACS.41 These models were then used in docking simulations. MD
simulations of calpain-2, taken from the calpastatin-bound calpain-2 structure, were used as
representative structures to dock cyclic peptides using AutoDock446 and Glide47-49. The lowest
energy conformations for the docked molecules clustered in the unprimed side of the active site
cleft and involved interactions with the C2-like domain of calpain (Figure 5).
The tendency for our initial cyclic peptide designs to also interact with the C2L domain
(green region in Fig. 5) in docking and MD experiments suggested that the synthesized peptides
could potentially be specific inhibitors of calpain, because this domain is absent from other
cysteine proteases.
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 5. Mimicking the active site of calpastatin-bound calpain-2 with cyclic peptides.
Close-up view of the calpastatin-bound calpain-2 active site. A representative cyclic peptide
docking result from Glide is shown for c*[PGLGK] (cyan), with polar hydrogens displayed and
residues labelled, while calpastatin is shown in purple. N and O atoms are coloured blue and red,
18
ACS Paragon Plus Environment

Page 19 of 49
Journal of Medicinal Chemistry
1
2
3
4
respectively. The structure is shown with a partial cut-away to give a clear view of the bound
5
6
inhibitors. Domain colours (C2L œ green, PC2 œ orange, PC1 œ yellow) are as shown in Fig.1.
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
Macrocyclic and cycle-tail motif peptidomimetic library
All peptides were made with Pro at the N-terminus to ensure efficient cyclization. Based
on previous work with tetramers and pentamers containing two D-amino acids, which tended to
undergo amidine formation, we split the library design into two groups. The first group, made of
cyclized pentamers (Figure 3A), was aimed at mimicking the inhibitory calpastatin ꢀ-turn
sequence at KLG. At the Lys, the side chain length requirement was investigated using the
increasingly shorter amine-bearing side chain residues Orn, Dab, and Dap. Based on the crystal
structure of c*[PGLGF], which positioned the ꢀ-turn around Gly-Leu, the basic residues were
placed in the fifth position. Finally, as it was previously seen that the solution phase structure of
c*[PGLGF] could index the turn by one residue and form the ꢀ-turn around Pro-Gly (Figure 3B),
basic residues were also placed in the third position.
Cycle-tail motif peptidomimetics comprised the second group of peptides formed from
the multicomponent reaction of aziridine aldehyde dimers, peptides, and isocyanides (Figure 4).
This set of inhibitor probes was also designed to screen permutations of KLG tripeptide sequence
and length of amine-bearing side chains. As the binding mode was less certain, the order of the
KLG tripeptide was also perturbed in a subset of the peptidomimetic probes. We anticipated that
our compounds would inhibit calpain with radically different properties from Abell‘s tripeptide
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 49
1
2
3
4
5
6
peptidomimetics (Figure 2), which relied on covalent binding to inhibit with high IC50s and were
not selective against cathepsins.24
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
Screening of peptide library using fluorescence-based protease assays
When the cyclized pentamers and peptidomimetics were screened using a fluorescence-
based enzyme activity assay against calpain-2, macrocyclic compounds demonstrated significant
levels of inhibition of calpain-2 (Figure 6A). Approximately 60% of the macrocyclic compounds
showed levels of inhibition of calpain-2 that were ≥ 25% at concentrations of 100 µM. Two
macrocyclic peptides, c*[PGALK] and c*[PGSGO], showed reproducibly high levels of
inhibition of calpain-2 (> 85%) at this concentration.
Levels of inhibition by the peptidomimetic set of compounds were similar to those of the
macrocyclic forms with approximately half of the cycle-tail motif peptidomimetics showing ≥
25% inhibition at concentrations of 100 µM. One peptidomimetic, am*[PLKG], also showed
reproducibly high levels of inhibition of calpain-2 (> 85%) at the 100 µM concentration, while
another, am*[PGLo], similarly showed high levels of inhibition of calpain-2 at concentrations of
100 µM, but with more variation (from 16-72%) (Figure 6B).
Full-length calpain enzyme has been found to undergo extensive autoproteolysis in vitro
upon activation60. The autoproteolysis of calpain was examined in the context of these
experiments to ensure apparent inhibition of calpain was not due to loss of activity enzyme.
Under the experimental conditions used and the short time frame analyzed for activity,
autoproteolysis was not found to be a significant source of inactivation of calpain-2 (see SI).
20
ACS Paragon Plus Environment

Page 21 of 49
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 49
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
ACS Paragon Plus Environment

Page 23 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Due to the high structural similarities of cysteine proteases, it was necessary to test any
potential inhibitors for protease selectivity. Inhibition assays using fluorogenic substrates were
performed against calpain-1 protease core, papain, and cathepsin L (Figure 8). We observed
potent but reduced inhibition of calpain-1. As the calpain-1 protease core construct lacks the
auxiliary domains of calpain, it gave us the opportunity to deduce if our inhibitors exhibited their
effect through the C2L domain. Indeed, we observed negligible inhibition of calpain-1 protease
core, which meant that our study peptides interacted with more than just the PC1 and PC2
domains of the protease core. The negative inhibition seen of the calpain-1 protease core by these
compounds indicates activation of the enzyme. Previous studies have suggested the inherent
flexibility of the protease core, caused by a lack of ancillary domains, results in lower activity
than the full-length calpain enzyme13. It is possible that the protease core domains could be
stabilized upon binding to some compounds, and thus display increased protease activity. It
should be noted that currently, the only recombinantly-expressing, active construct of the
protease core domains is that of calpain-1.
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
c*[PGALK] and c*[PGSGO] showed some specificity for calpain-2 over calpain-1
(respectively, c*[PGALK]: ~ 84% and ~ 45%, c*[PGSGO]: ~ 78% and ~ 38%). It is uncertain
why there could be differential inhibition between calpains with these macrocycles because of
the high sequence and structural homology of calpains-1 and -2, especially in the active site cleft
where the presumed competitive inhibitors would bind.
Papain and cathepsin L enzymes were used to investigate any differences based on
similarities in structural characteristics of the catalytic triad. The four selected lead compounds,
c*[PGALK], c*[PGSGO], am*[PLKG], and am*[PGLo], showed little to no inhibition of
papain, but high levels of inhibition (> 85%) of cathepsin L.
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 49
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
ACS Paragon Plus Environment

Page 25 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the orientation in the binding site, both of which are directly affected by peptide sequence and
the cyclization linker atoms. Moreover, the lack of protease core inhibition indicated that the
inhibitors were interacting with other regions such as the C2L domain. However, low selectivity
against cathepsin L implied that the inhibitors were not solely binding to the C2L domain.
Finally, peptides but am*[PGLo], shared a common selectivity profile, which warranted a closer
look by inhibitions kinetics.
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
Enzyme kinetics analyses and inhibition mechanisms.
To determine the mode of inhibition of the lead cyclic compounds with calpain-2,
extensive analyses of enzyme kinetics were performed using the fluorogenic substrate assay.
Using the four potent lead compounds, the initial rates of substrate cleavage by calpain-2 were fit
to the Michaelis-Menten plot for uninhibited and inhibited reactions (Figure 9A-D).
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 49
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
ACS Paragon Plus Environment

Page 27 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
parameters, it was determined that of the four lead compounds, am*[PGLo] functioned as a non-
competitive inhibitor, am*[PLKG] and c*[PGALK] showed competitive inhibition, and
c*[PGSGO] fit the mixed inhibitor category when tested with calpain-2. The compounds that
were found to be competitive inhibitors showed higher standard error of the parameters when fit
to the mixed inhibition equation. Thus, these data were refit to the competitive inhibition
equation and their parameters reported. The B27 WT peptide inhibitor based on calpastatin was
used as a control to ensure accurate data analyses, and showed competitive inhibition of
approximately 16.8 ± 1.7 nM, close to the previously reported value of 8.7 ± 2 nM.61
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
Table 1: Summary of inhibition kinetics after fitting data to the full mixed enzyme
inhibition equation.
Inhibition
Compound
Ki1 (ꢁM)
Ki2 (ꢁM)
model
competitive
competitive
competitive
mixed
B27 WT peptide
am*[PLKG]
c*[PGALK]
c*[PGSGO]
am*[PGLo]
16.8 x 10-3 ± 1.7 x 10-3 a
12.2 ± 0.8 a
n/a
n/a
16.6 ± 1.0 a
n/a
20.3 ± 2.0
45.6 ± 19.8
119.4 ± 37.9
153.7 ± 27.6
non-competitive
Calpain-2 cleavage of the fluorescent substrate (EDANS)-EPLFAERK-(DABCYL) was monitored over time.
Inhibition kinetics were obtained by fitting initial rate data to the full mixed enzyme inhibition equation. Reactions
were performed at 298 K in a reaction buffer consisting of 10 mM HEPES pH 7.4 and 10 mM DTT. Calpain activity
was initiated through the addition of 10 mM CaCl2. n/a denotes compounds whose fit of data to the full mixed
equation resulted in an extremely large Ki2 value with unreasonably high margin of error, and therefore exhibited
competitive inhibition. a Compounds were found to be competitive inhibitors and showed lower standard error for
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 49
1
2
3
4
5
6
parameters when fit to the competitive inhibition equation than when fit to the mixed inhibition equation. Their
parameters are thus reported according to the fit to the competitive equation.
7
8
9
While our initial efforts were to design competitive macrocyclic peptide inhibitors with
specificity for calpain through interactions with the C2L domain, only two of the four most
potent compounds exhibited competitive inhibition. The other two showed non-competitive
inhibition and mixed mode inhibition. Ideally, the macrocycle conformation could help form a
calpastatin-like ꢀ-turn, while the residues Leu and Lys, responsible for a key interaction between
calpastatin and calpain, could anchor the macrocycles in the intended binding site. It is apparent
that macrocyclic compounds without these crucial residues might then result in mixed inhibition,
as is the case with c*[PGSGO], a non-competitive inhibitor. Further studies may help elucidate
the role that ring size and flexibility may play in terms of inhibition potency. The cycle-tail motif
peptidomimetics display a novel peptide structure for calpain inhibition, yet there is seemingly
little information to be gathered as to a pattern for preferred residues (Figure 6B). It is possible
that, as is the case for c*[PGSGO], the peptide sequence can change the binding site of the
peptide and thus the mode of inhibition as well. Investigating these cycle-tail motif
peptidomimetics compounds further by structural studies will be crucial to defining the new
binding pockets and structure-activity relationships.
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
The most potent of all analyzed study compounds was the competitively-inhibiting
am*[PLKG] with a Ki of 12.2 ± 0.8 ꢁM. Ki values were calculated rather than IC50 values in
order to assess affinity for the enzyme as well as to determine the inhibition mechanism. The
level of inhibition seen with am*[PLKG] approaches the range of potency found with other
covalent and non-covalent peptidomimetic competitive inhibitors of calpain.26 Specifically, when
comparing these levels of inhibition to other inhibitors designed based on the calpain-calpastatin
28
ACS Paragon Plus Environment

Page 29 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
structure, we see that am*[PLKG], despite lacking covalent attachment to the enzyme, inhibits
calpain with equal potency as the irreversibly-binding .-helical inhibitory peptide 3c (Ki of 10.2
± 2.9 ꢁM).62 Most calpain inhibition studies report IC50 values instead of Ki values. The former
are easier to compute but are less rigorous than the latter. Although it is difficult to equate one to
the other, as the relationship between IC50 and Ki depends on inhibition mechanism as well as
reaction conditions,63,64 the larger macrocyclic inhibitors of calpain reported previously are
within the same order of magnitude as am*[PLKG],24 and show some selectivity for calpain.
Interestingly, am*[PLKG] as the most potent compound in this series retained the conserved
residues of the calpastatin ꢀ-turn structure, albeit in mixed order. It remains to be seen from an
enzyme-inhibitor complex structure if this is the reason for the potency for cycle-tail motif
peptidomimetics.
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
Molecular docking of c*[PGALK] NMR structures to calpain-2.
Because the macrocyclic peptide c*[PGALK] was found to competitively inhibit calpain-
2, we modelled its inhibition by docking studies. c*[PGALK] was modelled first with
Macromodel (simulated annealing and Monte Carlo multiple minimum (MCMM) with the
OPLS-2005 force field) and we considered the 30 lowest energy structures that fit the NMR data.
To accurately sample interaction of the side chains with the protein while accounting for the
hydrogen bonding network across the cycle backbone itself, all 30 lowest energy structures were
docked individually with torsional constraints on the backbone and peptide bonds while keeping
the side chains flexible. AutoDock4 and Glide provided the same structures as the most
favourable poses. Analysis of the representative top poses from Glide, ranked by lowest free
29
ACS Paragon Plus Environment