Conformationally restricted calpain inhibitors

Article abstract of DOI:10.1039/c5sc01158b

The cysteine protease calpain-I is linked to several diseases and is therefore a valuable target for inhibition. Selective inhibition of calpain-I has proved difficult as most compounds target the active site and inhibit a broad spectrum of cysteine prote

Full text of DOI:10.1039/c5sc01158b

Chemical
Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
Conformationally restricted calpain inhibitors†
S. E. Adams,a E. J. Robinson,b D. J. Miller,a P. J. Rizkallah,b M. B. Hallettb
Cite this: Chem. Sci., 2015, 6, 6865
a
*
and R. K. Allemann
The cysteine protease calpain-I is linked to several diseases and is therefore a valuable target for inhibition.
Selective inhibition of calpain-I has proved difficult as most compounds target the active site and inhibit a
broad spectrum of cysteine proteases as well as other calpain isoforms. Selective inhibitors might not only
be potential drugs but should act as tools to explore the physiological and pathophysiological roles of
calpain-I. a-Mercaptoacrylic acid based calpain inhibitors are potent, cell permeable and selective
inhibitors of calpain-I and calpain-II. These inhibitors target the calcium binding domain PEF(S) of
calpain-I and -II. Here X-ray diffraction analysis of co-crystals of PEF(S) revealed that the disulfide form
of an a-mercaptoacrylic acid bound within a hydrophobic groove that is also targeted by a calpastatin
inhibitory region and made a greater number of favourable interactions with the protein than the
reduced sulfhydryl form. Measurement of the inhibitory potency of the a-mercaptoacrylic acids and X-
ray crystallography revealed that the IC50 values decreased significantly on oxidation as a consequence
of the stereo-electronic properties of disulfide bonds that restrict rotation around the S–S bond.
Consequently, thioether analogues inhibited calpain-I with potencies similar to those of the free
sulfhydryl forms of a-mercaptoacrylic acids.
Received 31st March 2015
Accepted 20th August 2015
DOI: 10.1039/c5sc01158b
www.rsc.org/chemicalscience
approximately 50 mM Ca2+, whereas calpain-II requires approx-
imately 350 mM Ca2+ for activation.1
Introduction
Numerous physiological processes have been linked with
calpain-I and -II, including cell motility,8–10 apoptosis11,12 and
progression through the cell cycle13 but the precise roles of
these proteases remain poorly understood, which is at least in
part due to a lack of specic inhibitors that allow selective
knockout or knock-down of their cellular activities.14 Other
techniques such as microinjection of a surplus of the enzymes
or the release of calcium ions into cells have been used to
activate the enzymes to explore their cryptic roles in vivo. The
majority of synthetic inhibitors available to examine the role of
these isoforms are generic cysteine protease inhibitors that
react with the active site cysteine and hence show little
discrimination between calpain isoforms or indeed other
cysteine proteases such as caspases or cathepsins.15–17 The cal-
pain system includes an endogenous inhibitor, calpastatin
(CAST), a large protein that only binds to members of the cal-
Calpain-I and calpain-II are the two most studied members of a
family of calcium dependent cysteine proteases that currently
comprises een identied gene products in humans.1–3 These
heterodimeric proteases are composed of a large subunit with a
molecular mass of ꢀ80 000 and a small subunit of mass
ꢀ30 000. Calpain-I and calpain-II share a small subunit, which
consists of two domains, a penta-EF hand calcium binding
domain PEF(S) and a glycine rich domain that is thought to
interact with cellular membranes;4 the large subunits of cal-
pain-I and -II possess 62% sequence similarity in humans5 and
comprise four distinct domains, a N-terminal anchor helix, the
active site domain (CysPc), a domain that resembles the C2
membrane binding domains of phosphokinases and is hence
known as the C2L domain,2,6 and a second penta-EF hand
calcium binding domain known as PEF(L). PEF(L) is the domain
that determines the concentration of calcium required for
protease activation, which is the discriminating factor between
the two isoforms.1,7 Calpain-I is activated in vitro by
pain family that form
a heterodimeric complex, hence
discriminating between calpains and other cysteine proteases.
Specic cell permeable inhibitors of calpain(s) that target the
CAST binding sites could be used as valuable cell-biological
tools for the elucidation of the cryptic physiological and path-
ophysiological roles of calpain isoforms and potentially as
drugs to treat conditions such as autoimmune diseases,
ischemic stroke damage and cancer.14,18,19
A previous investigation led to the discovery of novel, iso-
form-selective inhibitors that target calpain through interac-
tions with allosteric binding sites.17 The a-mercaptoacrylic acids
aSchool of Chemistry, Cardiff University, Main Building, Park Place, Cardiff, UK, CF10
3AT. E-mail: allemannrk@cardiff.ac.uk; Fax: +44 (0) 29 208 74030; Tel: +44 (0) 29
2087 9014
bInstitute of Infection & Immunology, School of Medicine, Heath Campus, Cardiff, UK,
CF14 4XN
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c5sc01158b
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 6865–6871 | 6865

View Article Online
Chemical Science
Edge Article
that would otherwise hold the side chains of Leu 656 and Ile 653
of the inhibitory helix are nearby and should be available to
bind additional hydrophobic groups of a designed inhibitor
(Fig. 2). The close proximity of these lipophilic amino acids to
one another suggested that disulde versions of a-mercaptoa-
crylic acids might be able to mimic the binding mode of cal-
pastatin. Close inspection of the structure of the inhibitory
region C of the CAST domain IV bound to PEF(S) and molecular
docking experiments revealed the possibility that due to the
stereoelectronic constraints of sulfur–sulfur bonds the two
aromatic rings of a disulde linked linked form of 3 might t
into the two adjacent hydrophobic pockets of PEF(S); in
particular the aromatic groups of such a disulde based
compound may interact with the pockets that bind Leu 656 and
Leu 660 of calpastatin. To test this proposal, the oxidised form
of 3, (2Z,20Z)-2,20-disulfanediylbis(3-(6-bromoindol-3-yl)acrylic
acid), (4, Fig. 1), was generated and the single crystal X-ray
structure of 4 bound to PEF(S) solved.
Fig.
1 Structures of PD150606 (1), PD151746 (2), (Z)-3-(6-bro-
moindol-2-yl)-2-mercaptoacrylic acid (3) and (2Z,20Z)-2,20-disul-
fanediylbis(3-(6-bromoindol-3-yl)acrylic acid) (4).20,24
The rate of conversion of 3 to its disulde form was depen-
dent upon both the concentration of the compound and the
nature of the solution. UV-VIS spectroscopy showed that 3 (100
mM in 100 mM KiPO4, pH 7.0) was completely converted to its
oxidised form 4 in 3 hours at room temperature and was easily
reduced to the sulydryl form by addition of tris(2-carboxy-
ethyl)phosphine (TCEP) (see ESI†). For crystallography, 3 (50
mM) was dissolved in DMSO and le at room temperature for
24 hours, aer which time oxidation was complete according to
1H-NMR spectroscopy (see ESI†). Compound 4 was then added
to preformed crystals of PEF(S).
The electron density map revealed that the ligand adopts two
different conformations in each subunit of the PEF(S) homo-
dimer, therefore a total of four conformations of the ligand
were observed in the asymmetric unit of the crystal structure.
The unique geometry of the disulde bond is observed in all of
the conformations ensuring that both pockets of the hydro-
phobic groove of PEF(S) are explored. One of the 6-bromoindole
rings of 4 is bound in the hydrophobic binding pocket that also
binds 3. The aromatic ring is capable of moving within the
pocket, adopting two different conformations with the ring
ipped by approximately 180ꢁ relative to the other as observed
in chain A of PEF(S) (Fig. 2). The second indole ring of 4
interacts with a hydrophobic groove on the surface of PEF(S),
where the a-helix of CAST inhibitory region C binds with
PD150606 (1) and PD151746 (2) (Fig. 1)20 bind to the calcium
binding domain of calpain rather than the active site and show
modest selectivity for calpain-I over calpain-II.21,22 The hydro-
phobic pocket of PEF(S) targeted by these inhibitor is also
bound by the inhibitory region C of calpastatin domain IV,
where Leu 660 is embedded in the pocket.23 Based on
compounds 1 and 2, potent cell permeable inhibitors of cal-
pain-I that inhibit the cell spreading action of live neutrophils in
vitro were synthesised.24,25 A single co-crystal X-ray structure
showed that like PD150606 the new compounds bound to the
calcium binding domain PEF(S).24,26 The thiol and carboxylic
acid groups of the inhibitors are critical for activity20,27 but
somewhat surprisingly these functional groups point away from
the protein into solution. Here we report that oxidation of the
sulydryls of these a-mercaptoacrylic acids to form disuldes
leads to inhibitors with greatly enhanced potency, where one
half of the compound targets the hydrophobic CAST binding
groove of PEF(S). This previously unexplored mode of action
opens the way for the development of a new generation of stable
and selective inhibitors of calpain-I.
Results and discussion
The interaction of (Z)-3-(6-bromoindole-3-yl)-2-mercaptoacrylic PEF(S).21 Two different binding orientations of the ligand are
acid (3) (Fig. 1) with human PEF(S) was investigated by single observed, allowing for different protein residues to interact
crystal X-ray crystallography. Human PEF(S) was produced in E. with the ligand. In chain A, the second 6-bromoindole ring
coli, puried to apparent homogeneity, crystallised and soaked resides in a hydrophobic pocket where Leu 656 of CAST also
with 3 as previously described for other a-mercaptoacrylic binds (Fig. 2), whereas in chain B it targets the hydrophobic
acids.26,28 Similar to other PEF(S) a-mercaptoacrylic acid co- groove that binds the peptide backbone of CAST inhibitory
crystal structures, 3 bound to the hydrophobic pocket of PEF(S) region C (Fig. 2).23
in the region that also binds Leu 660 of the CAST domain IV
With the exception of a hydrogen bond between the
inhibitory region C. This is situated between the second and carboxylate of 3 and the side chain amide group of Gln175 in
fourth a-helices (Fig. 2).21–23,26 The hydrophobic and carbonyl chain B, van der Waals forces are the main interactions
portions of 3 match the spatial orientation of Leu 660, but many observed between 3 and the residues of both monomers of the
unexplored possibilities appear to exist for hydrophobic protein. A greater number of interactions form between 4 and
contacts corresponding to preceding residues in the a-helix of residues of PEF(S) than with 3 and the protein, including a
CAST inhibitory region C (Fig. 2).23 The hydrophobic pockets number of electrostatic interactions and hydrogen bonds. In
6866 | Chem. Sci., 2015, 6, 6865–6871
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
Fig. 2 (A) Surface representation of chain A of PEF(S) bound to 3 in the hydrophobic pocket that binds Leu 660 (PDB 4WQ2). (B) Inhibitory region
C of the CAST domain IV bound to PEF(S); Ile 653, Leu 656 and Leu 660 are highlighted (PDB 3BOW).23 (C) 4 bound to chain A of PEF(S); a single
binding mode of 4 is represented (PDB 4WQ3).
chain A, six hydrophilic interactions between the two confor- derivatives (Fig. 4). As observed previously,24 the position of the
mations of 4 and Arg130, His131 and Trp168 are observed halogen in the aromatic ring for phenyl based inhibitors was
(Fig. 3). An electrostatic bond forms between the carboxylate important for the inhibitory potency. All 2-substituted
group of 4 and the positively charged guanidinium group of compounds have IC50 values of ꢀ50 mM whereas compounds
Arg130. Hydrogen bonds are observed between the carboxylates that are substituted in the 3- or 4-position are about one order
of the ligand and the side-chain NH groups of His131 and of magnitude more potent. Also, the type of halogen is
Trp168 (Fig. 3). A different electrostatic interaction is seen important for the potency for the inhibitor in that F substituted
between 4 bound to chain B between a carboxylate group of the compounds are approximately 10-fold weaker inhibitors than
+
˚
ligand and the NH3 group of Lys172 with distances of 2.17 A the corresponding Br and I substituted inhibitors. This trend
˚
and 2.51 A for the two conformations (Fig. 3). Interactions form does not hold for Cl and Br substituents in the 2 position,
between the bromine atoms of 4 and hydrophilic groups in which bind with affinities of ꢀ60 mM and ꢀ50 mM, respectively
chain B; one bromide interacts with Arg130 and the second with (Fig. 4). The indole based a-mercaptoacrylic acid derivatives
˚
˚
His131 with Br–N distances of 3.31 A and 3.15 A, respectively showed IC50 values in the low micromolar range in the pres-
(Fig. 3). The greater number of interactions observed between ence of DTT and there was no obvious trend with respect to
compound 4 and PEF(S) suggests tighter binding to PEF(S). The position or type of halogen (Fig. 4). The IC50 values reported
structure of 4 bound to PEF(S) prompted a re-evaluation of the here are higher than previously reported for the same
inhibitory potency of the a-mercaptoacrylic acids previously compounds, most likely due to the presence of poorly dened
synthesised.24 Previously, no special measures had been taken mixtures of reduced and oxidised mercaptoacrylic acids in
to ensure that the compounds were in oxidised or reduced form previous experiments.14–17,20,24
during the assays. The oxidised and reduced compounds were
now assayed against calpain-I to determine their affinity for molar range, a 10- to 200-fold increase in potency, indicating
PEF(S). that the thiol group is not important for inhibition of calpain-I.
For the oxidised compounds IC50 values were in the nano-
The inhibitory properties of a series of oxidised and reduced The IC50 values for the phenyl derivatives range from 7.5 to 250
a-mercaptoacrylic acid derivatives towards calpain-I (Fig. 4) nM. The position of the halogen was important for the potency
were assessed in a FRET-based assay in the presence and of the inhibitor with a trend similar to that observed for the
absence of 10 mM dithiothreitol (DTT).24,25 To ensure complete reduced compounds. Iodo substituted compounds (17 and 23)
oxidation, compounds were le in the assay solution for 3 hours showed IC50 values of 25 nM and 7.5 nM, respectively (Fig. 4).
prior to performing the assay (vide supra and ESI†). In the With the exception of the iodo-substituent the type of halogen
absence of inhibitors there was no observable difference in appears to have little importance for IC50 values. The indole
enzyme activity when assayed under both oxidising and based disulde compounds show a 10- to 150-fold increase in
reducing conditions.
potency relative to the free sulydryls. IC50 values ranged from
When DTT was present, IC50 values were in the micromolar 10 to 28 nM, with the exception of 3, which was less potent (IC50
range for both phenyl and indole based a-mercaptoacrylic acid ¼ 100 nM) (Fig. 4). No discernible trend was observed with
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 6865–6871 | 6867

View Article Online
Chemical Science
Edge Article
Fig. 3 Comparison of the two conformations 4 (cyan) bound to chain A (yellow) of the PEF(S) homodimer, (A) and (B) respectively, and chain B
(green) of the homodimer, (C) and (D) respectively. Residues that form halogen bonds, hydrogen bonds and electrostatic interactions are
highlighted and important bonds are indicated by dashed lines, distances are shown in A˚ (PDB 4WQ2 and 4WQ3).
respect to the position or type of halogen relative to the potency substituents.31 The dihedral angles of the observed conforma-
of the inhibitor (Fig. 4). The most potent inhibitor in this series tions of 4 bound to PEF(S) are 30, ꢂ91, 128 and 169. The
was the disulde of PD150606 (23), which when tested against dihedral angle of ꢂ91 in chain A is close to the optimum
calpain-I gave an IC50 value of 7.5 nM, much more potent than dihedral angle and in this conformation 4 makes strong inter-
the Ki value of 0.21 mM, possibly due to the inhibition of calpain- actions with both hydrophobic pockets of PEF(S) (Fig. 5).
I being measured in a poorly dened redox buffer (Fig. 4).20 The Disulde bonds are on average 2.08 A, approximately 0.5 A
increased potency of PD150606 in the disulde form 23 led us to longer than a carbon–carbon single bond. The S–S bond length
˚
˚
32
˚
an examination of the co-crystal structure PEF(S) with this of 4 bound to PEF(S) ranges from 2.06 to 2.11 A. The slightly
inhibitor in oxidised form.21,22 The structure obtained shows a increased bond length allows for the second aryl group to
lower electron density around the model of the ligand 23 to that interact with the hydrophobic pocket that binds Leu 656 of
observed for 4, though upon comparison of the two structures calpastatin. To establish whether the properties of the disulde
both ligands adopt similar conformations and form analogous bond were indeed responsible for the strong inhibitory of
interactions with the protein (see ESI†).
compounds such as 4, one of the sulfur atoms was replaced with
The disulde bond is the only group in the structure of the a methylene group to generate a series of thioethers. The rela-
oxidised a-mercaptoacrylic acids that allows for exibility tively unhindered rotation around carbon–sulphur bonds
within the molecule. However in agreement with the design, should lead to a greater loss of entropy when they bind to PEF(S)
this exibility is limited due to the specic stereoelectronic and hence to reduced stability compared to the disulde
properties of S–S bonds. Rotation around the disulde bond is complexes.
restricted due to the repulsive interactions of the lone pairs of
Symmetric and asymmetric thioether compounds were syn-
each sulfur atom.29 Dihedral angles of aromatic disuldes are thesised in a four-step procedure (Fig. 6). A halogenated benz-
typically between 100 and 106ꢁ,29,30 although the dihedral angles aldehyde was reacted with methyl acrylate, followed by base
can be both greater and smaller depending upon the catalysed hydrolysis of the ester and treatment with aqueous
6868 | Chem. Sci., 2015, 6, 6865–6871
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
Fig. 4 IC50 values (mM) for the a-mercaptoacrylic acid compounds tested in FRET based inhibition assays in the presence (sulfhydryl) and
absence (disulfide) of 10 mM DTT.
measured were higher than those obtained for all the disuldes
examined here and indeed generally higher than the free sulf-
hydryls. Also, there was no discrimination between the
symmetric (40) and asymmetric (41, 42 and 43) thioether
derivatives with regards to the efficacy of these compounds
towards calpain-I.
Inhibition data for the thioether-based compounds supports
the importance of the stereoelectronic properties of the S–S
bond for the potency of the mercaptoacrylic acid based inhibi-
tors. The greater degree of exibility of the thioether bond leads
to higher entropic penalty when these compounds ‘lock’ into
position on PEF(S), which in turn produces higher IC50 values.
In addition, the decreased bond length of the C–S bond relative
to the S–S bond may lead to a suboptimal interaction of the
second aromatic ring with the hydrophobic pocket of PEF(S)
Fig. 5 Representation of 4 with a dihedral angle of ꢂ90.5ꢁ bound to
that binds Leu 656 of CAST. Crystallographic results support
this rationalisation of the low PEF(S) binding affinity of the
chain A of PEF(S).
thioethers. Single crystals of PEF(S) were individually soaked
with 40, 42 and 43 and the resulting crystals analysed. The X-ray
hydrogen bromide to generate an allylic bromide derivative, diffraction data indicated only partial occupancy of the cal-
which was used to alkylate an a-mercaptoacrylic acid.24 All IC50 pastatin binding site. Models representing compounds 40, 42
values of were in the micro-molar range between 1.8 and 10 mM and 43 were placed into each of the 2F0 ꢂ Fc maps revealing a
except for 41 (IC50 ¼ 100 mM) and DTT had no effect on the maximum occupancy of the ligands within the electron density
potency (Fig. 6). With an IC50 value of 1.8 mM the asymmetric 42 of 30% (see ESI†).
was the most potent thioether examined. The IC50 values
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 6865–6871 | 6869

View Article Online
Chemical Science
Edge Article
of 4 interact with the hydrophobic pockets that bind Leu 660
and Leu 656 of calpastatin. All disulde linked compounds
acted as potent inhibitors of calpain-I and were up to 200 times
more potent than their reduced counterparts. The dramatic
increase in potency of the oxidised inhibitors can be tentatively
explained by the stereoelectronic properties of S–S bonds such
as the reduced rotations around the S–S bond and the approx-
imately 90ꢁ dihedral angles. Accordingly, thioether analogues
were much less potent than their disulde counterparts.
These results open the way to the development of inhibitors
that combine the conformational restrictions of the S–S bond
with reduced sensitivity to reduction. Diselenides are charac-
terised by dihedral angles of 101–106ꢁ and bond lengths of
˚
approximately 2.3 A, values similar to those found in disuldes.
However, diselenides are more stable towards reduction than
disuldes.33–35 The incorporation of diselenide bonds could
therefore lead to highly potent drugs and cell biology tools that
are stable to the reducing conditions found in the cellular
environment.
Acknowledgements
This work was supported by the UK's Engineering and Physical
Sciences Research Council (EPSRC) and the Medical Research
Council (MRC) through grant G0701192. We would like to thank
Cardiff University and the EPSRC for the award of a Doctoral
Prize Fellowship to SEA from DTG grant number EP/K502819/1.
We are grateful to the EPSRC national mass spectrometry
service at Swansea for mass spectrometric analyses and Dr
Robert Jenkins and Mr Robin Hicks, Cardiff School of Chem-
istry for assistance with NMR spectroscopy and in-house mass
spectrometry. The authors would like to thank Diamond Light
Source for beamtime (proposal MX8096) and the staff of
beamlines I03 and I04-1 for assistance with crystal testing and
data collection.
Fig. 6 (A) Synthetic route to the thioether calpain inhibitors. (i)
DABCO, MeOH, 25 ꢁC, 72 h, (ii) NaOH, MeOH, 25 ꢁC, 16 h, (iii) HBr,
H2SO4, 25 ꢁC, 16 h, (iv) NEt3, MeCN, 25 ꢁC, 16 h. (B) Symmetric and
asymmetric compounds 40–43. (Z)-3-(4-Bromophenyl)-2-((((Z)-2-
(4-bromophenyl)-1-carboxyvinyl)thio)methyl)acrylic acid (40), (Z)-3-
(3-chlorophenyl)-2-((((Z)-2-(4-chlorophenyl)-1-carboxyvinyl)thio)-
methyl)acrylic acid (41), (Z)-3-(5-bromoindol-3-yl)-2-(((Z)-2-carboxy-
3-(3-chlorophenyl)allyl)thio)acrylic acid (42) and (Z)-3-(6-bro-
moindol-3-yl)-2-(((Z)-3-(4-bromophenyl)-2-carboxyallyl)thio)acrylic
acid (43).
References
1 D. E. Goll, V. F. Thompson, H. Li, W. Wei and J. Cong,
Physiol. Rev., 2003, 83, 731–801.
2 H. Sorimachi, S. Hata and Y. Ono, Proc. Jpn. Acad., Ser. B,
2011, 87, 287–327.
3 R. L. Campbell and P. L. Davies, Biochem. J., 2012, 447, 335–
351.
4 S. Ohno, Y. Emori and K. Suzuki, Nucleic Acids Res., 1986, 14,
5559.
5 S. Imajoh, K. Aoki, S. Ohno, Y. Emori, H. Kawasaki,
H. Sugihara and K. Suzuki, Biochemistry, 1988, 27, 8122–
8128.
6 P. Tompa, Y. Emori, H. Sorimachi, K. Suzuki and
P. Friedrich, Biochem. Biophys. Res. Commun., 2001, 280,
1333–1339.
7 H. Sorimachi, S. Hata and Y. Ono, J. Biochem., 2011, 150, 23–
37.
8 M. B. Hallett and S. Dewitt, Trends Cell Biol., 2007, 17, 209–
214.
Conclusions
The work described here reveals a novel mechanism for inhi-
bition of calpain-I by a-mercaptoacrylic acid derivatives. Based
on the X-ray crystal structure of PEF(S) bound to (Z)-3-(6-bro-
moindol-2-yl)-2-mercaptoacrylic acid (3), oxidised a-mercap-
toacrylic acids such as 4 were designed to target the full
calpastatin binding cavity. The restricted geometry of the
disulphide bond facilitated a larger number of favourable
binding interactions with PEF(S) so that the two aromatic rings
9 S. Dewitt and M. B. Hallett, J. Cell Biol., 2002, 159, 181–189.
6870 | Chem. Sci., 2015, 6, 6865–6871
This journal is © The Royal Society of Chemistry 2015

View Article Online
Edge Article
Chemical Science
10 N. Dourdin, A. K. Bhatt, P. Dutt, P. A. Greer, J. S. C. Arthur,
J. S. Elce and A. Huttenlocher, J. Biol. Chem., 2001, 276,
48382–48388.
11 S. Orrenius, B. Zhivotovsky and P. Nicotera, Nat. Rev. Mol.
Cell Biol., 2003, 4, 552–565.
12 S. M. Harwood, M. M. Yaqoob and D. A. Allen, Ann. Clin.
Biochem., 2005, 42, 415–431.
13 J. E. Schollmeyer, Science, 1988, 240, 911–913.
L. J. DeLucas and S. V. L. Narayana, Nat. Struct. Biol., 1997,
4, 539–547.
23 R. A. Hanna, R. L. Campbell and P. L. Davies, Nature, 2008,
456, 409–412.
24 S. E. Adams, C. Parr, D. J. Miller, R. K. Allemann and
M. B. Hallett, MedChemComm, 2012, 3, 566–570.
25 S. Mittoo, L. E. Sundstrom and M. Bradley, Anal. Biochem.,
2003, 319, 234–238.
14 D. J. Miller, S. E. Adams, M. B. Hallett and R. K. Allemann, 26 S. E. Adams, P. J. Rizkallah, D. J. Miller, E. J. Robinson,
Future Med. Chem., 2013, 5, 2057–2074.
M. B. Hallett and R. K. Allemann, J. Struct. Biol., 2014, 187,
236–241.
27 K. K. W. Wang and P. W. Yuen, US Pat., 5760048, 1998, pp.
546–547.
28 G. D. Lin, D. Chattopadhyay, M. Maki, E. Takano,
M. Hatanaka, L. DeLucas and S. V. L. Narayana, Acta
Crystallogr., Sect. D: Biol. Crystallogr., 1997, 53, 474–476.
15 I. O. Donkor, Curr. Med. Chem., 2000, 7, 1171–1188.
16 I. O. Donkor, Expert Opin. Ther. Pat., 2011, 21, 601–636.
17 M. Pietsch, K. C. H. Chua and A. D. Abell, Curr. Top. Med.
Chem., 2010, 10, 270–293.
18 S. J. Storr, N. O. Carragher, M. C. Frame, T. Parr and
S. G. Martin, Nat. Rev. Cancer, 2011, 11, 364–374.
19 Y. H. Huang and K. K. W. Wang, Trends Mol. Med., 2001, 7, 29 L. Pauling, The Nature of the Chemical Bond, Cornell
355–362. University Press, Ithica, New York, 3rd edn, 1960.
20 K. K. W. Wang, R. Nath, A. Posner, K. J. Raser, 30 C. C. Price and S. Oae, Sulfur Bonding, The Ronald Press
M. BurokerKilgore, I. Hajimohammadreza, A. W. Probert, Company, New York, 1st edn, 1962.
F. W. Marcoux, Q. H. Ye, E. Takano, M. Hatanaka, 31 H. E. Vanwart, A. Lewis, H. A. Scheraga and F. D. Saeva, Proc.
M. Maki, H. Caner, J. L. Collins, A. Fergus, K. S. Lee, Natl. Acad. Sci. U. S. A., 1973, 70, 2619–2623.
E. A. Lunney, S. J. Hays and P. W. Yuen, Proc. Natl. Acad. 32 L. S. Higashi, M. Lundeen and K. Seff, J. Am. Chem. Soc.,
Sci. U. S. A., 1996, 93, 6687–6692. 1978, 100, 8101–8106.
21 B. Todd, D. Moore, C. C. S. Deivanayagam, G. D. Lin, 33 D. Besse, F. Siedler, T. Diercks, H. Kessler and L. Moroder,
D. Chattopadhyay, M. Maki, K. K. W. Wang and
Angew. Chem., Int. Ed., 1997, 36, 883–885.
¨
34 W. H. H. Gunther, J. Org. Chem., 1966, 31, 1202–1205.
S. V. L. Narayana, J. Mol. Biol., 2003, 328, 131–146.
¨
22 G. D. Lin, D. Chattopadhyay, M. Maki, K. K. W. Wang, 35 W. H. H. Gunther, J. Org. Chem., 1967, 32, 3931–3933.
M. Carson, L. Jin, P. Yuen, E. Takano, M. Hatanaka,
This journal is © The Royal Society of Chemistry 2015
Chem. Sci., 2015, 6, 6865–6871 | 6871