- Email: [email protected]
Modulating caspase activity: beyond the active site
Available online at www.sciencedirect.com
ScienceDirect Modulating caspase activity: beyond the active site Jeremy Murray1 and Adam R Renslo2 Caspases are a family of aspartate-specific cysteine proteases that regulate cellular homeostasis through the mediation of apoptosis and inflammation. Despite keen interest in caspases as therapeutic targets for cancer, inflammatory, and neurodegenerative diseases, no active-site directed small molecule has yet succeeded in navigating human clinical trials. At the same time, recent biochemical and biophysical studies have revealed caspases to be highly dynamic proteases possessing a remarkable diversity of activation mechanisms. In addition, many caspases possess an allosteric circuit linking key active site loops with a distal allosteric site located at the dimer interface. Accordingly, small molecule binding at this allosteric site directly impacts structural organization of the active site and thus catalytic activity. Both cysteine-tethered and non-covalent reversible small molecules have recently been identified for these allosteric sites, with binding producing a variety of functional effects. Surprising new examples of caspase modulation have also been described recently, including a small molecule that binds caspase-6–substrate complexes uncompetitively and a short peptide that stabilizes an inactive, tetrameric form of procaspase-6. The confluence of recent biochemical, biophysical and pharmacological data has revealed exciting new avenues for the modulation of caspase activity via binding beyond the active site. Addresses 1 Department of Structural Biology, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080, United States 2 Small Molecule Discovery Center and Department of Pharmaceutical Chemistry, University of California, San Francisco, 1700 4th Street, San Francisco, CA 94158, United States Corresponding authors: Renslo, Adam R ([email protected], [email protected])
Current Opinion in Structural Biology 2013, 23:812–819 This review comes from a themed issue on Catalysis and regulation Edited by Ben M Dunn and Alexander Wlodawer For a complete overview see the Issue and the Editorial Available online 8th November 2013 0959-440X/$ – see front matter, # 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.sbi.2013.10.002
Introduction The modulation of caspases as a therapeutic strategy has attracted significant interest given the important roles played by caspases in apoptosis, inflammation, and neurodegeneration [1–3]. The substrate preferences of specific caspases have been known for some time and this has enabled the development of relatively selective Current Opinion in Structural Biology 2013, 23:812–819
inhibitors comprising a tetrapeptide with an electrophilic warhead function [4]. However, the presence in such inhibitors of both electrophilic functionality and an aspartate residue (or anionic surrogate) makes them poor lead compounds from a drug development perspective. Indeed, not a single active-site directed caspase inhibitor has received regulatory approval, and most of the compounds that have entered clinical trials have had their development terminated [3]. These costly failures have encouraged the exploration of alternative approaches for modulating caspase activity. Caspases are a highly dynamic class of proteases that exhibit a diverse array of activation mechanisms that differ according to whether a caspases is at the terminal step in the pathway (effector caspase) or initiates the signaling cascade (initiator caspases). Hence, the apoptotic ‘effector’ caspase-3 (C3) and caspase-7 (C7) are expressed as dimeric procaspases (P3, P7) that become activated on proteolytic cleavage by ‘initiator’ caspases such as C8 and C9. The initiator caspases by contrast are expressed as monomeric proenzymes that are activated upon assembly into larger complexes (e.g., the ‘apoptosome’ in the case of C9). Recent evidence suggests that C6 is unique among effector caspases in being dependent upon self-proteolysis for activation, with the intra-subunit cleavage sequence 190 TEVD193 positioned in the active site of P6, ready for activation [5]. Even the categorization of caspase forms as ‘active’ or ‘inactive’ is surprisingly complex in light of recent structural and biochemical findings that the zymogen and processed forms of C3 and C7 can each adopt either active or inactive conformations [6]. The dynamic nature of caspase structure in effect affords the drug discovery scientist with multiple ‘targets’ for intervention with small molecules. The first small molecules shown to target an inactive caspase conformation were compounds 1 and 2 (Figure 1), which bind an allosteric site at the dimer interface of either C3 or C7, forming disulfide bonds with native cysteine residues present at these sites [7]. Significantly, disulfide-bound 1 or 2 were found to trap the enzyme in a zymogen-like conformation, inhibiting the enzyme by an effectively competitive mode of inhibition (binding of 1 or 2 was mutually exclusive with substrate binding at the active site). Subsequently, an analogous disulfide-tethered inhibitor (3, Figure 1) was identified for C1 [8]. The X-ray structure of 3 bound to C1 (Figure 2) reveals two molecules of 3 (one for each small sub-unit) engaged in a stacking interaction that straddles the dimer interface. Finally, residues comprising an allosteric ‘circuit’ between key active site loops and the dimer interface have been www.sciencedirect.com
Modulating caspase activity: beyond the active site Murray and Renslo 813
Figure 1 Cl
SH F
HN N H
SH Cl
O
O
O 2 (DICA)
1 (FICA) F3C
S
O SH
N N
S
HN
O
N HN
HN
SH
4
3
Current Opinion in Structural Biology
Thiol-containing inhibitors 1 and 2 inhibit executioner caspases C3 and C7 allosterically. Thiols 3 and 4 allosterically inhibit inflammatory caspases C1 and C5 respectively.
identified for several caspases, helping to explain the mutually exclusive binding of allosteric and active-site directed molecules where compounds that bind at the dimer disrupt the hydrogen bond network that is essential for stabilization of the active form [9,10]. In the sections that follow, we will summarize recent reports of smallmolecules and peptides that bind outside the canonical active site, modulating caspase activity in unpredictable ways. The majority of the work described herein has appeared since the publication of an excellent and comprehensive review of this topic [2].
Small molecules that bind at the dimer interface The identification of functional allosteric sites in C1, C3, and C7 by Wells and co-workers [7,8,11] has inspired
additional efforts to identify allosteric modulators of these and other caspases. While the motivation for such work varies, a central theme is to discover chemical matter with functional and/or physicochemical properties that are distinct from traditional active-site inhibitors. Thus, allosteric inhibitors of P6 self-activation were sought as leads for neurodegenerative diseases [12], while activators of apoptotic P3 were sought for their possible relevance in cancer therapy [13]. A variety of screening approaches have now been applied to find allosteric compounds, including high-throughput screening, computational docking, fragment screening by SPR, and disulfide trapping/cysteine tethering. Befitting the diversity of discovery approaches is the diversity of the chemical matter that has resulted from these efforts (Figure 3). From these recent studies it is apparent that high-affinity, reversible ligands can be identified for these sites, and that a variety of functional effects (activation, inhibition, protein stabilization) can result from binding. Wells and co-workers [14] recently described a disulfide exchange screen against C5 that produced the allosteric C5 inhibitor 4 (Figure 1). While compound 4 tethered to both C1 and C5, the compound selectively inhibited C5, as determined with the fluorogenic substrate Ac-WEHDAMC. Interestingly, several additional C1/C5-binding fragments showed no functional effects on C1 or C5 activity, demonstrating that binding alone is insufficient for functional effect. Although no structure of the C5compound 4 complex was obtained, stoichiometric tethering to the small subunit of both wild type C5 and a penta-C-to-A-mutant implicated Cys341 as the site of disulfide formation. This residue is analogous to Cys331 in C1, the site of tethering by the allosteric inhibitor 3 (Figure 2). It is worth noting that the ten
Figure 2
(a)
(b)
C285 C331
C331
C285 Current Opinion in Structural Biology
X-ray crystal structure of C1 in complex with compound 3. (a) Surface representation of C1 with the A and B subunit colored violet and light blue, respectively and two molecules of 3 bound at the dimer interface and represented as spheres and colored green and grey respectively. The compounds bind in a buried cavity at the dimer interface. (b) Close up of compound 3 binding site showing the proximity of the catalytic cysteine (C285) and the tether to C331. www.sciencedirect.com
Current Opinion in Structural Biology 2013, 23:812–819
814 Catalysis and regulation
Figure 3
F
OH H N
OH H N
N
N
self-activation [5]. In addition to screening for active-site inhibitors, the research team sought to explore alternate approaches to C6 inhibition. Modulation via the putative allosteric site was attractive as it could potentially provide lead compounds with better prospects for crossing the Blood-Brain Barrier (BBB) in animals (i.e., reversible, non-ionic compounds). A library of 2300 fragments was screened against both C6 and P6 using SPR methods. The screen of P6 was most fruitful, yielding several fragments that were found by X-ray crystallography to bind between Tyr198A/B residues at the dimer interface. These structures are the first to confirm a small-molecule binding site in P6/C6 analogous to that in C1/C3/C5/C7.
HO H N N
N N 5
6 N
O O S N
S
S
O O S N
7
N N NH Cu Cl S S 8 Current Opinion in Structural Biology
Structures of compounds reported to bind reversibly to the dimer interface in caspases.
C5-binding fragments described in the most recent report are all significantly larger than C1 inhibitor 3. That the allosteric site in C5 can accommodate these much larger compounds suggests that it may be significantly larger or more dynamic than the allosteric site in C1. The solution of X-ray structures for 4 and its congeners bound to C5 would be of significant interest as this might shed light on the C1/C5 selectivity of 4 and perhaps also the lack of inhibitory activity for the other C5 binders identified. A team comprised of researchers from UCSF and Genentech recently reported on their efforts to identify allosteric modulators of C6/P6 [12]. Recent studies have implicated C3-mediated and C6-mediated proteolysis events in the pathogenesis of Huntington’s disease and Alzheimer’s [15,16]. Biochemical and structural studies have further demonstrated that C6 is atypical among effector caspases in relying on a unique mechanism of
Further optimization using a ‘fragment merging’ strategy produced compounds such as 5 (Figure 3) that exhibited KD values in the mid-nM range. The X-ray structure of 5 reveals the pyrimidine ring engaged in a stacking interaction with Tyr198A/B while the benzylic side chain projects off the pyridine ring and into a hydrophobic grove formed by the aliphatic side chain of Glu214 (Figure 4). The fluorine atom in 5 is apparently engaged in a C–F–C O multipolar interaction with the backbone carbonyl of Ala195. The hydrophobic groove bound by 5 is duplicated on the other side of the binding site and symmetrical analogs like 6 were found to occupy both of these grooves. Interestingly, the allosteric site in P6 comprises residues on the L2 loop that are in turn linked to the site of intramolecular cleavage that activates C6. While compounds like 5 were found to stabilize P6 in a temperature-dependent aggregation (Tagg) assay, they did not bind to active C6. This latter finding may indicate that C6 does not sample a zymogen-like conformation once it is fully activated and lends support to the notion of stabilizing zymogen-specific conformations as a means of inhibiting C6 activation. Notably, the compounds described in this work are the first reversible nM-affinity ligands for the allosteric site in a caspase and should spur further work in the same direction.
Figure 4
(a)
(b) C163A Y198 Y198 C163A A195 Current Opinion in Structural Biology
X-ray crystal structure of procaspase-6 in complex with 5. (a) Crystal structure of P6 in complex with compound 5. (b) Key binding interactions of compound 5 include stacking interactions with Y198 residues on either side of the dimer interface as well as a multipolar interaction with A195. Current Opinion in Structural Biology 2013, 23:812–819
www.sciencedirect.com
Modulating caspase activity: beyond the active site Murray and Renslo 815
The symmetrical bis-thiophene analog 7 was recently reported [13] to be an allosteric activator of P3. Compound 7 was identified by computational docking against the allosteric site of C3, the researchers reasoning that compounds that bound to this site should activate P3 by promoting the adoption of an active conformation. The symmetrical nature of compound 7 and its proposed docking pose mimics some aspects of the symmetrical, pairwise binding of sulfide analogs 1 and 2 in the allosteric site of C3. To evaluate their docking hits, the team employed a triple D-to-A mutant of P3 that is unable to become fully activated, but can presumably adopt active conformations with augmented catalytic activity against fluorogenic substrates. Of ten analogs examined, only compound 7 demonstrated a significant effect on P3 activity, affording 25-fold activation when employed at high mM concentrations. Compound 7 binds active C3 with a KD in the low mM range as determined by ITC, but binding of compound 7 to P3 could not be detected by this same approach. This result is echoed in a recent study [6] of conformation-specific antibodies shown to bind with low-nM affinity for DEVD-bound P3 (‘active’ conformation) but only mM affinity for apo P3 (zymogen conformation). Despite possessing nM-affinity for the active conformer, much higher (mM) concentrations of the antibody were required to stimulate catalytic activity in P3 (triple D-to-A mutant), suggesting a high energetic barrier for P3 activation. Also of note from this study is the advisable recommendation that sub-stoichiometric quantities of an irreversible active-site inhibitor (e.g., 10% Ac-DEVD-CMK) be employed in activation studies of procaspases with WT cleavage sites since the presence of even trace quantities of cleaved, active caspase can greatly complicate the interpretation of experimental results. Evidence for compound 7 binding at the allosteric site is based on an observed stoichiometry of one equivalent of 7 per C3 dimer (vs. two equivalents per dimer expected for an active-site inhibitor). Also, compound 7 failed to bind to a C3 construct bearing a mutation near the allosteric binding site (V266K). Overall, the authors propose that compound 7 binds P3 at the dimer interface, expelling the inter-subunit linker and stabilizing an active conformation of P3. Another very recent report [17] describes a structurally distinct class of small molecules that activate P3 and P7 in vitro at concentrations lower than are required for 7. These compounds were also found to induce apoptosis at low mM concentrations in several different cell lines, the exception being MCF-7 cells which lack active C3/P3. The notion of activating caspases with small molecules remains an intriguing one, though some caution is warranted given the lack of definitive structural information about the binding site of these activators, and the fact that two previously reported small-molecule activators of P3 were subsequently found to act via indirect mechanisms [18–20]. www.sciencedirect.com
An in vitro high-throughput screen designed to recapitulate the intrinsic, mitochondrial-mediated apoptosis pathway identified the copper complex 8 and related complexes as caspase inhibitors [21]. Several factors aside from their unusual organometallic composition are worth commenting on. First, the complexes appear to have pan-caspase activity, inhibiting both inflammatory and apoptotic caspases. In cells, 100 nM of compound 8 was able to block apoptosis following induction of either the intrinsic or extrinsic apoptosis pathway. The compound also decreased secretion of IL1b in J774 cells treated with lipopolysaccharide, a finding consistent with inhibition of C1. The pan-caspase activity of compound 8 is attributed to binding at the dimer interface, although the finding that 8 also inhibits unrelated cysteine proteases such as cathepsin C and papain suggests that multiple mechanisms of inhibition are likely for these compounds. Thus, some caution is warranted in interpreting the results of cellular assays with these complexes. Structural evidence is provided for binding of 8 to the dimer interface of C7. The 3.8 A˚ crystal structure of C7 complex reveals two equivalents of 8 bound in a central cavity at the C7 dimer interface (Figure 5). The low resolution of this structure complicates interpretation of the interactions the copper complex makes with the C7 central cavity. In comparison to the crystal structures of C7 bound to thiols 1 or 2, it appears that the binding of 8 results in increased conformational freedom of the loops at the dimer interface and the catalytic loops. The catalytic cysteine (Cys186) has been built in this crystallographic model but the residues that follow the Cys186 are not present in the model. It should be noted that at a resolution of 3.8 A˚, it is not clear whether the increased disorder is a result of the compound or the low resolution of the data, even though the spacegroups and unit cell dimensions of all three complex structures are the same. Dialysis experiments and inspection of the C7-8 X-ray structure indicate that compound 8 acts reversibly and does not form a covalent bond with the cysteine residue (C290) that tethers to 1 and 2; instead, the structure shows that each molecule of compound 8 sits above these two cysteine residues at an approximate distance of 2.8 A˚ from the Sg atom of C290 to the copper atom of the compound. Surprisingly, inhibition of C7 by complex 8 was found to be non-competitive; binding of 8 lowers Vmax but does not significantly affect the ability of C7 to bind substrate (Km). This is in contrast to thiols 1–3, the binding of which was shown to be mutually exclusive with substrate binding, resulting in effectively competitive inhibition.
Other approaches to modulating caspases The implication of C6 as a mediator of neurodegeneration has resulted in extensive efforts to identify inhibitors and probes of C6 activity [22,23]. In an approach complementary to the P6-targeted SPR screen described Current Opinion in Structural Biology 2013, 23:812–819
816 Catalysis and regulation
Figure 5
(a)
(b)
C186
C290
C290
C186 Current Opinion in Structural Biology
X-ray crystal structure of C7 in complex with 8. (a) Surface representation of C7 with the A and B subunit colored olive green and violet blue, respectively. Two molecules of 8 bound at the dimer interface are represented as spheres and colored green and cyan respectively (b) Close-up of the compound 8 binding site showing the proximity of the catalytic cysteine (C186) and C290.
above, researchers at Genentech sought to target C6/P6 using phage display techniques. The goal of this effort was to identify peptides that would bind to and stabilize a zymogen form of the enzyme. Four peptides were identified, one linear and three cyclic, that bound to P6 at single-digit micromolar concentration in a phage-competition binding ELISA [23]. After affinity maturation of one linear and one cyclic peptide, the researchers demonstrated that both types of peptides also inhibited the ability of the active form of the enzyme to cleave a fluorescent tetrapeptide substrate (VEID-AMC) with IC50 values of 8 mM and 4 mM for the linear and cyclic peptides, respectively. Interestingly, the linear peptide, (pep419), displayed non-competitive inhibition of the active form of C6 and enhanced the ability of both C6 and the P6 homodimer to tetramerize (unpublished data). The mechanism of action of these peptides is well explained by the co-crystal structure of procaspase-6 in complex with the linear peptide, pep419 (Figure 6). The structure shows that pep419 forms a disulfide-linked antiparallel b-strand peptide dimer that binds to a site on procaspase-6 that is remote from the active site. The majority of the interactions between pep419 and P6 are mediated by main chain atoms, however site-specific mutation made to either pep419 or the pep419 binding site on P6 suggest that Glu12 of pep419 and H126 on P6 mediate key interactions (Figure 6). Unpublished data indicate that procaspase-6 exists as a homotetramer under physiological conditions and pep419 binds to a previously uncharacterized homotetramerization interface. Thus, pep419 stabilizes the homotetramer form of P6 under conditions that disfavor homotetramerization and also Current Opinion in Structural Biology 2013, 23:812–819
induces homotetramerization of the active form of C6. These results suggest a novel mechanism of inhibition of caspases-6 via tetramer sequestration. The same UCSF-Genentech team that reported the dimer interface binder 5 (Figure 4) also recently reported [23] on a mechanistically novel inhibitor of active C6 (compound 9, Figure 7). The discovery of 9 began with a high-throughput screen against C6 using the divalent tetrapeptide substrate (VEID)2-Rho110. A hit compound comprising phenyalanine with an N-terminal aryl furan capping group became the subject of a chemical optimization campaign, ultimately yielding compound 9 which inhibited C6 in the screening assay with an IC50 in the low nanomolar range. Compound 9 is noteworthy for its exquisite selectivity for C6 and the lack of an electrophilic warhead or an aspartate mimic — both hallmarks of classic active-site directed caspase inhibitors. Although assumed initially to be an active-site inhibitor, further study of 9 revealed that the compound in fact exhibits an uncompetitive mechanism of C6 inhibition, with its potency dependent on both the nature and concentration of the tetrapeptide–fluorophore substrate. The compound does not simply sequester substrate however, since other caspases (C3 and C7) were uninhibited by 9, even when the same (VEID)2–R110 substrate was employed in the assay. The solution of an X-ray crystal structure of 9 bound to a zVEID-inhibited form of C6 revealed the unprecedented binding mode of this compound. In this structure, compound 9 (in purple, Figure 7) binds in a shallow groove formed by the backbone of the C6-bound tetrapeptide www.sciencedirect.com
Modulating caspase activity: beyond the active site Murray and Renslo 817
Figure 6
(a)
(b)
(d)
(c)
H126
4 Bits
3 2
E9 18
17
16
15
14
12 13
11
9
10
8
7
6
5
4
3
2
0 N
1
1 C
Current Opinion in Structural Biology
Crystal structure of P6 tetramer in complex with pep419. (a) P6 tetramer with chains A–D colored green, cyan, yellow and pink respectively. The four pep419 peptides are represented as cartoons and colored violet, magenta, dark green and orange. (b) Open book view of the P6 dimer showing the binding site of pep419 is distant from the dimer interface and the active site Cys (colored red). (c) pep419 consensus logo indicating the relative importance of each residue by relative font size. (d) Close up of the pep419 binding site showing the interaction between Glu12 and His126. H-bond interaction is represented as black dashes.
Figure 7
(a)
(b)
O
NC O
C163
O O
NH HO 9
Current Opinion in Structural Biology
X-ray crystal structure of compound 9 in complex with C6 that has been covalently modified with a zVEID suicide substrate. (a) Surface representation of the C6 dimer showing the covalent complex with zVEID as pink sticks and the uncompetitive inhibitor as purple sticks. (b) Close up of the uncompetitive inhibitor binding site. The covalent bond between the active site cysteine (C163) and zVEID substrate is shown. Residues involved in interactions are shown as sticks, H-bond interactions are shown as black dashes. www.sciencedirect.com
Current Opinion in Structural Biology 2013, 23:812–819
818 Catalysis and regulation
substrate (in pink). Compound 9 makes a hydrogen bond to the backbone Ile NH of the VEID substrate while also participating in a water mediated hydrogen bond to R220 of the L3 loop of C6. Thus, it is the specific conformation of VEID substrate formed on binding to C6 that constitutes the binding site of compound 9, while direct interactions between 9 and C6 active-site residues are also important. Computational studies employing a model of the VEID–R110–C6 complex furthermore suggested a role for the rhodamine dye itself in the binding affinity of 9 in the ternary complex. It is unlikely that uncompetitive inhibitors like 9 would be relevant in a therapeutic setting because the action of 9 is dependent on an artificial substrate not present in cellular contexts. However, this work does suggest the intriguing possibility of developing uncompetitive inhibitors that recognize more biologically relevant caspase–substrate complexes. Inhibition of caspases by this mechanism would presumably be highly specific and the chemical matter required would not require the electrophilic and anionic groups that have dogged development of classic active-site directed inhibitors.
References and recommended reading Papers of particular interest, published within the period of review, have been highlighted as: of special interest of outstanding interest 1.
McIlwain DR, Berger T, Mak TW: Caspase functions in cell death and disease. Cold Spring Harb Perspect Biol 2013, 5:a0086.
2.
Hacker HG, Sisay MT, Gutschow M: Allosteric modulation of caspases. Pharmacol Ther 2011, 132:180-195.
3.
MacKenzie SH, Schipper JL, Clark AC: The potential for caspases in drug discovery. Curr Opin Drug Discov Dev 2010, 13:568-576.
4.
Garcia-Calvo M, Peterson EP, Leiting B, Ruel R, Nicholson DW, Thornberry NA: Inhibition of human caspases by peptide-based and macromolecular inhibitors. J Biol Chem 1998, 273:32608-33261.
5.
Wang XJ, Cao Q, Liu X, Wang KT, Mi W, Zhang Y, Li LF, LeBlanc AC, Su XD: Crystal structures of human caspase 6 reveal a new mechanism for intramolecular cleavage selfactivation. EMBO Rep 2010, 11:841-847.
6.
Thomsen ND, Koerber JT, Wells JA: Structural snapshots reveal distinct mechanisms of procaspase-3 and -7 activation. Proc Natl Acad Sci USA 2013, 110:8477-8482. This paper presents several remarkable structures of caspases in various states of activation. The biochemical activity of procaspase-3 and -7 are studied, and distinct activation mechanisms are proposed for these important executioner caspases.
7.
Hardy JA, Wells JA: Searching for new allosteric sites in enzymes. Curr Opin Struct Biol 2004, 14:706-715.
8.
Scheer JM, Romanowski MJ, Wells JA: A common allosteric site and mechanism in caspases. Proc Natl Acad Sci USA 2006, 103:7595-7600.
9.
Datta D, Scheer JM, Romanowski MJ, Wells JA: An allosteric circuit in caspase-1. J Mol Biol 2008, 381:1157-1167.
Conclusions A decade has elapsed since disulfide-based fragment screening first identified an orphan allosteric site in C3/C7. In the intervening years, significant progress has been made in characterizing this site in several additional caspase family members, though endogenous ligands for these sites remain to be identified. Recent studies have afforded a better understanding of the allosteric circuitry that links this site to the active site, and a more subtle understanding of the various active and inactive conformations that can be adopted by both processed and unprocessed caspases. Much of this work is referenced or described herein, an underlying theme of which is the structural diversity and conformationally dynamic nature of this family of proteases. Proteins, peptides, and small molecules that interact with these various states are being identified at an increasing rate, and are affording surprising and unexpected insights into caspase modulation. At the same time, interest in caspases as therapeutic targets has only increased as new roles for these proteases in development and neurodegeneration have been proposed. Recent advances in computational and biophysical screening technology have enabled the application of new screening approaches and the result has been a variety of new chemical matter, as described herein. These molecules point to new avenues for caspase modulation but remain immature from a drug development perspective. Only a robust and sustained effort to improve their potencies and in vivo properties will produce molecules that can truly test the therapeutic potential of caspase modulation by allosteric mechanisms. Current Opinion in Structural Biology 2013, 23:812–819
10. Datta D, McClendon CL, Jacobson MP, Wells JA: Substrate and inhibitor-induced dimerization and cooperativity in caspase-1 but not caspase-3. J Biol Chem 2013, 288:9971-9981. 11. Hardy JA, Lam J, Nguyen JT, O’Brien T, Wells JA: Discovery of an allosteric site in the caspases. Proc Natl Acad Sci USA 2004, 101:12461-12466. 12. Murray J, Giannetti AM, Steffek M, Gibbons P, Hearn BR, Cohen F, Tam C, Pozniak C, Bravo B, Lewcock JW et al.: Tailoring Small Molecules for an Allosteric Site on Procaspse-6. ChemMedChem 2013. in press. This paper describes the discovery of small molecules that bind the dimer-interface site in procaspase-6 with nM affinity. This work demonstrates that high-affinity, non-covalent ligands can be identified for the dimer interface of caspases. 13. Schipper JL, MacKenzie SH, Sharma A, Clark AC: A bifunctional allosteric site in the dimer interface of procaspase-3. Biophys Chem 2011, 159:100-109. 14. Gao J, Wells JA: Identification of specific tethered inhibitors for caspase-5. Chem Biol Drug Des 2012, 79:209-215. 15. Simon DJ, Weimer RM, McLaughlin T, Kallop D, Stanger K, Yang J, O’Leary DD, Hannoush RN, Tessier-Lavigne M: A caspase cascade regulating developmental axon degeneration. J Neurosci 2012, 32:17540-17553. 16. LeBlanc AC: Caspase-6 as a novel early target in the treatment of Alzheimer’s disease. Eur J Neurosci 2013, 37:2005-2018. 17. Vickers CJ, Gonzalez-Paez GE, Umotoy JC, Cayanan-Garrett C, Brown SJ, Wolan DW: Small-molecule procaspase activators identified using fluorescence polarization. Chembiochem 2013, 14:1419-1422. 18. Zorn JA, Wille H, Wolan DW, Wells JA: Self-assembling small molecules form nanofibrils that bind procaspase-3 to promote activation. J Am Chem Soc 2011, 133:19630-19633. www.sciencedirect.com
Modulating caspase activity: beyond the active site Murray and Renslo 819
19. Zorn JA, Wolan DW, Agard NJ, Wells JA: Fibrils colocalize caspase-3 with procaspase-3 to foster maturation. J Biol Chem 2012 2012, 287:33781-33795. 20. Peterson QP, Goode DR, West DC, Ramsey KN, Lee JJ, Hergenrother PJ: PAC-1 activates procaspase-3 in vitro through relief of zinc-mediated inhibition. J Mol Biol 2009, 388:144-158.
22. Stanger K, Steffek M, Zhou L, Pozniak CD, Quan C, Franke Y, Tom J, Tam C, Krylova I, Elliott JM et al.: Allosteric peptides bind a caspase zymogen and mediate caspase tetramerization. Nat Chem Biol 2012, 8:655-660. Using phage-display, these workers identified a peptide that binds a previously uncharacterized tetramerization interface in procaspase-6. Interestingly, this peptide can induce tetramerization of active caspase-6, inhibiting the protease non-competitively.
21. Feldman T, Kabaleeswaran V, Jang SB, Antczak C, Djaballah H, Wu H, Jiang X: A class of allosteric caspase inhibitors identified by high-throughput screening. Mol Cell 2012, 47:585-595.
23. Heise CE, Murray J, Augustyn KE, Bravo B, Chugha P, Cohen F, Giannetti AM, Gibbons P, Hannoush RN, Hearn BR et al.: Mechanistic and structural understanding of uncompetitive inhibitors of caspase-6. PLoS ONE 2012, 7:e45086.
www.sciencedirect.com
Current Opinion in Structural Biology 2013, 23:812–819
ScienceDirect Modulating caspase activity: beyond the active site Jeremy Murray1 and Adam R Renslo2 Caspases are a family of aspartate-specific cysteine proteases that regulate cellular homeostasis through the mediation of apoptosis and inflammation. Despite keen interest in caspases as therapeutic targets for cancer, inflammatory, and neurodegenerative diseases, no active-site directed small molecule has yet succeeded in navigating human clinical trials. At the same time, recent biochemical and biophysical studies have revealed caspases to be highly dynamic proteases possessing a remarkable diversity of activation mechanisms. In addition, many caspases possess an allosteric circuit linking key active site loops with a distal allosteric site located at the dimer interface. Accordingly, small molecule binding at this allosteric site directly impacts structural organization of the active site and thus catalytic activity. Both cysteine-tethered and non-covalent reversible small molecules have recently been identified for these allosteric sites, with binding producing a variety of functional effects. Surprising new examples of caspase modulation have also been described recently, including a small molecule that binds caspase-6–substrate complexes uncompetitively and a short peptide that stabilizes an inactive, tetrameric form of procaspase-6. The confluence of recent biochemical, biophysical and pharmacological data has revealed exciting new avenues for the modulation of caspase activity via binding beyond the active site. Addresses 1 Department of Structural Biology, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080, United States 2 Small Molecule Discovery Center and Department of Pharmaceutical Chemistry, University of California, San Francisco, 1700 4th Street, San Francisco, CA 94158, United States Corresponding authors: Renslo, Adam R ([email protected], [email protected])
Current Opinion in Structural Biology 2013, 23:812–819 This review comes from a themed issue on Catalysis and regulation Edited by Ben M Dunn and Alexander Wlodawer For a complete overview see the Issue and the Editorial Available online 8th November 2013 0959-440X/$ – see front matter, # 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.sbi.2013.10.002
Introduction The modulation of caspases as a therapeutic strategy has attracted significant interest given the important roles played by caspases in apoptosis, inflammation, and neurodegeneration [1–3]. The substrate preferences of specific caspases have been known for some time and this has enabled the development of relatively selective Current Opinion in Structural Biology 2013, 23:812–819
inhibitors comprising a tetrapeptide with an electrophilic warhead function [4]. However, the presence in such inhibitors of both electrophilic functionality and an aspartate residue (or anionic surrogate) makes them poor lead compounds from a drug development perspective. Indeed, not a single active-site directed caspase inhibitor has received regulatory approval, and most of the compounds that have entered clinical trials have had their development terminated [3]. These costly failures have encouraged the exploration of alternative approaches for modulating caspase activity. Caspases are a highly dynamic class of proteases that exhibit a diverse array of activation mechanisms that differ according to whether a caspases is at the terminal step in the pathway (effector caspase) or initiates the signaling cascade (initiator caspases). Hence, the apoptotic ‘effector’ caspase-3 (C3) and caspase-7 (C7) are expressed as dimeric procaspases (P3, P7) that become activated on proteolytic cleavage by ‘initiator’ caspases such as C8 and C9. The initiator caspases by contrast are expressed as monomeric proenzymes that are activated upon assembly into larger complexes (e.g., the ‘apoptosome’ in the case of C9). Recent evidence suggests that C6 is unique among effector caspases in being dependent upon self-proteolysis for activation, with the intra-subunit cleavage sequence 190 TEVD193 positioned in the active site of P6, ready for activation [5]. Even the categorization of caspase forms as ‘active’ or ‘inactive’ is surprisingly complex in light of recent structural and biochemical findings that the zymogen and processed forms of C3 and C7 can each adopt either active or inactive conformations [6]. The dynamic nature of caspase structure in effect affords the drug discovery scientist with multiple ‘targets’ for intervention with small molecules. The first small molecules shown to target an inactive caspase conformation were compounds 1 and 2 (Figure 1), which bind an allosteric site at the dimer interface of either C3 or C7, forming disulfide bonds with native cysteine residues present at these sites [7]. Significantly, disulfide-bound 1 or 2 were found to trap the enzyme in a zymogen-like conformation, inhibiting the enzyme by an effectively competitive mode of inhibition (binding of 1 or 2 was mutually exclusive with substrate binding at the active site). Subsequently, an analogous disulfide-tethered inhibitor (3, Figure 1) was identified for C1 [8]. The X-ray structure of 3 bound to C1 (Figure 2) reveals two molecules of 3 (one for each small sub-unit) engaged in a stacking interaction that straddles the dimer interface. Finally, residues comprising an allosteric ‘circuit’ between key active site loops and the dimer interface have been www.sciencedirect.com
Modulating caspase activity: beyond the active site Murray and Renslo 813
Figure 1 Cl
SH F
HN N H
SH Cl
O
O
O 2 (DICA)
1 (FICA) F3C
S
O SH
N N
S
HN
O
N HN
HN
SH
4
3
Current Opinion in Structural Biology
Thiol-containing inhibitors 1 and 2 inhibit executioner caspases C3 and C7 allosterically. Thiols 3 and 4 allosterically inhibit inflammatory caspases C1 and C5 respectively.
identified for several caspases, helping to explain the mutually exclusive binding of allosteric and active-site directed molecules where compounds that bind at the dimer disrupt the hydrogen bond network that is essential for stabilization of the active form [9,10]. In the sections that follow, we will summarize recent reports of smallmolecules and peptides that bind outside the canonical active site, modulating caspase activity in unpredictable ways. The majority of the work described herein has appeared since the publication of an excellent and comprehensive review of this topic [2].
Small molecules that bind at the dimer interface The identification of functional allosteric sites in C1, C3, and C7 by Wells and co-workers [7,8,11] has inspired
additional efforts to identify allosteric modulators of these and other caspases. While the motivation for such work varies, a central theme is to discover chemical matter with functional and/or physicochemical properties that are distinct from traditional active-site inhibitors. Thus, allosteric inhibitors of P6 self-activation were sought as leads for neurodegenerative diseases [12], while activators of apoptotic P3 were sought for their possible relevance in cancer therapy [13]. A variety of screening approaches have now been applied to find allosteric compounds, including high-throughput screening, computational docking, fragment screening by SPR, and disulfide trapping/cysteine tethering. Befitting the diversity of discovery approaches is the diversity of the chemical matter that has resulted from these efforts (Figure 3). From these recent studies it is apparent that high-affinity, reversible ligands can be identified for these sites, and that a variety of functional effects (activation, inhibition, protein stabilization) can result from binding. Wells and co-workers [14] recently described a disulfide exchange screen against C5 that produced the allosteric C5 inhibitor 4 (Figure 1). While compound 4 tethered to both C1 and C5, the compound selectively inhibited C5, as determined with the fluorogenic substrate Ac-WEHDAMC. Interestingly, several additional C1/C5-binding fragments showed no functional effects on C1 or C5 activity, demonstrating that binding alone is insufficient for functional effect. Although no structure of the C5compound 4 complex was obtained, stoichiometric tethering to the small subunit of both wild type C5 and a penta-C-to-A-mutant implicated Cys341 as the site of disulfide formation. This residue is analogous to Cys331 in C1, the site of tethering by the allosteric inhibitor 3 (Figure 2). It is worth noting that the ten
Figure 2
(a)
(b)
C285 C331
C331
C285 Current Opinion in Structural Biology
X-ray crystal structure of C1 in complex with compound 3. (a) Surface representation of C1 with the A and B subunit colored violet and light blue, respectively and two molecules of 3 bound at the dimer interface and represented as spheres and colored green and grey respectively. The compounds bind in a buried cavity at the dimer interface. (b) Close up of compound 3 binding site showing the proximity of the catalytic cysteine (C285) and the tether to C331. www.sciencedirect.com
Current Opinion in Structural Biology 2013, 23:812–819
814 Catalysis and regulation
Figure 3
F
OH H N
OH H N
N
N
self-activation [5]. In addition to screening for active-site inhibitors, the research team sought to explore alternate approaches to C6 inhibition. Modulation via the putative allosteric site was attractive as it could potentially provide lead compounds with better prospects for crossing the Blood-Brain Barrier (BBB) in animals (i.e., reversible, non-ionic compounds). A library of 2300 fragments was screened against both C6 and P6 using SPR methods. The screen of P6 was most fruitful, yielding several fragments that were found by X-ray crystallography to bind between Tyr198A/B residues at the dimer interface. These structures are the first to confirm a small-molecule binding site in P6/C6 analogous to that in C1/C3/C5/C7.
HO H N N
N N 5
6 N
O O S N
S
S
O O S N
7
N N NH Cu Cl S S 8 Current Opinion in Structural Biology
Structures of compounds reported to bind reversibly to the dimer interface in caspases.
C5-binding fragments described in the most recent report are all significantly larger than C1 inhibitor 3. That the allosteric site in C5 can accommodate these much larger compounds suggests that it may be significantly larger or more dynamic than the allosteric site in C1. The solution of X-ray structures for 4 and its congeners bound to C5 would be of significant interest as this might shed light on the C1/C5 selectivity of 4 and perhaps also the lack of inhibitory activity for the other C5 binders identified. A team comprised of researchers from UCSF and Genentech recently reported on their efforts to identify allosteric modulators of C6/P6 [12]. Recent studies have implicated C3-mediated and C6-mediated proteolysis events in the pathogenesis of Huntington’s disease and Alzheimer’s [15,16]. Biochemical and structural studies have further demonstrated that C6 is atypical among effector caspases in relying on a unique mechanism of
Further optimization using a ‘fragment merging’ strategy produced compounds such as 5 (Figure 3) that exhibited KD values in the mid-nM range. The X-ray structure of 5 reveals the pyrimidine ring engaged in a stacking interaction with Tyr198A/B while the benzylic side chain projects off the pyridine ring and into a hydrophobic grove formed by the aliphatic side chain of Glu214 (Figure 4). The fluorine atom in 5 is apparently engaged in a C–F–C O multipolar interaction with the backbone carbonyl of Ala195. The hydrophobic groove bound by 5 is duplicated on the other side of the binding site and symmetrical analogs like 6 were found to occupy both of these grooves. Interestingly, the allosteric site in P6 comprises residues on the L2 loop that are in turn linked to the site of intramolecular cleavage that activates C6. While compounds like 5 were found to stabilize P6 in a temperature-dependent aggregation (Tagg) assay, they did not bind to active C6. This latter finding may indicate that C6 does not sample a zymogen-like conformation once it is fully activated and lends support to the notion of stabilizing zymogen-specific conformations as a means of inhibiting C6 activation. Notably, the compounds described in this work are the first reversible nM-affinity ligands for the allosteric site in a caspase and should spur further work in the same direction.
Figure 4
(a)
(b) C163A Y198 Y198 C163A A195 Current Opinion in Structural Biology
X-ray crystal structure of procaspase-6 in complex with 5. (a) Crystal structure of P6 in complex with compound 5. (b) Key binding interactions of compound 5 include stacking interactions with Y198 residues on either side of the dimer interface as well as a multipolar interaction with A195. Current Opinion in Structural Biology 2013, 23:812–819
www.sciencedirect.com
Modulating caspase activity: beyond the active site Murray and Renslo 815
The symmetrical bis-thiophene analog 7 was recently reported [13] to be an allosteric activator of P3. Compound 7 was identified by computational docking against the allosteric site of C3, the researchers reasoning that compounds that bound to this site should activate P3 by promoting the adoption of an active conformation. The symmetrical nature of compound 7 and its proposed docking pose mimics some aspects of the symmetrical, pairwise binding of sulfide analogs 1 and 2 in the allosteric site of C3. To evaluate their docking hits, the team employed a triple D-to-A mutant of P3 that is unable to become fully activated, but can presumably adopt active conformations with augmented catalytic activity against fluorogenic substrates. Of ten analogs examined, only compound 7 demonstrated a significant effect on P3 activity, affording 25-fold activation when employed at high mM concentrations. Compound 7 binds active C3 with a KD in the low mM range as determined by ITC, but binding of compound 7 to P3 could not be detected by this same approach. This result is echoed in a recent study [6] of conformation-specific antibodies shown to bind with low-nM affinity for DEVD-bound P3 (‘active’ conformation) but only mM affinity for apo P3 (zymogen conformation). Despite possessing nM-affinity for the active conformer, much higher (mM) concentrations of the antibody were required to stimulate catalytic activity in P3 (triple D-to-A mutant), suggesting a high energetic barrier for P3 activation. Also of note from this study is the advisable recommendation that sub-stoichiometric quantities of an irreversible active-site inhibitor (e.g., 10% Ac-DEVD-CMK) be employed in activation studies of procaspases with WT cleavage sites since the presence of even trace quantities of cleaved, active caspase can greatly complicate the interpretation of experimental results. Evidence for compound 7 binding at the allosteric site is based on an observed stoichiometry of one equivalent of 7 per C3 dimer (vs. two equivalents per dimer expected for an active-site inhibitor). Also, compound 7 failed to bind to a C3 construct bearing a mutation near the allosteric binding site (V266K). Overall, the authors propose that compound 7 binds P3 at the dimer interface, expelling the inter-subunit linker and stabilizing an active conformation of P3. Another very recent report [17] describes a structurally distinct class of small molecules that activate P3 and P7 in vitro at concentrations lower than are required for 7. These compounds were also found to induce apoptosis at low mM concentrations in several different cell lines, the exception being MCF-7 cells which lack active C3/P3. The notion of activating caspases with small molecules remains an intriguing one, though some caution is warranted given the lack of definitive structural information about the binding site of these activators, and the fact that two previously reported small-molecule activators of P3 were subsequently found to act via indirect mechanisms [18–20]. www.sciencedirect.com
An in vitro high-throughput screen designed to recapitulate the intrinsic, mitochondrial-mediated apoptosis pathway identified the copper complex 8 and related complexes as caspase inhibitors [21]. Several factors aside from their unusual organometallic composition are worth commenting on. First, the complexes appear to have pan-caspase activity, inhibiting both inflammatory and apoptotic caspases. In cells, 100 nM of compound 8 was able to block apoptosis following induction of either the intrinsic or extrinsic apoptosis pathway. The compound also decreased secretion of IL1b in J774 cells treated with lipopolysaccharide, a finding consistent with inhibition of C1. The pan-caspase activity of compound 8 is attributed to binding at the dimer interface, although the finding that 8 also inhibits unrelated cysteine proteases such as cathepsin C and papain suggests that multiple mechanisms of inhibition are likely for these compounds. Thus, some caution is warranted in interpreting the results of cellular assays with these complexes. Structural evidence is provided for binding of 8 to the dimer interface of C7. The 3.8 A˚ crystal structure of C7 complex reveals two equivalents of 8 bound in a central cavity at the C7 dimer interface (Figure 5). The low resolution of this structure complicates interpretation of the interactions the copper complex makes with the C7 central cavity. In comparison to the crystal structures of C7 bound to thiols 1 or 2, it appears that the binding of 8 results in increased conformational freedom of the loops at the dimer interface and the catalytic loops. The catalytic cysteine (Cys186) has been built in this crystallographic model but the residues that follow the Cys186 are not present in the model. It should be noted that at a resolution of 3.8 A˚, it is not clear whether the increased disorder is a result of the compound or the low resolution of the data, even though the spacegroups and unit cell dimensions of all three complex structures are the same. Dialysis experiments and inspection of the C7-8 X-ray structure indicate that compound 8 acts reversibly and does not form a covalent bond with the cysteine residue (C290) that tethers to 1 and 2; instead, the structure shows that each molecule of compound 8 sits above these two cysteine residues at an approximate distance of 2.8 A˚ from the Sg atom of C290 to the copper atom of the compound. Surprisingly, inhibition of C7 by complex 8 was found to be non-competitive; binding of 8 lowers Vmax but does not significantly affect the ability of C7 to bind substrate (Km). This is in contrast to thiols 1–3, the binding of which was shown to be mutually exclusive with substrate binding, resulting in effectively competitive inhibition.
Other approaches to modulating caspases The implication of C6 as a mediator of neurodegeneration has resulted in extensive efforts to identify inhibitors and probes of C6 activity [22,23]. In an approach complementary to the P6-targeted SPR screen described Current Opinion in Structural Biology 2013, 23:812–819
816 Catalysis and regulation
Figure 5
(a)
(b)
C186
C290
C290
C186 Current Opinion in Structural Biology
X-ray crystal structure of C7 in complex with 8. (a) Surface representation of C7 with the A and B subunit colored olive green and violet blue, respectively. Two molecules of 8 bound at the dimer interface are represented as spheres and colored green and cyan respectively (b) Close-up of the compound 8 binding site showing the proximity of the catalytic cysteine (C186) and C290.
above, researchers at Genentech sought to target C6/P6 using phage display techniques. The goal of this effort was to identify peptides that would bind to and stabilize a zymogen form of the enzyme. Four peptides were identified, one linear and three cyclic, that bound to P6 at single-digit micromolar concentration in a phage-competition binding ELISA [23]. After affinity maturation of one linear and one cyclic peptide, the researchers demonstrated that both types of peptides also inhibited the ability of the active form of the enzyme to cleave a fluorescent tetrapeptide substrate (VEID-AMC) with IC50 values of 8 mM and 4 mM for the linear and cyclic peptides, respectively. Interestingly, the linear peptide, (pep419), displayed non-competitive inhibition of the active form of C6 and enhanced the ability of both C6 and the P6 homodimer to tetramerize (unpublished data). The mechanism of action of these peptides is well explained by the co-crystal structure of procaspase-6 in complex with the linear peptide, pep419 (Figure 6). The structure shows that pep419 forms a disulfide-linked antiparallel b-strand peptide dimer that binds to a site on procaspase-6 that is remote from the active site. The majority of the interactions between pep419 and P6 are mediated by main chain atoms, however site-specific mutation made to either pep419 or the pep419 binding site on P6 suggest that Glu12 of pep419 and H126 on P6 mediate key interactions (Figure 6). Unpublished data indicate that procaspase-6 exists as a homotetramer under physiological conditions and pep419 binds to a previously uncharacterized homotetramerization interface. Thus, pep419 stabilizes the homotetramer form of P6 under conditions that disfavor homotetramerization and also Current Opinion in Structural Biology 2013, 23:812–819
induces homotetramerization of the active form of C6. These results suggest a novel mechanism of inhibition of caspases-6 via tetramer sequestration. The same UCSF-Genentech team that reported the dimer interface binder 5 (Figure 4) also recently reported [23] on a mechanistically novel inhibitor of active C6 (compound 9, Figure 7). The discovery of 9 began with a high-throughput screen against C6 using the divalent tetrapeptide substrate (VEID)2-Rho110. A hit compound comprising phenyalanine with an N-terminal aryl furan capping group became the subject of a chemical optimization campaign, ultimately yielding compound 9 which inhibited C6 in the screening assay with an IC50 in the low nanomolar range. Compound 9 is noteworthy for its exquisite selectivity for C6 and the lack of an electrophilic warhead or an aspartate mimic — both hallmarks of classic active-site directed caspase inhibitors. Although assumed initially to be an active-site inhibitor, further study of 9 revealed that the compound in fact exhibits an uncompetitive mechanism of C6 inhibition, with its potency dependent on both the nature and concentration of the tetrapeptide–fluorophore substrate. The compound does not simply sequester substrate however, since other caspases (C3 and C7) were uninhibited by 9, even when the same (VEID)2–R110 substrate was employed in the assay. The solution of an X-ray crystal structure of 9 bound to a zVEID-inhibited form of C6 revealed the unprecedented binding mode of this compound. In this structure, compound 9 (in purple, Figure 7) binds in a shallow groove formed by the backbone of the C6-bound tetrapeptide www.sciencedirect.com
Modulating caspase activity: beyond the active site Murray and Renslo 817
Figure 6
(a)
(b)
(d)
(c)
H126
4 Bits
3 2
E9 18
17
16
15
14
12 13
11
9
10
8
7
6
5
4
3
2
0 N
1
1 C
Current Opinion in Structural Biology
Crystal structure of P6 tetramer in complex with pep419. (a) P6 tetramer with chains A–D colored green, cyan, yellow and pink respectively. The four pep419 peptides are represented as cartoons and colored violet, magenta, dark green and orange. (b) Open book view of the P6 dimer showing the binding site of pep419 is distant from the dimer interface and the active site Cys (colored red). (c) pep419 consensus logo indicating the relative importance of each residue by relative font size. (d) Close up of the pep419 binding site showing the interaction between Glu12 and His126. H-bond interaction is represented as black dashes.
Figure 7
(a)
(b)
O
NC O
C163
O O
NH HO 9
Current Opinion in Structural Biology
X-ray crystal structure of compound 9 in complex with C6 that has been covalently modified with a zVEID suicide substrate. (a) Surface representation of the C6 dimer showing the covalent complex with zVEID as pink sticks and the uncompetitive inhibitor as purple sticks. (b) Close up of the uncompetitive inhibitor binding site. The covalent bond between the active site cysteine (C163) and zVEID substrate is shown. Residues involved in interactions are shown as sticks, H-bond interactions are shown as black dashes. www.sciencedirect.com
Current Opinion in Structural Biology 2013, 23:812–819
818 Catalysis and regulation
substrate (in pink). Compound 9 makes a hydrogen bond to the backbone Ile NH of the VEID substrate while also participating in a water mediated hydrogen bond to R220 of the L3 loop of C6. Thus, it is the specific conformation of VEID substrate formed on binding to C6 that constitutes the binding site of compound 9, while direct interactions between 9 and C6 active-site residues are also important. Computational studies employing a model of the VEID–R110–C6 complex furthermore suggested a role for the rhodamine dye itself in the binding affinity of 9 in the ternary complex. It is unlikely that uncompetitive inhibitors like 9 would be relevant in a therapeutic setting because the action of 9 is dependent on an artificial substrate not present in cellular contexts. However, this work does suggest the intriguing possibility of developing uncompetitive inhibitors that recognize more biologically relevant caspase–substrate complexes. Inhibition of caspases by this mechanism would presumably be highly specific and the chemical matter required would not require the electrophilic and anionic groups that have dogged development of classic active-site directed inhibitors.
References and recommended reading Papers of particular interest, published within the period of review, have been highlighted as: of special interest of outstanding interest 1.
McIlwain DR, Berger T, Mak TW: Caspase functions in cell death and disease. Cold Spring Harb Perspect Biol 2013, 5:a0086.
2.
Hacker HG, Sisay MT, Gutschow M: Allosteric modulation of caspases. Pharmacol Ther 2011, 132:180-195.
3.
MacKenzie SH, Schipper JL, Clark AC: The potential for caspases in drug discovery. Curr Opin Drug Discov Dev 2010, 13:568-576.
4.
Garcia-Calvo M, Peterson EP, Leiting B, Ruel R, Nicholson DW, Thornberry NA: Inhibition of human caspases by peptide-based and macromolecular inhibitors. J Biol Chem 1998, 273:32608-33261.
5.
Wang XJ, Cao Q, Liu X, Wang KT, Mi W, Zhang Y, Li LF, LeBlanc AC, Su XD: Crystal structures of human caspase 6 reveal a new mechanism for intramolecular cleavage selfactivation. EMBO Rep 2010, 11:841-847.
6.
Thomsen ND, Koerber JT, Wells JA: Structural snapshots reveal distinct mechanisms of procaspase-3 and -7 activation. Proc Natl Acad Sci USA 2013, 110:8477-8482. This paper presents several remarkable structures of caspases in various states of activation. The biochemical activity of procaspase-3 and -7 are studied, and distinct activation mechanisms are proposed for these important executioner caspases.
7.
Hardy JA, Wells JA: Searching for new allosteric sites in enzymes. Curr Opin Struct Biol 2004, 14:706-715.
8.
Scheer JM, Romanowski MJ, Wells JA: A common allosteric site and mechanism in caspases. Proc Natl Acad Sci USA 2006, 103:7595-7600.
9.
Datta D, Scheer JM, Romanowski MJ, Wells JA: An allosteric circuit in caspase-1. J Mol Biol 2008, 381:1157-1167.
Conclusions A decade has elapsed since disulfide-based fragment screening first identified an orphan allosteric site in C3/C7. In the intervening years, significant progress has been made in characterizing this site in several additional caspase family members, though endogenous ligands for these sites remain to be identified. Recent studies have afforded a better understanding of the allosteric circuitry that links this site to the active site, and a more subtle understanding of the various active and inactive conformations that can be adopted by both processed and unprocessed caspases. Much of this work is referenced or described herein, an underlying theme of which is the structural diversity and conformationally dynamic nature of this family of proteases. Proteins, peptides, and small molecules that interact with these various states are being identified at an increasing rate, and are affording surprising and unexpected insights into caspase modulation. At the same time, interest in caspases as therapeutic targets has only increased as new roles for these proteases in development and neurodegeneration have been proposed. Recent advances in computational and biophysical screening technology have enabled the application of new screening approaches and the result has been a variety of new chemical matter, as described herein. These molecules point to new avenues for caspase modulation but remain immature from a drug development perspective. Only a robust and sustained effort to improve their potencies and in vivo properties will produce molecules that can truly test the therapeutic potential of caspase modulation by allosteric mechanisms. Current Opinion in Structural Biology 2013, 23:812–819
10. Datta D, McClendon CL, Jacobson MP, Wells JA: Substrate and inhibitor-induced dimerization and cooperativity in caspase-1 but not caspase-3. J Biol Chem 2013, 288:9971-9981. 11. Hardy JA, Lam J, Nguyen JT, O’Brien T, Wells JA: Discovery of an allosteric site in the caspases. Proc Natl Acad Sci USA 2004, 101:12461-12466. 12. Murray J, Giannetti AM, Steffek M, Gibbons P, Hearn BR, Cohen F, Tam C, Pozniak C, Bravo B, Lewcock JW et al.: Tailoring Small Molecules for an Allosteric Site on Procaspse-6. ChemMedChem 2013. in press. This paper describes the discovery of small molecules that bind the dimer-interface site in procaspase-6 with nM affinity. This work demonstrates that high-affinity, non-covalent ligands can be identified for the dimer interface of caspases. 13. Schipper JL, MacKenzie SH, Sharma A, Clark AC: A bifunctional allosteric site in the dimer interface of procaspase-3. Biophys Chem 2011, 159:100-109. 14. Gao J, Wells JA: Identification of specific tethered inhibitors for caspase-5. Chem Biol Drug Des 2012, 79:209-215. 15. Simon DJ, Weimer RM, McLaughlin T, Kallop D, Stanger K, Yang J, O’Leary DD, Hannoush RN, Tessier-Lavigne M: A caspase cascade regulating developmental axon degeneration. J Neurosci 2012, 32:17540-17553. 16. LeBlanc AC: Caspase-6 as a novel early target in the treatment of Alzheimer’s disease. Eur J Neurosci 2013, 37:2005-2018. 17. Vickers CJ, Gonzalez-Paez GE, Umotoy JC, Cayanan-Garrett C, Brown SJ, Wolan DW: Small-molecule procaspase activators identified using fluorescence polarization. Chembiochem 2013, 14:1419-1422. 18. Zorn JA, Wille H, Wolan DW, Wells JA: Self-assembling small molecules form nanofibrils that bind procaspase-3 to promote activation. J Am Chem Soc 2011, 133:19630-19633. www.sciencedirect.com
Modulating caspase activity: beyond the active site Murray and Renslo 819
19. Zorn JA, Wolan DW, Agard NJ, Wells JA: Fibrils colocalize caspase-3 with procaspase-3 to foster maturation. J Biol Chem 2012 2012, 287:33781-33795. 20. Peterson QP, Goode DR, West DC, Ramsey KN, Lee JJ, Hergenrother PJ: PAC-1 activates procaspase-3 in vitro through relief of zinc-mediated inhibition. J Mol Biol 2009, 388:144-158.
22. Stanger K, Steffek M, Zhou L, Pozniak CD, Quan C, Franke Y, Tom J, Tam C, Krylova I, Elliott JM et al.: Allosteric peptides bind a caspase zymogen and mediate caspase tetramerization. Nat Chem Biol 2012, 8:655-660. Using phage-display, these workers identified a peptide that binds a previously uncharacterized tetramerization interface in procaspase-6. Interestingly, this peptide can induce tetramerization of active caspase-6, inhibiting the protease non-competitively.
21. Feldman T, Kabaleeswaran V, Jang SB, Antczak C, Djaballah H, Wu H, Jiang X: A class of allosteric caspase inhibitors identified by high-throughput screening. Mol Cell 2012, 47:585-595.
23. Heise CE, Murray J, Augustyn KE, Bravo B, Chugha P, Cohen F, Giannetti AM, Gibbons P, Hannoush RN, Hearn BR et al.: Mechanistic and structural understanding of uncompetitive inhibitors of caspase-6. PLoS ONE 2012, 7:e45086.
www.sciencedirect.com
Current Opinion in Structural Biology 2013, 23:812–819