Tunable allosteric library of caspase-3 identifies coupling between conserved water molecules and conformational selection

Joseph J. Maciag; Sarah H. Mackenzie; Matthew B. Tucker; Joshua L. Schipper; Paul Swartz; A. Clay Clark

doi:10.1073/pnas.1603549113

Edited by Ruth Nussinov, National Cancer Institute, Frederick, MD, and accepted by Editorial Board Member Gregory A. Petsko August 24, 2016 (received for review March 2, 2016)

Significance

The interconversion of states in the caspase-3 native ensemble is affected by binding of ligands that either stabilize or destabilize active-site loops. It is not clear how the ensemble is regulated in cells, aside from modulating levels of endogenous caspase inhibitors. We describe a library of caspase-3 variants with activities that vary by more than four orders of magnitude and show that removal of conserved water molecules may provide a strategy to design novel allosteric inhibitors that globally destabilize the active conformation within the ensemble. Our results suggest that posttranslational modifications fine-tune caspase activity by disrupting conserved water networks, and our database provides an approach to examine caspase signaling in cells by modifying caspase-3 activity while simultaneously maintaining endogenous enzyme levels.

Abstract

The native ensemble of caspases is described globally by a complex energy landscape where the binding of substrate selects for the active conformation, whereas targeting an allosteric site in the dimer interface selects an inactive conformation that contains disordered active-site loops. Mutations and posttranslational modifications stabilize high-energy inactive conformations, with mostly formed, but distorted, active sites. To examine the interconversion of active and inactive states in the ensemble, we used detection of related solvent positions to analyze 4,995 waters in 15 high-resolution (<2.0 Å) structures of wild-type caspase-3, resulting in 450 clusters with the most highly conserved set containing 145 water molecules. The data show that regions of the protein that contact the conserved waters also correspond to sites of posttranslational modifications, suggesting that the conserved waters are an integral part of allosteric mechanisms. To test this hypothesis, we created a library of 19 caspase-3 variants through saturation mutagenesis in a single position of the allosteric site of the dimer interface, and we show that the enzyme activity varies by more than four orders of magnitude. Altogether, our database consists of 37 high-resolution structures of caspase-3 variants, and we demonstrate that the decrease in activity correlates with a loss of conserved water molecules. The data show that the activity of caspase-3 can be fine-tuned through globally desolvating the active conformation within the native ensemble, providing a mechanism for cells to repartition the ensemble and thus fine-tune activity through conformational selection.

Caspase function in cell development and cell death results from a continuum of enzyme activity, in which an as-yet-undefined activity threshold is required for cell death. At subthreshold levels, caspase activity is important for a variety of physiological reactions (referred to as adaptive responses), including remodeling the cytoplasm (1), cell differentiation (2), neuron pruning (3), receptor endocytosis (4), macrophage function (5), and development of the eye lens (6) and inner ear (7). The roles of caspases in apoptosis are well known, but their roles in adaptive responses are less clear, particularly in regard to how cells set the threshold of caspase activity to limit apoptosis while ensuring sufficient activity for signaling and differentiation.

Cells use two general mechanisms to modify caspase activity, through modulating levels of active caspase or through allosteric mechanisms that change the distribution of conformations in the native ensemble, although the two are not mutually exclusive. Levels of caspase-3 are controlled by cleavage of the inactive zymogen to yield a dimer of protomers (Fig. 1A) (8, 9), and this process is responsive to several signaling pathways, such as transient expression of the Bad-Bax cascade (10) or phosphorylation of the zymogen (Fig. 1B) (11). Alternatively, inhibitor of apoptosis proteins (IAPs) affect levels of active caspase-3 by direct inhibition through active-site binding or through ubiquitination, leading to proteasome degradation (12 ⇓⇓–15). The interplay between these signaling pathways modulates total activity by changing the amount of active caspase-3.

Fig. 1.

Caspase allosteric regulation. (A) Active caspase-3 dimer showing active site loops L1–L4 and L2′ from each protomer. (B) Model for allosteric regulation of caspases. The procaspase native ensemble contains at least two states and favors the inactive conformations, whereas the mature caspase native ensemble contains several states, including the active conformation (state 1), a low-energy state with disordered active site loops (state 3), and a high-energy state with displaced catalytic groups (state 2). Maturation of the procaspase is influenced by PTMs or signaling cascades (Bad-Bax, for example). The active conformation (state 1) is selected by the binding of substrate, whereas inactive conformations (states 2 and 3) are selected by the binding of allosteric ligands or PTMs.

Caspase activity is also affected by posttranslational modifications (PTMs), most notably by phosphorylation, where several sites on mature caspases are modified (11, 16). Depending on the site of modification, the PTM inactivates the caspase, either by directly interfering with substrate binding or through allosteric mechanisms, so controlling kinase or phosphatase activities allows for a level of control over caspases. Consequently, PTMs link caspase activity to cellular signaling reactions that are important in the adaptive responses (17 ⇓–19).

The mature caspase energy landscape is an ensemble of states consisting of active and inactive conformations (Fig. 1B, brackets), where the free energies of the inactive states relative to that of the active state vary among caspases. Both caspase-2 and -3, for example, contain high-energy inactive states (Fig. 1B, state 2), whereas caspase-1, -6, and -7 contain low-energy, so-called “closed-loop” inactive states with disordered active-site loops (Fig. 1B, state 3) (20 ⇓⇓⇓⇓–25). Conformational selection by ligand binding to the active or allosteric sites affects the partitioning of active vs. inactive conformations, so the activity of the caspase reports on the relative distribution of active and inactive states within the ensemble (23, 25), providing a mechanism to fine-tune activity. An allosteric site in the dimer interface (Fig. 1 A and B), for example, prevents active-site loop formation when bound to small-molecule allosteric inhibitors (26, 27), demonstrating a common allosteric mechanism through steric clashes with active-site loops, and selecting the “disordered loop” conformation (Fig. 1B, state 3). Although the active (Fig. 1B, state 1) and disordered-loop (Fig. 1B, state 3) conformations are well-described structurally and appear to be well-populated thermodynamically, much less is known about the high-energy inactive state(s) within the caspase-3 ensemble, collectively referred to here as state 2 (Fig. 1B). Higher-energy inactive states have been observed in caspase-2 when inhibited by DARPins and demonstrate that small changes in positioning the catalytic groups are sufficient to allosterically inhibit the enzyme (28). In contrast, caspase-6 undergoes very large transitions when allosterically inhibited by zinc, where a short surface β-sheet (β¹–β³; Fig. 1A) undergoes a coil-to-helix transition that extends helix 3 and disrupts the catalytic histidine and cysteine (29). We showed previously that a high-energy inactive conformation in the caspase-3 ensemble results from transient rotation of helix 3 toward the dimer interface, reducing interface volume by ∼800 Å³ (Fig. 1A, H3), and resulting in disruption of the catalytic dyad, H121 and C163, increased mobility of active-site loop 1 (Fig. 1A, L1), and a narrower S1′ binding pocket (22, 23). Collectively, the data for caspase-2 and -3 suggest that conformational selection in the native ensemble decreases enzyme activity by populating high-energy states with mostly intact active sites, but with disrupted catalytic groups. Importantly, the ensemble may provide the cell with the means to reversibly fine-tune caspase activity by using combinations of common and/or unique sites of PTMs.

Here, we examine the role of conserved water molecules in the allosteric regulation of conformational states in the native ensemble. Starting with a database of 15 high-resolution structures of caspase-3, we used the detection of related solvent positions (DRoP) analysis (30) to identify 145 highly conserved water molecules. The positioning of the conserved waters correlates with regions of the protein that are also modified, suggesting that PTMs may fine-tune activity by disrupting the conserved water networks. We used saturation mutagenesis in an allosteric site to test this hypothesis, and we show that the activities of the mutants vary by more than four orders of magnitude. Our database of 37 high-resolution structures of caspase-3 variants shows that desolvating the enzyme decreases activity by repartitioning states within the native ensemble. Overall, the database establishes a means to fine-tune caspase-3 activity over several orders of magnitude through conformational selection that is facilitated by dehydrating the native state ensemble.

Results and Discussion

Global Analysis of Caspase-3 Solvation.

We performed a global analysis of solvation using the recently described DRoP program (30), which analyzes positions of water molecules in related X-ray crystal structures to cluster the waters in crystallographically related positions. One also determines the level of conservation for each water molecule as well as conserved contact positions on the protein. Our database consisted of 15 high-resolution (<2 Å) structures of wild-type caspase-3 and contained 4,995 water molecules (Fig. 2A and Table S1). After clustering and conservation analysis, we identified 450 unique clusters (Fig. 2B), which we grouped based on degree of conservation in the database: >93%, 80–93%, 66–79%, and <66% (Fig. 2C, blue, cyan, orange, and red, respectively). The conserved waters, defined as the group present in 14 of the 15 structures (>93%), comprise 145 water molecules and are the focus of the subsequent analyses. The results show that the conserved waters interact with active-site loops and helices on or near the protein surface (Fig. 2C, blue spheres, and Fig. 3A), particularly in the dimer interface, whereas the least-conserved waters (<66%) are located in the substrate-binding groove and chain termini (Fig. 2C, red spheres). We note that there was no effect on the conserved waters from crystal packing.

Fig. 2.

DRoP analysis of wild-type caspase-3. (A and B) Water molecules in 15 structures of caspase-3 overlaid onto the protomer (PDB ID code 2J30) (A), then organized into 450 clusters after DRoP analysis (B). (C) The 450 water clusters were scored based on level of conservation within the database: blue, >93%; cyan, 80–93%; orange, 66–79%; and red, <66%.

Fig. 3.

Conserved water molecules identified in DRoP analysis. (A) Regions of caspase-3 protomer that interact with water molecules, colored by conservation: Blue represents most conserved, and red represents least conserved. Active site loops L1–L4 are shown, as are surface helices H1–H5; β6 refers to β-strand 6, which forms the binding surface in the dimer interface with the second protomer. (B) Within the 450 water molecules, 145 waters were highly conserved in the database (>93%). The conserved waters were parsed into three categories based on number of hydrogen bonds with the protein: surface, channel, or buried. The number of waters in each category is indicated. (C) Percent of surface, channel or buried conserved waters that are affected by mutations of caspase-3. (D–F) Location of surface (D; red), channel (E; yellow), or buried (F; blue) waters overlaid onto caspase-3 protomer (PDB ID code 2J30).

View this table:

Table S1.

Wild-type caspase-3 structures used in DRoP analysis

Formation of the intact active site involves movement of active site loop 3 (L3) from a solvent-exposed position, in which the loop is mostly disordered, to the active position, where the loop forms the substrate-binding groove along the protein surface (Fig. 1A) (8, 9). Although the substrate-binding pocket is “dry”—that is, it does not contain conserved waters—several conserved waters are observed to connect the C-terminal side of L3, with helices 1 and 4 on the opposite side of the central cavity (Fig. S1B and described below). In addition, several conserved waters interact with the N-terminal side of L3, the so-called “elbow-loop” region (20) near the central cavity of the dimer interface (Fig. S1A), and the conserved waters connect L3 with the “loop-bundle” of L2, L2′, and L4 (Fig. 1A), which form a network of hydrogen bonds that stabilize the active conformation (31, 32). Steric clashes between the elbow-loop and L2′ in the closed-loop conformation is a primary source of allosteric regulation in maintaining the disordered-loop inactive conformation (Fig. 1B, state 3). The conserved waters at the N and C termini of L3 likely stabilize the active conformation of the loop through extensive hydrogen bonding. Identification of the conserved waters near L3 of caspase-3 is also consistent with an allosteric network in caspase-1, where side chains in this region are known to change H-bonding patterns in the transition from inactive to active states (33). The DRoP analysis also showed that many conserved waters bind to helices 1, 4, and 5 (Fig. 3A, blue regions). At present, no protein effectors have been observed to bind in this region of caspase-3. Interestingly, however, these helices are known to contain several sites of modification in caspases, including phosphorylation (11) and glutathionylation (34), as well as a zinc-binding site (29). The potential linkage between modifications at these sites and allosteric regulation of enzyme activity is currently not known.

Fig. S1.

Conserved water molecules (green spheres) near caspase allosteric sites. Conserved water molecules near the elbow loop of active site L3 and active site L2 (A); connecting helices 1 and 4 with active site L1 and L3 (B); near the C terminus of helix 3 and turn 9, which connects α-helix 3 and β-strand 4 and is a site of posttranslational modification (C); and near turn 6 connecting the dimer interface, helix 3, and the active site (D).

In addition to the active site loops, the C-terminal region of helix 3 (Fig. 3A, H3) also interacts with several conserved waters (Fig. S1C). Flexibility in this helix affects the population distribution in the native ensemble, where rotation of the helix toward the dimer interface narrows the S1′ binding pocket and disrupts the catalytic C163 and H121, thus abolishing activity (22, 23). Notably, the N terminus of helix 3, as well as three short β-strands near the active site (β¹–β³), which bind to conserved waters (Fig. S1D), undergoes extensive rearrangements in the transition from the procaspase-3 zymogen to the mature caspase-3 (see Fig. 7) (24). In addition, the same region in caspase-6 undergoes a coil-to-helix transition, which increases the distance between the catalytic histidine, located on β¹, and the catalytic cysteine, located at the base of active site L2. The extended helix form of caspase-6 is stabilized by zinc binding to an allosteric site near helix 5 (29). The resolution of caspase-6 structures is not sufficient to include in the DRoP analysis with caspase-3, so it is not clear whether similar conserved waters are present; however, if caspase-6 has conserved waters near β¹–β³ and turn 6, as observed in caspase-3 (Fig. S1D), then the sites would be abolished upon transition to the extended helix conformation.

We further parsed the set of 145 conserved waters into three groups based on hydrogen-bond contacts—namely, surface, channel, or buried, and the three groups contain 104, 14, and 27 waters, respectively (Fig. 3B). The conserved surface water molecules (Fig. 3D) are distributed throughout the protein, although notably the substrate-binding groove appears “dry” because the active-site waters are less conserved. The conserved channel water molecules are found primarily near helix 3 and connecting helices 1 and 4 (Fig. 3E), whereas buried water molecules (Fig. 3F) connect several active-site loops, as well as several loops near helices 3 and 4.

Interestingly, the buried water molecules are distributed similarly to the PTMs described in ref. 35. Aside from studies of zinc binding to caspase-6 (29) and phosphorylation of helix 5 in caspase-6 (36, 37), little is known of the mechanisms by which PTMs or metal binding affect enzyme activity. The positioning of the conserved buried water molecules in caspase-3 suggests that the PTMs may affect partitioning of the native ensemble by disrupting conserved water networks.

Saturation Mutagenesis in an Allosteric Site Provides a Library of Caspase Variants with a Broad Range of Activities.

To test the hypothesis that the inactive conformations in the native ensemble (Fig. 1B) have altered solvation, we examined changes in the conserved waters resulting from mutations in the allosteric site within the central cavity of the dimer interface. We chose this site because, as described (26, 27), the binding of allosteric inhibitors to this site prevents active-site loop formation via steric clashes with L3. The allosteric site of the central cavity is considered a common allosteric site in caspases (35), so it may report on a common mechanism for regulation of caspase activity, as opposed to a unique site that may report on a mechanism specific to caspase-3. In addition, two mutations of V266 at the center of the allosteric site suggested context-dependent changes in activity. For example, mutation of V266 to histidine abolished enzyme activity, primarily due to destabilizing helix 3 (22, 23, 38), whereas mutation of V266 to glutamate decreased activity only 10-fold in the mature caspase, but resulted in a substantial increase in activity of the zymogen, most likely due to expulsion of a linker from the dimer interface, allowing the active conformation to form in the zymogen (38, 39). Finally, other amino acids in or near the allosteric site have been mutated, and several high-resolution structures are available for the variants (Table S2).

View this table:

Table S2.

Caspase-3 mutants used in DRoP analysis

Using a saturation mutagenesis approach, we changed V266 to all other amino acids, resulting in 17 new mutants at this site, because V266H and V266E have been described. Activity measurements for the 20 proteins show that the substitution of V266 with glutamate is unique in increasing activity of the zymogen because all other mutations at this site had little effect on zymogen activity (Fig. 4A and Table S3). With the exception of V266H (described in ref. 38), V266Y, and V266N, all of which had no activity, the substitutions at V266 in the zymogen resulted in changes between 0.1-fold and threefold that of wild-type procaspase-3, with an average activity of 6.7 × 10² M⁻¹⋅s⁻¹ (excluding the three inactive variants and V266E).

Fig. 4.

Saturation mutagenesis of dimer interface allosteric site. (A and B) Relative enzyme activity for procaspase-3 (A) or caspase-3 (B) variants mutated at V266. Relative activity calculated as k_cat/K_M(mutant) ÷ k_cat/K_M(wild-type). Squares indicate no activity in the mutant. (C–E) Activity vs. volume change upon substitution of V266 in terms of enzyme specificity (C), k_cat (D), and K_M (E). For A–E, the V266 substitutions are indicated by the one-letter code for amino acids. (F–H) Changes in enzyme specificity (F), k_cat (G), and K_M (H) for the caspase-3 mutant database, containing 37 caspase-3 variants. The water molecules in the 37 caspase-3 mutants in the database were analyzed by DRoP, as described in the text, and the total number of conserved water molecules affected by the mutation is plotted vs. enzyme activity. The data were fit either to a semilog (F and G, solid lines) or linear equation (H, solid line), and results from the fits are described in the text.

View this table:

Table S3.

Enzymatic activity of (pro)caspase-3 V266 variants

In contrast to results for the zymogen, we observed large effects on the activity of mature caspase-3 when V266 was substituted with other amino acids (Fig. 4B). With the exceptions of V266H (described in ref. 38) and V266P, both of which exhibited no activity, the activities of the mutants varied by nearly four orders of magnitude, with V266F exhibiting the lowest measurable activity. The mutations resulted in increased K_M values of up to ∼10-fold, where the average K_M was 23.2 μM for this dataset, and large decreases in k_cat of up to ∼100-fold (Table S3), with the combined changes resulting in a large distribution of activities. As described for the V266E and V266H variants (38), there were no changes in dimer stability, so the changes in enzyme activity were not due to lower dimer formation. In addition, there was no correlation in enzyme specificity (Fig. 4C), k_cat (Fig. 4D), or K_M (Fig. 4E) with changes in side-chain volume upon substitution of V266. Likewise, there was no apparent correlation with side-chain chemistry, because changes in activity did not correlate with hydrophobicity, aromatic, small polar, or charged groups of side chains.

Structural and Dynamic Studies Show That Flexibility in the Allosteric Site Disrupts Conserved Waters and Propagates to the Active Site.

To further examine changes in enzyme activity for the V266 mutant dataset and potential changes in the conserved waters, we determined the structures of 13 of the 17 new variants by X-ray crystallography. Noting that structures of the V226H and V266E variants have been described (22, 39), we were therefore successful in obtaining a nearly complete structural description of the saturation mutagenesis dataset at V266, where only V266G, V266P, V266T, and V266R did not crystallize (Table S4). Of the 13 new V266X structures, 12 structures had resolution of <2Å and were used in the subsequent DRoP analysis (described below).

View this table:

Table S4.

Crystallographic parameters for V266X variants

During formation of the active conformation of caspase-3, R164, from active site L2, intercalates between Y197 (β-strand 6) and P201 (elbow-loop of L3) in the dimer interface (20). These are the movements that are prevented through binding of allosteric effectors in the central cavity (26, 27). In wild-type caspase-3, six water molecules, three from each protomer, form a hydrogen-bonding network across the dimer interface and stabilize R164 and R164′ (from the second protomer) (Fig. 5A). The water-mediated hydrogen-bonding network across the dimer interface is important for positioning active-site loop 2 and C163 for catalysis (22). Using the DRoP analysis, we identified two of the three water molecules as conserved (>93%) (Fig. 5A, Wat40 and Wat50), whereas the third water was present in >80% of the structures in our database of wild-type caspase-3 (Fig. 5A, Wat177). Our structures of the V266 variants show that substitution of valine with small side chains (A, S, and C) generally retains the H-bonding network in the interface (Fig. S2). In contrast to the small side chains, substitution of V266 with larger side chains generally disrupted the H-bonding network by displacing the conserved waters (Wat40, Wat50, and Wat177) in the interface (Fig. 5).

Fig. 5.

Comparison of wild-type and V266X caspase-3 dimer interfaces. (A) Caspase-3 structure (PDB ID code 2J30) highlighting residues in the dimer interface, including V266. Red spheres represent three water molecules (from each protomer) that form hydrogen bonds across the dimer interface. Two waters (Wat40 and Wat50) are conserved in >93% of the caspase-3 structures in the database, whereas Wat177 is conserved in >80% of the structures. (B–E) Comparison of wild-type caspase-3 (PDB ID code 2J30) (gray) with caspase-3 variants (yellow), V266F (B), V266Y (C), V266W (D), and V266H (E). In this series, F266 (B), Y266 (C), and W266 (D) are observed in two conformations.

Fig. S2.

Comparison of wild-type and V266X caspase-3 dimer interfaces: small/polar amino acids. Comparison of wild-type caspase-3 (PDB ID code 2J30) (gray) with caspase-3 variants (yellow), V266C (A), V266S (B), and V266A (C) is shown. C266 forms a disulfide bond, and S266 is observed in two conformations. Red/cyan spheres represent conserved water molecules.

Five of the V266 variants showed two conformations for the side-chain—V266S (Fig. S2B), V266D (Fig. S3C), V266F (Fig. 5B), V266Y (Fig. 5C), and V266W (Fig. 5D). The two conformations can be characterized generally based on their positions relative to R164. In the first conformation, the bulk of the side chain is positioned close to R164; in the second conformation, the bulk of the side chain is positioned away from R164 and closer to Y195 on β-strand 6. In the V266E variant (Fig. S3A), the larger side chain (compared with valine) displaces the conserved waters, but maintains the hydrogen-bonding network in the interface, whereas the side chain of aspartate, which is shorter than that of glutamate, is observed in two conformations, with longer hydrogen bonds between the waters and the side chain (Fig. S3C). The asparagine side chain is locked between these two conformations (Fig. S3D), as is the histidine side chain (Fig. 5E), and disrupts the conserved waters. Overall, either the changes in hydrogen bonding in the interface or increased side-chain dynamics, or both, may explain the lower activity in V266S, V266D, and V266N compared with V266E (Fig. 4B).

Fig. S3.

Comparison of wild-type and V266X caspase-3 dimer interfaces: large polar and charged amino acids. Dimer interface of wild-type caspase-3 (PDB ID code 2J30) (gray) compared with caspase-3 variants (yellow), V266E (A), V266Q (B), V266D (C), V266N (D), and V266K (E) is shown. In this series, D266 is observed in two conformations. Red spheres represent conserved water molecules.

In the case of V266I, the isoleucine is observed in conformation 1, where the bulk of the side chain is near R164, whereas the leucine side chain in V266L is observed in conformation 2, near Y195 (Fig. S4 A–D). The positioning of I266 results in a shift of Y197 and of Y195, on β-strand 5, away from β-strand 6, suggesting steric clashes in the interface. In contrast, the leucine side chain forms close van der Waals contacts with L266′ of the second protomer, with no observed shifts in Y195 or Y197 (Fig. S4 C and D). Although the V266I variant retains the water network in the interface, the less-optimal hydrophobic contacts across the interface, as well as potential steric clashes with the neighboring β-strand, may explain the lower activity (1.6 × 10³ M⁻¹⋅s⁻¹) compared with the V266L variant (9.0 × 10³ M⁻¹⋅s⁻¹). Finally, although the side chains for F266 (Fig. 5B), Y266 (Fig. 5C), and W266 (Fig. 5D) are observed in two conformations, the V266Y variant exhibits significantly higher activity than the two other mutants (Fig. 4B). This result is most likely because of formation of a hydrogen bond between the hydroxyl group of Y266 and the side chain of R164. Overall, the 15 V266X mutants showed no large structural changes that affected the active sites, so substitution of V266 in the dimer interface results in localized structural changes that affect conserved waters in the interface, side-chain fluctuations at 266, and residues on the neighboring β-strand 5. Because all of the structures contain an active-site inhibitor, the data show that the mutations have little global effects on the ground-state active conformation observed in the crystals (Fig. 1A, state 1).

Fig. S4.

Comparison of wild-type and V266X caspase-3 dimer interfaces: hydrophobic amino acids. Comparison of wild-type caspase-3 (PDB ID code 2J30) (gray) with caspase-3 variants (yellow), V266I (A and B), V266L (C and D), and V266M (E). Steric clashes are observed between I266 and I266′ (from the second protomer) (B), but L266 demonstrates tight packing with L266′ (D). Red spheres represent conserved water molecules.

We showed previously that the lower activity of mutants coupled to V266H correlated to changes in the mobility of several active-site loops (22, 23). We examined the changes in active-site loop mobility for the caspase-3 mutant database in two ways. First, from the structural databases, we determined the average B-factor for each residue in wild-type caspase-3 (Fig. S5 A and B), based on the 15 structures in the database. The results show that, relative to the average B-factor for all residues (16.4 ± 5.8), active-site loops L1 and the C-terminal region of L3 show increased mobility. Comparable data were determined for the caspase-3 mutant database (Table S2 and Fig. S5 C and D), where the average B-factor for all residues was 16.0 ± 6.0. A comparison of the two databases showed that each residue of the mutant caspase-3 demonstrates similar B-factors as those of wild-type caspase-3, because the difference is close to zero. However, active-site loops L1 and L3 are both less mobile in the caspase-3 mutants based on significantly lower B-factors compared with wild-type caspase-3 (Fig. S5 E and F). In addition, amino acids in the turn connecting helices 4 and 5, as well as the amino and carboxyl termini of the protomers, are more mobile in the caspase-3 mutants (Fig. S5 E and F).

Fig. S5.

B-factor analysis of caspase-3 databases. (A and B) Average B-factor of wild-type caspase-3 database from structures listed in Table S1. The large subunit (A) and small subunit (B) in the protomer are shown. (C and D) Average B-factor of mutant caspase-3 database from structures listed in Table S2, shown as large (C) and small (D) subunits in the protomer. In A–D, active-site loops and surface helices are labeled, and numbers refer to amino acid positions. The SD is shown by the red bars. (E and F) Change in B-factors for the two databases, calculated as ΔB-Factor=(B-Factor_mut – B-Factor_wild-type). In this calculation, a value below zero represents a lower B-factor in the mutant caspase-3 database, compared with wild-type caspase-3. Examples are shown for active site loops L1 (E) and L3 (F).

The second way in which we examined active-site loop mobility was to perform molecular dynamics (MD) simulations on six mutants—V266Y, V266F, V266D, V266E, V266Q, and V266N—because of the differences in enzyme activity and structural features described above. Data from 50-ns MD simulations of wild-type caspase-3 (22) provide the baseline for comparison with the V266 mutants and show that the largest fluctuations are observed in two active-site loops (L1 and L4) and turn 6, which contains E123 and E124 (Fig. S6 A and B). The positioning of turn 6 is important for activity because it contains residues that form part of the oxyanion hole during catalysis, as well as residues that contribute to the hydrogen-bonding network of the dimer interface. In addition, turn 6 connects β-strands in the surface β-sheet (β¹–β³) described above.

Fig. S6.

MD simulations of caspase-3 and selected mutants. (A) rmsfs of wild-type caspase-3 from 50-ns MD simulations. Active site loops and surface helices are labeled, and numbers refer to amino acid positions. (B) Regions identified in A mapped onto caspase-3 protomer (PDB ID code 2J30). Change in rmsf for six caspase-3 variants (V266E, V266D, V266Q, V266N, V266Y, and V266F) was calculated as Δrmsf = (rmsf_mut – rmsf_wild-type).

The MD simulations for the V266 variants show that the rms fluctuations (rmsfs) are generally within 1–2 Å of those observed for wild-type caspase-3 (Fig. S5C). Here, the data are reported as Δrmsf (rmsf_Mut – rmsf_WT), so values greater than zero show larger fluctuations in the mutant compared with those of wild-type caspase-3, whereas values less than zero reflect larger fluctuations in wild-type caspase-3 compared with those of the mutant. Comparisons of fluctuations in V266D (Fig. 6A) to those in V266E (Fig. 6B) and in V266F (Fig. 6C) to those in V266Y (Fig. 6D) show increased fluctuations in turn 6 that result in new, albeit transient, interactions between E123, in turn 6, and the catalytic H121. In wild-type caspase-3, H121 fluctuates between two conformations during the simulation, which likely reflects its movements during proton transfer of the catalytic reaction. In these conformations, the H121 side-chain fluctuates between ∼7 and ∼12 Å from E123. In the case of the V266D and V266F variants, movements in turn 6, along with the fluctuations in H121, result in shorter distances between H121 and E123, between 3.2 and 4.1 Å, so the electrostatic interactions between H121 and E123 may affect the catalytic activity by decreasing the rate of proton transfer. We note that the motions described for turn 6 do not appear to be coupled with changes in mobility of active-site loops L1, L3, or L4 described above. The position of the surface β-sheet (β¹–β³) and turn 6 is more similar to that of the inactive zymogen than to the active conformation (Fig. 7). In the active enzyme, turn 6 is in the down conformation (Fig. 7, blue), where E124 is positioned to H-bond with R164 in the central cavity of the dimer interface (see also Fig. 5A). In the up conformation observed in the zymogen (Fig. 7, orange), turn 6 is rotated away from the central cavity such that E124 is solvent-exposed, and the surface β-sheet, β¹–β³, is not well-formed. In the V266 variants, fluctuations in turn 6 displace the catalytic H121 by 2.7 Å (Fig. 7A), whereas the catalytic cysteine is moved 1.7 Å from the active position (Fig. 7B). In Fig. 7, representative data for V266D are shown in yellow. In addition, all variants showed increased fluctuations in helix 3, which, as described previously for V266H, also affect the positions of the catalytic groups through movements propagated through the surface β-sheet, β¹–β³ (22, 23). Together, the B-factor analysis shows that, on average, the caspase-3 variants have lower mobility in active site loops L1 and L3, whereas the MD data show that the lower activity of some V266 variants may result from fluctuations in turn 6 and active-site loop 2, culminating in displacement of the catalytic groups.

Fig. 6.

MD simulations of V266X variants demonstrate movements of amino acids in turn 6, E123 and E124, relative to the catalytic H121. (A–D, Left) V266D (A, Left), V266E (B, Left), V266F (C, Left), and V266Y (D, Left): gray indicates starting position from the crystal structure, and yellow indicates closest distance between E123 and H121. A–D, Right show 200 frames (at 250-ps intervals) of the 50-ns simulation to demonstrate increased flexibility in Turn 6, for V266D and V266F relative to V266E and V266Y, respectively.

Fig. 7.

Comparison of turn 6 and catalytic groups in the caspase-3 zymogen, active caspase-3, and V266D variant. (A) In the active caspase-3 (cyan) (PDB ID code 2J30), turn 6 is in the down position, where E124 hydrogen-bonds with R164 in the dimer interface (Fig. 5A). In the inactive zymogen (orange) (PDB ID code 4JQY), turn 6 is in the up position, resulting in a shift of H121 of 2 Å from the active position. In MD simulations of V266D (yellow), flexibility in turn 6 results in a 2.7-Å shift of H121 compared with the active position. (B) The catalytic C163 in the zymogen (orange) or in V266D (yellow) is 5.1 or 1.7 Å, respectively, from the active position (cyan). Data for V266D are described in Fig. 6.

Lower Activity Correlates with Changes in Solvation.

We performed the DRoP analysis on the V266 variants for which we obtained X-ray crystal structures of <2-Å resolution (Table S4). To expand the database, we also included 24 additional caspase-3 variants for which we have previously determined enzyme activity and high-resolution structures, providing a database of 37 caspase-3 variants containing a total of 10,736 water molecules (Table S2). After DRoP analysis, we determined the average B-factor for each conserved water molecule in the database (Fig. S7A) and determined which of the water molecules within the set of 145 conserved waters were absent in each mutant. The latter data are presented in two ways. First, we describe the total number of conserved waters displaced for each mutant, and, second, we quantified the disruption of each of the conserved waters in the caspase-3 mutant database. We report the latter results as relative disruption (Fig. S7), where a value of 1 signifies that the conserved water was disrupted in all caspases in the database, and a value of 0 signifies that the conserved water was present in all caspases in the database. We note that all three classes of the conserved waters (surface, channel, and buried) were affected by the mutations, with conserved surface waters demonstrating the largest changes (Fig. 3C). The database of 37 variants provides a range of enzyme activities, nearly four orders of magnitude (Fig. 4F), and the DRoP analysis shows that the change in enzyme activity correlates with a loss of conserved waters, where the lower catalytic efficiency (Fig. 4F) was due to lower k_cat (Fig. 4G) and higher K_M (Fig. 4H) values. Each point in Fig. 4 F–H represents one mutant in the caspase-3 database, where an average of 27.9 waters are displaced per mutant, with a broad range between zero (wild-type) and 70 waters. Fits of the data show that, when the 145 conserved waters are present (no change in the waters), values for k_cat, 0.65 s⁻¹ (Fig. 4G, range of 0.42–1.04 s⁻¹), and K_M, 9.7 μM (Fig. 4H, range of 9.3–10.2 μM), yield an enzyme specificity of 6.7 × 10⁴ M⁻¹⋅s⁻¹, which is similar to that obtained from a fit of the data in Fig. 4F, 1.65 × 10⁵ (range of 1.36 × 10⁵ to 2.01 × 10⁵) M⁻¹⋅s⁻¹ and agrees well with values previously determined for wild-type caspase-3 (Table S3). We note that the broad range of values in Fig. 4 F–H suggest that the data may reflect the heterogeneity of the ensemble and report on multiple inactive conformations rather than a single discrete state as represented in Fig. 1B (state 2). For example, an improved understanding of inactive conformations characterized by increased dynamics of turn 6 vs. those with increased dynamics of helix 3, as well as the dehydration associated with each state, may allow one to parse data such as those shown in Fig. 4 to resolve multiple inactive conformations.

Fig. S7.

B-factor analysis of conserved water molecules. (A) Average B-factor of 145 conserved water molecules as identified in DRoP analysis of the wild-type caspase-3 database (Table S1) and described in Fig. 3. The SD is shown by the red bars, and the average B-factor for the 145 conserved water molecules is 25.6 ± 7.9. (B) Disruption of 145 conserved water molecules in the mutant caspase-3 database (Table S2). Relative disruption refers to the frequency in which the water was disrupted in a mutant caspase-3. The results were scaled between 1 (disrupted in all structures) and 0 (present in all structures). (C) Relative disruption of conserved water molecules from panel B mapped onto the caspase-3 protomer (PDB ID code 2J30). (D) Cluster of conserved waters near helix 4 (wat61, wat74, and wat38; Fig. 8) connecting helix 4 to active site loop 3 (D211). (E) Cluster of conserved waters near helix 4 (wat61, wat74, and wat38; Fig. 8) also connect helix 4 to binding sites for the N and C termini. The cluster is positioned near sites of posttranslational modifications in caspases in the N and C termini, which also contain several conserved, and disrupted, water molecules. Both regions show increased mobility in the caspase-3 mutants (Fig. S5F). For C–E, the color of each water corresponds to the color scale shown in panel B, corresponding to the relative disruption: blue, <0.1; cyan, 0.11–0.25; green, 0.26–0.50; orange, 0.51–0.70; and red, 0.71–1.

Changes in the conserved water molecules also report on potential “hotspots” on the protein where water networks may be sensitive to conformational changes or to ligand binding. This concept is analogous to the solvent mapping of Ras, where binding of multiple small organic compounds provides “maps” of potential ligand-binding sites (30). To examine potential hotspots on caspase-3, we analyzed the relative disruption of each of the 145 conserved waters in the caspase-3 mutant database (Fig. S7B). We parsed the data into five bins, representing levels of disruption—<0.1 (blue), 0.11–0.25 (cyan), 0.26–0.5 (green), 0.51–0.7 (orange), and 0.71–1 (red) —and we mapped the relative disruption onto the conserved waters of the caspase-3 structure using the same color code (Fig. 8A and Fig. S7 B and C). The highly disrupted waters (0.71–1; red) were present primarily at the chain termini, and the conserved waters with low disruption (<0.1; blue) were primarily buried waters that are close to sites of posttranslational modifications. As shown in Fig. 8A (and Fig. S7 C–E), the remaining water molecules are found throughout the protein.

Fig. 8.

Hotspots of conserved water molecules affected by mutations in allosteric site of the dimer interface. (A) Overall perturbations of the 145 conserved water molecules. Blue, cyan, green, orange, and red represent perturbation levels, from low to high. (B) Five water molecules (red spheres) that connect the dimer interface (V266, Y197, and Y195) with helix 3 and turn 6 show perturbations in the mutational database. (C) Two additional regions show perturbations near active site loop 3 and helices 1 and 4 (Upper); and the C termini of helices 1 and 4 (Lower). Perturbations of waters in helices 1 and 4 are near sites of posttranslational modifications. In C, water molecules are colored as surface (red), channel (yellow), or buried (blue).

We highlight two examples that show significant changes in the conserved waters, where 20–60% of the structures report disrupted water networks: the helix 3-dimer interface cavity and two areas of the face containing helices 1, 4, and 5 (Fig. 8A). First, of 21 conserved water molecules that interact with the helix 3–β¹–β³–turn 6 region, 13 of those waters are disrupted in the mutant database. For example, several water molecules near turn 6 form part of an extensive hydrogen-bonding network between the dimer interface, turn 6, and the N terminus of helix 3. Five water molecules in this network are disrupted in the caspase-3 variants (Fig. 8B). From our MD simulations, we examined fluctuations in the set of conserved water molecules compared with the nonconserved waters and bulk solvent, and the results showed that the average rmsf for the conserved water molecules (9.1 ± 4.0 Å) is similar that of the nonconserved waters (10.2 ± 3.5 Å) or the bulk solvent (11.9 ± 2.6 Å). However, one observes variations within the clusters of conserved waters. For example, two of the five water molecules in the cluster near turn 6 have lower than average RMSF (Wat21 and Wat32; 4.2–5.8 Å), whereas three waters (Wat13, Wat48, and Wat65) demonstrate higher than average rmsf (11.5–17.2 Å). As shown in Fig. 8B, these five conserved water molecules connect the N terminus of helix 3 with amino acids on β-strand 5 of the dimer interface (Y195 and Y197) and turn 6 near the active site (E124). As described above, structural perturbations due to the mutations at V266 in the dimer interface are localized to side-chain perturbations at position 266 and residues on the neighboring β-strand 5. Our DRoP analysis and MD simulations of the conserved water molecules show that the structural perturbations in the interface are propagated through β-strand 5 to turn 6 by disruption of the conserved water molecules that connect the two regions. Loss of the water molecules likely results in the increased flexibility of turn 6 observed in MD simulations (Fig. 6), movements toward the up position observed in the inactive zymogen (24) (Fig. 7), and transient interactions between E123 (on turn 6) and the catalytic H121 (Fig. 6). This finding is consistent with the two conserved water molecules that hydrogen-bond to E124 in turn 6, Wat48 and Wat65, having the largest fluctuations in the cluster (rmsf of 11.5 and 17.2 Å, respectively).

Second, our DRoP analysis and MD simulations show that 15 of 24 conserved waters that bind to helices 1 and 4 are disrupted in the mutant database, with a cluster of conserved waters bound to the C terminus of helix 4 (Fig. 8C, Lower) and a cluster of conserved waters bound to the N terminus of helix 4 (Fig. 8C, Upper). At present, the data support two modes of action for the conserved water clusters connecting helix 4 with the active site. First, the change in hydration of the C terminus of helix 4 may propagate along the helix toward active-site loop 3 through ionic interactions with charged or polar groups on helices 1 and 4, culminating in a change in mobility of D211 in active-site loop 3 (Fig. S7D). One observes that 9 of the 12 conserved waters in this network are disrupted in the caspase-3 mutants. Second, the loss of the conserved water molecules at the C terminus of helix 4 may propagate to the active site of the second protomer. Because of the ∼180° rotation of the two protomers, the active sites are on approximately opposite ends of the dimer (Fig. 1A), so the region of the C termini of helices 1 and 4 and the turn between helices 4 and 5 (Fig. 8C, Lower) are connected to the active site of the second protomer through proximity to the loop bundle of active site loops L2, L2′, and L4 (Fig. 1A). The conserved water molecules near the C terminus of helix 4 and the turn between helices 4 and 5 also connect the helices to binding sites for the N and C termini of the protein (Fig. S7E), and we find that he conserved water molecules in these binding sites are also disrupted in the caspase-3 mutants. In addition, the N and C termini show increased mobility in the mutants, as does the adjacent loop 2′ (Fig. S5 E and F). Similar to the fluctuations described above, the conserved waters that hydrogen-bond to D228 show the largest rmsf in the cluster, 14.2–17.2 Å.

Regardless of whether the loss of conserved water molecules is propagated through intraprotomer or interprotomer contacts, or both, the changes culminate in lower mobility in active-site loop 3 (Fig. S5F). Several amino acids in the helix 1,4,5 network are known to be modified on caspases, including glutathionylation of caspase-3 (34). In addition, the region is close to other modification sites, including phosphorylation of the N terminus of several caspases (35), so the conserved waters may be important for bridging communication networks between sites of PTMs and the active sites. In addition, helix 5 contains a binding site for zinc (29), ionic interactions (40), and phosphorylation sites (11) that are important in allosteric regulation of several caspases. Overall, the data show that the entire face of caspase-3 containing helices 1, 4, and 5, which is on the opposite side of the central β-sheet from the allosteric site, is perturbed by mutations in the dimer interface and that the loss of conserved water molecules may be propagated to the active site in the same protomer through helix 4 or to the active site of the second protomer through contacts near the loop bundle, including loop L2′. Together, our data show that multiple regions of the protein were affected by mutations in the allosteric site of the dimer interface.

Conclusions

Protein solvation mediates folding, stabilizes native structures, and is critical for enzyme function (41). Conversely, disruption of protein–water interactions may result in unfavorable changes in stability or dynamics (41, 42). Aside from the well-known changes in solvation for the conformational transitions of hemoglobin (43), relatively little is known about the role of water in allosteric transitions (44 ⇓⇓–47). Our study provides insight into the intrinsic nature of allosteric mechanisms and the role of water molecules in conformational selection. The analysis of water molecules in the caspase-3 mutational database, combined with structural and dynamic studies, shows that disrupting conserved waters on the protomer is correlated with a shift in the population of states toward an inactive conformation (state 2 in Fig. 1B). The data further describe the inactive state as a conformation with a distorted active site, transient interactions that may decrease the rate of the proton transfer of the catalytic histidine, desolvated through removal of conserved water molecules, and with lower mobility in two active-site loops. We note that the analysis reports on changes in the conserved waters within the protomer, particularly regarding the allosteric site of the central cavity, and also suggests that interprotomer interactions propagate the effects of desolvating the helix 1,4,5 face of the protein. Although the effects of mutations on the conserved water molecules are global, in that changes are observed throughout the protein, the data show that conserved waters may report on communication between allosteric sites facilitated by the bridging water molecules. This fact suggests that, in addition to targeting the allosteric site of the dimer interface to stabilize the disordered-loop conformation (state 3 in Fig. 1B), ligands that remove conserved water molecules from the protomer could effectively fine-tune caspase-3 activity by shifting the population toward the high-energy inactive conformation (state 2 in Fig. 1B).

The disordered-loop conformation (Fig. 1B, state 3) appears to be a low-energy state because it is readily trapped by small drug compounds. It is also observed in the so-called closed-loop conformations of several caspases, where binding of active site loop L2′ in the dimer interface prevents insertion of the elbow loop of L3 and subsequent formation of the substrate binding groove (20, 21, 36, 37, 48, 49). Interestingly, the allosteric site of the dimer interface is not modified by PTMs in the cell, and a global analysis of caspase modifications (35), along with trapping high-energy states of caspase-2 (28) and -3 (22, 23), shows that the disordered-loop conformation is only one of potentially several inactive states in the ensemble. Studies of the high-energy inactive conformations of caspase-2 and -3 show that disruption of the catalytic groups can be achieved without large-scale disordering of the active site loops, as small shifts in the catalytic cysteine and histidine are sufficient to inactivate the enzyme. The high-energy inactive state may provide an advantage for regulation of caspase activity in the cell, particularly if the state is stabilized by posttranslational modifications, because it provides the cell with a means to reversibly control activity by coupling one or more modifications to increased population of the inactive state. Removal of the modification would result in an unstable high-energy state that would rapidly convert to the more stable active conformation.

The database described here provides a tunable allosteric library of caspase-3 variants with nearly four orders of magnitude change in activity, which we suggest will be useful for examining caspase-3 signaling in cells. In current transfection technologies, caspase activity is manipulated in cells through knockdown or knockout strategies coupled with expression of caspase variants. Levels of endogenous caspase activity are difficult to control, and interpretation of results from overexpressed caspase variants can be problematic. New genome-editing techniques, however, combined with our caspase-3 database, should provide tools to fine-tune caspase-3 activity while simultaneously maintaining endogenous protein levels.

Methods

Cloning, Expression, and Protein Purification.

Site-directed mutagenesis was performed as described with plasmids pHC332 (wild-type caspase-3) (50) and pHC33209 [caspase-3(D9A,D28A,D175A); uncleavable procaspase-3] (50, 51) and the primers shown in Table S5. Escherichia coli BL21(DE3)pLysS cells were transformed with each of the plasmids, and proteins were expressed and purified as described (22, 31, 38, 52, 53).

View this table:

Table S5.

Plasmids and primers used to construct V266 library

Enzyme Activity Assay.

The initial velocity of substrate cleavage was measured at 25 °C in the presence of varying concentrations of substrate (Ac-DEVD-AFC), as described (52). The final protein concentration for the active mutants was 10 nM, whereas a protein concentration of 100 nM was used for the largely inactive mutants. The total reaction volume was 200 µL. Briefly, substrate was added to the sample, which contained protein in activity assay buffer (150 mM Tris⋅HCl, pH 7.5, 100 mM DTT, 0.1% CHAPS, 50 mM NaCl, and 1% sucrose), and samples were immediately excited at 400 nm, while the fluorescence emission was measured at 505 nm for 60 s.

Analysis of Water Clusters.

Water molecules were analyzed by using DRoP, as described (30), and the web server dropinthemattoslab.org/. Briefly, all structures were first aligned to that of wild-type caspase-3, [Protein Data Bank (PDB) ID code 2J30], and then loaded onto the web server. The DRoP program returns a PDB file that contains clustered and renumbered waters, as well as the level of conservation of each water molecule in the database. The conserved waters were further characterized based on the number of hydrogen bonds to the protein and to other water molecules: buried, form at least three H-bonds to protein side chains or backbone atoms; channel, generally form two H-bonds to the protein and at least one H-bond to other channel or buried water molecules; surface, form one or two H-bonds with the protein. For analysis of the caspase-3 mutants in the I222 space group, DRoP analysis was carried out after alignment with wild-type caspase-3 (PDB ID code 2J30). For caspase-3 mutants in the C121 space group, which consists of a dimer of protomers in the asymmetric unit, one protomer was removed from the file before alignment with wild-type caspase-3. After DRoP analysis, the conserved water molecules in each mutant were inspected for displacement.

Crystallization and Data Collection.

Caspase-3 variants were crystallized in the presence of Ac-DEVD-CMK as described (22, 31), and most crystals appeared within ∼3 d, although some took as long as 3 wk to grow. Cryoprotectants included 20% (vol/vol) PEG 400/80% reservoir solution or 20% (vol/vol) MPD/80% reservoir solution. Datasets were collected, and structures were solved by molecular replacement using the wild-type caspase-3 structure for initial phasing (PDB ID code 2J30); structural models were refined by using Phenix (54), as described (23, 31, 39). Structures were determined for all V266X variants, except for V266G, V266P, V266R, and V266T. The 13 structures determined were within 0.095 Å rmsd from wild-type caspase-3. A summary of the data collection and refinement statistics is shown in Table S4.

MD Simulations.

MD simulations were performed for 50 ns with GROMACS (Version 4.5) (55) by using the Amber99 force field (56) and the TIP3P water model (57), as described (22, 23). MD simulations of conserved water molecules were performed as described in SI Methods.

SI Methods

The structural model that is used to represent the protein and crystal waters, both conserved and nonconserved positions, is caspase-3 (PDB ID code 2J30). Protein molecules were represented by using the extended atom model. The conserved waters were arranged in a segment labeled XWAT, and the nonconserved waters in a segment labeled OWAT. The model was placed in an equilibrated box of size 80 × 76 × 76 Å with 11,316 TIP3 water molecules (58). The equilibrated water segment was labeled SOLV. SOLV segment waters that are within 2.8 Å of protein, XWAT, or OWAT molecules were deleted, leaving 7,975 SOLV molecules. All minimization and dynamics calculations were performed by using the MD and mechanics program CHARMM (59). The system was then energy-minimized by using 20 steps of Steepest Descents minimization with alternately the protein, XWAT, OWAT, and SOLV residues not constrained, and then 500 steps of Adopted Basis Set Newton Raphson energy minimization were performed with no constraints applied. MD then were performed for 200,000 steps of propagation using the Verlet Leapfrog algorithm after a heating routine that starts the randomly assigned velocities at 10 K and increases the temperature by 5 K each 10 steps until the simulation temperature reaches 300 K. Trajectory collection/scaling (scaling refers to reassignment of random velocities around 300 K) was then performed with no scaling events in the 200,000 steps of trajectory. The time step was 2 fs, electrostatics were modeled by using a constant dielectric of 1.0 and allowing the water in the simulation box to establish the local dielectric. Image centering was checked each 10 steps, and a coordinate set was saved every 10 steps. The shake algorithm was applied to bonds containing hydrogen. rmsfs were calculated using the analysis routines in CHARMM 36B2.

Acknowledgments

We thank Denise Appel and Matthew Willson for technical expertise. We also thank the research agencies of North Carolina State University and the North Carolina Agricultural Research Service. This work was supported by National Institutes of Health Grant GM065970 (to A.C.C.). Use of the Advanced Photon Source was supported by the US Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract W-31-109-ENG-38.

Footnotes

↵¹Present address: Institute for Genome Science, Duke University, Durham, NC 27708.
↵²To whom correspondence should be addressed. Email: clay.clark{at}uta.edu.

Author contributions: J.J.M., S.H.M., J.L.S., and A.C.C. designed research; J.J.M., S.H.M., M.B.T., J.L.S., and P.S. performed research; J.J.M., S.H.M., P.S., and A.C.C. analyzed data; and J.J.M. and A.C.C. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission. R.N. is a Guest Editor invited by the Editorial Board.
Data deposition: The crystallography, atomic coordinates, and structure factors have been deposited in the Protein Data Bank, www.pdb.org (PDB ID codes 5I9B, 5I9T, 5IAB, 5IAE, 5IBC, 5IBR, 5IAJ, 5IBP, 5IAN, 5IAG, 5IAK, 5IAR, and 5IAS).
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1603549113/-/DCSupplemental.

References

↵
1. Crawford ED,
2. Wells JA
(2011) Caspase substrates and cellular remodeling. Annu Rev Biochem 80:1055–1087.
.
OpenUrl CrossRef PubMed
↵
1. Schwerk C,
2. Schulze-Osthoff K
(2003) Non-apoptotic functions of caspases in cellular proliferation and differentiation. Biochem Pharmacol 66(8):1453–1458.
.
OpenUrl CrossRef PubMed
↵
1. D’Amelio M,
2. Cavallucci V,
3. Cecconi F
(2010) Neuronal caspase-3 signaling: Not only cell death. Cell Death Differ 17(7):1104–1114.
.
OpenUrl CrossRef PubMed
↵
1. Han M-H, et al.
(2013) The novel caspase-3 substrate Gap43 is involved in AMPA receptor endocytosis and long-term depression. Mol Cell Proteomics 12(12):3719–3731.
.
OpenUrl Abstract/FREE Full Text
↵
1. Edelmann B, et al.
(2011) Caspase-8 and caspase-7 sequentially mediate proteolytic activation of acid sphingomyelinase in TNF-R1 receptosomes. EMBO J 30(2):379–394.
.
OpenUrl Abstract/FREE Full Text
↵
1. Zandy AJ,
2. Lakhani S,
3. Zheng T,
4. Flavell RA,
5. Bassnett S
(2005) Role of the executioner caspases during lens development. J Biol Chem 280(34):30263–30272.
.
OpenUrl Abstract/FREE Full Text
↵
1. Makishima T, et al.
(2011) Inner ear dysfunction in caspase-3 deficient mice. BMC Neurosci 12:102.
.
OpenUrl CrossRef PubMed
↵
1. Boatright KM,
2. Salvesen GS
(2003) Mechanisms of caspase activation. Curr Opin Cell Biol 15(6):725–731.
.
OpenUrl CrossRef PubMed
↵
1. MacKenzie SH,
2. Clark AC
(2012) Death by caspase dimerization. Adv Exp Med Biol 747:55–73.
.
OpenUrl CrossRef PubMed
↵
1. Jiao S,
2. Li Z
(2011) Nonapoptotic function of BAD and BAX in long-term depression of synaptic transmission. Neuron 70(4):758–772.
.
OpenUrl CrossRef PubMed
↵
1. Parrish AB,
2. Freel CD,
3. Kornbluth S
(2013) Cellular mechanisms controlling caspase activation and function. Cold Spring Harb Perspect Biol 5(6):1–24.
.
OpenUrl CrossRef
↵
1. Chai J, et al.
(2001) Structural basis of caspase-7 inhibition by XIAP. Cell 104(5):769–780.
.
OpenUrl CrossRef PubMed
↵
1. Riedl SJ, et al.
(2001) Structural basis for the inhibition of caspase-3 by XIAP. Cell 104(5):791–800.
.
OpenUrl CrossRef PubMed
↵
1. Eckelman BP,
2. Salvesen GS
(2006) The human anti-apoptotic proteins cIAP1 and cIAP2 bind but do not inhibit caspases. J Biol Chem 281(6):3254–3260.
.
OpenUrl Abstract/FREE Full Text
↵
1. Choi YE, et al.
(2009) The E3 ubiquitin ligase cIAP1 binds and ubiquitinates caspase-3 and -7 via unique mechanisms at distinct steps in their processing. J Biol Chem 284(19):12772–12782.
.
OpenUrl Abstract/FREE Full Text
↵
1. Kurokawa M,
2. Kornbluth S
(2009) Caspases and kinases in a death grip. Cell 138(5):838–854.
.
OpenUrl CrossRef PubMed
↵
1. Voss OH,
2. Kim S,
3. Wewers MD,
4. Doseff AI
(2005) Regulation of monocyte apoptosis by the protein kinase Cdelta-dependent phosphorylation of caspase-3. J Biol Chem 280(17):17371–17379.
.
OpenUrl Abstract/FREE Full Text
↵
1. Dix MM, et al.
(2012) Functional interplay between caspase cleavage and phosphorylation sculpts the apoptotic proteome. Cell 150(2):426–440.
.
OpenUrl CrossRef PubMed
↵
1. Alvarado-Kristensson M, et al.
(2004) p38-MAPK signals survival by phosphorylation of caspase-8 and caspase-3 in human neutrophils. J Exp Med 199(4):449–458.
.
OpenUrl Abstract/FREE Full Text
↵
1. Fuentes-Prior P,
2. Salvesen GS
(2004) The protein structures that shape caspase activity, specificity, activation and inhibition. Biochem J 384(Pt 2):201–232.
.
OpenUrl Abstract/FREE Full Text
↵
1. Romanowski MJ,
2. Scheer JM,
3. O’Brien T,
4. McDowell RS
(2004) Crystal structures of a ligand-free and malonate-bound human caspase-1: Implications for the mechanism of substrate binding. Structure 12(8):1361–1371.
.
OpenUrl CrossRef PubMed
↵
1. Walters J,
2. Schipper JL,
3. Swartz P,
4. Mattos C,
5. Clark AC
(2012) Allosteric modulation of caspase 3 through mutagenesis. Biosci Rep 32(4):401–411.
.
OpenUrl Abstract/FREE Full Text
↵
1. Cade C,
2. Swartz P,
3. MacKenzie SH,
4. Clark AC
(2014) Modifying caspase-3 activity by altering allosteric networks. Biochemistry 53(48):7582–7595.
.
OpenUrl CrossRef PubMed
↵
1. Thomsen ND,
2. Koerber JT,
3. Wells JA
(2013) Structural snapshots reveal distinct mechanisms of procaspase-3 and -7 activation. Proc Natl Acad Sci USA 110(21):8477–8482.
.
OpenUrl Abstract/FREE Full Text
↵
1. Clark AC
(2016) Caspase allostery and conformational selection. Chem Rev 116(11):6666–6706.
.
OpenUrl CrossRef PubMed
↵
1. Hardy JA,
2. Lam J,
3. Nguyen JT,
4. O’Brien T,
5. Wells JA
(2004) Discovery of an allosteric site in the caspases. Proc Natl Acad Sci USA 101(34):12461–12466.
.
OpenUrl Abstract/FREE Full Text
↵
1. Scheer JM,
2. Romanowski MJ,
3. Wells JA
(2006) A common allosteric site and mechanism in caspases. Proc Natl Acad Sci USA 103(20):7595–7600.
.
OpenUrl Abstract/FREE Full Text
↵
1. Schweizer A, et al.
(2007) Inhibition of caspase-2 by a designed ankyrin repeat protein: specificity, structure, and inhibition mechanism. Structure 15(5):625–636.
.
OpenUrl CrossRef PubMed
↵
1. Velázquez-Delgado EM,
2. Hardy JA
(2012) Zinc-mediated allosteric inhibition of caspase-6. J Biol Chem 287(43):36000–36011.
.
OpenUrl Abstract/FREE Full Text
↵
1. Kearney BM,
2. Johnson CW,
3. Roberts DM,
4. Swartz P,
5. Mattos C
(2014) DRoP: A water analysis program identifies Ras-GTP-specific pathway of communication between membrane-interacting regions and the active site. J Mol Biol 426(3):611–629.
.
OpenUrl CrossRef PubMed
↵
1. Feeney B,
2. Pop C,
3. Swartz P,
4. Mattos C,
5. Clark AC
(2006) Role of loop bundle hydrogen bonds in the maturation and activity of (Pro)caspase-3. Biochemistry 45(44):13249–13263.
.
OpenUrl CrossRef PubMed
↵
1. Witkowski WA,
2. Hardy JA
(2009) L2′ loop is critical for caspase-7 active site formation. Protein Sci 18(7):1459–1468.
.
OpenUrl CrossRef PubMed
↵
1. Datta D,
2. Scheer JM,
3. Romanowski MJ,
4. Wells JA
(2008) An allosteric circuit in caspase-1. J Mol Biol 381(5):1157–1167.
.
OpenUrl CrossRef PubMed
↵
1. Huang Z,
2. Pinto JT,
3. Deng H,
4. Richie JP Jr
(2008) Inhibition of caspase-3 activity and activation by protein glutathionylation. Biochem Pharmacol 75(11):2234–2244.
.
OpenUrl CrossRef PubMed
↵
1. Dagbay K, et al.
(2014) A multipronged approach for compiling a global map of allosteric regulation in the apoptotic caspases. Methods Enzymol 544:215–249.
.
OpenUrl CrossRef PubMed
↵
1. Velázquez-Delgado EM,
2. Hardy JA
(2012) Phosphorylation regulates assembly of the caspase-6 substrate-binding groove. Structure 20(4):742–751.
.
OpenUrl CrossRef PubMed
↵
1. Cao Q, et al.
(2012) Inhibitory mechanism of caspase-6 phosphorylation revealed by crystal structures, molecular dynamics simulations, and biochemical assays. J Biol Chem 287(19):15371–15379.
.
OpenUrl Abstract/FREE Full Text
↵
1. Pop C,
2. Feeney B,
3. Tripathy A,
4. Clark AC
(2003) Mutations in the procaspase-3 dimer interface affect the activity of the zymogen. Biochemistry 42(42):12311–12320.
.
OpenUrl CrossRef PubMed
↵
1. Walters J, et al.
(2009) A constitutively active and uninhibitable caspase-3 zymogen efficiently induces apoptosis. Biochem J 424(3):335–345.
.
OpenUrl Abstract/FREE Full Text
↵
1. Feeney B,
2. Pop C,
3. Tripathy A,
4. Clark AC
(2004) Ionic interactions near the loop L4 are important for maintaining the active-site environment and the dimer stability of (pro)caspase 3. Biochem J 384(Pt 3):515–525.
.
OpenUrl Abstract/FREE Full Text
↵
1. Levy Y,
2. Onuchic JN
(2006) Water mediation in protein folding and molecular recognition. Annu Rev Biophys Biomol Struct 35:389–415.
.
OpenUrl CrossRef PubMed
↵
1. Frauenfelder H, et al.
(2009) A unified model of protein dynamics. Proc Natl Acad Sci USA 106(13):5129–5134.
.
OpenUrl Abstract/FREE Full Text
↵
1. Colombo MF,
2. Rau DC,
3. Parsegian VA
(1992) Protein solvation in allosteric regulation: A water effect on hemoglobin. Science 256(5057):655–659.
.
OpenUrl Abstract/FREE Full Text
↵
1. Motlagh HN,
2. Wrabl JO,
3. Li J,
4. Hilser VJ
(2014) The ensemble nature of allostery. Nature 508(7496):331–339.
.
OpenUrl CrossRef PubMed
↵
1. Davydov DR,
2. Baas BJ,
3. Sligar SG,
4. Halpert JR
(2007) Allosteric mechanisms in cytochrome P450 3A4 studied by high-pressure spectroscopy: Pivotal role of substrate-induced changes in the accessibility and degree of hydration of the heme pocket. Biochemistry 46(26):7852–7864.
.
OpenUrl CrossRef PubMed
↵
1. Zhuravlev PI,
2. Papoian GA
(2010) Protein functional landscapes, dynamics, allostery: A tortuous path towards a universal theoretical framework. Q Rev Biophys 43(3):295–332.
.
OpenUrl CrossRef PubMed
↵
1. Salvay AG,
2. Grigera JR,
3. Colombo MF
(2003) The role of hydration on the mechanism of allosteric regulation: In situ measurements of the oxygen-linked kinetics of water binding to hemoglobin. Biophys J 84(1):564–570.
.
OpenUrl CrossRef PubMed
↵
1. Riedl SJ, et al.
(2001) Structural basis for the activation of human procaspase-7. Proc Natl Acad Sci USA 98(26):14790–14795.
.
OpenUrl Abstract/FREE Full Text
↵
1. Chai J, et al.
(2001) Crystal structure of a procaspase-7 zymogen: Mechanisms of activation and substrate binding. Cell 107(3):399–407.
.
OpenUrl CrossRef PubMed
↵
1. Pop C, et al.
(2001) Removal of the pro-domain does not affect the conformation of the procaspase-3 dimer. Biochemistry 40(47):14224–14235.
.
OpenUrl CrossRef PubMed
↵
1. Bose K,
2. Clark AC
(2001) Dimeric procaspase-3 unfolds via a four-state equilibrium process. Biochemistry 40(47):14236–14242.
.
OpenUrl CrossRef PubMed
↵
1. Bose K,
2. Pop C,
3. Feeney B,
4. Clark AC
(2003) An uncleavable procaspase-3 mutant has a lower catalytic efficiency but an active site similar to that of mature caspase-3. Biochemistry 42(42):12298–12310.
.
OpenUrl CrossRef PubMed
↵
1. Walters J,
2. Swartz P,
3. Mattos C,
4. Clark AC
(2011) Thermodynamic, enzymatic and structural effects of removing a salt bridge at the base of loop 4 in (pro)caspase-3. Arch Biochem Biophys 508(1):31–38.
.
OpenUrl CrossRef PubMed
↵
1. Adams PD, et al.
(2010) PHENIX: A comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr 66(Pt 2):213–221.
.
OpenUrl CrossRef PubMed
↵
1. Pronk S, et al.
(2013) GROMACS 4.5: A high-throughput and highly parallel open source molecular simulation toolkit. Bioinformatics 29(7):845–854.
.
OpenUrl Abstract/FREE Full Text
↵
1. Wang J,
2. Cieplak P,
3. Kollman PA
(2000) How well does a restrained electrostatic potential (RESP) model perform in calculating conformational energies of organic and biological molecules? J Comput Chem 21:1049–1074.
.
OpenUrl CrossRef
↵
1. Jorgensen WL,
2. Chandrasekhar J,
3. Madura JD,
4. Impey RW,
5. Klein ML
(1983) Comparison of simple potential functions for simulating liquid water. J Chem Phys 79:926–935.
.
OpenUrl CrossRef
↵
1. Jorgensen WL,
2. Jorgensen WL
(1982) Revised TIPS for simulations of liquid water and aqueous solutions. J Chem Phys 77:4156–4163.
.
OpenUrl CrossRef
↵
1. Brooks BR, et al.
(1983) CHARMM: A program for macromolecular energy, minimization, and dynamics calculations. J Comput Chem 4:187–217.
.
OpenUrl CrossRef
1. Erlanson DA, et al.
(2003) In situ assembly of enzyme inhibitors using extended tethering. Nat Biotechnol 21(3):308–314.
.
OpenUrl CrossRef PubMed
1. Ni C-Z,
2. Li C,
3. Wu JC,
4. Spada AP,
5. Ely KR
(2003) Conformational restrictions in the active site of unliganded human caspase-3. J Mol Recognit 16(3):121–124.
.
OpenUrl CrossRef PubMed
1. Ekici OD, et al.
(2006) Design, synthesis, and evaluation of aza-peptide Michael acceptors as selective and potent inhibitors of caspases-2, -3, -6, -7, -8, -9, and -10. J Med Chem 49(19):5728–5749.
.
OpenUrl CrossRef PubMed
1. Ganesan R, et al.
(2006) Exploring the S4 and S1 prime subsite specificities in caspase-3 with aza-peptide epoxide inhibitors. Biochemistry 45(30):9059–9067.
.
OpenUrl CrossRef PubMed
1. Ganesan R,
2. Mittl PRE,
3. Jelakovic S,
4. Grütter MG
(2006) Extended substrate recognition in caspase-3 revealed by high resolution X-ray structure analysis. J Mol Biol 359(5):1378–1388.
.
OpenUrl CrossRef PubMed
1. Ganesan R,
2. Jelakovic S,
3. Mittl PRE,
4. Caflisch A,
5. Grütter MG
(2011) In silico identification and crystal structure validation of caspase-3 inhibitors without a P1 aspartic acid moiety. Acta Crystallogr Sect F Struct Biol Cryst Commun 67(Pt 8):842–850.
.
OpenUrl CrossRef PubMed
1. Vickers CJ,
2. González-Páez GE,
3. Wolan DW
(2013) Selective detection of caspase-3 versus caspase-7 using activity-based probes with key unnatural amino acids. ACS Chem Biol 8(7):1558–1566.
.
OpenUrl CrossRef PubMed

Article Alerts

Email Article

Citation Tools

Request Permissions

Article Classifications

Biological Sciences
Biophysics and Computational Biology

Proceedings of the National Academy of Sciences: 113 (41)

Table of Contents

Submit

[1] ↵
Crawford ED,
Wells JA
(2011) Caspase substrates and cellular remodeling. Annu Rev Biochem 80:1055–1087.
.
OpenUrl CrossRef PubMed

[2] Crawford ED,

[3] Wells JA

[4] ↵
Schwerk C,
Schulze-Osthoff K
(2003) Non-apoptotic functions of caspases in cellular proliferation and differentiation. Biochem Pharmacol 66(8):1453–1458.
.
OpenUrl CrossRef PubMed

[5] Schwerk C,

[6] Schulze-Osthoff K

[7] ↵
D’Amelio M,
Cavallucci V,
Cecconi F
(2010) Neuronal caspase-3 signaling: Not only cell death. Cell Death Differ 17(7):1104–1114.
.
OpenUrl CrossRef PubMed

[8] D’Amelio M,

[9] Cavallucci V,

[10] Cecconi F

[11] ↵
Han M-H, et al.
(2013) The novel caspase-3 substrate Gap43 is involved in AMPA receptor endocytosis and long-term depression. Mol Cell Proteomics 12(12):3719–3731.
.
OpenUrl Abstract/FREE Full Text

[12] Han M-H, et al.

[13] ↵
Edelmann B, et al.
(2011) Caspase-8 and caspase-7 sequentially mediate proteolytic activation of acid sphingomyelinase in TNF-R1 receptosomes. EMBO J 30(2):379–394.
.
OpenUrl Abstract/FREE Full Text

[14] Edelmann B, et al.

[15] ↵
Zandy AJ,
Lakhani S,
Zheng T,
Flavell RA,
Bassnett S
(2005) Role of the executioner caspases during lens development. J Biol Chem 280(34):30263–30272.
.
OpenUrl Abstract/FREE Full Text

[16] Zandy AJ,

[17] Lakhani S,

[18] Zheng T,

[19] Flavell RA,

[20] Bassnett S

[21] ↵
Makishima T, et al.
(2011) Inner ear dysfunction in caspase-3 deficient mice. BMC Neurosci 12:102.
.
OpenUrl CrossRef PubMed

[22] Makishima T, et al.

[23] ↵
Boatright KM,
Salvesen GS
(2003) Mechanisms of caspase activation. Curr Opin Cell Biol 15(6):725–731.
.
OpenUrl CrossRef PubMed

[24] Boatright KM,

[25] Salvesen GS

[26] ↵
MacKenzie SH,
Clark AC
(2012) Death by caspase dimerization. Adv Exp Med Biol 747:55–73.
.
OpenUrl CrossRef PubMed

[27] MacKenzie SH,

[28] Clark AC

[29] ↵
Jiao S,
Li Z
(2011) Nonapoptotic function of BAD and BAX in long-term depression of synaptic transmission. Neuron 70(4):758–772.
.
OpenUrl CrossRef PubMed

[30] Jiao S,

[31] Li Z

[32] ↵
Parrish AB,
Freel CD,
Kornbluth S
(2013) Cellular mechanisms controlling caspase activation and function. Cold Spring Harb Perspect Biol 5(6):1–24.
.
OpenUrl CrossRef

[33] Parrish AB,

[34] Freel CD,

[35] Kornbluth S

[36] ↵
Chai J, et al.
(2001) Structural basis of caspase-7 inhibition by XIAP. Cell 104(5):769–780.
.
OpenUrl CrossRef PubMed

[37] Chai J, et al.

[38] ↵
Riedl SJ, et al.
(2001) Structural basis for the inhibition of caspase-3 by XIAP. Cell 104(5):791–800.
.
OpenUrl CrossRef PubMed

[39] Riedl SJ, et al.

[40] ↵
Eckelman BP,
Salvesen GS
(2006) The human anti-apoptotic proteins cIAP1 and cIAP2 bind but do not inhibit caspases. J Biol Chem 281(6):3254–3260.
.
OpenUrl Abstract/FREE Full Text

[41] Eckelman BP,

[42] Salvesen GS

[43] ↵
Choi YE, et al.
(2009) The E3 ubiquitin ligase cIAP1 binds and ubiquitinates caspase-3 and -7 via unique mechanisms at distinct steps in their processing. J Biol Chem 284(19):12772–12782.
.
OpenUrl Abstract/FREE Full Text

[44] Choi YE, et al.

[45] ↵
Kurokawa M,
Kornbluth S
(2009) Caspases and kinases in a death grip. Cell 138(5):838–854.
.
OpenUrl CrossRef PubMed

[46] Kurokawa M,

[47] Kornbluth S

[48] ↵
Voss OH,
Kim S,
Wewers MD,
Doseff AI
(2005) Regulation of monocyte apoptosis by the protein kinase Cdelta-dependent phosphorylation of caspase-3. J Biol Chem 280(17):17371–17379.
.
OpenUrl Abstract/FREE Full Text

[49] Voss OH,

[50] Kim S,

[51] Wewers MD,

[52] Doseff AI

[53] ↵
Dix MM, et al.
(2012) Functional interplay between caspase cleavage and phosphorylation sculpts the apoptotic proteome. Cell 150(2):426–440.
.
OpenUrl CrossRef PubMed

[54] Dix MM, et al.

[55] ↵
Alvarado-Kristensson M, et al.
(2004) p38-MAPK signals survival by phosphorylation of caspase-8 and caspase-3 in human neutrophils. J Exp Med 199(4):449–458.
.
OpenUrl Abstract/FREE Full Text

[56] Alvarado-Kristensson M, et al.

[57] ↵
Fuentes-Prior P,
Salvesen GS
(2004) The protein structures that shape caspase activity, specificity, activation and inhibition. Biochem J 384(Pt 2):201–232.
.
OpenUrl Abstract/FREE Full Text

[58] Fuentes-Prior P,

[59] Salvesen GS

[60] ↵
Romanowski MJ,
Scheer JM,
O’Brien T,
McDowell RS
(2004) Crystal structures of a ligand-free and malonate-bound human caspase-1: Implications for the mechanism of substrate binding. Structure 12(8):1361–1371.
.
OpenUrl CrossRef PubMed

[61] Romanowski MJ,

[62] Scheer JM,

[63] O’Brien T,

[64] McDowell RS

[65] ↵
Walters J,
Schipper JL,
Swartz P,
Mattos C,
Clark AC
(2012) Allosteric modulation of caspase 3 through mutagenesis. Biosci Rep 32(4):401–411.
.
OpenUrl Abstract/FREE Full Text

[66] Walters J,

[67] Schipper JL,

[68] Swartz P,

[69] Mattos C,

[70] Clark AC

[71] ↵
Cade C,
Swartz P,
MacKenzie SH,
Clark AC
(2014) Modifying caspase-3 activity by altering allosteric networks. Biochemistry 53(48):7582–7595.
.
OpenUrl CrossRef PubMed

[72] Cade C,

[73] Swartz P,

[74] MacKenzie SH,

[75] Clark AC

[76] ↵
Thomsen ND,
Koerber JT,
Wells JA
(2013) Structural snapshots reveal distinct mechanisms of procaspase-3 and -7 activation. Proc Natl Acad Sci USA 110(21):8477–8482.
.
OpenUrl Abstract/FREE Full Text

[77] Thomsen ND,

[78] Koerber JT,

[79] Wells JA

[80] ↵
Clark AC
(2016) Caspase allostery and conformational selection. Chem Rev 116(11):6666–6706.
.
OpenUrl CrossRef PubMed

[81] Clark AC

[82] ↵
Hardy JA,
Lam J,
Nguyen JT,
O’Brien T,
Wells JA
(2004) Discovery of an allosteric site in the caspases. Proc Natl Acad Sci USA 101(34):12461–12466.
.
OpenUrl Abstract/FREE Full Text

[83] Hardy JA,

[84] Lam J,

[85] Nguyen JT,

[86] O’Brien T,

[87] Wells JA

[88] ↵
Scheer JM,
Romanowski MJ,
Wells JA
(2006) A common allosteric site and mechanism in caspases. Proc Natl Acad Sci USA 103(20):7595–7600.
.
OpenUrl Abstract/FREE Full Text

[89] Scheer JM,

[90] Romanowski MJ,

[91] Wells JA

[92] ↵
Schweizer A, et al.
(2007) Inhibition of caspase-2 by a designed ankyrin repeat protein: specificity, structure, and inhibition mechanism. Structure 15(5):625–636.
.
OpenUrl CrossRef PubMed

[93] Schweizer A, et al.

[94] ↵
Velázquez-Delgado EM,
Hardy JA
(2012) Zinc-mediated allosteric inhibition of caspase-6. J Biol Chem 287(43):36000–36011.
.
OpenUrl Abstract/FREE Full Text

[95] Velázquez-Delgado EM,

[96] Hardy JA

[97] ↵
Kearney BM,
Johnson CW,
Roberts DM,
Swartz P,
Mattos C
(2014) DRoP: A water analysis program identifies Ras-GTP-specific pathway of communication between membrane-interacting regions and the active site. J Mol Biol 426(3):611–629.
.
OpenUrl CrossRef PubMed

[98] Kearney BM,

[99] Johnson CW,

[100] Roberts DM,

[101] Swartz P,

[102] Mattos C

[103] ↵
Feeney B,
Pop C,
Swartz P,
Mattos C,
Clark AC
(2006) Role of loop bundle hydrogen bonds in the maturation and activity of (Pro)caspase-3. Biochemistry 45(44):13249–13263.
.
OpenUrl CrossRef PubMed

[104] Feeney B,

[105] Pop C,

[106] Swartz P,

[107] Mattos C,

[108] Clark AC

[109] ↵
Witkowski WA,
Hardy JA
(2009) L2′ loop is critical for caspase-7 active site formation. Protein Sci 18(7):1459–1468.
.
OpenUrl CrossRef PubMed

[110] Witkowski WA,

[111] Hardy JA

[112] ↵
Datta D,
Scheer JM,
Romanowski MJ,
Wells JA
(2008) An allosteric circuit in caspase-1. J Mol Biol 381(5):1157–1167.
.
OpenUrl CrossRef PubMed

[113] Datta D,

[114] Scheer JM,

[115] Romanowski MJ,

[116] Wells JA

[117] ↵
Huang Z,
Pinto JT,
Deng H,
Richie JP Jr
(2008) Inhibition of caspase-3 activity and activation by protein glutathionylation. Biochem Pharmacol 75(11):2234–2244.
.
OpenUrl CrossRef PubMed

[118] Huang Z,

[119] Pinto JT,

[120] Deng H,

[121] Richie JP Jr

[122] ↵
Dagbay K, et al.
(2014) A multipronged approach for compiling a global map of allosteric regulation in the apoptotic caspases. Methods Enzymol 544:215–249.
.
OpenUrl CrossRef PubMed

[123] Dagbay K, et al.

[124] ↵
Velázquez-Delgado EM,
Hardy JA
(2012) Phosphorylation regulates assembly of the caspase-6 substrate-binding groove. Structure 20(4):742–751.
.
OpenUrl CrossRef PubMed

[125] Velázquez-Delgado EM,

[126] Hardy JA

[127] ↵
Cao Q, et al.
(2012) Inhibitory mechanism of caspase-6 phosphorylation revealed by crystal structures, molecular dynamics simulations, and biochemical assays. J Biol Chem 287(19):15371–15379.
.
OpenUrl Abstract/FREE Full Text

[128] Cao Q, et al.

[129] ↵
Pop C,
Feeney B,
Tripathy A,
Clark AC
(2003) Mutations in the procaspase-3 dimer interface affect the activity of the zymogen. Biochemistry 42(42):12311–12320.
.
OpenUrl CrossRef PubMed

[130] Pop C,

[131] Feeney B,

[132] Tripathy A,

[133] Clark AC

[134] ↵
Walters J, et al.
(2009) A constitutively active and uninhibitable caspase-3 zymogen efficiently induces apoptosis. Biochem J 424(3):335–345.
.
OpenUrl Abstract/FREE Full Text

[135] Walters J, et al.

[136] ↵
Feeney B,
Pop C,
Tripathy A,
Clark AC
(2004) Ionic interactions near the loop L4 are important for maintaining the active-site environment and the dimer stability of (pro)caspase 3. Biochem J 384(Pt 3):515–525.
.
OpenUrl Abstract/FREE Full Text

[137] Feeney B,

[138] Pop C,

[139] Tripathy A,

[140] Clark AC

[141] ↵
Levy Y,
Onuchic JN
(2006) Water mediation in protein folding and molecular recognition. Annu Rev Biophys Biomol Struct 35:389–415.
.
OpenUrl CrossRef PubMed

[142] Levy Y,

[143] Onuchic JN

[144] ↵
Frauenfelder H, et al.
(2009) A unified model of protein dynamics. Proc Natl Acad Sci USA 106(13):5129–5134.
.
OpenUrl Abstract/FREE Full Text

[145] Frauenfelder H, et al.

[146] ↵
Colombo MF,
Rau DC,
Parsegian VA
(1992) Protein solvation in allosteric regulation: A water effect on hemoglobin. Science 256(5057):655–659.
.
OpenUrl Abstract/FREE Full Text

[147] Colombo MF,

[148] Rau DC,

[149] Parsegian VA

[150] ↵
Motlagh HN,
Wrabl JO,
Li J,
Hilser VJ
(2014) The ensemble nature of allostery. Nature 508(7496):331–339.
.
OpenUrl CrossRef PubMed

[151] Motlagh HN,

[152] Wrabl JO,

[153] Li J,

[154] Hilser VJ

[155] ↵
Davydov DR,
Baas BJ,
Sligar SG,
Halpert JR
(2007) Allosteric mechanisms in cytochrome P450 3A4 studied by high-pressure spectroscopy: Pivotal role of substrate-induced changes in the accessibility and degree of hydration of the heme pocket. Biochemistry 46(26):7852–7864.
.
OpenUrl CrossRef PubMed

[156] Davydov DR,

[157] Baas BJ,

[158] Sligar SG,

[159] Halpert JR

[160] ↵
Zhuravlev PI,
Papoian GA
(2010) Protein functional landscapes, dynamics, allostery: A tortuous path towards a universal theoretical framework. Q Rev Biophys 43(3):295–332.
.
OpenUrl CrossRef PubMed

[161] Zhuravlev PI,

[162] Papoian GA

[163] ↵
Salvay AG,
Grigera JR,
Colombo MF
(2003) The role of hydration on the mechanism of allosteric regulation: In situ measurements of the oxygen-linked kinetics of water binding to hemoglobin. Biophys J 84(1):564–570.
.
OpenUrl CrossRef PubMed

[164] Salvay AG,

[165] Grigera JR,

[166] Colombo MF

[167] ↵
Riedl SJ, et al.
(2001) Structural basis for the activation of human procaspase-7. Proc Natl Acad Sci USA 98(26):14790–14795.
.
OpenUrl Abstract/FREE Full Text

[168] Riedl SJ, et al.

[169] ↵
Chai J, et al.
(2001) Crystal structure of a procaspase-7 zymogen: Mechanisms of activation and substrate binding. Cell 107(3):399–407.
.
OpenUrl CrossRef PubMed

[170] Chai J, et al.

[171] ↵
Pop C, et al.
(2001) Removal of the pro-domain does not affect the conformation of the procaspase-3 dimer. Biochemistry 40(47):14224–14235.
.
OpenUrl CrossRef PubMed

[172] Pop C, et al.

[173] ↵
Bose K,
Clark AC
(2001) Dimeric procaspase-3 unfolds via a four-state equilibrium process. Biochemistry 40(47):14236–14242.
.
OpenUrl CrossRef PubMed

[174] Bose K,

[175] Clark AC

[176] ↵
Bose K,
Pop C,
Feeney B,
Clark AC
(2003) An uncleavable procaspase-3 mutant has a lower catalytic efficiency but an active site similar to that of mature caspase-3. Biochemistry 42(42):12298–12310.
.
OpenUrl CrossRef PubMed

[177] Bose K,

[178] Pop C,

[179] Feeney B,

[180] Clark AC

[181] ↵
Walters J,
Swartz P,
Mattos C,
Clark AC
(2011) Thermodynamic, enzymatic and structural effects of removing a salt bridge at the base of loop 4 in (pro)caspase-3. Arch Biochem Biophys 508(1):31–38.
.
OpenUrl CrossRef PubMed

[182] Walters J,

[183] Swartz P,

[184] Mattos C,

[185] Clark AC

[186] ↵
Adams PD, et al.
(2010) PHENIX: A comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr 66(Pt 2):213–221.
.
OpenUrl CrossRef PubMed

[187] Adams PD, et al.

[188] ↵
Pronk S, et al.
(2013) GROMACS 4.5: A high-throughput and highly parallel open source molecular simulation toolkit. Bioinformatics 29(7):845–854.
.
OpenUrl Abstract/FREE Full Text

[189] Pronk S, et al.

[190] ↵
Wang J,
Cieplak P,
Kollman PA
(2000) How well does a restrained electrostatic potential (RESP) model perform in calculating conformational energies of organic and biological molecules? J Comput Chem 21:1049–1074.
.
OpenUrl CrossRef

[191] Wang J,

[192] Cieplak P,

[193] Kollman PA

[194] ↵
Jorgensen WL,
Chandrasekhar J,
Madura JD,
Impey RW,
Klein ML
(1983) Comparison of simple potential functions for simulating liquid water. J Chem Phys 79:926–935.
.
OpenUrl CrossRef

[195] Jorgensen WL,

[196] Chandrasekhar J,

[197] Madura JD,

[198] Impey RW,

[199] Klein ML

[200] ↵
Jorgensen WL,
Jorgensen WL
(1982) Revised TIPS for simulations of liquid water and aqueous solutions. J Chem Phys 77:4156–4163.
.
OpenUrl CrossRef

[201] Jorgensen WL,

[202] Jorgensen WL

[203] ↵
Brooks BR, et al.
(1983) CHARMM: A program for macromolecular energy, minimization, and dynamics calculations. J Comput Chem 4:187–217.
.
OpenUrl CrossRef

[204] Brooks BR, et al.

[205] Erlanson DA, et al.
(2003) In situ assembly of enzyme inhibitors using extended tethering. Nat Biotechnol 21(3):308–314.
.
OpenUrl CrossRef PubMed

[206] Erlanson DA, et al.

[207] Ni C-Z,
Li C,
Wu JC,
Spada AP,
Ely KR
(2003) Conformational restrictions in the active site of unliganded human caspase-3. J Mol Recognit 16(3):121–124.
.
OpenUrl CrossRef PubMed

[208] Ni C-Z,

[209] Li C,

[210] Wu JC,

[211] Spada AP,

[212] Ely KR

[213] Ekici OD, et al.
(2006) Design, synthesis, and evaluation of aza-peptide Michael acceptors as selective and potent inhibitors of caspases-2, -3, -6, -7, -8, -9, and -10. J Med Chem 49(19):5728–5749.
.
OpenUrl CrossRef PubMed

[214] Ekici OD, et al.

[215] Ganesan R, et al.
(2006) Exploring the S4 and S1 prime subsite specificities in caspase-3 with aza-peptide epoxide inhibitors. Biochemistry 45(30):9059–9067.
.
OpenUrl CrossRef PubMed

[216] Ganesan R, et al.

[217] Ganesan R,
Mittl PRE,
Jelakovic S,
Grütter MG
(2006) Extended substrate recognition in caspase-3 revealed by high resolution X-ray structure analysis. J Mol Biol 359(5):1378–1388.
.
OpenUrl CrossRef PubMed

[218] Ganesan R,

[219] Mittl PRE,

[220] Jelakovic S,

[221] Grütter MG

[222] Ganesan R,
Jelakovic S,
Mittl PRE,
Caflisch A,
Grütter MG
(2011) In silico identification and crystal structure validation of caspase-3 inhibitors without a P1 aspartic acid moiety. Acta Crystallogr Sect F Struct Biol Cryst Commun 67(Pt 8):842–850.
.
OpenUrl CrossRef PubMed

[223] Ganesan R,

[224] Jelakovic S,

[225] Mittl PRE,

[226] Caflisch A,

[227] Grütter MG

[228] Vickers CJ,
González-Páez GE,
Wolan DW
(2013) Selective detection of caspase-3 versus caspase-7 using activity-based probes with key unnatural amino acids. ACS Chem Biol 8(7):1558–1566.
.
OpenUrl CrossRef PubMed

[229] Vickers CJ,

[230] González-Páez GE,

[231] Wolan DW

Main menu

User menu

Search

Tunable allosteric library of caspase-3 identifies coupling between conserved water molecules and conformational selection

Significance

Abstract

Results and Discussion

Global Analysis of Caspase-3 Solvation.

Saturation Mutagenesis in an Allosteric Site Provides a Library of Caspase Variants with a Broad Range of Activities.

Structural and Dynamic Studies Show That Flexibility in the Allosteric Site Disrupts Conserved Waters and Propagates to the Active Site.

Lower Activity Correlates with Changes in Solvation.

Conclusions

Methods

Cloning, Expression, and Protein Purification.

Enzyme Activity Assay.

Analysis of Water Clusters.

Crystallization and Data Collection.

MD Simulations.

SI Methods

Acknowledgments

Footnotes

References

Citation Manager Formats

Article Classifications

You May Also be Interested in

Similar Articles

Articles

PNAS Portals

Information

Main menu

User menu

Search

Significance

Abstract

Results and Discussion

Global Analysis of Caspase-3 Solvation.

Saturation Mutagenesis in an Allosteric Site Provides a Library of Caspase Variants with a Broad Range of Activities.

Structural and Dynamic Studies Show That Flexibility in the Allosteric Site Disrupts Conserved Waters and Propagates to the Active Site.

Lower Activity Correlates with Changes in Solvation.

Conclusions

Methods

Cloning, Expression, and Protein Purification.

Enzyme Activity Assay.

Analysis of Water Clusters.

Crystallization and Data Collection.

MD Simulations.

SI Methods

Acknowledgments

Footnotes

References

Citation Manager Formats

Article Classifications

Sign up for Article Alerts

Jump to section

You May Also be Interested in

Similar Articles

Articles

PNAS Portals

Information