We have examined the enzymatic activity of an uncleavable procaspase-3 mutant (D9A/D28A/D175A), which contains the wild-type catalytic residues in the active site. The results are compared to those for the mature caspase-3. Although at pH 7.5 and 25 degrees C the K(m) values are similar, the catalytic efficiency (k(cat)) is approximately 130-fold lower in the zymogen. The mature caspase-3 demonstrates a maximum activity at pH 7.4, whereas the maximum activity of procaspase-3 occurs at pH 8.3. The pK(a) values of both catalytic groups, H121 and C163, are shifted to higher pH for procaspase-3. We developed limited proteolysis assays using trypsin and V8 proteases, and we show that these assays allow the examination of amino acids in three of five active site loops. In addition, we examined the fluorescence emission of the two tryptophanyl residues in the active site over the pH range of 2.5-9 as well as the response to several quenching agents. Overall, the data suggest that the major conformational change that occurs upon maturation results in formation of the loop bundle among loops L4, L2, and L2'. The pK(a) values of both catalytic groups decrease as a result of the loop movements. However, loop L3, which comprises the bulk of the substrate binding pocket, does not appear to be unraveled and solvent-exposed, even at lower pH.

During maturation, procaspase-3 is cleaved at D175, which resides in a linker that connects the large and small subunits. The intersubunit linker also connects two active site loops that rearrange following cleavage and, in part, form the so-called loop bundle. As a result of chain cleavage, new hydrogen bonds and van der Waals contacts form among three active site loops. The new interactions are predicted to stabilize the active site. One unresolved issue is the extent to which the loop bundle residues also stabilize the procaspase active site. We examined the effects of replacing four loop bundle residues (E167, D169, E173, and Y203) on the biochemical and structural properties of the (pro)caspase. We show that replacing the residues affects the activity of the procaspase as well as the mature caspase, with D169A and E167A replacements having the largest effects. Replacement of D169 prevents caspase-3 autoactivation, and its cleavage at D175 no longer leads to an active enzyme. In addition, the E173A mutation, when coupled to a second mutation in the procaspase, D175A, may alter the substrate specificity of the procaspase. The mutations affected the active site environment as assessed by changes in fluorescence emission, accessibility to quencher, and cleavage by either trypsin or V8 proteases. High-resolution X-ray crystallographic structures of E167A, D173A, and Y203F caspases show that changes in the active site environment may be due to the increased flexibility of several residues in the N-terminus of the small subunit. Overall, the results show that these residues are important for stabilizing the procaspase active site as well as that of the mature caspase.

The interface of the procaspase-3 dimer plays a critical role in zymogen maturation. We show that replacement of valine 266, the residue at the center of the procaspase-3 dimer interface, with glutamate resulted in an increase in enzyme activity of ~60-fold, representing a pseudoactivation of the procaspase. In contrast, substitution of V266 with histidine abolished the activity of the procaspase-3 as well as that of the mature caspase. While the mutations do not affect the dimeric properties of the procaspase, we show that the V266E mutation may affect the formation of a loop bundle that is important for stabilizing the active site. In contrast, the V266H mutation affects the positioning of loop L3, the loop that forms the bulk of the substrate binding pocket. In some cases, the amino acids affected by the mutations are >20 Å from the interface. Overall, the results demonstrate that the integrity of the dimer interface is important for maintaining the proper active site conformation.Programmed cell death is dependent on the maturation of the effector caspases from latent zymogens. While the procaspases are present at relatively high concentrations in the cell, it is not clear why they are enzymatically inactive. The mature caspase is a dimer of heterodimeric units containing a large subunit (17 kDa) and a small subunit (12 kDa). In the procaspase monomer, the structural units are organized as a short pro domain (28 amino acids), the large subunit, an intersubunit linker (25 amino acids), and the small subunit. Two monomers assemble into the procaspase homodimer that consists of a contiguous 12-stranded β-sheet core with several helices surrounding the β-sheet (see Figure 1). The procaspase is cleaved at D175, in the intersubunit linker, to remove the covalent connection between the subunits. The pro domain is then removed after cleavage at D9 and then at D28. Recently determined structures of procaspase-7 (1, 2) show that the core structural unit of the procaspase is comparable to that of the mature caspase (see Figure 1). The primary differences reside in five active site loops (for a review, see ref 3). Following maturation of procaspase-7, only the position of loop L1 (residues 52-67, caspase-3 numbering) remains unaltered. The data show that the procaspase is not enzymatically active because loop L3 (residues 198-213) is unraveled and positioned away from the active site, and the catalytic C163 is rotated away from solvent so that it cannot attack the substrate. The positioning of † This work was supported by a grant from the National Institutes of Health (GM065970) NIH-PA Author ManuscriptNIH-PA Author Manuscript NIH-PA Author ManuscriptC163 is likely due to the covalent connection between loop L2 (residues 163-175) and the intersubunit linker (residues 176-192). Following cleavage at D175, loop L3 moves more than 10 Å toward the protein core to form the substrate binding pocket. Loop L2 moves toward L4, and the intersubunit linker, now called loop L2′, flips 180° to interact with loops L2 and L...

Changes in ionic homeostasis are early events leading up to the commitment to apoptosis. Although the direct effects of cations on caspase-3 activity have been examined, comparable studies on procaspase-3 are lacking. In addition, the effects of salts on caspase structure have not been examined. We have studied the effects of cations on the activities and conformations of caspase-3 and an uncleavable mutant of procaspase-3 that is enzymatically active. The results show that caspase-3 is more sensitive to changes in pH and ion concentrations than is the zymogen. This is due to the loss of both an intact intersubunit linker and the prodomain. The results show that, although the caspase-3 subunits reassemble to the heterotetramer, the activity return is low after the protein is incubated at or below pH 4.5 and then returned to pH 7.5. The data further show that the irreversible step in assembly results from heterotetramer rather than heterodimer dissociation and demonstrate that the active site does not form properly following reassembly. However, active-site formation is fully reversible when reassembly occurs in the presence of the prodomain, and this effect is specific for the propeptide of caspase-3. The data show that the prodomain facilitates both dimerization and active-site formation in addition to stabilizing the native structure. Overall, the results show that the prodomain acts as an intramolecular chaperone during assembly of the (pro)caspase subunits and increases the efficiency of formation of the native conformation.

Global transition state regulator proteins represent one of the most diverse classes of prokaryotic transcription factors. One such transition state regulator, AbrB from Bacillus subtilis, is known to bind more than 60 gene targets yet displays specificity within this target set by binding each promoter with a different affinity. Microelectrospray ionization mass spectrometry (microESI-MS), circular dichroism, fluorescence, UV spectroscopy, and molecular modeling were used to elucidate differences among AbrB, DNA, and AbrB-DNA complexes. MicroESI-MS analysis of AbrB confirmed its stable macromolecular state as being tetrameric and verified the same stoichiometric state in complex with DNA targets. MicroESI-MS, circular dichroism, and fluorescence provided relative binding affinities for AbrB-DNA interactions in a qualitative manner. UV spectroscopy was used in a quantitative manner to determine solution phase dissociation constants for AbrB-DNA complexes. General DNA structural parameters for all known natural AbrB binding sequences were also studied and significant similarities in topological constraints (stretch, opening, and propeller twist) were observed. It is likely that these parameters contribute to the differential binding proclivities of AbrB. In addition to providing an improved understanding of transition state regulator-DNA binding properties and structural tendencies of target promoters, this comprehensive and corroborative spectroscopic study endorses the use of microESI-MS for rapidly ascertaining qualitative binding trends in noncovalent systems in a high-throughput manner.

We have examined the role of a salt bridge between Lys242 and Glu246 in loop L4 of procaspase 3 and of mature caspase 3, and we show that the interactions are required for stabilizing the active site. Replacing either of the residues with an alanine residue results in a complete loss of procaspase 3 activity. Although both mutants are active in the context of the mature caspase 3, the mutations result in an increase in K(m) and a decrease in kcat when compared with the wild-type caspase 3. In addition, the mutations result in an increase in the pK(a) value associated with a change in kcat with pH, but does not affect the transition observed for Km versus pH. The mutations also affect the accessibility of the active-site solvent as measured by tryptophan fluorescence emission in the presence of quenching agents and as a function of pH. We show that, as the pH is lowered, the (pro)caspase dissociates, and the mutations increase the pH-dependent instability of the dimer. Overall, the results suggest that the contacts lost in the procaspase as a result of replacing Lys242 and Glu246 are compensated partially in the mature caspase as a result of new contacts that are known to form on zymogen processing

Caspase recruitment domains (CARDs) are members of the death domain superfamily and contain six antiparallel helices in an alpha-helical Greek key topology. We have examined the equilibrium and kinetic folding of the CARD of Apaf-1 (apoptotic protease activating factor 1), which consists of 97 amino acid residues, at pH 6 and pH 8. The results showed that an apparent two state equilibrium mechanism is not adequate to describe the folding of Apaf-1 CARD at either pH, suggesting the presence of intermediates in equilibrium unfolding. Interestingly, the results showed that the secondary structure is less stable than the tertiary structure, based on the transition mid-points for unfolding. Single mixing and sequential mixing stopped-flow studies showed that Apaf-1 CARD folds and unfolds rapidly and suggest a folding mechanism that contains parallel channels with two unfolded conformations folding to the native conformation. Kinetic simulations show that a slow folding phase is described by a third conformation in the unfolded ensemble that interconverts with one or both unfolded species. Overall, the native ensemble is formed rapidly upon refolding. This is in contrast to other CARDs in which folding appears to be dominated by formation of kinetic traps.

Coupled with over-expression in host organisms, fusion protein systems afford economical methods to obtain large quantities of target proteins in a fast and efficient manner. Some proteases used for these purposes cleave C-terminal to their recognition sequences and do not leave extra amino acids on the target. However, they are often inefficient and are frequently promiscuous, resulting in non-specific cleavages of the target protein. To address these issues, we created a fusion protein system that utilizes a highly efficient enzyme and leaves no residual amino acids on the target protein after removal of the affinity tag. We designed a glutathione S-transferase (GST)-fusion protein vector with a caspase-3 consensus cleavage sequence located between the Nterminal GST tag and a target protein. We show that the enzyme efficiently cleaves the fusion protein without leaving excess amino acids on the target protein. In addition, we used an engineered caspase-3 enzyme that is highly stable, has increased activity relative to the wild-type enzyme, and contains a poly-histidine tag that allows for efficient removal of the enzyme after cleavage of the fusion protein. Although we have developed this system using a GST tag, the system is amenable to any commercially available affinity tag. KeywordsCaspase-3; Fusion protein; Protein expression; ProteolysisIn the past thirty years, a number of protein purification systems have been developed using fusion proteins to aid in the efficient purification and recovery of recombinant proteins from crude cell extracts or culture media. These systems incorporate amino-or carboxy-terminal proteins or peptides referred to as "tags." Many of the tags can be removed subsequent to binding to, or while concurrently bound to, a high affinity matrix. Matrices for binding the tags have generally incorporated immobilized metal for binding poly-histidine sequences, immobilized glutathione for binding glutathione S-transferase (GST),1 poly-alanine affinity matrices, antigenic epitopes to bind monoclonal antibodies, biotinylated resins for binding avidin or strep-avidin, carbohydrate-binding proteins, or even complete enzymes immobilized in a matrix for substrate binding [1]. While tags provide many advantages for aiding in the rapid recovery, stabilization, and increased scale of protein expression, many tags interfere with protein structure and function and must be removed after purification of the fusion protein. Removal of the tags usually occurs by cleavage of the fusion protein by a specific protease, such as thrombin. It is sometimes problematic, however, to remove the entire carrier protein sequence without leaving residual amino acids on the target protein. For example, the proteases most frequently used to remove carrier proteins are thrombin, factor Xa, and enterokinase. Factor Xa and enterokinase cleave C-terminal to their recognition sequences, leaving no residual amino acids. Thrombin, however, cleaves the amino acid sequence LVPR/GS between R and G, which leaves two amino ...