Calpains are a family of complex multi-domain intracellular enzymes that share a calcium-dependent cysteine protease core. These are not degradative enzymes, but instead carry out limited cleavage of target proteins in response to calcium signalling. Selective cutting of cytoskeletal proteins to facilitate cell migration is one such function. The two most abundant and extensively studied members of this family in mammals, calpains 1 and 2, are heterodimers of an isoform-specific 80 kDa large subunit and a common 28 kDa small subunit. Structures of calpain-2, both Ca2+-free and bound to calpastatin in the activated Ca2+-bound state, have provided a wealth of information about the enzyme's structure–function relationships and activation. The main association between the subunits is the pairing of their C-terminal penta-EF-hand domains through extensive intimate hydrophobic contacts. A lesser contact is made between the N-terminal anchor helix of the large subunit and the penta-EF-hand domain of the small subunit. Up to ten Ca2+ ions are co-operatively bound during activation. The anchor helix is released and individual domains change their positions relative to each other to properly align the active site. Because calpains 1 and 2 require ~30 and ~350 μM Ca2+ ions for half-maximal activation respectively, it has long been argued that autoproteolysis, subunit dissociation, post-translational modifications or auxiliary proteins are needed to activate the enzymes in the cell, where Ca2+ levels are in the nanomolar range. In the absence of robust support for these mechanisms, it is possible that under normal conditions calpains are transiently activated by high Ca2+ concentrations in the microenvironment of a Ca2+ influx, and then return to an inactive state ready for reactivation.

INTRODUCTION

The present review will focus on insights about calpains and their functioning that have largely been derived from recombinant protein approaches, calpain crystal structures and homology models based on these structures. There are excellent reviews of calpain referenced in the present review. In particular, we recommend reading the comprehensive text by Goll et al. [1] to acquire the necessary background to the calpain field.

BRIEF HISTORY OF CALPAINS

First discovery

The first reports on calpain came from two different groups in the 1960s who noted the presence of a calcium-activated proteolytic activity in soluble extracts from rat brain [2] and skeletal muscle [3]. In the 1970s, the enzyme was purified to homogeneity from skeletal muscle [4,5] and was called CANP (calcium-activated neutral protease). There ‘neutral’ referred to its optimal pH for activity, which is compatible with a cytoplasmic location. However, even at this early stage in calpain research, concern was expressed that “since this enzyme requires an extremely high and unphysiological concentration of Ca ions its physiological role remains unclear” [6]. This thread of concern that the [Ca2+] needed for half-maximal activation of calpain is orders of magnitude higher than the resting [Ca2+] inside the cell has been woven through the literature for three decades without conclusive resolution.

Origin of the name ‘calpain’

The first use of ‘calpain’ (referring to a calcium-dependent papain-like enzyme) was by Murachi et al. [7]. Around this time it became apparent that mammalian calpain could be resolved into two main isoforms (subsequently called μ- and m-) that differed in their calcium requirement [8]. The designations μ- and m- are short forms of micromolar and millimolar Ca2+-requiring respectively. These two isoforms were shown to have distinct (although similar) large (80 kDa) subunits, but a common small (28 kDa) subunit [9]. The early cDNA sequencing of calpains revealed the fusion of a papain-like cysteine protease with a calcium-binding protein and was a major step forward in appreciating the organization and complexity of these large proteases [10].

Genome sequencing reveals extent of the calpain family

Tissue-specific calpains, for example the muscle-specific isoform [11], were subsequently discovered. But it was not until the human genome sequencing effort approached fruition that the full extent of the mammalian calpain family was revealed [12,13]. A total of 15 human calpain genes have been listed, CAPN1–15, one of which (CAPN4) codes for the small subunit (Table 1). However, the criterion for designation as a calpain is generally agreed to be possession of a cysteine protease core sequence [14], even if some of these (such as the CAPN6 gene product) are not enzymatically functional because they are missing one or more of the catalytic triad residues. By this measure, the small subunit is not a calpain and therefore there are 14 calpain genes in humans with complete protease cores. An additional gene, CAPN16 for calpain-16, has been designated, but its predicted product only includes the first half of the protease core [15].

Table 1
Calpain nomenclature [120]

T-S, tissue-specific; U, ubiquitous; X, no; Y, yes.

Calpain Gene products Aliases PEF Tissue distribution pattern Expression (highest) 
Calpain-1 CAPN1 and CAPNS1 μ-Calpain, calpain I, μCANP Placenta, oesophagus, trachea 
Calpain-2 CAPN2 and CAPNS1 m-Calpain, calpain II, mCANP Kidney, lung, trachea 
Calpain-3 CAPN3 dimer? p94, nCL-1 (Lp82, Lp85, Rt88) T-S Skeletal muscle (lens, retina) 
Calpain-5 CAPN5 nCL-3, htra-3, CAPN5 Brain, kidney, liver 
Calpain-6 CAPN6 CAPN6 T-S Placenta 
Calpain-7 CAPN7 CAPN7, palBH  
Calpain-8 CAPN8 CAPN8, nCL-2 T-S Stomach, digestive tract 
Calpain-9 CAPN9 plus CAPNS1 CAPN9, nCL-4 T-S Digestive tract, heart 
Calpain-10 CAPN10 CAPN10 Heart 
Calpain-11 CAPN11 plus ? CAPN11 T-S Testis 
Calpain-12 CAPN12 plus ? CAPN12 T-S Hair follicle 
Calpain-13 CAPN13 plus ? CAPN13 Lung, testis 
Calpain-14 CAPN14 plus ? CAPN14  
Calpain-15 CAPN15 CAPN15, SolH Brain 
Calpain small subunit 1 CAPNS1 CAPN4, CPNS1* Heart, pancreas, kidney 
Calpain Gene products Aliases PEF Tissue distribution pattern Expression (highest) 
Calpain-1 CAPN1 and CAPNS1 μ-Calpain, calpain I, μCANP Placenta, oesophagus, trachea 
Calpain-2 CAPN2 and CAPNS1 m-Calpain, calpain II, mCANP Kidney, lung, trachea 
Calpain-3 CAPN3 dimer? p94, nCL-1 (Lp82, Lp85, Rt88) T-S Skeletal muscle (lens, retina) 
Calpain-5 CAPN5 nCL-3, htra-3, CAPN5 Brain, kidney, liver 
Calpain-6 CAPN6 CAPN6 T-S Placenta 
Calpain-7 CAPN7 CAPN7, palBH  
Calpain-8 CAPN8 CAPN8, nCL-2 T-S Stomach, digestive tract 
Calpain-9 CAPN9 plus CAPNS1 CAPN9, nCL-4 T-S Digestive tract, heart 
Calpain-10 CAPN10 CAPN10 Heart 
Calpain-11 CAPN11 plus ? CAPN11 T-S Testis 
Calpain-12 CAPN12 plus ? CAPN12 T-S Hair follicle 
Calpain-13 CAPN13 plus ? CAPN13 Lung, testis 
Calpain-14 CAPN14 plus ? CAPN14  
Calpain-15 CAPN15 CAPN15, SolH Brain 
Calpain small subunit 1 CAPNS1 CAPN4, CPNS1* Heart, pancreas, kidney 

NOMENCLATURE

Isoform nomenclature

The two abundant calpain isoforms found in nearly all tissues and cell types of mammals have for a long time been referred to as μ- and m-calpain to reflect the difference in the [Ca2+] required for half-maximal activation [16]. Unfortunately, this designation is somewhat misleading, because the difference in activating [Ca2+] is only one order of magnitude, being 3–50 μM for μ-calpain and 400–800 μM for m-calpain isolated from bovine skeletal muscle [1]. The [Ca2+] required for half-maximal activation of recombinant rat m-calpain with an N-terminally truncated small subunit, referred to extensively in the present review, was 350 μM [17]. Complications could arise in identifying orthologues on the sole basis of their Ca2+ requirement, and a more rational sequence-based nomenclature has been needed for some time, now that all calpain homologues have been identified in the human genome. Following extensive online discussion by the calpain community, we suggest in the present review that μ- and m-calpain be referred to as calpain-1 and -2 respectively. Calpain-1 is a heterodimer of the CAPN1 gene product, CAPN1, and the CAPNS1 (formerly CAPN4) gene product, CAPNS1. Thus calpain-1 can be fully described as CAPN1+CAPNS1 (Table 1) or CAPN1/S1 for short. Calpain-2 is a heterodimer of the CAPN2 gene product, CAPN2, and CAPNS1, again CAPN2/S1 for short. Calpain-3 is the isoform found predominantly in skeletal muscle [11]. It seems to exist as a homodimer of CAPN3 (CAPN3/3) [18]. It is widely known as p94, and splice variants of this isoform exist, such as p82 in lens [19]. This dual nomenclature, where p94 and p82 refer to molecular mass×10−3, might still be useful in combination with calpain-3 for distinguishing these variants. Another example of a calpain oligomer is G-calpain, which is associated with the gastrointestinal tract [20]. This enzyme appears to be a dimer of CAPN8 and CAPN9 and can be referred to as calpain-8/9 [21]. However, it is not presently clear whether other proteins, such as the small subunit, are present in this complex.

There is a second small subunit homologue, S2, in the human genome [22]. Although S2 could potentially form heterodimers with PEF (penta-EF-hand)-containing CAPN gene products, these variants have not been identified in vivo. A list of human calpain isoforms and their tissue specificity is given in Table 1.

Interpretation of calpain domains on the basis of structure

The first crystal structures of calcium-free m-calpain from rat [23] and human [24] (Figure 1) gave definition to the domains and their boundaries (Figure 2), but has resulted in two contradictory schemes for domain nomenclature. On the basis of sequence comparisons, Domain I was originally conceived of as a region of unknown function N-terminal to the papain-like region (Domain II). It was presumed to include a putative propeptide region that might need to be cleaved for enzyme activation. On the basis of the crystal structures, the papain-like region or protease core is more extensive than originally thought and was subsequently shown to have elements that bind Ca2+ ions and are not present in papain [25]. Indeed, the only portion of Domain I that is not structurally part of the protease core is the 19-residue α-helical N-terminal anchor [23], which is shown in Figure 1 making a contact with the small subunit. To avoid renumbering downstream domains, Hosfield et al. [23] called the two halves of the protease core Domains I and II (Figure 2). In another scheme [24], the two halves of the core are known as domains IIa and IIb.

Crystal structure of human Ca2+-free calpain-2

Figure 1
Crystal structure of human Ca2+-free calpain-2

Domain structure of human calpain-2 (PDB code 1KFU) produced in baculovirus-infected insect cells with the GR domain present [24]. The N-terminus of the large (80 kDa) subunits begins with the anchor helix (red) and leads into the protease core domains PC1 (orange) and PC2 (yellow), followed by the C2L domain (green) and then into the C-terminal PEF domain (light blue). This PEF domain is paired with the C-terminal PEF domain of the small (28 kDa) subunit (dark blue). The N-terminal GR domain is unstructured and electron density is lost 11 residues in from the C-terminal end of this domain. This structure is shown in preference to recombinant rat calpain-2 produced in E. coli (PDB code 1DF0) [23] which lacks some detail in flexible loop regions of the PC2 and C2L domains that can be seen in the human enzyme.

Figure 1
Crystal structure of human Ca2+-free calpain-2

Domain structure of human calpain-2 (PDB code 1KFU) produced in baculovirus-infected insect cells with the GR domain present [24]. The N-terminus of the large (80 kDa) subunits begins with the anchor helix (red) and leads into the protease core domains PC1 (orange) and PC2 (yellow), followed by the C2L domain (green) and then into the C-terminal PEF domain (light blue). This PEF domain is paired with the C-terminal PEF domain of the small (28 kDa) subunit (dark blue). The N-terminal GR domain is unstructured and electron density is lost 11 residues in from the C-terminal end of this domain. This structure is shown in preference to recombinant rat calpain-2 produced in E. coli (PDB code 1DF0) [23] which lacks some detail in flexible loop regions of the PC2 and C2L domains that can be seen in the human enzyme.

Rationalization of calpain domain boundaries and nomenclature

Figure 2
Rationalization of calpain domain boundaries and nomenclature

(A) The numerical domain nomenclature (I–IV) is displayed on a representation of the first complete sequence of a calpain large subunit, which is the chicken orthologue of human CAPN11 [10]. Domain colours are the same as those illustrated in Figure 1. (B) A redrawing of the domain boundaries necessitated by the crystal structure of rat calpain-2 [23]. Domain I was clearly not a structural entity in its own right. The first 19 residues form a helix that anchors the N-terminal end of the large subunit to the PEF domain of the small subunit, while the C-terminal remainder packs against the protease core. To avoid renumbering the downstream domains that continued into the small subunit, the two sub-domains of the protease core were designated D-I and D-II. These two sub-domains do behave functionally and structurally as separate entities, as for example when one rotates around the other during activation (see Figure 4). The extended linker region, residues 514–531 between D–III and D–IV, is represented as a thinner tube in grey. Following the structure determination of human calpain-2 [24], Strobl et al. [24] renamed the two protease core domains IIa and IIb, but without including residues 16–93 in with IIa (C). They retained domain I for the anchor helix. To avoid having two conflicting numerical schemes for the domains, a new system has been developed that identifies each domain by an acronym of its structural description [15,32]. Thus the protease core domains become PC1 and PC2; the C2-like domain is C2L; and the penta-EF-hand domains of the large and small subunits are PEF(L) and PEF(S) respectively. The glycine-rich (GR) domain of the small subunit was previously Domain V. The domain boundaries are indicated by residue numbers flanking each domain. The domain colour scheme (red to blue) borrows from the structural biology practice of representing a polypeptide chain in the colours of the rainbow to trace the relationship of any one section to either terminus. However, it is the reverse of the PyMOL tradition of having blue at the N-terminus so that this colour coding can still be applied to calpains without conflicting with domain colours.

Figure 2
Rationalization of calpain domain boundaries and nomenclature

(A) The numerical domain nomenclature (I–IV) is displayed on a representation of the first complete sequence of a calpain large subunit, which is the chicken orthologue of human CAPN11 [10]. Domain colours are the same as those illustrated in Figure 1. (B) A redrawing of the domain boundaries necessitated by the crystal structure of rat calpain-2 [23]. Domain I was clearly not a structural entity in its own right. The first 19 residues form a helix that anchors the N-terminal end of the large subunit to the PEF domain of the small subunit, while the C-terminal remainder packs against the protease core. To avoid renumbering the downstream domains that continued into the small subunit, the two sub-domains of the protease core were designated D-I and D-II. These two sub-domains do behave functionally and structurally as separate entities, as for example when one rotates around the other during activation (see Figure 4). The extended linker region, residues 514–531 between D–III and D–IV, is represented as a thinner tube in grey. Following the structure determination of human calpain-2 [24], Strobl et al. [24] renamed the two protease core domains IIa and IIb, but without including residues 16–93 in with IIa (C). They retained domain I for the anchor helix. To avoid having two conflicting numerical schemes for the domains, a new system has been developed that identifies each domain by an acronym of its structural description [15,32]. Thus the protease core domains become PC1 and PC2; the C2-like domain is C2L; and the penta-EF-hand domains of the large and small subunits are PEF(L) and PEF(S) respectively. The glycine-rich (GR) domain of the small subunit was previously Domain V. The domain boundaries are indicated by residue numbers flanking each domain. The domain colour scheme (red to blue) borrows from the structural biology practice of representing a polypeptide chain in the colours of the rainbow to trace the relationship of any one section to either terminus. However, it is the reverse of the PyMOL tradition of having blue at the N-terminus so that this colour coding can still be applied to calpains without conflicting with domain colours.

A more detailed comparison of the structures of the protease core of calpain with papain shows that there is only partial structural overlap between them. The protease core of calpain-2 is approximately 140 residues larger than papain, and they each contain structural elements not present in the other. Although the two structures do contain regions of nearly identical topology, they are not perfectly superimposable. The structural similarity that is observed between these and other members of the clan CA [26] cysteine protease family results in a substrate-binding site in calpain that shares features, such as the hydrophobic S2 subsite, with the smaller cysteine proteases. This active site similarity makes it more difficult to design calpain-specific inhibitors.

There was unanimity about the proximal domains, with Domain III referring to a β-sandwich structure that at the level of tertiary structure resembles the C2L (C2-like) domains found in a number of enzymes, such as protein kinase C and a phospholipase that transiently bind to membranes [27,28]. The fact that most C2 domains contain Ca2+ and phospholipid-binding sites led to the suggestion that the C2L domains of calpain might also bind these ligands. The C2L domain in calpain-2 contains two clusters of acidic residues: Glu392, Glu393, Glu394, Asp395, Glu396, Glu481 and Asp508; and Asp351, Asp398, Glu397, Asp399, Glu402 and Glu504, which supported the idea of a Ca2+-binding site through co-ordination and charge neutralization. However, no Ca2+ was observed to be bound in this region in the crystal structures of Ca2+-bound calpain when in complex with calpastatin [29,30]. The only other C2L calpain domain for which there is a crystal structure is that of human calpain-7 (PDB code 2QFE), which does not have the same cluster of acidic residues. Nor does this domain appear to bind Ca2+. Although the C2L domain does contain a similar-looking β-sandwich tertiary structure to the known C2 domains, the topology of the C2L domain in calpains is completely different from that of the previously described C2 domains (Figure 3). The antiparallel β-sheet containing the N-terminal end of the C2L domain of calpain is made up of β-strands 1, 8, 3 and 6. The equivalent sheet from the C2 domain has strands 5, 2, 1 and 8 side by side. The other sheets have strands 2, 7, 4 and 5, as opposed to strands 4, 3, 6 and 7. Thus, given the absence of sequence homology to C2 domains and the differences in topology, there is no a priori reason why the C2L domain of calpains should bind Ca2+ and phospholipids.

Domain topology diagrams for the synaptotagmin C2 domain [121] (A) and the calpain C2L domain (B)

Figure 3
Domain topology diagrams for the synaptotagmin C2 domain [121] (A) and the calpain C2L domain (B)

Strands of the β-sandwich are numbered and coloured in order from blue (N-terminus) through cyan, green, yellow, orange to red (C-terminus). The cylinder between β-strands 1 and 2 in (B) represents a short segment of α-helix.

Figure 3
Domain topology diagrams for the synaptotagmin C2 domain [121] (A) and the calpain C2L domain (B)

Strands of the β-sandwich are numbered and coloured in order from blue (N-terminus) through cyan, green, yellow, orange to red (C-terminus). The cylinder between β-strands 1 and 2 in (B) represents a short segment of α-helix.

The C-terminal domains of the large and small subunits of calpain (Domains IV and VI respectively) are PEF domains [31]. These α-helical folds of approximately 170 amino acids contain four EF hands that are able to bind Ca2+ and one EF-hand that does not bind Ca2+, but instead is involved in the dimerization contacts between the two subunits. The two PEF domains have a very similar sequence and structure. They show approximately 45% sequence identity and an RMSD (root mean square deviation) of approximately 1.5 Å (1 Å=0.1 nm) for Cα atoms. By comparison, the superimposition of the Ca2+-bound rat small subunit PEF domain on that of the Ca2+-free human structure also gives an RMSD of 1.5 Å.

Domain V is the N-terminal region of the small subunit that is structurally undefined but is notably rich in glycine. In the human small subunit, the 95 amino acid residues that comprise this domain contain 40 glycine and five proline residues, including two contiguous stretches of 11 and 20 glycines. For the nine known mammalian sequences, this region shows pairwise sequence identities ranging from 93% to 100% and all contain similar contiguous stretches of glycine residues.

Improved domain nomenclature

Following online discussion by the calpain community, a proposal has emerged to give the calpain domains descriptive names abbreviated as acronyms [15,32]. This renaming has the advantage of eliminating two competing misaligned numerical systems, while also providing a more informative clue to the identity of the domain than a number. It will also allow for the subsequent recognition and designation of domains N-terminal to the protease core i.e. that precede Domain I, such as those in calpains 7 and 15 [33]. Thus the protease core domains are PC1 and PC2 rather than I and II or IIa and IIb (Figure 2). Domain III becomes C2L for C2-like domain, and the penta-EF-hand domains IV and VI become PEF domains of the large (L) and small (S) subunits, respectively. Finally, Domain V is referred to as the GR (glycine-rich) domain.

Is there such a thing as a typical calpain?

Another terminology that should be reviewed is the description of some calpains as being ‘conventional’ or ‘typical’ (compared with ‘atypical’), or ‘classical’ (compared with ‘non-classical’). The perspective for ‘conventional’, ‘typical’ or ‘classical’ is similar to the PEF-containing calpains 1 and 2, which were the first calpain types to be characterized. The presence of a PEF domain is a significant structural feature in that this domain is typically involved in dimerization, either heterodimerization, usually with the small subunit, CAPNS1, or homodimerization [34] as in the case of calpain-3 [35,36]. Although the PEF-containing calpains make up eight of the 14 human calpains, they are largely confined to animals (metazoans) [15]. The ‘atypical’ non-PEF-containing calpains are in fact far more widespread and numerous (i.e. typical), occurring in animals, plants, fungi and even some bacteria [37]. Under these circumstances, the term ‘classical’ is preferable to ‘conventional’ or ‘typical’, meaning simply similarity to the first calpains that were characterized.

ROLES OF CALPAINS 1 AND 2

Physiological roles

The abundant heterodimeric calpains 1 and 2 are complex, intracellular proteases that require high levels of Ca2+ for activity. This presumably provides a safeguard against the overactivation of these potent proteases within the cell. Indeed, the Ca2+ concentrations for half-maximal activity of these two enzymes are several orders of magnitude higher than the multi-nanomolar resting [Ca2+] in the cell. Calpains 1 and 2 function in calcium signalling by making specific limited cuts in proteins to effect a change in the function of their targets rather than serving as degradative enzymes. The target proteins and processes in which calpains are implicated are too extensive to list in the present review, but have been covered by Goll et al. [1]. Perhaps the best documented of their many roles is in remodelling of the cytoskeleton for processes such as cell movement [38,39]. Even with this role there is uncertainty about the division of labour between the two main isoforms. For cells to move, the trailing edge must be detached from the substrate. Here there is a role for calpain(s) cutting the focal adhesion complexes for an easier ‘get away’ [40,41]. At the leading edge of the cells, where membrane ruffles, pseudopods and other protrusions occur, there must be extensive reorganization of the cytoskeleton to achieve movement, and again the major calpains are involved [42].

Calpains in disease

There is a case to be made that calpain-1 and -2's high calcium requirement for activity ensures that activation is temporally and spatially limited, occurring only near the epicentre of a Ca2+ influx, which is then quickly dissipated by diffusion and Ca2+ removal by various pumps [43]. However, there are situations in which Ca2+ levels in the cell rise out of control and cause prolonged widespread overactivation of calpain with off-target proteolysis leading to necrotic cell death [44]. This contributes, for example, to reperfusion injury when a blockage in the circulation is opened after heart attack, stroke or blunt trauma [45]. When this previously anoxic tissue is re-supplied with blood it undergoes a number of stresses that include the loss of calcium homoeostasis. There are many other situations, as in neurodegenerative diseases, muscular dystrophies, retinopathy and cataract, where calcium homoeostasis can be compromised, causing calpains to contribute to cell and tissue injury [4648]. In various animal models, the administration of low-molecular-mass calpain inhibitors can bring about some amelioration of the damage and has prompted the search and development of more specific and efficacious calpain inhibitors [4951].

ACTIVATION BY CALCIUM

The active site is misaligned in the absence of Ca2+

The Ca2+-free structures of rat and human calpain-2 [23,24] revealed that the catalytic triad residues (Cys115, His272 and Asn296; calpain-1 numbering) in the protease core were not in a position to produce an active enzyme. Unlike the structures of other cysteine proteases, the cysteine and histidine side chains were much too far apart (>8.5 Å) for the histidine to deprotonate the cysteine and activate it for nucleophilic attack on a substrate. The functional distance should be ~3.5 Å.

The calpain core is activated by Ca2+, even in the absence of other domains

The original concept of calpain activation by Ca2+ that flowed from the protein sequence imagined the C-terminal PEF domains as calmodulin-like regulators of the enzyme's activity that would, on binding Ca2+, induce a conformational change in the dimer, resulting in a propeptide cleavage event to autoproteolytically activate the enzyme. The real situation is far removed from this scenario. Unlike calmodulin, the PEF domains do not undergo major conformational changes upon binding Ca2+ [29,30], and calpains 1 and 2 do not have cleavable pro sequences. Unexpectedly, the protease core has its own two calcium-binding sites that act co-operatively to convert the core into a functioning cysteine protease [25].

The existence of these Ca2+-binding sites was suggested when partial proteolysis of active-site-inactivated (C105S) calpain-2 in the presence of divalent metal ions revealed that the protease core was selectively stabilized by Ca2+, even though this made the whole enzyme far more susceptible to digestion by exogenous proteases. No such effect was seen with Mg2+ [52]. Subsequently, it was reported that the protease core of human calpain-2 had very weak Ca2+-dependent proteolytic activity [53]. The recombinant calpain-1 core can be stably produced in Escherichia coli and has 5–10% of the activity of the whole enzyme [25]. Its crystal structure shows two Ca2+ ions binding to non-EF-hand sites in the core, one in each of the domains PC1 and PC2. Each Ca2+-binding site is made up of two flexible loops that supply backbone carbonyl and side-chain carboxy groups for co-ordinating the Ca2+ ions. The two Ca2+-binding sites act co-operatively to cause a large conformational change in the core that reorientates the two domains. Reorientation enables the core to form a functional active site cleft by a 25° rigid-body rotation of the two protease core domains relative to each other [54]. As shown in a stereo diagram, this movement brings the catalytic triad into register for peptide bond cleavage (Figure 4). The key movement is the repositioning of Cys115 (calpain-1 numbering) close enough to His272 for deprotonation of the former to occur. At the same time, a key tryptophan residue (Trp298) moves from an exposed position in the cleft to tuck into a hydrophobic patch formed by the rearrangement of the calcium-binding loops. The shape of the cleft between the core domains is radically different before and after activation (Figure 5). This process of activation, which can be conveniently monitored by intrinsic tryptophan fluorescence or hydrolysis of calpain substrates, displays obvious co-operativity [25]. Despite the absence of the other four domains, two of which bind eight Ca2+ ions, the observed [Ca2+] for half-maximal activation of the core is similar to that for the whole enzyme. Indeed, the calpain-1 core becomes a functional protease (mini-calpain) with very similar properties to the whole enzyme, including substrate specificity, inhibitor sensitivity and calcium requirements, but with a greatly reduced turnover number.

Movement of the protease core domains during activation

Figure 4
Movement of the protease core domains during activation

(A) Human calpain-2 protease core structure in the apo form (PDB code 1KFU) [24] shown as a stereo diagram in ribbon format. The catalytic triad residues and Trp288 are displayed to show their side chains in green. (B) Stereo diagram of the rat calpain-2 protease core structure in the Ca2+-bound form (PDB code 1KXR) [25] aligned to have its PC1 domain in the same orientation as in (A). Calcium ions are shown as purple spheres. Note the movement of the PC2 domain on binding Ca2+ and the closing of the distance between the catalytic cysteine and histidine residues.

Figure 4
Movement of the protease core domains during activation

(A) Human calpain-2 protease core structure in the apo form (PDB code 1KFU) [24] shown as a stereo diagram in ribbon format. The catalytic triad residues and Trp288 are displayed to show their side chains in green. (B) Stereo diagram of the rat calpain-2 protease core structure in the Ca2+-bound form (PDB code 1KXR) [25] aligned to have its PC1 domain in the same orientation as in (A). Calcium ions are shown as purple spheres. Note the movement of the PC2 domain on binding Ca2+ and the closing of the distance between the catalytic cysteine and histidine residues.

Formation of the calpain-2 catalytic cleft

Figure 5
Formation of the calpain-2 catalytic cleft

The catalytic cleft between the PC1 (orange) and PC2 (yellow) domains of human calpain-2 is shown (A) in the Ca2+-free form [24], and that of rat calpain-2 in the Ca2+-bound form is shown in (B). In both Figures, a section of a calpastatin inhibitory domain (grey) is displayed in the cleft for reference [29]. A section of the C2L domain (green) can be seen in the bottom left-hand corner.

Figure 5
Formation of the calpain-2 catalytic cleft

The catalytic cleft between the PC1 (orange) and PC2 (yellow) domains of human calpain-2 is shown (A) in the Ca2+-free form [24], and that of rat calpain-2 in the Ca2+-bound form is shown in (B). In both Figures, a section of a calpastatin inhibitory domain (grey) is displayed in the cleft for reference [29]. A section of the C2L domain (green) can be seen in the bottom left-hand corner.

The role of other calcium-binding sites

What then are the roles of the eight Ca2+ ions that bind to the two PEF domains? Rather than being the primary regulators of calpain activity as originally thought [55], we suggest that the PEF domains provide an additional level of safeguard against casual activation of calpain. The Ca2+-activated calpain structure [29,30] shows that the protease core is stabilized by contacts with the other domains, most notably the C2L and small subunit PEF domains. The lack of support by these domains in the protease core most probably explains why the enzymatic activity of the core is only 5–10% of that of the whole enzyme. In the protease core, the PC1 and PC2 domains can rotate about a pivot at residue Gly209, where both the phi and psi angles change as a result of the movement. Owing to this flexibility and the absence of supporting domains, the Ca2+-activated form is likely to be in equilibrium with the apo form.

If the core is supported in its active conformation by the other domains, it can be argued that these domains constrain the core when Ca2+ is absent and prevent spurious activation. Indeed, it is unlikely that the two core domains could rotate into the active conformation without being released from contact with the other domains. One constraint that is freed is the anchor helix, which bridges the N-terminus of the large subunit to the small subunit. It has been suggested that the co-ordination of Ca2+ by EF-hand 2 in the small subunit immediately opposite the anchor helix causes a charge repulsion of the basic residues in the helix [24,56]. Consistent with this idea, the anchor helix is both displaced and unstructured in the Ca2+-bound calpastatin-inhibited calpain-2 structure [29,30]. This allows the core to nestle down on to, and make even more intimate contacts with, the PEF domain of the small subunit (Figure 6).

Calcium-dependent activation of calpain-2 produces a more condensed enzyme with new contacts between the small subunit and the protease core

Figure 6
Calcium-dependent activation of calpain-2 produces a more condensed enzyme with new contacts between the small subunit and the protease core

In the space-filling view of the calpain-2 crystal structure in the apo-form (A), the anchor helix (red) makes contacts between the small subunit PEF (dark blue) and PC1 of the protease core (orange). The C2L domain (green) is clearly visible between these two domains. After activation, the anchor helix has been displaced and is disordered. PEF(S) has made new contacts with PC1 of the protease core that help stabilize the active conformation, and PC2 has moved up and over towards PC1. These new associations obscure that portion of C2L seen previously, although a new region projects to the right of PEF(S). Overall, the activated enzyme is more condensed, consistent with its lower apparent molecular mass on size-exclusion chromatography.

Figure 6
Calcium-dependent activation of calpain-2 produces a more condensed enzyme with new contacts between the small subunit and the protease core

In the space-filling view of the calpain-2 crystal structure in the apo-form (A), the anchor helix (red) makes contacts between the small subunit PEF (dark blue) and PC1 of the protease core (orange). The C2L domain (green) is clearly visible between these two domains. After activation, the anchor helix has been displaced and is disordered. PEF(S) has made new contacts with PC1 of the protease core that help stabilize the active conformation, and PC2 has moved up and over towards PC1. These new associations obscure that portion of C2L seen previously, although a new region projects to the right of PEF(S). Overall, the activated enzyme is more condensed, consistent with its lower apparent molecular mass on size-exclusion chromatography.

On the other side of the core there is an extensive series of electrostatic contacts between domains PC2 and C2L and an extended linker region that connects C2L to PEF (L). It had been suggested that subtle conformational changes derived from Ca2+ binding to the PEF domains might be transduced through the linker to the domain PC2/C2L contacts [23]. Although there is some experimental evidence for this mechanism from site-directed mutagenesis [23,57], the full-length structures of Ca2+-bound calpastatin-inhibited calpain-2 lend little support to the idea [29,30]. Instead, Ca2+ binding within PEF(L) may induce changes in the C2L/PEF domain interface itself. For example, binding of Ca2+ to the third EF-hand causes Glu626 to break an electrostatic interaction with Lys629. The latter is then free to form a hydrogen bond with the backbone carbonyl of Ile446 of C2L. The shift in the relative positions of C2L and PEF(L) upon Ca2+ binding results in an increase in the contact area between the two domains of approximately 25%. This shift also causes the linker region (residues 511–526) to become more flexible, such that some residues were no longer visible in the electron density map. In other words, there is no sign of tension in the linker, consistent with a pull being exerted from the PEF domains.

How do in vivo calcium signals activate calpain?

In vitro studies clearly show that calpain-1 and -2 can be fully activated without the involvement of extraneous factors or post-translational modification, provided that sub-millimolar to millimolar Ca2+ levels are present. However, it is not clear whether (and how) these concentrations are met in vivo. Several suggestions have been put forward for mechanisms that might lower the Ca2+ requirement for calpain activation within the cell. These include: the involvement of activator proteins [58,59], the effectiveness of which have not yet been substantiated by other laboratories [60]; small subunit dissociation [6164]; and covalent modification of calpain, including phosphorylation [65] and autoproteolysis [16,6669]. In the latter case, there is at least a structural explanation for how this might work. One of the ‘hot-spots’ for autoproteolysis is the anchor helix [70]. If the anchor helix were selectively cleaved off, then this modification might lower the energy barrier to activation by removing a key restraint on the movement of the core domains PC1 and PC2 (Figure 7). The same effect could be achieved by small subunit dissociation. However, given the many arguments in the next section against autoproteolysis and small subunit dissociation as physiological activation mechanisms, more consideration should be given to the idea that calpain is simply activated in the immediate vicinity of the epicentre of Ca2+ influx. This would both localize and limit the amount of calpain that is activated in the cell. Rapid diffusion of the localized high [Ca2+] down a huge concentration gradient would return calpain to an inactive state and limit the time over which calpain can be active once the signal ceases.

Hypotheses for the activation of calpains 1 and 2

Figure 7
Hypotheses for the activation of calpains 1 and 2

(A) In this scenario, calpain becomes activated by autolytic cleavage of the anchor helix (red rectangle). Loss of the helix (dotted red line) releases constraints on the protease core, which then allow PC1 and PC2 to reorientate to form the active site (indicated by the black ‘V’) on binding Ca2+. Domain colours are described in the legends to Figures 1 and 2. In this Figure, cleavage of the GR domain (dotted red line) of the small subunit is collateral damage and not obviously part of the activation process. For a movie of scenario A, please see http://www.BiochemJ.org/bj/447/bj4470335add.htm. (B) In this scenario, release of constraints on the calpain protease core is due to dissociation of the small subunit from the large subunit, with the small subunit forming a homodimer. Dissociation may or may not be accompanied by autoproteolysis of the anchor helix and GR domain. (C) Fully reversible Ca2+-dependent activation of calpain where the binding of Ca2+ to PEF(S) releases the anchor helix intact. Once the calcium ions dissociate and diffuse away, the apo structure reforms with the anchor helix constraining the core in its inactive form. (D) Consequences of prolonged activation in vivo or calpain purification in vitro. Initial autoproteolysis produces contemporaneous cuts in the anchor helix and C2L domain [70]. Extensive autoproteolysis leads to almost complete attenuation of calpain activity, extensive digestion of the C2L domain (see Figure 8), but preservation of both the protease core and a heterodimer of the PEF(L) and (S) domains. For an animated sequence of this Figure, please see http://www.BiochemJ.org/bj/447/0335/bj4470335padd.htm.

Figure 7
Hypotheses for the activation of calpains 1 and 2

(A) In this scenario, calpain becomes activated by autolytic cleavage of the anchor helix (red rectangle). Loss of the helix (dotted red line) releases constraints on the protease core, which then allow PC1 and PC2 to reorientate to form the active site (indicated by the black ‘V’) on binding Ca2+. Domain colours are described in the legends to Figures 1 and 2. In this Figure, cleavage of the GR domain (dotted red line) of the small subunit is collateral damage and not obviously part of the activation process. For a movie of scenario A, please see http://www.BiochemJ.org/bj/447/bj4470335add.htm. (B) In this scenario, release of constraints on the calpain protease core is due to dissociation of the small subunit from the large subunit, with the small subunit forming a homodimer. Dissociation may or may not be accompanied by autoproteolysis of the anchor helix and GR domain. (C) Fully reversible Ca2+-dependent activation of calpain where the binding of Ca2+ to PEF(S) releases the anchor helix intact. Once the calcium ions dissociate and diffuse away, the apo structure reforms with the anchor helix constraining the core in its inactive form. (D) Consequences of prolonged activation in vivo or calpain purification in vitro. Initial autoproteolysis produces contemporaneous cuts in the anchor helix and C2L domain [70]. Extensive autoproteolysis leads to almost complete attenuation of calpain activity, extensive digestion of the C2L domain (see Figure 8), but preservation of both the protease core and a heterodimer of the PEF(L) and (S) domains. For an animated sequence of this Figure, please see http://www.BiochemJ.org/bj/447/0335/bj4470335padd.htm.

AUTOPROTEOLYSIS

Autoproteolysis of calpains 1 and 2 is an intermolecular reaction

Being a multi-domain heterodimer with exposed linkers and loops (Figure 1), calpain is a prime target for digestion by proteases [52,71], including calpain itself [72,73]. However, unlike calpain-3, where there is clear evidence for at least one intramolecular cut site, it is obvious from the crystal structures of apo- and Ca2+-bound calpain-2 that none of the large subunit sequences cut by calpain-2 are within range of its own active site cleft [23,24,29,30]. The structural similarities between calpains 1 and 2 are so great that this will probably be also true for calpain-1. Therefore, although there is evidence that N-terminal autoproteolysis of calpain-1 and -2 does lower the calcium requirement for enzyme activation [16,6669], this anchor peptide cleavage must be an intermolecular event because the cleavage point is far removed from the active site cleft. Thus at least one calpain would have to be activated before it could activate others.

Autoproteolysis of calpains 1 and 2 is largely an artefact of purification

Other arguments can be made against autoproteolysis being a physiological activation mechanism. In the inactive state of calpain, the anchor peptide cleavage sites would be protected from hydrolysis by being part of an α-helix. It is only after activation and release from its contact with the small subunit PEF domain that the anchor helix becomes disordered and hence susceptible to proteolysis [29,30]. Although the anchor helix is cleaved early on during autoproteolysis, so too are many sites within the C2L domain that result in inactivation or greatly reduced activity of the enzyme, making this a risky method for calpain activation (Figure 7) [70]. Indeed, it could be argued that autoproteolysis of calpain is an artificial situation that is promoted in vitro when calpains are concentrated through purification. In vivo, the enzyme would be surrounded by other proteins, including its substrates, and would not naturally be concentrated to the same extent in the cell.

If the anchor peptide is not cut during activation, then the calpain molecules would be able to return to the ground state for additional cycles of activation and inactivation in response to calcium signals, rather than be consumed by autoproteolysis. Indeed, the idea of activation by anchor peptide removal seems to be a hangover from the time that the N-terminal region was thought to be a propeptide blocking the active site cleft as in some of the cathepsins, which the X-ray crystal structures have shown is clearly not the case [23,24].

SMALL SUBUNIT DISSOCIATION

Does the small subunit dissociate under physiological conditions?

Another incompletely substantiated claim that is deeply entrenched in the calpain literature is that the small ‘regulatory’ subunit dissociates from the large ‘catalytic’ subunit during activation or even as a calpain activation mechanism [6163,74,75], although this finding is disputed by others [76,77]. In theory, subunit dissociation could be as equally effective as autoproteolysis in releasing the constraints on realignment of the protease core imposed by the anchor peptide. Some researchers have even reported that N-terminal autolysis of calpain promotes subunit dissociation [56,78]. However, the advantage of a subunit dissociation activation mechanism without autoproteolysis would be its potential reversibility.

One reason why this ‘subunit dissociation mechanism’ is hard to accept from the perspective of structural biology is that the interface between the PEF domains of the small and large subunits is extensive and hydrophobic (Figure 8) [23,24]. It is difficult to imagine this interaction being broken and reformed without the involvement of a chaperone or a denaturing condition. Crystal structures of the small subunit homodimer in the presence [79,80] and absence [79] of Ca2+ only show small differences in conformation or area of contact between the apo and Ca2+-bound forms. This is also true for the calpain-2 heterodimer when it shifts from the calcium-free to the calcium-bound states [29,30]. Therefore, it seems unlikely that Ca2+ binding could perturb this PEF domain interaction enough to disrupt the pairing up of their fifth EF-hands. Furthermore, in the structure of the calpastatin-bound Ca2+-activated calpain-2, the protease core is nestled down on to, and makes new specific contacts with, the small subunit. This Ca2+-dependent compression of the enzyme is seen even in the absence of calpastatin, where the presence of both subunits in a 1:1 stoichiometry has been documented by SDS/PAGE ([29]; R. A. Hanna and P. L. Davies, unpublished work). Thus the heterodimer structure is stable in Ca2+ in the absence of calpastatin and its dimeric state is not a consequence of tethering by the extended calpastatin inhibitor contacting both subunits.

Contact surfaces between the calpain-2 subunits in the presence and absence of Ca2+

Figure 8
Contact surfaces between the calpain-2 subunits in the presence and absence of Ca2+

Contact surfaces between the large (upper) and small (lower) subunits are colour-coded and compared in the absence (left) and presence (right) of Ca2+. Atoms that make van der Waals contacts with a domain in the other subunit are coloured. Those in the large subunit are coloured (using the colour scheme of previous Figures) by the domain they belong to and those in the small subunit by the colour of the large subunit domain they touch. Purple spheres are Ca2+ ions. The small subunits (lower panels) have been rotated through 180° relative to the large subunits (upper panels).

Figure 8
Contact surfaces between the calpain-2 subunits in the presence and absence of Ca2+

Contact surfaces between the large (upper) and small (lower) subunits are colour-coded and compared in the absence (left) and presence (right) of Ca2+. Atoms that make van der Waals contacts with a domain in the other subunit are coloured. Those in the large subunit are coloured (using the colour scheme of previous Figures) by the domain they belong to and those in the small subunit by the colour of the large subunit domain they touch. Purple spheres are Ca2+ ions. The small subunits (lower panels) have been rotated through 180° relative to the large subunits (upper panels).

Some of the in vitro studies that show the stability of calpain-2 have been done with two important structural modifications. One is the absence of the N-terminal GR domain from the small subunit. This domain has no assigned function as of yet and, therefore, a role in activation cannot be ruled out. It was deliberately omitted from these constructs because it is unstable and rapidly proteolysed during production of the recombinant enzyme in E. coli. Where efforts have been made to include this domain in human calpain 2 expressed in insect cells, the domain appears to be disordered in the crystal structure [24]. The other modification has been the knockout of enzymatic activity by converting the catalytic cysteine residue to serine or alanine [81,82]. This ensures that there is no autoproteolysis. When the native enzyme undergoes autoproteolysis or proteolysis by exogenous proteases, the C2L domain is rapidly digested [52]. This releases the PEF domains of the two subunits as a stable heterodimer that does not dissociate or undergo rearrangement into homodimers [70]. This product is difficult to distinguish from the small subunit homodimer and might have been mistaken for it in some of the reports of subunit dissociation.

Having found fault with autoproteolysis and subunit dissociation as activation mechanisms, and having doubted protein activators and post-translational modifications because these regulatory mechanisms have not been reproduced by independent laboratories, there is an obligation to explain how calpain can be activated in the cell when resting calcium levels are three orders of magnitude lower than those need to achieve half-maximal activation. Quite simply, as argued previously [70,83,84], the extremely localized calcium concentration near the inflow from channels and pumps should be high enough to transiently activate calpain where needed – for example at a focal adhesion complex. By the time this calcium influx has diffused and registered cell-wide as a small increase in calcium concentrations, calpain will already have returned to its inactive state. We maintain that the physiological activation of calpain occurs in an extremely localized fashion, such that at any one time only a tiny fraction of calpain is activated in extremely localized areas for very brief moments. Indeed, the argument has been made that the requirement for high calcium levels acts as a biological safety device that helps prevent calpain overactivation under physiological conditions [85].

In support of this direct activation mechanism, we argue that calpain is probably not activated long enough during physiological calcium signalling to need inhibition by calpastatin. How else can one explain the striking observation that a calpastatin-knockout mouse is largely symptomless [86]? Calpastatin might only be needed as a stop gap to prevent prolonged calpain activity under conditions of cell stress [87,88]. This silencing of calpain appears to break down in pathological conditions such as ischaemic injury, because calpastatin is eventually proteolysed by its enzyme target [1] or during apoptosis by caspases [89]. In the absence of robust evidence for specific calpain activation mechanisms, it seems reasonable to propose the localized activation of calpain at the epicentres of Ca2+ (in)fluxes.

Aggregation of Ca2+-activated calpain

Another complication of working with purified calpain is its tendency to aggregate, particularly when divalent ion concentrations are high and monovalent ion concentrations are low [64]. Calpain large-subunit aggregation was reported as a consequence of autoproteolysis and subunit dissociation [63]. However, aggregation also occurs when calpain autoproteolysis is blocked by converting the active-site cysteine residue to a serine [64]. In forming the calpain–calpastatin complex for crystallography it was important to mix calpastatin with calpain before the addition of calcium, otherwise the enzyme precipitated [29]. It should be stressed that the inhibitor has no affinity for calpain until the protease is activated by Ca2+. Once calpastatin has bound to the Ca2+-activated calpain, the enzyme seems to be protected from aggregation and precipitation, which we attribute to the coverage of hydrophobic patches on calpain by the inhibitor. Aggregation of calpain in the presence of Ca2+ poses a problem for crystallography of the whole enzyme. Although physiological NaCl concentrations can largely prevent aggregation, this additive places one limit on the conditions under which the enzyme can be crystallized.

USEFULNESS OF ‘MINI-CALPAINS’

Inhibitor-bound calpain protease core structures

Problems with autoproteolysis and aggregation have complicated work on whole calpains 1 and 2 and have retarded attempts to obtain crystal structures of the Ca2+-activated enzymes in the absence of calpastatin. The discovery that domains PC1 and PC2 were resistant to proteolysis in the presence of Ca2+ but not in its absence [52] led to the expression and analysis of the calpain protease core as a functional ‘mini-calpain’ [25]. In particular, the protease core of calpain-1 has proved to be a useful reagent, because it mimics the function of the whole enzyme, but is not susceptible to autoproteolysis. This is illustrated when comparing the time course of hydrolysis of a hexapeptide FRET (fluorescence resonance energy transfer) substrate PLFAER by the calpain-1 core compared with the whole enzyme (Figure 9). The two enzymes have similar initial reaction rates. However, the whole enzyme rapidly approaches a plateau of activity, whereas the core maintains its enzyme activity for longer. As a result of this stability, it has been possible to co-crystallize the rat [9093] and human [94] calpain-1 protease core with a series of covalent cysteine protease inhibitors of varying degrees of reversibility and calpain specificity (Table 2). These have included compounds with epoxide [90,92], α-ketoamide [91,93] and aldehyde [90] warheads. The same strategy was used to examine leupeptin in the active site of human calpain-9 [95].

Effect of autoproteolysis on calpain reaction rate

Figure 9
Effect of autoproteolysis on calpain reaction rate

Line A, rate of hydrolysis of 10 μM FRET substrate (EDANS)-EPLFAERK-(DABCYL) [98] by 10 nM calpain-1 (Calbiochem). Line B, rate of hydrolysis of the same FRET substrate (10 μM) by 330 nM recombinant calpain-1 protease core [25]. The assays were performed at 22°C in 4 mM CaCl2, 10 mM Hepes, pH 7.4, and 10 mM DTT (dithiothreitol).

Figure 9
Effect of autoproteolysis on calpain reaction rate

Line A, rate of hydrolysis of 10 μM FRET substrate (EDANS)-EPLFAERK-(DABCYL) [98] by 10 nM calpain-1 (Calbiochem). Line B, rate of hydrolysis of the same FRET substrate (10 μM) by 330 nM recombinant calpain-1 protease core [25]. The assays were performed at 22°C in 4 mM CaCl2, 10 mM Hepes, pH 7.4, and 10 mM DTT (dithiothreitol).

Table 2
List of calpain structures with inhibitors bound at the active site
PDB code and reference Calpain Inhibitor abbreviation Inhibitor full name 
1TL9 [90Rat calpain-1 protease core Leupeptin N-Acetyl-Leu-Leu-Arg aldehyde 
1TLO [90Rat calpain-1 protease core E64 N-[N-[1-Hydroxycarboxyethyl-carbonyl]leucylamino-butyl]-guanidine 
2G8E [91Rat calpain-1 protease core SNJ-1715 (2S)-4-Methyl-2-(3-phenylthioureido)-N-((3S)-tetrahydro-2-hydroxy-3-furanyl)pentanamide 
2G8J [91Rat calpain-1 protease core SNJ-1945 ((1S)-1-((((1S)-1-Benzyl-3-(cyclopropylamino)-2,3-dioxopropyl)amino)carbonyl)-3-methylbutyl)carbamic acid 5-methoxy-3-oxapentyl ester 
2NQG [92Rat calpain-1 protease core WR18(S,S) 5-Azanylidyne-N-[(2S)-4-ethoxy-2-hydroxy-4-oxobutanoyl]-l-norvalyl-l-arginyl-l-tryptophanamide 
2NQI [92Rat calpain-1 protease core WR13(R,R) N~2~-[(2S)-2-{[(2R)-4-Ethoxy-2-hydroxy-4-oxobutanoyl]amino}pent-4-enoyl]-l-arginyl-l-tryptophanamide 
2R9C [93Rat calpain-1 protease core ZLAK-3001 Benzyl (S)-1-((2S,3S)-1-(3-(6-amino-9h-purin-9-yl) propylamino)-2-hydroxy-1-oxopentan-3-ylamino)-4-methyl-1-oxopentan-2-ylcarbamate 
2R9F [93Rat calpain-1 protease core ZLAK-3002 Benzyl [(1S)-1-{[(1S,2S)-1-Ethyl-2-hydroxy-3-{[3-(4-methylpiperazin-1-yl)propyl]amino}-3-oxopropyl]carbamoyl}-3-methylbutyl]carbamate 
1ZCM [94Human calpain-1 protease core ZLLYCH2N-[(benzyloxy)carbonyl]leucyl-N~1~-[3-fluoro-1-(4-hydroxybenzyl)-2-oxoxpropyl]leucinamide 
2P0R Human calpain-9 protease core Leupeptin N-Acetyl-Leu-Leu-Arg aldehyde 
3BOW [29Rat calpain-2 with calpastatin domain 4 CAST4 Calpastatin domain 4 
3DF0 [30Rat calpain-2 with calpastatin domain 1 CAST1 Calpastatin domain 1 
PDB code and reference Calpain Inhibitor abbreviation Inhibitor full name 
1TL9 [90Rat calpain-1 protease core Leupeptin N-Acetyl-Leu-Leu-Arg aldehyde 
1TLO [90Rat calpain-1 protease core E64 N-[N-[1-Hydroxycarboxyethyl-carbonyl]leucylamino-butyl]-guanidine 
2G8E [91Rat calpain-1 protease core SNJ-1715 (2S)-4-Methyl-2-(3-phenylthioureido)-N-((3S)-tetrahydro-2-hydroxy-3-furanyl)pentanamide 
2G8J [91Rat calpain-1 protease core SNJ-1945 ((1S)-1-((((1S)-1-Benzyl-3-(cyclopropylamino)-2,3-dioxopropyl)amino)carbonyl)-3-methylbutyl)carbamic acid 5-methoxy-3-oxapentyl ester 
2NQG [92Rat calpain-1 protease core WR18(S,S) 5-Azanylidyne-N-[(2S)-4-ethoxy-2-hydroxy-4-oxobutanoyl]-l-norvalyl-l-arginyl-l-tryptophanamide 
2NQI [92Rat calpain-1 protease core WR13(R,R) N~2~-[(2S)-2-{[(2R)-4-Ethoxy-2-hydroxy-4-oxobutanoyl]amino}pent-4-enoyl]-l-arginyl-l-tryptophanamide 
2R9C [93Rat calpain-1 protease core ZLAK-3001 Benzyl (S)-1-((2S,3S)-1-(3-(6-amino-9h-purin-9-yl) propylamino)-2-hydroxy-1-oxopentan-3-ylamino)-4-methyl-1-oxopentan-2-ylcarbamate 
2R9F [93Rat calpain-1 protease core ZLAK-3002 Benzyl [(1S)-1-{[(1S,2S)-1-Ethyl-2-hydroxy-3-{[3-(4-methylpiperazin-1-yl)propyl]amino}-3-oxopropyl]carbamoyl}-3-methylbutyl]carbamate 
1ZCM [94Human calpain-1 protease core ZLLYCH2N-[(benzyloxy)carbonyl]leucyl-N~1~-[3-fluoro-1-(4-hydroxybenzyl)-2-oxoxpropyl]leucinamide 
2P0R Human calpain-9 protease core Leupeptin N-Acetyl-Leu-Leu-Arg aldehyde 
3BOW [29Rat calpain-2 with calpastatin domain 4 CAST4 Calpastatin domain 4 
3DF0 [30Rat calpain-2 with calpastatin domain 1 CAST1 Calpastatin domain 1 

The established inhibitors and their derivatives typically occupy the unprimed side of the active site cleft in the S1, S2 and S3 subsites, where it is difficult to distinguish calpains from other cysteine proteinases. Inhibitors that could extend into the S4 subsite area would be in a region of calpain where contact with the C2L domain could prove to be discriminatory, since this domain is not present in other cysteine proteases such as the cathepsins. An indication of this is given in the structure of Senju Pharmaceutical's inhibitor SNJ-1945 bound to the calpain-1 core (Figure 10) [91]. Figure 10 shows that the long ether chain in the P3 position is equally distributed between two sites in the co-crystal structure, one of which would slightly clash if the C2L domain were present. The nine structures of the protease core of rat calpain-1 bound to a variety of inhibitors display a great degree of similarity. Excluding the flexible gating loops (residues 64–84 and 253–262) [25], the RMSDs between all pairs of structures vary between 0.09 Å and 0.5 Å for Cα atoms. After aligning the structures, the RMSDs of the gating loops themselves range from 0.21 Å for the most similar structures to 3.6 Å for the most different.

The co-crystal structure of the protease core of rat calpain-2 with the inhibitor SNJ-1945 covalently bound in the active site cleft

Figure 10
The co-crystal structure of the protease core of rat calpain-2 with the inhibitor SNJ-1945 covalently bound in the active site cleft

The enzyme is shown in surface representation with C atoms in white, S atoms in yellow, charged O atoms in dark red, uncharged O atoms in light red, charged N atoms in dark blue, and uncharged N atoms in light blue. The SNJ-1945 inhibitor (((1S)-1-((((1S)-1-benzyl-3-(cyclopropylamino)-2,3-dioxopropyl)amino)carbonyl)-3-methylbutyl)carbamic acid 5-methoxy-3-oxapentyl ester) bound to Cys115 is shown in stick representation with C atoms in green. Its cyclopropyl group (CP) projects into the primed site of the cleft and forces the gating loops (G) into the open configuration [91]. Hydrogen bond connections made by the inhibitor in the cleft are shown as blue dashed lines. The diethylene glycol group of SNJ-1945 in the P3 position is shown occupying two possible conformations (i and ii), one of which (i) would clash with the C2L domain in whole calpain.

Figure 10
The co-crystal structure of the protease core of rat calpain-2 with the inhibitor SNJ-1945 covalently bound in the active site cleft

The enzyme is shown in surface representation with C atoms in white, S atoms in yellow, charged O atoms in dark red, uncharged O atoms in light red, charged N atoms in dark blue, and uncharged N atoms in light blue. The SNJ-1945 inhibitor (((1S)-1-((((1S)-1-benzyl-3-(cyclopropylamino)-2,3-dioxopropyl)amino)carbonyl)-3-methylbutyl)carbamic acid 5-methoxy-3-oxapentyl ester) bound to Cys115 is shown in stick representation with C atoms in green. Its cyclopropyl group (CP) projects into the primed site of the cleft and forces the gating loops (G) into the open configuration [91]. Hydrogen bond connections made by the inhibitor in the cleft are shown as blue dashed lines. The diethylene glycol group of SNJ-1945 in the P3 position is shown occupying two possible conformations (i and ii), one of which (i) would clash with the C2L domain in whole calpain.

The primed side of the active site cleft is more different in calpain than its clan members, although this side is not traditionally targeted by inhibitors. Several observations suggest that there is potential for the development of primed-side calpain inhibitors. When inhibitors have extended into this side of the cleft they seem to have easily pushed aside the gating loops that initially appeared to occlude this part of the substrate-binding site [93]. In the example of the inhibitor SNJ-1715 co-crystallized with the calpain-1 core, a molecule of 2-(N-morpholino)ethanesulfonic acid buffer was found in the primed side of the cleft [91]. The series of primed side-extended α-ketoamide inhibitors developed by the Powers laboratory included one with an adenyl group that made a stacking interaction with the side chain of Trp298 [93]. Although this tryptophan residue is conserved in other cysteine proteases, the regions around the residue are quite distinct to calpain. The adenyl group could be a useful base for extending functional groups outwards to contact these calpain-specific sites. This arrangement with compounds occupying both sides of the cleft presaged the fit of the inhibitory B-subdomain of calpastatin, where peptide sequences on either side of a central loop make specific intimate (but relatively independent) contacts with the primed and unprimed sides of the cleft [29,30].

One of the features of most of the calpain inhibitors described above that may be responsible for their poor specificity is their inherent flexibility. An approach that has been used to try to circumvent this problem is one in which the inhibitors contain a macrocyclic moiety [96]. The goals of introducing the macrocycle were to force the inhibitors into a predominantly β-strand conformation, and to rigidify the inhibitor, thus minimizing the entropic penalty of binding. A side effect of this strategy may be that a more rigid inhibitor will have a reduced ability to accommodate different active site structures, thus improving specificity. Various macrocyclic calpain inhibitors synthesized by Abell et al. [96] were tested for their relative inhibition of ovine calpains 1 and 2. The most selective inhibitor showed a 7-fold lower IC50 value for calpain-2 than for calpain-1.

Use of the calpain core in screening for optimal substrate sequences

Another important use of the calpain core has been to probe the sequence preference of the enzyme. As calpain cleaves a great variety of exposed sequences from many different proteins with little obvious pattern, other than a strong preference for leucine residues or other branched aliphatic amino acids in the P2 position, it has been assumed that the enzyme has a minimal cleavage sequence specificity. However, it is possible that these sequences are cut at very different rates. When the calpain-1 core was used to lightly digest a degenerate dodecapeptide library [97], an analysis of the early cut sites revealed a distinct sequence preference on the primed sides of the scissile bond [98]. When this preference was built into the second stage degenerate library, the scissile bond was readily cleaved with a characteristic preference for branched aliphatic amino acids in the P2 position. The use of the protease core by Cuerrier et al. [98] avoided any contamination by internal autoproteolysis fragments that might have confused the analysis. The effectiveness of the cleavage sequence was shown by its rapid hydrolysis when presented as a FRET substrate. This PLFAER consensus sequence was cleaved much more rapidly than a consensus sequence based on a literature survey of known calpain cleavage sequences and more rapidly again than that of the calpain-specific cleavage site in spectrin.

The non-core domains have ancillary roles in calpain activity

The calcium-bound structure of the free calpain-1 core [25] is virtually the same as that of the homologous calpain-2 core within the calpastatin-bound Ca2+-activated whole enzyme [29]. Indeed, the molecular replacement solution of the latter successfully began with the calpain-1 protease core as a search model. Since the core is functional in the absence of the other calpain domains and is activated by the same calcium concentrations as the whole enzyme, there is reason to question the function of these other domains. It is likely that the C2L domain and the PEF domains of the large and small subunit are present to provide additional levels of regulation and control over a potent intracellular protease. A particularly graphic example of this is the inhibition of activated calpain where the endogenous unstructured inhibitor, calpastatin, interacts with all five domains of the enzyme. For this, domains C2L and PEF(L) and (S) play various roles. They increase the specificity and strength of calpastatin binding, and help direct the inhibitory portion of the molecule into the active site cleft (see below).

Rather than pushing the calpain core towards an active form when the calcium signal is registered, the adjacent domains seem to be involved in restraining movement of the core domains when they are in the inactive, apo, form. This restraint takes the form of a circular arrangement of the five domains in which the small subunit PEF domain makes contact with both ends of the large subunit, through its N-terminal anchor helix and through the pairing of the C-terminal fifth EF-hands. As mentioned earlier, there is a logical mechanism by which the anchor helix can be released during calcium signalling. As the calcium ions bind to the small subunit, the Ca2+ that occupies EF-hand 2 is directly opposite the anchor peptide binding site and might help its release by charge repulsion of basic amino acids in the helix [56]. In support of this model, the elution position of the small subunit PEF dimer from an ion-exchange column shifts enormously in a comparison with the Ca2+-bound and Ca2+-free forms (Figure 11). The fully Ca2+-loaded PEF(S) dimer fails to bind to QAE-Sepharose at pH 7.6, but after being stripped of Ca2+ by the chelating agent EDTA, the homodimer binds tightly to the resin and is eluted by ~0.2 M NaCl. The two PEF domains can each bind four Ca2+ ions. The first few Ca2+ ions bind with similar affinity to the Ca2+ pair binding to the protease core, as judged by the smooth sigmoidal relationship between% activity and calcium ion concentration [17]. Thus the whole activation process seems to be a concerted, co-operative process involving ten Ca2+ ions rather than just the two needed to activate the core.

The influence of Ca2+ on the elution of PEF(S) dimer from an anion exchanger

Figure 11
The influence of Ca2+ on the elution of PEF(S) dimer from an anion exchanger

In the presence of millimolar Ca2+ at pH 7.6, a PEF(S) dimer fails to bind to a Mono-Q anion exchange column (line A). After stripping the eight Ca2+ ions from the dimer using EDTA, the apo form of the protein binds tightly to the column and is eluted (line B) by a salt gradient (dotted line) at a NaCl concentration of 0.2 M.

Figure 11
The influence of Ca2+ on the elution of PEF(S) dimer from an anion exchanger

In the presence of millimolar Ca2+ at pH 7.6, a PEF(S) dimer fails to bind to a Mono-Q anion exchange column (line A). After stripping the eight Ca2+ ions from the dimer using EDTA, the apo form of the protein binds tightly to the column and is eluted (line B) by a salt gradient (dotted line) at a NaCl concentration of 0.2 M.

Why the calpain-2 core is less active

When the protease core of calpain-2 was produced and tested, it had very little proteolytic activity [99]. This core's crystal structure showed that a key α-helix, which supports the core domains in calpain-1, had denatured and caused distortion to the PC1 domain and the active site cleft. Trp106, which is a key residue in maintaining the hydrophobic core of PC1, was displaced into the active site cleft, thereby partially obstructing substrate access to the catalytic cysteine residue. The basis for the structural collapse is the presence of a glycine residue in the α-helix where an alanine residue resides in calpain-1. The negative effects of this helix breaker were demonstrated by mutating the glycine residue to alanine in the calpain-2 core. The G203A mutant core gained activity, although only approximately 10% of that of the calpain-1 core. Sequence comparisons of calpain isoforms show that some have a glycine residue in this position and others have alanine. It is not clear what the functional significance of the sequence difference is, but it is likely that the activity of the glycine-containing isoforms would be silenced by autoproteolysis once the core was separated from the C2L domain, where most of the autoproteolytic sites occur.

The protease cores of other calpains

Given the size and diversity of the calpain family, remarkably few whole enzyme structures have been solved. This is despite the human calpains having been a target of the Structural Genomics Consortium (http://www.thesgc.org). The problem seems to lie in the production of recombinant protein. Rat calpain-2 has been produced in E. coli [82,100]; human calpain-2 was initially made in insect cells by using a baculovirus vector [101], but has since been produced in E. coli [102]. However, the GR domain has either been deliberately deleted or remains unseen in the crystal structure [24]. Although rat calpain-1 is poorly produced in E. coli, a chimaera of rat calpains 1 and 2 was made in sufficient quantities to solve its structure [103]. The strategy used to achieve a useful yield of chimaeric enzyme was to replace both ends of the calpain-1 large subunit that interacts with the common small subunit with the equivalent regions from calpain-2. In this way, 85% of the calpain-1 large subunit structure was solved, albeit with some uncertainty about whether or not formation of the chimaera had caused some distortions.

Since the protease core and PEF domains of the major calpains seem to be relatively stable proteins that are not prone to proteolysis, the problem with producing the whole enzyme may lie with the presence of the C2L domain, which is very susceptible to proteolysis. Therefore one route to some structural information about other calpain isoforms has been to produce individual domains or domain combinations, particularly the protease core.

The protease core of calpain-3 was prepared in E. coli and was stable when the active site cysteine residue was converted into a serine [104]. In addition to the protease cores of rat calpains 1 and 2, structures of protease cores have been solved for: human calpain-1 and -9 in the presence of Ca2+ [95]; the Ca2+-bound G203A mutant of human calpain-1 with the inhibitor ZLLYCH2F bound to the active site cysteine residue [94]; and for Ca2+-bound human calpains 8 and 9 with leupeptin bound to the active site cysteine residue [95]. The individual domain structures of all of these protease cores are very similar to each other, with pairwise RMSD values for C-α atoms ranging from 0.25 to 0.7 Å (omitting the flexible gating loops). Similar to rat calpain-2, the protease cores of human calpains 1 and 8 contain a glycine residue at the position corresponding to Gly203 of rat calpain-2 and also show a disturbance of the helix in that area. Alignment of the complete protease core structures shows that they are all very similar, with the exception of the structure of the human calpain-9 protease core. If the latter structure is omitted, the pairwise RMSD values range from 0.38 to 0.85 Å, again without the flexible gating loops being included. Curiously, although the conformation of the leupeptin-bound protease core of calpain-9 is very similar to the protease core of Ca2+-activated calpain-1 with or without leupeptin bound (RMSD for C-α atoms of approximately 0.5 Å), the structure of the Ca2+-bound protease core of calpain-9 in the absence of leupeptin shows that the PC2 domain is rotated 48° away from the active conformation and in the opposite direction from that demonstrated by the calcium-free structures (Figure 4). In other words, the active site in this structure is even farther from the active conformation than is the calcium-free calpain-2 structure. These observations reinforce the idea that the protease core can be quite flexible in the absence of the other domains of the full-length structure. This may help explain the relatively low activity demonstrated by the protease cores – if only a subset of the total population in solution is in the active conformation at any instant. Also, it fits with the observation that the crystallization of the protease core is assisted by the presence of inhibitors in the active site cleft that serve to make contacts between the PC1 and PC2 domains.

NO SOONER ACTIVATED THAN INHIBITED BY CALPASTATIN?

Domain structure of calpastatin

Calpastatin, the endogenous inhibitor of calpain, is the product of a single gene with no known homologues [105]. A number of isoforms are produced by alternative splicing and transcription initiation. The archetypical calpastatin is a protein of approximately 70 kDa. It is composed of an N-terminal L subunit of unknown function followed by four independent inhibitory domains, each of which contains three subdomains known as A, B and C. The intact protein is capable of simultaneously binding to and inhibiting four calpain molecules [106]. Calpastatin has been shown to be intrinsically unstructured, although NMR studies show that the A and C subdomains exhibit a propensity for being α-helical [107] and the central region of the B subdomain displays evidence for the formation of a β-turn [107,108]. The B subdomain is the only one that exhibits any inhibitory activity. A 27-residue fragment spanning this region, known as the B-27 peptide, has been shown to specifically inhibit calpain [109,110]. A structure of a 19-residue fragment of the C subdomain (peptide C) in complex with a small subunit PEF homodimer showed that the amphipathic α-helix of peptide C binds in a hydrophobic groove on PEF(S) that becomes more open when the domain binds Ca2+ [111]. By homology, the A subdomain was predicted to bind to a similar groove on PEF(L). These structures facilitated a model made by these authors in which the ends of calpastatin are tethered by binding to the PEF domains in such a way that the inhibitory B peptide could interact with the protease core.

Mechanism of action

The crystal structures of the first and fourth domains of calpastatin in complex with Ca2+-bound calpain-2 show exactly how calpastatin is able to inhibit calpain without itself being cut [29,30]. The A and C subdomains, as predicted and demonstrated respectively, form α-helical regions that interact with the PEF domains of the large and small subunits respectively. The extended B subdomain binds across the surface of the C2L domain and passes through the active site between the PC1 and PC2 domains. The B subdomain sequence forms tight interactions with the enzyme surface on either side of the catalytically active cysteine residue. That is, the inhibitory peptide region lies within primed and unprimed sides of the cleft, but in the vicinity of that cysteine residue, the backbone of calpastatin loops out away from the sulfhydryl group (Figure 12B). The loop contains a series of β-turns corresponding to those identified by NMR of the isolated inhibitor in solution [107,108]. It is these turns beginning with one at the conserved glycine residue in the equivalent of the P1 position that enable the inhibitor to bend out of the cleft away from the cysteine residue.

Binding and inhibition of calpain-2 by an inhibitory domain of calpastatin

Figure 12
Binding and inhibition of calpain-2 by an inhibitory domain of calpastatin

(A) The structure of Ca2+-activated rat calpain-2 [29] is shown in surface representation with CAST4 represented in ribbon format (purple). Ca2+ atoms are shown as mauve spheres. (B) Close-up view of the active site cleft region bounded by the dotted lines in (A) showing the calpastatin B peptide looping out around the active site cysteine. N and O atoms of calpastatin are shown in blue and red respectively. Key residues from the core and inhibitor are indicated in stick representation. Note the stacking between Pro620 (calpastatin) and Trp288 (calpain).

Figure 12
Binding and inhibition of calpain-2 by an inhibitory domain of calpastatin

(A) The structure of Ca2+-activated rat calpain-2 [29] is shown in surface representation with CAST4 represented in ribbon format (purple). Ca2+ atoms are shown as mauve spheres. (B) Close-up view of the active site cleft region bounded by the dotted lines in (A) showing the calpastatin B peptide looping out around the active site cysteine. N and O atoms of calpastatin are shown in blue and red respectively. Key residues from the core and inhibitor are indicated in stick representation. Note the stacking between Pro620 (calpastatin) and Trp288 (calpain).

OTHER CALPAIN FAMILY MEMBERS

Calpain-3

This calpain [11] has been intensively studied, partly because of its abundance in skeletal muscle and because inactivating mutations in its gene (CAPN3) cause the genetic disease LGMD2A (limb girdle muscular dystrophy type 2A) [112].

Three extra sequences that are not in CAPN2 increase the CAPN3 molecular mass from 80 kDa to 94 kDa (hence the nickname p94). NS is a long N-terminal sequence that takes the place of the N-terminal anchor helix in CAPN2. IS1 (insertion sequence 1) is a 48-residue segment within the PC2 domain. IS2 lies in the region connecting the C2L and PEF(L) domains. It appears to bind calpain-3 to titin of the sarcomere [113]. Despite calpain-3's abundance in muscle, it has not been possible to produce the native enzyme without autoproteolysis [114]. The same was true for attempts to produce recombinant enzyme in COS cells [18]. Kinbara et al. [18] reported that the active form of the enzyme was initially cut into two large fragments of 60 kDa and 58 kDa, with further processing to a 55 kDa fragment. The insertion sequences IS1 and IS2 were identified as sites of autoproteolysis. Consistent with this, a tissue-specific isoform of calpain-3 (Lp82) that lacks these insertion sequences has been characterized from rat lens; and this form of the enzyme is relatively resistant to autoproteolysis [19]. Studies performed with the protease core of calpain-3 suggest that IS1 is a propeptide and that the first proteolytic cut made by this calpain is an intramolecular autolytic cleavage near the N-terminal end of IS1 [104,115]. Once this cut is made, the calpain-3 core can hydrolyse exogenous substrates and it also makes a distal cut towards the C-terminal end of IS1. The residues that constitute the two non-EF-hand Ca2+-binding sites of the core are conserved in calpain-3 (Table 3) [25], and the rate of autolytic cleavage is accelerated by increasing the Ca2+ concentration. However, it has been reported that Na+ can also activate the enzyme [116].

Table 3
Conservation of calcium-binding residues in human calpains

The conservation of Ca2+-binding residues in human calpains (compared with rat calpain-2) was evaluated on the basis of a sequence alignment. Residues that contribute side-chain oxygen atom(s) are indicated with an asterisk, others contribute only a main-chain oxygen atom. Italic residues are conserved, bold residues contribute one side-chain oxygen atom (asparagine or glutamine), underlined residues are not conserved. Note: since the two aspartate residues (2nd and 4th positions) in the PC2 Ca2+-binding site only contribute one oxygen atom each, if a sequence had a glutamate, asparagine or glutamine instead of an aspartate, that residue was still considered conserved.

 PC1 domain PC2 domain 
Calpain     
Rat calpain-2 D E E D D 
Human calpain-1 D E E D D 
Human calpain-2 D E E D D 
Human calpain-3 D E E D D 
Human calpain-5 D E E D D 
Human calpain-6 Q E E E D 
Human calpain-7 S E R E N 
Human calpain-8 D E E D D 
Human calpain-9 D E E D D 
Human calpain-10 Q V C E E 
Human calpain-11 D E E D D 
Human calpain-12 D E E D D 
Human calpain-13 D E E D D 
Human calpain-14 D L E D D 
Human calpain-15 D Q S D E 
 PC1 domain PC2 domain 
Calpain     
Rat calpain-2 D E E D D 
Human calpain-1 D E E D D 
Human calpain-2 D E E D D 
Human calpain-3 D E E D D 
Human calpain-5 D E E D D 
Human calpain-6 Q E E E D 
Human calpain-7 S E R E N 
Human calpain-8 D E E D D 
Human calpain-9 D E E D D 
Human calpain-10 Q V C E E 
Human calpain-11 D E E D D 
Human calpain-12 D E E D D 
Human calpain-13 D E E D D 
Human calpain-14 D L E D D 
Human calpain-15 D Q S D E 

There is evidence that calpain-3 forms a homodimer during purification [18]. Given the natural tendency for PEF domains to dimerize [31], we anticipated that these domains might be the site of dimerization. Expression of the calpain-3 PEF domain in E. coli produced a very stable homodimer and a model was developed with the two active sites located at opposite ends of the molecule [36]. Once the structure of calpain-2 was solved, it was possible to homology model most of calpain-3 (all except NS, IS1 and IS2, for which there are still no homologues). One reason for doing the modelling was to see if any of the more obscure LGMD2A point mutations could be explained in terms of the three-dimensional structure [117]. After the calcium-activated structure became available [29,30], this modelling exercise was repeated [118]. Some of the LGMD2A mutations that could not be rationalized in the context of the apo structure made sense when the apo- and Ca2+-bound structures were compared. The two modelling studies showed that many of the LGMD2A point mutations are to residues that provide contacts between domains and are likely to be important for the folding of calpain-3 and its transitioning between the active and inactive forms of the enzyme. One of the best examples of this type became apparent when the structure of the Ca2+-bound calpain-1 core structure was solved [25]. The human enzyme is inactivated by an arginine to glycine mutation. In the Ca2+-bound structure of the rat calpain-1 core, Arg104 makes a double salt bridge to Glu333, which helps to stabilize the activated conformation.

Is calcium signalling involved in the activation of most calpains?

The discovery of the two calcium-binding sites in the calpain protease cores of calpains 1 and 2 is also significant in the sense that they provide an activation mechanism for the majority of calpains that lack PEF domains. Most of the calcium co-ordinating residues in PC1 and PC2 are conserved throughout the human calpains (Table 3). In some cases, an aspartate or glutamate residue is replaced with a side chain that can offer one co-ordination position (asparagine or glutamine). The presence of calcium-binding sites in the protease core domains is particularly important in linking their activation to calcium signalling, because the only other domain common to most calpains is the C2L domain. Although it was long speculated that this domain might bind Ca2+ [119], there is no evidence for this in Ca2+-activated CAST (calpastatin inhibitory domain)-bound rat calpain-2 [29,30]. Also, when the crystal structure of the C2L domain of human calpain-7 (PDB code 2QFE) was solved in the presence of Ca2+, there were no calcium ions bound to the protein.

FUTURE STUDIES

Many more structures to be done

Calpain structures and those of their constituent domains have been extremely useful in elucidating the enzyme's activation by calcium and inhibition by calpastatin. Structures of the calpain core bound to low molecular mass inhibitors and of whole calpain bound to calpastatin are paving the way for the structure-guided design of more potent inhibitors with increased specificity for calpains over other cysteine proteases. However, few new structures have emerged in recent years despite a considerable effort expended on human calpains by the Structural Genomics Consortium in Toronto. One of the difficulties lies in producing recombinant calpains. Until recently, the only isoform which has been reliably produced in high yield from bacteria is rat calpain-2 [82]. Even then it was not possible to make it with an intact GR domain. The human orthologue differs by only 7% and yet until recently it has been difficult to produce in E. coli [102]. Unlike its calpain-2 paralogue, rat calpain-1 was produced in E. coli with very poor yield, although as stated earlier the amount produced could be increased by making a chimaera of the two enzymes where the N- and C-terminal regions of CAPN1 were replaced by CAPN2 [103].

The absence of the GR domain from the present structures and most in vitro studies leaves considerable uncertainty about its role. Because of this omission, one cannot be completely confident about any aspect of calpain activation. Interestingly, the GR domain's presumed location in the structure of calpastatin-bound active calpain would place it in position to make additional contacts with that portion of calpastatin lying between the B and C sub-domains that is presently not visible in the electron density because of polypeptide flexibility (Figure 12)

CONCLUSION

Calpains are complex, multi-domain, calcium-dependent cysteine proteases. They are intracellular enzymes responsible for limited cleavage of target proteins during calcium signalling, and as such their activity is tightly regulated. The recently published structures of calcium-activated calpain-2 bound to different repeats of its natural inhibitor, calpastatin, has been an important milestone in the quest for knowledge about calpains. Together with the earlier solved apo-structure of this calpain isoform, we now have an appreciation for how Ca2+ switches the protease from the inactive to the active form, and how the latter is then inhibited by calpastatin. Along the way, the ability to produce and crystallize an active calpain protease core free from the complications of autoproteolysis has paved the way for structure-based design of calpain-specific low molecular mass inhibitors. These compounds are needed to better understand the cellular functions of calpain and as leads in drug development to combat the over-activation of calpain seen in many diseases.

Abbreviations

     
  • C2L

    C2-like

  •  
  • FRET

    fluorescence resonance energy transfer

  •  
  • GR domain

    glycine-rich domain

  •  
  • PC domain

    protease core domain

  •  
  • LGMD2A

    limb girdle muscular dystrophy type 2A

  •  
  • NS

    N-terminal sequence

  •  
  • IS

    insertion sequence

  •  
  • PEF

    penta-EF-hand

  •  
  • RMSD

    root mean square deviation

We thank past and present members of the Davies lab for their contributions to the structural biology of the calpain field. Kristin Low, Jordan Chou and Olivia Macleod supplied previously unpublished work and/or assisted with the production of Figures.

FUNDING

We thank the Canadian Institutes for Health Research for financial support of our research. P.L.D. holds a Canada Research Chair in Protein Engineering.

References

References
1
Goll
D. E.
Thompson
V. F.
Li
H.
Wei
W.
Cong
J.
The calpain system
Physiol. Rev.
2003
, vol. 
83
 (pg. 
731
-
801
)
2
Guroff
G.
A neutral, calcium-activated proteinase from the soluble fraction of rat brain
J. Biol. Chem.
1964
, vol. 
239
 (pg. 
149
-
155
)
3
Huston
R. B.
Krebs
E. G.
Activation of skeletal muscle phosphorylase kinase by Ca2+. II. Identification of the kinase activating factor as a proteolytic enzyme
Biochemistry
1968
, vol. 
7
 (pg. 
2116
-
2122
)
4
Dayton
W. R.
Goll
D. E.
Zeece
M. G.
Robson
R. M.
Reville
W. J.
A Ca2+-activated protease possibly involved in myofibrillar protein turnover. Purification from porcine muscle
Biochemistry
1976
, vol. 
15
 (pg. 
2150
-
2158
)
5
Dayton
W. R.
Reville
W. J.
Goll
D. E.
Stromer
M. H.
A Ca2+-activated protease possibly involved in myofibrillar protein turnover. Partial characterization of the purified enzyme
Biochemistry
1976
, vol. 
15
 (pg. 
2159
-
2167
)
6
Ishiura
S.
Sugita
H.
Nonaka
I.
Imahori
K.
Calcium-activated neutral protease. Its localization in the myofibril, especially at the Z-band
J. Biochem.
1980
, vol. 
87
 (pg. 
343
-
346
)
7
Murachi
T.
Tanaka
K.
Hatanaka
M.
Murakami
T.
Intracellular Ca2+-dependent protease (calpain) and its high-molecular-weight endogenous inhibitor (calpastatin)
Adv. Enzyme Regul.
1980
, vol. 
19
 (pg. 
407
-
424
)
8
Mellgren
S. I.
[Alzheimer-type dementia. Possible relation to hypofunction of the cholinergic central nervous system]
Tidsskr. Nor. Laegeforen.
1980
, vol. 
100
 (pg. 
1355
-
1356
)
9
Wheelock
M. J.
Evidence for two structurally different forms of skeletal muscle Ca2+-activated protease
J. Biol. Chem.
1982
, vol. 
257
 (pg. 
12471
-
12474
)
10
Ohno
S.
Emori
Y.
Imajoh
S.
Kawasaki
H.
Kisaragi
M.
Suzuki
K.
Evolutionary origin of a calcium-dependent protease by fusion of genes for a thiol protease and a calcium-binding protein?
Nature
1984
, vol. 
312
 (pg. 
566
-
570
)
11
Sorimachi
H.
Imajoh-Ohmi
S.
Emori
Y.
Kawasaki
H.
Ohno
S.
Minami
Y.
Suzuki
K.
Molecular cloning of a novel mammalian calcium-dependent protease distinct from both m- and mu-types. Specific expression of the mRNA in skeletal muscle
J. Biol. Chem.
1989
, vol. 
264
 (pg. 
20106
-
20111
)
12
Dear
N.
Matena
K.
Vingron
M.
Boehm
T.
A new subfamily of vertebrate calpains lacking a calmodulin-like domain: implications for calpain regulation and evolution
Genomics
1997
, vol. 
45
 (pg. 
175
-
184
)
13
Dear
T. N.
Boehm
T.
Identification and characterization of two novel calpain large subunit genes
Gene
2001
, vol. 
274
 (pg. 
245
-
252
)
14
Croall
D. E.
Ersfeld
K.
The calpains: modular designs and functional diversity
Genome Biol.
2007
, vol. 
8
 pg. 
218
 
15
Sorimachi
H.
Hata
S.
Ono
Y.
Impact of genetic insights into calpain biology
J. Biochem.
2011
, vol. 
150
 (pg. 
23
-
37
)
16
Cong
J.
Goll
D. E.
Peterson
A. M.
Kapprell
H. P.
The role of autolysis in activity of the Ca2+-dependent proteinases (mu-calpain and m-calpain)
J. Biol. Chem.
1989
, vol. 
264
 (pg. 
10096
-
10103
)
17
Graham-Siegenthaler
K.
Gauthier
S.
Davies
P. L.
Elce
J. S.
Active recombinant rat calpain II. Bacterially produced large and small subunits associate both in vivo and in vitro
J. Biol. Chem.
1994
, vol. 
269
 (pg. 
30457
-
30460
)
18
Kinbara
K.
Ishiura
S.
Tomioka
S.
Sorimachi
H.
Jeong
S. Y.
Amano
S.
Kawasaki
H.
Kolmerer
B.
Kimura
S.
Labeit
S.
Suzuki
K.
Purification of native p94, a muscle-specific calpain, and characterization of its autolysis
Biochem J.
1998
, vol. 
335
 (pg. 
589
-
596
)
19
Ma
H.
Fukiage
C.
Azuma
M.
Shearer
T. R.
Cloning and expression of mRNA for calpain Lp82 from rat lens: splice variant of p94
Invest. Ophthalmol. Vis. Sci.
1998
, vol. 
39
 (pg. 
454
-
461
)
20
Hata
S.
Doi
N.
Kitamura
F.
Sorimachi
H.
Stomach-specific calpain, nCL-2/calpain 8, is active without calpain regulatory subunit and oligomerizes through C2-like domains
J. Biol. Chem.
2007
, vol. 
282
 (pg. 
27847
-
27856
)
21
Hata
S.
Abe
M.
Suzuki
H.
Kitamura
F.
Toyama-Sorimachi
N.
Abe
K.
Sakimura
K.
Sorimachi
H.
Calpain 8/nCL-2 and calpain 9/nCL-4 constitute an active protease complex, G-calpain, involved in gastric mucosal defense
PLoS Genet.
2010
, vol. 
6
 pg. 
e1001040
 
22
Schad
E.
Farkas
A.
Jekely
G.
Tompa
P.
Friedrich
P.
A novel human small subunit of calpains
Biochem. J.
2002
, vol. 
362
 (pg. 
383
-
388
)
23
Hosfield
C. M.
Elce
J. S.
Davies
P. L.
Jia
Z.
Crystal structure of calpain reveals the structural basis for Ca2+-dependent protease activity and a novel mode of enzyme activation
EMBO J.
1999
, vol. 
18
 (pg. 
6880
-
6889
)
24
Strobl
S.
Fernandez-Catalan
C.
Braun
M.
Huber
R.
Masumoto
H.
Nakagawa
K.
Irie
A.
Sorimachi
H.
Bourenkow
G.
Bartunik
H.
, et al. 
The crystal structure of calcium-free human m-calpain suggests an electrostatic switch mechanism for activation by calcium
Proc. Natl. Acad. Sci. U.S.A.
2000
, vol. 
97
 (pg. 
588
-
592
)
25
Moldoveanu
T.
Hosfield
C. M.
Lim
D.
Elce
J. S.
Jia
Z.
Davies
P. L.
A Ca2+ switch aligns the active site of calpain
Cell
2002
, vol. 
108
 (pg. 
649
-
660
)
26
Rawlings
N. D.
Barrett
A. J.
Bateman
A.
MEROPS: the database of proteolytic enzymes, their substrates and inhibitors
Nucleic Acids Res.
2012
, vol. 
40
 (pg. 
D343
-
D350
)
27
Corbalan-Garcia
S.
Gomez-Fernandez
J. C.
The C2 domains of classical and novel PKCs as versatile decoders of membrane signals
Biofactors
2010
, vol. 
36
 (pg. 
1
-
7
)
28
Rizo
J.
Sudhof
T. C.
C2-domains, structure and function of a universal Ca2+-binding domain
J. Biol. Chem.
1998
, vol. 
273
 (pg. 
15879
-
15882
)
29
Hanna
R. A.
Campbell
R. L.
Davies
P. L.
Calcium-bound structure of calpain and its mechanism of inhibition by calpastatin
Nature
2008
, vol. 
456
 (pg. 
409
-
412
)
30
Moldoveanu
T.
Gehring
K.
Green
D. R.
Concerted multi-pronged attack by calpastatin to occlude the catalytic cleft of heterodimeric calpains
Nature
2008
, vol. 
456
 (pg. 
404
-
408
)
31
Kretsinger
R. H.
EF-hands embrace
Nat. Struct. Biol.
1997
, vol. 
4
 (pg. 
514
-
516
)
32
Ono
Y.
Sorimachi
H.
Calpains: an elaborate proteolytic system
Biochim. Biophys. Acta
2012
, vol. 
1824
 (pg. 
224
-
236
)
33
Sorimachi
H.
Suzuki
K.
The structure of calpain
J. Biochem.
2001
, vol. 
129
 (pg. 
653
-
664
)
34
Ravulapalli
R.
Campbell
R. L.
Gauthier
S. Y.
Dhe-Paganon
S.
Davies
P. L.
Distinguishing between calpain heterodimerization and homodimerization
FEBS J.
2009
, vol. 
276
 (pg. 
973
-
982
)
35
Kinbara
K.
Sorimachi
H.
Ishiura
S.
Suzuki
K.
Skeletal muscle-specific calpain, p49: structure and physiological function
Biochem. Pharmacol.
1998
, vol. 
56
 (pg. 
415
-
420
)
36
Ravulapalli
R.
Diaz
B. G.
Campbell
R. L.
Davies
P. L.
Homodimerization of calpain 3 penta-EF-hand domain
Biochem. J.
2005
, vol. 
388
 (pg. 
585
-
591
)
37
Sorimachi
H.
Hata
S.
Ono
Y.
Calpain chronicle – an enzyme family under multidisciplinary characterization
Proc. Jpn. Acad. Ser. B
2011
, vol. 
87
 (pg. 
287
-
327
)
38
Franco
S. J.
Huttenlocher
A.
Regulating cell migration: calpains make the cut
J. Cell Sci.
2005
, vol. 
118
 (pg. 
3829
-
3838
)
39
Lebart
M. C.
Benyamin
Y.
Calpain involvement in the remodeling of cytoskeletal anchorage complexes
FEBS J.
2006
, vol. 
273
 (pg. 
3415
-
3426
)
40
Chan
K. T.
Bennin
D. A.
Huttenlocher
A.
Regulation of adhesion dynamics by calpain-mediated proteolysis of focal adhesion kinase (FAK)
J. Biol. Chem.
2010
, vol. 
285
 (pg. 
11418
-
11426
)
41
Wells
A.
Huttenlocher
A.
Lauffenburger
D. A.
Calpain proteases in cell adhesion and motility
Int. Rev. Cytol.
2005
, vol. 
245
 (pg. 
1
-
16
)
42
Cortesio
C. L.
Perrin
B. J.
Bennin
D. A.
Huttenlocher
A.
Actin-binding protein-1 interacts with WASp-interacting protein to regulate growth factor-induced dorsal ruffle formation
Mol. Biol. Cell
2010
, vol. 
21
 (pg. 
186
-
197
)
43
Brini
M.
Carafoli
E.
Calcium pumps in health and disease
Physiol. Rev.
2009
, vol. 
89
 (pg. 
1341
-
1378
)
44
McCall
K.
Genetic control of necrosis - another type of programmed cell death
Curr. Opin. Cell Biol.
2010
, vol. 
22
 (pg. 
882
-
888
)
45
Inserte
J.
Barrabes
J. A.
Hernando
V.
Garcia-Dorado
D.
Orphan targets for reperfusion injury
Cardiovasc. Res.
2009
, vol. 
83
 (pg. 
169
-
178
)
46
Bertipaglia
I.
Carafoli
E.
Calpains and human disease
Subcell. Biochem.
2007
, vol. 
45
 (pg. 
29
-
53
)
47
Biswas
S.
Harris
F.
Dennison
S.
Singh
J. P.
Phoenix
D.
Calpains: enzymes of vision?
Med. Sci. Monit.
2005
, vol. 
11
 (pg. 
RA301
-
RA310
)
48
Paquet-Durand
F.
Johnson
L.
Ekstrom
P.
Calpain activity in retinal degeneration
J. Neurosci. Res.
2007
, vol. 
85
 (pg. 
693
-
702
)
49
Pietsch
M.
Chua
K. C.
Abell
A. D.
Calpains: attractive targets for the development of synthetic inhibitors
Curr. Top. Med. Chem.
2010
, vol. 
10
 (pg. 
270
-
293
)
50
Carragher
N. O.
Calpain inhibition: a therapeutic strategy targeting multiple disease states
Curr. Pharm. Des.
2006
, vol. 
12
 (pg. 
615
-
638
)
51
Donkor
I. O.
Calpain inhibitors: a survey of compounds reported in the patent and scientific literature
Expert Opin. Ther. Pat.
2011
, vol. 
21
 (pg. 
601
-
636
)
52
Moldoveanu
T.
Hosfield
C. M.
Jia
Z.
Elce
J. S.
Davies
P. L.
Ca2+-induced structural changes in rat m-calpain revealed by partial proteolysis
Biochim. Biophys. Acta
2001
, vol. 
1545
 (pg. 
245
-
254
)
53
Hata
S.
Sorimachi
H.
Nakagawa
K.
Maeda
T.
Abe
K.
Suzuki
K.
Domain II of m-calpain is a Ca2+-dependent cysteine protease
FEBS Lett.
2001
, vol. 
501
 (pg. 
111
-
114
)
54
Moldoveanu
T.
Jia
Z.
Davies
P. L.
Calpain activation by cooperative Ca2+ binding at two non-EF-hand sites
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
6106
-
6114
)
55
Suzuki
K.
Emori
Y.
Ohno
S.
Imahori
S.
Kawasaki
H.
Miyake
S.
Structure and function of the small (30K) subunit of calcium-activated neutral protease (CANP)
Biomed. Biochim. Acta
1986
, vol. 
45
 (pg. 
1487
-
1491
)
56
Nakagawa
K.
Masumoto
H.
Sorimachi
H.
Suzuki
K.
Dissociation of m-calpain subunits occurs after autolysis of the N-terminus of the catalytic subunit, and is not required for activation
J. Biochem.
2001
, vol. 
130
 (pg. 
605
-
611
)
57
Alexa
A.
Bozoky
Z.
Farkas
A.
Tompa
P.
Friedrich
P.
Contribution of distinct structural elements to activation of calpain by Ca2+ ions
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
20118
-
20126
)
58
Melloni
E.
Averna
M.
Salamino
F.
Sparatore
B.
Minafra
R.
Pontremoli
S.
Acyl-CoA-binding protein is a potent m-calpain activator
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
82
-
86
)
59
Melloni
E.
Michetti
M.
Salamino
F.
Pontremoli
S.
Molecular and functional properties of a calpain activator protein specific for mu-isoforms
J. Biol. Chem.
1998
, vol. 
273
 (pg. 
12827
-
12831
)
60
Farkas
A.
Nardai
G.
Csermely
P.
Tompa
P.
Friedrich
P.
DUK114, the Drosophila orthologue of bovine brain calpain activator protein, is a molecular chaperone
Biochem. J.
2004
, vol. 
383
 (pg. 
165
-
170
)
61
Yoshizawa
T.
Sorimachi
H.
Tomioka
S.
Ishiura
S.
Suzuki
K.
Calpain dissociates into subunits in the presence of calcium ions
Biochem. Biophys. Res. Commun.
1995
, vol. 
208
 (pg. 
376
-
383
)
62
Anagli
J.
Vilei
E. M.
Molinari
M.
Calderara
S.
Carafoli
E.
Purification of active calpain by affinity chromatography on an immobilized peptide inhibitor
Eur. J. Biochem.
1996
, vol. 
241
 (pg. 
948
-
954
)
63
Li
H.
Thompson
V. F.
Goll
D. E.
Effects of autolysis on properties of mu- and m-calpain
Biochim. Biophys. Acta
2004
, vol. 
1691
 (pg. 
91
-
103
)
64
Pal
G. P.
Elce
J. S.
Jia
Z.
Dissociation and aggregation of calpain in the presence of calcium
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
47233
-
47238
)
65
Glading
A.
Bodnar
R. J.
Reynolds
I. J.
Shiraha
H.
Satish
L.
Potter
D. A.
Blair
H. C.
Wells
A.
Epidermal growth factor activates m-calpain (calpain II), at least in part, by extracellular signal-regulated kinase-mediated phosphorylation
Mol. Cell. Biol.
2004
, vol. 
24
 (pg. 
2499
-
2512
)
66
Baki
A.
Tompa
P.
Alexa
A.
Molnar
O.
Friedrich
P.
Autolysis parallels activation of mu-calpain
Biochem. J.
1996
, vol. 
318
 (pg. 
897
-
901
)
67
Brown
N.
Crawford
C.
Structural modifications associated with the change in Ca2+ sensitivity on activation of m-calpain
FEBS Lett.
1993
, vol. 
322
 (pg. 
65
-
68
)
68
Elce
J. S.
Hegadorn
C.
Arthur
J. S.
Autolysis, Ca2+ requirement, and heterodimer stability in m-calpain
J. Biol. Chem.
1997
, vol. 
272
 (pg. 
11268
-
11275
)
69
Suzuki
K.
Sorimachi
H.
Yoshizawa
T.
Kinbara
K.
Ishiura
S.
Calpain: novel family members, activation, and physiologic function
Biol. Chem. Hoppe Seyler
1995
, vol. 
376
 (pg. 
523
-
529
)
70
Chou
J. S.
Impens
F.
Gevaert
K.
Davies
P. L.
m-Calpain activation in vitro does not require autolysis or subunit dissociation
Biochim. Biophys. Acta
2011
, vol. 
1814
 (pg. 
864
-
872
)
71
Thompson
V. F.
Lawson
K. R.
Barlow
J.
Goll
D. E.
Digestion of mu- and m-calpain by trypsin and chymotrypsin
Biochim. Biophys. Acta
2003
, vol. 
1648
 (pg. 
140
-
153
)
72
Crawford
C.
Willis
A. C.
Gagnon
J.
The effects of autolysis on the structure of chicken calpain II
Biochem. J.
1987
, vol. 
248
 (pg. 
579
-
588
)
73
DeMartino
G. N.
Huff
C. A.
Croall
D. E.
Autoproteolysis of the small subunit of calcium-dependent protease II activates and regulates protease activity
J. Biol. Chem.
1986
, vol. 
261
 (pg. 
12047
-
12052
)
74
Yoshizawa
T.
Sorimachi
H.
Tomioka
S.
Ishiura
S.
Suzuki
K.
A catalytic subunit of calpain possesses full proteolytic activity
FEBS Lett.
1995
, vol. 
358
 (pg. 
101
-
103
)
75
Suzuki
K.
Sorimachi
H.
A novel aspect of calpain activation
FEBS Lett.
1998
, vol. 
433
 (pg. 
1
-
4
)
76
Zhang
W.
Mellgren
R. L.
Calpain subunits remain associated during catalysis
Biochem. Biophys. Res. Commun.
1996
, vol. 
227
 (pg. 
891
-
896
)
77
Dutt
P.
Arthur
J. S.
Croall
D. E.
Elce
J. S.
m-Calpain subunits remain associated in the presence of calcium
FEBS Lett.
1998
, vol. 
436
 (pg. 
367
-
371
)
78
Kitagaki
H.
Tomioka
S.
Yoshizawa
T.
Sorimachi
H.
Saido
T. C.
Ishiura
S.
Suzuki
K.
Autolysis of calpain large subunit inducing irreversible dissociation of stoichiometric heterodimer of calpain
Biosci. Biotechnol. Biochem.
2000
, vol. 
64
 (pg. 
689
-
695
)
79
Blanchard
H.
Grochulski
P.
Li
Y.
Arthur
J. S.
Davies
P. L.
Elce
J. S.
Cygler
M.
Structure of a calpain Ca2+-binding domain reveals a novel EF-hand and Ca2+-induced conformational changes
Nat. Struct. Biol.
1997
, vol. 
4
 (pg. 
532
-
538
)
80
Lin
G. D.
Chattopadhyay
D.
Maki
M.
Wang
K. K.
Carson
M.
Jin
L.
Yuen
P. W.
Takano
E.
Hatanaka
M.
DeLucas
L. J.
Narayana
S. V.
Crystal structure of calcium bound domain VI of calpain at 1.9 Å resolution and its role in enzyme assembly, regulation, and inhibitor binding
Nat. Struct. Biol.
1997
, vol. 
4
 (pg. 
539
-
547
)
81
Arthur
J. S.
Gauthier
S.
Elce
J. S.
Active site residues in m-calpain: identification by site-directed mutagenesis
FEBS Lett.
1995
, vol. 
368
 (pg. 
397
-
400
)
82
Elce
J. S.
Hegadorn
C.
Gauthier
S.
Vince
J. W.
Davies
P. L.
Recombinant calpain II: improved expression systems and production of a C105A active-site mutant for crystallography
Protein Eng.
1995
, vol. 
8
 (pg. 
843
-
848
)
83
Davies
P. L.
Campbell
R. L.
Moldoveanu
T.
Messerschmidt
A.
Calpain
The Handbook of Metalloproteins
2004
Chichester
John Wiley & Sons
(pg. 
489
-
500
)
84
Mellgren
R. L.
Huang
X.
Fetuin A stabilizes m-calpain and facilitates plasma membrane repair
J. Biol. Chem.
2007
, vol. 
282
 (pg. 
35868
-
35877
)
85
Friedrich
P.
The intriguing Ca2+ requirement of calpain activation
Biochem. Biophys. Res. Commun.
2004
, vol. 
323
 (pg. 
1131
-
1133
)
86
Takano
J.
Tomioka
M.
Tsubuki
S.
Higuchi
M.
Iwata
N.
Itohara
S.
Maki
M.
Saido
T. C.
Calpain mediates excitotoxic DNA fragmentation via mitochondrial pathways in adult brains: evidence from calpastatin mutant mice
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
16175
-
16184
)
87
Higuchi
M.
Tomioka
M.
Takano
J.
Shirotani
K.
Iwata
N.
Masumoto
H.
Maki
M.
Itohara
S.
Saido
T. C.
Distinct mechanistic roles of calpain and caspase activation in neurodegeneration as revealed in mice overexpressing their specific inhibitors
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
15229
-
15237
)
88
Rao
M. V.
Mohan
P. S.
Peterhoff
C. M.
Yang
D. S.
Schmidt
S. D.
Stavrides
P. H.
Campbell
J.
Chen
Y.
Jiang
Y.
Paskevich
P. A.
, et al. 
Marked calpastatin (CAST) depletion in Alzheimer's disease accelerates cytoskeleton disruption and neurodegeneration: neuroprotection by CAST overexpression
J. Neurosci.
2008
, vol. 
28
 (pg. 
12241
-
12254
)
89
Wang
K. K.
Posmantur
R.
Nadimpalli
R.
Nath
R.
Mohan
P.
Nixon
R. A.
Talanian
R. V.
Keegan
M.
Herzog
L.
Allen
H.
Caspase-mediated fragmentation of calpain inhibitor protein calpastatin during apoptosis
Arch. Biochem. Biophys.
1998
, vol. 
356
 (pg. 
187
-
196
)
90
Moldoveanu
T.
Campbell
R. L.
Cuerrier
D.
Davies
P. L.
Crystal structures of calpain-E64 and -leupeptin inhibitor complexes reveal mobile loops gating the active site
J. Mol. Biol.
2004
, vol. 
343
 (pg. 
1313
-
1326
)
91
Cuerrier
D.
Moldoveanu
T.
Inoue
J.
Davies
P. L.
Campbell
R. L.
Calpain inhibition by α-ketoamide and cyclic hemiacetal inhibitors revealed by X-ray crystallography
Biochemistry
2006
, vol. 
45
 (pg. 
7446
-
7452
)
92
Cuerrier
D.
Moldoveanu
T.
Campbell
R. L.
Kelly
J.
Yoruk
B.
Verhelst
S. H.
Greenbaum
D.
Bogyo
M.
Davies
P. L.
Development of calpain-specific inactivators by screening of positional scanning epoxide libraries
J. Biol. Chem.
2007
, vol. 
282
 (pg. 
9600
-
9611
)
93
Qian
J.
Cuerrier
D.
Davies
P. L.
Li
Z.
Powers
J. C.
Campbell
R. L.
Cocrystal structures of primed side-extending α-ketoamide inhibitors reveal novel calpain-inhibitor aromatic interactions
J. Med. Chem.
2008
, vol. 
51
 (pg. 
5264
-
5270
)
94
Li
Q.
Hanzlik
R. P.
Weaver
R. F.
Schonbrunn
E.
Molecular mode of action of a covalently inhibiting peptidomimetic on the human calpain protease core
Biochemistry
2006
, vol. 
45
 (pg. 
701
-
708
)
95
Davis
T. L.
Walker
J. R.
Finerty
P. J.
Jr
Mackenzie
F.
Newman
E. M.
Dhe-Paganon
S.
The crystal structures of human calpains 1 and 9 imply diverse mechanisms of action and auto-inhibition
J. Mol. Biol.
2007
, vol. 
366
 (pg. 
216
-
229
)
96
Abell
A. D.
Jones
M. A.
Coxon
J. M.
Morton
J. D.
Aitken
S. G.
McNabb
S. B.
Lee
H. Y.
Mehrtens
J. M.
Alexander
N. A.
Stuart
B. G.
, et al. 
Molecular modeling, synthesis, and biological evaluation of macrocyclic calpain inhibitors
Angew. Chem., Int. Ed. Engl.
2009
, vol. 
48
 (pg. 
1455
-
1458
)
97
Turk
B. E.
Huang
L. L.
Piro
E. T.
Cantley
L. C.
Determination of protease cleavage site motifs using mixture-based oriented peptide libraries
Nat. Biotechnol.
2001
, vol. 
19
 (pg. 
661
-
667
)
98
Cuerrier
D.
Moldoveanu
T.
Davies
P. L.
Determination of peptide substrate specificity for mu-calpain by a peptide library-based approach: the importance of primed side interactions
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
40632
-
40641
)
99
Moldoveanu
T.
Hosfield
C. M.
Lim
D.
Jia
Z.
Davies
P. L.
Calpain silencing by a reversible intrinsic mechanism
Nat. Struct. Biol.
2003
, vol. 
10
 (pg. 
371
-
378
)
100
Elce
J. S.
Expression of m-calpain in Escherichia coli
Methods Mol. Biol.
2000
, vol. 
144
 (pg. 
47
-
54
)
101
Masumoto
H.
Yoshizawa
T.
Sorimachi
H.
Nishino
T.
Ishiura
S.
Suzuki
K.
Overexpression, purification, and characterization of human m-calpain and its active site mutant, m-C105S-calpain, using a baculovirus expression system
J. Biochem.
1998
, vol. 
124
 (pg. 
957
-
961
)
102
Hata
S.
Ueno
M.
Kitamura
F.
Sorimachi
H.
Efficient expression and purification of recombinant human m-calpain using an Escherichia coli expression system at low temperature
J. Biochem.
2012
, vol. 
151
 (pg. 
417
-
422
)
103
Pal
G. P.
De Veyra
T.
Elce
J. S.
Jia
Z.
Crystal structure of a micro-like calpain reveals a partially activated conformation with low Ca2+ requirement
Structure
2003
, vol. 
11
 (pg. 
1521
-
1526
)
104
Diaz
B. G.
Moldoveanu
T.
Kuiper
M. J.
Campbell
R. L.
Davies
P. L.
Insertion sequence 1 of muscle-specific calpain, p94, acts as an internal propeptide
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
27656
-
27666
)
105
Wendt
A.
Thompson
V. F.
Goll
D. E.
Interaction of calpastatin with calpain: a review
Biol. Chem.
2004
, vol. 
385
 (pg. 
465
-
472
)
106
Hanna
R. A.
Garcia-Diaz
B. E.
Davies
P. L.
Calpastatin simultaneously binds four calpains with different kinetic constants
FEBS Lett.
2007
, vol. 
581
 (pg. 
2894
-
2898
)
107
Kiss
R.
Kovacs
D.
Tompa
P.
Perczel
A.
Local structural preferences of calpastatin, the intrinsically unstructured protein inhibitor of calpain
Biochemistry
2008
, vol. 
47
 (pg. 
6936
-
6945
)
108
Ishima
R.
Tamura
A.
Akasaka
K.
Hamaguchi
K.
Makino
K.
Murachi
T.
Hatanaka
M.
Maki
M.
Structure of the active 27-residue fragment of human calpastatin
FEBS Lett.
1991
, vol. 
294
 (pg. 
64
-
66
)
109
Betts
R.
Anagli
J.
The β- and γ-CH2 of B27-WT's Leu11 and Ile18 side chains play a direct role in calpain inhibition
Biochemistry
2004
, vol. 
43
 (pg. 
2596
-
2604
)
110
Betts
R.
Weinsheimer
S.
Blouse
G. E.
Anagli
J.
Structural determinants of the calpain inhibitory activity of calpastatin peptide B27-WT
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
7800
-
7809
)
111
Todd
B.
Moore
D.
Deivanayagam
C. C.
Lin
G. D.
Chattopadhyay
D.
Maki
M.
Wang
K. K.
Narayana
S. V.
A structural model for the inhibition of by calpastatin: crystal structures of the native domain VI of calpain and its complexes with calpastatin peptide and a small molecule inhibitor
J. Mol. Biol.
2003
, vol. 
328
 (pg. 
131
-
146
)
112
Richard
I.
Broux
O.
Allamand
V.
Fougerousse
F.
Chiannilkulchai
N.
Bourg
N.
Brenguier
L.
Devaud
C.
Pasturaud
P.
Roudaut
C.
, et al. 
Mutations in the proteolytic enzyme calpain 3 cause limb-girdle muscular dystrophy type 2A
Cell
1995
, vol. 
81
 (pg. 
27
-
40
)
113
Hayashi
C.
Ono
Y.
Doi
N.
Kitamura
F.
Tagami
M.
Mineki
R.
Arai
T.
Taguchi
H.
Yanagida
M.
Hirner
S.
, et al. 
Multiple molecular interactions implicate the connectin/titin N2A region as a modulating scaffold for p94/calpain 3 activity in skeletal muscle
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
14801
-
14814
)
114
Sorimachi
H.
Toyama-Sorimachi
N.
Saido
T. C.
Kawasaki
H.
Sugita
H.
Miyasaka
M.
Arahata
K.
Ishiura
S.
Suzuki
K.
Muscle-specific calpain, p94, is degraded by autolysis immediately after translation, resulting in disappearance from muscle
J. Biol. Chem.
1993
, vol. 
268
 (pg. 
10593
-
10605
)
115
Rey
M. A.
Davies
P. L.
The protease core of the muscle-specific calpain, p94, undergoes Ca2+-dependent intramolecular autolysis
FEBS Lett.
2002
, vol. 
532
 (pg. 
401
-
406
)
116
Ono
Y.
Ojima
K.
Torii
F.
Takaya
E.
Doi
N.
Nakagawa
K.
Hata
S.
Abe
K.
Sorimachi
H.
Skeletal muscle-specific calpain is an intracellular Na+-dependent protease
J. Biol. Chem.
2010
, vol. 
285
 (pg. 
22986
-
22998
)
117
Jia
Z.
Petrounevitch
V.
Wong
A.
Moldoveanu
T.
Davies
P. L.
Elce
J. S.
Beckmann
J. S.
Mutations in calpain 3 associated with limb girdle muscular dystrophy: analysis by molecular modeling and by mutation in m-calpain
Biophys. J.
2001
, vol. 
80
 (pg. 
2590
-
2596
)
118
Garnham
C. P.
Hanna
R. A.
Chou
J. S.
Low
K. E.
Gourlay
K.
Campbell
R. L.
Beckmann
J. S.
Davies
P. L.
Limb-girdle muscular dystrophy type 2A can result from accelerated autoproteolytic inactivation of calpain 3
Biochemistry
2009
, vol. 
48
 (pg. 
3457
-
3467
)
119
Tompa
P.
Emori
Y.
Sorimachi
H.
Suzuki
K.
Friedrich
P.
Domain III of calpain is a Ca2+-regulated phospholipid-binding domain
Biochem. Biophys. Res. Commun.
2001
, vol. 
280
 (pg. 
1333
-
1339
)
120
Ravulapalli
R.
Regulatory domains of the human calpain family
Ph.D. Thesis
2009
Canada
Department of Biochemistry, Queen's University, Kingston
121
Fuson
K. L.
Montes
M.
Robert
J. J.
Sutton
R. B.
Structure of human synaptotagmin 1 C2AB in the absence of Ca2+ reveals a novel domain association
Biochemistry
2007
, vol. 
46
 (pg. 
13041
-
13048
)

Author notes

1

This review is dedicated to the memory of Darrel Goll and Koichi Suzuki, two luminaries of the calpain field who lit the way for others to follow.

Supplementary data