m-Calpain activation in vitro does not require autolysis or subunit dissociation

Biochimica et Biophysica Acta 1814 (2011) 864–872

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / b b a p a p

m-Calpain activation in vitro does not require autolysis or subunit dissociation Jordan S. Chou a, Francis Impens b,c, Kris Gevaert b,c, Peter L. Davies a,⁎ a b c

Department of Biochemistry, Queen's University, Kingston, ON, Canada K7L 3N6 Department of Medical Protein Research, VIB, B-9000 Ghent, Belgium Department of Biochemistry, Ghent University, B-9000 Ghent, Belgium

a r t i c l e

i n f o

Article history: Received 6 December 2010 Received in revised form 25 March 2011 Accepted 12 April 2011 Available online 28 April 2011 Keywords: Calpain activation model Autolysis Subunit dissociation Mass spectrometry Calcium requirement Chromatography

a b s t r a c t Calpains are Ca2+-dependent, intracellular cysteine proteases involved in many physiological functions. How calpains are activated in the cell is unknown because the average intracellular concentration of Ca2+ is orders of magnitude lower than that needed for half-maximal activation of the enzyme in vitro. Two of the proposed mechanisms by which calpains can overcome this Ca2+ concentration differential are autoproteolysis (autolysis) and subunit dissociation, both of which could release constraints on the core by breaking the link between the anchor helix and the small subunit to allow the active site to form. By measuring the rate of autolysis at different sites in calpain, we show that while the anchor helix is one of the first targets to be cut, this occurs in the same time-frame as several potentially inactivating cleavages in Domain III. Thus autolytic activation would overlap with inactivation. We also show that the small subunit does not dissociate from the large subunit, but is proteolyzed to a 40–45 k heterodimer of Domains IV and VI. It is likely that this autolysis-generated heterodimer has previously been misidentified as the small subunit homodimer produced by subunit dissociation. We propose a model for m-calpain activation that does not involve either autolysis or subunit dissociation. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Calpains are a family of intracellular cysteine proteases that are found in both vertebrates and invertebrates. They catalyze the limited cleavage of protein substrates when activated by Ca2+, essentially converting local cellular Ca2+ signals into a wide range of cellular responses, including cell motility, cell cycle regulation, transcriptional regulation, signal transduction, and apoptosis (reviewed in Ref.[1]). The direct or indirect involvement of calpains in many diseases, including but not limited to traumatic brain injury, Alzheimer's disease, myocardial ischaemic injury, limb-girdle muscular dystrophy type 2A, gastric cancer, and type 2 diabetes mellitus [2–8], emphasizes the importance of calpains' roles and functions. μ- and m-Calpains, the two most studied calpain isoforms, are ubiquitously expressed and are heterodimers of an 80 k large subunit (Domains I–IV) and 28 k small subunit (Domains V–VI). Domains I and II, the “protease core”, contain the catalytic triad residues: the active site cysteine in Domain I and the histidine/asparagine pair in Domain II. Domain III, the C2-like domain, has been suggested to play a role in regulating calpain activity via electrostatic interactions with other domains [9], and interactions with cell membranes by phospholipid binding [10]. The last domain in the large subunit, Domain IV, has five EF-hands. The first four EF-hands contain calcium binding sites while

⁎ Corresponding author. Tel.: +1 613 533 2983; fax: +1 613 533 2497. E-mail address: [email protected] (P.L. Davies). 1570-9639/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.bbapap.2011.04.007

the fifth EF-hand, the one closest to the C terminus, is involved in heterodimerization of the large and small subunits of calpain [9,11–14]. Domain V, at the N terminus of the small subunit of calpain, is glycinerich and may have an unordered structure. Domain VI of the small subunit is homologous to Domain IV of the large subunit and forms a tight interaction with Domain IV through the pairing of their penta-EFhands; thus Domains IV and VI are often termed the PEF-domains. The crystal structure of the apo-form of m-calpain lacking Domain V [9,12] revealed that the catalytic triad is misaligned prior to activation. In the absence of Ca2+, the anchor helix at the N terminus of Domain I of the large subunit interacts with Domain VI of the small subunit. Along with the penta-EF-hand heterodimerization between Domains IV and VI, the anchor helix effectively circularizes all six domains and helps to structurally restrain calpain in an inactive conformation. In the presence of sufficient Ca2+, relief of structural constraints on the protease core allows the rotation of these two domains relative to each other, resulting in realignment of the catalytic triad to form the catalytic cleft [15]. However, the resting intracellular Ca2+ concentration across the cytoplasm (~0.1 μM) is orders of magnitude lower than that required for half-maximal activity of either μ-calpain (~50 μM) or m-calpain (~500 μM) in vitro[1,16]. One frequently cited explanation for overcoming this Ca2+ barrier in the cell is autoproteolysis (autolysis) of the anchor helix (Fig. 1A), which is claimed to lower the Ca2+ concentration needed for activation [17–22]. Subunit dissociation, where the small subunit separates from the large subunit and forms small subunit homodimers (Fig. 1B), is another published and widely quoted mechanism that is thought to facilitate calpain activation [23–28]. In both cases, removal of either the anchor

J.S. Chou et al. / Biochimica et Biophysica Acta 1814 (2011) 864–872

helix or the small subunit could in principle relieve structural constraints on the protease core, allowing the realignment of the core domains required for activation. Furthermore, researchers have recently reported that the two mechanisms are not mutually exclusive, claiming that autolysis induces subunit dissociation [29,30]. However, other groups have found that the small subunit does not dissociate [31,32]. Mutation of the small subunit EF-hands affects the Ca2+ requirement of the whole enzyme, which suggests that it remains bound to the large subunit during activation [11]. Recently, the Ca2+-bound structure of active m-calpain [13,14] showed that the small subunit is present in the activated enzyme and that it makes new heterodimeric interactions after activation. Given the disagreement in the literature about the contribution of autolysis and subunit dissociation to calpain activation, we have set out to investigate these two processes. Careful separation and analysis of end-stage calpain autolysis products showed that the PEF-domains remained together as heterodimers without any sign of dissociation or reorganization into homodimers. Mass spectrometric analysis of the early autolysis products showed that the anchor helix was extensively cleaved during the first minute of Ca2+ addition but numerous sites in Domain III were also cut during this initial phase of autolysis. Since cleavage of Domain III in m-calpain can release inactive protease core [33], and cuts here occur within the same time-frame as anchor helix cleavage, autolysis is unlikely to have evolved as a mechanism for calpain activation.

865

set to ArgC/P with up to one missed cleavage allowed, fixed trideuteroacetate modification of lysine residues, variable trideutero-acetate modification of peptide N-termini, variable oxidation of methionine residues to methionine-sulfoxide, a precursor mass tolerance of 10 ppm and a fragment mass tolerance of 0.5 Da. Two additional searches were performed to allow identification of neo-N-terminal peptides resulting from calpain autolysis. In these, Mascot's semi-ArgC/P protease settings were used and further peptide-centric databases holding all possible neo-N-termini from rat proteins were generated using the DBToolkit software as described previously [37,38]. Following Mascot database searches, peptides trideutero-acetylated at their α-amine were used to report calpain autolysis sites. 2.3. Size-exclusion chromatography m-Calpain (0.37 mg) was autolyzed for more than 3 h in 0.5 mL of Buffer B (50 mM Tris–HCl pH 7.6, 10 mM β-mercaptoethanol, and 1 mM CaCl2) to generate fully autolyzed calpain products. The digest was then run into a Superdex 75 10/300 GL column (S-75) that was equilibrated in Buffer B and coupled to the AKTA FPLC system (GE Healthcare Life Sciences). The column flow rate was set at 1 mL/min and 1 mL fractions were collected. Fractions corresponding to the peak seen at 280 nm were pooled, and 500 μL of this aliquot was concentrated using Vivaspin 500 (GE Healthcare Life Sciences) at 15,000 ×g before analysis by SDS-PAGE.

2. Materials and methods 2.4. Phenyl sepharose chromatography 2.1. m-Calpain protein purification and autolysis time course Rat calpain 80 k large subunit and truncated 21 k small subunit were co-expressed and purified as previously described [34]. Purified mcalpain protein (70 μg) was diluted into 100 μL of Buffer A (50 mM Tris– HCl pH 7.6, and 10 mM β-mercaptoethanol) and autolysis was initiated by the addition of CaCl2 to a final concentration of 1 mM. Aliquots (7 μg) were removed from the reaction at various time points (1, 2, 5, 10, 15, 20, 40, and 180 min) and mixed with 10 μL of SDS loading buffer + 25 mM EDTA to stop the reaction. Time point samples were then analyzed by 10% (w/v) SDS-PAGE. Autolysis samples prepared for mass spectrometry analysis were digested in 50 mM HEPES pH 7.4, and 10 mM dithiothreitol without any obvious difference in the pattern of cleavage products. 2.2. m-Calpain autolysis site determination by mass spectrometry Time course samples were analyzed following modification of N-terminal peptides as previously described [35]. Briefly, aliquots containing ~1.2 nmol of calpain and its autodigestion products were dissolved in 500 μl 2 M guanidinium hydrochloride in 50 mM sodium phosphate pH 8.0. Trideuteroacetate-N-hydroxysuccinimide (AcD3NHS, final concentration of 12.5 mM) was added to each sample to modify α- and ε-amines during a reaction for 2 h at 30 °C. Hydroxylamine (final concentration of 75 mM) and glycine (final concentration of 25 mM) were added and samples were incubated for 20 min at room temperature to reverse potential O-trideutero-acetylation on Ser, Thr or Tyr residues and quench non-reacted NHS-ester, respectively. Samples were desalted on NAP-5 columns (GE Healthcare Life Sciences) and recollected in 1 mL of 20 mM ammonium bicarbonate pH 8.0. Samples were boiled for 5 min, put on ice for 5 min and digested overnight at 37 °C with 2 μg trypsin (Promega Sequencing Grade Modified Trypsin). Without any pre-enrichment for N-terminal peptides, 100 μL of each sample was acidified with 2 μL formic acid and used for LC-MS/MS analysis on a Thermo LTQ-Orbitrap XL mass spectrometer (2 μL of each sample was injected) which was operated as described before [36]. Recorded MS/MS spectra were searched using a locally installed version of Mascot algorithm in the rat subsection of the Swiss-Prot database. The following search parameters were used: the protease was

Fully autolyzed calpain (0.7 mg) was loaded onto a Phenyl FF (low sub) column 1 mL (GE Healthcare Life Sciences) in Buffer C (25 mM Tris– HCl pH 7.6, 600 mM NaCl, 10 mM β-mercaptoethanol, and 1 mM CaCl2) at a flow rate of 1 mL/min. The column was washed with 2 column volumes of Buffer C and then eluted with a gradient from 0 to 100% Buffer D (25 mM Tris–HCl pH 7.6, 400 mM NaCl, 10 mM β-mercaptoethanol, and 2 mM EDTA) over 20 column volumes. Fractions (1 mL) containing protein peaks as detected by absorbance 280 nm were concentrated in Vivaspin 500 tubes and analyzed by SDS-PAGE. 2.5. Anion-exchange chromatography Calpain (1.5 mg) was autolyzed in 2.0 mL of Buffer B before the fully autolyzed products were injected into a Mono Q 5/50 GL column (GE Healthcare Life Sciences). The column was washed with 10 column volumes of Buffer B before elution with a gradient from 0 to 100% of Buffer B + 500 mM NaCl at 1 mL/min. Fractions (1 mL) corresponding to protein peaks were collected, concentrated in Vivaspin 500 tubes, and analyzed by SDS-PAGE. Pure truncated small subunit (DVI) was diluted into Buffer B, loaded, and eluted off the Mono Q column in an identical manner. 2.6. Trypsin digestion and identification of calpain peptides by MALDITOF mass spectrometry After SDS-PAGE analysis, protein bands were carefully excised and destained in 50 mM ammonium bicarbonate in 50% acetonitrile. The samples were then reduced with DTT (dithiothreitol), alkylated with iodoacetamide, digested with 6 ng/μL trypsin (Promega Sequencing Grade Modified Trypsin) for 5 h at 37 °C, and recovered from the gel with three extractions of 1% formic acid in 2% acetonitrile. Extracted peptides were mixed into the matrix solvent (5 mg/mL of α-cyano-4-hydroxycinnamic acid in 70% acetonitrile, 0.1% trifluoroacetic acid, and 10 mM di-ammonium citrate) in a 1:1 ratio. A spot of this final preparation at 1 pmol/μL was air dried and MALDI-TOF mass spectra were acquired and analyzed on a Voyager DePro (Applied Biosystems Corporation) using Data Explorer Version 5.1 software

866

J.S. Chou et al. / Biochimica et Biophysica Acta 1814 (2011) 864–872

(Applied Biosystems Corporation). Spectra were acquired in positiveion reflector mode. Peptide peaks with a signal to noise ratio of greater than 2:1 were selected for SwissProt database searching. Search

parameters were set at trypsin with up to one missed cleavage, fixed carbamidomethyl modification of cysteine, variable oxidation of methionine, and mass accuracy of 50 ppm.

J.S. Chou et al. / Biochimica et Biophysica Acta 1814 (2011) 864–872

867

3. Results 3.1. Calpain is rapidly autolyzed upon Ca2+ addition In the absence of Ca2+, the inactivity of m-calpain was represented by two prominent protein bands on SDS-PAGE: one at 80 k corresponding to the large subunit (Domains I–IV) and the other at 21 k corresponding to the N-terminally truncated small subunit (Domain VI) (Fig. 2, lane 0). Within 1 min of Ca2+ addition, calpain autolysis products became visible. As autolysis progressed, the amount of 80 k large subunit present by visual estimation gradually decreased to about half after 5 min and to 5–10% after 20 min. At the same time, autolysis products at 40 k and several around 20 k gradually appeared and became more prominent during the digestion. After 40 min, the reaction was almost complete as indicated by the near disappearance of the large subunit and appearance of considerable amounts of 20-k and 40-k autolysis products. As can be seen from the digestion pattern, many sites on m-calpain were susceptible to autolysis; however, unlike calpain 1, autolysis of m-calpain's anchor helix could not be seen by SDS-PAGE as a shift in mobility of the 80 k band. Therefore, an alternative method was necessary to monitor its loss.

Fig. 2. Autolysis profile of rat m-calpain shown by SDS-PAGE. Numbers above each lane of the protein gel represent individual time points (min) of the digest after Ca2+ addition. Molecular weights for two of the protein standards (M) are shown on the lefthand side of the gel. Equal amounts of protein (7 μg) were applied in each lane. 0 = control (no Ca2+); A = autolyzed for 3 h.

3.2. Identification of autolysis sites — Domain III is a hotspot for autolysis 3.3. Using autolyzed calpain to explore subunit dissociation To determine which sites in m-calpain were susceptible to autolysis, neo-N-termini exposed at various time points by autolysis were masstagged at their α-amine by trideutero-acetylation. Subsequent trypsin digestion and LC-MS/MS analysis resulted in the identification of trideutero-acetylated peptides indicative for calpain autolysis sites (Fig. 3A). By counting the number of trideutero-acetylated peptides starting at every position in the protein sequence (spectral counts), it can be seen that three sites in the anchor helix (residues 4, 10 and 15), numerous sites in Domain III, and one site in Domain IV (residue 612) experienced autolysis within 1 min. Analysis of longer time points revealed that the anchor helix was autolyzed early on, since the rate of cleavage at residues 4 and 10 was already declining after 1 min. Whereas, at most sites in Domain III, the rate of cleavage was steady or increased slightly. At residue 388 there was a significant increase in cutting over time, with a redoubling of the spectral counts at each interval from 1 to 40 min. Some sites (250, 514 and 526), located outside of Domain III, only became susceptible to cleavage after 40 min as if they were exposed by earlier autolysis at other sites. Ion intensity measurement, based on the sum of peak areas of the extracted ion chromatograms of trideutero-acetylated peptides starting at the same position, was used as a different quantification method to evaluate m-calpain autolysis sites (Fig. 3B). Again, it could be seen that both the anchor helix and Domain III experienced autolysis within 1 min of Ca2+ addition. While autolysis in the anchor helix slowed down, sites in Domain III, especially at residue 388, continued to be rapidly digested with time, in line with the spectral counts analysis. Not many differences were observed with the latter method except for the autolysis site at residue 360 in Domain III. Even though this site was not identified by the spectral counts analysis, it gave a strong ion intensity signal (see Supplementary data for complete tables of spectral counts and ion intensities). A graphical comparison of the autolysis sites determined by the two different quantification methods shows that most of the sites cluster in Domain III (Fig. 3C).

When calpain was autolyzed for more than 3 h there was no trace of the 80 k large subunit or indeed of any fragments larger than 40 k (Fig. 2, lane A). Completely autolyzed calpain consisted of two main products: the protease core (Domains I and II) at around 40 k, and several proteins similar in size to the small subunit at around 20 k. To explore the possibility that subunit dissociation was facilitated by autolysis, several chromatographic methods were explored to cleanly fractionate the different autolysis products for identification. 3.4. The 20 k calpain autolysis products behave as dimers m-Calpain autolysis products were analyzed by size-exclusion chromatography to see if any of the 20 k products were monomers. A Superdex-75 column was chosen for this purpose because it has an optimal separation range between 3 and 70 k. Completely autolyzed calpain was loaded onto the S-75 column in the presence of Ca2+ and eluted as one broad peak at approximately 45 k (Fig. 4A). No peaks were detected around 80 k or 20 k. SDS-PAGE analysis of this peak showed that no autolysis peptides were missing (Fig. 4B). Although unsuccessful in resolving the autolyzed calpain, size-exclusion chromatography demonstrated that all the autolysis products were extremely close in size, as indicated by the breadth of the one elution peak. In addition, our study confirmed previous observations that the globular 20 k peptides were dimers, not monomers that were close in size to the putative protease core at 40 k. 3.5. Phenyl sepharose chromatography separates the protease core from the dimers The previously solved active m-calpain structure with calpastatin bound [13,14] revealed the formation of hydrophobic clefts in the PEF Domains (IV and VI) that bind calpastatin when m-calpain becomes active in the presence of Ca2+. Taking advantage of these exposed

Fig. 1. Previously proposed models of calpain activation. Five of the six m-calpain domains are labeled as DI–DIV and DVI (Domain V was not seen because an N-terminally truncated small subunit was used for this study). Autolysis is represented by solid arrows, the 5th EF-hand pairing interaction between Domains IV and VI is represented as (=), and the dissociation of this interaction is indicated by dotted arrows. A) Activation by autolysis. Upon activation by Ca2+, the anchor helix is autolyzed, resulting in relief of structural constraints and formation of catalytic cleft (indicated by a shallow ‘V’). Autolysis is the cause of general activation. B) Subunit dissociation. Upon activation by Ca2+, the relief of structural constraints on the protease core (DI + II) originates from the small subunit (DVI) dissociating from the large subunit. A pair of small subunits then interacts with one another to form the small subunit homodimer. C) Our calpain activation model. Upon activation by Ca2+, the anchor helix is released allowing the catalytic cleft to be formed. The disordered anchor helix and DIII are extensively autolyzed resulting in two main products: the protease core (DI + DII) and the heterodimers (DIV + DVI). Autolysis is the result of general activation.

868

J.S. Chou et al. / Biochimica et Biophysica Acta 1814 (2011) 864–872

Fig. 3. Representations of autolysis sites in rat m-calpain. Residue numbers are according to the corresponding SwissProt entry: Q07009. A) Spectral counts represent the number of trideutero-acetylated peptides per autolysis site and show the relative abundance of the different sites in each sample. Only sites with spectral counts greater than 1 for at least one of the time points analyzed are shown. B) Ion intensity, another quantification method, is based on the sum of the peak areas of the extracted ion chromatograms of trideuteroacetylated peptides starting at the same position. Only sites with ion intensities above the 1 × 107 threshold for at least one of the time points are shown. C) Graphical domain representation of the sites of autolysis determined by the two quantification methods. Longer arrows represent sites that experience more autolysis. All four autolysis time points (1, 5, 10 and 40 min) indicated extensive autolysis activity in Domain III.

clefts, hydrophobic chromatography was used to fractionate the calpain autolysis products. Autolyzed calpain was loaded onto the Phenyl Sepharose column in the presence of Ca2+. The unbound material washed off with Ca2+-containing running buffer as a sharp peak, but the bound proteins could only be eluted with EDTA (Fig. 5A). The two resulting protein peaks showed a distinct separation of autolyzed calpain products: the sharp wash peak represented a homogeneous sample of the protease core, while the broad EDTA-

eluted peak corresponded to the 20-k protein dimers as seen by SDSPAGE (Fig. 5B). Even with extremely shallow gradients of 1 mM CaCl2 to 1 mM excess EDTA (represented by − 1 mM CaCl2) over the course of 20 column volumes (Fig. 5A), we were unable to sub-fractionate the dimers. The very instant Ca2+ was chelated, the dimer fraction eluted as one peak. However, using this method, we were successful in separating the putative protease core from the putative PEFdomain dimers.

J.S. Chou et al. / Biochimica et Biophysica Acta 1814 (2011) 864–872

Fig. 4. Size-exclusion chromatography analysis of autolyzed rat m-calpain. A) Superdex 75 column elution profile of m-calpain that had undergone complete autolysis. The elution volumes of the protein markers BSA (67 k) and chymotrypsinogen A (25 k) are indicated by dotted lines. Fractions corresponding to the peak at ~ 45 k were pooled (S-75), concentrated, and analyzed by SDS-PAGE. B) SDS-PAGE analysis of the completely autolyzed sample before (A) and after (S-75) it was run on the column. M = protein markers; 2 = intact m-calpain.

869

Fig. 5. Hydrophobic interaction chromatography analysis of autolyzed rat m-calpain. A) Fully autolyzed m-calpain chromatographed on Phenyl Sepharose. After initial washing, the Ca2+ in the column was slowly chelated by the addition of EDTA in the gradient. Negative molar amounts of CaCl2 = excess EDTA. B) SDS-PAGE analysis of the completely autolyzed sample before (A) and after (W and E) fractionation. M = protein markers; W = Wash; E = Elute.

3.6. The PEF-domains in autolyzed calpain remain as heterodimers The PEF-domains IV and VI of m-calpain have a slight difference in pI. This slight difference should be sufficient to resolve homodimers of these domains from a heterodimer. m-Calpain autolysis products were loaded onto a Mono Q anion exchange column in the presence of Ca2+ and washed briefly before gradual elution with increasing [NaCl] (Fig. 6A). The four peaks (E1, E2a, E2b, and E3) were eluted at NaCl concentrations of 110 mM, 275 mM, 295 mM, and 350 mM respectively. These four fractions and the wash peak that did not stick to the column were collected for SDS-PAGE analysis. Other chromatography runs at slightly different pH values gave similar profiles (data not shown). A separate identical run using truncated small subunit (DVI) instead of autolyzed calpain showed that, unlike the autolysis products, the small subunit does not bind to the column. SDS-PAGE analysis of these samples showed that each peak was a mixture of 2–4 protein species (Fig. 6B). Peaks E1, E2a and E2b each contained 3–4 bands that were in the range of ~ 20 k. The three in peak E1 were on average smaller than those in peaks E2a and E2b, although there was some overlap in sizes. The three protein bands in the E2a peak and the four in E2b covered the same size range. Peak E3 contained the putative protease core that appeared to have been

cleanly separated from the PEF-domain dimers, but it did contain one other band running at approximately 20 k. It is important to note that no protein was detected in the unbound fraction (Wash) either by visual inspection of the SDS-PAGE or by Bradford protein assay. In addition, a sample of the small subunit (DVI) was included as reference on the gel. As can be seen, several protein bands in peaks E2a and E2b were larger in size compared to the 21 k small subunit, and therefore, were likely to have originated from the large subunit of calpain. Each protein band in the four peaks was carefully excised from the gel, digested with trypsin, and identified by MS-MALDI peptide mass fingerprinting (see Supplementary data for tables of peptides obtained). To be thorough, protein bands that ran close together on the SDS-PAGE “doublets” were excised and analyzed separately. Since our results showed that the doublets were always derived from the same domain, we hypothesize that the lower band was just a more autolyzed version of the higher one. From the peptide identification arrows in Fig. 6B, it can be seen that autolysis products corresponding to Domains IV and VI are always found together. The putative protease core in E3 did indeed contain both Domains I and II, but with no trace of Domain III. The minor band in E3 running at ~20 k proved to be Domain I. Overall, the peptides

870

J.S. Chou et al. / Biochimica et Biophysica Acta 1814 (2011) 864–872

Fig. 7. Sequence coverage of rat calpain 2. All of the autolysis peptides identified by MSMALDI peptide mass fingerprinting are underlined. Notice that there were no Domain III peptides identified. Lysines and arginines are bolded.

Fig. 6. Ion-exchange chromatography analysis of autolyzed rat m-calpain. A) Profile of autolyzed m-calpain chromatographed on a Mono Q column (solid line) and eluted by a NaCl gradient (dotted line). A second identical run was performed using pure small subunit (dashed line) B). SDS-PAGE analysis of the completely autolyzed sample before (A) and after (eluted peaks E1, E2a, E2b, E3) fractionation. Protein band identification by MS-MADLI peptide mass fingerprinting: triangular arrowheads = Domain IV; circular arrowheads = Domain VI; diamond arrow = Domains I and II (protease core); dotted arrow = Domain I. M = protein markers; 2 = intact calpain 2; DVI = small subunit; W = wash.

obtained from mass fingerprinting covered approximately 50% of the entire rat m-calpain sequence (Fig. 7). Not a single peptide identified by MS-MALDI originated from Domain III, presumably due to extensive autolysis in that region. If the absence of Domain III is factored in, then the mass fingerprinting covered approximately 70% of the remainder of the m-calpain sequence. 4. Discussion 4.1. m-calpain autolysis is the result of activation, not the cause Some cysteine proteases, including (but not limited to) cathepsins B, K, and papain, have a propeptide located within the enzyme's catalytic cleft that must be removed by intramolecular autolysis for the enzyme to be activated (reviewed in Ref.[39]). Over the years, several groups have argued that the major calpains (isoforms μ- and m-) also require autolysis for activation in the cell, because cleavage of the anchor helix lowered the Ca2+ requirement for activation [17–22,29]. This effect was mainly studied in μ-calpain [1,40,41] where the anchor peptide is long enough that its removal causes a shift in the mobility of the large subunit that can be observed on SDS-PAGE. Subsequently, it was shown that the

anchor peptide region of μ-calpain was longer than that of m-calpain because it also contains a mitochondrial import sequence [42]. Using results from both past literature and new experimental evidence reported in this paper, we argue that m-calpain large subunit autolysis has no part to play in activation. 1) The previously solved structures of the rat [9] and human [12] inactive m-calpain isoforms revealed that the anchor helix is situated well outside of the catalytic cleft. One of the first autolytic cleavages of the anchor helix occurs right after residue 9, as shown by both our mass spectrometry data and previously by N-terminal sequencing [18]. Even if the ~20-residue-long anchor helix became fully extended, the calpain structures show that this N-terminal peptide is not long enough to loop back into the cleft to be cut at this site by intramolecular cleavage. 2) m-Calpain can be activated without autolysis. The recently determined Ca2+-activated m-calpain structures in complex with a calpastatin inhibitory domain were solved using an enzymatically dead mutant [13,14]. Even in the absence of autolysis, m-calpain was still able to adopt what appears to be its active conformation. In this case, release of the anchor helix, rather than cleavage, allowed the catalytic cleft to be formed. 3) Autolysis is too risky as an activation mechanism, because many cut sites are potentially inactivating. Various loops and extended regions in other domains, particularly Domain III, are likely be more susceptible to autolysis than the unreleased anchor helix, which should be protected from proteolysis by its alpha-helical secondary structure. Only after calpain activation with release of the anchor helix from the small subunit should it become susceptible to proteolysis. Our in vitro study showed that, in addition to the released, unstructured anchor helix, there were many autolysis sites located in Domain III that were cleaved within the first minute of

J.S. Chou et al. / Biochimica et Biophysica Acta 1814 (2011) 864–872

digestion. This confirms previous reports that Domain III is very susceptible to proteolysis both by calpain [43] and by exogenous proteases [44,45]. Recombinant rat m-calpain [44] and calpain purified from bovine muscle [45] are both cut by the exogenous proteases trypsin and chymotrypsin mainly in Domain III with a few cuts in the anchor helix. Although the cut sites identified from these two studies differ from the ones we identified due to the different cleavage specificities of the three enzymes (trypsin, chymotrypsin and calpain), the overall picture is clear: Domain III is prone to proteolysis. Indeed in our in vitro study, there was such extensive autolysis of Domain III after 3 h that no residual traces of this domain were detected on gels by mass spectrometry. 4) The autolysis process is irreversible, thus unfavorable as an activation mechanism. It would be more advantageous for the cell if calpain avoided autolysis in vivo and could be used more than once. 5) m-Calpain autolysis is most likely an in vitro phenomenon. Removal of all other proteins during purification results in calpain concentrations being much higher than those found in the cell. Also, the enzyme is surrounded by other calpain molecules, which become its substrates rather than the thousands of other proteins that it would normally be next to in the cell, including its natural targets in the cytoskeleton. In at least one previous study [46], substrate cleavage precedes autolysis of the large subunit, showing that the presence of substrates may well delay autolysis. 6) Lastly, the very first calpain molecule cannot be activated if activation requires autolysis. Domain III autolysis sites, like those in the anchor helix, are remote from the active site and can only be cut by intermolecular autolysis due to structural considerations. From this perspective, we agree with the observation that mcalpain autolysis is entirely intermolecular [47], and not intramolecular like that of calpain-3 [48]. This begs the question that if autolysis was intermolecular and is required for activation, how is the very first calpain molecule activated in the presence of Ca2+? We conclude from these arguments that autolysis is not required for m-calpain activation. 4.2. The case against calpain subunit dissociation A number of research groups have previously claimed that, in the presence of Ca2+, calpain's small subunit dissociates away from the large subunit, thereby resulting in the formation of active calpain [23–28]. These dissociated small subunits then form homodimers. Recent work from our laboratory using a combination of size-exclusion chromatography and multi-angle light scattering showed that an enzymatically dead calpain mutant was intact both with and without Ca2+ present (Hanna et al. unpublished). Analysis of calpain that was aggregated and precipitated in the presence of excess Ca2+ revealed that the small subunit was still associated with the large subunit in a 1:1 stoichiometry. Therefore, subunit dissociation was not seen for inactive calpain lacking Domain V. Two reports published after the first calpain structures became available [9,12] have stated that subunit dissociation occurs in the presence of autolysis but is not needed for activation of calpain [29,30]. In this study, we worked with active m-calpain and its autolyzed products to see if subunit dissociation was observed in vitro in the presence of Ca2+ and autolysis. The most convincing evidence that the small subunit of calpain does not dissociate from the rest of the enzyme came from anionexchange chromatography. Pure small subunit homodimer [49] did not bind to the column, while autolyzed m-calpain products did bind with differing affinities. This suggested that none of the autolyzed species were the small subunit homodimer expected from subunit dissociation. For further verification, the different autolyzed products separated by anion-exchange chromatography and then SDS-PAGE were analyzed and identified by peptide mass fingerprinting. The fact that the identified peptides originating from Domain IV and Domain VI were always found together, and none of the fractions corre-

871

sponded to peptides solely from the small subunit, confirmed our earlier results that the small subunit stayed with the large subunit. This study shows that subunit dissociation does not occur during autolysis of recombinant m-calpain lacking Domain V. 4.3. Mechanism of m-calpain activation Since there is no proof that autolysis and subunit dissociation are required for activation, we propose a revised model for m-calpain activation independent of either. When m-calpain is in the presence of sufficient Ca2+ to occupy many of the ten known binding sites, the anchor helix is released, thereby allowing the catalytic cleft to be formed through the concerted binding of two Ca2+ to the protease core after which the enzyme becomes active (Fig. 1C). Judging by the sigmoidal shape of calpain activity plotted as a function of [Ca2+], this is a highly cooperative process [50]. In enriched preparations of calpain, the disordered anchor helix and Domain III are then intermolecularly autolyzed by neighboring calpains, resulting in two main products: the protease core (Domains I and II) and the heterodimer (Domains IV and VI). Domain III can be so extensively autolyzed that it is reduced to the level of small peptides. In contradiction of some previous reports by other groups, we maintain that autolysis occurs only after calpain has become active. Even then calpain autolysis is probably not a physiological response but a pathological one. We suggest that under normal conditions, small pockets of the cell's calpain are transiently activated in the immediate vicinity of a calcium influx. Calpain is deactivated by the loss of Ca2+ to diffusion before the enzyme can find and cut another calpain in an intermolecular reaction. If high levels of Ca2+ persist, calpain might be bound and silenced by calpastatin until the Ca2+ levels fall. While not ruling out post-translational modification and calpain binding partner(s) as activation aides, our model describes the simplest scenario where calpains in the immediate vicinity of a localized pulse of Ca2+ are transiently activated and then deactivated before intermolecular autolysis can do any damage to the enzyme. As shown by size-exclusion, Phenyl Sepharose, and anion-exchange chromatographies, Domain IV of the large subunit has strikingly similar size, hydrophobicity, and charge when compared to Domain VI of the small subunit. Differentiating between homodimers and heterodimers of these two domains was difficult: antibodies targeting one domain cross-reacted with the other due to their structural similarities and affinity tags were cleaved by autolysis (data not shown). It would not be surprising if researchers have previously misinterpreted the autolysisgenerated heterodimer product identified in this study as the homodimer from subunit dissociation. Supplementary materials related to this article can be found online at doi:10.1016/j.bbapap.2011.04.007. Acknowledgments We are grateful to David McLeod of the Protein Function Discovery Facility at Queen's University for mass spectrometry analyses and to Sherry Gauthier for technical assistance. This work was funded by research grants to PLD from the Canadian Institutes for Health Research [74681] and to KG from the Fund for Scientific ResearchFlanders (Belgium) [project number G.0048.08], the Concerted Research Actions [project BOF07/GOA/012] from Ghent University, and the Interuniversity Attraction Poles [IUAP06]. JSC is the recipient of a Heart & Stroke Master's Studentship Award, F.I. is a Postdoctoral Fellow of the Research Foundation — Flanders (FWO-Vlaanderen), and PLD holds a Canada Research Chair in Protein Engineering. References [1] D.E. Goll, V.F. Thompson, H. Li, W. Wei, J. Cong, The calpain system, Physiol. Rev. 83 (2003) 731–801. [2] Y. Huang, K.K. Wang, The calpain family and human disease, Trends Mol. Med. 7 (2001) 355–362.

872

J.S. Chou et al. / Biochimica et Biophysica Acta 1814 (2011) 864–872

[3] S. Sandmann, M. Yu, T. Unger, Transcriptional and translational regulation of calpain in the rat heart after myocardial infarction — effects of AT(1) and AT(2) receptor antagonists and ACE inhibitor, Br. J. Pharmacol. 132 (2001) 767–777. [4] I. Richard, O. Broux, V. Allamand, F. Fougerousse, N. Chiannilkulchai, N. Bourg, L. Brenguier, C. Devaud, P. Pasturaud, C. Roudaut, et al., Mutations in the proteolytic enzyme calpain 3 cause limb-girdle muscular dystrophy type 2A, Cell 81 (1995) 27–40. [5] Y. Yoshikawa, H. Mukai, F. Hino, K. Asada, I. Kato, Isolation of two novel genes, down-regulated in gastric cancer, Jpn. J. Cancer Res. 91 (2000) 459–463. [6] Y. Horikawa, N. Oda, N.J. Cox, X. Li, M. Orho-Melander, M. Hara, Y. Hinokio, T.H. Lindner, H. Mashima, P.E. Schwarz, L. del Bosque-Plata, Y. Oda, I. Yoshiuchi, S. Colilla, K.S. Polonsky, S. Wei, P. Concannon, N. Iwasaki, J. Schulze, L.J. Baier, C. Bogardus, L. Groop, E. Boerwinkle, C.L. Hanis, G.I. Bell, Genetic variation in the gene encoding calpain-10 is associated with type 2 diabetes mellitus, Nat. Genet. 26 (2000) 163–175. [7] K.E. Saatman, J. Creed, R. Raghupathi, Calpain as a therapeutic target in traumatic brain injury, Neurotherapeutics 7 (2010) 31–42. [8] M.V. Rao, P.S. Mohan, C.M. Peterhoff, D.S. Yang, S.D. Schmidt, P.H. Stavrides, J. Campbell, Y. Chen, Y. Jiang, P.A. Paskevich, A.M. Cataldo, V. Haroutunian, R.A. Nixon, Marked calpastatin (CAST) depletion in Alzheimer's disease accelerates cytoskeleton disruption and neurodegeneration: neuroprotection by CAST overexpression, J. Neurosci. 28 (2008) 12241–12254. [9] C.M. Hosfield, J.S. Elce, P.L. Davies, Z. Jia, Crystal structure of calpain reveals the structural basis for Ca(2+)-dependent protease activity and a novel mode of enzyme activation, EMBO J. 18 (1999) 6880–6889. [10] P. Tompa, Y. Emori, H. Sorimachi, K. Suzuki, P. Friedrich, Domain III of calpain is a Ca2+–regulated phospholipid-binding domain, Biochem. Biophys. Res. Commun. 280 (2001) 1333–1339. [11] P. Dutt, J.S. Arthur, P. Grochulski, M. Cygler, J.S. Elce, Roles of individual EF-hands in the activation of m-calpain by calcium, Biochem. J. 348 (Pt 1) (2000) 37–43. [12] S. Strobl, C. Fernandez-Catalan, M. Braun, R. Huber, H. Masumoto, K. Nakagawa, A. Irie, H. Sorimachi, G. Bourenkow, H. Bartunik, K. Suzuki, W. Bode, 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. 97 (2000) 588–592. [13] R.A. Hanna, R.L. Campbell, P.L. Davies, Calcium-bound structure of calpain and its mechanism of inhibition by calpastatin, Nature 456 (2008) 409–412. [14] T. Moldoveanu, K. Gehring, D.R. Green, Concerted multi-pronged attack by calpastatin to occlude the catalytic cleft of heterodimeric calpains, Nature 456 (2008) 404–408. [15] T. Moldoveanu, C.M. Hosfield, D. Lim, J.S. Elce, Z. Jia, P.L. Davies, A Ca(2+) switch aligns the active site of calpain, Cell 108 (2002) 649–660. [16] P. Friedrich, The intriguing Ca2+ requirement of calpain activation, Biochem. Biophys. Res. Commun. 323 (2004) 1131–1133. [17] T. Edmunds, P.A. Nagainis, S.K. Sathe, V.F. Thompson, D.E. Goll, Comparison of the autolyzed and unautolyzed forms of mu- and m-calpain from bovine skeletal muscle, Biochim. Biophys. Acta 1077 (1991) 197–208. [18] N. Brown, C. Crawford, Structural modifications associated with the change in Ca2+ sensitivity on activation of m-calpain, FEBS Lett. 322 (1993) 65–68. [19] J. Cong, D.E. Goll, A.M. Peterson, H.P. Kapprell, The role of autolysis in activity of the Ca2+-dependent proteinases (mu-calpain and m-calpain), J. Biol. Chem. 264 (1989) 10096–10103. [20] K. Suzuki, H. Sorimachi, T. Yoshizawa, K. Kinbara, S. Ishiura, Calpain: novel family members, activation, and physiologic function, Biol. Chem. Hoppe Seyler 376 (1995) 523–529. [21] A. Baki, P. Tompa, A. Alexa, O. Molnar, P. Friedrich, Autolysis parallels activation of mu-calpain, Biochem. J. 318 (Pt 3) (1996) 897–901. [22] J.S. Elce, C. Hegadorn, J.S. Arthur, Autolysis, Ca2+ requirement, and heterodimer stability in m-calpain, J. Biol. Chem. 272 (1997) 11268–11275. [23] J. Anagli, E.M. Vilei, M. Molinari, S. Calderara, E. Carafoli, Purification of active calpain by affinity chromatography on an immobilized peptide inhibitor, Eur. J. Biochem. 241 (1996) 948–954. [24] G.P. Pal, J.S. Elce, Z. Jia, Dissociation and aggregation of calpain in the presence of calcium, J. Biol. Chem. 276 (2001) 47233–47238. [25] T. Yoshizawa, H. Sorimachi, S. Tomioka, S. Ishiura, K. Suzuki, Calpain dissociates into subunits in the presence of calcium ions, Biochem. Biophys. Res. Commun. 208 (1995) 376–383.

[26] T. Yoshizawa, H. Sorimachi, S. Tomioka, S. Ishiura, K. Suzuki, A catalytic subunit of calpain possesses full proteolytic activity, FEBS Lett. 358 (1995) 101–103. [27] H. Li, V.F. Thompson, D.E. Goll, Effects of autolysis on properties of mu- and m-calpain, Biochim. Biophys. Acta 1691 (2004) 91–103. [28] K. Suzuki, H. Sorimachi, A novel aspect of calpain activation, FEBS Lett. 433 (1998) 1–4. [29] H. Kitagaki, S. Tomioka, T. Yoshizawa, H. Sorimachi, T.C. Saido, S. Ishiura, K. Suzuki, Autolysis of calpain large subunit inducing irreversible dissociation of stoichiometric heterodimer of calpain, Biosci. Biotechnol. Biochem. 64 (2000) 689–695. [30] K. Nakagawa, H. Masumoto, H. Sorimachi, K. Suzuki, Dissociation of m-calpain subunits occurs after autolysis of the N-terminus of the catalytic subunit, and is not required for activation, J. Biochem. 130 (2001) 605–611. [31] W. Zhang, R.L. Mellgren, Calpain subunits remain associated during catalysis, Biochem. Biophys. Res. Commun. 227 (1996) 891–896. [32] P. Dutt, J.S. Arthur, D.E. Croall, J.S. Elce, m-Calpain subunits remain associated in the presence of calcium, FEBS Lett. 436 (1998) 367–371. [33] T. Moldoveanu, C.M. Hosfield, D. Lim, Z. Jia, P.L. Davies, Calpain silencing by a reversible intrinsic mechanism, Nat. Struct. Biol. 10 (2003) 371–378. [34] J.S. Elce, C. Hegadorn, S. Gauthier, J.W. Vince, P.L. Davies, Recombinant calpain II: improved expression systems and production of a C105A active-site mutant for crystallography, Protein Eng. 8 (1995) 843–848. [35] P. Van Damme, F. Impens, J. Vandekerckhove, K. Gevaert, Protein processing characterized by a gel-free proteomics approach, Methods Mol. Biol. 484 (2008) 245–262. [36] B. Ghesquiere, N. Colaert, K. Helsens, L. Dejager, C. Vanhaute, K. Verleysen, K. Kas, E. Timmerman, M. Goethals, C. Libert, J. Vandekerckhove, K. Gevaert, In vitro and in vivo protein-bound tyrosine nitration characterized by diagonal chromatography, Mol. Cell. Proteomics 8 (2009) 2642–2652. [37] L. Martens, J. Vandekerckhove, K. Gevaert, DBToolkit: processing protein databases for peptide-centric proteomics, Bioinformatics (Oxford England) 21 (2005) 3584–3585. [38] P. Van Damme, L. Martens, J. Van Damme, K. Hugelier, A. Staes, J. Vandekerckhove, K. Gevaert, Caspase-specific and nonspecific in vivo protein processing during Fas-induced apoptosis, Nat. Methods 2 (2005) 771–777. [39] B. Wiederanders, Structure–function relationships in class CA1 cysteine peptidase propeptides, Acta Biochim. Pol. 50 (2003) 691–713. [40] R.M. Murphy, E. Verburg, G.D. Lamb, Ca2+ activation of diffusible and bound pools of mu-calpain in rat skeletal muscle, J. Physiol. 576 (2006) 595–612. [41] K. Hitomi, Y. Uchiyama, I. Ohkubo, M. Kunimatsu, M. Sasaki, M. Maki, Purification and characterization of the active-site-mutated recombinant human mu-calpain expressed in baculovirus-infected insect cells, Biochem. Biophys. Res. Commun. 246 (1998) 681–685. [42] R. Badugu, M. Garcia, V. Bondada, A. Joshi, J.W. Geddes, N terminus of calpain 1 is a mitochondrial targeting sequence, J. Biol. Chem. 283 (2008) 3409–3417. [43] T. Nishimura, D.E. Goll, Binding of calpain fragments to calpastatin, J. Biol. Chem. 266 (1991) 11842–11850. [44] T. Moldoveanu, C.M. Hosfield, Z. Jia, J.S. Elce, P.L. Davies, Ca(2+)-induced structural changes in rat m-calpain revealed by partial proteolysis, Biochim. Biophys. Acta 1545 (2001) 245–254. [45] V.F. Thompson, K.R. Lawson, J. Barlow, D.E. Goll, Digestion of mu- and m-calpain by trypsin and chymotrypsin, Biochim. Biophys. Acta 1648 (2003) 140–153. [46] T.C. Saido, S. Nagao, M. Shiramine, M. Tsukaguchi, T. Yoshizawa, H. Sorimachi, H. Ito, T. Tsuchiya, S. Kawashima, K. Suzuki, Distinct kinetics of subunit autolysis in mammalian m-calpain activation, FEBS Lett. 346 (1994) 263–267. [47] T. Zalewska, V.F. Thompson, D.E. Goll, Effect of phosphatidylinositol and insideout erythrocyte vesicles on autolysis of mu- and m-calpain from bovine skeletal muscle, Biochim. Biophys. Acta 1693 (2004) 125–133. [48] B.E. Garcia Diaz, S. Gauthier, P.L. Davies, Ca2+ dependency of calpain 3 (p94) activation, Biochemistry 45 (2006) 3714–3722. [49] H. Blanchard, Y. Li, M. Cygler, C.M. Kay, J. Simon, C. Arthur, P.L. Davies, J.S. Elce, Ca (2+)-binding Domain VI of rat calpain is a homodimer in solution: hydrodynamic, crystallization and preliminary X-ray diffraction studies, Protein Sci. 5 (1996) 535–537. [50] K. Graham-Siegenthaler, S. Gauthier, P.L. Davies, J.S. Elce, Active recombinant rat calpain II. Bacterially produced large and small subunits associate both in vivo and in vitro, J. Biol. Chem. 269 (1994) 30457–30460.