The Calpain Cascade μ-CALPAIN ACTIVATES m-CALPAIN

m-Calpain, which usually requires near-millimolar Ca2+ for activation, undergoes autolysis at 25 μM Ca2+ in the presence of μ-calpain. m-Calpain in itself exhibits no sign of autolysis around this Ca2+ concentration. Half-maximal rate of the reaction occurs at 30 μM Ca2+, showing that it is μ-calpain that catalyzes the limited proteolysis of m-calpain in an intermolecular reaction (“heterolysis”). This heterolytic step is accompanied by the activation of m-calpain: μ- and m-calpain preincubated together at 25 μM Ca2+ show significantly higher activity than the sum of activities of μ- and m-calpains preincubated separately. m-Calpain is sensitized to Ca2+ by μ-calpain-mediated activation: the half-maximal value of 160 μM for activation is lowered to 64 μM, which is similar to the shift found in m-calpain autoactivation. We suggest that these in vitro observations are relevant in vivo and the calpain cascade may play an important role in coordinating the functioning of calpains in living cells.

Calpain (Ca 2ϩ -activated cysteine protease) is one of the mediators of intracellular Ca 2ϩ signal in animal cells (1,2). It has two well characterized ubiquitous isoforms,and m-calpain. Both consist of a large (80-kDa) catalytic and a small (30-kDa) regulatory subunit but differ considerably in their Ca 2ϩ sensitivities: -calpain is activated at micromolar, m-calpain at millimolar free Ca 2ϩ concentration. At the resting cytoplasmic Ca 2ϩ concentration, both forms are inactive. Mandatory for their activation is an elevation in free Ca 2ϩ concentration, which results in an autolytic cleavage at the N terminus of either the large (-calpain) or the small (m-calpain) subunit (3,4). This limited autolysis, which entails significant sensitization to Ca 2ϩ (5), is considered to be the key mechanistic step in the activation of calpains (6,7).
The mode of m-calpain action in vivo is rather puzzling as activation of the isolated enzyme requires very high, nonphysiological Ca 2ϩ concentrations. Typically, Ca 2ϩ concentrations where half-maximal rate of autolysis or activation occurs range from 200 M (8) to 1.25 mM (9), a finding that motivated the search for factors other than Ca 2ϩ contributing to m-calpain activation. Numerous effectors which lower the Ca 2ϩ concen-tration of activation have been identified, including proteins (10), nucleic acids (11), and lipids (12). The prevailing activation model, partially drawing on such studies, is that activation involves translocation of calpain to the plasma membrane where interaction with lipid head groups and/or membrane proteins along with high local Ca 2ϩ concentration at Ca 2ϩ channels would suffice (2). Experimental evidence in support of this model, however, is not conclusive.
In this work we raise the possibility of an alternative means of m-calpain activation. It is known that calpain autolysis may proceed by an intermolecular reaction (13), i.e. via the autolytic cleavage of native calpain by another, already activated, calpain molecule. In principle, this would allow for a similar reaction betweenand m-calpains as they coexist in most tissues (14) and share significant structural homology (15). Our results demonstrate that this reaction indeed takes place in vitro. We discuss the possible implications of this novel finding and propose that such a "calpain cascade" may explain the activation of m-calpain at least in some tissues.

EXPERIMENTAL PROCEDURES
Calpain Preparation-Human erythrocyte -calpain was purified as described previously (16). Pig kidney m-calpain was isolated by essentially the same method, with slight modifications. Pig kidneys, kept on ice, were used as fresh as possible, usually within one-half hour after slaughter. They were minced and homogenized with 4 volumes of homogenization buffer using a commercial blender. The homogenate was centrifuged at 25,000 ϫ g for 30 min, and the supernatant was applied to the DEAE-cellulose column. m-Calpain was eluted with a linear gradient of 50 to 500 mM NaCl; activity appeared at 215 mM NaCl. Phenyl-Sepharose chromatography was then carried out as described in Ref. 17. In the subsequent step, gradient elution was used again: m-calpain appeared at 350 mM NaCl from the Q-Sepharose column. The rest of the procedure was used without modifications.and m-calpain were dialyzed into the same batch of buffer to ensure that the EGTA concentration of the reaction mixtures be independent of their original composition.
Calpain Assay-Calpain activity was determined under initial rate conditions essentially as described in Ref. 18 in the following reaction mixture: 10 mM HEPES, pH 7.5, 1 mM EGTA, 0.5 mM dithioerythritol, 0.1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine (calpain buffer), various concentrations of Ca 2ϩ , and 66 g/ml [ 3 H]casein (19) in a total volume of 75 l. The reaction was started by the addition of calpain and terminated after 3 min at 30°C by adding 58 l of 30% trichloroacetic acid and 118 l of 3.7 mg/ml bovine serum albumin. Trichloroacetic acid-soluble peptides were separated from precipitated casein by centrifugation (15,000 ϫ g ϫ 5 min), and 200 l of the supernatant was tested for radioactivity. Calpain activity was calculated from the difference in counts between the sample and a control prepared without Ca 2ϩ .
Preparation of E-64 1 -inactivated Calpains-For control purposes, calpains were irreversibly inactivated by E-64 applied at a molar excess of 350 (-calpain) or 250 (m-calpain). -Calpain was incubated with the inhibitor for 4 h at 0°C in the presence of 400 M Ca 2ϩ whereas m-calpain for 1 h at room temperature in the presence of 1 mM Ca 2ϩ . Unreacted E-64 was removed by overnight dialysis at 4°C against calpain buffer. Residual activity of calpains following this treatment was less than 1%.
Determination of the Apparent First Order Rate Constant of Autolysis or Heterolysis-m-Calpain was incubated either in the absence or in the presence of -calpain at various Ca 2ϩ concentrations at 30°C. Aliquots were withdrawn at various times, boiled for 5 min with SDS sample buffer, and run on SDS-PAGE. Optical density of the 30-kDa band was determined by densitometry. Natural logarithm of the optical density values was plotted against incubation time, and the apparent first order rate constant of the reaction was determined as the slope of the linear function fitted to these data points.
Other Procedures-SDS-PAGE was carried out according to Laemmli (20). The intensity of Coomassie Brilliant Blue-stained bands was determined by densitometry on a Bio-Rad GelDoc 1000 videodensitometer. Optical density of the bands was proportional to protein content over the calpain concentration range used, as checked in control experiments. The free Ca 2ϩ concentration of reaction mixtures was calculated with the program LIGANDY by using the stability constant log K app Ca 2ϩ ϭ 6.2266 for Ca 2ϩ -EGTA.
Materials-Erythrocyte concentrate was purchased from the National Institute of Hematology and Blood Transfusion (Budapest). Pig kidney was purchased from a local slaughterhouse. Casein (prepared according to Hammarsten) and dithioerythritol were from Merck, HEPES was from Serva, and Q-Sepharose and Blue-Sepharose were from Pharmacia Biotech Inc. [ 3 H]Formaldehyde was a DuPont NEN product. All other chemicals were from Sigma. Buffers and solutions were prepared with ion-exchanged distilled water.

RESULTS
Autolysis of m-calpain requires very high Ca 2ϩ concentrations. At 25 M Ca 2ϩ , virtually no conversion of its 30-kDa small subunit can be seen in 4 min either in the presence or absence of E-64-inactivated -calpain (Fig. 1). -Calpain, however, significantly facilitates this interconversion. The small subunit of m-calpain is converted into smaller fragments within 1 min at the same Ca 2ϩ concentration. The effect is typically catalytic: at ato m-calpain molar ratio of 1:5 and 25 M Ca 2ϩ , the reaction rate is commensurable with that of m-calpain autolysis attained at 400 M Ca 2ϩ . A rate acceleration can be observed even at as low a molar ratio as 1:20 (data not shown). Upon complete conversion of the 30-kDa subunit, the main proteolytic products are at 28 and 24 kDa with a minor band appearing at 18 kDa. This latter band is more conspicuous at higher -calpain concentrations (data not shown). The reaction thus seems to be somewhat different from m-calpain autolysis where conversion to the 18-kDa form is more direct, even though transient intermediates have been observed (21). The apparent first order rate constant of the reaction at 1:5to m-calpain molar ratio was determined at various Ca 2ϩ concentrations. Half maximal rate is observed around 30 M Ca 2ϩ (Fig. 2) which is characteristic of -calpain. The curve converges to that measured with m-calpain alone because at high Ca 2ϩ concentrations m-calpain activity dominates in the system. The Ca 2ϩ requirement of m-calpain autolysis is significantly higher, with an estimated half-maximal value at or above 500 M. Autolysis of m-calpain becomes immeasurably fast at high Ca 2ϩ concentrations, which prevents an accurate determination of the maximal autolytic rate constant.
Autolysis of m-calpain has been claimed to have two functional consequences: activation of the enzyme and sensitization to Ca 2ϩ (5,21). Our results indicate that the reaction catalyzed by -calpain has similar consequences. Activation of m-calpain is demonstrated in Fig. 3. and m-calpain at a molar ratio of 1:5 were preincubated for various times at 25 M Ca 2ϩ , and their activity was measured at 100 M Ca 2ϩ . The activity of -calpain, measured separately in the presence of E-64-inactivated m-calpain, was subtracted from this joint activity to get the contribution of m-calpain. The m-calpain activity calculated in this way increases with time and becomes almost 3 times that of m-calpain preincubated in the absence of -calpain. In control experiments, -calpain was also preincubated with heat-inactivated m-calpain or casein to ensure that this increment in activity was not due to inhibition of -calpain autodegradation by m-calpain. A similar time course of -calpain activity was observed in both instances, indicating that the increase in m-calpain activity is genuine.
The -calpain-catalyzed conversion of m-calpain also results in a significant sensitization to Ca 2ϩ (Fig. 4). Half-maximal activity of m-calpain is seen at 160 M Ca 2ϩ while that of m-calpain activated with -calpain at 64 M Ca 2ϩ . This shift is of the same magnitude as that brought about by autolysis of m-calpain from the same source (22). These results show that autolysis and heterolysis are mechanistically closely related. DISCUSSION Autolytic removal of the N-terminal segment of the large subunit with -calpain, and of the small subunit with m-calpain, is instrumental in the activation of these enzymes. We have shown in kinetic experiments previously for -calpain that the two reactions run in close parallel (7), which renders autolysis an adequate marker of activation. In fact, autolysis has become virtually synonymous to activation over the years and has been used to demonstrate activation of calpain in vivo  Fig. 1. Optical density of the 30-kDa band was determined by densitometry, and the apparent first order rate constant of its conversion was calculated as given under "Experimental Procedures." Half-maximal rate is estimated at or above 500 M Ca 2ϩ in the absence and at 30 M Ca 2ϩ in the presence of -calpain. Data points represent the average of three experiments. (23,24). Mechanistic studies of autolysis have revealed that it may proceed in both intra-and intermolecular reactions. We asked the question whether intermolecular autolysis (i.e. heterolysis) existed betweenand m-calpains.
Our results show that this reaction does take place in vitro: -calpain catalyzes the limited proteolytic interconversion of m-calpain small subunit at a low Ca 2ϩ concentration where m-calpain exhibits no autolytic activity in itself. Apparently, the reaction is not due to sensitization of m-calpain to Ca 2ϩ by a direct interaction with -calpain as the latter is applied in substoichiometric amounts, and its inactivation completely abolishes the reaction. The presence of activating factors, possibly altering Ca 2ϩ sensitivity of m-calpain (cf. Refs. 10 -12), in the -calpain preparation can also be excluded on this ground. Furthermore, the Ca 2ϩ sensitivity of this heterolytic reaction is characteristic of -calpain: half-maximal rate is at 30 M (cf. Refs. 25 and 26). This value provides good evidence that -calpain is responsible for the reaction, because autolysis of mcalpain by itself requires much higher Ca 2ϩ concentrations (EC 50% about 500 M). The end products of autolysis and heterolysis are apparently identical as judged by SDS-PAGE: both result in a truncated, 18-kDa small subunit. The course of heterolysis, however, is somewhat different from that of autolysis: it proceeds via two relatively stable intermediates of about 28 and 24 kDa. The prevalence of these intermediates may permit the demonstration of the calpain cascade in vivo.
We have checked if activation accompanies the heterolysis of m-calpain. Indeed, m-calpain is activated by -calpain at a low Ca 2ϩ concentration.and m-calpain preincubated at 25 M Ca 2ϩ exhibit an activity significantly greater than the sum of activities ofand m-calpain measured separately. This activity increment could not be due to subtracting spuriously low -calpain activities from the ( ϩ m) curve, because control experiments showed that the decrease in -calpain activity during incubation was not influenced by inactivated m-calpain or casein. The loss of -calpain activity results from autodegradation of the enzyme which follows its autolytic activation (7). The fast initial rise in activity, however, is not seen here as it cannot be resolved in the present assay which includes a 3-min incubation of the enzyme with the substrate.
These observations constitute a compelling body of evidence for the in vitro operation of a "calpain cascade," i.e. the activation of m-calpain by -calpain. As for the physiological relevance of this cascade, its functioning in vivo is to be tested. While this needs further studies, there are several considerations that make this organization likely and attractive. In the first place, it may alleviate the extreme Ca 2ϩ demand for m-calpain activity. The coexistence and similar subcellular localization of the two ubiquitous calpains in many tissues render their physical contact possible. The well-known extracellular proteolytic cascades such as those in fibrinolysis and complement activation have two important functional characteristics: (i) they accomplish significant amplification, which results in a large response to a small stimulus, and (ii) they provide multiple control to prevent unwanted and deleterious activation. One may speculate that the calpain cascade possesses similar functional attributes. The calpain cascade, along with other factors (10 -12), may provide for controlled m-calpain activation at low Ca 2ϩ concentrations. This control may be tissue-specific as distribution ofand m-calpains is not uniform (14). In tissues where -calpain is scarce, activation by the calpain cascade may be limited or even negligible. In cells where -calpain abounds, the calpain cascade may dominate in m-calpain function. It is also conceivable, that the calpain cascade not only activates but also down-regulates m-calpain due to degradation of the activated enzyme. Such a mechanism has been attributed to -calpain in vivo (27). The calpain cascade may enable a further, more subtle, way of m-calpain control. Calpain action is thought to be controlled by cellular localization, e.g. by translocation to the plasma membrane (2), to the nucleus (28), or to mitotic chromosomes (29). Localized activation of -calpain may bring about a spatially restricted activation of m-calpain, leaving other parts of the cell unaffected.
In conclusion, the calpain cascade may contribute in various and incisive ways to coordinatingand m-calpain action. Further work is needed to assess the physiological implications.
Acknowledgment-We thank Csilla Farkas for her assistance in preparing the figures. -Calpain activity was also determined separately under identical conditions and was subtracted from this joint activity to obtain the contribution of activated m-calpain (q); half-maximal activity occurs at 64 M Ca 2ϩ . m-Calpain in itself (E) requires significantly higher Ca 2ϩ concentration for activity (EC 50% ϭ 160 M). Data points are averages of duplicate experiments.