Ca2+-dependent Proteolytic Activity in Crab Claw Muscle EFFECTS OF INHIBITORS AND SPECIFICITY FOR MYOFIBRILLAR PROTEINS*

The claw closer muscle of the Bermuda land crab, Gecarcinus lateralis, undergoes a sequential atrophy and restoration during each molting cycle. We describe here the role of Ca2+-dependent proteinases in the turn- over of myofibrillar protein in normal anecdysial (in-termolt) claw muscle. Crab Ca2+-dependent proteinase degrades the myofibrillar proteins actin, myosin heavy and light chains, paramyosin, tropomyosin, and tro-ponin-T and -I. Ca2+-dependent proteinase activity in whole homogenates and 90,000 X g supernatant fractions from muscle homogenates has been characterized with respect to Ca2+ requirement, substrate specificity, and effects of proteinase inhibitors. The enzyme is inhibited by antipain, leupeptin, E-64, and iodoacetamide; it is insensitive to pepstatin A. The Ca2+-depend- ent proteinase is a sarcoplasmic cysteine proteinase that shows maximal activation at 1 mM Ca2+ at neutral pH. Since approximately 28% of the activity remains at 1.5 PM Ca2+, the enzyme is partially active at phys- iological Ca2+ concentrations. The specificity of crab Ca2+-dependent proteinase was examined with native myosin with normal ATPase activity as well as with radioiodinated myosin and radioiodinated hemolymph proteins. Hydrolysis of ‘251-myosin occurs in two phases, both Ca2+-dependent: The present study has characterized Ca*+-dependent proteinase activity in crab claw muscle with respect to substrate specificity, effective Ca*+ concentration, and effects of proteinase inhibitors. These results show similarities between crustacean and vertebrate Ca2+-dependent proteinases thereby establishing this crustacean system as a simple and convenient model for the role of Ca2’-dependent proteolysis in myofibrillar protein turnover and its manifes-tation in the structure of the sarcomere.

The claw closer muscle of the Bermuda land crab, Gecarcinus lateralis, undergoes a sequential atrophy and restoration during each molting cycle. We describe here the role of Ca2+-dependent proteinases in the turnover of myofibrillar protein in normal anecdysial (intermolt) claw muscle. Crab Ca2+-dependent proteinase degrades the myofibrillar proteins actin, myosin heavy and light chains, paramyosin, tropomyosin, and troponin-T and -I. Ca2+-dependent proteinase activity in whole homogenates and 90,000 X g supernatant fractions from muscle homogenates has been characterized with respect to Ca2+ requirement, substrate specificity, and effects of proteinase inhibitors. The enzyme is inhibited by antipain, leupeptin, E-64, and iodoacetamide; it is insensitive to pepstatin A. The Ca2+-dependent proteinase is a sarcoplasmic cysteine proteinase that shows maximal activation at 1 mM Ca2+ at neutral pH. Since approximately 28% of the activity remains at 1.5 PM Ca2+, the enzyme is partially active at physiological Ca2+ concentrations. The specificity of crab Ca2+-dependent proteinase was examined with native myosin with normal ATPase activity as well as with radioiodinated myosin and radioiodinated hemolymph proteins. Hydrolysis of '251-myosin occurs in two phases, both Ca2+-dependent: 1) heavy chain (M, = 200,000) is cleaved into four large fragments (M, = 160,000, 110,000, 73,000, 60,000) and numerous smaller fragments; light chain (M, = 18,000) is cleaved to a 15,000-Da fragment; 2) the fragments produced in the first phase are hydrolyzed to acid-soluble material. Although radioiodinated native hemolymph proteins are not susceptible to the Ca2+-dependent proteinase, those denatured by carboxymethylation are degraded. These data suggest that crab Ca2+-dependent proteinase is involved in turnover of myofibrillar protein in normal muscle and muscle undergoing proecdysial atrophy.
The role of lysosomal and nonlysosomal pathways in the turnover of myofibrillar proteins has been explored in a number of systems during both steady state and pathological conditions (see Refs. 1-3 for reviews). A Ca2+-dependent proteinase (4) active at neutral pH has been implicated in the nonlysosomal hydrolysis of myofibrillar proteins. In verte- brate muscle maintained in uitro, increased intracellular levels of Ca" stimulate protein turnover (5, 6) and are associated with dissolution of myofilaments (7) and release of a-actinin from Z lines (8). Incubation of muscle with Ca2+ ionophore increases the release of filaments from myofibrils (9). Increased levels of ca*+-dependent proteinase activity have been observed in dystrophic muscle (10) and in muscle caused to atrophy by vitamin E deficiency (11) or hyperthyroidism (12).
The claw closer muscle of the Bermuda land crab Gecarcinus lateralis experiences a cyclical atrophy and restoration during each intermolt period (13). Muscle protein decreases 40% during proecdysis, resulting in a 4-fold reduction in myofibril cross-sectional area (14). The muscle is restored following ecdysis (13). We have examined the role of Ca2+-dependent proteinases in this molt-related atrophy (see Ref. 15 for review). The calcium content of atrophic muscle homogenates is twice that of homogenates from normal muscle (16). The activity of the crab Ca*+-dependent proteinase, which degrades several myofibrillar proteins including actin and myosin, is more than 2-fold greater in proecdysial muscle (16).
A review on mammalian Ca*'-dependent proteinases mentions the existence of Ca2+-dependent proteinase activity in muscle from the horseshoe crab (18); otherwise, the occurrence of Ca*+-dependent proteinases in invertebrate muscle is largely unexplored. The present study has characterized Ca*+dependent proteinase activity in crab claw muscle with respect to substrate specificity, effective Ca*+ concentration, and effects of proteinase inhibitors. These results show similarities between crustacean and vertebrate Ca2+-dependent proteinases thereby establishing this crustacean system as a simple and convenient model for the role of Ca2'-dependent proteolysis in myofibrillar protein turnover and its manifestation in the structure of the sarcomere. and gentamicin from Schering. NaIz5I was obtained from Amersham Corp. and chloramine-T from Mallinckrodt Chemical Works.
Methods SDS-Polyacrylamide Gel Electrophoresis-Proteins were dissolved in SDS sample buffer (16). Protein was measured by fluorescence emission at 338 nm (excitation X = 286 nm); bovine serum albumin served as protein standard (16).
SDS-PAGE was done using a discontinuous gel system (16). Separating gels (7.5 and 10%) were made from a 30% acrylamide and 1% N,N'-methylenebisacrylamide stock. Bio-Rad low molecular weight standards and purified crab muscle myosin (16) were used as protein standards.
Carboxymethylation of Crab Hemolymph Proteins-Hemolymph proteins were denatured by carboxymethylation (29). A 2.5-ml solution of hemolymph proteins (10 mg/ml) in 100 mM sodium bicarbonate, pH 8.0, was brought to 10 mM in 2-mercaptoethanol, and 2.5 ml of a solution containing 8 M guanidine hydrochloride, 100 mM sodium phosphate, pH 8.0, and 3 mM EDTA were added. After 15 min at room temperature, 150 p1 of 1 M iodoacetic acid, pH 8.0, was added and the solution stirred in the dark for 20 min. After addition of 35 p1 of 2-mercaptoethanol, the solution was dialyzed against 2 liters of 20 mM Hepes-NaOH, pH 7.5,lOO mM KC], 5 mM MgC12,5 mM NaN3, 1 mM dithiothreitol, 1 mM EGTA, and 50 pg/ml of gentamicin (Buffer A). Most of the denatured protein formed a precipitate; about 13% of the protein remained in solution. SDS-PAGE showed that the composition of proteins in solution was the same as noncarboxymethylated hemolymph protein (data not shown). This soluble fraction was radioiodinated as described.
Preparation of Enzyme Extracts and Assays for Proteolytic Actiuity-Claw muscle (3-6 g) from anecdysial animals was homogenized in 75 ml of Buffer A in a Sorvall Omnimixer at % speed for 1 min. Homogenates were dialyzed for 48 h against four 2-liter changes of Buffer A containing 0.1 mM EGTA to remove free amino acids. They were assayed directly for Caz+-dependent proteinase activity.
In other experiments a soluble fraction (90,000 X g supernatant fraction) was used as a source of Ca2+-dependent proteinase activity. After homogenization, myofibrils were pelleted by centrifugation at 16,000 X g for 30 min. Supernatant fractions were dialyzed as above against Buffer A containing 0.1 mM EGTA and centrifuged at 90,000 X g for 1-3 h.
After all other solutions were autoclaved, dithiothreitol, dissolved in sterile water, was added. Unless otherwise stated, all operations were performed at 0-4 "C.
Reactions were incubated at pH 7.5 to suppress the activity of acid hydrolases present in claw muscle (31). Since 10 X stock solutions of compounds to be tested were added to dialyzed homogenates or 90,000 Buffer A were diluted 10 or 20% in the experiments below. X g supernatant fractions, the concentrations of the components of Hydrolysis of Radioiodinated Protein-Reaction mixtures (1.4 ml) contained 90,000 X g supernatant fraction (2.1-3.6 mg of protein/ml) and radioiodinated (2-3 X lo6 cpm) myosin, casein, or hemolymph proteins. After incubation at 37 "C, a 200-pl sample of the mixture was precipitated in cold 6% trichloroacetic acid; 150 p1 of the supernatant fraction was dissolved in 4 ml ACS (Amersham COT.) and counted with a Packard Tri-Carb 300C liquid scintillation counter. The acid-soluble radioactivity at 0-h incubation was less than 5% of the total label.
For SDS-PAGE, the composition of reaction mixtures (700 p l ) was the same as that described above except that they contained 10-30 X lo6 cpm of radioiodinated protein. Samples were dissolved in 3 volumes of SDS sample buffer.
Effect of Ca2+ Concentration-Muscle was homogenized and dialyzed as described above, except that in the last two changes of Buffer A, adjusted to pH 7.0, 10 g of Chelex 100 replaced EGTA. Reaction mixtures (222 pl) contained dialyzed 90,000 X g supernatant fractions (2.8-3.8 mg of protein/ml) and 1251-casein (4 x lo5 cpm).
CaCI,/EGTA buffers contained 100 p~ EGTA; the concentration of CaC12 was varied to obtain free Ca2+ concentrations from 0.01 pM to 10 mM, calculated from the equation of Bremel and Weber (32) using K = 1.9 x lo-' M at pH 7. After incubation for 1 h at 37 "C, protein was precipitated in cold 6% trichloroacetic acid and soluble radioactivity in supernatant fractions was measured as above.
In a second series, the effects of inhibitors on hydrolysis of '"1casein and '251-myosin were examined. Reaction mixtures (1 ml) contained 90,000 X g supernatant fraction (2.3-3.2 mg of protein/ ml), 5 mM CaClZ, radioiodinated protein (2 X lo6 cpm), and either 100 p~ leupeptin, 100 p~ E-64, or 10 mM EGTA, and incubated at 37 "C. Since hydrolysis of '251-myosin was linear between 1 and 4 h (see Fig. 4), the effects of inhibitors on Ca2+-dependent proteinase activity were monitored during this time interval. Hydrolysis of ' "Icasein was linear during the first hour of incubation. The specific radioactivities in these experiments were 4 X 10' cpm/pg of myosin and 1.6 X lo5 cpm/pg of casein.
Reconstitution Experiments-Low speed pellets (16,000 X g), which were washed four times in 100 ml of Buffer A and resuspended in 2 volumes of Buffer A, and 90,000 X g supernatant fractions were used in reconstitution experiments.
Three sets of reaction mixtures were used pellet alone (50 pl of resuspended pellet containing between 1.1 and 2.1 mg of protein and 200 p1 of Buffer A), supernatant fraction alone (200 pl containing between 1.0 and 1.7 mg of protein and 50 pl of Buffer A), and pellet (50 pl) and supernat.ant fraction (200 pl) combined. To each mixture 50 mM CaCL was added to a final concentration of 5 mM. After incubation at 37 "C, protein was precipitated in cold 6% trichloroacetic acid, and acid-soluble primary amines were determined by fluorescamine. For SDS-PAGE, mixtures were dissolved in 2 volumes of SDS sample buffer.
To determine the proteolytic patterns at low Caz+ concentrations, pellets were washed with Chelex 100-treated Buffer A and supernatant fractions were dialyzed against 4 2-liter changes of Buffer A, adjusted to pH 7.0, as described above, except that 5 g of Chelex 100 replaced EGTA. Compositions of reaction mixtures were the same as above with 5 mM Ca"; CaClZ/EGTA buffer stock was added to obtain free CaZc concentrations of 10 and 100 p~. After incubation for 96 h at 37 "C, mixtures were dissolved in 2 volumes of SDS sample buffer for SDS-PAGE.

RESULTS AND DISCUSSION
Washed pellets from low speed spins contained the myofibrillar proteins characteristic of arthropod muscle ( LC on one dimensional SDS-polyacrylamide gels. Washing pellets in low salt buffer followed by low speed centrifugation extracted some myosin and enriched pellets in proteins associated with thin filaments. As a result, the large amount of actin often obscured troponin-T (Fig. L4). In addition to muscle proteins, the 90,000 x g supernatant fraction contained a collection of six proteins (M, = 68,000-90,000) from the hemolymph, in all likelihood the six subunits of crustacean hemocyanin (34). These proteins are seen in the hemolymph of land crabs electrophoresed under similar conditions (16).
In the presence of 5 mM Ca2+ there was no detectable release of trichloroacetic acid-soluble material from myofibrillar proteins incubated alone (Fig. 2 ) . SDS-PAGE confirmed that there was no apparent degradation of myofibrillar proteins in these mixtures (Fig. L4, lane e). Supernatant fractions incubated alone showed only a small amount of proteolysis (Fig. 2 ) , which was associated with hydrolysis of residual amounts of actin and a soluble protein larger than myosin HC (Fig. lB, lane b, arrow) but not other soluble proteins. However, large amounts of acid-soluble primary amines were released when myofibrillar protein and supernatant fractions were incubated together (Fig. 2 ) . SDS-PAGE demonstrated that myosin HC, paramyosin, actin, troponin-T and -I, and tropomyosin were hydrolyzed (Fig. L4, lane b).
As myosin HC was degraded, two new bands appeared at 160,000 and 110,000 (Fig. L4, lane b, arrowheads). Although the degradation of myosin LC was not detected in reconstitution experiments, hydrolysis of '2sII-myosin (see below) as well as previous studies (16) showed that LC is susceptible to Ca2+-dependent proteolysis. All six hemolymph proteins re- mained undegraded, showing the specificity of the Ca2+-dependent proteinase for myofibrillar proteins. EGTA appeared to inhibit proteolysis completely (Fig. lA, lane c).
A t lower Ca2' concentrations the proteolytic pattern was essentially the same as that at 5 mM, although the lower proteolytic activity necessitated prolonged incubations of 96 h. Myosin HC was degraded as fragments appeared at 160,000 and 110,000 (Fig. 3, lanes b and c). Actin, troponin-T, and tropomyosin were also hydrolyzed while degradation of paramyosin, troponin-I, and myosin LC (not shown) was not evident under these conditions. However, given that these three proteins showed little degradation at 5 mM Ca2+ (Fig.  L4, lane b), it is not surprising that hydrolysis could not be detected at low Ca2+-dependent proteinase activities. Paramyosin would not be as susceptible to hydrolysis as myosin since paramyosin constitutes the core of the thick filament. Degradation of myofibrillar proteins was completely inhibited by EGTA (Fig. 3, lane a ) ; hemolymph proteins remained unhydrolyzed in low Ca2+ (Fig. 3, lanes b and c ) . Although the concentration of Ca2+ in these fibers is not known, the low Ca2+ concentrations selected were probably within the normal physiological range, particularly since troponin and sarcoplasmic reticulum would bind Ca", which would reduce the amount of free Cap+ available for the Ca2+-dependent proteinase. Thus these data show that contractile proteins are degraded at physiological levels of Ca2+ and that substrate specificity is similar over a 500-fold range in ca2+ concentration.
These reconstitution experiments demonstrated that the Ca2+-dependent proteinase is localized in the sarcoplasm and is not closely associated with either microsomes or myofibrils, since activity remained in 90,000 X g supernatants of muscle homogenates. Crustacean muscle contains very small amounts of connective tissue, vascular elements, and nerves; satellite cells are absent (14). Both hemolymph and epidermis are significant contaminants of muscle. However, epidermis contains less than 5% of the Ca*+-dependent proteinase activity as that found in the same amount of claw muscle while no Ca2+-dependent proteinase activity was detected in homogenized hemolymph (data not shown). Since hemolymph contains hemocytes, which are also found in connective tissue (35), it is unlikely that these phagocytic cells contribute to Ca2-dependent proteolysis in muscle homogenates. Thus it appears that most of the Ca2+-dependent proteinase activity originates in claw muscle fibers. Large molecular weight fragments appeared (arrowheads, M , = 160,000 and 110,000) as myosin heavy chain (MHC) was degraded (lanes b and c). A, actin; HP, hemolymph proteins; MLC, myosin light chain; P, paramyosin; TM, tropomyosin; TNI, troponin-I; TNT, troponin-T.
Proteolytic activity in 90,000 x g supernatant fractions of muscle homogenates increased with increasing concentrations of Ca2+ from 0.10 p M to 1 mM (Fig. 4); further increases to 5 and 10 mM Ca2+ depressed proteolysis. This broad response to Ca2+ suggests there is more than one form of crab Ca2+dependent proteinase and that each form differs in Ca'+ sensitivity. Supporting data are the presence of two peaks of Ca"-dependent proteinase activity following aminohexyl-Sepharose chromatography (data not shown). Since substantial Ca"-dependent proteinase activity remained at 1.5 p~ Ca2+ (28% of the rate at 1 mM Ca"), it appears that a significant portion of Ca*+-dependent proteinase functions a t physiological levels of Ca'+ in the sarcoplasm and is involved in myofibrillar protein turnover in uiuo. The short incubation time used (1 h), combined with the remarkable stability of crab Ca'+-dependent proteinase (Ref. 16, see below), minimized possible autolysis to a form with greater Ca2+ sensitivity.
Ca2+-dependent proteinase activity in whole muscle homogenates showed a similar response to Ca2+, except that maximum activity occurred a t 5 mM rather than 1 mM Ca2+ (36) The difference between the two preparations probably re-sulted from the absence in 90,000 X g supernatants of sarcoplasmic reticulum and troponin, both of which would sequester Caz+.
Group-specific inhibitors have been used to characterize various vertebrate proteolytic enzymes in vivo and in uitro.
Pepstatin inhibits aspartic proteinases such as cathepsin D (37); leupeptin, antipain, and E-64 inhibit cysteine proteinases, including vertebrate muscle Ca2+-dependent proteinases (18,(21)(22)(23)26). Sulfhydryl reagents such as iodoacetic acid, iodoacetamide, and N-ethylmaleimide also inhibit vertebrate Ca2+-dependent proteinases (17,18,21), demonstrating the requirement of a thiol group for catalysis. Inhibitors of cysteine proteinases also reduce crab muscle Ca"-dependent proteinase activity. Table I shows the effects of several inhibitors on Ca2+-dependent proteinase activity in muscle homogenates. Although both 100 p~ antipain and leupeptin inhibited proteolysis by about 50% neither was as effective as EGTA, which inhibited more than 90%. Neither in 90,000 X g supernatants of muscle homogenates (pH 7.0) as a function of Ca2+ concentration. Supernatants were dialyzed against two 2-liter changes of Buffer A containing 100 p~ EGTA followed by two 2-liter changes of Buffer A in which 10 g of Chelex 100 replaced EGTA to remove free Ca2+. Average of four experiments, standard deviation indicated by oertical bars.  pepstatin A (100 PM), which was dissolved in 1% Me2S0, nor 1% Me2S0 alone had any effect on activity. Iodoacetamide (5 mM) inhibited Ca2+-dependent proteinase activity by 54%. E-64 (100 PM) was as effective an inhibitor as excess EGTA; as little as 10 PM yielded the same amount of inhibition as higher concentrations (data not shown). The greater effectiveness of E-64 as compared to other substrate analogs, such as leupeptin and antipain, possibly resulted from its ability to bind irreversibly to the Ca2+-dependent proteinase (38).

Effects of inhibitors on Ca'+-dependent proteolysis in whole muscle homogenutes
The effects of inhibitors on hydrolysis of radioiodinated myosin and casein in 90,000 x g supernatant fractions were qualitatively similar to those observed on proteolysis in muscle homogenates in that both E-64 and EGTA inhibited hydrolysis to a greater extent than leupeptin (Table 11). E-64 was again as effective as excess EGTA in inhibiting proteolysis. These similarities between the degradation of radioiodinated substrates and unlabeled myofilaments in the presence of specific inhibitors show that the more sensitive assay conditions employing '*'I-casein to characterize the catalytic properties of crab Ca2+-dependent proteinase are valid.
The specificity of crab Cat+-dependent proteinase was examined with radioiodinated substrates. 12'I-myosin was degraded in the presence of 5 mM Ca2+, resulting in the release of acid-soluble radioactivity; 10 mM EGTA completely inhibited hydrolysis (Fig. 5). Hydrolysis of '251-myosin occurred in two phases; there was a slow release of label during the first 10-15 min, followed by a more rapid hydrolysis. Preincubation for 20 min at 37 "C before addition of Ca2+ did not change this pattern. Under identical conditions, radioiodinated native hemolymph proteins were not hydrolyzed (Fig. 5).
Autoradiographs of SDS-polyacrylamide gels showed that when Ca2+ was present in the reaction mixture, both I2'Imyosin HC and LC were degraded into several major fragments and many minor fragments (Fig. 6, A and B). The molecular weights of the major fragments obtained from the hydrolysis of "'I-myosin HC were 160,000, 110,000, 73,000, and 60,000 Da (Fig. 6A, lanes b and c ) . The 160,000 and 110,000 fragments correspond to fragments observed in 10% gels (Fig. l A , lane b). These same fragments had been sized previously at 140,000 and 105,000 Da (16); electrophoresis on 7.5% gels (Fig. 6A) provided a more accurate estimate of molecular weights above 100,000. The two most prominent bands (M, = 73,000 and 60,000) were not observed previously (16) since their molecular weights and mobilities in acrylamide gels are similar to those of hemolymph proteins (Fig. 1B). In addition to the major fragments, numerous minor fragments, ranging in molecular weight from 25,000 to 100,000, appeared. "'I-Myosin LC was degraded to a 15,000 molecular weight fragment (Fig. 6B, lane b). After prolonged exposure (22 h) to the proteinase, the major fragments produced during the early phase of digestion were hydrolyzed without the   appearance of additional fragments (Fig. 6B, lane c). The autoradiographs showed that the short lag period before release of acid-soluble radioactivity resulted from the generation of large fragments before their subsequent hydrolysis to acidsoluble material (Fig. 5). It was possible that the fragments generated from Ca2+-dependent proteolysis in the 90,000 X g supernatant fractions were subsequently degraded by other proteolytic enzymes active a t neutral pH. T o determine whether the Ca'+-dependent proteinase was responsible for the complete hydrolysis of the '""I-myosin, reaction mixtures were incubated in the presence of 5 mM Ca2+ for 2 h during which most of the myosin was degraded. EGTA was then added to one mixture to a final concentration of 10 mM. The release of acid-soluble radioactivity was measured and compared to the release from a mixture in which 10 mM EGTA was present throughout the incubation period. Both hydrolytic phases were Ca'+-dependent since EGTA inhibited hydrolysis whether or not myosin was first cleaved into fragments (Fig. 7). Furthermore, the hydrolytic rates were similar in both after the addition of EGTA, 0.2 ng/h in mixtures with EGTA throughout the incubation and 0.3 ng/h in mixtures in which EGTA was added a t 2 h. Proteolysis in mixtures containing Ca'+ was 10fold greater (3.0 ng/h). Ca'+-independent proteolysis usually amounted to less than 10% of the total proteolytic activity when Ca2+ is present (Tables I and 11). Since there was little degradation of myofibrillar proteins in reactions lacking Ca'+ (Figs. 1,2, and 9), we conclude that other neutral proteinases play a relatively minor role in turnover of contractile proteins in claw muscle and that crab Ca'+-dependent proteinase can completely degrade both myosin HC and LC to acid-soluble material.
Previous studies (16), as well as results presented here ( Figs.  1 and 3), showed that hemolymph proteins were not degraded by crab Ca'+-dependent proteinase. Experiments with radioiodinated native hemolymph proteins confirmed our earlier observations; these proteins were not hydrolyzed by crab muscle Ca2+-dependent proteinase (Fig. 8, lane e ) . However, radioiodinated hemolymph proteins denatured by carboxymethylation were susceptible to Caz+-dependent proteolysis (lune b), which was inhibited by EGTA (lune c).
Since hemolymph proteins, which are not normally hydrolyzed by crab Ca2+-dependent proteinase, became susceptible to proteolysis when denatured, it seemed possible that degradation of "'1-myosin resulted from denaturation with chloramine-T. In fact, "'1-myosin had no ATPase activity (data not shown). However, it is clear from reconstitution experi-FIG. 7. Release of acid-soluble radioactivity from '*"Imyosin (4 X 1 0 ' cpm/pg) incubated with 90,000 X g supernatant fraction and 5 m M Ca2+ at 37 "C. EGTA was added to a final concentration of 10 mM a t 0 (0) and 2 h ((3). An equivalent volume of water was added a t 2 h to a third reaction mixture (0). Acid-soluble radioactivity a t 0 h was approximately 8 X 109 cpm.  (lanes a, b, d, and e)  cpm. The specific activities were 9 X IO4 cpm/pg for 'Z51-native hemolymph protein and 2.2 X IO5 cpmlpg for '*'II-denatured hemolymph protein.
ments ( Figs. 1 and 3) and earlier studies (16) that HC and LC in native myosin in myofibrillar fractions and whole muscle homogenates are degraded by Ca'+-dependent proteinase. To ascertain that native myosin was hydrolyzed, purified crab muscle myosin was used as substrate. This myosin preparation had a K+-EDTA ATPase activity of 1.3 pmol of ATP. mg" . min", which is comparable to Limulus muscle myosin (39). A Ca'+-dependent proteinase, partially purified by organomercurial-Sepharose chromatography, contained two major proteins (M, = 95,000 and 40,000). Incubation of that enzyme preparation with purified myosin cleaved HC to numerous fragments only in the presence of Ca" (Fig. 9). Thus, native myosin, either purified or as an integral part of a myofilament, is susceptible to Ca'+-dependent proteolysis in vitro.
Although crustacean claw muscle Ca2+-dependent proteinase resembles the cysteine proteinases purified from vertebrate muscle, there are some important differences. In the presence of Ca2+, vertebrate Ca'+-dependent proteinase undergoes extensive autolysis. The Ca'+-dependent proteinase in chicken skeletal muscle is completely inactivated in 4 h at 25 "C (40); rabbit skeletal muscle Ca'+-dependent proteinase has a half-life of 8.1 min at 37 "C (41). In contrast, the crustacean Ca2+-dependent proteinase remains active even after incubation at 37 "C for 48 h (16). The enzymes from vertebrates differ from crab Ca'+-dependent proteinase with respect to substrate specificity. Vertebrate muscle Ca'+dependent proteinase degrades myosin HC, troponin, tropomyosin, and C protein (19). The crab Ca'+-dependent proteinase hydrolyzes actin and myosin LC as well as myosin HC, troponin, tropomyosin, and paramyosin (data described here and in Ref. 16). C protein is absent from arthropod muscle (33).
The differences in substrate specificity may result from different catalytic properties of the proteinases or different susceptibilities of crustacean and vertebrate myofibrillar proteins. A Ca'+-dependent proteinase purified from Ehrlich ascites tumor cells that degrades the intermediate filament proteins vimentin and desmin but not myofibrils, actin, tubulin, azocasein, bovine serum albumin, ovalbumin, fibrinogen, neurofilaments, or histones is an example of the former (42). It appears that Ca'+-dependent proteinases, which occur in various tissues in addition to muscle (19,20), have specialized functions tailored to a particular cell type.
The many ultrastructural and biochemical similarities between crab claw muscle atrophy and atrophies induced by In the presence of Ca2+, myosin HC was degraded to at least 16 fragments (arrowheads). Actin ( A ) was a minor contaminant of the myosin preparation.
disuse, denervation, and disease in vertebrate muscle make crab claw muscle a useful model in studying protein turnover and its effect on myofibril organization. As in other atrophies, the reduction in protein mass is correlated with a decrease in myofibril cross-sectional area (14) and an increase in Ca2+dependent proteinase activity (16). Vertebrate Ca2+-dependent proteinases hydrolyze several myofibrillar proteins and release a-actinin from Z lines, suggesting that they remove filaments from the contractile apparatus, and function in the initial steps of myofilament degradation (43). Supporting evidence is the release of myofilaments during digestion of myofibrils with Ca2+-dependent proteinase (9). In addition to the other myofibrillar proteins degraded by vertebrate Ca2+dependent proteinases, crab Ca2+-dependent proteinase also degrades actin and myosin LC, which suggests its involvement in both removal and breakdown of myofilaments. Since it operates a t neutral pH and is partially activated by low levels of Ca2+, we propose that the crab Ca2+-dependent proteinase plays an important role in myofibrillar protein turnover in muscle from both molting and nonmolting animals.

D L Mykles and D M Skinner
specificity for myofibrillar proteins.