Two Distinct Ca2’ Proteases (Calpain I and Calpain 11) Purified Concurrently by the Same Method from Rat Kidney*

Two molecular species of calpain (Ca2+-dependent cysteine proteinases) were concurrently purified from rat kidney, both to homogeneity. Calpain I and calpain I1 having low and high Ca2+ requirements, respec- tively, were clearly separated on DEAE-cellulose chromatography at pH 7.5, and thereafter they were purified by separate but almost identical procedures which included (NH4)2S04 fractionation and successive chro-matographies on TSK-Gel G 3000 SWG, blue Sepha- rose CL-GB, and DEAE-Bio-Gel A. The purification folds and activity yields were 6170-fold and 17.8% for calpain I and 4160-fold and 11.9% for calpain 11. Ca2+ concentrations for half-maximal activation were 2 pM for calpain I and 200 p~ for calpain 11. The specific activity of calpain I1 on casein as the substrate was more than twice higher than that of calpain I. Both enzymes are heterodimers, each composed of 80,000-Da and 25,000-Da subunits. The amino acid composi- tions of calpain I and calpain I1 are very similar but not identical. Calpain I1 is more acidic (PI 4.6) than calpain I (PI 5.3). This paper is the first to describe parallel isolation and characterization of low and high Ca2+-requiring proteases from one single nonmuscular tissue. heat treatment (100 "C, 5 min), and gel filtration. The details of the purification and characterization will be described (27). elsewhere. The unit calpastatin

P To whom correspondence should be addressed. malian and avian tissues that the generic terms calpain' I for the enzyme with low Ca2+ requirement, calpain I1 for the enzyme with high Ca2+ requirement, and calpastatin for the inhibitor protein were proposed (34). Chronologically, calpain I1 was discovered in brain (35) and skeletal muscle (20) much earlier than calpain I (26)(27)(28)(29). Several reports on the purification of calpain I1 mainly from muscular tissues (36)(37)(38)(39)(40)(41)(42) have appeared. On the other hand reports on the purification of calpain I are limited in number and in scope (43-46), leaving physicochemical properties of calpain I still less elucidated, Also unclear is the relationship between calpain I and 11. To clarify these problems it is of utmost importance to purify the two enzymes from one organ. We have purified calpain I and calpain I1 in parallel from rat kidney by almost identical methods. In this report, purification methods and some comparisons of similarities and dissimilarities of the two enzymes are described.

EXPERIMENTAL PROCEDURES
Materials-Hammarsten-grade casein, monoiodoacetic acid, and calcium chloride were obtained from E. Merck, Darmstadt, Germany. DEAE-cellulose (DE 52) was purchased from Whatman, Springfield, U.K. Sephadex G-150, blue Sepharose CL-GB, Pharmalyte, molecular weight protein standards for gel filtration, and electrophoresis and isoelectric point protein standards were products of Pharmacia, Uppsala, Sweden. DEAD-Bio-Gel A was obtained from Bio-Rad; TSK-Gel G 3000 SWG column and high performance liquid chromatography apparatus were products of Toyo-Soda, Tokyo, Japan. Milli Q water purification system was obtained from Millipore Corp., Bedford, MA. Leupeptin and antipain were generous gifts from Dr. H. Umezawa, Research Institute for Microbial Chemistry, Tokyo, Japan. W-7* and W-5 were obtained from Rikaken, Nagoya, Japan. Other reagent grade chemicals were obtained from Wako Pure Chemicals, Osaka, or Nakarai Chemicals, Kyoto, Japan.

II
homogenizing buffer using a Potter-Elvehjem Teflon-glass homogenizer. The homogenate was ultracentrifuged at 105,000 X g for 90 min. The supernatant was dialyzed overnight against buffer A containing 50 mM NaCl. The dialysate was called crude extract.
Chromatographic Procedures-Purification of calpain I and calpain I1 required four different column chromatographic procedures in the following sequence: DEAE-cellulose, TSK-Gel G 3000 SWG, blue Sepharose CL-GB, and DEAE-Bio-Gel A. All chromatographic procedures were performed at 4 "C, except that TSK-Gel G 3000 SWG was used at room temperature. Flow rate, column dimensions, buffer systems and the amounts of the protein loaded to the columns used in this study are given under ''Results" and in the legends to the figures.
Assay of Calpain-Calpain activity was determined with casein as substrate. Each incubation mixture having a final volume of 1.0 ml contained 0.4% casein, 100 mM imidazole-HCI buffer, pH 7.5, 5 mM cysteine, 0.1 mM CaCl, for calpain I, and 5.0 mM CaC1, for calpain 11. After an incubation for 30 min at 30 "C the reaction was terminated by adding 1 ml of 5% trichloroacetic acid. Acid-soluble products were determined colorimetrically by the method of Ross and Schatz (47), for which 0.4 ml of the filtrate was diluted with the reagents to a total volume of 2.8 ml and absorbance at 750 nm was read against the blank. The reaction carried out without CaClz was taken as the blank, in which CaC1, was added after the trichloroacetic acid precipitation. One unit of calpain was defined as such quantities of enzyme that increased the absorbance at 750 nm by 1.0 after 30 min of incubation at 30 "C.
Electrophoresis-Polyacrylamide slab gel electrophoresis with SDS was carried out by the method of Laemmli (48). Rod gel electrophoresis without detergent was performed after the method of Davis (49).
Isoelectric focusing was carried out in thin layer agarose gel with 2.5% Pharmalyte, with a pH grandient of 4.0 to 6.5. The anode electrolyte was 0.05 M H,S04 and the cathode electrolyte was 1 M NaOH. Isoelectric point was estimated using a PI calibration kit from Pharmacia. All gels were stained wit,h Coomassie brilliant blue.
Determination of Free Calcium-Distilled water was further purified by Milli Q water purification system. Purified water which showed more than 18 megaohm cm was used. Ca2*-EGTA buffers were prepared by adding varying amounts of CaCI, to a mixture of 110 mM imidazole-HCI, pH 7.3, 5 mM 2-mercaptoethanol and 1 mM EGTA. A dissociation constant of 5.5 X M (50) was used for calculations of free calcium concentrations.
Amino Acid Composition-Desalted and lyophilized calpain I and calpain I1 (each approximately 10 Gg) were added to 0.2 ml of 6 N HC1, sealed under reduced pressure, and hydrolyzed at 105 "C for 48 and 70 h. Half-cystine was determined by performic acid oxidation according to the method of Moore (51) and tryptophan was determined spectrophotometrically in 6 M guanidine hydrochloride (52). Amino acid analysis was performed on a Hitachi model 835 amino acid analyzer (53).
Protein Determination-Protein concentration was determined by the method of Lowry et ai. (54), using crystalline bovine serum albumin as the standard.
published elsewhere. The unit of calpastatin was defined as previously

RESULTS
Enzyme Purification-The two ea2' proteases were purified by almost identical methods. Crude extract from rat kidney was applied to a column of DE 52 (3.2 x 17 cm) pre-equilibrated with buffer A containing 50 mM NaCl. The column was washed extensively with the same buffer until the absorbance at 280 nm decreased to a negligible level. No calpain activity was observed in the flow-through fraction. The adsorbed protein was eluted with a linear gradient of NaC1, 50 to 400 mM, in a total volume of 2 liters (Fig. 1).
Two peaks of Caz+ proteases appeared, one at 120 mM NaCl and the other at 250 mM NaCl. The former was assigned to calpain I and the latter to calpain I1 judging from their Ca'+ requirements. The activity profile of calpain I was biphasic because of co-existing calpastatin (endogenous inhibitor protein). By this chromatography, calpain I and calpain I1 were completely separated. The following purification steps were carried out separately for each of these two proteases. The peak fractions for each showing calpain I or I1 activity were collected and concentrated by Amicon PM-10 membrane. The concentrates were brought to 30% (NH&SO, saturation and the mixture was left for 1 h at 0 "C. After centrifugation at 10,000 rpm for 30 min, the pellets were discarded and the supernatant were made to 45% saturation with (NH,),SO, for 1 h at 0 "C. The pellets were collected by centrifugation at 10,000 rpm for 30 min and then resuspended in 4 ml of buffer B. The solutions were clarified by ultracentrifugation at 100,000 X g for 10 min and immediately applied to a column of TSK-Gel G 3000 SWG, equilibrated with buffer E. Calpain activity was eluted as a sharp symmetrical peak (Fig. 2, A and B ) . The fractions containing calpain activities were collected and dialyzed overnight against buffer A containing 50 m M NaC1. The dialysate (12 ml, 16.1 mg of protein for calpain I and 12 ml, 3.4 mg of protein for calpain 11) was applied to a column of blue Sepharose CL-GB, pre-equilibrated with the same buffer. The column was washed with the same buffer until the absorbance a t 280 nm decreased to near zero, and then the protease was eluted by buffer A containing 1 M urea ( Fig. 2, C and D ) . The enzymatically active fractions were combined and dialyzed against buffer C containing 50 mM NaCl in the case of calpain I and buffer C containing 100 mM NaCl in the case of calpain 11. For the final step of purifica- Step" activitv  11) in a total volume of 200 ml. Calpain activity was eluted as a single peak (Fig. 2 E and F ) . As summarized in Table I, the final calpain preparations were purified over 6100-fold with a 17.8% yield (calpain I) and 4100-fold with a 11.9% yield (calpain 11) compared to the respective activities in the crude materials.
EFFLUENT VOLUME (relative value) Characterization of the Purified Enzymes-Comparisons of molecular nature and physical properties of calpain I and calpain I1 are summarized in Table 11. The molecular weights of both species were determined by comparing their elution positions on Sephadex G-150 with those of molecular weight standard proteins. Calpain I was eluted at the position for an apparent molecular weight of 110,000 and calpain I1 at the position for 115,000 molecular weight. Rod gel electrophoresis of the purified enzymes without detergent showed one protein band for each. Calpain I1 was electrophoresed more rapidly than calpain I (Fig. 3 A ) . When unstained gels were sliced into 5-mm discs which were then crushed and extracted with 0.5 ml of 0.2 M imidazole-HC1 buffer, pH 7.5, a t 4 "C for 16 h, the calpain I and calpain I1 activities were found only at the respective positions for the stained protein bands with an average recovery of 6% activity. SDS-polyacrylamide gel electrophoresis revealed that both enzymes composed of 80,000-Da and 25,000-Da subunits and the pattern were indistinguishable whether the two enzymes were electrophoresed individually or simultaneously (Fig. 3 B ) . Calpain I had an Calpain I was more resistant to heating than calpain 11.
Thus, when both enzymes were incubated at 58 "C for 10 min, calpain I1 activity was completely lost while more than 30% of calpain I activity remained. Calpain I had an optimal pH of 7.0 to 7.5 and calpain I1 7.5 to 8.0 (graphic data not shown).
Both enzymes are inhibited by leupeptin, antipain, and iodoacetic acid (data not shown) as previously reported (27). On the other hand, 100 PM of W-7 and W-5, both calmodulin inhibitors (56), had no inhibitory effect on both protease activities.
Amino acid compositions of calpain I and calpain I1 are shown in Table IV. The two enzyme proteins were found to have very similar compositions; in particular, both have almost equal amounts of basic and aromatic amino acids. However, some differences exist between the two types; calpain I1 has more glutamic acid and isoleucine, whereas calpain I has more proline and valine. Table IV also contains the reported amino acid compositions of calpain I1 from porcine (3) and z I and Calpain I I 8887 chicken (37) skeletal muscle for comparison. The data for calpain I1 from rat kidney are in general agreement with those from skeletal muscle.

DISCUSSION
This is the first paper reporting the purification and characterization of both calpain I (having low Ca2+ requirement) and calpain I1 (having high Ca2+ requirement) from a nonmuscular parenchymatous organ to apparent homogeneity. From rat kidney we have concurrently purified the two Ca2+ proteases and established the following. 1) Both enzymes can be prepared to homogeneous state by almost identical procedures in spite of their distinctly different elution positions in the first-step ion exchange chromatography (Figs. 1 and 2 and Table I). 2) Both enzymes are heterodimers, each composed of 80,000-Da and 25,000-Da subunits. The two enzymes are indistinguishable on SDS-polyacrylamide gel electrophoresis (Fig. 3 B ) . 3)Calpain I can be half-maximally activated at a free Ca2+ concentration of 2 PM (Fig. 4), which is the lowest value ever reported. 4) The specific activity of calpain 11 was more than twice higher than that of calpain I. 5 ) Amino acid compositions of the two proteases are very similar but not identical (Table IV).
The primary objective of the present st.udy was to purify calpain I and calpain I1 concurrently from one single tissue, possibly by almost identical procedures for both enzymes. Our interest centered on nonmuscular tissues. Comparison of calpain I and calpain I1 has been reported only with those from muscular tissues (43,45,46), and even in these reports the two enzymes were not necessarily isolated concurrently and the isolation methods often differed between both species. We chose rat kidney as the suitable source of the enzymes, because it had been known that this tissue contained calpain I and calpain I1 abundantly and in almost equal quantities (28). We then established a new method for the purification of the two enzymes in parallel, whereby almost identical procedures for both were employed. Using the crude extract, calpain I was eluted from DEAE-cellulose column a t 120 mM NaCl and calpain I1 at 250 mM NaCl (Fig. 1). Once completely separated by ion exchange chromatography, both enzymes behaved much alike on the following purification steps (Table I). Total activities after DEAE-cellulose chromatography increased over the levels for those of the crude extract, because coexisting calpastatin (an endogenous inhibitor protein) and/or unknown protease inhibitor(s) had been removed (Table I). The most effective step in the purification was blue Sepharose CL-GB chromatography. Why the proteases bind to a blue Sepharose column is unknown, but perhaps hydrophobic interactions may play a significant role. Calpain I seems to be more hydrophobic, since it was eluted from a blue Sepharose column more slowly than calpain I1 (Fig. 2,C and D ) . Our purification method is simple, time saving, and brings a high yield (Table I). Starting from 80 rats, we could prepare concurrently homogeneous preparations of the two enzymes as rapid as within 1 week. The previous investigators (43,45,46), who compared calpain I and calpain I1 from one kind of muscle, employed rather different methods for purification of each one of the two types of calpain. They might have overlooked a possibility of utilizing almost identical procedures for both enzymes.
Upon nondenaturing gel electrophoresis, each purified enzyme gave a single band with a high mobility toward the anode for calpain I1 (Fig. 3A). In denaturing gels, both enzymes gave two bands each, corresponding to 80,000 Da and 25,000 Da. The patterns were indistinguishable (Fig. 3B). These results mean that both proteases are heterodimers having equal molecular weights and equal subunit structures, Rat Kidney Calpain I and Calpain II but they differ in electric charges, calpain I1 being more acidic.
Isoelectric points of 5.3 for calpain I and 4.6 for calpain I1 are also consistent with the data of the DEAE-cellulose chromatography, in which calpain I was eluted much earlier than calpain I1 at pH 7.5. Dayton e.? al. (43) reported that porcine skeletal muscle calpain I1 was more acidic than calpain I, but they did not determine their isoelectric points. Available evidence indicates that a native molecule of calpain I1 is a heterodimer (12, 14,[36][37][38][39][40][41][42]. On the other hand, the molecular nature of calpain I has not been made as clear as that of calpain 11. Our present data clearly demonstrate that it is also a heterodimer whose subunit structure is indistinguishable from that of calpain 11. We have experienced more than eight runs of the preparation of calpain I from rat kidney, and we have always found, upon SDS-gel electrophoresis, only two distinct bands of 80,000 Da and 25,000 Da which showed a molar ratio of 1 to 1 as determined by densitometry. With calpain 11, however, sometimes a 27,000-Da band also appeared besides 80,000-Da and 25,000-Da bands which were always seen. We could not separate the 27,000-Da protein from the 25,000-Da protein by any method, while the densitometric determination revealed the molar ratio of 80,000-Da protein versus a sum of 27,000and 25,000-Da proteins to be 1 to 1. It seems, therefore, reasonable to consider that calpain I1 of rat kidney may have microheterogeneity in its smaller subunit either determined genetically or produced by post-translational proteolysis. Inclusion of 1 mM phenylmethylsulfonyl fluoride in the buffers used for the purification of calpain I1 did not alter the appearance of 27,000-Da band. The microheterogeneity of the product may be responsible for some blurring of the banding pattern seen in nondenaturing gel of calpain I1 (Fig. 3A).
When calpain I was first discovered from canine cardiac muscle (as), it attracted much attention because of its distinctly low requirement for Ca2+ compared to previously known calpain 11. However, even with calpain I it was reported that 40-50 PM Ca2+ was required for its half-maximal activity.
By highly purifying calpain I from rat kidney and conducting the most careful experiments, we have been able to record the lowest value, 2 PM Ca2+, for the half-maximal activity of the protease (Fig. 4, solid circles). This value is close to the physiological Ca'+ concentration in cells. The Ca'+ requirement curve shown in the figure almost coincides with that of calmodulin (57). However, the activation of calpain I should not be calmodulin-dependent, because caplain, either I or 11, was not inhibited at all by calmodulin inhibitors, W-7 and W-5, and because the purified calpain preparations showed no trace of a 17,000-Da band upon SDS-gel electrophoresis (Fig. 3B). When casein was used as substrate, the specific activity of rat kidney calpain I1 was found to be 2.24 times higher than that of calpain I (Table I). Fig. 5 shows that an endogenous inhibitor protein, calpastatin, of the same origin exhibited approximately twice stronger inhibition of calpain I1 compared with calpain I. Such apparent difference in sensitivity to calpastatin between the two species of calpain had earlier been noted even when crude preparations were used (27). However, we have now become able to explain the recorded differences in terms of the difference in specific activity between calpain I and calpain 11. Ordinarily, the experiments were carried out with calpain preparations showing activities of equal units (for example, 0.4 unit each for Fig. 5 ) , which implies that at least twice as much enzyme molecules were present in a reaction mixture in the case of calpain I as in the case of calpain 11. It is, therefore, concluded that on molecular basis calpain I and calpain I1 are almost equally sensitive to calpastatin under the experimental conditions we employed. However, we have not yet been able to demonstrate how many molecules of calpain I or I1 can combine with 1 molecule of calpastatin. Takahashi-Nakamura et al. (58) reported that a 34,000-Da subunit of calpastatin from rabbit skeletal muscle combined with 1 molecule of calpain 11. Table IV shows close similarity of the amino acid compositions between calpain I and calpain I1 from rat kidney. While these compositions are not very different from those of calpain species of other origins (Table IV), they differ vastly from those of calpastatin (55) and calmodulin (59). Our present data show, however, nonidentity of amino acid compositions between calpain I and calpain I1 (Table IV), thus excluding the possibility of an interconversion by post-translational modifications such as phosphorylation (43) and limited proteolysis without removal of small peptides. Limited proteolysis in vitro with removal of small peptides was known to reduce the Ca'+ requirement of calpain I1 (42,60), but the product enzyme with low Ca'+ requirement was found to be different from the naturally occurring calpain I (45,61). Using peptide mapping and immunological cross-reactivity tests, Wheelock (62) has shown that porcine skeletal muscle calpain I and calpain I1 are composed of each one of mutually identical 30,000-Da subunit and nonidentical 80,000-Da catalytic subunit. Whether this will also apply to the case of rat kidney enzymes must await further investigations.