Active Catalytic Fragment of Ca2+/Calmodulin-dependent Protein Kinase I1 PURIFICATION, CHARACTERIZATION, AND STRUCTURAL ANALYSIS*

We report the purification and characterization of an active catalytic fragment of Ca2+/calmodulin-de-pendent protein kinase 11, derived from autophosphorylation and subsequent limited chymotryptic digestion of the purified rat forebrain soluble kinase. The purified fragment was completely Ca2+/calmodulin-inde-pendent, existed as a monomer, and phosphorylated synapsin I at the same sites as does the native form of Ca2+/calmodulin-dependent protein kinase 11. Kinetic studies with the purified fragment revealed a more than 10-fold increase in V,,, and a 50% decrease in K , for synthetic peptide substrates, compared with native Ca2’/calmodulin-dependent protein kinase 11. No 32P-labeled autophosphorylated residues were detected in the purified active fragment, indicating that the autophosphorylation sites were not contained within this fragment. Comparative studies of this active fragment (30 kDa) and its inactive counterpart (32-kDa fragment) revealed certain structural details of both fragments. Calmodulin-overlay study, immu- noblot analysis, and direct amino acid sequencing suggest that both fragments contain the entire NH2-ter- minal catalytic domain and were generated by distinct cleavage within the regulatory domain. The putative cleavage sites for both fragments are discussed. Comparison of kinetic parametus for natiue CaM kinase 11 and for thr active fragmrnt Kinase activity was measured in the presence of Ca"/calmodulin for native CaM kinase I1 (130 ng) or in the ahsence of Ca"/calmodulin for the purified active fragment (15 ng). CaMK-(281-291) peptide (16-268 pM final concentration) or Syn I-site 3 peptide (10-328 p ~ ) was used as a suhstrate. Incuhation was performed for 1.5 s at 30 "C. Kinetic parameters, K,, and V,,,,,, were derived from double-reciprocal plots. The data represent the mean f S.E. of four independent experiments. The data for the turnover numher (k,,,,) and catalytic efficiency (kc;L,/Kmt) were calculated hy using the mean values of K,, and V,,,;,, for each substrate. The turnover numher of native CaM kinase I1 was calculated in terms of each subunit of the enzyme.

Recent studies have shown that a catalytically active, Ca2+/ calmodulin-independent fragment of CaM kinase I1 can be produced by limited proteolysis after autophosphorylation (30,31) and it has been used in some in vitro studies (30)(31)(32)(33)(34), but the fragment has not been purified to homogeneity. Limited proteolysis of CaM kinase I1 without prior autophosphorylation has been shown to generate a slightly larger, inactive fragment (31). The structural relationships between this inactive fragment, the active fragment, and the native form of CaM kinase I1 have not been determined. Sitedirected mutagenesis of the brain a subunit cDNA clone has produced forms of the kinase only partially independent of Ca2'/calmodulin (23,35).
We now report a procedure for the generation and purification of an active fragment of rat brain soluble CaM kinase 11. Comparative studies of this fragment and the inactive counterpart provided structural details of both fragments, which should help in understanding the autoregulatory mechanism of CaM kinase I1 in situ.

EXPERIMENTAL PROCEDURES
Materiak-CaM kinase I1 was purified from rat forebrain essentially as described (12), except that hydroxylapatite chromatography was added to the purification procedure. Synapsin I was prepared from bovine brain as described (36). Catalytic subunit of cyclic AMPdependent protein kinase and calmodulin were gifts from A. Horiuchi and A. C. Nairn (The Rockefeller University). [y-"'P]ATP and lZsIlabeled calmodulin were purchased from Du Pont-New England Nuclear. TLCK-treated a-chymotrypsin was purchased from Worthington. Aprotinin was purchased from Miles (West Haven, CT).
A synthetic peptide (CaMK-(281-291)) (corresponding to residues 281-15391 291' of the a subunit (20)), a CaM kinase I1 inhibitory peptide (CaMKIP) (a nonphosphorylatable analog corresponding to residues 281-302 of the a subunit in which Thr-286 is substituted by alanine (5)), and a bovine synapsin I phosphorylation site 3 peptide (Syn Isite 3) (corresponding to residues 587-609 of bovine synapsin I with a tyrosine residue added to the NH2 terminus (37,38)) were prepared by the Protein Sequencing Facility of the Rockefeller University and the Protein and Nucleic Acid Chemistry Facility of Yale University. The concentrations of peptides were determined by amino acid analysis.
Protein Kinase Assays-Kinase activity was measured as described (12), except that the peptides CaMK-(281-291) or Syn I-site 3 were used as substrate and the reactions were carried out for 1 min. Ca'+/ calmodulin-independent activity was measured in the presence of 1 mM EGTA without Ca2+/calmodulin. For kinetic studies, 200 p M [y-'"PIATP was used, and the reactions were carried out for 15 s. Kinetic parameters were derived from double-reciprocal plots.
One-dimensional Phosphopeptide Analysis of Synapsin I-Analysis was performed as described (39).
Generation of the Active and the Inactive Fragments of CaM Kinase ZI-Generation of the active fragment of CaM kinase I1 was performed by a modification of a procedure described previously (30).
CaM kinase I1 was first autophosphorylated for 10 min a t 0 "C in a reaction mixture containing 50 mM HEPES, pH 7.5, 10 mM magnesium acetate, 1 mM EGTA, 1.5 mM CaCI2, 25 pg/ml calmodulin, and 500 p~ ATP. After 10 min, TLCK-treated a-chymotrypsin was added to the incubation mixture without terminating the autophosphorylation reaction. Continued autophosphorylation during proteolysis resulted in an increased production of the active fragment, compared with the condition in which autophosphorylation was terminated before proteolysis. The incubation with a-chymotrypsin was continued for 60 min at 0 "C. The reaction was terminated by addition of an excess amount of aprotinin (10-fold excess by weight), EGTA, and EDTA (final concentration, 20 mM). A protease (a-chymotrypsin):substrate (CaM kinase 11) ratio of 1:2 was employed in order to provide for complete digestion of the kinase, generation of maximal Ca'+/calmodulin-independent activity, and total elimination of Ca2+/ calmodulin-dependent activity. Generation of the inactive fragment was performed by proteolysis without prior autophosphorylation.
CaM kinase I1 was incubated with a-chymotrypsin for 60 min a t 0 "C in 50 mM HEPES, pH 7.5, containing 1 mM EGTA, and the reaction was terminated by addition of an excess amount of aprotinin (10-fold excess by weight). The same protease:substrate ratio was used as that for the generation of the active fragment. In some experiments, inactive fragment was generated under the same incubation conditions as were used for the generation of the active fragment, except that CaCI2 was excluded from the reaction mixture.
Antibody Production and Immunoblot Analysis-Polyclonal antibodies (G-301) were raised against a synthetic peptide corresponding to amino acid residues 281-302 of the a subunit of rat brain CaM kinase I1 after conjugation to bovine thyroglobulin. Antibodies were affinity-purified with the aid of a peptide-CH-Sepharose 4B column. Immunoblot analysis was performed as described (41) using alkaline phosphatase-conjugated goat anti-rabbit IgG (1:5000 dilution) as a secondary antibody (Promega Biotec, Madison, WI).
Analysis of the NH2-terminal Amino Acid Sequences-Samples (5-10 pg of protein) were subjected to SDS-PAGE, electrophoretically transferred to polyvinylidene difluoride (PVDF) membrane (Millipore Corp., Bedford, MA), and then directly employed for amino acid sequence analysis by the method of Matsudaira (42).
In Situ Cyanogen Bromide (CNBr) Cleavage-After immobilization onto PVDF membrane as described above, proteins were subjected to in situ CNBr cleavage according to the method of Simpson and Nice (43) and then directly subjected to amino acid sequence analysis. In certain samples, o-phthalaldehyde (OPA) was used as a blocking reagent according to a modification of the method of Hulmes et al. (44).
Other Methods-SDS-PAGE was carried out according to the method of Laemmli (45). The composition of 5 X SDS sample buffer was 0.2 M Tris/HCl, pH 6.8, containing 40% (w/v) sucrose, 15% SDS, a trace amount of bromphenol blue and 10% 2-mercaptoethanol.
Protein bands on gels were visualized by using Coomassie Brilliant ' The numbering of amino acid residues in this communication is based on the sequence of the a subunit of rat brain CaM kinase I1 (14) unless otherwise noted.
Blue R-250. Protein concentration was measured by Peterson's modification (46) of the method of Lowry et al. (47) with bovine serum albumin as standard.

RESULTS
Generation and Purification of the Active 30-kDa Fragment-The active fragment of CaM kinase I1 was generated as described under "Experimental Procedures." The complete reaction sequence was performed at 0 "C rather than at 30 "C for the following reasons. At 0 "C, Ca2+/calmodulin-independent kinase activity rose dramatically for 30 min and remained constant for 150 min after the addition of a-chymotrypsin. In contrast, when the reactions were performed at 30 "C, the Ca2+/calmodulin-independent kinase activity reached a maximum value at 15 min, which was only half of that observed at 0 "C, and then declined gradually (data not shown). As shown in Fig 11. CaM kinase I1 (0.6 mg) was first autophosphorylated as described under "Experimental Procedures" and subsequently digested by TLCK-treated a-chymotrypsin (0.3 mg) for 60 min at 0 "C. The reaction was terminated by the addition of an excess amount of aprotinin, EGTA, and EDTA. The reaction mixture was applied to a fast protein liquid chromatography Mono Q HR 5/5 anion-exchange column equilibrated with 20 mM Tris/HCl, pH 7.5, containing 0.5 mM EGTA, 0.5 mM EDTA, and 10 mM 2-mercaptoethanol. After a 5-column volume wash with the same buffer, proteins were eluted with a linear gradient of NaCl (0-1.0 M) and fractions of 0.5 ml were collected. A, a 2 0 4 aliquot of the reaction mixture prior to proteolysis (lane 1 ) and after proteolysis (lane 2) and of column fractions 49-53 shown in B (lanes 3-7, respectively) was mixed with 5 pl of 5 X SDS sample buffer and subjected to SDS-PAGE using 12% acrylamide gels. Protein bands on gels were visualized by using Coomassie Brilliant Blue R-250. a, p subunit of CaM kinase 11; b, N subunit of CaM kinase 11; c, the active 30-kDa fragment; d, calmodulin; e, achymotrypsin; f, aprotinin. B, elution profile from the Mono Q HR 5/5 column. Each fraction was assayed for kinase activity in the absence of Ca'+/calmodulin using 11 @M CaMK-(281-291) as substrate (filled circles). The solid line shows absorbance a t 280 nm, and the broken line shows NaCl concentration. The large absorbance peak observed in fractions 46-49 is due to ATP and/or ADP. Tris/HCl, pH 7.5, 0.5 mM EGTA, 0.5 mM EDTA, 50% glycerol) at -20 "C, the purified fragment retained 60-80% of the initial activity after 1 week.
Characterization of the Purified Active Fragment-The purified active fragment migrated with a molecular mass of 30 kDa on SDS-PAGE under reducing conditions (Fig. 1A). The fragment eluted a t a position corresponding to a molecular mass of a 24-kDa protein on a Superose 12 HR 10/30 gel filtration column (data not shown), clearly indicating the monomeric structure of the fragment. It phosphorylated synapsin I, with a stoichiometry close to 2.0, like native CaM kinase I1 (data not shown). Fig. 2 shows phosphorylation of synapsin I (Fig. 2 A ) and a one-dimensional phosphopeptide map of phosphorylated synapsin I (Fig. 2R). The active fragment and native CaM kinase I1 produced a 35-kDa phosphopeptide (Fig. 2B, lanes I and 2), which contains phosphorylation sites 2 and 3 of synapsin I (37,39,48). A 10-kDa phosphopeptide was produced by cyclic AMP-dependent protein kinase (Fig. 2B, lane 3), which contains phosphorylation site 1 (37,39,48). Phosphorylation of synapsin I by the active fragment was inhibited by CaMKIP (data not shown). These results indicate that the substrate specificity of native CaM kinase I1 is conserved in the active fragment. Kinetic parameters of the purified active fragment were compared with those of native CaM kinase I1 (Table I), using the synthetic peptide substrates CaMK-(281-291) and Syn Isite 3. The reactions were performed under conditions in which maximal enzyme activity was observed, i.e. in the presence of Ca'+/calmodulin for native CaM kinase I1 and in the absence of Ca'+/calmodulin for the active fragment. The active fragment exhibited a 40-50% decrease in K,,, and a 10-13-fold increase in V,,,,,, compared with the native enzyme. The overall catalytic efficiency showed 19-and 12-fold increases for CaMK-(281-291) and Syn I-site 3, respectively, compared with those values of native CaM kinase 11. An increase in V,,,,, for synthetic peptide substrate was in agreement with previous observations using unpurified reaction mixtures containing the active fragment (30,31). In contrast, when synapsin I was used as substrate, both K,,, and V,,, increased, and the catalytic efficiency was virtually unchanged (data not shown). Comparative Studies of the Active 30-kDa Fragment and the Inactive 32-kDa Fragment-As previously observed (31), proteolysis of CaM kinase I1 by a-chymotrypsin without prior autophosphorylation produced a slightly larger fragment that had no kinase activity. The inactive fragment migrated with a molecular mass of 32 kDa on SDS-PAGE under reducing conditions (Fig. 3A, lane 4 ) . It was completely devoid of catalytic activity even in the presence of Ca'+/calmodulin (Fig. 3B, lane 4 ) . In contrast, the total Ca"/calmodulinindependent kinase activity of the active fragment was increased more than 10-fold after proteolysis of the autophosphorylated kinase and had no residual Ca'+/calmodulin-dependent kinase activity (Fig. 3I3, lanes I and 2). In fact, the active fragment showed 38% less kinase activity in the presence than in the absence of Ca'+/calmodulin. The presence of calmodulin alone did not affect the activity, but Ca2+ alone reduced the activity to the same extent as Ca"/calmodulin. This apparent "suppression" of the kinase activity by Ca" was also observed in the purified preparations of the active fragment (data not shown).
Since the major autophosphorylation site, Thr-286, is located within the regulatory domain, the presence or absence of autophosphorylated residues in the active fragment provides important structural information. The absence of the major autophosphorylation site in the active fragment has been suggested (31). We tested for the presence of' this autophosphorylation site, using [ -y-:'"P]ATP in the autophosphorylation reaction (Fig. 4). After proteolysis, several "'P-labeled bands smaller than 30 kDa were detected (Fig. 4 A , lane 2), but no radioactivity was detected in the purified 30-kDa protein band, even after a long exposure of the autoradiogram (Fig. 4, A, lanes 3-6, and R). This demonstrates that the specific threonyl residue that had been autophosphorylated in native CaM kinase I1 (Thr-286) was not contained within the active fragment.
Neither the active 30-kDa fragment nor the inactive 32-kDa fragment bound "'I-labeled calmodulin in the presence or absence of Ca2+ in gel overlay studies (Fig. 5). The absence of "'I-calmodulin binding to the active fragment was in agreement with previous observations (30). A faintly reactive 24-kDa protein band was detected in the active 30-kDa fragment reaction mixture after a long exposure of the autoradiogram (Fig. 5, lane 2), but was not detected in a purified fraction of the active 30-kDa fragment (data not shown). These results indicate that the critical residues required for high affinity binding to calmodulin (residues 305-309) (49) are absent in both the active 30-kDa and the inactive 32-kDa fragments.
In order to further define the COOH-terminal region of the active 30-kDa and the inactive 32-kDa fragments, immunoblot analysis was performed using affinity-purified anti-CaM kinase I1 peptide antibodies (G-301), which recognize an epitope corresponding to amino acid residues 281-302 of the a subunit. For this experiment, the purified forms of 30-kDa and 32-kDa fragments were used. The inactive 32-kDa fragment was purified on a Mono S HR 5/5 cation-exchange column. The initial buffer was 20 mM HEPES, pH 7.5, containing 0.5 mM EGTA, 0.5 mM EDTA, and 10 mM 2-mercaptoethanol, and proteins were eluted with a linear gradient of NaCl (0-1.0 M). The inactive 32-kDa fragment was identified by absorbance a t 280 nm and SDS-PAGE (data not shown). The antipeptide antibodies detected the inactive 32-kDa fragment, but not the active 30-kDa fragment (Fig. 6, lanes 2 and  3). This indicates that at least a portion of the corresponding TAME I Comparison of kinetic parametus for natiue CaM kinase 11 and for thr active fragmrnt Kinase activity was measured in the presence of Ca"/calmodulin for native CaM kinase I1 (130 ng) or in the ahsence of Ca"/calmodulin for the purified active fragment (15 ng). CaMK-(281-291) peptide (16-268 pM final concentration) or Syn I-site 3 peptide (10-328 p~) was used as a suhstrate. Incuhation was performed for 1.5 s at 30 "C. Kinetic parameters, K,, and V,,,,,, were derived from double-reciprocal plots. The data represent the mean f S.E. of four independent experiments. The data for the turnover numher (k,,,,) and catalytic efficiency (kc;L,/Kmt) were calculated hy using the mean values of K,, and V,,,;,, for each substrate. The turnover numher of native CaM kinase I1 was calculated in terms of each subunit of the enzyme.  2 and 4 ) . Incubation was terminated hy addition of 0.25 volume of 5 X SDS sample huffer either before (lanrs I and 3) or after (lanes 2 and 4 ) proteolysis, and the whole sample was subjected to SDS-PAGE using 12% acrylamide gels. Protein hands on gels were visualized by using Coomassie Brilliant Blue R-250. a, the active 30-kDa fragment; b, the inactive :32-kDa lragment. R, both fragments were generated under the same conditions as in A , except that the incubation was terminated by addition of an excess amount of aprotinin, EGTA, and EDTA. peptide sequence (281-302) is absent in the active 30-kDa fragment but is retained in the inactive 32-kDa fragment. The reactivity of the inactive 32-kDa fragment with this antibody provides clear evidence that this fragment was derived from native CaM kinase 11.
In order to determine whether the limited chymotryptic digestion resulted in cleavage of residues from the NHr terminus of the kinase subunits, the purified active 30-kDa and inactive 32-kDa fragments were subjected to SDS-PAGE and electrophoretically transferred to PVDF membranes, which were then subjected to direct amino acid sequence analysis. No sequence was obtained from either sample. The NH2terminal residue of the LY subunit of the postsynaptic density CaM kinase I1 was previously reported to be acetylated (Nacetyl-alanine) and thus blocked to Edman degradation (50).
We also subjected the purified cy and /3 subunits of soluble CaM kinase I1 to NHr-terminal amino acid sequence analysis after transfer to PVDF membranes, and no sequence was obtained from either subunit. This indicated that the NH2terminal amino acids of both the active 30-kDa and the inactive 32-kDa fragments, as well as those of the CY and /3 subunits of native CaM kinase 11, were blocked. Taken together, these data suggest that n-chymotrypsin did not cleave the NH,-terminal portion of native CaM kinase I1 under the conditions employed in this study, and that both fragments contained the intact NH?-terminal catalytic domain.
Since there are a limited number of methionine residues in the LY and p subunits of native CaM kinase I1 from the NH, terminus through the end of the regulatory domain, we performed in situ CNBr cleavage of the active 30-kDa and the inactive 32-kDa fragments after transfer to PVDF membranes, and subjected the peptide mixtures to direct amino acid sequence analysis. Fig. 7 shows a summary of the sequence analysis. For the active 30-kDa fragment, amino acid residues corresponding to the sequences of peptide 1 (starting from Gly-130) and peptide 2 (starting from Leu-252) were detected a t successive cycles of Edman degradation, but those corresponding to peptide 3 (starting from His-282) were not detected. When OPA treatment was performed after the fourth cycle, only the residues corresponding to the sequence of peptide 2 were detected, and there was a significant drop in the yield of the released amino acid after the 20th cycle. For the inactive 32-kDa fragment, amino acid residues corresponding to the sequences of peptides 1-3 were detected at each cycle. For the sequence of peptide 3, amino acids through Phe-293 were clearly detected, and small amounts of amino acids were also detected through Leu-299, although not unequivocally (data not shown). No amino acid residues corresponding to the sequence of peptide 4 (starting from Leu-308) were detected in either the active 30-kDa or the inactive 32-kDa fragment. These results suggest that the COOH terminus of the active 30-kDa fragment is Ile-271 (n)/Val-272 (p), and that the major portion of the inactive 32-kDa fragment terminates a t Phe-293, with a minor portion (roughly 15-20%) of the fragment extending to an alternative site of chymotryptic cleavage a t Leu-299.

DISCUSSION
In this report, we described a procedure for the generation and purification of the active catalytic fragment of rat brain soluble CaM kinase 11. The proteolytic generation of the active fragment required prior autophosphorylation as described previously (31). Autophosphorylation was reported not to be necessary to produce a catalytic fragment from CaM kinase I1 derived from postsynaptic densities (51). The disparity of these observations may be due to a difference in the autoregulatory mechanism of the soluble and postsynaptic density CaM kinase 11, or may be due to some other fundamental difference in these two forms of CaM kinase 11. The purified active fragment should provide a useful tool for the  2 and 5 ) , and the inactive 32-kDa fragment (lanes 3 and 6 ) , both of which were generated from 4 pg of CaM kinase I1 under the conditions described in the legend to Fig. 3, were subjected to SDS-PAGE using 12% acrylamide gels, and then electrophoretically transferred to a nitrocellulose membrane. The membrane was incubated with ""I-labeled calmodulin in the presence (/arms 1-3) or ahsence (lanes 4 -6 ) of Ca", and the binding of ".'I-labeled calmodulin was visualized by autoradiography. Kinetic properties of the purified active fragment were compared with those of the native, soluble form of CaM kinase 11. For synthetic peptide substrates, the K,, was decreased and the turnover number was increased; the overall catalytic efficiency of the active fragment increased more than 10-fold. The increase in the turnover number could be attributed to the monomeric structure of the active fragment, in which accessibility of substrate molecules might be increased, compared with the multimeric structure of the native kinase. Interestingly, the K,,, for synapsin I was increased, whereas those for synthetic peptide substrates were decreased. The affinity of each substrate for CaM kinase I1 may be partly regulated by the tertiary structure of the holoenzyme, and some portion of the regulatory or association domain may play an important role for interaction with protein substrates such as synapsin I. Comparative studies of the active 30-kDa fragment and the inactive 32-kDa fragment provided certain structural details of both fragments (Fig. 8).
1) The purified active 30-kDa fragment did not contain autophosphorylated residues. Initial autophosphorylation OCcurs a t Thr-286, which is responsible for the generation of Ca"'/calmodulin-independent activity in the autophosphorylated kinase (20-22). The reaction time employed in the current study may have been long enough for autophosphorylation to occur not only at Thr-286 but also at Ser-279 (21). The COOH-terminal cleavage site for the active 30-kDa frag-  16). X-NH indicates an unidentified blocked NH2 terminus. The numbering of amino acid residues is based on the sequence of the a subunit (14). Peptides 1-4 indicate expected CNBr-derived peptides whose NH2 termini are not blocked to Edman degradation. The lower panel shows the summary of the results of the sequence analysis of the active 30-kDa and the inactive 32-kDa fragments after in situ CNBr cleavage. Samples were prepared as described under "Experimental Procedures." 1-20 indicate the cycle numbers of the sequence analysis. The arrow indicates the location where OPA treatment was performed. X indicates that no amino acid was clearly detected. Amino acids identified at each cycle are aligned to fit the predicted sequence of CNBr-derived peptides. Only the residues that were unequivocally identified are shown. ment must be to the NHz-terminal side of Thr-286 and probably of Ser-279. This is supported by the fact that the active 30-kDa fragment did not bind calmodulin, and anti-CaM kinase I1 peptide antibodies, which recognize an epitope corresponding to amino acid residues 281-302 of the a subunit, did not react with this fragment.
2) The inactive 32-kDa fragment did not bind calmodulin, but it was recognized by the anti-CaM kinase I1 peptide antibodies. Thus, the calmodulin-binding region (residues 296-309) (52), and more specifically, the "critical" residues required for high affinity binding to calmodulin (residues 305-309) (49), are not contained within the inactive 32-kDa fragment. This fragment, however, does contain a considerable portion of residues 281-302.
3) Since both the active 30-kDa and the inactive 32-kDa fragments contain the intact NH2-terminal catalytic domain of the native subunits, it is clear that each fragment was generated by distinct cleavage within the regulatory domain. The results of amino acid sequence analysis after in situ CNBr cleavage indicate that the COOH terminus of the active 30-kDa fragment is Ile-271 (a)/Val-272 (p). They also suggest that the COOH terminus of the major portion of the inactive 32-kDa fragment is Phe-293 and that a small portion of it may extend to Leu-299. The suggested amino acid sequences of both fragments (30-kDa fragment: 1-271; 32-kDa fragment: major, 1-293 and minor, 1-299) are consistent with the results of functional studies described above. Autophosphorylation of the kinase prior to proteolysis should have induced a critical conformational change of native CaM kinase 11, which in turn exposed another preferred cleavage site for a-chymotrypsin and thus led to the generation of the smaller active 30-kDa fragment instead of the larger inactive 32-kDa fragment. The heterogeneity of the COOH terminus of the inactive 32-kDa fragment may reflect the difference of the susceptibility to chymotrypsin between the CY and / 3 subunits of the native enzyme.
Recent investigations have revealed the existence of an autoinhibitory sequence within the regulatory region of CaM kinase I1 (30,52,53), which may be involved in the regulation of CaM kinase I1 activity. Other protein kinases, such as myosin light chain kinase and protein kinase C, contain analogous regions termed pseudosubstrate sequences (54)(55)(56)(57)(58). Based on the studies using synthetic peptides, the autoinhibitory region of CaM kinase I1 has been proposed to reside within amino acid residues 281-309 (30,52,53). Binding of Ca2+/calmodulin to this region and subsequent autophosphorylation at Thr-286 seem to relieve the inhibitory constraint of this region on the catalytic domain. Thus, the autoinhibitory region may be responsible for maintaining the enzyme in an inactive form in the absence of Ca2+/calmodulin. The results of this study indicate that the entire autoinhibitory sequence is not present within the active 30-kDa fragment and that the inactive 32-kDa fragment retains at least a portion of the autoinhibitory sequence (residues 281-293 or 281-299). The observation that the inactive 32-kDa fragment cannot be activated even in the presence of Ca2+/calmodulin suggests that the presence of the retained portion of the autoinhibitory sequence, and perhaps additional residues on the NHz-terminal side of this sequence (residues 272-2801, maintains the enzyme in the inactive form. These results provide direct support for the concept that at least a certain portion of the proposed autoinhibitory sequence is responsible for the autoregulation of CaM kinase I1 in situ.