Inactivation of Calmodulin-dependent Protein Kinase IV by Autophosphorylation of Serine 332 within the Putative Calmodulin-binding Domain*

When brain calmodulin-dependent protein kinase IV is incubated with calmodulin-dependent protein kinase IV kinase under the phosphorylation conditions in the presence of Ca 2 (cid:49) /calmodulin, rapid initial incorporation of 1 mol of phosphate into 1 mol of the enzyme by the action of the kinase kinase occurs, resulting in marked activation of the enzyme, and the subsequent incorpo- ration of more than 3 mol of phosphate by autophosphorylation occurs, resulting in no significant change in the activity (Okuno, S., Kitani, T., and Fujisawa, H. (1994) J. Biochem. ( Tokyo ) 116, 923–930; Okuno, S., Kitani, T., and Fujisawa, H. (1995) J. Biochem. ( Tokyo ) 117, 686–690). After the maximal phosphorylation, the continued incu- bation in the presence of excess EGTA resulted in additional autophosphorylation of the enzyme, leading to a complete loss of the Ca 2 (cid:49) /calmodulin-dependent activity, while causing no significant change in the Ca 2 (cid:49) / calmodulin-independent activity. The amino acid se- quence analysis revealed that the autophosphorylation after removal of Ca 2 (cid:49) occurred on Ser 332 , Ser 333 , Ser 337 , and Ser 341 . Analysis by site-directed mutagenesis clearly showed that the autophosphorylation site responsible for the inactivation is Ser 332 . Thus, calmodulin-depend-ent protein kinase IV activated by the kinase kinase may lose its Ca 2 (cid:49) /calmodulin-dependent activity by autophosphorylation incubation EGTA to a final concentration of 0.24 m M , and the incubation continued for an additional 60 min. The reaction was stopped by the addition of ATP and ice-cold trichloroacetic acid at final concentrations of 2 m M and 10%, respectively. The precipitate obtained on centrifugation was washed by sonication once in 1.6 ml and then seven times in 1.0 ml of ice-cold acetone, and it was air-dried at room The resulting dissolved in 15 (cid:109) l of 20 m M Tris-HCl (pH 8.0) containing 8 M urea and after incubation for 30 min at 37 °C, the mixture was diluted with 31 (cid:109) l of 20 m M Tris-HCl (pH 8.0). After addition of 2.4 (cid:109) g (7 (cid:109) l) of lysyl endopeptidase ( Achromobacter protease I), the mixture was incubated at 30 °C for 4 Another 2.4 (cid:109) g (7 (cid:109) l) of lysyl endopep- tidase was added and the mixture was incubated for a further 14 h at 30 °C. Then 6 (cid:109) l of 0.24 M dithiothreitol (a final concentration of 20 m M was added under an argon atmosphere. the mixture was added 5 (cid:109) l of 0.56 M iodoacetamide m M and the mixture was incubated for 30 min at in the dark. Some experiments were carried out on a

Phosphoamino Acid Analysis-CaM-kinase IV phosphorylated with [␥-32 ]ATP was precipitated by the addition of ice-cold trichloroacetic acid to give a final concentration of 10%. The precipitate obtained on centrifugation was washed twice with ice-cold acetone and hydrolyzed with 6 N HCl for 2 h at 110°C. The hydrolysate was electrophoresed on a cellulose-precoated thin layer plate in pH 3.5 buffer consisting of pyridine/glacial acetic acid/H 2 O (10/100/1890), and 32 P radioactivity, and amino acids were located by autoradiography and staining with ninhydrin, respectively (23).
Lysyl Endopeptidase Digestion of Phosphorylated CaM-kinase IV-Approximately 100 g (10 g/ml) of CaM-kinase IV was incubated with 2 g (0.2 g/ml) of CaM-kinase IV kinase, in a final volume of 10 ml, in the phosphorylation mixture containing 0.5 mM ATP or [␥-32 P]ATP (6.3 ϫ 10 4 cpm/nmol) at 30°C. After incubation for 60 min, EGTA was added to a final concentration of 0.24 mM, and the incubation was continued for an additional 60 min. The reaction was stopped by the addition of ATP and ice-cold trichloroacetic acid at final concentrations of 2 mM and 10%, respectively. The precipitate obtained on centrifugation was washed by sonication once in 1.6 ml and then seven times in 1.0 ml of ice-cold acetone, and it was air-dried at room temperature. The resulting protein was dissolved in 15 l of 20 mM Tris-HCl (pH 8.0) containing 8 M urea and after incubation for 30 min at 37°C, the mixture was diluted with 31 l of 20 mM Tris-HCl (pH 8.0). After addition of 2.4 g (7 l) of lysyl endopeptidase (Achromobacter protease I), the mixture was incubated at 30°C for 4 h. Another 2.4 g (7 l) of lysyl endopeptidase was added and the mixture was incubated for a further 14 h at 30°C. Then 6 l of 0.24 M dithiothreitol (a final concentration of 20 mM) was added under an argon atmosphere. To the mixture was added 5 l of 0.56 M iodoacetamide (40 mM), and the mixture was incubated for 30 min at room temperature in the dark. Some experiments were carried out on a scale of 1/20.
Reverse-phase HPLC Fractionation of Phosphopeptide-The mixture digested by lysyl endopeptidase as described above was filtered through a 0.22-m filter and then loaded onto a C 18 reverse-phase HPLC column (0.46 ϫ 25 cm, TSK gel ODS-80Ts, Tosoh) equilibrated with 20 mM triethylamine phosphate (pH 3). The column was eluted successively, at a flow rate of 0.5 ml/min, with a linear gradient of 0 -45% (v/v) acetonitrile in the same buffer for 50 min and then 45-90% acetonitrile for 5 min and with 90% acetonitrile for 5 min. Peptide and phosphopeptide peaks were monitored spectrophotometrically at A 215 or A 230 using an on-line UV monitor, Tosoh UV-8000, and radiometrically by Cerenkov counting using an on-line detector, Raytest Ramona 90, respectively. The radioactive fractions corresponding to each peak were pooled and subjected to the second HPLC on the same column. The second column was equilibrated with 20 mM triethylamine acetate (pH 6.5) and eluted with a linear gradient of 0 -7.2% (v/v) acetonitrile in the equilibration buffer for 5 min, then 7.2-16.2% acetonitrile for 90 min, and then 16.2-90% acetonitrile for 5 min, and with 90% acetonitrile for 5 min, or it was equilibrated with 0.1% trifluoroacetic acid and eluted with a linear gradient of 0 -40% (v/v) acetonitrile in 0.1% trifluoroacetic acid for 40 min and then 40 -100% acetonitrile for 5 min, and with 100% acetonitrile for 5 min.
Modification of Phosphoserine Residues for Sequence Analysis-The phosphopeptide purified as described above was dried in a 1.5-ml Eppendorf tube by centrifugation under vacuum and solubilized in 50 l of the modification mixture, consisting of 60 l of ethanethiol, 200 l of water, 200 l of dimethyl sulfoxide, 80 l of ethanol, and 65 l of 5 N NaOH, as described by Meyer et al. (24,25). After the mixture was incubated for 1 h at 50°C under an argon atmosphere and then cooled, 5 l of acetic acid was added. The sample was stored at Ϫ20°C until subjected to amino acid sequence analysis with an Applied Biosystems model 477A protein/peptide sequenator with an on-line model 120A PTH analyzer.
Other Analytical Procedures-SDS-polyacrylamide gel electrophoresis was carried out according to the method of Laemmli (30). Gel overlay assay by biotinylated calmodulin was carried out as described by Kincaid et al. (31). The concentration of proteins was determined by the method of Lowry et al. (32), as modified by Peterson (33) with bovine serum albumin as a standard. CaM-kinase IV (10 g/ml) expressed in Sf9 cells was incubated with 0.2 g/ml CaM-kinase IV kinase in the phosphorylation mixture containing nonradioactive ATP at 30°C, as described under "Experimental Procedures" (circles). At 60 min, EGTA was added to a final concentration of 0.24 mM (triangles), and at 65 min, EDTA was added to a final concentration of 10 mM (squares). At the indicated times, 2-l aliquots were withdrawn, and the CaM-kinase IV activity was determined in the presence of Ca 2ϩ (closed symbols) or EGTA (open symbols) for 1 min, as described under "Experimental Procedures." B, CaM-kinase IV was incubated under the same conditions as described above, except that radioactive ATP (3.7 ϫ 10 5 cpm/nmol) was used. At the indicated times, 4-l aliquots were withdrawn and the incorporation of [ 32 P]phosphate into protein was determined as described under "Experimental Procedures." The reproducibility of the data was confirmed by five independent experiments.

RESULTS
Activation of CaM-kinase IV by CaM-kinase IV Kinase followed by Inactivation upon Autophosphorylation in the Absence of Ca 2ϩ -When recombinant rat brain CaM-kinase IV␣ expressed in insect Sf9 cells was incubated with CaM-kinase IV kinase purified from rat cerebral cortex under the phosphorylation conditions in the presence of Ca 2ϩ /calmodulin, rapid marked activation and rather slow phosphorylation of the enzyme occurred as shown in Fig. 1, in agreement with our earlier observation (8). After the phosphorylation had reached a max-imum level, the extent of the phosphorylation being approximately 4.2 mol of phosphate/mol of the enzyme, the addition of excess EGTA to remove free Ca 2ϩ resulted in an additional fairly slow incorporation of phosphate into the enzyme, the extent finally reaching approximately 7.2 mol of phosphate/mol of the enzyme (Fig. 1B), indicating that 3 mol of phosphate was incorporated into 1 mol of the enzyme in the presence of EGTA. The phosphorylation after removal of Ca 2ϩ was accompanied by loss of the Ca 2ϩ /calmodulin-dependent activity but by no significant loss of the Ca 2ϩ /calmodulin-independent activity of the enzyme (Fig. 1A) and after incubation for 5 h, the enzyme became an almost completely Ca 2ϩ /calmodulin-independent form whose activity was 15 times higher than that of the original enzyme. When excess EDTA was added to the incubation mixture to chelate free Mg 2ϩ at 5 min after the addition of EGTA, both the phosphorylation and inactivation of the enzyme in the presence of EGTA were completely blocked instantaneously. These results, taken together, suggest that the phosphorylation of CaM-kinase IV in the presence of EGTA resulted in the almost complete loss of the Ca 2ϩ /calmodulin-dependent activity, although not affecting the Ca 2ϩ /calmodulin-independent activity. Similar results were obtained with CaM-kinase IV purified from rat brain (data not shown).
To characterize the mechanism of the phosphorylation causing the inactivation of the enzyme, effect of varying the concentration of the enzyme on the phosphorylation rate was investigated as shown in Fig. 2. The plot of the logarithm of the phosphorylation rate versus the logarithm of the enzyme concentration (van't Hoff plot) (34, 35) gave a straight line with a slope of 1.2 (approximately 1), suggesting that the phosphorylation of CaM-kinase IV in the absence of Ca 2ϩ causing the enzyme inactivation occurs through an intramolecular autophosphorylation mechanism. The fact that the rate of the phosphorylation after removal of Ca 2ϩ was not affected by increasing the concentration of CaM-kinase IV kinase (data not shown) provides support for the autophosphorylation mechanism.
Identification of Autophosphorylation Sites in the Absence of Ca 2ϩ -Analysis of the phosphorylated amino acids in CaMkinase IV which had been phosphorylated by incubation with CaM-kinase IV kinase in the presence of Ca 2ϩ /calmodulin or phosphorylated after the addition of EGTA, as shown in Fig. 3, indicated that CaM-kinase IV was much more strongly phos-  phorylated on serine residues than on other phosphorylatable amino acids in either case. In order to identify the phosphorylation sites of CaM-kinase IV after removal of Ca 2ϩ , phosphorylated CaM-kinase IV was digested with lysyl endopeptidase and subjected to reverse-phase HPLC (Fig. 4), and the purified phosphopeptides were analyzed by a protein sequenator (Fig.  5). The HPLC analysis (in triethylamine phosphate, pH 3) of the enzyme, which had been phosphorylated with radioactive ATP for 60 min in the presence of CaM-kinase IV kinase and Ca 2ϩ /calmodulin and then for another 60 min in the presence of excess EGTA, revealed four radioactive peaks, a big broad peak eluting at a retention time around 48 min, E2, C2, and C3, as shown in Fig. 4B. The big broad peak was resolved into three peaks, ATP, E1, and C1, by the second HPLC (in 0.1% trifluoroacetic acid) (Fig. 4F). In contrast, the enzyme phosphorylated for 60 min in the presence of CaM-kinase IV kinase and Ca 2ϩ /calmodulin gave three peaks, a big broad peak at 48 min, C2, and C3, upon HPLC (pH 3), as shown in Fig. 4C, and the big broad peak was also resolved into three peaks, ATP, E1, and C1, by the second HPLC (data not shown). Thus, the radioac-tive peak E2 appeared only when the enzyme had been phosphorylated in the absence of Ca 2ϩ . The enzyme, which had been incubated with nonradioactive ATP for 60 min in the presence of CaM-kinase IV kinase and Ca 2ϩ /calmodulin and then phosphorylated with radioactive ATP for 60 min in the presence of EGTA, produced two peaks, a big but rather sharp peak at 48 min and E2, upon HPLC (pH 3) as shown in Fig. 4D, confirming that E2 contained phosphopeptide(s) which was phosphorylated only after removal of Ca 2ϩ . The big peak was resolved into two peaks, ATP and E1, by the second HPLC as shown in Fig.  4G. These results, taken together, indicate that phosphorylation of C1, C2, and C3 occurred only in the initial 60-min incubation of the enzyme in the presence of CaM-kinase IV kinase and Ca 2ϩ /calmodulin, but did not occur in the subsequent 60-min incubation in the presence of EGTA, that phosphorylation of E1 occurred during both the incubations and that phosphorylation of E2 did not occur in the initial incubation but occurred in the subsequent incubation in the absence of Ca 2ϩ . When the radioactive fractions corresponding to E2, whose phosphorylation was possibly responsible for the inacti- A and B, brain CaM-kinase IV was incubated with CaM-kinase IV kinase in the phosphorylation mixture containing radioactive ATP for 60 min and then for another 60 min in the presence of excess EGTA. C, brain CaM-kinase IV was incubated with CaM-kinase IV kinase in the mixture containing radioactive ATP for 60 min. D, brain CaM-kinase IV was incubated with CaM-kinase IV kinase in the mixture containing nonradioactive ATP for 60 min and then for another 60 min in the mixture containing radioactive ATP in the presence of excess EGTA. After phosphorylation, the protein was digested with lysyl endopeptidase, and the resulting phosphopeptides were fractionated by HPLC in the triethylamine phosphate/acetonitrile solvent system, as described under "Experimental Procedures." E and F, the radioactive fractions eluted as a big peak at about 48 min in B were subjected to the second HPLC in the trifluoroacetic acid/acetonitrile solvent system. G, the radioactive fractions eluted as a big peak at about 48 min in D were subjected to the second HPLC in the trifluoroacetic acid/acetonitrile solvent system. H, the radioactive fractions corresponding Peak E2 in B were subjected to the second HPLC in the triethylamine acetate/acetonitrile solvent system. Peptide (A, E, and inset in H) and phosphopeptide (B-D and F-H) peaks were monitored spectrophotometrically and radiometrically, respectively, as described under "Experimental Procedures." The broken lines indicate the concentrations of acetonitrile. The recoveries of radioactivity upon HPLC were 80 -97%. vation of the enzyme, were subjected to HPLC in triethylamine acetate, pH 6.5, six contiguous radioactive peaks, E2-1, E2-2, E2-3, E2-4, E2-5, and E2-6, were obtained (Fig. 4H). Similar results were obtained with recombinant CaM-kinase IV expressed in Sf9 cells, except that a small radioactive peak, C2 (Fig. 4, B and C), was not observed with the recombinant enzyme, suggesting that C2 was derived from the amino-terminal serine-rich segment (MSCAGNDQAAASGSSSGSGGI-FRSPAAK) of CaM-kinase IV␤ (36), because only this segment of 28 amino acids is deleted in CaM-kinase IV␣. Fig. 5 shows the results of amino acid sequence analysis of phosphopeptides, C1, C3, E1, E2-4, and E2-6, by an automated protein sequenator. The sequences of C1 (Fig. 5A), C3 (Fig. 5B), E1 (Fig. 5C), and E2-4 (Fig. 5D) and E2-6 ( Fig. 5E) were assigned to the proteolytic peptides corresponding to Arg 435 -Lys 442 (C1), Val 4 -Lys 34 (C3), Ala 351 -Lys 362 (E1), and Ala 328 -Lys 350 (E2-4 and E2-6) of the predicted amino acid sequence of CaM-kinase IV␣ (37,38), respectively (Fig. 8). Since C1 contains only one serine residue, the phosphorylation site in C1 may be Ser 437 , consistent with our earlier observation (39). To determine the precise serine residues phosphorylated in phosphopeptides C3, E1, E2-4, and E2-6, each phosphopeptide was treated with alkaliethanethiol to convert specifically the phosphoserine residue to S-ethylcysteine prior to sequence analysis (24,25). As shown in Fig. 5, Ser 8 , Ser 11 , and Ser 15 in C3 (Fig. 5B), Ser 353 in E1 (Fig.   5C), Ser 332 , Ser 333 , Ser 337 , and Ser 341 in E2-4 (Fig. 5D), and Ser 333 , Ser 337 , and Ser 341 in E2-6 ( Fig. 5E) were identified as the phosphorylation sites. Although amino acid sequence analysis suggested that phosphopeptides E2-1, E2-2, E2-3, and E2-5 were the proteolytic peptide corresponding to Ala 328 -Lys 350 , significant amounts of S-ethylcysteine were not detected, owing to low abundance of these peptides. The results thus far obtained suggest that the phosphorylation of one or some of Ser 332 , Ser 333 , Ser 337 , and Ser 341 may be responsible for the inactivation of the enzyme.
Analysis by Site-directed Mutagenesis-In order to establish which serine residue of Ser 332 , Ser 333 , Ser 337 , and Ser 341 is responsible for the enzyme inactivation upon autophosphorylation in the absence of Ca 2ϩ , these four serine residues were replaced with alanine to eliminate their phosphorylation and with aspartic acid to mimic their phosphorylation. Fig. 6 shows the time course of changes in the enzyme activity of the recombinant wild-type and mutant enzymes during the incubation with CaM-kinase IV kinase in the presence of Ca 2ϩ /calmodulin followed by the incubation in the presence of excess EGTA. The reactions were carried out using the same amounts (2.8 l) of the E. coli extracts. The activity of the wild-type enzyme was markedly increased with incubation under the phosphorylation conditions in the presence of CaM-kinase IV kinase and Ca 2ϩ / calmodulin, and the subsequent addition of excess EGTA re- sulted in a time-dependent loss of Ca 2ϩ /calmodulin-dependent activity without affecting the Ca 2ϩ /calmodulin-independent activity (Fig. 6A), in accord with the result obtained with the recombinant Sf9 enzyme (Fig. 1A). In contrast, the mutant enzyme S332A, in which Ser 332 was replaced with alanine, was markedly activated by incubation with CaM-kinase IV kinase under the phosphorylation conditions but the subsequent addition of EGTA caused no significant decrease in the enzyme activity (Fig. 6B), indicating the involvement of the phosphorylation of Ser 332 in the inactivation of the enzyme after removal of Ca 2ϩ . The incubation of the mutant enzyme S332D, in which Ser 332 was replaced with aspartic acid, with CaM-kinase IV kinase under the phosphorylation conditions resulted in a similar activation of Ca 2ϩ /calmodulin-independent activity to that of the wild-type enzyme, but caused no marked activation of Ca 2ϩ /calmodulin-dependent activity (Fig. 6C). These results suggest that Asp 332 causes the enzyme to become insensitive to stimulation by Ca 2ϩ /calmodulin by mimicking phospho-Ser 332 , confirming the importance of Ser 332 in the enzyme inactivation. Replacements of Ser 333 , the amino acid residue next to Ser 332 , with alanine ( Fig. 6D) or aspartic acid (Fig. 6E), and Ser 337 and Ser 341 with alanine or aspartic acid (data not shown), did not affect the time-dependent changes in the enzyme activity essentially. Thus, the phosphorylation of Ser 332 appears to cause loss of the Ca ϩ /calmodulin-dependent activity without affect-ing the Ca 2ϩ /calmodulin-independent activity.
Since Ser 332 is located within the putative calmodulin-binding domain of CaM-kinase IV (40, 41) (Fig. 8), the loss of the Ca 2ϩ /calmodulin-dependent activity of the wild-type enzyme autophosphorylated in the presence of EGTA and the mutant S332D was thought to be due to loss of Ca 2ϩ /calmodulin binding. Calmodulin overlay analysis (Fig. 7) showed that the wildtype enzyme, unphosphorylated or phosphorylated in the presence of Ca 2ϩ /calmodulin, and the mutant enzyme S333D bound calmodulin in the presence of Ca 2ϩ , but the wild-type enzyme phosphorylated after addition of EGTA and the mutant S332D did not significantly bind calmodulin even in the presence of Ca 2ϩ . Thus, the autophosphorylation of CaM-kinase IV on Ser 332 occurring only in the absence of Ca 2ϩ appears to cause loss of the ability of the enzyme to bind calmodulin, thereby leading to loss of its Ca 2ϩ /calmodulin-dependent activity. The shift in mobility on SDS-polyacrylamide gel electrophoresis of the enzyme upon phosphorylation in the presence of Ca 2ϩ / calmodulin observed in Fig. 7A (lane 2) has been reported previously (42). DISCUSSION CaM-kinase IV is thought to play important roles in the functioning of Ca 2ϩ in the central nervous system, along with another Ca 2ϩ -responsive multifunctional protein kinase, CaM- kinase II, and therefore the regulation of its activity is very important. Discovery of CaM-kinase IV kinase in the brain (6) provided insight into the mechanism by which CaM-kinase IV may be regulated, and phosphorylation of CaM-kinase IV by CaM-kinase IV kinase has recently been demonstrated to cause a marked activation of the enzyme (8). Unlike CaM-kinase IV, CaM-kinase II is activated upon Ca 2ϩ /calmodulin-dependent autophosphorylation at Thr 286 (9 -11). Thus, the two Ca 2ϩresponsive multifunctional protein kinases occurring abundantly in the brain are activated upon phosphorylation, by two contrasting mechanisms. After activation by Ca 2ϩ /calmodulindependent autophosphorylation at Thr 286 , CaM-kinase II undergoes autophosphorylation at Thr 305 in the absence of Ca 2ϩ , resulting in decrease in the Ca 2ϩ /calmodulin-dependent activity without decrease in the Ca 2ϩ /calmodulin-independent activity (12)(13)(14)(15)(16). The present study demonstrates that CaM-kinase IV loses its Ca 2ϩ /calmodulin-dependent activity by a similar autophosphorylation mechanism.
When CaM-kinase IV was incubated with CaM-kinase IV kinase under the phosphorylation conditions in the presence of Ca 2ϩ /calmodulin until the phosphorylation reached a maximum level (4.2 mol of phosphate/mol of enzyme), both the Ca 2ϩ /calmodulin-dependent and independent activities were rapidly activated (Fig. 1) and phosphorylation of several serine residues (Ser 8 , Ser 11 , Ser 15 , etc.) in the segment of Val 4 -Lys 34 , Ser 437 , and Ser 353 was observed (Figs. 4 and 5). Although threonine residues (Thr 196 and Thr 200 ) also have been reported to be phosphorylated in the presence of Ca 2ϩ /calmodulin (42,43), our phosphoamino acid analysis (Fig. 3) could not detect phosphothreonine, probably owing to a low ratio of phosphothreonine to phosphoserine in the phosphorylated enzyme preparation. When the incubation was continued after the ad-dition of EGTA to remove free Ca 2ϩ until the additional phosphorylation reached a maximum (additionally 3 mol of phosphate/mol of enzyme), the Ca 2ϩ /calmodulin-dependent activity was decreased and finally completely lost (Fig. 1) and phosphorylation of Ser 332 , Ser 333 , Ser 337 , Ser 341 , and Ser 353 was observed (Figs. 4 and 5). Thus, among many phosphorylation sites, four serine residues, Ser 332 , Ser 333 , Ser 337 , and Ser 341 , were phosphorylated only after removal of Ca 2ϩ .
Replacement of Ser 332 with alanine by site-directed mutagenesis completely blocked the inactivation by the incubation after removal of Ca 2ϩ (Fig. 6B), probably by elimination of the phosphorylation site responsive for the inactivation. Replacement of Ser 332 with aspartic acid made the enzyme possess only very low activity of the Ca 2ϩ /calmodulin-dependent activity even after activation by CaM-kinase IV kinase (Fig. 6C), probably by the action of Asp 332 mimicking phospho-Ser 332 . Replacement of Ser 341 , Ser 337 , and even Ser 333 next to Ser 332 with alanine or aspartic acid had essentially no effect on the time course of the enzyme activity (Fig. 6, D and E). Thus, among four serine residues which were phosphorylated in the presence of EGTA, the phosphorylation of only Ser 332 appears to cause the inactivation of Ca 2ϩ /calmodulin-dependent activity, although the possibility of a little involvement of other phosphorylation sites in the inactivation cannot be excluded, because the mutant enzyme S332D activated by CaM-kinase IV kinase exhibited very little but significant Ca 2ϩ /calmodulin-dependent activity, which was lost on incubation in the presence of EGTA (Fig. 6C). The rate of the inactivation appears to occur more slowly than that of the phosphorylation, judging from the result of Fig. 1, suggesting that the phosphorylation of Ser 332 occurs relatively slowly. The finding that only Ser 332 of the four serine residues which become phosphorylatable after removal of Ca 2ϩ was not phosphorylated in phosphopeptide E2-4 ( Fig.  5E), although all the four serine residues were phosphorylated in E2-6 ( Fig. 5D), also suggests the slow phosphorylation of Ser 332 . The inactivation rate of the mutant enzyme S333A was much higher than that of the wild-type enzyme, and that of S333D was also significantly higher than that of the wild-type enzyme, as shown in Fig. 6, indicate that the adjacent amino acids affects the phosphorylation rate of Ser 332 . As shown in Fig. 8, alignment of the amino acid sequences of CaM-kinase IV and CaM-kinase II showed that Ser 332 may be located within or near the calmodulin-binding domain of CaM-kinase IV. The fact that replacement of Ser 332 with aspartic acid strongly blocked the calmodulin binding indicates that Ser 332 is located within the calmodulin-binding domain. In contrast, replacement of Ser 333 , the residue next to Ser 332 , with aspartic acid did not affect the calmodulin binding, indicating that this amino acid residue is not involved in the calmodulin binding.  3 and 7), and 0.001 g of CaMkinase IV kinase (lanes 4 and 8) were subjected to SDS-polyacrylamide gel electrophoresis on 7.5% gel. B, 0.03-l aliquots of crude extracts of bacteria carrying wild-type enzyme (lanes 1 and 4), mutant S332D (lanes 2 and 5), and S333D (lanes 3 and 6) were electrophoresed as described above. After electrophoresis, separate proteins were transferred onto polyvinylidene difluoride membranes. The membranes were blocked with 5% non-fat milk in 50 mM Tris-HCl (pH 7.5) containing 150 mM NaCl for 30 min at 24°C and then incubated with 25 g/ml biotinylated calmodulin for 60 min, followed by incubation with 2 g/ml avidin conjugated with peroxidase for 2 h, in the presence of 1 mM CaCl 2 (lanes 1-4 in A and lanes 1-3 in B)  A similar regulatory mechanism by which phosphorylation of serine or threonine residue located within a calmodulin-binding domain abolishes calmodulin binding, thereby leading to a loss of the Ca 2ϩ /calmodulin-dependent activity, is also observed with CaM-kinase II; autophosphorylation of Thr 305 within the calmodulin-binding domain in CaM-kinase II leads to a loss of the Ca 2ϩ /calmodulin-dependent activity (12)(13)(14)(15)(16). This together with the fact that Ser 332 is conserved in rat (36,38), mouse (41), and human (44 -46) CaM-kinase IV suggests that the inactivation of Ca 2ϩ /calmodulin-dependent activity by autophosphorylation at a serine or threonine residue located within a calmodulin-binding domain in the absence of Ca 2ϩ may be a rather common and physiologically important regulatory mechanism for Ca 2ϩ /calmodulin-dependent protein kinases.