Calcium/Calmodulin-dependent Protein Kinase IV Is Cleaved by Caspase-3 and Calpain in SH-SY5Y Human Neuroblastoma Cells Undergoing Apoptosis*

We have previously demonstrated cleavage of α-spectrin by caspase-3 and calpain during apoptosis in SH-SY5Y neuroblastoma cells (Nath, R., Raser, K. J., Stafford, D., Hajimohammadreza, I., Posner, A., Allen, H., Talanian, R. V., Yuen, P., Gilbertsen, R. B., and Wang, K. K. (1996)Biochem. J. 319, 683–690). We demonstrate here that calcium/calmodulin-dependent protein kinase IV (CaMK IV) is cleaved during apoptosis by caspase-3 and calpain. We challenged SH-SY5Y cells with the pro-apoptotic agent thapsigargin. Western blot analysis revealed major CaMK IV breakdown products of 40, 38, and 33 kDa. Digestion of control SH-SY5Y lysate with purified caspase-3 produced a 38-kDa CaMK IV fragment; digestion with purified calpain produced a major fragment of 40 kDa. Pretreatment with carbobenzoxy-Asp-CH2OC(O)-2,6-dichlorobenzene or Z-Val-Ala-Asp-fluoromethylketone was able to block the caspase-3-mediated production of the 38-kDa fragment both in situ and in vitro. Calpain inhibitor II similarly blocked formation of the calpain-mediated 40-kDa fragment both in situ and in vitro. Digestion of recombinant CaMK IV by other caspase family members revealed that only caspase-3 produces a fragmentation pattern consistent to that seen in situ. The major caspase-3 and calpain cleavage sites are respectively identified as PAPD176*A and CG201*A, both within the CaMK IV catalytic domain. Furthermore, calmodulin-stimulated protein kinase activity decreases within 6 h in thapsigargin-treated SH-SY5Y. The loss of activity precedes cell death.

(1) has focused research on the role of ICE-like cysteine proteases (caspases) in apoptosis. To date, more than 10 caspases have been discovered and linked to mammalian apoptosis (2)(3)(4). The overall common features of caspases are the conservation of the active site QACXG pentapeptide, where X is R, Q, or G, and the requirement for an Asp residue in the P 1 position (5,6).
Calpain is a cysteine protease with dozens of substrates; it is activated through an increase in intracellular Ca 2ϩ (22). The two major calpain isoforms,and m-, differ in the amount of Ca 2ϩ required for activation. Calpain is triggered in necrosis (23) and in many cell types undergoing apoptosis, including SH-SY5Y (12), probably as a consequence of a loss of Ca 2ϩ homeostasis (24). CaM-binding proteins are particularly vulnerable to cleavage by calpain (25,26). We recently showed that the CaM-binding protein ␣-spectrin is proteolytically cleaved in apoptotic SH-SY5Y cells by both caspase and calpain activation (12). We suspected that other Ca 2ϩ /CaM-binding proteins may be similarly vulnerable to cleavage during apoptosis. The present study examines the fate of Ca 2ϩ /CaM-dependent protein kinase IV in SH-SY5Y cells in response to apoptotic challenge.
Several reports have demonstrated that inhibition of Ca 2ϩ / calmodulin-dependent protein kinase activity is associated with apoptosis. Inhibition of CaM kinase activity with CaM kinase-specific inhibitors induces apoptosis in NIH 3T3 cells (45) and sensitizes etoposide-resistant cells to apoptotic challenge (46). Thymic T cells from transgenic mice expressing a catalytically inactive form of CaMK IV showed defects in survival and activation (47).
In this study, we found that CaMK IV was proteolytically cleaved during apoptosis. The CaMK IV fragmentation was mediated by both caspase-3 and calpain. We also report that loss of Ca 2ϩ /CaM-dependent kinase activity is an early event in neuroblastoma cells undergoing apoptosis.
Cell Culture and Treatment-SH-SY5Y cells were grown on 12-well plates to confluence (roughly 2 million cells/well) at 37°C, 5% CO 2 in a humidified atmosphere with Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. At the beginning of the experiment, cultures were washed three times with serum-free Dulbecco's modified Eagle's medium. As indicated, cells were pretreated for 1 h with protease inhibitors. The cultures were then challenged with 0.5 M staurosporine or 2 M thapsigargin and maintained for indicated time, when protein was extracted. Cerebellar granule neurons and mixed cortical cells were isolated as described previously (12).
Protein Extraction and Western Blotting-Total protein was extracted by lysing cells with 2% SDS/Tris buffer, precipitating proteins with trichloroacetic acid and solubilizing with Tris base as described previously (48). Protein concentration was determined with a modified Lowry (Bio-Rad D-C protein assay kit). Equal amounts of protein were loaded on each lane and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE; 4 -20% acrylamide gradient gel; Novex) with a Tris/glycine running buffer. The separated proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (0.2 m) by semidry electrotransfer (Bio-Rad semidry transfer unit) for 2 h at 20 V. The blots were probed with primary antibody, a biotinylated secondary antibody, and avidin-conjugated alkaline phosphatase (Amersham Pharmacia Biotech). The immunoblots were developed with nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate.
Cell Extract and Purified CaMK IV Digestion in Vitro with Caspases and Calpain-Total protein was extracted from untreated confluent SH-SY5Y cells by the Triton X-100 method (12). SH-SY5Y cell extract (30 g of protein) was digested with 2.5 g of mature purified recombinant caspases (supplied by Dr. Robert Talanian, BASF) or 1 g of purifiedor m-calpain (Calbiochem) in 100 mM Hepes buffer (pH 7.4 at room temperature), 10 mM dithiothreitol, 10% (v, v) glycerol, 1 mM EGTA (for caspases), or 1 mM Ca 2ϩ (for calpain) for indicated times. The digestion was halted by the addition of SDS-containing sample buffer for PAGE. Alternatively, 1.5 g of purified, recombinant CaMK IV was subject to digestion under identical conditions. Samples were subjected to SDS-PAGE, transferred to PVDF, and probed with indicated antibody.
N-terminal Sequencing of Digested Recombinant CaMK IV-Purified, recombinant mouse CaMK IV fragments produced by digestion with purified caspase-3 or -calpain were subjected to Edman degradation to obtain N-terminal sequences (in-house at Parke-Davis). Mouse CaMK IV (20 g, a gift from Dr. Thomas Soderling) was incubated for time indicated at room temperature with purified recombinant caspase-3 (5 g) or purified -calpain (3 g). The digestion was halted with the addition of SDS-PAGE sample buffer. Samples were subjected to SDS-PAGE and transferred in a CAPS/methanol buffer to PVDF membrane using the method of Matsudaira (49). The membranes were stained with 0.1% Coomassie in 50% methanol until bands appeared. The stained bands were excised and subjected to N-terminal sequencing.
Assay for CaM-stimulated Protein Kinase Activity-SH-SY5Y cells were treated as described, and cells were collected by scraping and centrifugation. Cells washed once with cold Tris-buffered saline plus 1 mM EDTA. Cell lysates were made by resuspending the cell pellet in homogenization buffer (20 mM Tris-HCl, pH 7.5, 0.5 mM EGTA, 1 mM EDTA, 10 mM sodium pyrophosphate, 0.4 mM sodium molybdate, 2 mM dithiothreitol, 0.5% Triton X-100, and protease inhibitors) (50). Cells were kept on ice 10 min with vortexing and then sonicated for 15 s. Total CaM kinase activity in 15 g of sample was assayed by phosphorylation of the CaM kinase-selective substrate syntide-2 (Pro-Leu-Ala-Arg-Thr-Leu-Ser-Val-Ala-Gly-Leu-Pro-Gly-Lys-Lys, 40 M), in the presence of Ca 2ϩ (0.5 mM) and calmodulin (1.5 g/assay tube) with 5 Ci/ assay tube [␥-32 P]ATP (ICN). Kinase reactions were carried out for 5 min at 30°C and halted by spotting onto P-81 phosphocellulose paper (Whatman) and washing in 75 mM phosphoric acid (3 ϫ 10 min). The extent of phosphorylation was quantified by a Beckman scintillation counter.
Cell Death Measurement-SH-SY5Y cell viability was assessed by measuring release of the cytosolic enzyme, lactate dehydrogenase (LDH) into the culture medium (25 l samples). Quantification of LDH release was done using the Cytotox 96 colorimetric LDH assay kit (Promega), following the manufacturer's directions.

CaMK IV, but Not CaMK II (␣ or ␤), Is Expressed in SH-SY5Y Human Neuroblastoma Cells-To determine whether
CaMK II or CaMK IV are expressed in SH-SY5Y cells, we subjected control lysate from SH-SY5Y cells to Western blotting with antibodies to CaMK IV (Fig. 1A, top), and to the major neuronal isoforms of CaMK II: ␣ (Fig. 1A, bottom) and ␤ (Fig.  1A, middle). Rat cerebral cortical and cerebellar granule cell extracts were used as positive controls for the presence of CaMK II ␣ and ␤. In our hands, human CaMK IV appeared as a roughly 55-kDa doublet (probably ␣ and ␤ CaMK IV isoforms) with rat CaMK IV running slightly higher. The predicted molecular mass for human CaMK IV is 52 kDa (27). Although neither ␣ nor ␤ isoform of CaMK II is detectable in SH-SY5Y cells, the presence of other CaMK II isoforms, such as ␦ or ␥, or CaM kinase I cannot be disregarded.
To validate our antibodies, purified recombinant CaMK II ␣ and CaMK IV were each Western blotted with antibodies to CaMK II ␣ and CaMK IV (Fig. 1B). The CaMK IV antibody reacted with only a 55-kDa protein, whereas the CaMK II ␣ antibody reacted with only a 50-kDa protein. These molecular masses are consistent with masses for these proteins observed in situ (Fig. 1A). No cross-reactivity occurred, demonstrating the specificity of the antibodies used.
Time Course of CaMK IV, Poly(ADP-ribose) Polymerase, and ␣-Spectrin Fragmentation in Apoptotic SH-SY5Y Cells-To investigate the fate of CaMK IV in neuronal cells undergoing apoptosis, we challenged SH-SY5Y cell with 2 M thapsigargin, a well established pro-apoptotic agent (51-53). Thapsigargin increases intracellular Ca 2ϩ concentration through inhibition of the endoplasmic reticulum Ca 2ϩ /Mg 2ϩ -ATPase (54). Whole cell lysate was prepared and subjected to Western blot analysis. Treatment with thapsigargin resulted in the appearance of two major CaMK IV fragments, 40 kDa and 38 kDa, and two minor fragments, 50 kDa and 33 kDa ( Fig. 2A, top panel). The 40-kDa fragment is seen faintly in control lysate, but increases by 3 h of thapsigargin treatment. The 38-kDa band is readily apparent 3 h after start of thapsigargin treatment.
We compared the time course for CaMK IV breakdown with those of PARP ( Fig. 2A, middle panel) and ␣-spectrin ( Fig. 2A, bottom panel) in thapsigargin-treated SH-SY5Y cells. ␣-Spectrin (19,55,56) and PARP (57,58) are caspase substrates cleaved in many cell types undergoing apoptosis, including SH-SY5Y (11,12). Caspase-3 activity mediates the formation of the 120-kDa ␣-spectrin breakdown product (BDP) and the 85-kDa PARP BDP. Densitometric analysis was performed on the fragments (Fig. 2B). CaMK IV, PARP, and ␣-spectrin fragments were apparent beginning at 3 h post-treatment. The early appearance of a greater amount the 120-kDa ␣-spectrin BDP is consistent with reports that ␣-spectrin cleavage is one of the early events of apoptosis (56). The increase of the CaMK IV fragments occurs in a time frame similar to that of PARP and ␣-spectrin.
Caspase and Calpain Inhibitors Alter the CaMK IV Breakdown Pattern in SH-SY5Y Cells Undergoing Staurosporineand Thapsigargin-mediated Apoptosis-We wanted to confirm that CaMK IV would be cleaved under a different apoptotic paradigm. SH-SY5Y cells were challenged with staurosporine, a nonspecific kinase inhibitor and well established pro-apoptotic agent (10,11,57,59). Staurosporine treatment produced the same breakdown pattern as thapsigargin (Fig. 3A, second lane versus fifth lane).
We have previously demonstrated that pretreatment with the pan-caspase inhibitor Z-D-DCB blocks formation of the 120-kDa ␣-spectrin fragment (12). Here, we examine its ability to prevent CaMK IV cleavage during apoptosis. Z-D-DCB prevented formation of the 38-kDa CaMK IV fragment in both staurosporine-and thapsigargin-treated SH-SY5Y cells (Fig.  3A). Ability of Z-D-DCB to block formation of a subset of CaMK IV fragments suggests that caspase is in part, but not fully, mediating CaMK IV proteolysis.
In staurosporine-treated, but not thapsigargin-treated, cells, formation of the 50-and 33-kDa CaMK IV fragments was attenuated. This discrepancy may be a result of the different mechanisms that lead to apoptosis in staurosporine-versus thapsigargin-treated cells or differences in time to onset for apoptosis. Staurosporine activates caspase-3 much more rapidly than thapsigargin in these cells (data not shown). Furthermore, because staurosporine is a protein kinase inhibitor, it may directly interact with CaMK IV and alter susceptibility to cleavage.
Because calpain is also activated in SH-SY5Y cells undergoing apoptosis (12), we investigated its role in CaMK IV fragmentation. We inhibited calpain in two ways: by incubation with 2 mM EGTA and by treatment with CalpInh II before the onset of apoptosis. Pretreatment with EGTA inhibits formation of the 40-kDa fragment in thapsigargin-treated SH-SY5Y (Fig.  3A). In thapsigargin-and staurosporine-induced apoptosis, pretreatment with CalpInh II completely blocked formation of the 40-kDa fragment but did not affect the 38-kDa fragment. Blockade of the 33-kDa fragment was more variable; in some experiments it was attenuated by Z-D-DCB, in others by Calp-Inh II.
To further confirm the role of caspases in the proteolysis of CaMK IV in SH-SY5Y cells undergoing apoptosis, we pretreated with another commonly used caspase inhibitor, Z-VADfmk (60), and then challenged for 24 h with thapsigargin. In thapsigargin-treated cells, Z-VAD-fmk acted identically to Z-D-DCB in its ability to block formation of the 38-kDa CaMK IV BDP (Fig. 3B).
Characterization of CaMK IV Breakdown in Vitro-To examine the effect of specific caspases or calpain on CaMK IV fragmentation, cell lysate (30 g) from control SH-SY5Y cells was digested with purified recombinant caspase-3 (␤ form) or -calpain. Digestion with caspase-3 or caspase-1 produced a 38-kDa CaMK IV breakdown product (Fig. 4). Digestion with caspases in the presence of Z-D-DCB eliminated the 38-kDa fragment, whereas CalpInh II had no effect on it. -Calpain produced CaMK IV fragments of 40 kDa and several fragments of 30 -33 kDa (Fig. 4, last lane). The presence of 40-and 33-kDa fragments in the control lysate probably reflects basal activation of endogenous calpain.
We also probed the digested lysate with ␣-spectrin antibody as a positive control for caspase and calpain activity. The appearance of 150-and 120-kDa ␣-spectrin fragments is consistent with caspase-3 activation as shown previously. Digestion with -calpain produced only a 150-kDa fragment. The calpainmediated 150-kDa fragment is generated at a cleavage site, which is distinct from the caspase-mediated 150-kDa fragment (12).
Purified, Recombinant CaMK IV Is Cleaved by Calpain and Caspase-3 in Pattern Consistent with in Situ Results-Purified recombinant CaMK IV was incubated with purified caspase-3 or m-calpain as described under "Experimental Procedures" (Fig. 5A). Caspase-3 digestion produced fragments of 50 and 38 kDa, whereas calpain produced a major 40-kDa fragment in addition to minor fragments. A number of different caspases can co-exist in mammalian cells as part of a caspase cascade (61). We wanted to examine the CaMK IV cleavage pattern produced by various members of the caspase family. We used purified recombinant caspase-1 (ICE), caspase-2 (ICH-1), caspase-3 (CPP32), caspase-4 (ICH-2), and caspase-6 (Mch-2) to digest purified CaMK IV for 1 or 4 h (Fig. 5B). Only caspase-3 and caspase-1 produced the 38-kDa fragment that appears in situ during apoptosis. However, caspase-1 produced additional fragments of 47 and 45 kDa, which were not observed in situ.
Cleavage Sites of Caspase-3-and Calpain-mediated CaMK IV Fragments-To determine the caspase-3-and calpain-mediated CaMK IV cleavage sites, we performed N-terminal sequencing on the major fragments using Edman degradation as described under "Experimental Procedures." Table I summarizes the N-terminal sequence data.
Multiple sequences were detected within the major Coomassie-stained fragments. Intact 55-kDa CaMK IV has an Nterminal sequence of V 4 TVPSXPSS, suggesting that the recombinant protein is truncated upon expression in Escherichia coli (Table I). In the case of both caspase-3 and m-calpain digestion, the 38 -40-kDa bands also contained this truncated N terminus (V 4 TVPSXPSS, Table I). Because the CaMK IV antibody was made to residues 1-271, the fragments seen in Western blots (Figs. [2][3][4][5] were likely this N terminus.
In the case of calpain, we detected a 38-kDa fragment with an N-terminal sequence of TPGYXAPEIL, corresponding to a cleavage of CG 201 T 202 PGYCAPEIL, which co-migrated with the 40-kDa N-terminal fragment. A smaller fragment of about 33 kDa with an N terminus of LVPDYXIDGS was also sequenced. This fragment indicates a calpain cleavage site at EN 23 L 24 VPDYWIDGS.
On the other hand, caspase-3 produced a major fragment with an N-terminal sequence of APLKIADFXL, corresponding to cleavage at PAPD 176 A 177 PLKIADFGL of CaMK IV. The N-terminal fragment of the caspase digested CaMK IV is therefore slightly smaller than the calpain-mediated fragment, reflecting what we observed in situ (Figs. 2 and 3); a 38-kDa fragment produced by caspase and a 40-kDa fragment by calpain. A second caspase cleavage site was found in the 33-kDa fragment with an N terminus of GSNRDPLGDF, correspond- ing to a cleavage at YWID 31 G 32 SNRDPLGDF. Fig. 6 is a schematic of the CaMK IV molecule with the caspase-3 and mcalpain cleavage sites marked and fragment sizes indicated.

Ca 2ϩ /CaM-stimulated Protein Kinase Activity Reduction Is an Early Event of Thapsigargin-mediated Apoptosis in SH-
SY5Y Cells-We investigated whether CaM kinase IV proteolysis could be correlated with physiological changes in cells undergoing apoptosis. We treated SH-SY5Y cells for various times with thapsigargin and monitored CaM-stimulated protein kinase activity in the cell homogenate using the CaM kinase-selective substrate syntide-2. This peptide has been used to measure CaMK IV activity in vitro (43,62,63). Syntide-2 is homologous to phosphorylation site 2 of glycogen synthase (64). In order to prevent changes in the phosphorylation state and further degradation of the CaM kinase, the homogenization buffer contained protein phosphatase inhibitors and protease inhibitors. Because syntide-2 phosphorylation may reflect nonspecific kinase activity, we determined Ca 2ϩ /CaMstimulated protein kinase activity by calculating the difference between phosphorylation in the presence or absence of Ca 2ϩ / CaM. Additionally, although we were unable to detect CaMK II ␣ or ␤ in SH-SY5Y cells (Fig. 1), the influence on syntide-2 phosphorylation of other CaM kinases that may be expressed in these cells, such as CaMK II␥, could not be discounted.
A 30% loss in CaMK activity in SH-SY5Y cells treated with thapsigargin was apparent at 6 h and increased through 24 h, when 55% of activity was lost (Fig. 7A). Importantly, the loss in CaMK activity preceded loss of cell viability as measured by LDH leaked into the culture medium (Fig. 7A).
Next, we investigated the effects of Z-D-DCB and CalpInh II on CaM kinase activity in thapsigargin-treated SH-SY5Y cells. We chose 6 h of thapsigargin treatment because at that time there is a significant decrease in CaM kinase activity with little increase in LDH release. Loss of CaM-stimulated protein kinase activity at 6 h is not attributable to loss in cell viability. Additionally, by 6 h, although CaMK IV fragmentation is evident, there is very little PARP breakdown. Pretreatment with the pan-caspase inhibitor Z-D-DCB for 1 h prevented the thapsigargin-mediated loss of CaM-stimulated protein kinase activity at 6 h (Fig. 7B). However, pretreatment with CalpInh II did not reverse the loss of activity. Neither Z-D-DCB nor CalpInh II alone affected CaM kinase activity.
CaM Kinase II␣ and CaM Kinase Kinase Are Also Caspase Substrates in Neuronal Cells Undergoing Apoptosis-Because the percent loss in CaM kinase activity is greater than the loss of intact CaMK IV, we suspected that other components of the CaM kinase signal transduction pathway are affected by apoptosis. We examined the fate of the upstream activator of CaMK IV, CaM kinase kinase (CaMKK), in SH-SY5Y cells undergoing apoptosis (Fig. 8A). Western blot analysis of CaMKK from control and staurosporine-treated SH-SY5Y cells revealed a caspase-dependent digestion of CaMKK to a 58-kDa fragment. The fragmentation was blocked by pretreatment with Z-D-DCB, but not CalpInh II.
Because we could not discount the possibility that CaM kinases that we were unable to detect may be expressed in SH-SY5Y, we examined the vulnerability of CaMK II␣ to fragmentation in neurons undergoing apoptosis (Fig. 8B). Rat mixed cortical neurons were treated for 24 h with staurosporine in the presence or absence of Z-D-DCB or CalpInh II. A 35-kDa CaMK II␣ BDP was observed that was inhibited by Z-D-DCB but not CalpInh II.

DISCUSSION
This is the first report describing caspase-3-and calpainmediated CaMK IV degradation in apoptotic cells. We identify both caspase-and calpain-mediated CaMK IV cleavage sites. This is also the first report demonstrating a loss in Ca 2ϩ / calmodulin-dependent kinase activity in cells undergoing apoptosis.
Mounting evidence has clearly shown that caspase-3 plays a key role in mammalian apoptosis. Caspase-3 is activated by a wide range of apoptotic challenges in a variety of cell types (11,18,65). We have previously demonstrated its activation in apoptotic SH-SY5Y (11,12). An increasing number of caspase-3 substrates have now been identified. These substrates all share the DXXD consensus site preferred by caspase-3. However, Talanian et al. (66) recently reported that caspase-3 and the closely related caspase-7 (Mch-3) would also accept other less preferred residues in the P 4 position in peptidic substrates. The sites of caspase-3-mediated CaMK IV cleavage have an Asp in the P 1 but not in the P 4 position. CaMK IV cleavage is not as efficient as PARP and ␣-spectrin cleavage ( Fig. 2A), which may be a result of the absence of Asp in the P 4 position of the caspase-3 cleavage sites. Caspase-3-mediated fragmentation of CaMK IV is an early event in SH-SY5Y cells undergoing apoptosis. In thapsigargin-treated SH-SY5Y cells, the 38-kDa CaMK IV BDP (cleavage site PAPD 176 APLK) appears in a time frame consistent with caspase-3-mediated PARP and ␣-spectrin fragmentation (Fig. 2).
We found that CaM-stimulated protein kinase activity decreased over time in thapsigargin-treated SH-SY5Y cells (Fig.  7A). CaMK II activity has recently been shown to decrease in cultured neurons exposed to excitotoxic insult (67). In addition, inhibition of CaM kinase activity with CaM kinase-specific inhibitors induces apoptosis in fibroblasts (45) and sensitizes etoposide-resistant cells to apoptotic challenge (46). To address whether fragmentation of CaMK IV by caspase-3 and calpain has physiological consequences in the apoptosis cascade, we considered expressing a caspase/calpain cleavage-resistant CaMK IV mutant. Owing to the presence of multiple cleavages by both proteases, this strategy would not be feasible.
One reason for the loss of CaM-stimulated protein kinase activity in apoptotic SH-SY5Y cells may be proteolytic fragmentation. CaMK IV was cleaved within the catalytic domain (Fig. 6). The possibility also exists that other CaM kinases are expressed in SH-SY5Y cells. Tombes et al. (45) have presented data that CaMK II␥ may be involved in apoptosis in NIH 3T3 cells. We propose that multiple CaM kinases are FIG. 5. Purified recombinant CaMK IV digested with caspase-3 and m-calpain produces cleavage patterns consistent with in situ breakdown. A, purified, recombinant CaMK IV (5 g) was either untreated (Con) or digested for 1 h with 1.5 g of caspase-3 or for 4 min with 1 g of m-calpain. B, purified recombinant CaMK IV was untreated (Con) or digested with 1.5 g of caspase-3, caspase-1, caspase-4, or caspase-6 for 1 or 4 h. Digests were subjected to SDS-PAGE and probed with antibody against CaMK IV.
susceptible to proteolytic fragmentation in neurons undergoing apoptosis. In this report, we show that CaM kinase II␣, which has significant sequence homology to CaMK IV within the catalytic domain, is also cleaved by caspases in apoptotic cortical neurons (Fig. 8A). Thus, the loss of activity of other CaM kinases may contribute to the loss of total CaM-stimulated protein kinase activity.
Because most of the parent CaMK IV remained intact, even after 24 h of thapsigargin treatment, the 60% loss in CaMstimulated protein kinase activity (Fig. 7A) is not likely to be caused by CaMK IV fragmentation alone. The activity decrease could also be attributed to events occurring upstream from CaMK IV in the CaM kinase cascade. CaMK IV activity is enhanced by the Ca 2ϩ /calmodulin-dependent CaM kinase kinase (CaMKK) (68). We report here that CaMKK is cleaved by caspase(s) in apoptotic SH-SY5Y cells (Fig. 8B). The caspasemediated CaMKK cleavage may further contribute to the loss of CaMK IV activity. This is a subject for future investigation. CaMK IV activity is also altered by autophosphorylation on Ser 12 -Ser 13 at the N terminus (69). Changes in phosphorylation state as a result of apoptotic challenge may also lead to decreases in CaMK activity.
Caspase inhibition eliminated the loss of CaM-stimulated protein kinase activity, whereas calpain inhibition had no effect (Fig. 7B). Thus, the caspase-mediated fragmentation appears to be predominant over the calpain pathway in neuronal cells undergoing apoptosis, as we described previously with staurosporine-mediated apoptosis in SH-SY5Y and lowpotassium-mediated apoptosis in rat cerebellar granule neu-rons (12). The attenuation of CaMKK cleavage by caspase inhibition, but not calpain inhibition (Fig. 8A), may explain why caspase inhibition alone is sufficient to prevent loss of  CaM-stimulated protein kinase activity (Fig. 7B). Further investigation of the activity changes in purified CaMK IV and CaMKK in response to digestion by caspases and calpain are planned.
CaMK IV is involved in gene transcription factor phosphorylation (34,35,63) leading to the expression of immediate early genes. In addition, CaMK IV has recently been reported to inhibit type I adenylyl cyclase, a Ca 2ϩ /CaM-dependent neurospecific enzyme (70) and activate transforming growth factor ␤, which is involved in cell cycle regulation (38). The degradation of CaMK IV and loss of activity in apoptotic neurons may be functionally related to the morphological and nuclear changes accompanying apoptosis. Although its physiological role is not well defined, CaMK IV expression is restricted primarily to the brain and thymus. CaMK IV may regulate important neuronal processes that are altered as the neuron undergoes apoptosis.