Requirements for calcium and calmodulin in the calmodulin kinase activation cascade.

We have previously purified and cloned rat brain Ca/calmodulin-dependent protein kinase kinase (CaM-KK), and the 68-kDa recombinant CaM-KK activates in vitro both CaM-kinase IV (CaM-K IV) and CaM-K I (Tokumitsu, H., Enslen, H., and Soderling, T. R.(1995) J. Biol. Chem. 270, 19320-19324). In the present study we have determined that activation of CaM-K IV through phosphorylation of Thr by CaM-KK is triggered by elevated intracellular Ca in intact cells and requires binding of Ca/CaM to both enzymes. An expressed fragment of CaM-K IV (CaM-K IV), which contains the activating phosphorylation site (Thr) but not the autoinhibitory domain or the CaM-binding domain, still required Ca/CaM for phosphorylation by wild-type CaM-KK. A truncated mutant of CaM-KK (CaM-KK) phosphorylated CaM-K IV in a Ca/CaM-independent manner, but this constitutively active CaM-KK required Ca/CaM for phosphorylation and activation of wild-type CaM-K IV. These results demonstrate that binding of Ca/CaM to both CaM-K IV and CaM-KK is required for the CaM-kinase cascade. Both CaM-KK and CaM-K IV appear to have similar Ca/CaM requirements with EC values of approximately 100 nM. Studies using co-expression of CaM-K IV with CaM-KK in COS-7 cells demonstrated that CaM-KK rapidly activated both total and Ca/CaM-independent activities of wild-type CaM-K IV, but not the Thr Ala mutant, upon ionomycin stimulation.

GTC ACG-3Ј and 5Ј-CCT CCT CTT GCA-3Ј as primers, was digested with NheI and BstBI and subcloned into pRSETB vector (Invitrogen). After digesting this plasmid with NdeI and filling in, a 0.32-kilobase pair His-tagged (Met-Arg-Gly-Ser-His 6 -Gly-Ala-Ser) fragment was obtained by digestion with BstBI and subcloning into pME18s plasmid (DNAX Research Institute) with the 3Ј-portion of CaM-K IV (BstBI-XbaI fragment) (13). Thr 196 3 Ala mutants were made by using a site-specific plasmid DNA mutagenesis kit (5 Prime 3 3 Prime, Inc.) and pME-His-CaM-K IV as a target plasmid with a mutagenic oligonucleotide (5Ј-AAG TGC TCA TGA AGG CAG TGT GTG GAA CCC CGG G-3Ј). Mutant plasmids were selected by nucleotide sequencing. For expression of CaM-KK in E. coli we made a HIS-tagged CaM-KK construct (pRSET-CaM-KK). The 5Ј portion of CaM-KK was made by PCR to create a unique NheI site using 5Ј-ACT GCT AGC ATG GAG CGC AGT CCA-3Ј and 5Ј-TTC AGG ATC AGG TCT TT-3Ј. The NheI-XbaI fragment was subcloned into NheI/NcoI sites of pRSETB vector with the 3Ј portion of CaM-KK (XbaI-NcoI fragment). Truncated His-tagged CaM-KK construct (pRSET-CaM-KK  ) was made by replacement of BstXI-EcoRI fragment of pRSET-CaM-KK and BstXI-NotI fragment of pME-CaM-KK 1-434 construct which was made by PCR using 5Ј-TTG GCG CCG CTC ACA CCT CCT CCT CAG TCA CCT CT-3Ј and 5Ј-AGG AAA GAC CAG CGG AAA-3Ј and subcloning of the PCR fragment (XbaI/NotI digest) into XbaI/NotI digested pME-CaM-KK (wild type) (18). For expression of CaM-K IV 178 -246 and CaM-K IV 1-60 , these fragments were amplified by PCR using combination of 5Ј-AAG GAT CCG AAA ATT GCT GAT TTT GGA CTT3Ј and 5Ј-GTA AAG CTT TCA TCC TAC AGA CCA CAT GTC-3Ј, and 5Ј-AAG GAT CCG ATG CTC AAA GTC ACG GTG CCC-3Ј and 5Ј-GGT AAG CTT TCA TTT GCA TCT GTA CAC AAT-3Ј, respectively, as PCR primers followed by digestion with BamHI/HindIII and subcloning into BamHI/HindIII sites of pRSETB vector. Mutant fragments were made by site-directed mutagenesis as described above and selected by sequencing.
Transient Expression and Stimulation of COS-7 Cells-COS-7 cells were maintained in Dulbecco's modified Eagle medium containing 10% fetal calf serum. Cells were subcultured in 6-cm dishes for 12 h before transfection. The cells were then transferred to serum-free medium and treated with a mixture of either pME18s plasmid DNA (4 g, mock) or plasmid containing His-tagged CaM-K IV cDNA (4 g, pME-His-CaM-K IV), either with or without 0.8 g of pME-CaM-KK and 30 g of LipofectAMINE reagent (Life Technologies, Inc.) in 2.4 ml of serum-free medium. After 5 h of incubation, 2.4 ml of medium containing 20% fetal calf serum were added to the cells, and the incubation was continued for another 20 h. After changing to fresh medium, cells were cultured an additional 20 h. Prior to stimulation by ionomycin, the cells were cultured for 2 h in serum-free medium (Dulbecco's modified Eagle's medium, Life Technologies, Inc.). At the indicated times (1-10 min), stimulation was terminated by aspirating the medium, and the cells were immediately frozen with liquid N 2 . The COS cells were lysed for 30 min at 4°C with 1 ml of lysis buffer A (150 mM NaCl, 20 mM Tris-HCl, pH 7.5, 10% glycerol, 1% Nonidet P-40, 10 mM pyrophosphate, 50 mM NaF, 1 mM sodium vanadate, 1 M microcystin-LR, 0.5 mM PMSF, 10 mg/liter trypsin inhibitor, 10 mg/liter leupeptin, 10 mg/liter antipain, 10 mg/liter aprotinin and 1 mM benzamidine). After centrifugation at 10,000 ϫ g for 20 min, 20 l of Ni 2ϩ resin (50% slurry, ProBond TM resin (Invitrogen)) were added to the supernatant, and the mixture was incubated for 1 h at 4°C. After precipitation of the resin by centrifugation and removal of the supernatant, the resin was washed three times with 1 ml of lysis buffer and once with buffer A containing 0.3 M NaCl. The resin was transferred into 0.45-m filter unit (ULTRAFREE-MC, Millipore Corp.), and CaM-K IV was eluted with 50 l of 20 mM Tris-HCl (pH 7.5), 0.15 M NaCl, 250 mM imidazole, 1 mM EDTA, and 1 mM EGTA. Five l and 15 l of eluted CaM-K IV were used for the protein kinase assay as described below and for the immunoblotting, respectively.
Expression and Purification of Recombinant CaM-KKs and CaM-K IV Fragments-Escherichia coli (JM109) transformed with pRSET-CaM-KK (wild type), -CaM-KK 1-434 , -CaM-K IV 178 -246 , or -CaM-K IV 1-60 were grown in ampicillin-containing SOB medium (200 ml) to an OD 600 ϭ 0.3. After adding 1 mM isopropyl-1-thio-␤-Dgalactopyranoside, the cells were grown for an additional 1 h, infected with M13/T7 phage at an multiplicity of infection of 5 plaque-forming units/cell, and cultured for 5 h. Cells were collected by centrifugation at 10,000 ϫ g for 10 min and frozen in liquid N 2 . The cell pellet was resuspended in 15 ml of lysis buffer A as described above, sonicated, and centrifuged at 10,000 ϫ g for 15 min. Two ml of Ni 2ϩ resin were added to the supernatant, and the mixture was incubated for 1 h on ice and then applied to a Poly-Prep chromatography column (Bio-Rad). After washing the column extensively with lysis buffer A and buffer with 0.3 M NaCl, elution was carried out with 50 mM Tris-HCl (pH 7.5), 250 mM imidazole, 1 mM EGTA, 1 mM EDTA, 0.1 mM PMSF, and 1 mM benzamidine. After dialysis against 50 mM Tris-HCl (pH 7.5), 0.2 mM EGTA, 0.2 mM EDTA, 0.2 mM PMSF, eluates were applied to Q-Sepharose (0.5 ml) and eluted with 0.1 M NaCl containing buffer. Recombinant CaM-K IV 178 -246 and CaM-K IV 1-60 was stored frozen at Ϫ20°C. Recombinant wild-type CaM-KK was further purified by CaM-Sepharose column chromatography as previous described (13). The truncation mutant CaM-KK  was further purified by Q-Sepharose chromatography as follows. The fractions eluted from the Ni 2ϩ column were dialyzed against buffer B (50 mM Tris-HCl (pH 7.5), 1 mM EGTA, 1 mM EDTA, 0.1 mM PMSF, 1 mM benzamidine) and applied to Q-Sepharose (1-ml column) which was equilibrated with buffer B. After washing the column with buffer B, elution was carried out by a linear NaCl gradient (0 -0.6 M NaCl). (36 nM) at 30°C in 50 mM HEPES (pH 7.5), 10 mM magnesium acetate, 1 mM DTT, 400 M ATP, and either 1 mM CaCl 2 , 10 M CaM or 1 mM EGTA. The reaction was terminated at 5 min by a 25-fold dilution at 5°C with 50 mM HEPES (pH 7.5), 2 mg/ml bovine serum albumin, 10% ethylene glycol, and 1 mM EDTA. CaM-K IV activity was measured at 30°C in a 25-l assay containing 50 mM HEPES (pH 7.5), 10 mM magnesium acetate, 1 mM DTT, 1 M microcystin-LR, 400 M [␥-32 P]ATP (1000 -2000 cpm/pmol), 40 M syntide-2, 43 nM recombinant CaM kinase IV (or 5 l of Ni 2ϩ resin purified CaM-K IV), and either 1 mM CaCl 2 , 1 M CaM (total activity) or 1 mM EGTA (Ca 2ϩ /CaM-independent activity). The reaction was initiated by the addition of CaM-K IV and terminated at 5 min by spotting aliquots (15 l) onto phosphocellulose paper (Whatman P-81) followed by washing in 75 mM phosphoric acid (25).
In Vitro Phosphorylation of CaM-K IV Fragments-Either recombinant CaM-K IV-wild type, CaM-K IV 178 -246 , or CaM-K IV 1-60 was incubated at 30°C for 15 min with either 36 nM of recombinant CaM-KK or 22 nM of recombinant CaM-K IV in a solution containing 50 mM HEPES (pH 7.5), 10 mM magnesium acetate, 1 mM DTT, 100 M [␥-32 P]ATP (400 M for phosphorylation of recombinant CaM-K IV-wild type) and either 1 mM CaCl 2 , 1 M CaM (10 M CaM for phosphorylation of recombinant wild-type CaM-K IV) or 1 mM EGTA. Reactions were terminated by adding 2.5 l of SDS-PAGE sample buffer, and the samples were immediately subjected to SDS-PAGE. After staining and destaining, the gels were exposed to autoradiography. The bands corresponding to the fragment were excised and 32 P incorporation was measured.
Other Methods-Anti-human CaM-K IV rabbit IgG (26) was raised against the COOH-terminal sequence ( 436 GLAEEKLTVEEA 447 ) of human CaM-K IV (27). SDS-PAGE was carried out with either Trisglycine (28) or the Tricine system (29). Protein concentrations were measured by the Bradford (30) method using bovine serum albumin as a standard. The protein concentration of partially purified truncated CaM-KK 1-434 was adjusted by immunoreactivity against anti-CaM-K II antibody (18) using purified recombinant CaM-KK as a standard.

CaM-KK Is a Ca 2ϩ /CaM-dependent Enzyme That Phosphorylates Thr 196 in CaM-K IV-
We previously demonstrated the activation of CaM-K IV and CaM-K I by an extract of COS-7 cells expressing CaM-KK (18). To further characterize CaM-KK, we have expressed it in E. coli and purified it on CaM-Sepharose. The recombinant CaM-KK, like the brain CaM-KK (13), bound to CaM-Sepharose in the presence of Ca 2ϩ and was eluted in the presence of EGTA (not shown) to yield a highly purified 68-kDa protein (Fig. 1A). Incubation of the purified CaM-KK with recombinant CaM-K IV in the presence of Mg 2ϩ / ATP enhanced by 10-fold or more both total and Ca 2ϩ -independent activities of CaM-K IV (Fig. 1B). This activation of CaM-K IV by CaM-KK was totally dependent on the presence of Ca 2ϩ /CaM.
In order to determine whether the required Ca 2ϩ /CaM was binding to CaM-KK, CaM-K IV, or both, we needed a substrate of CaM-KK which itself does not bind Ca 2ϩ /CaM. We constructed a poly(His) 6 -tagged fragment of CaM-K IV (residues 178 -246) which contains the activating phosphorylation site (i.e. Thr 196 ) (23) but not the AID or CaM-binding domain (residues 304 -328). The His/CaM-K IV 178 -246 was expressed in E. coli and purified for in vitro phosphorylation by CaM-KK. Using this fragment of CaM-K IV eliminates any ambiguity due to autophosphorylation by activated CaM-K IV. As shown in Fig. 2A (lower panel), His/ CaM-K IV 178 -246 was phosphorylated by recombinant CaM-KK in a completely Ca 2ϩ /CaM-dependent manner. Phosphoamino acid analysis (not shown) showed that this phosphorylation was exclusively on Thr, and mutation of Thr 196 to Ala abolished approximately 90% of 32 P incorporation into His/CaM-K IV 178 -246 by CaM-KK ( Fig. 2A). These results demonstrate that recombinant CaM-KK is Ca 2ϩ /CaM-dependent and directly phosphorylates Thr 196 of CaM-K IV. The His/CaM-K IV 178 -246 was not phosphorylated by CaM-K IV itself (data not shown).
It has also been reported that phosphorylation of Ser residues in the NH 2 terminus of CaM-K IV by purified rat brain CaM-KK is responsible for activation of CaM-K IV (31). We tested the ability of CaM-KK to phosphorylate this domain by expressing a His/CaM-K IV 1-60 construct. However, His/ CaM-K IV 1-60 was not significantly phosphorylated by CaM-KK, but it was phosphorylated by CaM-K IV, especially after its activation by CaM-KK (Fig. 2B). These results suggest that the observed phosphorylation of the NH 2 -terminal Ser residues in CaM-K IV may have been due to autophosphorylation subsequent to activation of CaM-K IV by CaM-KK. ( Fig. 3A, lower panel). However, when wild-type CaM-K IV was the substrate, its phosphorylation (Fig. 3B, insert) and activation (Fig. 3B) by CaM-KK 1-434 required Ca 2ϩ /CaM. The above results establish that the CaM-kinase cascade requires binding of Ca 2ϩ /CaM to both CaM-KK and CaM-K IV, and they also show that CaM-KK has an AID and CaM-binding domain COOH-terminal of residue 434.
Since binding of Ca 2ϩ /CaM to both CaM-KK and CaM-K IV was required for the CaM-kinase cascade to operate, we tested whether CaM-KK and CaM-K IV had different requirements for Ca 2ϩ /CaM. If CaM-KK needed much higher concentrations of Ca 2ϩ /CaM than CaM-K IV, it was possible that small elevations of intracellular Ca 2ϩ might selectively activate CaM-K IV, and higher elevations of Ca 2ϩ would be required to trigger the CaM-kinase cascade. However, it appears that the Ca 2ϩ / CaM requirement (EC 50 ) of both CaM-KK and CaM-K IV is approximately 100 nM (Fig. 4). (12,13,23). To directly demonstrate that CaM-KK can activate CaM-K IV in intact cells, we transfected His-tagged CaM-K IV with or without co-transfected CaM-KK into COS-7 cells. At various times after ionomycin stimulation to allow Ca 2ϩ influx, cells were frozen in liquid N 2 , lysed, and His-tagged CaM-K IV was purified from the cell lysate by a Ni 2ϩ resin. As shown in Fig. 5A, both total and Ca 2ϩ -independent activities of CaM-K IV were elevated upon ionomycin stimulation when CaM-KK was co-transfected (circles). In the absence of CaM-KK co-transfection, ionomycin treatment did not alter CaM-K IV activities. The increases in total (3-fold) and Ca 2ϩ -independent (6-fold) activities by CaM-KK and ionomycin stimulation occurred within 4 min but returned to near basal values within 10 min. This biphasic activation phenomena was not due to changes of CaM-K IV expression levels (Fig. 5B).

Ca 2ϩ -dependent Activation of CaM-K IV by CaM-KK in COS-7 Cells-It is well established that CaM-K IV can be activated by CaM-KK in vitro
It was recently reported that mutation of Thr 196 to Ala in CaM-K IV blocks the increase in its total activity normally generated in vitro by purified procine brain CaM-K I kinase (23). The data of Fig. 2 confirm that Thr 196 in HIS/CaM-K IV 178 -246 was phosphorylated by CaM-KK. To test whether Thr 196 is also the phosphorylation/activation site in intact cells, we introduced His-tagged CaM-K IV T196A mutant into COS-7 cells with CaM-KK. Fig. 6 shows that the Thr 196 3 Ala mutant had basal Ca 2ϩ /CaM-dependent activity, but it was not activated by CaM-KK in response to ionomycin treatment.
Mutation of Thr 196 to Ala not only prevents the increase in total activity of CaM-K IV generated by CaM-KK (23), but it also blocked the increase in Ca 2ϩ -independent activity (Fig. 6). It is possible that phosphorylation of Thr 196 is directly responsible for increasing both total and Ca 2ϩ -independent activities. Alternatively, phosphorylation of Thr 196 may be a prerequisite for phosphorylation of some other site, either by CaM-KK or through autophosphorylation by activated CaM-K IV, which generates Ca 2ϩ -independent activity. Since Thr 308 of CaM-K IV is analogous to the phosphorylation site (Thr 286 ) in CaM-K II which generates Ca 2ϩ -independent activity, Thr 308 was considered a likely candidate. Furthermore, mutation of HMDT 308 to DEDD converts CaM-K IV into a Ca 2ϩ -independent species (13). We therefore made CaM-K IV mutants Thr 308 3 Ala and Ser 337 3 Ala and examined their in vitro activation by CaM-KK. Both mutant CaM-K IV species showed normal increases in both total and Ca 2ϩ -independent activities upon phosphorylation by CaM-KK (data not shown). DISCUSSION Extensive in vitro studies have documented the phosphorylation and activation of CaM-K IV by CaM-KK (12,13,23). This report extends these studies by 1) establishing a requirement for binding of Ca 2ϩ /CaM to both the CaM-KK and the substrate CaM-K IV and 2) demonstrating in COS-7 cells with transfected CaM-KK the Ca 2ϩ -dependent activation of co-transfected CaM-K IV through phosphorylation of Thr 196 .
The requirement for binding of Ca 2ϩ /CaM to CaM-KK is consistent with the previous observation that CaM-KK can be purified on CaM-Sepharose (13) and that the expressed CaM-KK binds Ca 2ϩ /CaM using the gel overlay technique (18). However, since both CaM-KK and CaM-K IV are apparently CaM-dependent, it was not clear whether the requirement for Ca 2ϩ /CaM in the CaM-kinase cascade was due to binding of Ca 2ϩ /CaM to either CaM-KK, CaM-K IV, or both. By constructing a His-tagged construct containing the phosphorylation site in CaM-K IV (Thr 196 ) but lacking the AID and CaM-binding domain (residues 304 -328), we demonstrated that phosphorylation of Thr 196 in this CaM-independent substrate by wild-type CaM-KK still required Ca 2ϩ /CaM (Fig. 2). This indicates a requirement for Ca 2ϩ /CaM in this reaction for the activation of CaM-KK. This conclusion was substantiated by using a truncated form of CaM-KK, CaM-KK 1-434 , lacking the putative AID and CaM-binding domain. This CaM-KK 1-434 phosphorylated the His/CaM-K IV 178 -246 in a Ca 2ϩ /CaM-independent manner (Fig. 3), confirming the requirement for Ca 2ϩ /CaM of wild-type CaM-KK. Furthermore, the requirement for Ca 2ϩ /CaM in the phosphorylation of wild-type CaM-K IV by CaM-KK 1-434 confirmed the requirement for Ca 2ϩ /CaM-binding to the substrate CaM-K IV. Thus, binding of Ca 2ϩ /CaM to both CaM-KK and CaM-K IV is required for this kinase cascade. The binding of Ca 2ϩ /CaM to CaM-KK is presumably required for neutralization of an AID COOH-terminal of residue 434. Studies are currently in progress to further define the AID and CaMbinding domain in CaM-KK. Extensive studies have established the existence of adjacent and sometimes overlapping AIDs and CaM-binding domains in numerous other kinases activated by Ca 2ϩ /CaM (32). Presumably Ca 2ϩ /CaM must bind to CaM-K IV and remove its AID to expose the activation loop Thr 196 which is probably within the catalytic cleft.
Our studies confirm with recombinant CaM-KK the observation made with purified porcine brain CaM-K I kinase that mutation of Thr 196 to Ala blocks the increase in total activity of CaM-K IV (23). We also observed that the increase in Ca 2ϩindependent activity was also absent in this mutant (Fig. 6). It is interesting that this single site mutation blocks the increases in both total and Ca 2ϩ -independent activities of CaM-K IV, whereas phosphorylation of the equivalent site (Thr 177 ) in CaM-K I produces only an increase in total activity (34). Both of these phosphorylation sites are within the "activation loops" that require phosphorylation for activation of numerous protein kinases. One possibility is that the AID in CaM-kinase IV lies within the catalytic cleft and makes an inhibitory interaction with the activation loop. This could explain why removal of the AID through binding of Ca 2ϩ /CaM is required to expose Thr 196 for phosphorylation by CaM-KK. Phosphorylation of Thr 196 could prevent the inhibitory interaction of the activation loop with the AID in the absence of Ca 2ϩ /CaM, thereby generating Ca 2ϩ -independent activity. Another possibility is that phosphorylation of Thr 196 in CaM-K IV generates elevated total kinase activity which then allows autophosphorylation on another site to account for the elevated Ca 2ϩ -independent activity. If such were the case, a phosphorylation site in the AID would seem likely since introduction of negative charge in the AID generates Ca 2ϩ -independent activity of CaM-K IV (13). To test this possibility we mutated Thr 308 , which is equivalent to the autophosphorylation site in CaM-K II (Thr 286 ) that generates Ca 2ϩ -independent activity. However, CaM-KK increased both total and Ca 2ϩ -independent activities of the Thr 308 3 Ala mutant. Similar results were obtained with the Ser 337 3 Ala mutant. Thus, it is still not clear whether phosphorylation of Thr 196 alone is sufficient for increasing Ca 2ϩ -independent activity of CaM-K IV.
A major purpose of this study was to test whether this CaM-kinase cascade, which has been demonstrated in vitro, could also be observed in intact cells. When COS-7 cells were transfected with His-tagged CaM-K IV alone, there was no effect of Ca 2ϩ -mobilization through ionomycin treatment on the His/CaM-K IV activity subsequently assayed in vitro. However, when CaM-KK was co-transfected with His/CaM-K IV, then ionomycin-treatment resulted in a 3-6-fold increases in total and Ca 2ϩ -independent His/CaM-kinase IV activities (Fig.  5A). These changes in CaM-K IV activities were not due to changes in amounts of expressed CaM-K IV (Fig. 5B). Interestingly, the activation of CaM-K IV by ionomycin was transient, peaking at 4 min and returning to near basal values at 10 min. This biphasic nature is consistent with phosphorylation of CaM-K IV followed by dephosphorylation. That phosphorylation was required for the activation was demonstrated by the fact that the Thr 196 3 Ala mutant of CaM-K IV was not activated by CaM-KK upon ionomycin treatment (Fig. 6). The fact that only a 3-4-fold activation of CaM-K IV was observed in the intact cells compared to the 10-fold or greater activation in vitro is probably due to the endogenous protein phosphatases in the COS-7 cells that limit both the extent and duration of activation. These results are similar to the activation of CaM-K IV in Jurkat cells upon stimulation of the CD3 receptor (33). The CD3-mediated activation of CaM-K IV is due to its phosphorylation since this activation can be reversed in vitro by treatment with protein phosphatase 2A but not protein phosphatase 1 (26), the same specificity as for reversal of in vitro activation of CaM-K IV by CaM-KK. Furthermore, CD3mediated activation of CaM-K IV is transient, presumably due to endogenous protein phosphatase 2C which can inactivated CaM-K IV (33), and the in situ activation can be further augmented 2-3-fold by subsequent in vitro treatment with purified CaM-KK (26). From these studies we concluded that the CD3dependent activation of CaM-K IV in Jurkat cells is probably mediated by CaM-KK.
It is puzzling why adjacent steps in a kinase cascade should both require the same activator, i.e. Ca 2ϩ /CaM. Our initial thought was that perhaps the CaM-KK would have a much lower affinity for activation by Ca 2ϩ /CaM. Thus, low levels of elevated intracellular Ca 2ϩ might selectively activate CaM-K IV and CaM-K I, whereas much higher level of Ca 2ϩ would be required for activation of CaM-KK to initiate the cascade. However, the data of Fig. 4 show that both kinases have very similar requirements for Ca 2ϩ /CaM. Of course, the effect of activation of different substrates of CaM-KK may confer differences in their Ca 2ϩ dependencies. CaM-K IV, which has been activated by CaM-KK, can maintain sustained activity in the absence of continued Ca 2ϩ because of its considerable Ca 2ϩindependent activity. This is not true for CaM-K I which does not generate Ca 2ϩ -independent activity upon activation by CaM-KK. While this manuscript was under review, a paper appeared which shows a requirement for binding of Ca 2ϩ /CaM to both CaM-KI and CaM-KI kinase and for binding of AMP to both AMP-kinase and AMP-kinase kinase in those cascades (35). Lastly, it is possible there may be unidentified CaM-KK substrates which themselves are not regulated by Ca 2ϩ /CaM. We are currently searching for additional physiological pathways which may be regulated by CaM-KK.