In situ Ca2+ dependence for activation of Ca2+/calmodulin-dependent protein kinase II in vascular smooth muscle cells.

Activation of Ca2+/calmodulin (CaM)-dependent protein kinase II (CaM kinase II) and development of the Ca2+/CaM-independent (autonomous) form of the kinase was investigated in cultured vascular smooth muscle (VSM) cells. Within 15 s of ionomycin (1 μM) exposure 52.7 ± 4.4% of the kinase became autonomous, a response that was partially maintained for at least 10 min. This correlated with 32P phosphorylation of CaM kinase II δ-subunits in situ and was abolished by pretreatment with the CaM kinase II inhibitor KN-93. The in situ Ca2+ dependence for generating autonomous CaM kinase II was determined in cells selectively permeabilized to Ca2+ and depleted of sarcoplasmic reticulum Ca2+ by pretreatment with thapsigargin. Analysis of the resulting curve revealed an EC50 (concentration producing 50% of maximal response) of 692 ± 28 nM [Ca2+]i, a maximum of 68 ± 2% of the total activity becoming autonomous reflecting nearly complete activation of CaM kinase II and a Hill slope of 3, indicating a highly cooperative process. Based on this dependence and measured [Ca2+]i responses in intact cells, increases in autonomous activity stimulated by angiotensin II, vasopressin and platelet-derived growth factor-BB (4.6-, 2-, and 1.7-fold, respectively) were unexpectedly high. In intact cells stimulated by ionomycin, the correlation between autonomous activity and [Ca2+]i resulted in a parallel curve with an EC50 of 304 ± 23 nM [Ca2+]i. This apparent increase in Ca2+ sensitivity for generating autonomous activity in intact VSM cells was eliminated by thapsigargin pretreatment. We conclude that alteration of [Ca2+]i over a physiological range activates CaM kinase II in VSM and that this process is facilitated by release of Ca2+ from intracellular pools which initiates cooperative autophosphorylation and consequent generation of autonomous CaM kinase II activity.

Mobilization of intracellular free calcium ([Ca 2ϩ ] i ) 1 following agonist-stimulated phospholipid hydrolysis is a major signal transduction pathway in vascular smooth muscle (VSM) cells (1). [Ca 2ϩ ] i controls contractile activity in smooth muscle (2) as well as signaling pathways related to cell growth and differen-tiation (3,4) and replication (5,6). Calcium-mediated activation of a number of cellular enzymes requires the ubiquitous calcium-binding protein, calmodulin (CaM) (7). One such enzyme is Ca 2ϩ /CaM-dependent protein kinase II (CaM kinase II), which is expressed in high amounts in the brain and in lesser amounts in most peripheral tissues (8,9). CaM kinase II has been implicated in diverse Ca 2ϩ -mediated processes, including neurotransmitter production and release (10,11), regulation of smooth muscle myosin light chain kinase activity (12), VSM cell migration (13), oocyte fertilization (14), and gene expression (15,16).
CaM kinase II is a large multimer (ϳ600 kDa) composed of 8 -10 individual kinase subunits of 54 -60 kDa size. Four distinct kinase subunits (␣, ␤, ␦, ␥) and a number of variants arising by alternative splicing have been cloned to date (9,17,18). A distinguishing feature of CaM kinase II is that it undergoes autophosphorylation in the presence of Ca 2ϩ /CaM on a specific conserved threonine residue (Thr 286 in the ␣-subunit) which results in the generation of Ca 2ϩ /CaM-independent (or "autonomous") kinase activity (9). 70 -80% of the total Ca 2ϩ / CaM-dependent kinase activity may become autonomous in vitro under optimal autophosphorylation conditions (19). Autophosphorylation on Thr 286 has also been reported to result in a 1000-fold increase in the affinity of the kinase subunits for calmodulin (20). Thr 286 autophosphorylation is predicted to be cooperative, since it has been shown to occur by an intersubunit intraholoenzyme reaction which requires that both the phosphorylating subunit and the substrate subunit have bound Ca 2ϩ /CaM (21).
Autophosphorylation-dependent generation of autonomous activity and "calmodulin trapping" could result in CaM kinase II activity which in vivo would outlast transient increases in [Ca 2ϩ ] i and enable the kinase to respond in a frequency-dependent manner to repetitive transient increases in [Ca 2ϩ ] i (20,21). However, previous in vitro studies indicated that at saturating calmodulin concentrations, CaM kinase II activation and phosphorylation of exogenous substrates required a relatively high concentration of free Ca 2ϩ (K a Ͼ 700 nM) (22,23). Given that: 1) inactive CaM kinase II has a low affinity for calmodulin, 2) CaM kinase II autophosphorylation is predicted to be a cooperative process requiring at least two activated subunits per holoenzyme, and 3) intact cells have phosphatase activities that are capable of reversing autophosphorylation; activation of CaM kinase II in vivo and subsequent autophosphorylation with generation of autonomous activity may be relatively insensitive to gradual or sustained small increases in activator Ca 2ϩ which would be expected in response to many physiological stimuli.
CaM kinase II is present in cultured rat aortic VSM cells (about 1 ⁄10th the activity found in comparable brain extracts) and is comprised mainly of the ␦ 2 -subunit variant (18). In the present study, we assessed the ability of Ca 2ϩ -mobilizing stimuli to activate CaM kinase II in VSM, resulting in the auto-phosphorylation-dependent generation of the autonomous form of the kinase. The Ca 2ϩ dependence for generating autonomous CaM kinase II activity was determined in cells that were made selectively permeable to Ca 2ϩ . Despite a relatively low sensitivity for [Ca 2ϩ ] in the development of autonomous CaM kinase II activity under these conditions (EC 50 (concentration producing 50% of maximal response) ϭ 692 nM), all of the Ca 2ϩmobilizing agents tested were able to significantly stimulate the development of autonomous CaM kinase II activity in intact cells indicating in situ activation of CaM kinase II. Our findings further suggest an important role for intracellular pools of Ca 2ϩ in the activation of CaM kinase II and the subsequent initiation of cooperative autophosphorylation and generation of autonomous activity.
Cell Culture-VSM cells were dispersed from the medial layer of the thoracic aorta of Sprague-Dawley rats (150 -250 g) using collagenase and elastase (Worthington) according to the method of Geisterfer et al. (24). The cells were grown in an equal mixture of Dulbecco's modified Eagle's medium and Ham's F-12 medium containing antibiotic/antimycotic (Life Technologies, Inc.) and 10% fetal calf serum (Hyclone, Logan, UT) under standard culture conditions. Cells were subcultured weekly into 60-and 100-mm 2 dishes and grown to confluence before use in the experiments. VSM cells from the third to tenth passage were used in this study.
Kinase Assay-Cells were stimulated with various Ca 2ϩ -mobilizing agents at 37°C in HEPES-buffered Hanks' balanced salt solution (HBSS, pH 7.4) while still attached to the culture dish. At appropriate times the reaction was stopped by aspiration of the media and addition of ice-cold lysis buffer containing an equal mixture of Ca 2ϩ ,Mg 2ϩ -free HBSS (Life Technologies, Inc.) and Buffer A. Buffer A was composed of 50 mM MOPS (pH 7.4), 2 mM EGTA, 100 mM sodium fluoride, 100 mM sodium pyrophosphate, 2 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride, 0.4 unit/ml aprotinin and 1% Nonidet P-40. After 5-10 min on ice, crude lysates were aspirated and centrifuged at 17,000 ϫ g for 10 min. The resulting supernatant was kept on ice until the kinase assay was performed. Protein content in the lysates was determined using the Bio-Rad Protein reagent with bovine serum albumin as the standard. Total CaM kinase II activity in the lysate was assayed in 25 l containing 10 mM MOPS (pH 7.4), 10 mM magnesium chloride, 3 mM EGTA, 4 mM calcium chloride, 400 nM calmodulin, 0.2 mM [␥-32 P]ATP (400 -1000 cpm/pmol), 20 M autocamtide-2 (KKALRRQETVDAL, Ref. 19) as substrate, and 0.5-2.5 g of lysate protein. To determine the Ca 2ϩ /CaM-independent (autonomous) activity in the same lysates CaCl 2 and CaM were omitted from the kinase assay mixture. The reactions were carried out at 30°C in a shaking incubator for 3 min and terminated by precipitation of the phosphorylated peptide on Whatman P-81 paper (Whatman). The papers were rinsed thoroughly in 75 mM phosphoric acid, and the adherent radioactivity was quantified in ReadySafe LS mixture (Amersham Corp.) by liquid scintillation counting. Kinase activity was expressed as nanomoles of P i transferred to the substrate/min/mg lysate protein. Autonomous (Ca 2ϩ /CaM-independent) kinase activity was expressed as a percent of total Ca 2ϩ /CaMdependent activity from the same samples.
[ 32 P]PO 4 Labeling and Immunoprecipitation of CaM Kinase II-Cellular ATP pools were labeled for 16 h by addition of 100 -200 Ci/ml [ 32 P]orthophosphoric acid in phosphate-free medium under standard culture conditions. Cells were equilibrated at 37°C in HEPES-buffered HBSS for 5 min before stimulation with 1 M ionomycin for up to 10 min, at which time the medium was aspirated and cells were disrupted with ice-cold lysis buffer. Equivalent amounts of lysate protein were incubated with 3 l/ml CK2-DELTA antiserum precomplexed to protein A-agarose beads (18) for 2 h at 4°C unless indicated otherwise. Beadimmune complexes were washed twice with 500 l of lysis buffer and solubilized in SDS sample buffer. Proteins were resolved electrophoretically on 8% SDS-polyacrylamide gels, transferred electrophoretically onto Immobilon membranes (Millipore), and visualized by autoradiography.
125 I-CaM Overlay-Immunoprecipitated lysate proteins were resolved by SDS-polyacrylamide gel electrophoresis and transferred to Immobilon membranes as indicated above. Membranes were incubated for 1 h with 125 I-CaM (1 Ci/5 ml) at room temperature in a buffer containing 50 mM Tris (pH 7.4), 150 mM NaCl (Tris-buffered saline), 1 mM CaCl 2 and 5% non-fat dry milk. Following thorough rinsing of membranes with Ca 2ϩ -containing Tris-buffered saline, calmodulinbinding proteins in immunoprecipitated samples were identified by exposing membranes to x-ray film for up to 72 h. Control overlay procedures were performed under identical conditions except that 1 mM CaCl 2 was replaced with 1 mM EGTA.
[Ca 2ϩ ] i Measurements Using Fura-2-VSM cells were exposed to Fura-2/AM for 1 h at 37°C in HBSS containing 1 mg/ml bovine serum albumin and 2 mg/ml dextrose to load cells with the dye. The cells were washed with HBSS to remove unincorporated Fura-2, suspended in HBSS using 0.25 mg/ml trypsin (Life Technologies, Inc.), and then washed once with HBSS containing 1 mg/ml soybean trypsin inhibitor. Intracellular Fura-2 fluorescence was monitored in suspended cells maintained at 37°C using a SPEX DM3000 Fluorolog spectrometer (SPEX Industries, Camden, NJ ] i in a stepped manner and at each steady-state a 40 l aliquot of cell suspension was quickly removed from the cuvette and disrupted in an equal volume of lysis buffer (4°C). Cell lysates were processed, and total and Ca 2ϩ /CaM-independent kinase activities were determined as indicated above. Measured [Ca 2ϩ ] i was plotted against the Ca 2ϩ /CaM-independent activity (expressed as a percent of total Ca 2ϩ /CaM-dependent activity) to construct an intracellular Ca 2ϩ dependence curve for in situ activation and generation of the autonomous form of the kinase.
In a separate protocol, Fura-2-loaded cells were suspended in normal HBSS (1.8 mM CaCl 2 ), and [Ca 2ϩ ] i was elevated using single, noncumulative ionomycin additions (0.01-10 M) followed by rapid sampling of the cell suspension at various time points and kinase assay. In a third protocol, intracellular pools of calcium were first depleted with 10-min treatment of 1 M thapsigargin and then cells were challenged with different concentrations of ionomycin before cell lysis and kinase assay. Total and Ca 2ϩ /CaM-independent kinase activity was then plotted against measured [Ca 2ϩ ] i to produce in situ Ca 2ϩ dependence curves for generation of autonomous CaM kinase II activity in cells with or without intact intracellular Ca 2ϩ pools.
Data Analysis-Grouped data represent the mean Ϯ S.E. of three or more experiments, and statistical comparisons between means were performed using one-way analysis of variance followed by Bonferroni's post hoc test using GraphPad Instat software (San Diego, CA). The Ca 2ϩ dependence plots were subjected to nonlinear regression analyses using GraphPad Prism software (San Diego, CA). Statistical comparisons among curves were performed according to Motulsky and Ransnas, (26). Briefly, two curve fits were considered statistically different if they described the data better than a single curve fitted to the pooled data from both sets. This was achieved by carrying out an F-test to compare the variance of the pooled data set with the sum of the variance from the individual data sets. In all comparisons p Ͻ 0.05 was taken to indicate statistical significance.

CaM Kinase II Activation and Autophosphorylation in Situ-
The synthetic peptide autocamtide-2 has been reported to be a relatively specific substrate for CaM kinase II compared with other multifunctional kinases such as protein kinase C and cAMP-dependent protein kinase (19). In representative experiments, total Ca 2ϩ /CaM-dependent autocamtide-2 kinase activity in cell lysates from unstimulated VSM cells was found to be 8.81 Ϯ 0.74 nmol/min/mg protein, while Ca 2ϩ /CaMindependent activity in the same samples was 0.87 Ϯ 0.14 nmol/min/mg of protein (n ϭ 10). A hallmark of CaM kinase II is activation-dependent autophosphorylation leading to development of Ca 2ϩ /CaM-independent or autonomous kinase ac-tivity (9). Exposure of intact VSM cells to the Ca 2ϩ ionophore ionomycin (1 M) resulted in a rapid increase in the level of Ca 2ϩ /CaM-independent autocamtide-2 kinase activity assayed in vitro, from 11.4 Ϯ 1.9% of the total (Ca 2ϩ /CaM-dependent) kinase activity at rest to a peak value (15 s) of 52.7 Ϯ 4.4% of total activity, a response which was partly sustained for the duration of the 10 min exposure to the ionophore (Fig. 1A).
Experiments were carried out to determine if these increases in Ca 2ϩ /CaM-independent autocamtide-2 kinase activity correlated with phosphorylation of CaM kinase II subunits and to confirm that the activity measured was in fact due to CaM kinase II. Ionomycin stimulation of cells with 32 P-labeled ATP pools resulted in incorporation of 32 P into a 52-53-kDa protein which was isolated from the lysates by immunoprecipitation with an antipeptide antibody (CK2-DELTA) specific for the carboxyl terminus of the CaM kinase II ␦-subunit (Fig. 1B). The CK2-DELTA and another CaM kinase II subunit nonselective antibody have previously been shown to detect a band of this size on immunoblots of rat aortic VSM fractions (18). In the present study, the identity of this band as a CaM kinase II subunit was confirmed by overlay with 125 I-calmodulin (Fig.  1B, rightmost lane) and by immunoblotting with a different subunit nonselective anti-CaM kinase II antibody (obtained from K. Smith and R. Colbran, Vanderbilt University; data not shown). Phosphorylation of CaM kinase II subunits was maximal within 15-30 s of ionomycin addition, corresponding with peak increases in Ca 2ϩ /CaM-independent kinase activity, and persisted for the duration of the stimulation (Fig. 1B). In addition to the 53-kDa band a high molecular weight protein (Ͼ140 kDa), and proteins with estimated sizes of 56 and 60 kDa in the immunoprecipitates were also phosphorylated. The 56and 60-kDa bands may be minor CaM kinase II subunits, since corresponding bands could be detected with longer exposure of the 125 I-CaM overlay.
Immunoprecipitation with CK2-DELTA depleted Ca 2ϩ /CaMdependent and -independent autocamtide-2 activity by 80 -90% from lysates of ionomycin-stimulated cells, confirming the specificity of the autocamtide-2 substrate based assay for CaM kinase II and indicating that the stimulated Ca 2ϩ /CaM-independent activity was due to autonomous CaM kinase II (Fig. 2). While Ca 2ϩ /CaM-independent autocamtide-2 kinase activity was low in resting cells, the CK2-DELTA antibody also immu-noprecipitated 60% of this activity, consistent with a small amount of CaM kinase II autophosphorylation which was observed in unstimulated cells (Fig. 1B). The residual Ca 2ϩ /CaMdependent activity remaining following immunoprecipitation could be due to other kinases such as CaM kinase IV which do not undergo autophosphorylation-dependent transition to a Ca 2ϩ /CaM-independent form (27). Further confirmation that Ca 2ϩ /CaM-independent autocamtide-2 kinase activity in the VSM cell lysates was due to autonomous CaM kinase II was obtained by pretreating the cells with KN-93, a CaM kinase II inhibitor which interferes with calmodulin binding to the kinase subunit (10). In the experiments shown in Fig. 3, stimulation of VSM cells for 15 s with ionomycin (1 M) resulted in a 7.9-fold increase in Ca 2ϩ /CaM-independent activity without altering total CaM kinase II activity in cells. Pretreatment of cells with 30 M KN-93 completely prevented the ionomycininduced increase in Ca 2ϩ /CaM-independent kinase activity. Based on the experiments shown in Figs. 1-3 we concluded that autocamtide-2 kinase activity was indicative of CaM kinase II in the VSM cell lysates and that Ca 2ϩ /CaMindependent kinase activity was due to the autophosphorylated or autonomous form of the kinase.
In Situ [Ca 2ϩ ] i Dependence for Generating Autonomous CaM Kinase II-Vascular smooth muscle cells were loaded with the fluorescent Ca 2ϩ indicator Fura-2, depleted of Ca 2ϩ by incubating with thapsigargin (1 M) media in Ca 2ϩ -free HBSS containing 2 mM EGTA, and selectively permeabilized to Ca 2ϩ by addition of ionomycin (1 M). Stepped increases in free [Ca 2ϩ ] i were obtained by cumulative additions of CaCl 2 to the media (Fig. 4, inset) and quantified using the Fura-2 fluorescence. After each step change in [Ca 2ϩ ] i a 40-l aliquot of the cell suspension was removed (indicated by the vertical deflections in the tracing) and assayed for Ca 2ϩ /CaM-dependent and -independent CaM kinase II activity. The resulting data established the in situ Ca 2ϩ dependence for the generation of autonomous CaM kinase II activity (Fig. 4). Nonlinear curve fitting to a sigmoidal (logistic) function resulted in an EC 50 of 692 Ϯ 28 nM [Ca 2ϩ ] i with 68 Ϯ 2% of the total activity being autonomous at maximal [Ca 2ϩ ] i (goodness of fit r 2 ϭ 0.915). The calculated Hill slope for the curve was 3, consistent with a highly cooperative process for autophosphorylation of the kinase subunits by an intraholoenzyme intersubunit reaction (9,28 (Fig. 5B), although in these cells the response was generally smaller than that obtained with angiotensin II. Platelet-derived growth factor (PDGF-BB; 40 ng/ml) stimulated slower and relatively small increases in average [Ca 2ϩ ] i (Fig. 5C), whereas 1 M ionomycin elicited a rapid somewhat transient increase in [Ca 2ϩ ] i which was partially sustained (Fig. 5D). By quantitatively comparing these [Ca 2ϩ ] i responses with the [Ca 2ϩ ] dependence for generating autonomous CaM kinase II shown in Fig. 4, it would not be expected that resting levels of [Ca 2ϩ ] i would support significant autonomous CaM kinase II activity or that addition of these receptor agonists would substantially increase autonomous activity. In contrast to this prediction, resting levels of autonomous activity in the intact VSM cells ranged from 4.4 to 12.5% of total activity, and each stimulus significantly increased autonomous CaM kinase II activity with kinetics which paralleled increases in [Ca 2ϩ ] i (Fig. 5). In the case of angiotensin II or vasopressin, transient increases in autonomous CaM kinase II activity to 41 Ϯ 4% (n ϭ 5) and 22.4 Ϯ 2.2% (n ϭ 3) of total activity, respectively, were reached within 15 s (Figs. 5, A and B). PDGF also increased autonomous CaM kinase II activity, but at a slower rate and to a lesser extent (14.7 Ϯ 1.1% autonomous activity at 60 s), as compared with angiotensin II and vasopressin (Fig. 5C). Angiotensin II-induced increases in autonomous CaM kinase II activity were concentration-dependent with an estimated EC 50 of 6 nM (data not shown), which is similar to the reported dissociation constant of the agonist for its receptor (29,30). Analysis of variance predicted that the EC 50 obtained with this protocol was statistically different from that of the curve in Fig.  3 (F (4, 139) ϭ 60.600, p Ͻ 0.001). The apparent Ca 2ϩ dependence obtained with this protocol was quantitatively consistent with the increases in [Ca 2ϩ ] i and autonomous CaM kinase II activity observed in intact VSM cells in response to addition of the physiological stimuli shown in Fig. 5.

Role of Intracellular Ca 2ϩ Pools in the Generation of Autonomous Activity-Increasing
In order to explain the apparent higher Ca 2ϩ sensitivity for generating autonomous CaM kinase II in intact VSM cells, we considered the possibility that average [Ca 2ϩ ] i calculated using the Fura-2 technique did not reflect locally high [Ca 2ϩ ] i , produced in response to ionomycin and peptide agonists that release intracellular pools of Ca 2ϩ (3,30,31), which could support the cooperative autophosphorylation of CaM kinase II. To test this hypothesis, intact cells in normal Ca 2ϩ HBSS were pretreated with 1 M thapsigargin to deplete intracellular Ca 2ϩ pools and the protocol shown in Fig. 6A was repeated using ionomycin as a stimulus. The resulting apparent Ca 2ϩ dependence ( Fig. 6B; goodness of fit r 2 ϭ 0.940) was significantly to the right of the curve obtained from VSM cells with intact intracellular pools of Ca 2ϩ (EC 50 ϭ 616 Ϯ 37 nM; F (4, 62) ϭ 29.278, p Ͻ 0.001, n ϭ 8), although the maximal increase in autonomous activity, Hill slope, and the total Ca 2ϩ /CaM-dependent kinase activity were not significantly altered by thapsigargin treatment (Fig. 6B). The apparent [Ca 2ϩ ] i dependence for generating autonomous CaM kinase II activity described by this data set was not significantly different from that obtained using the Ca 2ϩ step protocol shown in Fig. 4. A 10-min treatment of VSM cells with 1 M thapsigargin also abolished a 3-fold increase in autonomous activity elicited by 0.1 M angiotensin II (n ϭ 3, data not shown). These results are consistent with the contribution of intracellular Ca 2ϩ pools in the activation of CaM kinase II and the cooperative generation of autonomous kinase activity.

DISCUSSION
The present study was undertaken as a first approach toward assessing the activation of CaM kinase II in intact VSM cells. Ca 2ϩ /CaM-independent CaM kinase II activity (autonomous activity) was measured in VSM cell lysates for two reasons: 1) generation of autonomous activity provides an index of CaM kinase II activation in the intact cell prior to lysis, and 2) the appearance of autonomous CaM kinase II activity in situ has been hypothesized to be of functional significance.
Evidence that the ionomycin-induced Ca 2ϩ /CaMindependent autocamtide-2 kinase activity was due to autophosphorylation-dependent generation of autonomous CaM kinase II includes: 1) the activity was efficiently immunoprecipitated with a CaM kinase II ␦-subunit-specific antibody; 2)  Fig. 3, which is included to facilitate comparison, and the dotted lines are 95% confidence intervals for the curves. B, the same experiment was repeated in cells pretreated with 1 M thapsigargin (to deplete intracellular Ca 2ϩ pools) for 10 min prior to stimulation with ionomycin. The dotted lines are the 95% confidence interval for the curve, and the dashed curve is of the open circles from A above. The data are from eight independent experiments and goodness of fit r 2 ϭ 0.940 for the curve. due to phosphorylation on additional serine/threonine sites (9) accompanied by dephosphorylation of Thr 286 . The functional significance of these additional phosphorylation events is largely unknown.
It was previously estimated that the free [Ca 2ϩ ] for halfmaximal activation of CaM kinase II in vitro under conditions of saturating CaM was relatively high, in the range of 700-2000 nM (22,23). Because generation of autonomous CaM kinase II activity results from intersubunit phosphorylation between activated (Ca 2ϩ /CaM bound) subunits (9,27), a similar Ca 2ϩ dependence would be predicted for generation of autonomous activity in vivo. Other factors in intact cells could act to further decrease the apparent Ca 2ϩ sensitivity for CaM kinase II activation and generation of autonomous activity, such as limiting concentrations of free calmodulin and/or protein phosphatase activities that are capable of reversing CaM kinase II autophosphorylation. The calculated EC 50 of 692 nM Ca 2ϩ in selectively permeabilized VSM cells is the first direct estimation of the in situ Ca 2ϩ dependence for generating autonomous CaM kinase II activity. Significantly, a positive Hill slope of 3 for the relationship indicates that Ca 2ϩ -dependent generation of autonomous CaM kinase II activity in situ is a highly cooperative process, consistent with in vitro data indicating cooperativity in the intersubunit intraholoenzyme autophosphorylation reaction (21). This value also provides an estimate of the Ca 2ϩ sensitivity for activation of CaM kinase II, which is a prerequisite for generating autonomous activity. Tansey et al. (32) reported a similar half-maximal [Ca 2ϩ ] i (ϳ500 nM) for phosphorylation and desensitization of myosin light chain kinase in permeabilized bovine tracheal smooth muscle cells, a process thought to be mediated by CaM kinase II.
Under optimal autophosphorylation conditions in vitro, we have found that autonomous activity of the CaM kinase II holoenzyme composed of recombinant ␦ 2 -subunits could reach as much as 70% of the total Ca 2ϩ /CaM-dependent activity. 2 Similar maximal levels of independent activity have been observed by others using various purified and recombinant forms of CaM kinase II (19). To our knowledge, it is not known to what extent the difference in maximal autonomous activity compared with total activity reflects the stoichiometry of holoenzyme autophosphorylation, specifically on Thr 286 in the subunits, and/or whether autophosphorylated kinase has a lower activity for substrates than Ca 2ϩ /CaM-activated kinase. Therefore, as a marker of Ca 2ϩ /CaM-dependent activation of CaM kinase II in vivo, the level of autonomous activity may actually underestimate the full extent of CaM kinase II activation. Similar maximal levels of autonomous activity were also generated in the selectively permeabilized VSM cells (Fig.  4), or transiently in cells with intact Ca 2ϩ pools in response to ionomycin stimulation (Figs. 1A and 6A), indicating a near complete activation of kinase subunits in situ. This suggests that in VSM cells under these conditions, free CaM is not a limiting factor and protein phosphatases do not significantly antagonize the maximal development of autonomous activity. Comparable increases in autonomous CaM kinase II activity in situ were reported previously only in KCl-depolarized PC12 cells (33). In most other studies where this approach has been used to assess CaM kinase II activation, small increases in autonomous activity (2-fold or less) were reported in response to stimuli which raise [Ca 2ϩ ] i (34,35), indicating either lower levels of CaM kinase II activation and/or incomplete control of CaM kinase II subunit dephosphorylation during cell lysis.
Average resting VSM cell [Ca 2ϩ ] i was found to be between 100 and 150 nM, and the physiological agonists angiotensin II, vasopressin, and PDGF produced transient increases in [Ca 2ϩ ] i to values which were typically in a range of 200 -400 nM. These results are similar both qualitatively and quantitatively to previously published data obtained in VSM cells (31). Based on the quantitative relationship between [Ca 2ϩ ] and generation of autonomous CaM kinase II activity (Fig. 4), these levels of [Ca 2ϩ ] i should have produced minimal autonomous CaM kinase II activity. However, a small amount of Ca 2ϩ /CaM-independent activity was measurable in lysates from unstimulated intact cells (about 3-12% of the total activity of which about 60% was immunoprecipitated with the CaM kinase II antibody), and all of the stimuli tested produced significant transient increases in autonomous CaM kinase II activity, with angiotensin II producing the largest increases, to as much as 40% of the total kinase activity. A significant factor in interpreting these data was that the in situ Ca 2ϩ dependence curve for generating autonomous CaM kinase II resulted from experiments using cells which were depleted of intracellular Ca 2ϩ and which required Ca 2ϩ influx from the extracellular medium to activate the kinase. In the case of experiments examining agonist-stimulated increases in CaM kinase II activity, intracellular pools of Ca 2ϩ were intact and the stimuli used (angiotensin II, vasopressin, and PDGF) are known to mobilize Ca 2ϩ from inositol trisphosphate-sensitive intracellular pools (36 -38). Ionomycin was useful in determining the effect of intracellular Ca 2ϩ pools on the apparent Ca 2ϩ dependence for generating autonomous CaM kinase II activity, since it is able to mobilize Ca 2ϩ from both intracellular and extracellular pools to elevate [Ca 2ϩ ] i (3). The relationship between ionomycin-induced increases in [Ca 2ϩ ] i and autonomous CaM kinase II activity in cells with intact [Ca 2ϩ ] i pools predicted an apparent EC 50 for [Ca 2ϩ ] i of 304 nM for the generation of autonomous kinase activity. This apparent Ca 2ϩ sensitivity was consistent with the physiological agonist-induced increases in [Ca 2ϩ ] i and autonomous kinase activity. Prior depletion of intracellular Ca 2ϩ pools with thapsigargin markedly decreased the apparent Ca 2ϩ sensitivity for ionomycin-induced generation of autonomous CaM kinase II activity and blocked angiotensin II-induced transients in [Ca 2ϩ ] i and generation of autonomous activity.
At least three factors could contribute to the observed differences in Ca 2ϩ sensitivities for generating autonomous CaM kinase II in cells with or without intact pools of intracellular Ca 2ϩ . 1) In the protocol using ionomycin-permeabilized and Ca 2ϩ -depleted cells, Ca 2ϩ added to the extracellular medium might be expected to equilibrate uniformly through the cytoplasm allowing Fura-2 to faithfully report [Ca 2ϩ ] i . This protocol should then result in a reasonable estimation of the Ca 2ϩ sensitivity for activation of CaM kinase II and generation of autonomous activity. In cells with intact intracellular Ca 2ϩ pools, locally high [Ca 2ϩ ] i resulting from sarcoplasmic reticulum release in response to ionomycin or physiological stimuli may contribute to an average [Ca 2ϩ ] i reported by Fura-2. Several studies have documented spatial inhomogeneities in [Ca 2ϩ ] i using microscopic imaging of single cell Ca 2ϩ transients (39,40). Models of Ca 2ϩ signaling in VSM, which take into account restricted diffusional spaces underlying the plasma membrane and surrounding sarcoplasmic reticulum, suggest that micromolar concentrations of [Ca 2ϩ ] i could develop and persist for short periods of time in these spaces in response to Ca 2ϩ -mobilizing stimuli (41). 2.) Assuming that locally high concentrations of Ca 2ϩ are produced in intact cells, a second factor contributing to the observed variable Ca 2ϩ sensitivity is the cooperative nature of the process for generating autonomous CaM kinase II activity. 3) CaM kinase II itself could be localized in discrete subcellular regions in close proximity to the sarcoplasmic reticulum and released Ca 2ϩ . In this case, the effects of locally high [Ca 2ϩ ] i and a cooperative process for generating autonomous activity would be maximized by CaM kinase II holoenzymes optimally positioned for activation in response to stimuli that mobilize intracellular calcium. The fact that Fura-2 reports average [Ca 2ϩ ] i levels in intact cells and that inhomogeneities in [Ca 2ϩ ] i could also exist in unstimulated cells may partially explain the small amount of Ca 2ϩ /CaM-independent autocamtide-2 kinase activity which could be attributed to autonomous CaM kinase II in resting VSM cells. These inhomogeneities in [Ca 2ϩ ] could exist within all cells such that a small fraction of the kinase within each cell became active (reflected by the resting autonomous activity) and/or within a small subset of "leaky" cells where [Ca 2ϩ ] i is maximally elevated and essentially all of the CaM kinase II is active and autophosphorylated.
These arguments are supported indirectly by reports of localization and/or association of CaM kinase II with the sarcoplasmic reticulum of cardiac and skeletal muscle, where it is proposed to play a role in regulation of cytosolic Ca 2ϩ by modulating Ca 2ϩ -ATPase activity (42,43). In contrast, MacNicol and Schulman (44) demonstrated greater activation of CaM kinase II in PC12 cells by K ϩ -induced depolarization, which produced influx of Ca 2ϩ , than by ionomycin, which releases Ca 2ϩ from endoplasmic reticulum pools, even though the two agents raised average [Ca 2ϩ ] i to a similar extent. Thus, unlike in VSM cells, CaM kinase II in PC12 cells appears to be more sensitive to activation by Ca 2ϩ from extracellular sources, which is consistent with a role for the enzyme in neurotransmitter metabolism and secretion during depolarizationinduced Ca 2ϩ influx (10,11). Discrete localization and activation of CaM kinase II in subcellular compartments may also be a mechanism for regulating the function of this enzyme which otherwise shows little substrate specificity in vitro (8,9).
The physiological significance of the autonomous form of CaM kinase II is not known. Development of the autonomous form of CaM kinase II has been proposed as a mechanism by which activity of CaM kinase II could be maintained even after [Ca 2ϩ ] i returns to resting values (45). However, in VSM cells the Ca 2ϩ /CaM-independent activity did not appear to significantly outlast the Ca 2ϩ transient generated by physiological agonists, indicating that once activator Ca 2ϩ falls, phosphatases rapidly dephosphorylate and inactivate CaM kinase II subunits. This doesn't negate the possible importance of this mechanism in maintaining kinase activity and trapping calmodulin on a shorter time scale such as might be encountered in depolarizing nerve terminals (20). Another possible role for autonomous CaM kinase II is that it may phosphorylate specific substrates (8).
In summary, this study demonstrates that activation of CaM kinase II and generation of autonomous CaM kinase II activity occurs in situ over a range of [Ca 2ϩ ] between 0.2 and 2 M. Generation of autonomous CaM kinase II activity in situ is highly cooperative, a reflection of the unique structure of the CaM kinase II holoenzyme and the requirement for intersubunit autophosphorylation on Thr 286 to generate autonomous activity. Locally high concentrations of [Ca 2ϩ ] i in intact VSM cells, produced by agonists that release Ca 2ϩ from intracellular pools, result in the cooperative generation of autonomous CaM kinase II activity over a range of average cell [Ca 2ϩ ] i of 0.1-1 M.