Glucose Activates the Multifunctional Ca2+/Calmodulin-dependent Protein Kinase I1 in Isolated Rat Pancreatic Islets*

The influence of the insulin secretagogues, glucose and K+, to activate the multifunctional, Ca2+/calmodu-lin-dependent protein kinase I1 (CaM kinase 11) in iso- lated rat pancreatic islets has been examined. Glucose (28 mM) and K+ (40 mM) were demonstrated to induce a 1.89 2 0.19- and 1.75 f 0.16-fold increase, respec- tively, in phosphorylation of a subunit of CaM kinase I1 immunoprecipitated by an anti-CaM kinase IIa an- tibody. In intact islets, glucose and K+ also induced the generation of an autonomous, Ca2+/calmodulin-inde-pendent protein kinase I1 activity characteristic of autophosphorylated enzyme. Maximal activation, 2.9 2 0.2- and 3.0 f 0.5-fold for glucose and K+, respec- tively, relative to basal glucose control, was achieved at 2.5-5 min followed by a decline to near basal levels by 20 min. Glucose induced the production of autonomous CaM kinase I1 activity that, in terms of -fold stimulation, correlated closely with the extent of insulin release over a glucose concentration range of 3- 28 mM. This stimulated activity was completely prevented by an inhibitor of glucose metabolism, manno- heptulose. These

The influence of the insulin secretagogues, glucose and K+, to activate the multifunctional, Ca2+/calmodulin-dependent protein kinase I1 (CaM kinase 11) in isolated rat pancreatic islets has been examined. Glucose (28 mM) and K+ (40 mM) were demonstrated to induce a 1.89 2 0.19-and 1.75 f 0.16-fold increase, respectively, in phosphorylation of a subunit of CaM kinase I1 immunoprecipitated by an anti-CaM kinase IIa antibody. In intact islets, glucose and K+ also induced the generation of an autonomous, Ca2+/calmodulin-independent protein kinase I1 activity characteristic of autophosphorylated enzyme. Maximal activation, 2.9 2 0.2-and 3.0 f 0.5-fold for glucose and K+, respectively, relative to basal glucose control, was achieved at 2.5-5 min followed by a decline to near basal levels by 20 min. Glucose induced the production of autonomous CaM kinase I1 activity that, in terms of -fold stimulation, correlated closely with the extent of insulin release over a glucose concentration range of 3-28 mM. This stimulated activity was completely prevented by an inhibitor of glucose metabolism, mannoheptulose. These data demonstrate that the exposure of islets to stimulatory glucose concentrations activates CaM kinase 11. The close correlation of enzyme activation with insulin secretion is consistent with the hypothesis that CaM kinase I1 plays an important role in the regulation of insulin secretion or related &cell processes. D-Glucose is the primary physiological stimulator of insulin secretion from the pancreatic @-cell in man and rodents and induces biphasic secretion from isolated islets that mimics physiological secretion in vivo (Hedeskov, 1980). It is clear that the metabolism of glucose is required for secretion, and it is thought that the first enzyme in this process, glucokinase, serves as the intracellular glucose recognition molecule (Meglasson and Matschinsky, 1986). Subsequent to metabolism, the principal response of the @-cell to glucose is an increased intracellular Caz+, contributed to, in the most part, by the stimulation of Caz+ influx (Prentki and Matschinsky, 1987;Wollheim and Sharp, 1981). The cellular mechanism linking * This work was supported by Juvenile Diabetes Foundation International and American Heart Association Texas Affiliate grants (to R. A. E.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ll To whom correspondence should be addressed Dept. of Biochemistry and Molecular Biology, University of North Texas Health Science Center at Fort Worth, 3500 Camp Bowie Blvd., Fort Worth, TX 76107-2699. these processes is thought to involve an increased intracellular ATP:ADP ratio as the result of glucose metabolism and the depolarization of the @-cell via the closing of ATP-sensitive K+ channels (KATP channels)' (Rajan et al., 1990). Cell depolarization promotes the opening of L-type voltage-dependent Ca2+ channels on the plasma membrane allowing the influx of Ca2+ down its concentration gradient (Rajan et al., 1990).
In contrast to the recent progress in the understanding of these early events of the secretion process, the cellular events linking increased Ca2+ concentrations to the release of hormone are unknown. Several reports support a role of the cytoskeleton in insulin secretion (Howell and Tyhurst, 1984;Stutchfield and Howell, 1984;Wang et al., 1990), which has led to the hypothesis that cytoskeletal-associated, Ca2+-activated proteins, particularly protein kinases, may regulate insulin secretion (Ashcroft and Hughes, 1990). One enzyme implicated in this process is the Ca2+/calmodulin-dependent protein kinase I1 (CaM kinase 11).
An increasing body of evidence has now established CaM kinase I1 as one of the major protein kinases orchestrating intracellular responses to extracellular signals (Hanson and Schulman, 1992). It is widespread in tissues and is characterized by the ability to phosphorylate a diverse array of substrates in vitro and in situ that contribute to a variety of cellular functions that include the regulation of cell metabolism, ion flux, Ca2+ homeostasis, and cytoskeletal function (Schulman and Hanson, 1993;Colbran et al., 1989). CaM kinase I1 activity is found in greatest concentration in neurons where it is involved in the regulation of neurotransmitter biosynthesis and release (Benfenati et al., 1992). In the forebrain, CaM kinase I1 exists as an oligomer of 8-12 similar subunits including an a-subunit of 50-54 kDa and a /?-subunit of 58-60 kDa in a a:@ ratio of 3-4:l (Lin et al., 1987;Hanson and Schulman, 1992). This ratio, however, varies in other brain tissues (Miller and Kennedy, 1985;McGuinness et al., 1985). Other isoforms, @' (generated by alternative splicing of the @ gene) 7 , and 6, have been identified by cloning and sequencing analysis, but no additional cellular function has been detected in these subunits (Tobimatsu and Fugisawa, 1989;Schulman and Hanson, 1993).
Activation of CaM kinase I1 by Ca2+ and calmodulin in vitro results in rapid enzyme autophosphorylation (Schulman and Hanson, 1993). The autophosphorylated enzyme is no longer dependent on Ca2+ and calmodulin for activity and performs further autophosphorylation or the phosphorylation of exogenous substrates in the absence of these cofactors (Saitoh and Schwartz, 1985;Miller and Kennedy, 1986; 'The abbreviations used are: KAT^ channel, ATP-sensitive K+ channel; CaM kinase 11, Ca2+/calmodulin-dependent protein kinase 11; MES, 2-(N-morpholino)ethanesulfonic acid PIPES, piperazine-N,N"bis(2-ethanesulfonic acid).

4947
Schworer et al., 1986). Enzyme autophosphorylation and generation of autonomous kinase I1 activity have since been reported in intact cells in response to agonists that increase intracellular Ca2+ concentration (Gorelick et al., 1988;Fukunaga et al., 1989;MacNicol et al., 1990;Jefferson et al., 1991;Ocorr and Schulman, 1991). Isolated pancreatic islets have been shown to possess a cytoskeletal-associated Ca2+-and calmodulin-dependent protein kinase activity (referred to as P53 kinase (Harrison and Ashcroft, 1982;Ashcroft and Hughes, 1990;Landt et al., 1982)), the major substrates of which are endogenous proteins of M, 53,000 and 57,000 (Colca et al., 1983a). This P53 kinase has been purified to near homogeneity from RINm5F cells and the enzyme shown to possess properties reminiscent of the rat brain CaM kinase I1 including a low affinity for calmodulin (relative to other calmodulin-dependent enzymes) and the ability to phosphorylate known substrates of CaM kinase 11.' Furthermore, a P53 kinase in islets has recently been shown to possess kinetic properties similar to rat brain CaM kinase I1 (Hughes et al., 1993). Evidence from which a possible role of CaM kinase I1 in insulin secretion can be evaluated is, however, purely circumstantial. In a preliminary study, the phosphorylation of the 53,000 molecular weight protein was increased, though only modestly, in glucosestimulated islets (Colca et al., 1983a). The diabetogenic agent, alloxan, was further shown to inhibit P53 kinase activity and glucose-induced insulin secretion from islets in parallel (Colca et aL, 1983b). In more recent efforts, the putative inhibitors of CaM kinase 11, KN-62 (Wenham et al., 1992;Li et al., 1992), and KN-93 (Niki et al., 1993) have been shown to suppress insulin secretion. Experiments utilizing these pharmacological agents are, however, confused by additional effects of alloxan and KN-62 to influence glucokinase (Lenzen et al., 1988) and Ca2+ channel activity, respectively (Li et al., 1992).
The objective of the current study was to evaluate the ability of glucose to activate CaM kinase I1 in isolated islets by the induction of enzyme autophosphorylation and autonomous kinase activity. The correlation of CaM kinase I1 activation to glucose-induced insulin secretion over a similar concentration range supports the hypothesis that CaM kinase I1 plays an important role in glucose-regulated insulin secretion or related /3-cell function.

EXPERIMENTAL PROCEDURES
Materials-Male Wistar rats were purchased from Harlan Sprague-Dawley (Indianapolis, IN) and maintained on Tekland Rodent Diet (Indianapolis, IN) ad libitum for 7-10 days prior to use. CMRL-1066, glutamine, streptomycin, and fetal calf serum were purchased from Life Technologies, Inc., and Hanks' balanced salt solution was from Whittaker Bioproducts (Walkersville, MD). Ficoll, 12-0-tetradecanoylphorbol-13-acetate, ATP (disodium salt), and protein A-Sepharose were purchased from Sigma. Collagenase P was purchased from Boehringer Mannheim, and glucose (dextrose) was from the National Bureau of Standards (Gaithersburg, MD). Radiochemicals (i.e. [3zP] orthophosphoric acid and [y3*P]ATP) were purchased from DuPont NEN. Autocamtide-2, sequence KKALRRQETVDAL (Hanson et al., 1989), was synthesized by Bio-Synthesis, Inc. (Lewisville, TX). Rabbit anti-rat brain CaM kinase IIa polyclonal antibody was a gift from Dr. Paul Kelly (University of Texas Health Science Center, Houston, TX). All other chemicals were of the finest reagent grade available.
Isolation of Islets-Pancreatic islets were isolated from male rats by collagenase P digestion and subsequent enrichment by centrifugation on a Ficoll gradient as described previously (Johnson et al., 1992;McDaniel et al., 1984). Following isolation, islets were cultured in CMRL-1066 containing 5.5 mM glucose and supplemented with 1% L-glutamine, 10% heat-inactivated fetal calf serum, 50 units/ml streptomycin, and 100 pg/ml penicillin under an atmosphere of 95% air, 5% COP until use the same day. Preparation of Anti-CaM Kinase ZZ Antibody-Protein A Conjugate-Anti-CaM kinase I1 antibody was conjugated to preswelled, washed protein A-Sepharose by incubation at a ratio of 10.5 pl of antisera/4 mg of dry agarose for 16 h at 4 'C with constant agitation. The conjugate was washed five times prior to use.
Immunoprecipitation of CUM K i m e ZZ-Isolated islets were washed three times in Krebs-Ringer Hepes buffer (25 mM Hepes pH 7.4,115 mM NaCl, 24 mM NaHC03, 5 mM KCl, 2.5 mM CaCl2, 1 mM MgC12) containing 3 mM glucose and 0.1% bovine serum albumin (KRB basal medium). Four hundred islets were individually selected under a stereomicroscope (Meiji, San Jose, CA) and placed in polyallomer microcentrifuge tubes. The bathing medium was removed using a drawn out Pasteur pipette and replaced with KRB basal medium (300 pl) containing 400 pCi of [32P]orthophosphoric acid. The islets were preincubated for 90 min at 37 'C under an atmosphere of 95% air, 5% COS with gentle agitation every 30 min to resuspend the islets.
Incubations were then initiated by the addition of 131 pl of KRB basal medium (control) or KRB containing stimulatory concentrations of glucose or KC1 to yield final concentrations of 28 and 40 mM, respectively. In the latter case, appropriate adjustments (i.e. the elimination of a compensatory concentration of NaCl) were made to maintain iso-osmotic balance. Incubations were continued for 2.5 min at 37 "C. Reactions were terminated by brief centrifugation (5 s), aspiration of incubation medium, and immediate freezing in solid COZ. Islets were then lysed by sonication (10 pulses, setting 3, 30% duty cycle; Branson Ultrasonics, Danbury, CT) in 100 pl of ice-cold homogenization buffer (50 mM MES, pH 7.2, 250 mM sucrose, 1 mM EDTA, 50 mM sodium pyrophosphate, 50 mM NaF). An equal volume (100 pl) of a second buffer (100 mM Tris-HC1, pH 7.2, 150 mM NaC1, 50 mM sodium pyrophosphate, 50 mM NaF, 0.1% SDS, 1% Triton X-100, 1% deoxycholate) was immediately added and the homogenate vortexed and then centrifuged for 2 min (4 "C) at 8,000 X g. Unincorporated nucleotides were removed by repeated (2 times) concentration of the supernatant using a Centricon-30 column (Amicon Inc., Beverly, MA) and subsequent washing in homogenate buffer (200 pl). Immunoprecipitation from the retentate was performed by the addition of 200 pl of anti-CaM kinase I1 antibody-protein A complex (see above) and constant rotation on a spin wheel for 2 h at 4 "C. CaM kinase 11-antibody conjugate was sedimented by centrifugation (Sorvall Microspin; 12,000 X g, 5 min) and washed 4 times in homogenate buffer and 1 time in 20 mM NaHPO,, pH 7.2, 2.0 mM EDTA, 0.01% SDS, 0.05% Tween 20 buffer by resuspension and centrifugation. After the final wash, 50 pl of SDS sample buffer (186 mM Tris-HC1, pH 6.7, 9 mM SDS, 6 mM 2-mercaptoethanol, 15% glycerol, 0.01% bromphenol blue) was added and the sample boiled for 2 min. Dissociated protein A-Sepharose was removed by centrifugation and a portion (15-25 pl) of the supernatant of each sample subjected to SDS-polyacrylamide gel electrophoresis on a 10% gel. The dried gel was developed by autoradiography.
Assay of CUM Kinase ZZ Activity-CaM kinase I1 activity was assayed by modification of previously described methods (Waldman et al., 1990;MacNicol et al., 1990;MacNicol and Schulman, 1992;Jefferson et al., 1991). Islets were washed and counted into polyallomer tubes (400/tube) as described above. Islets were then preincubated in 500 pl of basal KRB medium (3 mM glucose) for 10 min at 37 "C. Incubations were initiated by the removal of the bathing medium and replacement with KRB medium containing basal or stimulatory concentrations of glucose, 8-28 mM, or 40 mM KCl. Incubations were continued for 0-20 min at 37 "C in a shaking water bath. To terminate the incubation, the islets were briefly (5 8 ) sedimented by centrifugation and the bathing medium removed and collected for the assay of insulin content. Ice-cold homogenization buffer (20 mM Tris-HC1, pH 7.5, 0.5 mM EGTA, 1.0 mM EDTA, 2.0 mM dithiothreitol, 10 mM sodium pyrophosphate, 0.4 mM ammonium molybdate, 100 pg/ml leupeptin, 200 pl) was immediately added to the islets and the sedimentation process repeated. The buffer supernatant was discarded. Fresh homogenization buffer (75 pl) was then added and the islets lysed by sonication (10 pulses, setting 3, 30% duty cycle). The resultant homogenate was used immediately for the assay of CaM kinase I1 activity. The protein concentration of the homogenate ranged from 1.5 to 2.2 mg/ml. CaM kinase I1 activity was assayed in a reaction mixture containing 50 mM PIPES, pH 7.0, 10 mM M&lZ, 0.1 mg/ml bovine serum albumin (fraction V), 10 p~ autocamtide-2, 20 p M ATP (specific activity, 40 Ci/mmol), and either 0.5 mM CaClZ, 5 pg/ml calmodulin for Ca2+-stimulated activity or 0.9 mM EGTA for Ca2+-independent activity. Total reaction volume was 50 pl. The assay was initiated by the addition of 10 pl of islet homogenate and continued for 30 s at 30 "C. The reaction was terminated by the addition of ice-cold trichloroacetic acid (25 pl, 15%). Tubes were placed on ice for 20 min to precipitate large proteins, which were then sedimented by centrifugation for 1 min at 12,000 X g (Sorvall Microspin). Thirty-five pl of the resulting supernatant was spotted onto 5-cm by 2-cm strips of phosphocellulose paper (Whatman P-81). Strips were washed 5 times in 500 ml of distilled H20, dried at 110 "C for 15 min, and 32Pi incorporation into autocamtide-2 determined by Cerenkov radiation (Beckman Instruments). Initial experiments were performed to established conditions for linearity of activity with respect to protein concentration and time. Total CaM kinase I1 activity in islet homogenates by this method was calculated as approximately 3 nmol of Pi incorporated per min/mg of protein.
In the described experiments, 32Pi incorporation into autocamtide-2 in the absence of Ca2+/calmodulin (autonomous CaM kinase I1 activity) is expressed as a percentage of incorporation in the presence of these cofactors (Ca2+-dependent CaM kinase I1 activity).
Insulin Secretion-For static secretion experiments, islets were counted (20/tube) into 12 X 75-mm siliconized borosilicate tubes and preincubated for 30 min at 37 "C with gentle shaking in KRB basal medium containing 3 mM glucose and 0.1% bovine serum albumin (200 pl) under an atmosphere of oxygen/C02 (95:5%). The medium was replaced with fresh KRB basal medium alone or supplemented with stimulatory concentrations of glucose (8, 11, 17, or 28 mM) and the incubation continued for a further 30 min. The incubation was terminated by the removal of the medium. Insulin content of incubation media was determined by a double antibody radioimmunoassay (Morgan and Lazarow, 1963).
Protein Determination-Protein concentration was estimated by the method of Bradford (1976) using bovine serum albumin as standard.
Statistical Treatment ofData-Statistical significance was assessed by Student's t test.

RESULTS
Autophosphorylation of CaM Kinase 11-In vitro studies in neuronal tissues have established that the activation of CaM kinase I1 is accompanied by enzyme autophosphorylation and the resultant generation of autonomous kinase activity (Hanson and Schulman, 1992). Other studies have extended these observations to demonstrate that this is a valid determinant of enzyme activation in situ (Gorelick et al., 1988;Fukunaga et al., 1989;MacNicol et al., 1990;Jefferson et al., 1991;Ocorr and Schulman, 1991). CaM kinase I1 autophosphorylation occurs in PC12 cells in response to bradykinin (MacNicol et al., 1990) and in both PC12 and GH3 cells in response to depolarizing concentrations of K' (MacNicol et al., 1990;Jefferson et al., 1991). In the current study, an immunoprecipitation procedure using polyclonal anti-CaM kinase IIa antibodies was employed to determine whether the phosphorylation status of CaM kinase I1 is increased in islets exposed to stimulatory concentrations of the insulin secretagogues, K' (40 mM) and glucose (28 mM). Such concentrations of K' induce &cell depolarization and subsequent influx of Ca2+, which results in a marked, but transient, increased insulin secretion response (Henquin and Lambert, 1974). As such, K' is thought to mimic the cellular effects of glucose to open voltage-dependent Ca2+ channels. It was reasoned that such an increase in intracellular concentration of Ca2+ should be sufficient to promote the activation of CaM kinase I1 and was therefore used as a positive control with which the effects of glucose could be compared.
Islets were prelabeled with [32P]orthophosphoric acid as described under "Experimental Procedures" and subsequently stimulated by K' (40 mM) or glucose (28 mM). Control incubations using KRB basal medium (3 mM glucose, 5 mM K' ) were conducted in parallel. After stimulation for 2.5 min, CaM kinase I1 was immunoprecipitated from islet homogenates and subjected to electrophoresis and autoradiography. As demonstrated in Fig. 1, a phosphoprotein of M, -54,000, consistent with that expected of the a-subunit of CaM kinase I1 (CaM kinase IIa) (Lin et al., 1987), was immunoprecipitated by these procedures. Radioactive phosphate (32Pi) incorporation into this protein was increased by stimulation of the islets with either 40 mM K' or 28 mM glucose as demonstrated by the increased intensity of this band by autoradiography. Quantification of 32Pi incorporation by densitometry (Optimus Imager using BIOMED software) demonstrated that K' (40 mM) and glucose (28 mM) increased phosphate incorporation into CaM kinase I1 by 175 f 15 and 189 f 19%, respectively. These data are consistent with secretagogueinduced autophosphorylation of CaM kinase I1 in islets.
Generation of Autonomous CaM Kinase 11 Activity-A hallmark of autophosphorylated CaM kinase I1 is the possession of Ca2+/calmodulin-independent (autonomous) kinase activity (Schulman and Hanson, 1993). Therefore, in order to confirm that K+-and glucose-induced enzyme phosphorylation was the result of enzyme autophosphorylation, the ability of these secretagogues to induce autonomous enzyme activity in islets was assessed. Autonomous CaM kinase I1 activity was assayed in islet homogenates using a synthetic peptide, autocamtide-2, incorporating the autophosphorylation sequence (RQETVD) of the a-subunit of rat brain CaM kinase I1 (Hanson et al., 1989).

Potassium-induced Increase in Autonomous CaM
Kinase II Activity-The effects of K' on CaM kinase I1 activation were studied over a time period in which K'-induced insulin secretion occurs (0-20 min) (Henquin and Lambert, 1974). As illustrated in Fig. 2 activity in K+-stimulated islets increased 2.9 f 0.4-fold ( n = 6) relative to control (3 mM glucose, 5 mM K+) autonomous activity (Fig. 2). Autonomous CaM kinase I1 activity peaked at 5 min at 3.0 f 0.5-fold of control ( n = 8) and remained elevated relative to control at 10 min a t 2.7 k 0.2-fold of control ( n = 8), respectively, before returning to the basal level at 20 min. This effect in islets was very similar to the effects of K+ on the activation of CaM kinase I1 observed in PC12 (MacNicol et al., 1990) andGH3 cells (Jefferson et al., 1991). In this experiment and the subsequent experiments described, the total amount of CaM kinase I1 activity in islet homogenates remained constant. These data demonstrate that K+ stimulates CaM kinase I1 activity in isolated islets, and this activation is likely due to increased Ca2+ influx as the result of cell depolarization.
Glucose-induced Increase in Autonomous CaM Kinase 1 1 Actiuity-The principal objective of this study was to determine whether glucose activates CaM kinase I1 in isolated islets. Using conditions established in the previous experiments, the ability of glucose to promote the generation of autonomous CaM kinase I1 activity was assessed. Glucose at a concentration that maximally stimulates insulin secretion from isolated islets (28 mM) induced a marked increase in autonomous CaM kinase I1 activity (Fig. 3A). In basal conditions, autonomous CaM kinase I1 activity represented 4.2 f 1.6% of activity achieved in the presence of Ca2+ and calmodulin; this value did not vary significantly over the time period studied (0-20 min). In a manner similar to K', glucose induced a rapid increase in autonomous CaM kinase I1 activity, which, at peak stimulation a t 2.5 min, was 11.0 f 0.9% of Ca2+/calmodulin-dependent activity (2.9 k 0.2-fold induction over 3 mM glucose control, n = 4). This stimulation was similar quantitatively to the stimulation of enzyme phosphorylation induced by glucose (Fig. 1) and autonomous CaM kinase I1 activity induced by depolarizing concentrations of K+ (Fig. 2). As was the case in K+-stimulated islets, autonomous activity in glucose-stimulated islets remained elevated relative to control over 5-10 min (Fig. 3A). Autonomous CaM kinase I1 activity in these islets was not significantly different from control at 20 min. Cumulative glucose (28 mM)-induced insulin secretion, monitored in these islet incubations, increased steadily throughout the period of the study relative to control (3 mM glucose) (Fig. 3B).
Correlation of Glucose-induced Autonomous CaM Kinase 1 1 Islets (400/tube) were incubated with glucose (28 mM, 0 ) or control (3 mM glucose, 0) for the times indicated. A, Ca2+-dependent and autonomous CaM kinase I1 activity was calculated as described under "Experimental Procedures." Autonomous activity is expressed as a percentage of Ca2+-dependent activity. B, insulin content of incubation medium from the same islets was determined by radioimmunoassay. Values are means f S.E. for three to six separate determinations except where indicated. *, p < 0.004; **, p < 0.007; ***, p < 0.010 (unpaired Student's t test).
Activity and Insulin Secretion-To further characterize the effect of glucose to activate CaM kinase 11, islets were incubated at the optimal time point (2.5 min) in medium containing increasing concentrations of the nutrient secretagogue (3-28 mM). Glucose dose dependently stimulated the production of autonomous CaM kinase I1 activity (Fig. 4A). The threshold of activation occurred between 8 and 11 mM glucose, and, at a concentration considered to be near maximal for glucose (28 mM) (Wollheim and Sharp, 1981), activation achieved was 2.68 f 0.15-fold of control. This sigmoidal-like relationship correlated very closely with glucose-induced insulin secretion from isolated islets (Fig. 4B). Using values at 28 mM glucose as an approximation of maximal activation, the concentrations of glucose required to produce half-maximal increases were 14 and 17 mM for insulin release and CaM kinase I1 activation, respectively.
Effect of Mannoheptulose-To further determine whether the activation of CaM kinase I1 by glucose was dependent on the metabolism of glucose, the ability of mannoheptulose, which is known to suppress glucose oxidation by the inhibition of glucokinase (Ashcroft et al., 1970), to prevent this effect was studied. In these experiments, glucose (17 mM) induced a 1.6 f 0.2-fold activation of CaM kinase I1 relative to control (3 mM glucose) as determined by the appearance of autonomous kinase activity. Mannoheptulose (25 mM), when added in addition to glucose (17 mM), completely prevented glucoseinduced activation of CaM kinase 11; under these conditions, autonomous CaM kinase I1 activity was 68 -t 3% of control (3 mM glucose alone). These data suggest that glucose-induced activation of CaM kinase I1 is absolutely dependent on its metabolism and, therefore, support a specific effect of glucose. Furthermore, since insulin secretion induced by glucose is similarly obliterated by co-incubation with mannoheptulose (Ashcroft et al., 1970), the mechanisms in the regulation of secretion and activation of CaM kinase I1 likely involve the same components.

DISCUSSION
CaM kinase I1 has long been implicated in the regulation of insulin secretion, but a definitive evaluation of this hypothesis has been hindered by the lack of information concerning the function and regulation of this enzyme in the @cell. In particular, the endogenous substrate(s) for CaM kinase I1 in islets is (are) not known. In other pharmacological approaches, some putative inhibitors of CaM kinase I1 have proven not to have desired selectivity. Moreover, until recently, there have not been methods of sufficient sensitivity to detect the activation of islet CaM kinase I1 in situ in response to insulin secretagogues. The recent surge of information regarding the regulation of CaM kinase I1 by autophosphorylation and generation of autonomous CaM kinase I1 activity has provided insight into possible cellular functions of the enzyme and has further allowed an experimental means for the assessment of its activation in intact cells.
The principal observation of this study was that glucose induced the increased phosphorylation of an cy-like subunit of CaM kinase I1 and the generation of autonomous CaM kinase I1 activity in isolated islets. These observations provide the first clear evidence that glucose activates CaM kinase I1 in isolated islets. The increased phosphorylation of CaM kinase I1 is thought to represent enzyme autophosphorylation since: 1) autonomous kinase activity is a hallmark characteristic of CaM kinase I1 phosphorylated at the "autonomy" site (e.g. Thr-286 in the a-subunit); and 2) the effect of glucose to increase enzyme phosphorylation and the generation of autonomous activity was quantitatively similar and was mimicked closely by stimulatory K'. Depolarizing K+ has previously been shown to induce autophosphorylation of CaM kinase I1 in PC12 (MacNicol et al., 1990) and GH, cells (Jefferson et al., 1991). A definitive demonstration of CaM kinase I1 autophosphorylation in response to glucose would require the identification, by phosphopeptide analysis, of the phosphorylation at a site that corresponds to the autonomy site on rat brain kinase. This was not attempted in this study because of the low amount of material and subsequent signal generated from isolated islets.
That the activation of CaM kinase I1 occurs in the @-cell of the islet is suggested in preliminary studies that demonstrate that depolarizing K' and glyceraldehyde induce similar enzyme autophosphorylation in the clonal @-cell line, RINm5F (Landt, 1992). Furthermore, the immunoprecipitation of CaM kinase I1 further provides immunological evidence for the presence of this enzyme in isolated islets (this study) and pcells (Landt, 1992) and complements enzymatic evidence reported previously. The precise molecular structure and isoenzymic form of CaM kinase I1 possessed by the @-cell of the islet has yet to be determined.
The activation of CaM kinase I1 by glucose and K+ as measured by the generation of autonomous CaM kinase I1 activity was essentially identical. Thus, the maximal effects of both secretagogues were quantitatively similar (approximately %fold in each case) and were achieved after the same exposure time (2.5-5 min). Furthermore, autonomous activity persisted in glucose-and K+-treated islets for a period of at least 10 min. These observations suggest that the mechanism by which these secretagogues activate CaM kinase I1 involve similar or identical components and are consistent with the currently proposed mechanism of glucose-induced insulin secretion. Depolarization of the @-cell by elevated extracellular concentrations of K+ promotes Ca2+ influx through the opening of L-type Ca2+ channels (Prentki and Matschinsky, 1987). Similarly, glucose is thought to depolarize the cell as the result of its metabolism and subsequent closing of KAT^ channels (Ashcroft, 1988;Cook et al., 1988). That glucose metabolism is required for the activation of CaM kinase I1 was demonstrated by the ability of the glucokinase inhibitor, mannoheptulose, to prevent glucose generation of autonomous activity. A striking observation from this study was the close correlation in islets of the concentration dependence of glucoseinduced insulin secretion with CaM kinase I1 activation. The threshold of activation was estimated at between 8 and 11 mM in each case, and the glucose concentrations required to elicit a half-maximal stimulation (14-17 mM) were similar. Such a close correlation suggests that similar biochemical mechanisms lead to the activation of enzyme and secretion but further argues for a role of CaM kinase I1 in the mediation of glucose-induced insulin secretion. A temporal correlation of CaM kinase I1 activation with insulin secretion is, however, less obvious. The rapid activation of CaM kinase I1 by glucose (maximum at 2.5 min) correlates with the first phase of glucose-induced insulin secretion, which is transient and also peaks approximately 2.5-5 min after initial exposure of the islet to glucose (Easom et al., 1990). A tentative assignment of CaM kinase I1 to the regulation of the first phase of secretion is further supported in this study by the essentially identical activation of this enzyme by K'. Depolarizing K' (Henquin and Lambert, 1974) and other secretagogues that result in transient influx of Ca2+ such as tolbutamide (Henquin, 1980) and the Ca2+ ionophore A23187 (Zawalich et al., 1983) elicit a transient secretion response reminiscent of the first phase of glucose-induced insulin secretion. The interpretation that CaM kinase I1 may function to mediate the first phase secretion is, however, difficult to reconcile with the results of a recent report that demonstrated that the CaM kinase I1 inhibitor, KN-62, failed to inhibit Ca2+-induced insulin secretion from detergent-or electro-permeabilized @cells (HIT) (Li et al., 1992). Ca2+ induces a transient insulin secretion response in permeabilized islets resembling first phase secretion and induces the phosphorylation of a 54,000 molecular weight protein likely to be CaM kinase I1 (Jones et al., 1992). We have since confirmed this observation in electro-permeabilized islets indicating that this result is not due to unique characteristics of HIT cells,3 although there remains the possibility that insulin secretion is fundamentally altered in the permeabilized @-cell.
As an alternate possibility, the persistence of autonomous CaM kinase I1 activity beyond the period of first phase secretion could be active in the initiation and/or propagation of the second phase of secretion. Such a role could be mediated by the proposed cross-talk of CaM kinase I1 with other cellular pathways, such as a protein kinase C pathway (MacNicol and R. M. Wenham, L. C. Craig, and R. A. Easom, unpublished observations. Schulman, 1992). The activation of P-cell CaM kinase I1 may represent a mechanism whereby the message provided by the initial burst of Ca2+ influx is prolonged, as has been proposed as a function of CaM kinase I1 in other cell types (Schulman and Hanson, 1993;MacNicol et al., 1990). A third alternative is that the activation of CaM kinase I1 plays an important role in other @-cell functions that are coordinately regulated with secretion in response to glucose. It is anticipated that the identification of the intracellular substrates for CaM kinase I1 in islets will provide important clues as to the precise role of this enzyme in the regulation of P-cell function by glucose.
In summary, the data presented in this study clearly demonstrate that glucose activates CaM kinase I1 in isolated rat islets and that the extent of activation correlates closely with insulin secretion. These data suggest an important role of CaM kinase I1 in the glucose regulation of P-cell function related to insulin secretion.