Suppression of adenylate kinase catalyzed phosphotransfer precedes and is associated with glucose-induced insulin secretion in intact HIT-T15 cells.

Adenine nucleotide metabolism was characterized in intact insulin secreting HIT-T15 cells during the transition from non-stimulated (i.e. 0.2 mM glucose) to the glucose-stimulated secretory state. Metabolic dynamics were monitored by assessing rates of appearance of 18O-labeled phosphoryls of endogenous nucleotides in cells incubated in medium enriched in [18O]water. Most prominent of the metabolic alterations associated with stimulated insulin secretion was the suppression in the rate of adenylate kinase (AK)-catalyzed phosphorylation of AMP by ATP. This was manifest as a graded decrease of up to 50% in the rate of appearance of β-18O-labeled species of ADP and ATP and corresponded to the magnitude of the secretory response elicited over a range of stimulatory glucose concentrations. The only nucleotide exhibiting a significant concentration change associated with suppression of AK activity was AMP, which decreased by about 50%, irrespective of the glucose concentration. Leucine-stimulated secretion also decreased the rate of AK-catalyzed phosphotransfer. This secretory stimulus-related suppression of AK-catalyzed phosphotransfer occurs within 45 s of glucose addition, precedes insulin secretion, depends on the internalization and metabolism of glucose, and is independent of membrane depolarization and the influx of extracellular calcium. The secretory stimulus-induced decrease in AK-catalyzed phosphotransfer, therefore occurs prior to or at the time of K+ATP channel closure but it is not associated with or a consequence of events occurring subsequent to K+ATP channel closure. These results indicate that AK-catalyzed phosphotransfer may be a determinant of ATP to ADP conversion rates in the K+ATP channel microenvironment; secretory stimuli-linked decreased rates of AK-catalyzed ADP generation from ATP (and AMP) would translate into an increased probability of ATP-liganded and, therefore, closed state of the channel.

Adenine nucleotide metabolism was characterized in intact insulin secreting HIT-T15 cells during the transition from non-stimulated (i.e. 0.2 mM glucose) to the glucose-stimulated secretory state. Metabolic dynamics were monitored by assessing rates of appearance of 18

Olabeled phosphoryls of endogenous nucleotides in cells incubated in medium enriched in [ 18 O]water.
Most prominent of the metabolic alterations associated with stimulated insulin secretion was the suppression in the rate of adenylate kinase (AK)-catalyzed phosphorylation of AMP by ATP. This was manifest as a graded decrease of up to 50% in the rate of appearance of ␤-18 Olabeled species of ADP and ATP and corresponded to the magnitude of the secretory response elicited over a range of stimulatory glucose concentrations. The only nucleotide exhibiting a significant concentration change associated with suppression of AK activity was AMP, which decreased by about 50%, irrespective of the glucose concentration. Leucine-stimulated secretion also decreased the rate of AK-catalyzed phosphotransfer. This secretory stimulus-related suppression of AKcatalyzed phosphotransfer occurs within 45 s of glucose addition, precedes insulin secretion, depends on the internalization and metabolism of glucose, and is independent of membrane depolarization and the influx of extracellular calcium. The secretory stimulus-induced decrease in AK-catalyzed phosphotransfer, therefore occurs prior to or at the time of K ATP ؉ channel closure but it is not associated with or a consequence of events occurring subsequent to K ATP ؉ channel closure. These results indicate that AK-catalyzed phosphotransfer may be a determinant of ATP to ADP conversion rates in the K ATP ؉ channel microenvironment; secretory stimulilinked decreased rates of AK-catalyzed ADP generation from ATP (and AMP) would translate into an increased probability of ATP-liganded and, therefore, closed state of the channel.
Although there has been considerable progress in developing an understanding of how a glucose signal is transduced to elicit an insulin secretory response in pancreatic islets (for review, see Ref. 1), some major aspects of the overall mechanism have not been elucidated. Well established is an absolute requirement for glucose to be metabolized, this leads to membrane depolarization due to a decreased conductance of K ϩ by ATPsensitive K ϩ (K ATP ϩ ) channels in the ␤-cell plasma membrane. This membrane potential change results in the influx of Ca 2ϩ via L-type voltage-dependent Ca 2ϩ channels. How the metabolism of glucose is coupled to bringing about an increased frequency of K ATP ϩ channel closures is not known nor is the mechanism by which K ATP ϩ channel behavior is controlled. From in vitro studies it has been established that the K ϩ conductance by this channel is suppressed when it is liganded with ATP which increases the probability of its "closed" status (2)(3)(4)(5)(6); when liganded with ADP the "open" state predominates and the K ϩ conductance increases (7,8). How the transition from ATP-to ADP-liganded status of the channel is achieved is not understood.
One currently held view of how glucose effects a more closed (ATP-liganded) state of this channel is through changing the intracellular concentration of ATP or the ATP/ADP ratio (for reviews, see Refs. 1 and 6). The basic premise is that by enhancing glycolytic flux, cytosolic ATP concentration increases and this promotes ATP liganding to the K ATP ϩ channels. Opposition to this concept is severalfold. Ghosh et al. (9) found no significant changes in ␤-cell ATP concentration or that of any other adenine nucleotide when they examined nucleotide levels during glucose-induced insulin secretion in a perfused rat pancreas system. The concept can also be challenged on theoretical grounds. For example, the intracellular ATP concentration (e.g. 3-5 mM) is over 100-fold greater than the K i ATP value for K ATP ϩ channels (e.g. 15 M (2)) and whether any additional increase of an apparently saturating ATP concentration would alter the liganded status of the channel can be seriously questioned. Additionally, the rate of ATP generation is generally conceded to be governed by its rate of utilization rather than driven by the availability of a metabolizable substrate.
Since altered ATP and/or ADP concentrations are not readily detectable nor correlated with secretory stimulus-induced changes in K ATP ϩ channel operation, we reasoned that the dynamic transitions of the open/closed states of the channel may also be related to a dynamic rather than a static characteristic of adenine nucleotide metabolism. This was examined by assessing the kinetic behavior of adenine nucleotide metabolism in intact HIT T-15 cells stimulated to secrete insulin by glucose or other secretagogues. Enzyme-catalyzed phosphotransfer velocities were monitored by measuring [ 18 O]phosphoryl exchange rates (10). HIT-T15 cells, an SV-40 transformed Syrian hamster pancreatic ␤-cell line (11) were chosen because they: 1) secrete insulin in response to glucose, sulfonylureas, and other metabolic fuel secretagogues (11)(12)(13), 2) possess K ATP ϩ channels with characteristics similar to these channels in isolated pancreatic ␤ cells (14,15), and 3) provide a sufficient cell mass to permit analysis by the [ 18 O]phosphoryl oxygen exchange procedure.
The results show that stimulus-induced insulin secretion is associated with a marked and glucose concentration-dependent suppression of AK 1 -catalyzed phosphoryl transfer manifest as a reduced rate of AMP phosphorylation by ATP which translates into a decreased rate of ATP conversion to ADP. This occurs when K ATP ϩ channel conductance is predicted to be diminished and could account for extending the duration of the ATP-liganded state of the K ATP ϩ channel or a closely related regulatory component.

EXPERIMENTAL PROCEDURES
HIT Cell Cultures-HIT cells were grown and maintained in RPMI 1640 culture media supplemented with 10% fetal bovine serum, under 5% CO 2 , 95% O 2 at 37°C, as described previously (16). All studies were performed on HIT cell passages between 70 and 75 which have been previously shown to secrete insulin in response to glucose (17). HIT cells were subcultured at a density of 15-20 ϫ 10 6 cells in 100-mm Corning culture dishes 2-3 days before each study. Sixteen hours before each experiment the RPMI 1640 culture media was exchanged with fresh culture media.
Labeling of Endogenous Nucleotide Phosphoryls with 18  O]water-enriched KRB and immediate addition of ice-cold 0.5 M perchloric acid. While on ice cells were scraped from the surface, transferred along with the perchloric acid to a test tube, and then sonicated. The acidified sonicated cell suspension was centrifuged at 14,500 ϫ g for 10 min to remove precipitated protein. These acid extracts were then neutralized with 2 M KHCO 3 , the precipitated KClO 4 was removed by centrifugation, and the supernatant was evaporated to dryness in a SpeedVac (Savant). The protein pellet was dissolved in 1 M NaOH and the protein concentration was determined by the BCA method (Pierce). Cellular concentrations of AMP, ADP, ATP, and creatine phosphate were determined by enzymatic fluorometric analysis (18). The cellular levels of ATP, GTP, UTP, and ADP were also determined by UVabsorption upon their elution from Mono Q high performance liquid chromatograph.
Purification and Isotopic Analysis of 5Ј-Nucleotide Phosphoryls-The purification and analytical procedure for determining the 18 O in the phosphoryls of the 5Ј-nucleotides, orthophosphate, and creatine phosphate has been previously described (10). The only modification of this procedure was the use of a Mono Q HR 5/5 FPLC column equilibrated with triethylammonium bicarbonate, pH 8.8, instead of AG MP-1 chromatography for the purification of the 5Ј-nucleotides. All the 5Ј-nucleotides bind to Mono Q resin in 10 mM triethylammonium bicarbonate and are sequentially eluted by increasing the triethylammonium bicarbonate concentration to 1 M.
Presentation of Experimental Results-The appearance of 18 O in the phosphoryls of the 5Ј-nucleotides is presented as the percentage of phosphoryl oxygens that have been replaced with 18 O during the indicated time of incubation. The percentage of nucleotide phosphoryl oxygens replaced by 18  Insulin release, lactate production, and nucleotide levels are presented as the mean Ϯ S.D. Lactate production is presented as the sum of nanomoles of cellular lactate plus the nanomoles of lactate determined in the extracellular media. Statistical significance was determined by Student's t test.

Glucose-induced Insulin Secretion and Lactate Production by
HIT Cells-HIT T-15 cells, a clonal cell line of Syrian hamster pancreatic islet ␤ cells (11), secrete insulin in response to glucose (and other secretagogues) in a concentration-dependent manner. These cells are about 10 times more sensitive to glucose than normal islets. Although the reason for this sensitivity difference has not been defined it has been suggested to result from an alteration in glucose transport (20,21) and/or an expanded role for hexokinase-mediated glucose utilization (22). This transformed cell line behaves identically with islet ␤ cells in almost all other respects. Fig. 1A shows the insulin secretory response in HIT cells at 5 min as a function of glucose concentration. Half-maximal and maximal secretion occur at approximately 1.6 and 2.8 mM glucose, respectively. The time course of insulin secretion and lactate production by a nearly maximal stimulatory (2.8 mM) compared to a minimally stimulatory (0.2 mM) concentration of glucose is shown in Fig. 1, B and C. With 0.2 mM glucose as the stimulus there is a small increase in the insulin as well as lactate generation but only during the first 1 to 2.5 min not thereafter. In the presence of 2.8 mM glucose, insulin secretion is increased 12-fold relative to the 1-min time period by 2.5 min and this stimulated rate is sustained for at least the next 2.5 min. Enhanced secretion although at a submaximal rate occurs for the ensuing 35 min with 2.8 mM glucose but not with 0.2 mM glucose in the medium (not shown). An increase in lactate production is measurable at 45 s which precedes the detectable increase in insulin secretion occurring sometime after 1 min. These experiments were carried out under conditions identical to those in which glucose-induced changes in nucleotide metabolism were examined.
Cellular Nucleotide Levels during Glucose-induced Insulin Secretion-HIT cells were incubated with increasing glucose concentrations for 5 min, after an initial 71-min preincubation in KRB containing 0.2 mM glucose. After acid extraction 5Јnucleotides concentrations were determined by enzymatic fluorometric assay (18). In these experiments, increasing the glucose concentration from 0.2 to 2.8 mM led to increased insulin release from 131.3 Ϯ 15.5 to 359.8 Ϯ 19.6 microunits of insulin/mg of protein (not shown). During this 5-min incubation with glucose concentrations from 0.2 to 2.8 mM there was no apparent change in cellular ATP concentration (Table I). The concentrations of the individual nucleotides GTP, UTP, CTP, and of creatine phosphate were also found to be unchanged (not shown). Although there appeared to be a decrease of 15-30% in ADP levels (Table I) which resulted in an apparent increase in ATP to ADP ratios (Table I) neither of these changes were determined to be statistically significant nor did they show any gradations that corresponded to the magnitude of the glucose stimulus. The cellular concentration of AMP exhibited a consistent and statistically significant decrease of approximately 50% when HIT cells were incubated with glucose concentrations greater than 0.2 mM (Table I). This decline in AMP concentration corresponds with the observation to be described below of a glucose-induced suppression of the rate of adenylate kinase-catalyzed phosphorylation of AMP. However, this decrease in AMP concentration also did not exhibit any gradations related to the glucose concentration.
Dynamics of ATP and ADP Metabolism during Glucose-in-    crease in cellular ATP concentration because the increased rate of P i labeling with 18 O is indicative of an increased rate of ATP hydrolytic consumption which is undoubtedly followed by a commenserately increased rate of ATP regeneration to provide for the constancy of the overall cellular ATP concentration.
In contrast to the increase in the rates of [␥]ATP and P i labeling, the transition to the higher glucose concentration decreased the rate of 18 O-labeled ␤-phosphoryl appearance in both ATP and ADP (Fig. 3, A and B). The decrease in [␤]ATP and [␤]ADP 18 O-labeling occurs within 45 s (Fig. 3) and precedes enhanced insulin secretion which occurs after 1 min (Fig.  1B) coincident with the increase in lactate production (Fig. 1C). The appearance of 18 O-labeled ␤-phosphoryls in ADP arises as a result of AK-catalyzed transfer of isotopically labeled ␥-phosphoryls of endogenous ATP to AMP and then it can appear as the ␤-phosphoryl of ATP as a result of the subsequent phosphorylation of [␤- 18 18 O min Ϫ1 mg of protein Ϫ1 or by 48% with a 2.8 mM glucose stimulus. These estimates of absolute phosphotransferase velocities require three successive temporally displaced measurements which were obtained in the experiment for which these velocities were calculated. AK catalysis in subsequent experiments were assessed by the percentage of the total cellular ADP and ATP with 18 O-labeled ␤-phosphoryls without modeling the results and therefore without correcting for the enhanced 18 O labeling of [␥]ATP. Therefore, the percentage 18 O labeling values reported for the subsequent experiments minimize the magnitude of the decrease in AK catalyzed phosphotransfer induced by glucose.
Glucose-induced changes in the rate of appearance of 18 Olabeled ␤-phosphoryls of ATP and ADP were further characterized by monitoring the appearance of 18 O-labeled ␤-phosphoryls for 2-min labeling durations representing the initial 2 min, the second 2 min, or the third 2-min interval after the transition from 0.2 to 2.8 mM glucose. Table II Table II also confirm that this glucose-induced suppression of [␤- 18 O]phosphoryl appearance in ATP and ADP occurs very early (i.e. between 0 and 2 min) after 2.8 mM glucose addition and precedes enhanced insulin secretion which was not detectable in this experiment until the 2-4-min interval.
Glucose-concentration Dependence of Suppressed AK-catalyzed Phosphotransfer Activity Relative to Stimulated Insulin Release-The suppression of [␤- 18 O]phosphoryl appearance in both ATP and ADP occurs over a very similar range of glucose concentrations that enhances insulin secretion. Fig. 4 shows that increasing the glucose concentration from 0.2 to 2.8 mM glucose, results in incrementally enhanced insulin secretion and correspondingly greater suppression in the appearance of 18 O-labeled ␤-phosphoryls in ATP and ADP during a 5-min period of glucose stimulation. Glucose-induced suppression of AK-catalyzed phosphotransfer is shown in Fig. 4C to occur decrementally over a range of glucose concentrations up to at least 2.8 mM glucose during a 1-min period of stimulation. It is important to note that insulin secretion was not enhanced during this initial 1-min period of stimulation by this range of glucose concentrations (not shown for this experiment; also see Figs. 1B and 6).

P Incorporation into Adenine Nucleotide ␤-Phosphoryls Is
Also Attenuated by Glucose-The incorporation of 32 P into cellular adenine nucleotides was monitored during glucose-induced insulin secretion from HIT cells as an independent, semiquantitative means to further investigate the glucose-induced changes in nucleotide metabolism uncovered by the 18 O labeling procedure. The time-dependent studies of the incorporation of 32 P into adenine nucleotide during glucose-induced insulin secretion were performed as described for the 18 O labeling experiments except that the KRB media contained 0.25 mCi of 32 P i . In these experiments the transition from 0.2 to 2.8 mM glucose led to a ϳ2.5-fold increase in both insulin release and lactate production by 5 min (not shown, but similar to the data shown in Fig. 1). Increasing the glucose concentration to 2.8 mM increased the rate of total 32 P i incorporation into cellular ATP (Fig. 5A). Analysis of the individual phosphoryls  Fig. 3 were obtained from the same cells used to determine insulin secretion and lactate production shown in Fig. 1, B and C, and the appearance of [␥-18 O]ATP in Fig. 2. showed that increasing the glucose concentration to 2.8 mM glucose resulted in increased incorporation of 32 P into the ␥-phosphoryl of ATP but diminished incorporation of 32 P into the ␤-phosphoryl of ATP. This confirms the observations made with the 18 O labeling procedure with respect to both the greater rate of [␥]ATP turnover and suppression of AK-catalyzed phosphorylation of AMP by ATP. The addition of 2.8 mM glucose also inhibited the incorporation of 32 P into the cellular pool of ADP (Fig. 5B). The inhibition of 32 P labeling of the ␤-phosphoryls in both ATP and ADP is detectable within 1 min after glucose addition and precedes enhanced insulin release (not shown). No 32 P incorporation was detected in the ␣-phosphoryls of ATP or ADP. These results confirm the [ 18 O]phosphoryl labeling results showing that glucose-induces a suppression of AK-catalyzed phosphotransfer.
Glucose Uptake and Metabolism Are Required for Both Glucose-induced Insulin Secretion and the Suppression of AK-catalyzed Phosphotransfer-Cytochalasin B, a glucose transport inhibitor, was used to further characterize the glucose-induced suppression of AK catalysis associated with stimulating insulin secretion. As observed previously, upon the addition of 2.8 mM glucose, HIT cells respond with a greater than 2-fold increase in both insulin secretion and lactate production, with accompanying suppression of the appearance of 18 (Table III). The addition of 10 g/ml cytochalasin B was sufficient to lower lactate production below the level observed with 0.2 mM glucose and this same level was also achieved with 2.8 mM glucose in cells treated with cytochalasin B (Table III). This inhibition of glucose transport by cytochalasin B also prevented the stimulated insulin release ordinarily observed with 2.8 mM glucose (Table III). Cytochalasin B also completely blocked the ability of 2.8 mM glucose to suppress 18 O-labeled ␤-phosphoryl appearance in ATP and ADP. Cytochalasin B treatment caused no significant changes in the cellular nucleotide concentrations (not shown).

O-labeled [␤]ATP and [␤]ADP
Iodoacetate was used to examine whether impairment of glucose metabolism would interfere with glucose-stimulated insulin secretion as well as the suppression of [␤- 18 O]phosphoryl appearance in adenine nucleotides. Iodoacetate (0.9 mM) inhibited the ability of 2.8 mM glucose to stimulate insulin secretion, which coincided with an iodoacetate-induced decrease in glycolytic rate as indicated by the suppression of lactate production (Table IV). Iodoacetate also prevented the suppression of 18 O-labeled phosphoryl appearance in [␤]ATP and [␤]ADP ordinarily seen with stimulation by 2.8 mM glucose, although the blockade by iodoacetate was not complete in the case of the ␤-phosphoryl of ATP (Table IV). Iodoacetate also reduced the ATP concentration to 60 and 30% of the control levels in cells stimulated with 0.2 and 2.8 mM glucose, respectively (not shown). These decreases in cellular ATP concentration are undoubtedly related to the inhibitory effect of iodoacetate on glycolysis but the reason for the greater decrease with the higher glucose concentration was not established.
Effect of Leucine or Arginine on Insulin Secretion and AKcatalyzed Phosphotransfer-Leucine is the most potent physiological non-glucose stimulator of insulin release. With 0.2 mM glucose in the medium, the addition of 20 mM leucine induced a 2-fold increase in insulin release without altering the rate of lactate production (Table IV). The addition of leucine also caused a decrease in AK-catalyzed phosphotransfer as indicated by the decreased appearance of [␤- 18 O]ATP and [␤- 18 O]ADP that is qualitatively similar to the effect produced by 2.8 mM glucose (Table IV). Unlike leucine, the amino acid arginine, which is a potentiator of insulin release with stimulatory concentrations of glucose greater than 0.2 mM, did not stimulate insulin release or decrease the appearance of [␤- 18 (Table IV). These results suggest that only the metabolizable (i.e. glycolysis and/or tricarboxylic acid cycle) insulin secretagogues, such as glucose and leucine, lead to suppression of AK-catalyzed phosphotransfer manifest as a reduction in [␤- 18 (Fig. 6). The addition of 40 mM KCl even after 1-min increased insulin release strikingly, presumably as a result of its action to directly depolarize the HIT cell plasma membrane causing an influx of extracellular calcium. This K ϩ -induced release of insulin occurred without affecting AK-catalyzed phosphotransfer (Fig. 6). These results indicate that the suppression of AK-catalyzed phosphotransfer elicited by a metabolizable secretagogue does not result from and therefore probably occurs prior to membrane depolarization.
The addition of the sulfonylurea, glipizide (1 M), stimulated insulin secretion both in the absence and presence of stimulatory concentrations of glucose (Table III). The mechanism of glipizide-induced insulin secretion is thought to involve a direct inhibition of the K ATP ϩ channel, leading to ␤ cell membrane  18

O in the ␥-phosphoryls of ATP and ␤-phosphoryls of ATP and ADP during glucose-induced insulin secretion
HIT cells were incubated with either 0.2 or 2.8 mM glucose for either 2, 4, or 6 min. At the specified times the incubation media was removed and either 0.2 or 2.8 mM glucose in KRB medium containing 12% atom excess of [ 18 O]water was added for 2 min. At the end of the 2-min incubation the KRB medium was removed and the cells were acid extracted and the nucleotide phosphoryls examined for 18 O content as described under "Experimental Procedures." Total insulin secretion is presented as the sum of insulin secreted in the first glucose incubation (2-  Extracellular calcium is required for glucose induced-insulin release but its involvement is believed to be manifest distal to stimulus-induced membrane depolarization. Extracellular calcium concentrations of 0, 2.5, and 5 mM were examined to determine whether the influx of extracellular calcium was responsible for the suppression of AK-catalyzed production of  (Table V). When the cells were incubated with calcium-free KRB supplemented with 5 mM EGTA the cells no longer secreted insulin in response to 0.8 or 2.8 mM glucose, however, the affect of glucose to suppress the appearance of 18 O-labeled ␤-phosphoryls in ATP and ADP in a glucose concentration dependent manner was preserved. The higher levels of insulin found in the Ca 2ϩ -free medium (EGTA supplemented) probably results from this Ca 2ϩ -free condition to permeabilize the cells. These results show that the effect of glucose to suppress AK-catalyzed phosphotransfer is not dependent on the influx of extracellular Ca 2ϩ . DISCUSSION The experimentation described here provides the first information about the dynamics of high energy phosphoryl metabolism in intact insulin secreting cells. Our results show that stimulatory concentrations of glucose cause a relatively rapid and graded suppression of AK-catalyzed phosphotransfer closely corresponding to the magnitude of the glucose stimulus Squares represent total 32 P incorporated into ATP or ADP. Circles, triangles, and diamonds represent 32 P incorporated into the ␥, ␤, or ␣ phosphoryls of the indicated nucleotides, respectively. Incorporation of 32 P into the nucleotides was determined as described under "Experimental Procedures." A representative experiment of two identical experiments that yielded similar results is shown. Each value represents the mean Ϯ S.D. of triplicate samples. and the secretory response. This suppression of AK activity precedes the release of insulin and occurs within the glucose concentration range that stimulates insulin release. This glucose-induced suppression of AK-catalyzed phosphotransfer and the release of insulin is prevented by either cytochalasin B or iodoacetate indicating that both glucose entry and metabolism are required. The attenuation in AK phosphotransfer is independent of membrane depolarization because it did not occur with either high KCl concentrations or glipizide. Moreover, exclusion of calcium from the medium did not prevent the suppression of AK activity, indicating that this event is independent of the influx of extracellular calcium subsequent to membrane depolarization. This dissection of the stimulus-secretion coupling at least provisionally identifies the K ATP ϩ channel and its greater frequency of closure to be the event most likely related to the suppression of AK-catalyzed phosphotransfer. In perifused rat islets, it has been reported that glucose stimulation also increases [ 32 P]ATP and reduces [ 32 P]ADP (23,24), suggesting that the glucose-induced suppression in AK catalytic activity is not a unique feature of the HIT cell but also occurs in the intact islet.
The mechanism by which glucose causes a change in the composition of adenine nucleotides so they may serve as effectors of the K ATP ϩ channel is poorly understood and highly controversial. Although it is tacitly assumed that the determinant is a change in the cellular ATP and/or ADP concentration there are numerous arguments that oppose this view. Whether glucose even effectively alters total ATP concentrations in the ␤ cell is controversial. For example, it has been shown that stim-ulatory concentrations of glucose increase ATP concentrations in islets only if they have been first attenuated by restricting glucose from the islet (25)(26)(27). However, when islets are maintained in non-stimulatory concentrations of glucose, and then stimulated with higher concentrations of glucose, there are marginal or no changes detected in total cellular ATP concentration (9,27,28). In addition, there are no changes in ATP/ ADP ratios in response to high glucose stimulation (9,(27)(28)(29)(30) unless HIT cells (31) or islets were previously fuel restricted (27,29,32,33). Our observations that HIT cells maintained in low glucose concentrations and exposed to stimulatory concentrations of glucose do not significantly change their intracellular levels of ATP and only marginally alter their ADP levels, is consistent with the latter observations cited (9,(27)(28)(29)(30) and do not support the concept that increases in the concentration of intracellular ATP or the ATP/ADP ratio serves as the signal for regulating the K ATP ϩ channel. As in islets (28), the only adenine nucleotide in HIT cells to undergo a significant change in response to increased glucose concentration was AMP. Although this decreased AMP level corresponds with and may be a critical determinant of the observed glucose-induced decrease in AK-catalyzed phosphotransfer, two important aspects of this altered AMP level remain undefined. What metabolic alteration underlies this decrease in AMP and why is the decrease a relatively constant 50% irrespective of the stimulatory glucose concentration, when suppression of AK catalysis is graded relative to the glucose concentration? Only speculation can be offered at this point. The only source of AMP that could account for its relatively rapid rate of AK-catalyzed phosphorylation is Until recently there has been little or no regulatory importance attributed to AK-catalyzed phosphotransfer. In intact skeletal muscle, evidence has been provided that AK operates as a high energy phosphoryl transfer system that also regulates the rate of glycolytic ATP generation so that it closely corresponds with the rate of ATP utilization by specific energy consuming processes (19,34). The feature common to the control of muscle glycolysis and nutrient-stimulated insulin secretion is that both processes are regulated by ATP and its metabolic product ADP. In the case of muscle glycolysis, specific enzymes including phosphofructokinase, aldolase, and glycer-aldehyde phosphate dehydrogenase (35)(36)(37) are allosterically inhibited by ATP and this inhibition is relieved or enzyme activities are stimulated by ADP. The adenine nucleotide-sensitive counterpart in the insulin secretory system is the K ATP ϩ channel; it is also inhibited (i.e. closed) as a result of ATP liganding and stimulated (i.e. opened) by ADP.
Within this frame of reference AK could function in the ␤ cell as a high energy phosphoryl transfer system that could also regulate the composition of adenine nucleotide species at the K ATP ϩ site. The glucose-induced changes in AK phosphoryl transfer in the HIT cell temporally correlate with the closure of the K ATP ϩ channel. In the HIT cell, 2.8 mM glucose decreased AK activity within 45 s. Eddlestone et al. (14) have reported that in the HIT cell closure of the K ATP ϩ channel commences within 1-3 min and a new steady-state of K ATP ϩ channel activity is reached within 2-9 min after the addition of glucose (14). In addition, the changes in AK activity occur within the same glucose concentration range required for insulin secretion and for closure of the K ATP ϩ channel (14). Eddlestone et al. (14) have reported that in HIT cells there is a 50% reduction in K ATP ϩ channel closure in response to 0.45 mM glucose and a maximum number of closures with 8.0 mM glucose. The link between suppression of AK-catalyzed phosphotransfer and alterations in K ATP ϩ channel conductance is further supported by the observations that leucine stimulates insulin release, decreases AK-catalyzed phosphotransfer, and leads to the closure of K ATP ϩ channel in the HIT cell (14). That AK is physically close to the K ATP ϩ channel can be concluded from the recent report that AMP addition to an ATP-inhibited K ATP ϩ channel in an isolated inside-out patch of ␤ cell membrane results in rapid opening of the channel (8). In addition, AK activity has been measured in isolated plasma membranes from HIT cells and isolated islets. 2 The activity of AK at the channel site could be envisioned to regulate the duration of the ATP-or ADP-liganded state by the rate of its catalytic action to transform ATP to ADP with AMP serving as the critical reactant (i.e. ATP ϩ AMP 3 2 ADP). The duration of the ATP-liganded state would depend on the rate that AMP is generated and made available, probably through the AK catalyzed transfer system, to the site or the microenvironment of the K ATP ϩ channel. The decreased rate of AK-cata-  lyzed phosphotransfer observed in the experiments reported here that precedes stimulation of secretion would coincide with a diminished rate of generation of AMP (i.e. 2 ADP 3 AMP ϩ ATP), transfer of AMP, and/or subsequent conversion of ATP to (2)ADP at the channel site. This would extend the duration of the ATP-liganded state and decrease the duration of the ADPliganded state, which would result in K ATP ϩ channel closure, membrane depolarization, calcium influx, and insulin secretion.