Desensitization of Mitochondrial Ca2+ and Insulin Secretion Responses in the Beta Cell*

The role of mitochondria in the desensitization of insulin secretion was investigated. In rat pancreatic beta cells, both insulin secretion and mitochondrial [Ca2+] increases were desensitized following two challenges with the mitochondrial substrate methyl succinate. In the beta cell line INS-1, similar results were observed when a 5-min interval separated two 5-min pulses. In contrast, ATP generation monitored in luciferase-expressing INS-1 cells was stimulated to the same extent during both exposures to methyl succinate. Succinate, like α-glycerophosphate, activates the electron transport chain at complex II. As a consequence, the mitochondrial membrane hyperpolarizes, promoting ATP synthesis and Ca2+ influx into the mitochondria through the uniporter. The mitochondrial desensitization was further studied in permeabilized INS-1 cells. Increasing extramitochondrial [Ca2+] from 100 to 500 nm enhanced succinate oxidation 4-fold. At 500 nm Ca2+, 1 mm succinate caused a blunted mitochondrial [Ca2+] increase upon the second, compared with the first, stimulation. These effects were mimicked by α-glycerophosphate, and there was cross-desensitization between the two compounds. Succinate hyperpolarized the mitochondrial membrane during both the first and second applications. This suggests that the uniporter itself, rather than the respiratory chain, is desensitized. These results emphasize the key role of the mitochondria not only in the stimulation of insulin secretion, but also in its desensitization.

Desensitization is a common feature of cell biology in general and of insulin secretion in particular. However, the molecular mechanism of desensitization toward nutrient stimuli is poorly understood. Nesher and Cerasi (1) first observed that successive short stimuli with glucose or arginine in the isolated perfused rat pancreas resulted in the inhibition of the insulin secretory response to the second stimulus. Insensitivity of the pancreatic beta cell to glucose was reported in pancreata taken from hyperglycemic rats (2) and is found in several diabetic animal models (3). A reduced responsiveness of the pancreatic beta cell to glucose has also been described after prolonged exposure of beta cells to hexose in vitro (4,5) or in human subjects (6). This desensitization phenomenon is distinguished from glucose toxicity, the latter being irreversible, whereas the former implies a reversible state of cellular refractoriness due to repeated exposures to an agonist (7). Desensitization can occur at any of the multiple steps coupling glucose recognition to insulin secretion, including the exocytotic process itself, as shown in permeabilized cells exposed to repeated Ca 2ϩ pulses (8).
In the pancreatic beta cell, mitochondrial metabolism plays a pivotal role in the generation of signals coupling glucose recognition to insulin secretion (9 -13). The main trigger of exocytosis is an increase in cytosolic Ca 2ϩ concentration (for a review, see Ref. 12). In addition, Ca 2ϩ controls several other cellular functions, among them mitochondrial metabolism (14 -16). An increase in mitochondrial Ca 2ϩ concentration ([Ca 2ϩ ] m ), 1 following an elevation in cytosolic Ca 2ϩ concentration ([Ca 2ϩ ] c ), participates in the activation of the respiratory chain through stimulation of Ca 2ϩ -sensitive NADH-generating dehydrogenases (15)(16)(17)(18)(19)(20). NADH and FADH 2 transfer reducing equivalents to the respiratory chain, thereby ensuring adequate ATP synthesis (15). Transfer of reducing equivalents to the electron transport chain increases the mitochondrial membrane potential (⌬⌿ m ), which enhances the driving force for mitochondrial Ca 2ϩ uptake mediated by a low affinity uniporter (21). This ⌬⌿ m -dependent Ca 2ϩ entry permits an amplification of [Ca 2ϩ ] m , relative to [Ca 2ϩ ] c , further favoring the stimulation of the aforementioned dehydrogenases (22,23). On the other hand, the hyperpolarization of the mitochondrial membrane exerts a negative feedback by lowering the oxygen consumption and the rate of H ϩ cycling (24,25). In glucose-stimulated beta cells, insulin secretion is initiated by the activation of mitochondrial metabolism, leading to an increase in [Ca 2ϩ ] c (10,26,27). Subsequently, the rise in [Ca 2ϩ ] m appears to be essential for the maintenance of metabolism-secretion coupling (12,13). The partial reduction of glucose oxidation by blockade of the [Ca 2ϩ ] c increase (17, 28) may reflect a need for permissive [Ca 2ϩ ] c levels in optimal glucose-stimulated insulin secretion (29).
Using cells stably expressing the calcium-sensitive photoprotein aequorin targeted to the mitochondria, we have previously shown that desensitization of insulin secretion is associated with a parallel loss of the [Ca 2ϩ ] m response (23). These findings and other recent studies point to a pivotal role for the mitochondria in metabolism-secretion coupling (11, 17, 20, 30 -32), not only as a relay in the metabolic cascade, but also as a primary source of an as yet unidentified factor triggering insulin exocytosis (13). The existence of this putative mitochondrial factor is further suggested by studies showing impaired glucose-stimulated insulin secretion in insulinoma cells depleted of the mitochondrial genome (33,34).
To study the involvement of the mitochondria in the desen-

EXPERIMENTAL PROCEDURES
Cell Culture-INS-1 cells were cultured in RPMI 1640 medium as described previously (23,35,36). Stable clones of INS-1 cells expressing mitochondrial aequorin (22) were established (INS-1/EK3) as detailed elsewhere (23) and cultured in the presence of 250 g/ml G418 (Promega, Madison, WI) for continuous selection of cells expressing the plasmid with the associated neomycin resistance. Clonal INS-1 lines expressing cytosolic luciferase under the control of doxycycline-dependent transcriptional transactivator were established (INS-r3-LUC7) (36). Following two successive stable transfections, resistant clones were cultured with 250 g/ml G418 and 100 g/ml hygromycin B (Calbiochem). Pancreatic islet cells were isolated by collagenase digestion from male Wistar rats weighing ϳ200 g (17) and cultured free floating in RPMI 1640 medium for 2-4 days.
Transient Transfection of Primary Cells-Rat pancreatic islet cells were isolated as described above, trypsinized, and seeded on 13-mm diameter extracellular matrix-coated coverslips (Eldan, Jerusalem, Israel) at 4 ϫ 10 5 cells/ml in RPMI 1640 medium. Two days later, the cells were transfected with 10 l of LipofectAMINE (Gibco BRL, Basel, Switzerland) and 1 g of vector encoding mitochondrially targeted aequorin as described previously (17). Three days later, the cells were used for the experiments. This transfection procedure resulted in 10 -15% of cells being transfected as judged by immunofluorescence of the hemagglutinin tag incorporated at the N terminus of aequorin (22,23).
Measurements of Luminescence and Insulin Secretion-Luciferaseor aequorin-expressing cells were seeded on 13-mm diameter coverslips 3-5 days prior to analysis and maintained in the same medium as described above except for the addition of G418 and hygromycin. For intact cell experiments, cells were seeded on plastic polyornithinetreated coverslips at a density of 4 ϫ 10 5 cells/ml. For permeabilized cell experiments, cells were seeded at 2 ϫ 10 5 cells/ml on A431 extracellular matrix-coated coverslips as described above. Prior to luminescence measurements, cells were maintained in glucose-and glutamine-free RPMI 1640 medium plus 10 mM HEPES for 2-5 h at 37°C. This period also served to load aequorin-expressing cells with 2.5 M coelenterazine (Molecular Probes, Inc., Eugene, OR), the prosthetic group of aequorin (23). Luminescence was measured by placing the coverslip in a 0.5-ml thermostatted chamber at 37°C ϳ5 mm from the photon detector. We used a photomultiplier apparatus (EMI 9789, Thorn-EMI, Middlesex, United Kingdom), and data were collected every second on a computer photon-counting board (EMI C660) prior to calibration as described previously for [Ca 2ϩ ] m (23). The cells were perifused constantly at a rate of 1 ml/min, and where indicated, 1-min fractions of the effluent were collected for insulin measurements. Suspensions of islet cells were perifused with the same buffers as INS-1 cells using a perifusion apparatus (23). Intact cells were perifused with HEPES-balanced Krebs-Ringer bicarbonate buffer (135 mM NaCl, 3.6 mM KCl, 10 mM HEPES, pH 7.4, 2 mM NaHCO 3 , 0.5 mM NaH 2 PO 4 , 0.5 mM MgCl 2 , 1.5 mM CaCl 2 , and 2.8 mM glucose) plus 10 M beetle luciferin (Promega) for luciferaseexpressing cells. Luciferase luminescence was used for the monitoring of [ATP] in living cells as described previously (36). Permeabilized cells were perifused with the intracellular buffer described above. For insulin secretion experiments, 0.1% bovine serum albumin (Sigma) was added to buffers as carrier, and insulin was measured by radioimmunoassay using rat insulin as a standard (35).
Mitochondrial Membrane Potential-⌬⌿ m was measured as described (13,39). Briefly, after a culture period in glucose-free RPMI 1640 medium, cells were loaded with 10 g/ml rhodamine 123 for 10 min at 37°C. For cell suspension measurements, after centrifugation, the cells were permeabilized as described above and transferred to a fluorometer cuvette, and the fluorescence excited at 490 nm was measured at 530 nm at 37°C with gentle stirring in an LS-50B fluorometer (Perkin-Elmer, Buckinghamshire, United Kingdom). For measurements on attached cells, the cells grown on A431-coated glass coverslips were loaded with rhodamine 123 prior to permeabilization (see above). Cells were then placed in a thermostatted microincubator (Medical Systems Corp., Greenvale, NY) on an inverted microscope (Nikon Diaphot) with a 40ϫ oil immersion objective. Fluorescence excitation was filtered at 485 nm, and emission was split at 505 nm and further filtered at 530 nm (Omega Optical Inc., Brattleboro, VT). The signal was recorded at 100 Hz with a photomultiplier (Nikon P100S) and a computerized acquisition system (40). The cell layer was perifused at 1 ml/min with the 500 nM free Ca 2ϩ intracellular buffer (see above) supplemented with 0.1 g/ml rhodamine 123.
Succinate Oxidation to CO 2 in Permeabilized INS-1 Cells-INS-1 cells were seeded at 4 ϫ 10 5 cells/2 ml on 35-mm diameter dishes coated with A431 extracellular matrix as described above. Cells were maintained 3-4 days prior to the experiment in the standard RPMI 1640 medium to subconfluency. Attached cells were then incubated in glucose-and glutamine-free RPMI 1640 medium plus 10 mM HEPES for 2 h at 37°C, transferred to a thermostatted glass chamber, and permeabilized according to the procedure described above. Cells were then washed with the corresponding intracellular buffer adjusted to either 100 or 500 nM free Ca 2ϩ and preincubated for 10 min in that buffer. Succinate oxidation was initiated by replacing the buffer with 1 ml of the respective fresh ones containing 1 mM [2,3-14 C]succinate (NEN Life Science Products; 0.1 Ci/chamber). After a 1-h incubation at 37°C in sealed chambers, 0.5 ml of 0.1 M HCl was added onto the cell layers to stop the reaction, and 1 ml of benzethonium hydroxide (Sigma) was injected into the bottom of the chamber to bind the CO 2 liberated by the cells (41). Following an overnight incubation at room temperature, 14 CO 2 production was measured in benzethonium extracted with 5 ml of EtOH and counted in an LS6500 liquid scintillation counter (Beckman Instruments).
Statistical Analysis-Where applicable, values are expressed as the mean Ϯ S.E., and significance of difference was calculated by Student's t-test for unpaired data. Traces without S.E. values are representative of at least three independent experiments.

RESULTS
Insulin Secretion in Islets-Rat pancreatic islets were maintained in culture for 2-4 days prior to the experiments. Stimulation of insulin secretion with 16.7 mM glucose for 10 min was repeated after a 10-min interval of perifusion at 2.8 mM glucose. This revealed that the secretory response was desensitized during the second stimulation, displaying ϳ50% reduction (Fig. 1A). The tricarboxylic acid cycle intermediate succinate, rendered cell-permeant by the ester binding of a methyl group (42), also produced a desensitization of the insulin exocytotic response with a pattern similar to that produced by glucose (Fig. 1B). Finally, KCl was used to raise [Ca 2ϩ ] c by membrane depolarization (29,17). Again, the second of two exposures to 20 mM KCl revealed a blunted insulin secretory response (Fig. 1C).
[Ca 2ϩ ] m in Primary Pancreatic Cells-Primary rat pancreatic cells were transiently transfected with the cDNA encoding mitochondrially targeted aequorin. Monitoring of [Ca 2ϩ ] m in these cells showed that 5 mM methyl succinate increased [Ca 2ϩ ] m during the first stimulation, but not during a second one repeated 5 min later ( Fig. 2A). This desensitization was also observed by raising [Ca 2ϩ ] c through depolarization of the plasma membrane induced by 20 mM KCl (Fig. 2B). Contrary to clones stably expressing aequorin, the low expression levels after transient transfection (13) do not permit a reliable calibration since the total photon emission was 10 -20-fold less in the later case. Therefore, [Ca 2ϩ ] m is expressed as photons emitted per second. insulin secretion demonstrated that both parameters exhibited an attenuated response when 5 mM methyl succinate was added to the perifusion 5 min after the first stimulation (Fig. 3, B and C, respectively). The [Ca 2ϩ ] m desensitization was not due to aequorin consumption or deleterious effects on the cells, as the [Ca 2ϩ ] m response to methyl succinate was completely restored after an interval of 30 min between the two pulses (Fig. 3D). The addition of 5 mM methyl succinate to INS-1 cells produced an increase in cytosolic ATP, and the same rise could be elicited 5 min later to the same extent during a second exposure to methyl succinate without any significant desensitization (Fig.  3A). Additional time points for the [Ca 2ϩ ] m increases and recovery of the secretory responses have already been documented using glucose as a stimulus (23). Moreover, glucose, which also increases cytosolic ATP levels (36), did not exhibit any desensitization using the protocol of Fig. 2A. The ATP response to 12.8 mM glucose was ϩ23.3 Ϯ 2.0 and ϩ24.3 Ϯ 2.2% during the first and second applications, respectively (not significant, n ϭ 4).
[Ca 2ϩ ] m in Permeabilized INS-1 Cells-The aequorin-expressing cells were then permeabilized with Staphylococcus ␣-toxin, which forms very small holes (2-3-nm diameter) in the plasma membrane (38,8). In this preparation, the cytosolic composition and hence the mitochondrial environment can be controlled. The permeabilized cells were perifused with an intracellular type buffer containing a permissive free Ca 2ϩ concentration of 500 nM and 10 mM ATP. The first addition of 1

FIG. 2. Desensitization of [Ca 2؉ ] m increases in rat islet cells.
Rat islet cells were transiently transfected with the cDNA encoding mitochondrially targeted aequorin and used 3 days later. Cells were perifused at 37°C with HEPES-balanced Krebs-Ringer bicarbonate buffer and exposed for 5 min to 5 mM methyl succinate (met-Suc) (A) or 20 mM KCl (B). Stimulation was repeated after a 5-min interval. The traces are representative of at least three independent experiments. mM succinate induced a large transient peak in [Ca 2ϩ ] m , but the second pulse 5 min later was ineffective (Fig. 4A). The desensitization phenomenon was also observed with ␣-glycerophosphate, which transfers reducing equivalents from the cytosol to the same site (complex II) in the electron transport chain as succinate (Fig. 4B). Glycerophosphate has been shown to produce ATP in isolated islet mitochondria (43). More important, using 5-min intervals, succinate desensitized the effect of ␣-glycerophosphate on [Ca 2ϩ ] m and vice versa (Fig. 4, C and  D). This latter effect shows that the desensitization mechanism appears to be located downstream of the oxidation of FADH 2 . It should be noted that when Ca 2ϩ was substituted with the Ca 2ϩ surrogate Sr 2ϩ in the intracellular type buffer, a very similar desensitization of the mitochondrial [Sr 2ϩ ] increase was ob-served upon repeated succinate stimulation. 2 As for intact cells, the desensitization was not an irreversible process due to a toxic effect or to the loss of functional aequorin since resensitization was observed after 30 min using either succinate or ␣-glycerophosphate (Fig. 5, A and B, respectively).
To examine whether a [Ca 2ϩ ] m increase per se causes desensitization to subsequent challenges, the free Ca 2ϩ concentration of the buffer was varied. To this end, the extramitochondrial [Ca 2ϩ ] in permeabilized cells was kept at permissive 500 nM levels or raised to 1. followed by a second phase that tended to stabilize to the level of extramitochondrial Ca 2ϩ . The second exposure to 1.3 M Ca 2ϩ showed a complete desensitization of the first transient increase in [Ca 2ϩ ] m above the equilibration value between the two compartments (Fig. 6). Imposing the same experimental protocol except for a shortening of the exposures to 1.3 M Ca 2ϩ to 30 s instead of 5 min did not result in any desensitization of the [Ca 2ϩ ] m increase. 2 The latter observation is in agreement with results reported for permeabilized HeLa cells expressing mitochondrial aequorin (44).

Effect of Inhibitors of the Electron Transport Chain on [Ca 2ϩ ] m in Permeabilized INS-1
Cells-Succinate dehydrogenase generates FADH 2 , with the subsequent transfer of electrons to complex II of the electron transport chain (45). In permeabilized cells, the effect of succinate on the increase in [Ca 2ϩ ] m was not affected by the presence of 100 M rotenone, which blocks complex I of the respiratory chain (Fig. 7A). On the other hand, the succinate-induced [Ca 2ϩ ] m increase was completely abolished by 10 M antimycin A, an inhibitor of complex III (Fig. 7B). This suggests that the desensitization occurs between complex II and the uniporter, the latter mediating Ca 2ϩ uptake in the mitochondria.
Effect of Free Ca 2ϩ Concentration on ⌬⌿ m and [Ca 2ϩ ] m in Permeabilized INS-1 Cells-Ca 2ϩ influx into the mitochondria through the uniporter is driven by the hyperpolarization of the mitochondrial membrane under conditions of permissive [Ca 2ϩ ] c . The hyperpolarization occurs by the transfer of reducing equivalents to the electron transport chain and the resulting extrusion of protons. We next tried to discriminate between the respiratory chain and the uniporter as the site of desensitization. For this purpose, we studied the effect of succinate on [Ca 2ϩ ] m in permeabilized cells perifused with nonpermissive (resting) or permissive free [Ca 2ϩ ] (100 and 500 nM, respectively). Under both conditions, succinate was efficient in hyperpolarizing the mitochondrial membrane (Fig. 8, A and B). The dissipation of the proton gradient by carbonyl cyanide p-triflu- oromethoxyphenylhydrazone (1 M) completely depolarized ⌬⌿ m , indicating the polarized state of the mitochondrial membrane. We then monitored [Ca 2ϩ ] m using these two conditions sequentially. As expected, the first addition of 1 mM succinate did not increase [Ca 2ϩ ] m when the permeabilized cells were perifused with 100 nM Ca 2ϩ . One min later, the [Ca 2ϩ ] of the buffer was clamped at 500 nM, which raised the [Ca 2ϩ ] m base line to ϳ300 nM. Four min later, thus 5 min after the first stimulation, the addition of succinate induced a large increase in [Ca 2ϩ ] m (Fig. 8C). This strongly suggests that the uniporter itself is undergoing desensitization since during both succinate exposures the mitochondrial membrane was hyperpolarized due to the activation of the electron transport chain. To support this contention further, ⌬⌿ m was recorded on an attached permeabilized cell preparation under similar conditions as those used for [Ca 2ϩ ] m in Fig. 4. Two successive 5-min exposures to 1 mM succinate were separated by a washing period of 5 min (Fig. 8D). Both exposures to succinate induced a hyper-polarization of ⌬⌿ m of similar magnitude, taking into account the slight drift of the base line.
Effect of Extramitochondrial Ca 2ϩ on Succinate Oxidation to CO 2 in Permeabilized INS-1 Cells-The hyperpolarizing action of succinate on ⌬⌿ m is catalyzed by succinate dehydrogenase, a Ca 2ϩ -independent enzyme (45). By contrast, CO 2 formation from succinate requires a complete turn of the tricarboxylic acid cycle, which involves the two Ca 2ϩ -sensitive enzymes NAD-isocitrate dehydrogenase and ␣-ketoglutarate dehydrogenase (15). As shown in Fig. 9, [2,3-14 C]succinate oxidation to 14 CO 2 was stimulated 4-fold (p Ͻ 0.01) by an increase in the extramitochondrial [Ca 2ϩ ] from 100 to 500 nM in the permeabilized INS-1 cells.

DISCUSSION
Mitochondria play a key role in the metabolism-secretion coupling of the pancreatic beta cell (9 -11, 13, 20). Evidence for a desensitization of this organelle is presented here, and the phenomenon may account for the documented desensitization of stimulated insulin secretion observed in pancreatic beta cells (1,4,5) and derived cell lines (23,46). In the present study, repeated exposures of rat islets to stimulatory glucose concentrations led to attenuated insulin exocytosis. This desensitized secretory response was also observed with the mitochondrial substrate methyl succinate or with KCl-induced depolarization of the plasma membrane. In primary islet cells, desensitization of the mitochondria was indicated by the impaired rise of [Ca 2ϩ ] m during the second exposure to methyl succinate or high potassium. This suggests that desensitization of Ca 2ϩ entry into the mitochondria can be evoked by either tricarboxylic acid cycle intermediates or simply by increasing the [Ca 2ϩ ] c . Nevertheless, to be considered a pure mitochondrial effect, the latter condition implies that the [Ca 2ϩ ] c increase would reach the same value during the second exposure to KCl or at least a level well above the threshold of the uniporter (400 M) (21,16). Although desensitization of the [Ca 2ϩ ] c response to KCl occurs in INS-1 cells, it still attains micromolar concentrations during the second pulse (23). The [Ca 2ϩ ] c reduction is less marked than that of [Ca 2ϩ ] m and therefore probably plays only a minor role in the mitochondrial desensitization. In intact INS-1 cells stimulated with methyl succinate, the blunted insulin secretion correlated with an inhibited increase in [Ca 2ϩ ] m upon a second exposure. In contrast, methyl succinate-induced generation of ATP, reflecting the activation of oxidative phosphorylation, did not display any desensitization, as demonstrated in luciferase-expressing INS-1 cells. The cellular responses to glucose are also desensitized with respect to [Ca 2ϩ ] m and insulin secretion (23), but not in terms of ATP generation (see "Results"). This dichotomy between two mitochondrial parameters, [Ca 2ϩ ] m and ATP generation, can be explained by reduced Ca 2ϩ uptake into the mitochondrial matrix, despite a fully activated respiratory chain. To investigate the underlying mechanism, we have used permeabilized cells to clamp extramitochondrial [Ca 2ϩ ] at a fixed permissive level of 500 nM. This was chosen to ascertain Ca 2ϩ uptake by the uniporter (16). Under these conditions, the succinate-induced increase in [Ca 2ϩ ] m was completely desensitized during the second stimulation. This inhibitory effect takes place downstream of complex II and is apparently not due to altered transport of succinate into the mitochondria. Indeed, the desensitizing effect of succinate could be reproduced with ␣-glycerophosphate. This latter compound transfers reducing equivalents from the glycolytic intermediate dihydroxyacetone phosphate to the same complex II of the electron transport chain without being transported into the mitochondrial matrix (47). Thus, the desensitization evoked by both of the FADH 2 -producing substances (succinate and ␣-glycerophosphate) is very similar, and a common mode of action is underscored by a clear cross-desensitization. In addition, succinate-induced increases in [Ca 2ϩ ] m were blocked by inhibiting complex III with antimycin A, but not by rotenone, which blocks complex I. We therefore conclude that the site of desensitization is located downstream of complex II either in the electron transport chain or at the uniporter through which Ca 2ϩ flows into the mitochondria. The desensitization does not appear to be due to inhibition of the respiratory chain, the activation of which was not impaired. This is demonstrated by the hyperpolarization of ⌬⌿ m irrespective of extramitochondrial Ca 2ϩ . ] m increase, but not to the equilibration of the ion between the extra-and intramitochondrial compartments, which suggests two independent pathways for mitochondrial Ca 2ϩ uptake. Moreover, the desensitization requires a complete activation involving a new steady state. Indeed, very short applications (Ͻ1 min) of Ca 2ϩ in the micromolar range do not lead to desensitization of [Ca 2ϩ ] m responses during a second stimulation in permeabilized cells (44). 2 Although the molecular nature of the mitochondrial Ca 2ϩ uniporter has not been identified, it appears to have properties similar to those of Ca 2ϩ channels of the plasma membrane (48). It may therefore be speculated that the desensitization evoked by an increase in [Ca 2ϩ ] m could involve a mechanism similar to that described for L-type Ca 2ϩ channels (49,50). Such Ca 2ϩ channel desensitization has also been reported in insulin-secreting cells (51). It is conceivable that the high frequency of the [Ca 2ϩ ] c oscillations (two to five/min) observed in glucosestimulated beta cells (12,17,52) serves to prevent desensitization of mitochondrial metabolism. It may be important to optimize the activity of the Ca 2ϩ -sensitive dehydrogenases of the mitochondria to ensure the continuous production of metabolic coupling factors. We show here that succinate oxidation, reflecting tricarboxylic acid cycle activity, is stimulated by extramitochondrial [Ca 2ϩ ] in the physiological concentration range (500 nM). Such an effect was previously reported for the oxidation of pyruvate and its conversion to citrate (20).
The consensus model of metabolism-secretion coupling in the beta cell attributes a key role to ATP produced by the mitochondria (9,10,31). However, as clearly demonstrated by repeated stimulation with methyl succinate, ATP generation is not sufficient for the triggering of insulin secretion. Hence, in intact INS-1 cells, ATP production was preserved in the face of blunted [Ca 2ϩ ] m and secretory responses during the second application of methyl succinate. This will result in diminished activation of the mitochondrial Ca 2ϩ -sensitive dehydrogenases (15), the stimulation of which is required for full activation of the mitochondrial metabolism. An unidentified mitochondrial factor, distinct from ATP, has been proposed to participate in the triggering of insulin exocytosis (13). Its generation requires both a rise in [Ca 2ϩ ] m and the provision of carbons to the tricarboxylic acid cycle (13). Thus, we speculate that during the desensitization of the beta cell, despite normal ATP generation, this mitochondrial factor is missing due to insufficient elevation of [Ca 2ϩ ] m . As a consequence of deficient generation of coupling factors, insulin secretion is impaired. The nature of the coupling factors of mitochondrial origin remains to be established.