The Effects of Cerulenin, an Inhibitor of Protein Acylation, on the Two Phases of Glucose-Stimulated Insulin Secretion

Male Wistar rats weighing 300–400 g were killed by CO₂ asphyxiation, the pancreata were surgically removed, and islets were isolated by collagenase digestion in HEPES-buffered Krebs-Ringer bicarbonate solution (18,19). The islets were then subjected to a 45-min static incubation at 37°C in the presence or absence of 100 μmol/l cerulenin before they were transferred to perifusion chambers. A 40-min equilibration period was followed by perifusion under the indicated conditions, in which samples for insulin measurement were taken. Release rates were determined by radioimmunoassay using a charcoal separation technique (20). Since it has been shown that cerulenin covalently modifies sulfhydryl residues contained in the active sites of enzymes involved in fatty acid metabolism and therefore cannot be easily washed out (21), we added the compound only during the 45-min static incubations and not during the subsequent perifusion period. There was no evidence of reversibility of the effect of cerulenin over 90 min (21).

In initial control studies, 100 μmol/l cerulenin had no effect on the rate of basal insulin secretion in the presence or absence of extracellular Ca²⁺ (data not shown). This concentration of the inhibitor was then used throughout the study. When insulin secretion was stimulated by glucose or the combination of 10 mmol/l glutamine and 10 mmol/l 2-aminobicyclo(2.2.1)heptane (BCH; a nonmetabolizable analog of leucine, which activates glutamate dehydrogenase), cerulenin had a powerful inhibitory effect (Fig. 1A and B, respectively). Both the first and second phases of release were profoundly inhibited. The effect on the first phase suggested that the operation of either the K_ATP channels or voltage-dependent Ca²⁺ channels had been compromised. To determine if this was the case, the effect of cerulenin on the response to a depolarizing concentration of KCl was studied. As can be seen from the results in Fig. 2, cerulenin had no effect on the response to KCl. Thus, the Ca²⁺ channels are operating normally in response to depolarization, and the subsequent secretory response to elevated [Ca²⁺]_i is also unaffected. It is concluded that cerulenin either prevents closure of the K_ATP channels in response to nutrient stimulation or is actually activating the channels. Because of this effect, and because the K_ATP channel–independent pathway acts in synergy with increased [Ca²⁺]_i, it was necessary to study the effect of cerulenin on the K_ATP channel–independent pathway in isolation. This was achieved by activating the K_ATP channels with diazoxide, depolarizing the cell with KCl, and then measuring the effect of 16.7 mmol/l glucose on insulin release independently of the K_ATP channels. The results are shown in Fig. 3. The control response to 40 mmol/l KCl in the presence of 250 μmol/l diazoxide was monophasic, with the peak secretion rate after 1 min and a return to basal levels after 10–15 min. The K_ATP channel–independent response to glucose under these conditions was a large increase in release over and above that of KCl alone. Peak rates were achieved after 10 min, and insulin secretion remained at high levels throughout the course of the experiment. Cerulenin strongly inhibited this response. The initial rise in insulin secretion to the KCl peak at 1 min was unaffected. However, from minute 2, the secretion rate began to decline as the glucose response was inhibited. Thus, cerulenin inhibits the K_ATP channel–independent pathway of glucose signaling. Because the main purpose of these studies is to determine whether acylation is involved in stimulus-secretion coupling, the K_ATP channel–independent effect of palmitate was studied under the same conditions with KCl and diazoxide (22). The results of these experiments are shown in Fig. 4. It can be seen that the response to palmitate was also completely blocked by cerulenin.

Next, we studied the K_ATP channel–independent effect of glucose under Ca²⁺-free conditions (in the presence of 1 mmol/l EGTA) and also took advantage of the marked responses to both glucose and palmitate that occur in the absence of extracellular Ca²⁺ (23). 12-O-Tetradecanoylphorbol-13-acetate (TPA; 100 nmol/l) and forskolin (6 μmol/l) applied at 10 min caused a small increase in insulin release as anticipated. After 20 min, the addition of 11.1 mmol/l glucose (the maximally effective concentration under these conditions) caused a threefold increase in insulin secretion, which was totally blocked by cerulenin (Fig. 5 A). Similar results were obtained when the inhibitor was tested against the response to 10 μmol/l free palmitate (Fig. 5 B). Palmitate (in the presence of TPA and forskolin) induced a rise in the rate of insulin secretion that rose to three times the basal rate. As was the case with glucose, cerulenin completely abolished the response.

In this study, we provide further evidence that nutrient-stimulated insulin release and specifically the nutrient effects on the K_ATP channel–independent pathways of secretion can be almost completely eliminated by the irreversible protein acylation inhibitor cerulenin. Strikingly, this antibiotic does not affect secretory responses elicited by nonnutrients like KCl, which increases intracellular Ca²⁺, or the stimulation caused by the combination of TPA and forskolin. Caution has to be paid to possible other specific (by inhibition of acylation) and/or nonspecific effects that cerulenin might exert in the β-cell, where it clearly inhibits the glucose effect on the first phase of secretion. The inhibition of the first phase of secretion by cerulenin must be due to a failure of glucose to increase [Ca²⁺]_i because the response to increased [Ca²⁺]_i by KCl is not affected by cerulenin. Such an effect might be exerted directly on the K_ATP channel or indirectly by virtue of an effect on glucose metabolism. In this respect, however, it is important to note that cerulenin treatment does not affect ATP production, glucose oxidation, or the effect of glucose to suppress palmitate oxidation (17). It is possible, therefore, that cerulenin has an effect on the K_ATP channel itself. The fact that cerulenin has no effect on the secretory response to a depolarizing concentration of KCl and raised [Ca²⁺]_i demonstrates that cerulenin has no effect on Ca²⁺ stimulus–secretion coupling. Its inhibitory effects are therefore exerted on the nutrient augmentation pathways.

Additional potential targets for inhibition by cerulenin are enzymatic activities like fatty acid synthase (23), although its activity is very low in the β-cell, and acetyl-CoA carboxylase (24). Whereas an inhibition of acetyl-CoA carboxylase cannot be excluded as a cause of the inhibition of the response to glucose (although the glucose suppression of fatty acid oxidation was not affected by cerulenin), we were able to bypass this enzyme in the experiments in which palmitate was used as the stimulus in both the presence and the absence of Ca²⁺ and in which cerulenin still exerted its full inhibitory capacity. Other described effects of the compound, e.g., reduced apoptosis (25) and inhibition of RNA (26) and protein synthesis (23), are not likely to explain our results because of the short exposure time to cerulenin and the lack of effect on the KCl response.

Protein acylation is a post-translational event, which usually links palmitate in the form of a fatty acyl-CoA, as the preferred substrate, to a cysteine residue through a thioester linkage in a variety of proteins (e.g., G_α subunits, ras [27], Ca²⁺ channels [28], etc.), and it has only recently been established that an enzymatic activity (protein S-acyltransferase) is likely to represent the predominant mechanism for thioacylation (29). Because it is a reversible modification with dynamic cycles of acylation and deacylation, it is capable of playing a role in signal transduction. One possible target for acylation has been recently identified as a PKC of 80 kDa whose translocation to membrane bilayers was facilitated upon palmitoylation (30). Moreover, acylation has also been shown to occur on proteins directly linked to exocytosis like synaptotagmin (31), which is a putative Ca²⁺ sensor, and SNAP-25 (synaptosomal-associated protein-25) (32), a component of the SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) complex, and it therefore might reflect a distal event in secretion. This assumption is supported by a study in HIT-T15 cells, where an enhancement of Ca²⁺-induced insulin secretion by long chain-acyl CoAs was reported (33). However, nonspecific interactions of acyl-CoAs at high concentrations with cell membranes and subsequent facilitation of fusion processes cannot yet be excluded. In preliminary studies where rat pancreatic islets were labeled with [³H]palmitate and subjected to SDS-PAGE and fluorography, we were able to distinguish several discrete bands, of which one had a molecular weight of 25,000, the same as SNAP-25. Obviously, more detailed studies are needed to determine an involvement of protein acylation in the course of the K_ATP channel–independent pathways.

This work was supported by grant R01-DK54243 from the National Institutes of Health.

The authors are grateful for the excellent technical assistance of Dr. Troitza K. Bratanova-Tochkova.