Adenosine and the regulation of insulin secretion by isolated rat islets of Langerhans.

The effect of adenosine in insulin secretion and adenylate cyclase activity of rat islets of Langerhans was investigated. Adenosine inhibited insulin secretion stimulated by glucose, glucagon, prostaglandin E2, tolbutamine and theophylline. Adenosine decreased basal adenylate cyclase activity of the islets as well as that stimulated by glucagon prostaglandin E2 and GTP, although fluoride-stimulated activity was not affected. Neither insulin secretion nor adenylate cyclase activity of the islets was affected by adenine, AMP or ADP. The inhibitory effect of adenosine on adenylate cyclase activity was not altered by either phenoxybenzamine (alpha-adrenergic blocker) or propranolol (beta-adrenergic blocker), suggesting that the effect is not mediated through the adrenergic receptors of the islet cells. These results suggest that the intracellular concentration of adenosine in the beta-cell may play a role in regulating insulin secretion and that this effect may be mediated via alterations in the activity of adenylate cyclase in the beta-cell.     

Effects of glucose, glucose metabolites and calcium ions on adenylate cyclase activity in homogenates of mouse pancreatic islets.

The effects of glucose, a series of glucose metabolites, nicotinamide nucleotides, Ca2+ and p-chloromercuribenzenesulphonate on adenylate cyclase activity in homogenates of mouse pancreatic islets were studied. The basal activity of the adenylate cyclase was approx. 6 pmol of cyclic AMP formed/30 min per microng of DNA at 30 degrees C. The enzyme activity was stimulated by some 150% by fluoride. Starvation of the animals for 48h had no effect on either the basal or the fluoride-stimulated activity. The adenylate cyclase activity was increased by 40-50% when 17 mM-glucose, 10 micronM-phosphoenolpyruvate or 10 micronM-pyruvate was added to the assay medium. The effect of glucose was unchanged in the presence of 17 mM-mannoheptulose, and mannoheptulose alone had no effect. The other glycolytic intermediates, and the coenzymes NAD+, NADH and NADPH, at concentrations up to 1 mM were without any detectable effect on the rate of formation of cyclic AMP. The insulin secretagogue p-chloromercuribenzenesulphonate inhibited the adenylate cyclase markedly even at a concentration of 10 micronM. Calculated concentrations of free Ca2+ of 10 micronM and 0.1 mM inhibited adenylate cyclase by 29 and 71% respectively. It is concluded that both glucose itself and phosphoenolpyruvate and/or pyruvate are true activating ligands for islet and adenylate cyclase and that inhibition of the cyclase by Ca2+ may be of physiological significance.     

Islet cyclic AMP levels are not lowered during alpha 2-adrenergic inhibition of insulin release

Epinephrine-induced changes in insulin release and cyclic AMP levels were measured simultaneously in isolated rat islets. Forskolin was used to enhance islet cyclic AMP levels. Forskolin (30 microM) stimulated adenylate cyclase activity 10-fold in islet homogenates and raised cyclic AMP levels 5-fold in intact islets (both at low and high glucose). Insulin release was enhanced by forskolin only at high glucose. Epinephrine (0.1 microM) inhibited glucose- and forskolin-induced insulin release to basal rates. At the same time epinephrine potentiated forskolin-elevated cyclic AMP levels. In contrast epinephrine attenuated forskolin-stimulated adenylate cyclase activity in islet homogenates. At low glucose, both alpha 2- and beta-adrenergic blockade counteracted the epinephrine potentiation, each by 50%. At high glucose the effect was mainly beta-adrenergic in nature. The actions of epinephrine in the presence of a beta-blocker were mimicked by the alpha 2-agonist clonidine. Despite the variations in cyclic AMP levels stimulated insulin release was always inhibited by activation of alpha 2-receptors. Finally, insulin release stimulated by exogenous cyclic AMP was abolished by epinephrine. These results suggest that epinephrine inhibits insulin release at a step distal to the generation of cyclic AMP.     

Inhibition of hepatic adenylate cyclase by NADH

Adenylate cyclase activity in isolated rat liver plasma membranes was inhibited by NADH in a concentration-dependent manner. Half-maximal inhibition of adenylate cyclase was observed at 120 microM concentration of NADH. The effect of NADH was specific since adenylate cyclase activity was not altered by NAD+, NADP+, NADPH, and nicotinic acid. The ability of NADH to inhibit adenylate cyclase was not altered when the enzyme was stimulated by activating the cyclase was not altered when the enzyme was stimulated by activating the Gs regulatory element with either glucagon or cholera toxin. Similarly, inhibition of Gi function by pertussis toxin treatment of membranes did not attenuate the ability of NADH to inhibit adenylate cyclase activity. Inhibition of adenylate cyclase activity to the same extent in the presence and absence of the Gpp (NH) p suggested that NADH directly affects the catalytic subunit. This notion was confirmed by the finding that NADH also inhibited solubilized adenylate cyclase in the absence of Gpp (NH)p. Kinetic analysis of the NADH-mediated inhibition suggested that NADH competes with ATP to inhibit adenylate cyclase; in the presence of NADH (1 mM) the Km for ATP was increased from 0.24 +/- 0.02 mM to 0.44 +/- 0.08 mM with no change in Vmax. This observation and the inability of high NADH concentrations to completely inhibit the enzyme suggest that NADH interacts at a site(s) on the enzyme to increase the Km for ATP by 2-fold and this inhibitory effect is overcome at high ATP concentrations.     

Adenylate cyclase in membrane fractions of RIN-A2-cells: studies with forskolin, NaF, GppNHp and NEM

In crude membrane fractions of rat pancreatic islets and of RIN-A2-cells, forskolin and NaF stimulated adenylate cyclase activity. Basal and stimulated enzyme activity was approximately 3 to 6 fold higher in membranes of RIN-A2-cells than in membranes of islet cells. In RIN-A2-cells GppNHp and NEM inhibited forskolin-stimulated enzyme activity. The inhibitory effect of GppNHp could be reduced by NEM. It is suggested that the adenylate cyclase system of RIN-A2-cells contains inhibitory and stimulatory N-proteins and that there are critical thiols related to Ni, Ns and/or the catalytic unit. Thus, membrane fractions of RIN-A2-cells may be an appropriate model for studies on the adenylate cyclase system of insulin-producing cells.     

A possible role of adenylate cyclase in the longterm dietary regulation of insulin secretion from rat islets of Langerhans

1. Adenylate cyclase activity and patterns of insulin release in response to various concentrations of glucose were determined in islets of Langerhans isolated from starving, fed, or glucose-loaded rats. 2. Basal and glucagon-stimulated activities of adenylate cyclase were lower in islets from starved than from fed rats. The minimum glucose concentration required for stimulation of insulin secretion was higher, whereas the maximum secretory response to glucose was lower, in islets from starved than from fed rats. 3. Adenylate cyclase activity in islets of Langerhans obtained from fed rats loaded with glucose by intermittent intravenous or intraperitoneal injections over 5h was significantly higher than that seen in islets from normal fed rats. Islets obtained from glucose-loaded rats required a lower glucose concentration for stimulation of insulin secretion and attained a higher maximal response to glucose stimulation than those derived from fed rats. 4. Incubation in vitro of islets isolated from normal fed rats, for periods of 1 to 24h in the presence of high concentrations of glucose resulted in an activation of adenylate cyclase that occurred progressively from 2 to 7h and which was maintained during 24h of incubation. The increase of adenylate cyclase activity in isolated islets incubated for 4h in the presence of glucose was not prevented by addition of cycloheximide or actinomycin D. Galactose or 2-deoxyglucose was ineffective in increasing adenylate cyclase activity, and pyruvate (20mm) was less effective than glucose. 5. It is suggested that glucose or a glucose metabolite may exert long-term effects on islet cell adenylate cyclase.     

Characterization of an ATP-Mg2+-dependent guanine nucleotide-stimulated adenylate cyclase from Neurospora crassa

A novel adenylate cyclase activity was found in crude homogenates of Neurospora crassa. The adenylate cyclase had substantial activity with ATP-Mg2+ as substrate differing significantly from the strictly ATP-Mn2+-dependent enzyme characterized previously. Additionally, the ATP-Mg2+-dependent activity was stimulated two- to fourfold by GTP or guanyl-5'-yl-imido-diphosphate (Gpp(NH)p). We propose that the ATP-Mg2+-dependent, guanine nucleotide-stimulated activity is due to a labile regulatory component (G component) of the adenylate cyclase which was present in carefully prepared extracts. The adenylate cyclase had a pH optimum of 5.8 and both the catalytic and G component were particulate. The Km for ATP-Mg2+ was 2.2 mM in the presence of 4.5 mM excess Mg2+. Low Mn2+ concentrations had no effect on adenylate cyclase activity whereas high concentrations of Mn2+ or Mg2+ stimulated the enzyme. Maximal Gpp(NH)p stimulation required preincubation of the enzyme in the presence of the guanine nucleotide and the K1/2 for Gpp(NH)p stimulation was 110 nM. Neither fluoride nor any of a variety of glycolytic intermediates or hormones, including glucagon, epinephrine, and dopamine, had an effect on ATP-Mg2+-dependent adenylate cyclase activity. However, the enzymatic activity was stimulated not only by GTP but also by 5'-AMP and was inhibited by NADH.     

Cytochemical localization of adenylate cyclase in the dense tubule system of human blood platelets stimulated by forskolin, prostacyclin and prostaglandin D2

Platelets were briefly fixed in paraformaldehyde/glutaraldehyde and then incubated with 5'-adenylyl imidodiphosphate under conditions suitable for the cytochemical detection of adenylate cyclase activity. The adenylate cyclase activity of these platelets retains the ability to respond to prostaglandins E1, D2, I2 (prostacyclin), forskolin and fluoride. Sites of stimulated adenylate cyclase activity were localized cytochemically by the reaction of lead with the reaction product imidodiphosphate to form deposits of lead imidodiphosphate that are visible in the electron microscope. Reaction product deposition was seen only in the dense tubule system of human platelets when the incubation medium contained forskolin, prostacyclin, or prostaglandin D2 at concentrations known to stimulate the enzyme in intact platelets. Epinephrine, an antagonist of adenylate cyclase inhibited the cytochemical reaction stimulated by prostacyclin. The fact that the cytochemical reaction was induced by agonists that stimulate the enzyme through two different types of prostaglandin receptors and by forskolin, which acts distal to the receptors, confirms that the method specifically detects adenylate cyclase. The presence of adenylate cyclase in the dense tubules may be significant for the regulation of intracellular Ca2+ and arachidonic acid metabolism by this membrane system.     

Adenylate cyclase system of bovine adrenal plasma membranes.

The adenylate cyclase system present in a preparation enriched in plasma membranes derived from bovine adrenal cortex was investigated in considerable detail. This system is stimulated by adrenocorticotropic hormone (ACTH), by biologically active analogs of this hormone, and by fluoride ion. The preparation contains sodium-potassium- and magnesium-dependent ATPases that are markedly inhibited by 50 mM sodium fluoride. Incorporation of a pyruvate phosphokinase ATP generating system into the adenylate cyclase assay medium provided constant substrate levels. In the presence of the ATP generating system, the rate of cyclic AMP formation (basal, fluoride, and ACTH-activated) was proportional to enzyme concentration and was linear with time. Proportionality with respect to enzyme concentration as concerned the hormone-activated adenylate cyclase was achieved only when the ratio of hormone to enzyme protein was kept constant. The temperature optimum of the adenylate cyclase, basal or activated, was approximately 30 degrees. Michaelis-Menten kinetics were observed when the ratio of Mg2+ to ATP was approximately 6:1. Both calcium and ethylene glycol bis(beta-aminoethyl ether)-N,N'-tetraacetic acid completely inhibited the adenylate cyclase system at concentrations of 5 and 0.5 mM, respectively. GTP was inhibitory at concentrations of 10-2 M but had little effect at lower concentrations. Freezing in liquid nitrogen and storage at -60 degrees exerted little effect on the fluoride-stimulated enzyme but lowered hormone stimulated activity. Preincubation in the presence of ACTH afforded a high degree of stabilization of the enzyme system while preincubation with a biologically inactive analog afforded no protection.     

Calcium-dependent adenylate cyclase of pituitary tumor cells

Effects of Ca2+ and calmodulin on the adenylate cyclase activity of a prolactin and growth hormone-producing pituitary tumor cell strain (GH3) were examined. The adenylate cyclase activity of homogenates was stimulated approx. 60% by submicromolar free Ca2+ concentrations and inhibited by higher (microM range) concentrations of the cation. A 2-3-fold stimulation of the activity in response to Ca2+ was observed at physiologic concentrations of KCl, with both the stimulatory and inhibitory responses occurring at respectively higher free Ca2+ concentrations. Calmodulin in incubations at low KCl concentrations increased the enzyme activity at all Ca2+ concentrations tested. In incubations conducted at physiologic KCl concentrations, both the inhibitory and stimulatory responses to Ca2+ were shifted by calmodulin to lower respective concentrations of the cation, without significant change occurring in the maximal rate of enzymic activity at optimal free Ca2+ X Mg2+ concentrations in the incubation also influenced the Ca2+ concentration dependence of adenylate cyclase; at high Mg2+ more Ca2+ was required to obtain maximal activity. Trifluoperazine inhibited adenylate cyclase of GH3 cells only in the presence of Ca2+; as Ca2+ concentrations in the assay were increased, higher drug concentrations were required to inhibit the enzyme. Ca2+ was also observed to reduce the extent of enzyme destabilization which occurred during pretreatments at warm temperatures. Vasoactive intestinal polypeptide and phorbol myristate acetate, which stimulate prolactin secretion in intact GH3 cells, enhanced enzyme activity 4- and 2.5-fold, respectively, without added Ca2+. Increasing free Ca2+ concentrations reduced the enhancement by VIP and eliminated the stimulation by PMA.     

The role of Ca-2+ and cyclic adenosine 3':5'-monophosphate in insulin release induced in vitro by the divalent cation ionophore A23187.

Insulin release from isolated perifused pancreatic islets was stimulated by the divalent ionophore A23187 in the absence of exogenous glucose. In addition, A23187 produced a 2-fold elevation of cyclic adenosine 3':5'-monophosphate (cAMP) levels in isolated perifused islets. The elevation of cAMP levels coincided with peak insulin release. Ionophore-induced insulin release was unaffected by pretreatment of the islets with theophylline (5 mM). Stimulation of insulin release produced by the ionophore occurred either in the presence or absence of extracellular Ca-2+; however, cAMP accumulation required the presence of extracellular Ca-2+. The ionophore (10 muM) had no effect on adenylate cyclase activity of homogenates of isolated islets. The results of this study are interpreted as indicating that intracellular Ca-2+ has an essential role in the insulin releasing mechanism, whereas the cAMP system has a modulatory effect on this process.     

Inhibitory effect of clonidine upon adenylate cyclase activity,cyclic AMP production,and insulin release in rat pancreatic islets

Conflicting opinions were recently expressed concerning the possible effect of ₂-adrenergic agonists upon cyclic AMP production in pancreatic islets. In the present: study, clonidine inhibited glucose-induced insulin release from rat pancreatic islets, this inhibitory effect being abolished by idazoxan. Clonidine did not suppress the capacity of forskolin to augment glucose-induced insulin release. In a particulate subcellular fraction derived from the islets, adenylate cyclase was activated by calmodulin (in the presence of Ca²⁺), NaF, GTP,, L-arginine, and forskolin, and slightly inhibited by clonidine. The inhibitory action of clonidine upon basal adenylate cyclase activity was more pronounced in islet crude homogenates. The inhibitory effect of clonidine was antagonized by forskolin whether in the particulate fraction or crude homogenate. At variance with the modest effects of glucagon, D-glucose, L-arginine, or a tumor-promoting phorbol ester upon cyclic AMP production by intact islets, forskolin caused a six-fold increase in cyclic AMP production. Clonidine inhibited cyclic AMP production by intact islets, whether in the absence or presence of forskolin. It is proposed that the inhibitory action of clonidine upon insulin release is attributable , in part at least, to inhibition of adenylate cyclase.     

Catecholamine inhibition of Ca2+-induced insulin secretion from electrically permeabilised islets of Langerhans

Noradrenaline (1-10 microM) inhibited Ca2+-induced insulin secretion from electrically permeabilised islets of Langerhans with an efficacy similar to that for inhibition of glucose-induced insulin secretion from intact islets. The inhibition of insulin secretion from permeabilised islets was blocked by the alpha 2-adrenoreceptor antagonist, yohimbine. Adenosine 3',5'-cyclic monophosphate (cAMP) did not relieve the noradrenaline inhibition of Ca2+-induced secretion from the permeabilised islets, although noradrenaline did not affect the secretory responses to cAMP at substimulatory (50 nM) concentrations of Ca2+. These results suggest that catecholamines do not inhibit insulin secretion solely by reducing B-cell adenylate cyclase activity, and imply that one site of action of noradrenaline is at a late stage in the secretory process.     

Effects of guanosine on insulin secretion and adenylyl and guanylyl cyclase activities of isolated rat islets of Langerhans

The effect of guanosine on insulin secretion adenylyl and guanylyl cyclase activities of isolated rat islets of Langerhans was investigated. Guanosine (1–100 μM) inhibited glucose, tolbutamide, theophylline and prostaglandin E₂-stimulated insulin secretion although it failed to affect glucagon stimulated secretion. Prostaglandin E₂-stimulated adenylyl cyclase of islets was inhibited by guanosine although guanosine had no effect on basal, fluoride, glucagon or GTP-stimulated activity. Guanosine markedly decreased basal guanylyl cyclase activity of islets.These results suggest that guanosine may affect insulin release by inhibiting adenylyl and guanylyl cyclase activities in the ß-cell thereby decreasing the intracellular concentrations of cyclic nucleotides.This effect may be important in modulating the secretory response of the islets to a variety of hormonal agents.     

In vivo regulation of cyclic AMP phosphodiesterase in Xenopus oocytes. Stimulation by insulin and insulin-like growth factor 1

Cyclic AMP phosphodiesterase activity was measured in vivo after microinjection of [3H]cAMP into intact Xenopus oocytes. This activity was inhibited by extracellular application of methylxanthines, and the dose-dependent inhibition of phosphodiesterase activity correlated with the abilities of isobutylmethylxanthine and theophylline to inhibit oocyte maturation induced by progesterone, with IC50 values of approximately 0.3 and 1.5 mM, respectively. Insulin stimulated in vivo phosphodiesterase activity measured after microinjection of 200 microM [3H]cAMP in a time- and dose-dependent fashion without affecting phosphodiesterase activity measured after microinjection of 2 microM [3H]cAMP. Although progesterone alone had no effect on in vivo phosphodiesterase activity, low concentrations of progesterone (0.01 microM) accelerated the time course of insulin stimulation of both phosphodiesterase activity and oocyte maturation. The EC50 for stimulation of in vivo phosphodiesterase activity by insulin correlated with the IC50 for inhibition of oocyte membrane adenylate cyclase activity measured in vitro (2 and 4 nM, respectively). Twenty-fold higher concentrations of insulin were required to stimulate oocyte maturation. In contrast, insulin-like growth factor 1 stimulated in vivo phosphodiesterase, inhibited in vitro adenylate cyclase, and induced oocyte maturation at concentrations of 0.3-1.0 nM. These results demonstrate a dual regulation of oocyte phosphodiesterase and adenylate cyclase by insulin and insulin-like growth factor 1.     

Adenylate cyclase in silkworm. Properties of the enzyme in pupal fat body.

Adenylate cyclase was assayed in a sonicated preparation of silkworm pupal fat body. The adenylate cyclase was found mostly in the particulate fraction. The activity depended upon either Mg2+ or Mn2+, and the degree of stimulation by Mn2+ was 2 times greater than that by Mg2+ compared at the saturating concentrations. In the presence of Mg2+, the enzyme was inhibited by both EGTA and high concentrations of Ca2+, showing biphasical response to Ca2+. The enzyme was stimulated several-fold by NaF. The enzyme exhibited typical Michaelis-Menten kinetics and Km values were 0.13 mM for MgATP and 0.086 mM for MnATP.     

Calmodulin-mediated adenylate cyclase from mammalian sperm

Calmodulin (CaM), the calcium binding protein that modulates the activity of a number of key regulatory enzymes, is present at high levels in sperm. To determine whether CaM regulates adenylate cyclase in mammalian sperm, the actions of EGTA and selected CaM antagonists on a solubilized adenylate cyclase from mature equine sperm were examined. The activity of equine sperm adenylate cyclase was inhibited by EGTA in a concentration-dependent manner with a half-maximal inhibitory concentration (IC50) of 2 mM. Equine sperm adenylate cyclase was also inhibited in a concentration-dependent manner by the CaM antagonists chlorpromazine and calmidazolium (IC50 = 400 and 50 microM, respectively). The inhibition of enzyme activity by these agents correlated with their known potency and specificity as anti-CaM agents. The activity of the enzyme in the presence of 200 microM calmidazolium was restored by the addition of authentic CaM (EC50 = 15 microM); full activity was restored by the addition of 50 microM CaM. La3+, an ion that dissociates CaM from tightly bound CaM-enzyme systems, inhibited equine sperm adenylate cyclase (IC50 = 1 mM). Incubation of equine sperm adenylate cyclase with La3+ dissociated endogenous CaM from the enzyme so that most of the enzyme bound to a CaM-Sepharose column equilibrated with Ca2+. Specific elution of CaM-binding proteins from the CaM-Sepharose column with EGTA yielded a CaM-depleted adenylate cyclase fraction that was stimulated 2-fold by the addition of exogenous CaM.     

Stimulation and inhibition of rat basophilic leukemia cell adenylate cyclase by forskolin

The influence of the diterpene, forskolin, was studied on adenylate cyclase activity in membranes of rat basophilic leukemia cells. Forskolin increased basal adenylate cyclase activity maximally 2-fold at 100 microM. However, adenylate cyclase activity stimulated via the stimulatory guanine nucleotide-binding protein, Ns, by fluoride and the stable GTP analog, guanosine 5'-O-(3-thiotriphosphate), was inhibited by forskolin. Half-maximal and maximal inhibition occurred at about 1 and 10 microM forskolin, respectively. The inhibition occurred without an apparent lag phase, whereas the enzyme stimulation by forskolin was preceded by a considerable lag period. The inhibition was not affected by treating intact cells or membranes with pertussis toxin and proteolytic enzymes, respectively, which have been shown in other cell types to prevent adenylate cyclase inhibition mediated by the guanine nucleotide-binding regulatory component, Ni. The forskolin inhibition of the stable GTP analog-activated adenylate cyclase was impaired by increasing the Mg2+ concentration and was reversed into a stimulation by Mn2+. Under optimal inhibitory conditions, forskolin even decreased basal adenylate cyclase activity. Finally, forskolin largely reduced the apparent affinity of the rat basophilic leukemia cell adenylate cyclase for its substrate, MgATP, which reduction resulted in an apparent inhibition at low MgATP concentrations and a loss of the inhibition at higher MgATP concentrations. The data indicate that forskolin can cause both stimulation and inhibition of adenylate cyclase and, furthermore, they suggest that the inhibition may not be mediated by the Ni protein, but may be caused by a direct action of forskolin at the adenylate cyclase catalytic moiety.     

Adenylate cyclase activity in isolated rat islets of Langerhans Effects of agents which alter rates of insulin secretion

Adenylate cyclase activity was estimated inhomogenates of rat islets of Langerhans. by measurement of the conversion of [α-³²P]ATP to adenosine cyclic 3′,5′-[³²P]monophosphate. Islet cell adenyulate cyclase activity was stimulated by the addition to the homogenates of glucagon, fluoride, prostaglandins E₁ or E₂ GTP or CTP although not by UTP, TTP, GDP, or GMP. Adrenaline, noradrenaline and isoproterenol were each found to inhibit the activity, the order of potency at a concentration of 10^?4 M being adrenaline > noradrenaline > isoproterenol. The effects of these agents were not altered by β-blackade with propanolol but could be preventived by α-blockade with phenoxybenzamine. The following agents, present at concentrations previously shown to increase rates of insulin secretion from rat islets of Langerhans, were ineffective in altering adenylate cyclase activity when tested in the presence or absence of 0.1 mM GTP: glucose, glibenclamide, xylitol leucine, arginine, or potassium. These results suggest that the activity of adenylate cyclase in the B cells of rat islets of Langerhans may play an important role in mediating the direct effects of hormones and adrenergic agents on insulin release, although the short term effects of substrates such as glucose or amino acids on the release process do not appear to be mediated through alterations in the activity of this enzyme.     

Regulation of insulin release from isolated islets of Langerhans of the rat in pregnancy. The role of adenosine 3':5'-cyclic monophosphate

1. The concentrations of cyclic AMP were compared in islets of Langerhans isolated from the pancreases of normal female and pregnant rats and were higher in islets in pregnancy. 2. There was also a significant increase in adenylate cyclase activity in homogenates of islets from pregnant rats compared with those from normal rats. 3. Increased cyclic AMP concentration in islets from pregnant rats was reflected in increased protein kinase activity. When the cyclic AMP-dependent protein kinase activity was increased by 3-isobutyl-1-methylxanthine this stimulated activity was significantly greater in pregnancy. 4. Insulin-secretion studies with islets from normal and pregnant rats showed that theophylline or 3-isobutyl-1-methylxanthine, which raise intracellular cyclic AMP concentrations, caused a significantly greater insulin secretion in pregnancy. 5. It was also found that in the presence of a glucose concentration too low to stimulate insulin secretion, the latter could be induced if the cyclic AMP concentrations were raised sufficiently with 3-isobutyl-1-methylxanthine. 6. It is suggested that the higher cyclic AMP concentrations observed in islets in pregnancy mediate the greater insulin-secretory capacity, as well as the greater sensitivity of these islets to low glucose concentrations.