The Stimulus-Secretion Coupling of Amino Acid-induced Insulin Release INFLUENCE OF A NONMETABOLIZED ANALOG OF LEUCINE ON THE METABOLISM OF GLUTAMINE IN PANCREATIC ISLETS*

L-Glutamine causes a dose-related enhancement of insulin release evoked, in rat pancreatic islets, by the nonmetabolized analog of leucine, 2-aminobicyclo-[2,2,l]heptane-2-carboxylic acid (BCH). The influence of BCH upon the metabolism of L-glutamine was inves-tigated. In the islets exposed to L-glutamine, BCH de- creased the deamidation of glutamine, but stimulated the oxidative deamination of glutamate, increased the rate of generation and islet content of 2-ketoglutarate, and augmented the oxidation of ~&J-'~C]glutamine. BCH antagonized the sparing action of L-glutamine upon the oxidation of endogenous fatty acids. The stim- ulation of insulin release by the association of L-gluta-mine and BCH was commensurate with the estimated increase in O2 consumption and coincided with an increase in the islet NADPH/NADP' ratio, net uptake of 46Ca, and cyclic AMP concentration. It is concluded that insulin release evoked by these amino acids is causally linked to an increase in catabolic fluxes, the secretagogues acting in the islet cells as a fuel (glutamine) or enzyme activator (BCH). The nonmetabolized analog of L-leucine,


MATERIALS AND METHODS
L-Glutamine and P(-+)BCH were obtained from Sigma and Calbiochem AG (Lucerne, Switzerland), respectively.
All experiments were performed with pancreatic islets isolated from the pancreas of fed albino rats. The methods used to measure insulin release (6), the islet production of NH,' , glutamate, 2ketoglutarate, malate, oxaloacetate, aspartate, pyruvate, and alanine (4), the oxidation of exogenous nutrients (8), the output of "COZ from islet prelabeled with [U-'4C]palmitate (9, lo), the islet content of adenine and pyridine nucleotides (Il), the net uptake of 45Ca (12), and the concentration of cyclic AMP in the islets and their incubation medium (13) were all described in prior publications.
Briefly, insulin release was measured in groups of 8 islets each incubated for 90 min in a bicarbonate-buffered medium (1.0 m l ) containing bovine albumin (5.0 mg/ml) and equilibrated against a mixture of COZ (5%) and Oz (95%) (6). Metabolic variables were measured in groups of 15-50 islets each incubated in 40-100 p1 of the same bicarbonate-buffered medium. All nucleotides and metabolites were measured in the islets and incubation media treated as a whole. Hence, no distinction was made between the islet content and output of metabolic end products. The net uptake of 45Ca was measured by incubating the islets for 90 min in the presence of 45Ca and then submitting them to repeated washes in order to remove extracellular radioactivity (12). For measuring the cyclic AMP content of the islets and media, groups of 10 islets each were preincubated for 60 min in media (0.2 m l ) containing D-glucose, 5.6 mM, and then incubated for 60 min in media (0.1 ml) containing the required amino acids and 3isobutyl-I-methykanthine (1.0 m). After incubation, the islets and media were mixed with and sonicated in 0.1 ml of trichloroacetic acid (lo%, w/v). The cyclic AMP was then extracted, acetylated, and eventually measured by radioimmunoassay (13).
In each individual experiment, control and experimental values were obtained in close to equal number of distinct batches of islets all derived from the same initial preparation. In a few instances (e.g. measurement of NADPH and NADP'), all four experimental conditions (i.e. no substrate, L-glutamine alone, P(-+)BCH alone, and both L-glutamine and P(&)BCH) were not tested within the same experiment(s). In such a case, the experimental values were normalized relative to the mean basal value found within the same experimentb) and converted back to absolute values taking into account the overall mean basal reading derived from all available experiments.
For measurement of glutaminase activity (EC 3.5.1.2), groups of 450 islets each were sonicated (twice for 5 s ) in 0.3 ml of a bicarbonatebuffered medium (Na+, 24, K+, 120, Mg2+, 1, Ca2+, 1, C1-, 124, HCOC, 24 mM) containing bovine albumin (2 mg/ml) and equilibrated against a mixture of Oz and COz (95/5, v/v; pH 7.4). An aliquot of the homogenate (15 pl) was mixed with an equal volume of a solution containing L-glutamine (20 mM), and when required 2-ketoglutarate (up to 2 RIM), and prepared in the same buffer. After a 20-min 90 "C. After centrifugation, an aliquot (20 p1) of the supernatant was incubation at 37 "C, the reaction was stopped by heating for 4 min at mixed with an equal volume of a solution which consisted of a Tris-3754 by guest on March 22, 2020 http://www.jbc.org/ Downloaded from HCl buffer (100 n"; pH 7.6) containing EDTA (10 m~) , 2-ketoglutarate (20 m), ADP (2 mM), NADH (2 n " ) , and glutamate dehydrogenase (20 pg/ml). After a 30-min incubation at room temperature and addition of HC1 (0.3 ml, 0.3 N), the tubes were maintained for 20 min at room temperature and the fluorescence of NAD' was induced by adding NaOH (1.0 ml; 6 N) and heating the tubes at 60 "C for 10 min. The readings were corrected for the blank value (no homogenate) obtained at the same concentration of L-glutamine and 2-ketoglutarate, and expressed as nmol/60 &/islet by reference to appropriate NH4+ standards treated in the same manner as the reaction mixture. All results are expressed as the mean f S.E. together with the number of individual observations. The statistical significance of differences between mean values was tested by use of Student's t-test. The S.E. on the sum or difference between mean values was calculated according to Snedecor (14).

RESULTS
Insulin Release- Fig. 1 illustrates the enhancing action of L-glutamine, in increasing concentrations, upon insulin release evoked by P(+)BCH (20 m). At an 1.0 mM concentration, L-glutamine doubled the insulinotropic action of P(+)BCH. This near physiological concentration of L-glutamine was used in all further studies. The effect of increasing concentrations of P(f)BCH upon insulin release in the present system was previously characterized (2).
Glutamine Deamidation-The metabolic changes evoked by either L-glutamine (1 m) or P(k)BCH (20 m) relative to basal value have been described and analyzed in two recent publications (2,15). Therefore, emphasis is here given to the metabolic situation found in the simultaneous presence of both amino acids. Nevertheless, the basal data and those collected in the presence of either L-glutamine or ,f3(+)BCH are also presented for purpose of comparison.
Although P(f)BCH significantly augmented NH,' production in islets not exposed to any other nutrient, the leucine analog decreased, significantly albeit slightly, NH4+ production in the presence of L-glutamine (Table I, first line). In the sole presence of L-glutamine, the rate of NH4+ production was linear with time, values of 114.0 f 5.1 and 225.7 f 11.0 pmol/ islet being recorded after 60-and 120-min incubation, respectively. The relative magnitude of the BCH-induced decrease in NH4+ production was virtually identical after 60-and 120min incubation (data not shown). The decrease in total NH,' production contrasted with the fact that, in the islets exposed to L-glutamine, P(+)BCH augmented the production of NH,' at the level of the reaction catalyzed by glutamate dehydrogenase ( Fig. 2).
BCH also significantly decreased the islet content and/or output of glutamate in the islets exposed to L-glutamine (Table I, second line). The BCH-induced decrease in gluta-mate production exceeded the value expected from the increase in the catabolism of this amino acid (Fig. 2). Thus, the measurement of both NH4+ and glutamate production suggested that BCH somehow impairs the deamidation of glutamine in the islets. As judged from the production of either NH4+ (after correction for the BCH-induced increase in the oxidative deamination of glutamate) or glutamate (after correction for the BCH-induced changes in the production of metabolites derived from glutamate), P(-+)BCH decreased the deamidation of glutamine by 11.0-24.2 pmol/60 min/islet, representing a mean decrease of 22.3 f 5.3% ( p < 0.001) relative to the rate of glutamine deamidation in the absence of BCH (ie. 79 pmol/60 min/islet; see Ref. 15).
The finding that BCH inhibited glutamine deamidation and the knowledge that BCH augments the production rate and islet content of 2-ketoglutarate ( Fig. 2) led us to explore the influence of 2-ketoglutarate upon glutaminase activity in an islet homogenate. As illustrated in Fig. 3, 2-ketoglutarate caused a dose-related decrease in glutaminase activity.
Conversion of Glutamate to 2-Ketoglutarate-In the presence of L-glutamine, P(-+)BCH failed to significantly affect the conversion of glutamate (derived from exogenous L-glutamine) to 2-ketoglutarate by transamination reactions, as judged from the islet content and/or output of alanine and aspartate (Table I, (rectangles) and production of metabolites (parentheses) evoked by /J(*)BCH (20 mm) in islets exposed to Gglutamine (1 mm). All values are expressed as pmol/60 min/islet and correspond to the mean values given in Table I, taking into account the BCHinduced mean increment in '?O, output (29.0 pmol/60 min/islet) from islets exposed to L-[U-'4C]glutamine. The effect of P(+)BCH upon the deamidation of glutamine represents the mean (and its range of variation) of two values based, respectively, on the decrease in NHs' production (after correction for the increase in the oxidative deamination of glutamine) and on the decrease in glutamate production (after correction for the changes in the production of its metabelites).

TABLE I Effects of L-glutamine and P(&)BCH upon the islet output and/or content
however, for P(-t)BCH to decrease the production of aspartate in the islets exposed to L-glutamine (p < 0.08). The rate of glutamate conversion to 2-ketoglutarate by oxidative deamination was obviously increased by /?(+)BCH. In the islets exposed to L-glutamine, this was documented, inter alia, by an increase in both the islet content of 2ketoglutarate (Table I, third line) and oxidation rate of L-[U-'*C]glutamine (Fig. 4). According to Fig. 2, the BCH-induced increase in the oxidative deamination of glutamate averaged 11 pmol/60 min/islet. This should be compared with the rate of the same reaction in islets exposed solely to L-glutamine, i.e. about 14 pmol/60 min/islet (15). Such a comparison indicates that P(+)BCH increased by approximately 80% the flow rate through the reaction catalyzed by glutamate dehydrogenase.
Catabolism of 2-Ketoglz&arczte-In the islets exposed to Lglutamine, /3(+)BCH failed to significantly affect the islet content and/or output of malate and pyruvate, the trend being here toward an increase in the content and/or output of these metabolites in response to /3(*)BCH (Table I, 4th and 7th lines).
BCH increased the oxidation of L-[U-'4C]glutamine by the islets. This increase amounted to +67.9 f 15.0% (degree of freedom, 42; p < 0.001) after a 30-min incubation (Fig. 4). Thereafter, however, the stimulant action of /3(+-)BCH upon glutamine oxidation seemed to fade out. Indeed, between the 30th and 120th min of incubation, the lines characterizing the time course for "?Oz output in the presence and absence of P(&)BCH, respectively, ran in parallel fashion, with a mean difference in elevation of 5.8 pmol/islet (expressed as L-glutamine residues).
Over a 30-min incubation, menadione (0.01 mM) augmented the oxidation of L-glutamine (1.0 111~) from 8.7 + 0.6 to 11.1 f 0.8 pmol/30 min/islet in the absence of /?(2t)BCH, and from 14.4 + 1.2 to 19.0 f 1.0 pmol/30 min/islet in the presence of P(+)BCH (n = 7 to 14, p < 0.025 in both cases). The relative magnitude of the menadione-induced increment in L-glutamine oxidation was thus similar in the presence or absence of P(+)BCH, with an overall mean values of + 30.3 f 8.5%.

Oxidation of Endogenous
Fatty Acids-According to Fig.  2, /?(+)BCH would augment 02 utpake by islets exposed to L-glutamine by 26.6 + 5.9 pmol/60 min/islet. This represents no more than 5.5% of the basal 02 consumption in the present system (16) and would hardly account for the synergistic effect of the two amino acids upon insulin release. This apparent discrepancy between respiratory and secretory rates led us to investigate whether /?(+)BCH and/or L-glutamine interfere with the oxidation of endogenous fatty acids, which may act as a fuel to cover part of the energy expenditure by islets deprived of exogenous nutrient (9,10). The data summarized in Table II indicate that L-glutamine decreased by approximately 50% the output of 14C02 from islets prelabeled with [U-'4C]palmitate, whereas /3(*)BCH slightly augmented 14C02 output (p < 0.05). In the presence of both amino acids, the output of 14C02 was lower than in the absence of exogenous nutrient (p < 0.05), but much higher than in the sole presence of L-glutamine (p < 0.001). These findings indicate that /?(+)BCH antagonizes the sparing action of L-glutamine upon the oxidation of endogenous fatty acids (4).
If, as proposed elsewhere (7,16), the latter oxidation accounts for 40% of the basal respiratory rate (486 f 43 pmol of OJ60 min/islet), the effect of P(+)BCH upon WO2 output   Effects of L-glutamine and /3(*)BCH upon the islet content of adenine nucleotides, pyridine nucleotides, and 45Ca All measurements were performed after a 30-min incubation, except for the net uptake of 45Ca which was measured after 90 rnin of incubation and the concentration of cyclic AMP which was measured after a 60-min incubation. The statistical indices in the footnotes refer to the effect of L-glutamine (second column) or /3(f)BCH (third column) relative to basal value (fust column), and to the effect of both amino acids (fourth column) relative to results found in the sole presence of /3(+)BCH (third column).

mM L-glutamine, no B(&)BCH
No L-glutamine, 1.0 mM L-glutamine, 20 In the presence of both amino acids, the increase in ATP content, ATP/ADP ratio, and adenylate charge was less marked than in the sole presence of P(+-)BCH. The NADPH/ NADP+ ratio was strikingly increased in the presence of both amino acids. As expected (12,18), there was a close parallelism between the effect of these amino acids upon 45Ca net uptake and either cyclic AMP concentration (Table 111) or insulin release ( r = 0.9706; p < 0.03).

DISCUSSION
L-Glutamine causes a dose-related enhancement of insulin release evoked by P(+)BCH (Fig. 1). This enhancing action coincides with a more reduced state of the NADPH/NADP' ratio, a stimulation of 45Ca net uptake, and an increase in the concentration of cyclic AMP (Table 111). These findings are compatible with the view that the generation of reducing equivalents, the net uptake of Ca2', the activity of adenylate cyclase, and the release of insulin are somehow related in the process of nutrient-induced insulin release (11,12,18).
Our metabolic data indicate that P(k)BCH dramatically augmented the oxidative deamination of glutamate derived from exogenous glutamine (Fig. 2), in good agreement with the knowledge that P(-)BCH activates glutamate dehydrogenase in the islets (3). The activation of the enzyme coincided with an increased oxidation of ~-[U-'~C]glutamine (Fig. 4). The latter effect was most marked during the initial period of exposure to P(k)BCH. It was previously shown that, in the absence of exogenous glutamine, the effect of P(-t)BCH upon the consumption of 02, oxidation of endogenous glutamate and net uptake of 45Ca also tends to fade out during prolonged exposure to the leucine analog (2, 19).
In the absence of P(+)BCH, the redox state of pyridine nucleotides and the concentration of ATP participate in the regulation of glutamine oxidation by the islets (15). Such was apparently also the case in the presence of P(+)BCH, since menadione, which lowers the islet content of NADPH and NADH ( l l ) , augmented to the same extent the oxidation of ~-[U-'~C]glutamine whether in the absence or presence of P(f)BCH. It should be stressed, therefore that P(k)BCH augmented glutamine oxidation, despite a concomitant increase in the islet ATP content and NADPH/NADP+ ratio.
In addition to the activation of glutamate dehydrogenase, the influence of P(+)BCH upon islet metabolism was characterized by two rather unexpected features.
First, P(+)BCH inhibited the conversion of exogenous Lglutamine to glutamate. This is unlikely to be due to a decrease in the uptake of L-glutamine by the islet cells, since L-glutamine and P(f)BCH are not transported by the same carrier system (20). A possible explanation for the decrease in glutamine deamidation would be that the BCH-induced increase in the mitochondrial generation of %ketoglutarate results in inhibition of glutaminase, an enzyme reported to be also located in mitochondria (21,22). This explanation is consistent with the fact that, in the islets as in other tissues (23), 2-ketoglutarate indeed inhibits glutaminase (Fig. 3).
The second unexpected feature consisted in the fact that P(+)BCH augmented the oxidation of endogenous fatty acids and counteracted the inhibitory effect of L-glutamine upon such an oxidation (Table 11). In our opinion, this is a most important finding. Indeed, it was consistently observed that the capacity of nutrient secretagogues to stimulate insulin release cannot be adequately explained by their capacity to be metabolized in the islets, if one ignores the effect of these secretagogues upon the utilization of endogenous nutrients (4,7,9,10,24). Likewise, in the present experiments, the influence of P ( f ) B C H upon the metabolism of exogenous L-glutamine was not sufficient to account for the synergistic effects of these two amino acids upon insulin release. However, when allowance was made for the changes in the oxidation of endogenous fatty acids, the estimated increase in O2 uptake and the observed value for insulin release yielded values close to those expected from the usual relationship between these two variables (17).
We now wish to discuss the mechanism by which L-glutamine and P(+)BCH affect the oxidation of endogenous fatty acids. It is unlikely that the islet content in ATP (or oxidized pyridine nucleotides) represents a major regulatory factor.
Thus, both P(?)BCH and L-leucine prevent the fall in ATP normally seen in islets deprived of exogenous nutrient (2,24) and, nevertheless, these two amino acids exert opposite effects upon the output of 14C02 from islets prelabeled with [U-"C] palmitate (24). An alternative explanation would be that the availability of oxaloacetate as an acceptor of acetyl-coA residues represents a rate-limiting factor in the oxidation of endogenous fatty acids. This explanation is supported by the following considerations. After correction for the value found in the presence of antimycin A, the oxidation of endogenous fatty acids is decreased by 62.0 f 5.2 and 92.9 k 6.1% in the presence of L-glutamine, 1.0 and 10.0 mM, respectively (4). This coincides with an increase in aspartate production of 11.2 f 2.3 and 16.4 f 2.2 pmol/60 min/islet (4). Thus, by diverting oxaloacetate to aspartate, L-glutamine may well decrease the availability of oxaloacetate for circulation in the Krebs cycle. The latter view is supported by both the observation that the islet content of oxaloacetate is indeed slightly reduced at the high concentration of L-glutamine (4) and the knowledge that, in the islets as in other tissues (25), L-glutamine decreases the flow rate in the segment of the Krebs cycle between oxaloacetate and 2-ketoglutarate (4). In mirror of such a situation, P(+)BCH increases the generation and further catabolism of 2-ketoglutarate without increasing aspartate production and, hence, may increase the availability of oxaloacetate. Last, in the presence of both L-glutamine and P(?)BCH and relative to the situation found in the sole presence of L-glutamine, the generation and further catabolism of 2-ketoglutarate is again increased (Fig. 2) whereas the production of aspartate tends to be decreased, thus leaving more oxaloacetate available for circulation through the Krebs cycle.
In conclusion, the present work demonstrates that the enhancing action of L-glutamine upon BCH-induced insulin release is attributable to the changes evoked by these two amino acids in the metabolism of nutrients in the islet cells. The present study affords direct support to the concept that the insulin secretory response to nutrients is tightly related to their capacity to stimulate catabolic events in the islet cells ( 5 ) . In this respect, nutrients may act as a fuel and/or an enzyme activator (26). In other words, the combined effect of L-glutamine and P(+)BCH upon both islet metabolism and insulin release illustrates that the substrate-site and regulatory-site hypotheses for insulin release, as f i s t defined by Randle et al. (27), are not necessarily exclusive of one another.