Insulin Exocytosis and Glucose-mediated Increase in Cytoplasmic

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 1996 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 271, No. 32, Issue of August 9, pp. 19074 –19079, 1996 Printed in U.S.A.

Insulin Exocytosis and Glucose-mediated Increase in Cytoplasmic Free Ca21 Concentration in the Pancreatic b-Cell Are Independent of Cyclic ADP-ribose* (Received for publication, March 26, 1996, and in revised form, May 20, 1996)

Dominic-Luc Webb, Md. Shahidul Islam‡, Alexandre M. Efanov, Graham Brown, Martin Ko¨hler, Olof Larsson, and Per-Olof Berggren§ From the Rolf Luft Center for Diabetes Research, Department of Molecular Medicine, Karolinska Institute, Karolinska Hospital, S-171 76 Stockholm, Sweden

Stimulation of pancreatic b-cells by glucose gives rise to an increase in the cytoplasmic free calcium concentration ([Ca21]i) and exocytosis of insulin. Cyclic adenosine 5*-diphosphate ribose (cADPR), a metabolite of b-NAD1, has been reported to increase [Ca21]i in pancreatic b-cells by releasing Ca21 from inositol 1,4,5trisphosphate-insensitive intracellular stores. In the present study, we have examined the role of cADPR in glucose-mediated increases in [Ca21]i and insulin exocytosis. Dispersed ob/ob mouse b-cell aggregates were either pressure microinjected with fura-2 salt or loaded with fura-2 acetoxymethyl ester, and [Ca21]i was monitored by microfluorimetry. Microinjection of b-NAD1 into fura-2-loaded b-cells did not increase [Ca21]i nor did it alter the cells’ subsequent [Ca21]i response to glucose. Cells microinjected with the cADPR antagonist 8NH2-cADPR increased [Ca21]i in response to glucose equally well as those injected with cADPR. Finally, the ability of cADPR to promote exocytosis of insulin in electropermeabilized b-cells was investigated. cADPR on its own did not increase insulin secretion nor did it potentiate Ca21-induced insulin secretion. We conclude that cADPR neither plays a significant role in glucosemediated increases in [Ca21]i nor interacts directly with the molecular mechanisms regulating exocytosis of insulin in normal pancreatic b-cells. Stimulation of insulin secretion from pancreatic b-cells by glucose is accompanied by an increase in the [Ca21]i.1 According to the currently accepted model, glucose metabolism increases the intracellular ATP/ADP ratio leading to closure of ATP-sensitive potassium channels (KATP channel), cell depolarization, opening of voltage-gated Ca21 channels, an increase in [Ca21]i, and ultimately exocytosis of insulin (1). In addition to Ca21 entry through voltage-gated Ca21 channels, release of Ca21 from intracellular Ca21 pools increases * This work was supported by Grants 03x-09890, 19F-10698, 04x09891, and 19x-00034 from the Swedish Medical Research Council, the Juvenile Diabetes Foundation International, the Swedish Diabetes Association, the Nordic Insulin Foundation, NOVO Industry, the Swedish Medical Society, and Funds of the Karolinska Institute. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ Researcher at the Swedish Medical Research Council. § To whom correspondence should be addressed. Tel.: 46 8-7295731, 7292197; Fax: 46 8-303458, 7293658, 7295731. 1 The abbreviations used are: [Ca21]i, cytoplasmic free calcium concentration; b-NAD1, b-nicotinamide-adenine dinucleotide (oxidized form); cADPR, cyclic adenosine 59-diphosphate ribose; 8NH2-cADPR, 8-amino-cyclic adenosine 59-diphosphate ribose; Ins(1,4,5)P3, inositol 1,4,5-trisphosphate; pCa, -log[Ca21]; CCh, carbamylcholine.

[Ca21]i. The most well known second messenger that releases Ca21 from intracellular stores is inositol 1,4,5-trisphosphate (Ins(1,4,5)P3), formed after activation of phospholipase C-linked receptors (2). The presence of additional, nonmitochondrial, Ca21 pools, which are insensitive to Ins(1,4,5)P3, has been discussed for a considerable time (3, 4). More recently, we have shown that the sulfhydryl reagent thimerosal, in the presence of mitochondrial blockers, releases Ca21 from an Ins(1,4,5)P3-insensitive pool in the pancreatic b-cell (5). The significance and mechanisms of regulation of this pool are poorly understood. One possible endogenous modulator is cyclic adenosine 59-diphosphate ribose (cADPR), a metabolite of b-NAD1. cADPR mobilizes Ca21 from intracellular pools in sea urchin eggs and several mammalian cell types (6). cADPR is synthesized and degraded by the human leukocyte antigen, CD38 (7). Enzymatic activity for the formation and degradation of cADPR appears to be present in pancreatic islets, as evidenced by the presence of comparatively high levels of mRNA for CD38 in these cells (8). The physiological significance of cADPR-mediated Ca21 signaling in mammalian cells is still a major controversy and is currently a topic of many investigations. It has been reported that cADPR, but not Ins(1,4,5)P3, mobilizes Ca21 from microsomes prepared from rat islets and releases insulin from digitonin-permeabilized rat islets (9). Consequently, cADPR has been postulated to play an important role in the stimulussecretion coupling mechanism in pancreatic b-cells. The proposed model, as it has been presented (10, 11), is difficult to reconcile with the established roles of the KATP channel and the voltage-gated Ca21 channel in the insulin secretory process. A further problem with the proposed model is the apparent inconsistency in the time required for cADPR synthesis, 5–15 min (12), compared with the time for the glucose-induced rise in [Ca21]i to occur, about 1– 4 min. Moreover, it has not always been possible to demonstrate cADPR-mediated Ca21 release from insulin-secreting cells (13, 14). In the present report, we have investigated in detail a possible role for cADPR in glucosemediated increase in [Ca21]i and in the molecular mechanisms directly promoting insulin exocytosis in normal pancreatic b-cells. MATERIALS AND METHODS

Isolation of Islets and Preparation of b-Cells—Pancreatic islets from 10 –12-month-old ob/ob mice from a local non-inbred colony were isolated (one mouse per isolation) by collagenase digestion and dispersed into small cell clusters by briefly shaking in a Ca21- and Mg21-deficient medium, as described previously (15). Cells were then washed and plated onto glass coverslips and incubated at 37 °C in RPMI 1640 medium containing 11 mM glucose and supplemented with fetal calf serum (10%, v/v), penicillin (40 i.u./ml), streptomycin (40 mg/ml), and 2% (v/v) L-glutamine. Chemicals—8NH2-cADPR and some samples of cADPR were the

19074

Effects of cADPR on [Ca21]i and Insulin Exocytosis generous gifts of Dr. T. F. Walseth. ADPR and all other samples of cADPR were from Sigma. Fura-2 pentapotassium salt was from Molecular Probes, Eugene, OR. b-NAD1 (crystallized free acid) was from Boehringer Mannheim GmbH, Germany. RPMI medium and fetal calf serum were from ICN, Costa Mesa, CA. Trypsin, glutamine, and penicillin/streptomycin were from HyClone, Cramlington, U. K. All other chemicals were from Sigma. Measurement of [Ca21]i by Microfluorimetry—Cells attached to coverslips were loaded with fura-2 by incubating in basal medium containing (in mM): NaCl 125, KCl 5.9, MgCl2 1.2, CaCl2 1.28, Hepes 25, glucose 3, BSA 0.1% (pH 7.4 with NaOH), and fura-2 acetoxymethyl ester (fura-2am) 1 mM, for 45 min at 37 °C. Coverslips were washed twice with the buffer and mounted as the bottom of an open chamber placed on the stage of an inverted epifluorescence microscope (Zeiss, Axiovert 35M). The stage of the microscope was thermostatically controlled to maintain a temperature of 37 °C in the chamber. The microscope was connected to a SPEX fluorolog-2 CM1T11I system for dual wavelength excitation fluorimetry. The excitation wavelengths generated by two monochromators were directed to the cells by a dichroic mirror. The emitted light selected by a 500 –530-nm filter was monitored with an IBM-compatible PC interfaced with a photomultiplier attached to the microscope. The excitation wavelengths were alternated at a frequency of 1 Hz, and the data acquisition time for each wavelength was 0.33 s. The emissions at the excitation wavelength of 340 nm (F340) and 380 nm (F380) were used to calculate the fluorescence ratio (F340/F380), yielding relative changes in [Ca21]i. Cells were optically isolated by means of a diaphragm situated between the cells and the photomultiplier and recorded using a 403 objective lens (Zeiss, Plan Neofluar). Microinjection—Cells were allowed to attach to coverslips by incubating for at least 4 h. Microinjection was then performed essentially as described previously (16), using an Eppendorf 5242 pressure microinjector and a 5170 micromanipulator fitted to the stage of the microscope. Pipettes were made from borosilicate capillary tubes, 1.6 3 1.28 3 7.5 mm (Hilgenberg, FRG), using a pipette puller manufactured in this laboratory. Injection tips had an outer diameter of less than 0.5 mm with a taper of ;5°. The microinjection solution consisted of (in mM): potassium gluconate 140, NaCl 5, MgCl2 1, EGTA 10, and Hepes 25 (pH 7.00 with KOH). Following each experiment where fura-2 was injected, another experiment using fura-2am-loaded aggregates of the same approximate size was performed with the same parameter settings on the recording equipment. This was done to verify that the F340 and F380 in the microinjected cells were comparable with those in fura-2am-loaded cells and to determine if cells were damaged by microinjection. A transient elevation in [Ca21]i occurred at the start of the experiments, which represented an injection artifact partly due to differences in [Ca21] in the injection solution and the intracellular mileau. Comparison of confocal microscopic images of fura-2 salt-injected cells against fura-2amloaded cells revealed identical distribution and fluorescence intensity of the dye in both the nuclear and cytosolic regions (data not shown). These experiments were used to relate concentrations of injected fura-2 salt to the actual concentration of dye in fura-2am-loaded cells. The minimum concentration of intracellular fura-2 in fura-2am-loaded ob/ob b-cells has been estimated from a previous study to be about 100 mM (17). When fura-2 salt was microinjected at a concentration 250 –500 times this minimum intracellular concentration, fluorescence intensities comparable with those in fura-2am-loaded cells were achieved. Hence, any substance coinjected with the dye was predicted to also be diluted approximately 250 –500-fold. Except for b-NAD1, test substances were dissolved in the microinjection solution along with fura-2 salt. To ensure that the desired concentrations were reached within the cell subsequent to microinjection, the concentration of fura-2 was 25 mM and that of the test substances was 1000 times the desired final intracellular concentration. b-NAD1 was injected without fura-2 into fura2am-loaded cells, because it was necessary to prepare it at concentrations near its solubility limit, and in some cases, a precipitate formed when it was dissolved in the presence of fura-2. Compartmentalization of fura-2 under our conditions was negligible. The mean volume of an ob/ob b-cell was estimated to be 1.1 6 0.2 pl (n 5 77), as calculated from capacitance measurements of whole-cell patch-clamp recordings. From these data the effective microinjection volume was calculated to be 4.4 fl (i.e. less than 1% of the cell volume). Criteria used to define a successful microinjection were (a) the F340 and F380 in injected cells were at least twice that of the neighboring uninjected cells; (b) the base line fluorescence was stable over time similar to that of fura-2am-loaded cells; (c) the F340/F380 was similar to that of fura-2am-loaded cells. Damaged cells registered high F340/ F380, as a result of leakage of extracellular Ca21 into the cells, and lost

19075

FIG. 1. Microfluorimetric measurement of [Ca21]i in b-cells microinjected with b-NAD1. Fura-2am-loaded b-cells, in a basal perifusion buffer containing 1.28 mM CaCl2 and 3 mM glucose, were microinjected with 20 mM (pipette concentration) b-NAD1, increasing the intracellular concentration of free b-NAD1 by ;80 mM. Fura-2 fluorescence at 340 (F340) and 380 (F380) nm was then recorded within 15 s after microinjection. F340 and F380 fluorescences were measured by microfluorimetry without stimulating the cells for 10 min. After 10 min, cells were stimulated with 200 mM carbamylcholine (CCh) for 30 s, indicating that CCh, through formation of Ins(1,4,5)P3, but not b-NAD1, was able to induce an increase in [Ca21]i. Cells were then stimulated with 11 mM glucose (Glu) until peak response was reached, indicating that injected cells were still able to metabolize glucose and mediate an increase in [Ca21]i. Hence, all enzymatic activity required for glucose-mediated increase in [Ca21]i was retained in the microinjected b-cells. The initial decline in the F340/F380 during the first 150 s of the recording is typical of the recovery phase in injected b-cells and is also seen in the injected controls. This recovery phase does not mask possible changes in [Ca21]i, however, as evidenced by the CCh responses during this phase in Fig. 2. The cells were also challenged with 25 mM KCl at the end of each experiment to indicate that the cells were able to efficiently regulate [Ca21]i for the entire duration of the experiments. This trace is representative of five recordings. fluorescence quickly, due to leakage of dye. Such cells were discarded. For all traces, the diaphragm and other settings on the microfluorimetry equipment were kept the same. Following completion of these experiments, calibrations were made to determine the range of changes in [Ca21]i. Applying the Grynkiewicz et al. equation (18) and correcting for variations in light output by the two independent monochromators, it was determined that the [Ca21]i ranged from ; 50 nM, at a basal glucose concentration of 3 mM, to 300 nM, when fully stimulated by 11 mM of the sugar. Insulin Secretion—Following isolation and dispersion of b-cells as described above, cells were washed three times in a permeabilization buffer consisting of (in mM) potassium glutamate 140, NaCl 5, MgCl2 1, EGTA 10, Hepes 25, and 0.025% albumin (w/v). pH was adjusted to 7.00 with KOH, and pCa was adjusted to 7.00 with CaCl2. The actual pCa values in the incubation buffers were checked by using a Ca21-selective electrode (model 93–20, Orion Research Inc., Boston, MA). Cells were subsequently electropermeabilized in this buffer by 5 pulses of a 3 kV/cm electric field. This preparation of permeabilized cells was aliquoted into five portions which were then centrifuged and resuspended in a modified permeabilization buffer containing 2 mM MgATP, 2 mM creatine phosphate, 10 units/ml creatine phosphokinase, and pCa ranging from 8 to 4, and pH 7.0 (pH and pCa adjusted with KOH and CaCl2, respectively). Ca21 standard solutions were prepared according to the methods of Tsien and Rink (19). Each of the five aliquots of cells were then split into three aliquots for a total of 15 fractions. Each of the five Ca21 concentrations received either 2 mM cADPR, 2 mM ADPR, or no additional treatment as control. ADPR, which differs from cADPR by only one H2O molecule as it is the open-chain hydrolyzed form of cADPR, served as an additional control. Cells were then seeded into microtiter well plates with each parameter at each Ca21 concentration being divided into five wells. Hence, each preparation had five data points per parameter at each of the Ca21 concentrations. Cells were incubated for 15 min at 37 °C and 5.0% CO2. Insulin release was measured by radioimmunoassay.

19076

Effects of cADPR on [Ca21]i and Insulin Exocytosis

FIG. 2. Microfluorimetric measurement of [Ca21]i in b-cells microinjected with 8NH2-cADPR. A, b-cells were microinjected with a mixture of 0.34 mM 8NH2-cADPR and fura-2 salt and recorded by microfluorimetry. Cells were initially stimulated with CCh to verify responsiveness to stimulation during the recovery phase. Thereafter, cells were stimulated with 11 mM glucose and at the end of the experiment with 25 mM KCl. B, b-cells microinjected with fura-2 alone in the injection buffer gave glucose responses comparable with 8NH2-cADPR-injected cells. The cells were first challenged with CCh in order to maintain a consistent protocol. C, b-cells loaded with fura-2am gave glucose responses that were not significantly different from those of the injected cells. Fluorescence intensities and, hence, concentrations of fura-2 in fura-2-injected cells and fura-2am-loaded cells were comparable. From this, it was possible to calculate a dilution of ;250-fold for microinjected substances. It was then also possible to compare 8NH2-cADPR and cADPR-injected cells to fura-2am-loaded cells. Note that in all of these recordings it is possible to clearly distinguish the changes in [Ca21]i in response to CCh from the recovering base lines. The traces are representative of 10 recordings. See also Tables I and II.

Calculations and Statistics—For microinjection experiments, three parameters were used to compare initial [Ca21]i responses to glucose of test conditions against controls: 1) elapsed time from glucose stimulation until initiation of glucose induced [Ca21]i increase; 2) height of response; and 3) slope, which would identify any changes in the kinetics of the glucose responses. The response pattern beyond the initial rise showed considerable variability, accounted for by variability in cell preparations. Therefore, it was only meaningful to compare the initial responses with glucose. Statistical significance of glucose responses was judged by Student’s t test for paired data and represents the comparison of 8NH2-cADPR or cADPR-injected cells against 1) fura-2-injected control cells and 2) fura-2am-loaded control cells. For insulin secretion data, each preparation of cells, representing one mouse each, was taken as one unit of n (i.e. five preparations yielded n 5 5). Data are presented as standard mean 6 S.E., and p , 0.05 was taken as the significance level.

RESULTS

To test whether b-NAD1 can induce Ca21 release from intracellular Ca21 pools through metabolism to cADPR, we microinjected the substance into b-cells loaded with fura-2am and monitored changes in [Ca21]i. In these experiments, the concentration of b-NAD1 in the pipette was 20 mM, which is estimated to increase the free cytosolic b-NAD1 concentration by ;80 mM. Microfluorimetry recording was started within 15 s after microinjection and continued without stimulation for 10 min. As shown in Fig. 1, no changes in [Ca21]i were observed in five out of five experiments from three preparations. Such microinjection of b-NAD1 also did not affect the cells’ subsequent responses to stimulation by glucose (11 mM), carbamylcholine (200 mM), or depolarization with KCl (25 mM) in all five

Effects of cADPR on [Ca21]i and Insulin Exocytosis TABLE I Glucose-induced increase in [Ca21]i in b-cells microinjected with 8NH2-cADPR or cADPR Pancreatic b-cells in a basal buffer containing 3 mM glucose were microinjected with 8NH2-cADPR (top part) or cADPR (bottom part) in the presence of fura-2 and immediately recorded by microfluorimetry. Injected cells were challenged with 11 mM glucose. The pattern of [Ca21]i increase in injected cells was compared against cells that were injected with fura-2 containing vehicle only and cells loaded with fura2am. In no case was there any statistical difference between the experimental group and the controls with regard to any of the three parameters we examined. The time parameter represents the time elapsed from switching to 11 mM glucose containing buffer until the initial rise in F340/380. Height is the initial increase in F340/380 from base line to peak in response to 11 mM glucose. Slope is the rate of the initial increase in F340/380 multiplied by 1000 (for ease of reading). Values are reported as standard mean 6 S.E. of n experiments on at least five preparations of cells. The actual [Ca21]i ranged from ;50 nM, at a basal glucose concentration of 3 mM, to 300 nM, when fully stimulated by 11 mM of the sugar, as estimated by the Grynklewicz et al. equation (18). The differences in the time parameter of the top part of table compared with the bottom part of table were due to the use of tubings of different dimensions, which were used to introduce buffer into the recording chamber. This did not complicate the results, however, since the same tubings were used in each series of experiments (top and bottom parts representing separate series of experiments). Parameter

Time Height Slope

Parameter

Time Height Slope

8NH2-cADPR injected (n 5 10)

Vehicle only (n 5 10)

Fura-2am loaded (n 5 10)

64 6 6 0.44 6 0.11 11.49 6 3.81

62 6 12 0.40 6 0.09 10.39 6 3.08

cADPR injected (n 5 14)



208 6 59 0.55 6 0.08 12.80 6 4.04

205 6 59 0.45 6 0.11 11.5 6 4.00

68 6 12 0.49 6 0.05 8.86 6 1.57

204 6 62 0.45 6 0.11 7.38 6 2.67

experiments, indicating that the cells were intact following microinjection. We then examined whether cADPR is involved in glucosestimulated increase in [Ca21]i by two approaches. In the first approach, we injected a mixture of fura-2 and the cADPR antagonist, 8NH2-cADPR (20, 21), into the cells and then stimulated them with glucose. The concentration of 8NH2-cADPR in the pipette was 0.34 or 1 mM, and intracellular concentrations of the substance after injection were estimated to increase by ;1.4 and 4.0 mM, respectively. Again, all cells injected with 8NH2-cADPR responded to glucose by an increase in [Ca21]i (Fig. 2), which was not significantly different from controls (Table I, top). As seen in Table II, top, these cells also responded normally to carbamylcholine (200 mM) and depolarization by KCl (25 mM). It is noteworthy that the glucose response is also not abolished by prior depletion of the Ins(1,4,5)P3sensitive pool, indicating that neither of these intracellular pools are involved in the [Ca21]i increase promoted by the sugar. The rapid return to the base line of the [Ca21]i, following the brief KCl stimulations (15 s) at the end of the experiments, indicates that the microinjected cells remained intact for the entire duration of the experimental period. In the second approach, we injected a mixture of fura-2 and cADPR into the cells and then stimulated them with glucose. We reasoned that prior exposure to cADPR would deplete the putative cADPR-sensitive Ca21 stores or desensitize the cADPR-mediated Ca21 release mechanism. In these experiments, the concentration of cADPR in the pipette was 1 mM, and the intracellular concentration of cADPR after injection was estimated to increase by ;4.0 mM. As shown in Fig. 3, when cells were stimulated with glucose about 1 min after injection,

19077

TABLE II Carbamylcholine and KCl-induced increase in [Ca21]i in b-cells microinjected with 8NH2-cADPR or cADPR Pancreatic b-cells in a basal buffer containing 3 mM glucose were microinjected with 8NH2-cADPR (top) or cADPR (bottom) in the presence of fura-2 and immediately recorded by microfluorimetry. Values for fura-2am are included for comparison and indicate that the microinjection technique did not significantly affect the responses to CCh and KCl. These responses were not affected by inhibition of the putative cADPR receptor by 8NH2-cADPR, when compared with cells injected by cADPR or vehicle alone. Values represent peak [Ca21]i elicited by CCh and KCl and are standard mean 6 S.E. of n experiments on at least five preparations of cells. The actual [Ca21]i ranged from ;50 nM at the base line to 225 nM when fully stimulated by either CCh or KCl, as determined by the Grynkiewicz et al. equation (18). Parameter

CCh KCl

Parameter

CCh KCl

8NH2-cADPR injected (n 5 10)



0.35 6 0.03 0.33 6 0.04

0.32 6 0.05 0.34 6 0.05

0.32 6 0.03 0.36 6 0.04

cADPR injected (n 5 14)



0.29 6 0.08 0.31 6 0.05

0.30 6 0.06 0.31 6 0.05

0.28 6 0.08 0.31 6 0.06

they responded by an increase in [Ca21]i which was not significantly different from controls (Table I, bottom). Again, as expected, these cells also responded normally to carbamylcholine and depolarization with KCl (Table II, bottom). Finally, we tested the possibility that cADPR might induce insulin secretion by a mechanism that is not dependent on an increase in [Ca21]i or by potentiating Ca21-induced insulin secretion. We calculated the units of insulin secreted per unit time from the untreated controls and obtained basal values of ;3.4 milliunits/106 cells/15 min at pCa 5 8, which increased to a maximum of ;4.5 milliunits/106 cells/15 min at pCa 5 5. These values were similar to published data for glucose responses in normal mice and rats (22, 23). As shown in Fig. 4, increasing pCa from 8 to 5 induced an ;1.6-fold rise in insulin secretion (p , 0.005, n 5 five preparations) with a basal secretion at pCa 5 8, maximum secretion at pCa 5 5, and an EC50 at pCa 5 ;6. cADPR did not significantly alter insulin secretion at any of these Ca21 concentrations, compared with untreated controls or ADPR-treated cells. DISCUSSION

One major finding of this study was that increasing the free cytoplasmic concentration of b-NAD1 in normal pancreatic b-cells, by microinjection of the free acid, does not release Ca21 from intracellular stores even when b-NAD1 is given sufficient time for enzymatic conversion to cADPR. This is consistent with our previous demonstration that cADPR does not induce significant Ca21 release in this type of endocrine cell (13). In that study, we used electropermeabilized cells and the wholecell mode of the patch-clamp technique, which left open the possibility that the cADPR effect was not observed because cytosolic factors necessary for mediation of the effect, such as calmodulin, were diluted. In the present study, by microinjecting b-NAD1 into intact b-cells, we have excluded this possibility. Dilution of factors necessary for cADPR activity seems unlikely since all factors required for glucose-mediated increase in [Ca21]i (i.e. glycolytic enzymes) were apparently unaltered. This is evidenced by the fact that the cells responded to glucose following injection of b-NAD1. The glucose response requires the presence of glycolytic enzymes in the cytoplasm in order to subsequently generate ATP which ultimately results in closure of KATP channels, membrane depolarization, and

19078


FIG. 3. Microfluorimetric measurement of [Ca21]i in b-cells microinjected with cADPR. A, cells were microinjected as in Fig. 2 with a mixture of cADPR and fura-2. Thereafter, cells were stimulated with 11 mM glucose and at the end of the experiment with 25 mM KCl. Again, CCh stimulation was merely for consistency of the protocol and indicated that the cells were measurably responsive to stimulation during the recovery phase. B, b-cells microinjected with fura-2 alone in the injection buffer gave glucose responses that were not significantly different from those of cADPR-injected cells. The second KCl response was included to verify that the initial KCl response was not simply due to washout of glucose. C, b-cells loaded with fura-2am gave comparable glucose responses to the injected cells. Note that in all of these recordings it is possible to clearly distinguish the responses to CCh from the recovering base lines. The traces are representative of 14 recordings. See also Tables I and II.

opening of voltage-gated Ca21 channels leading to the observed Ca21 influx. Had the glycolytic enzymes been diluted to any appreciable extent, the Ca21 influx would not have been observed or would have been markedly reduced. Since the mRNA for CD38 is in abundance in the pancreatic islet (8), it is unlikely that the lack of effect of b-NAD1 could be explained by dilution of the enzyme. Moreover, despite the fact that the cells were unable to respond to b-NAD1, they were able to release Ca21 from intracellular stores when stimulated with carbamylcholine, which acts through formation of Ins(1,4,5)P3. Another major finding in the present study was that glucosemediated increase in [Ca21]i is unaltered whether intact b-cells are microinjected with cADPR or 8NH2-cADPR. Hence, the stimulus-response mechanism of the b-cell was both functional and unaltered, at least with respect to changes in [Ca21]i, whether or not cADPR was permitted to bind to its putative intracellular receptor and whether or not the putative receptor was desensitized prior to glucose stimulation. It appears that cADPR is not an essential component of the stimulus-secretion

coupling mechanism with respect to the well established changes in [Ca21]i linked to glucose stimulation of the normal pancreatic b-cell. Finally, we examined whether cADPR may act as a classical cyclic nucleotide by inducing insulin release via a direct mechanism that is not dependent on an increase in [Ca21]i. We have also excluded this possibility by demonstrating that cADPR cannot on its own induce insulin secretion nor can it potentiate Ca21-induced insulin secretion. Although we cannot exclude that electropermeabilization causes leakage of some intracellular components, as discussed earlier, this method has been used successfully in many investigations of the molecular mechanisms directly involved in the regulation of exocytosis. The applied electropermeabilization protocol has been reported to create clean holes in the plasma membrane, leave the intracellular structure unagitated, and prevent leakage of molecules with molecular masses .5 kDa (24). Assuming this is correct, calmodulin, which is believed to be an essential cofactor for cADPR action (25) and has a molecular mass of 15–19 kDa,


19079

demonstrating that cADPR does not function as an intracellular signal in the b-cell stimulus-secretion coupling mechanism. Acknowledgment— We are grateful to Dr. George E. N. Kass and Dr. Sek C. Chow for useful discussions and to Dr. Timothy Walseth for preparing cADPR and 8NH2-cADPR. Thanks also to Kjell Bernehammar for statistical insights. We thank Olav Nordli for technical assistance related to microfluorimetry equipment. REFERENCES

FIG. 4. Ca21-induced insulin secretion in electropermeabilized b-cells. Electropermeabilized b-cells were incubated for 15 min in the presence of 2 mM cADPR, 2 mM ADPR, or untreated at Ca21 concentrations ranging from pCa 5 8 to pCa 5 4. The increase in insulin secretion was solely due to the increases in pCa. Compared with ADPR-treated cells and untreated controls, cADPR did not on its own increase basal insulin secretion at pCa 5 8. cADPR also did not alter Ca21-induced insulin secretion at pCa 5 6 (EC50 for insulin secretion) or at pCa 5 5 (maximum insulin secretion). Each bar is the mean 6 S.E. obtained from five separate preparations (n 5 5) and is expressed as the percentage of control insulin release at basal Ca21, pCa 5 8.

would not be expected to leak from cells permeabilized by our method. Thus, these experiments negate the possibility that the lack of effect of cADPR could be due to leakage of calmodulin or insufficient [Ca21]i. Our findings are in clear contrast to those of Takasawa et al. (9) who first postulated the cADPR-mediated insulin secretion model from which other hypotheses were derived (10, 11). Conversely, our present findings are in close agreement with our earlier findings (13) and those of Rutter et al. (14), clearly

1. Berggren, P.-O., Arkhammar, P., Islam, M. S., Juntti-Berggren. L., Khan, A. Kindmark, H., Ko¨hler, M., Larsson, O., Nilsson, T., Sjo¨holm, Å., Szecowka, J., and Zhang, Q. (1993) Adv. Exp. Med. Biol. 334, 24 – 45 2. Berridge, M. J. (1993) Nature 361, 315–325 3. Streb, H., Irvine, R. F., Berridge, M. J., and Shultz, I. (1983) Nature 306, 67– 69 4. Nilsson, T., Arkhammar, P., Hallberg, A., Hellman B., and Berggren, P.-O. (1987) Biochem. J. 248, 329 –336 5. Islam, M. S., Rorsman, P., and Berggren, P.-O. (1992) FEBS Lett. 296, 287–291 6. Lee, H. C., Walseth, T. F., Bratt, G. T., Hayes, R. N., and Clapper, D. L. (1989) J. Biol. Chem. 264, 1608 –1615 7. Howard, M., Grimaldi, C. J., Bazan, J. F., Lund, F. E., Santos-Argumedo, L., Parkhouse, R. M. E., Walseth, T. F., and Lee, H. C. (1993) Science 262, 1056 –1059 8. Takasawa, S., Tohgo, A., Noguchi, N., Koguma, T., Nata, K., Sugimoto, T., Yonekura, H., and Okamoto, T. (1993) J. Biol. Chem. 268, 26052–26054 9. Takasawa, S., Nata, K., Yonekura, H., and Okamoto, H. (1993) Science 259, 370 –373 10. Galione, A. (1993) Science 259, 325–326 11. Takasawa, S., Nata, K., Yonekura, H., and Okamoto, H. (1993) Science 262, 585 12. Clapper, D. L., Walseth, T. F., Dargie, P. J., and Lee, H. C. (1987) J. Biol. Chem. 262, 9561–9568 13. Islam, M. S., Larsson, O., and Berggren, P.-O. (1993) Science 262, 584 –586 14. Rutter, G. A., Theler, J.-M., Li, G., and Wollheim, C. B. (1994) Cell Calcium 16, 71– 80 15. Arkhammar, P., Nilsson, T., Rorsman, P., and Berggren, P.-O. (1987) J. Biol. Chem. 262, 5448 –5454 16. Kass, G. E. N., Webb, D.-L., Chow, S. C., Llopis, J., and Berggren, P.-O. (1994) Biochem. J. 302, 5–9 17. Arkhammar, P., Nilsson, T., and Berggren, P.-O. (1989) Cell Calcium 10, 17–27 18. Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985) J. Biol. Chem. 260, 3440 –3450 19. Tsien, R. Y., and Rink, T. J. (1980) Biochim. Biophys. Acta 599, 623– 638 20. Walseth, T. F., and Lee, H. C. (1993) Biochim. Biophys. Acta 1178, 235–242 21. Lee, H. C., Aarhus, R., and Walseth, T. F. (1993) Science 261, 352–355 22. Zhang, A., Gao, Z.-Y., Gilon, P., Nenquin, M., Drews G., and Henquin, J.-C. (1991) J. Biol. Chem. 266, 21649 –21656 ¨ stenson, C.-G., Ling, Z.-C., and Grill, V. (1995) Endocrinology 23. Zhou Y.-P., O 136, 3546 –3551 24. Knight, D. E., and Scrutton, M. C. (1986) Biochem. J. 234, 497–506 25. Lee, H. C., Aarhus, R., Graeff, R., Gurnack, M. E., and Walseth, T. F. (1994) Nature 370, 307–309