Subcellular distribution and function of Rab3A, B, C, and D isoforms in insulin-secreting cells

(1)

Article

Reference

Subcellular distribution and function of Rab3A, B, C, and D isoforms in insulin-secreting cells

IEZZI-BAKHTIARI, Mariella, et al.

Abstract

Insulin-secreting cells express four GTPases of the Rab3 family. After separation of extracts of INS-1 cells on a sucrose density gradient, the bulk of the A, B, and C isoforms was recovered in the fractions enriched in insulin-containing secretory granules. Rab3D was also mainly associated with secretory granules, but a fraction of this isoform was localized on lighter organelles. Analyses by confocal microscopy of immunostained HIT-T15 cells transfected with epitope-tagged constructs confirmed the distribution of the Rab3 isoforms.

Transfection of HIT-T15 cells with GTPase-deficient mutants of the Rab3 isoforms decreased nutrient-induced insulin release to different degrees (D>B>A>C), while overexpression of Rab3 wild types had minor or no effects. Expression of the same Rab3 mutants in PC12 cells provoked an inhibition of K+-stimulated secretion of dense core vesicles, indicating that, in beta-cells and neuroendocrine cells, the four Rab3 isoforms play a similar role in exocytosis.

A Rab3A/C chimera in which the carboxyterminal domain of A was replaced with the corresponding region of C inhibited insulin secretion as [...]

IEZZI-BAKHTIARI, Mariella, et al . Subcellular distribution and function of Rab3A, B, C, and D isoforms in insulin-secreting cells. Molecular Endocrinology , 1999, vol. 13, no. 2, p. 202-12

DOI : 10.1210/mend.13.2.0228 PMID : 9973251

Available at:

http://archive-ouverte.unige.ch/unige:35090

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

Isoforms in Insulin-Secreting Cells

Mariella Iezzi, Ge´rard Escher, Paolo Meda, Anne Charollais, Giulia Baldini, Franc¸ois Darchen, Claes B. Wollheim, and Romano Regazzi

Division de Biochimie Clinique (M.I., C.B.W.) De´partement de Me´dicine Interne and De´partement de Morphologie (P.M., A.C.) Universite´ de Gene`ve

Geneva, Switzerland 1211

Institut de Biologie Cellulaire et de Morphologie (G.E., R.R.) Universite´ de Lausanne

Lausanne, Switzerland 1005

Department of Anatomy and Cell Biology (G.B.) Columbia University

College of Physicians and Surgeons New York, New York 10032

Institut de Biologie Physico-Chimique (F.D.)

Centre Nationale de la Recherche Scientifique ERS 575 Paris, France 75005

Insulin-secreting cells express four GTPases of the Rab3 family. After separation of extracts of INS-1 cells on a sucrose density gradient, the bulk of the A, B, and C isoforms was recovered in the fractions enriched in insulin-containing secretory granules.

Rab3D was also mainly associated with secretory granules, but a fraction of this isoform was local- ized on lighter organelles. Analyses by confocal microscopy of immunostained HIT-T15 cells trans- fected with epitope-tagged constructs confirmed the distribution of the Rab3 isoforms. Transfection of HIT-T15 cells with GTPase-deficient mutants of the Rab3 isoforms decreased nutrient-induced in- sulin release to different degrees (D>B>A..C), while overexpression of Rab3 wild types had minor or no effects. Expression of the same Rab3 mu- tants in PC12 cells provoked an inhibition of K¹- stimulated secretion of dense core vesicles, indi- cating that, in b-cells and neuroendocrine cells, the four Rab3 isoforms play a similar role in exo- cytosis. A Rab3A/C chimera in which the carboxy- terminal domain of A was replaced with the corre- sponding region of C inhibited insulin secretion as Rab3A. In contrast, a Rab3C/A chimera containing the amino-terminal domain of C was less potent and reduced exocytosis as Rab3C. This suggests that the degree of inhibition obtained after trans-

fection of the Rab3 isoforms is determined by dif- ferences in the variable amino-terminal region.

(Molecular Endocrinology 13: 202–212, 1999)

INTRODUCTION

Pancreatic hormones released by the cells of the islets of Langerhans play a pivotal role in the regulation of nutrient disposal and metabolism. In fact, insulin secretion from pancreaticb-cells is an essential require- ment for the achievement of blood glucose homeosta- sis. Although the molecular details of the process of insulin exocytosis remain to be fully clarified, during the last few years several important components of the machinery permitting the targeting of secretory vesicles to the plasma membrane have been identified (1).

According to current models, the basic components controlling the targeting and fusion of secretory vesicles with the plasma membrane are largely conserved between cell types and between species (2, 3). The docking of transport vesicles to the appropriate membrane is thought to be specified by pairing of proteins located on the vesicle membrane termed v-SNAREs (vesicular SNAP receptors) with their specific partners on the acceptor membrane termed t-SNAREs (target SNAP receptors) (3). Pancreaticb-cells and clonal insulin-secreting cell lines express the v-SNAREs, VAMP-2 and cellubrevin, and the t-SNAREs, SNAP-25

0888-8809/99/$3.00/0 Molecular Endocrinology

202

(3)

and syntaxin-I (4–7). Each of these proteins has been demonstrated to play a role in insulin exocytosis using clostridial neurotoxins or with inactivating antibodies (5–8).

Rab GTPases represent a large family of homolo- gous Ras-like GTP-binding proteins that direct the vectorial movement of secretory vesicles. These regulatory proteins act as molecular switches that flip between two conformational states, the active GTP- bound and an inactive GDP-bound form (9). The activated form of Rab GTPases has been proposed to catalyze the formation of the v-SNARE/t-SNARE complex. Under resting conditions, t-SNAREs are unable to interact efficiently with v-SNAREs as they are bound to members of the Sec-1 protein family (10, 11). Stud- ies performed in yeast suggest that activated Ras-like GTPases located on transport vesicles disrupt the association between t-SNAREs and Sec-1-like proteins on the acceptor membrane (12). The displacement of Sec-1-like proteins would result in the assembly of t-SNAREs and v-SNAREs, leading to the docking of the vesicle to the target membrane (12).

The members of the Rab3 subfamily are the best candidates for controlling docking and fusion of Golgi- derived secretory vesicles with the plasma membrane (9, 13). Four isoforms of Rab3 (Rab3A, -B, -C, and -D) have been identified so far. At the protein level Rab3 isoforms display about 80% amino acid identity and differ almost exclusively in two short variable domains located at the amino- and at the carboxy terminus.

Rab3A and -C are primarily expressed in neuronal and neuroendocrine cells, while Rab3B and -D are more abundant outside the nervous system (9, 13). The precise function of the different Rab3 isoforms remains to be established. Differences in the subcellular localization of Rab3 isoforms (14–16) and in their functional involvement in the exocytotic process (17) have been reported. On the other hand, the relatively mild phenotype observed in Rab3A-deficient mice (18) and the ability of distinct isoforms to interact with the same effectors (17, 19, 20) suggest at least some degree of redundancy between Rab3 proteins.

Initial evidence for the possible implication of Rab3 proteins in insulin secretion came from studies performed in permeabilized cells. In this cell preparation, the introduction of peptides mimicking the putative effector-binding domain of Rab3 proteins stimulates insulin release (21–23). Subsequently, it was found that Rab3A is associated with the membrane of secretory granules and that the overexpresssion of a GTPase-deficient mutant of this isoform inhibits nutrient-stimulated insulin secretion (24). Insulin-secreting cells also express Rab3B, -C, and -D (24). Conse- quently, fine tuning of insulin secretion may result from the interplay between different isoforms.

In this study, we compared the subcellular distribution and the functional role of the four Rab3 isoforms in exocytosis. Our results indicate that a fraction of each Rab3 isoform is associated with dense-core in-

sulin-containing secretory granules and that all of them, if kept in a GTP-bound form, inhibit exocytosis.

RESULTS

We have previously demonstrated that theb-cell line INS-1 expresses four Rab3 isoforms and that Rab3A is associated with dense-core insulin-containing secretory granules (24). In a first attempt to evaluate the localization of Rab3B, -C, and -D in insulin-secreting cells, we determined the subcellular distribution of each Rab3 isoform using a continuous sucrose density gradient (0.45–2M) (24, 25). INS-1 cells contain a large number of secretory granules (26) and are, therefore, well suited for this type of biochemical characterization. As demonstrated in thetop panelof Fig. 1, when homogenates of INS-1 cells are centrifuged at equilibrium, dense-core insulin-containing secretory granules are recovered in the fractions corresponding to 1.3–1.8Msucrose, while synaptophysin, a marker of g-aminoisobutyric acid-containing synaptic-like mi-

Fig. 1. Separation of INS-1 Organelles on a Sucrose Density Gradient

A postnuclear supernatant of INS-1 cells disrupted by nitrogen cavitation was loaded on a continuous sucrose density gradient (0.45–2Msucrose) and centrifuged for 18 h at 110,0003g. The distribution of the position of the organelles was assessed by analyzing aliquots of each fraction of the gradient by Western blotting followed by densitometric scanning of the films (synaptophysin, Na¹/K¹, ATPase, BHKp23, calreticulin) or by RIA (insulin). Sucrose concentration was calculated from the refractive index of the fractions. Thetop panel shows the distribution of synaptic-like microvesicles (synaptophysin) and dense-core insulin-containing granules.

Thelower panelillustrates the distribution of plasma membranes (Na¹/K¹ATPase), Golgi complex (BHKp23), and endoplasmic reticulum (calreticulin).

(4)

crovesicles, is recovered at 0.8–1Msucrose. The distribution of plasma membranes, endoplasmic reticulum, and the Golgi complex is shown in thelower panel of Fig. 1. The plasma membrane marker Na¹/K¹- ATPase was recovered in the fractions containing 0.9–

1.2Msucrose; BHKp23, a protein associated with the cis-side of the Golgi complex (27), was concentrated in the fractions containing 1.2–1.3Msucrose; calreticulin, a resident protein of the endoplasmic reticulum (28) was found in the fractions containing 1.3–1.4 M

sucrose.

The distribution of the four Rab3 isoforms in the sucrose gradient is given in Fig. 2. A fraction of Rab3A, -B, and -C was recovered between 1.3 and 1.8 M

sucrose consistent with an association with insulin- containing granules. These GTPases were also detected in lighter fractions (0.5–0.6 M sucrose), most likely reflecting the presence of a soluble pool (29).

These Rab3 isoforms were barely detectable in fractions enriched in synaptic-like microvesicles and plasma membrane (Fig. 2). The distribution of Rab3D was, in part, different from that of the other Rab3 isoforms. As was the case for Rab3A, -B, and -C, Rab3D was mainly associated with organelles display- ing the same density as insulin-containing secretory granules (Fig. 2). However, in contrast to the other isoforms, a significant portion of Rab3D was recovered in fractions containing 0.8–1.1Msucrose (Fig. 2).

In addition, Rab3D was almost absent from light fractions, suggesting that the soluble pool of this GTPase is very small.

The subcellular localization of Rab3A, -B, -C, and -D was also investigated after transient transfection of the hamsterb-cell line HIT-T15. To follow the expression of the transfected proteins, a myc epitope was inserted at the amino terminus of each Rab3 isoform.

Immunostaining followed by conventional fluorescence microscopy revealed that 70–75% of the transfected HIT-T15 cells contained detectable levels of insulin in the four series of experiments (Table 1). A minority of these cells expressed levels of Rab3 isoform that were clearly detectable at the immunofluorescence level (Table 1). In virtually all cells transfected with myc-tagged Rab3A, Rab3B, and Rab3C, the GTPases were found in the cellular compartment containing immunoreactive insulin (Table 1). In most of the cells transfected with myc-tagged Rab3D, the GTPase colocalized with insulin. However, Rab3D was also observed within cytoplasmic compartments of insulin- positive cells that did not contain detectable levels of the hormone or in cells in which insulin was not detectable (Table 1). Similar results were obtained after transfection of INS-1 cells with Rab3A, -B, -C, and -D (not shown). These results confirm the biochemical data obtained with the sucrose gradient and suggest that Rab3D has a subcellular distribution different, at least in part, from the other isoforms of Rab3.

The subcellular localization of Rab3 isoforms in HIT- T15 cells was analyzed by high-resolution confocal microscopy. Using this technique the myc-tagged

Rab3 isoforms (Fig. 3, A–D) as well as insulin (Fig. 3, E–H) were localized on punctate structures in the cytoplasmic compartment. Mathematical comparison of the images obtained in the fluorescein and the rhodamine channels demonstrates that myc-tagged Rab3 proteins colocalize with insulin-containing secretory granules (Fig. 3, I–L). As a negative control, HIT-T15 cells were also transfected with myc-tagged Rab5, another member of the Rab family that is normally associated with early endosomes (30). As expected, myc-tagged Rab5 did not colocalize with insulin (data not shown). In contrast to the other isoforms and in agreement with the results in Table 1, in some cells myc-tagged Rab3D was also detected in cellular compartments that did not contain detectable levels of insulin (Fig. 4).

We have previously determined the role of Rab3A in insulin exocytosis by transiently coexpressing this GTPase and human proinsulin in the hamster b-cell line HIT-T15 (24). This line displays a higher transfection efficacy than INS-1 cells, which makes it more suitable for transient transfection experiments. The release of human C-peptide by the transfected hamster cells was used as a reporter for exocytosis. Using this system we found that the overexpression of a GTPase-deficient mutant of Rab3A, in which Glu⁸¹is replaced by Leu (Q81L), leads to the inhibition of nutrient-induced secretion (24). To assess the involvement of Rab3B, -C, and -D in exocytosis, HIT-T15 cells were cotransfected with human proinsulin and with the wild-type or with the Q81L mutants of the four Rab3 isoforms. As shown in theupper panel of Fig. 5, the isoforms were expressed at comparable levels. None of the constructs affected the content or the basal release of human C peptide (not shown). The overexpression of wild-type Rab3A, -B, and -C had no sig- Fig. 2. Subcellular Distribution of Rab3 Isoforms

Subcellular fractions of INS-1 cells were obtained using the sucrose density gradient described in Fig. 1. Aliquots of fractions 3–16 were analyzed by Western blotting using antibodies directed against the four Rab3 isoforms. The figure shows the distribution of Rab3 isoforms after densitometric scanning of the autoradiographic films. Sucrose concentration was calculated from the refractive index of the fractions.

(5)

nificant effect on stimulated human C peptide release, while exocytosis triggered by nutrients and bombesin was slightly decreased by the overexpression of Rab3D (Fig. 5, lower panel). In contrast to the wild type, the mutants of the four Rab3 isoforms deficient in GTPase activity (Q81L) inhibited exocytosis elicited by secretagogues (Fig. 5,lower panel). Despite the fact that all Rab3 Q81L constructs were expressed at a similar level, the Rab3 isoforms inhibited exocytosis with different efficacy. On average the most potent isoform was Rab3D followed in order by Rab3B, and -A. C Peptide secretion in cells transfected with Rab3C Q81L was also significantly different from the secretion measured in control cells (P,0.05; n520;

pairedttest). However, the inhibitory effect of Rab3C Q81L was significantly smaller (P,0.05; n59; paired t test) than the effect of the other Rab3 isoforms. A similar variation in the degree of inhibition was observed if secretion was triggered by depolarizing po- tassium concentrations (not shown). In HIT-T15 cells, differences exist between the level of expression of endogenous Rab3 isoforms (24). Thus, we examined whether the efficacy with which the four isoforms inhibit exocytosis correlates with the ratio between endogenous and transfected proteins. As estimated by immunoblotting compared with the endogenous protein, Rab3A was overexpressed about 100-fold (Rab3A is barely detectable in HIT-T15 cells), while Rab3B, -C, and -D were overexpressed about 5- to 7-fold. Thus, the ratio between the amount of endogenous and transfected proteins cannot explain the differences between the four Rab3 isoforms.

We also tested for the effect on exocytosis of another mutant of Rab3 that is deficient in guanine nu- cleotide binding. This mutant, in which Asn¹³⁵is replaced by Ile (N135I), displays a very high dissociation rate for both GDP and GTP (31). Rab3A N135I and Rab3D N135I were well expressed in HIT-T15 cells (Fig. 6,upper panel) and inhibited exocytosis triggered by nutrients and bombesin with an efficacy similar to Rab3A Q81L and Rab3D Q81L, respectively (Fig. 6, lower panel). We also attempted to investigate the effect on exocytosis of Rab3B N135I and Rab3C N135I. However, the level of expression of these two

mutants achieved in transfected HIT-T15 cells was too low to permit the interpretation of the results (data not shown).

Permeabilized cells have been used as a model system to measure effects on exocytosis that are independent of membrane depolarization, channel activity, and soluble second messengers (32). We found that the release of C peptide, resulting from the incubation of Streptolysin-O-permeabilized HIT-T15 cells with 100m^MGTPgS or with 10m^MCa²¹, was strongly reduced by the overexpression of the Q81L and N135I mutants of Rab3A (Fig. 7). Thus, the inhibition caused by the overexpression of Rab3 Q81L and Rab3 N135I is likely to reflect an effect of the mutants on the exocytotic process rather than an effect on the signal- ing pathway that triggers hormone secretion.

Stable expression of wild-type Rab3B or of Rab3B N135I has been reported to enhance Ca²¹-evoked secretion in the neuroendocrine cell line PC12 (17).

Differences between neuroendocrine cells and HIT- T15 cells could potentially explain the discrepancy between our data and the results obtained with PC12 cells. To assess whether this was the case, we analyzed exocytosis in PC12 cells transiently cotransfected with the Q81L mutants of the four Rab3 isoforms and with human GH (hGH). Exocytosis from the fraction of cells expressing the Rab3 constructs was determined by measuring hGH release (33). As was the case in HIT-T15 cells, the GTPase-deficient mutants of the four Rab3 isoforms inhibited stimulus- induced secretion (Fig. 8).

The GTPase-deficient mutant of Rab3C is less efficient than the corresponding mutants of Rab3A, -B, and -D. Rab3 isoforms contain two short variable domains at the amino- and at the carboxy terminus of the protein. To address whether either one of these domains is responsible for the differences observed between isoforms, we generated chimeric constructs in which the amino- or the carboxy terminus of Rab3A Q81L was replaced by the corresponding region of Rab3C (Fig. 9). After transfection, all the chimeric proteins were expressed at comparable levels (Fig. 10, upper panel). The Rab3A chimera containing the car- boxy terminus of Rab3C (Rab3A/C) inhibited exocyto- Table 1. Evaluation of Insulin and Rab3 Immunoreactivity in Clones of Transfected HIT Cells

Rab3 isoform No. of Cells (No. of experiments)

Insulin-Positive Cells (% of total)

Myc-Positive Cells (% of insulin positive cells)

Colocalization of Insulin and myc Immunoreactivity

(% of transfected cells)

Rab3A 1400 (2) 1050 (75.0) 52 (5.0) 50 (96.1)

Rab3B 2570 (2) 1829 (71.2) 207 (11.3) 206 (99.5)

Rab3C 1831 (2) 1283 (70.1) 49 (3.8) 47 (95.9)

Rab3D 2450 (2) 1730 (70.6) 119 (6.9) 67 (56.3)

HIT-T15 cells transfected with myc-tagged wild-type Rab3A, -B, -C, and -D were plated on glass coverslips and immunolabeled with a polyclonal antibody against insulin and with a monoclonal antibody against human c-myc. The number of cells in the field was estimated by counting the nuclei. The number of insulin-containing cells was established by scoring those cells that showed a granular, rhodamine labeling of cytoplasm clearly above background level. The number of Rab3-transfected cells was established by scoring those cells that showed a fluorescein labeling of cytoplasm clearly above background level.

(6)

Fig. 3. Immunolabeling of Transfected Clones for Insulin and Rab3 Isoforms

HIT-T15 cells were transiently transfected with myc-tagged Rab3A (panels A, E, and I), Rab3B (panels B, F, and J), Rab3C (panels C, G, and K), and Rab3D (panels D, H, and L). The cells were analyzed by confocal microscopy after double immunofluorescence with a monoclonal antibody against the myc epitope (revealed using fluorescein-conjugated antibodies) (panels A–D) and with a polyclonal antibody against insulin (detected using rhodamine-conjugated antibodies) (panels E–H). Panels I–L show the regions of colocalization between transfected Rab3 proteins and insulin granules. The colocalization images were obtained as described inMaterials and Methodsby plotting all pixels whose intensities in both channels were simultaneously above the 40% level (representing,10% of the total pixel count).

(7)

sis to the same extent as Rab3A Q81L (Fig. 10,lower panel). Similar results were obtained with two Rab3A chimeric constructs containing, respectively, the carboxy terminus of Rab3B and Rab3D (data not shown).

In contrast, the Rab3A chimera with the amino terminus of Rab3C (Rab3C/A) was much less efficient, de- creasing stimulated secretion like Rab3C. As a control, we also tested the corresponding Rab3A chimeric constructs containing the amino terminus of Rab3B and Rab3D. The inhibition of exocytosis obtained with these constructs was not significantly different from that of Rab3A, -B, or -D (data not shown). The localization of all the chimeric constructs analyzed by immunofluorescence was identical to that of Rab3A (data not shown).

DISCUSSION

The members of the Rab3 family are thought to control the targeting and/or the fusion of Golgi-derived vesicles with the plasma membrane (9, 13). However, the precise subcellular distribution and the respective role of the different Rab3 isoforms in exocytosis are still unclear. In bovine chromaffin cells, antibodies against Rab3A and Rab3C were found to stain vesicle structures, whereas an antibody against Rab3B stained the plasma membrane (16). In polarized epithelial cells, Rab3B was localized to the apical pole of the cells near the tight junctions (34). However, the same investiga- tors observed that transfected epitope-tagged Rab3B colocalized with dense-core granules of PC12 cells (17). We have previously demonstrated that Rab3A is associated with insulin-containing secretory granules of pancreaticb-cells (24). In this study, using double immunofluorescence experiments analyzed by confocal microscopy, we demonstrate that myc-tagged Rab3B, -C, and -D are also associated with insulin- containing granules and that, clearly, none of the isoforms are localized on the plasma membrane. It is very unlikely that our morphological observations are bi- ased by the overexpression of the proteins. In fact, first, these findings are corroborated by subcellular fractionation studies indicating that the endogenous Rab3 isoforms display a similar distribution. Second, Fig. 4. Myc-Tagged Rab3D Is Not Exclusively Associated

with Secretory Granules

HIT-T15 cells transiently transfected with myc-tagged Rab3D were analyzed by confocal microscopy after double immunofluorescence with a monoclonal antibody against the myc epitope (panel A) and with a polyclonal antibody against insulin (panel B). Thearrowsindicate examples of subcellular regions containing high levels of myc-tagged Rab3D and undetectable levels of insulin.

Fig. 5. Effect of Wild-Type and GTPase-Deficient Mutants of Rab3 Isoforms on Insulin Secretion

HIT-T15 cells were transiently cotransfected with human proinsulin and with wild-type or Q81L mutants of the four Rab3 isoforms. Two days after transfection, the cells were incubated in the presence or in the absence of 10 mMglu- cose, 5 mMleucine, 5 mMglutamine, and 100 nMbombesin for 30 min. The cells were then collected and analyzed by Western blotting using an antibody against the myc epitope.

Thetop panelshows the results of a representative experiment. The amount of human C-peptide secreted by the cells in response to the stimuli was measured by ELISA. Thelower panelshows the mean6^SEof three to five representative experiments. In control cells the mixture of nutrients and bombesin elicited a 4- to 8-fold increase in C peptide release.

(8)

overexpression of about 100-fold of Rab3A (that is expressed at low level in HIT-T15 cells) does not affect the targeting of the protein to the secretory granules.

In addition, we found the same localization in cells expressing low levels of the exogenous protein and in cells producing high levels of the transfected Rab3 isoforms. Third, when Rab5, another member of the Rab family known to be associated with early endosomes, was overexpressed in HIT-T15 cells, the protein was not associated with secretory granules.

In contrast to the other isoforms, Rab3D was also found in cells containing undetectable levels of insulin or within cytoplasmic compartments of insulin-positive cells not containing detectable levels of the hormone. It has been previously reported that Rab3A and Rab3D are associated with different vesicular compartments (14, 15, 35). The present data confirm a partial difference in the localization between the two isoforms but indicate that a considerable fraction of Rab3D is associated with insulin-containing granules.

At present, the precise nature of the vesicular structures containing Rab3D but devoid of insulin is un- known. In CHO cells, Rab3D cofractionates with a population of post-Golgi storage vesicles slightly larger than synaptic vesicles that have been identified with labeled glycosaminoglycan chains (36). When combined with our results, it is tempting to speculate that the fraction of Rab3D not associated with insulin- containing secretory granules may be located on a similar type of storage vesicles.

The presence of more than one isoform within the same cell raises the question whether Rab3 proteins are redundant or whether they play a specific role in the exocytotic process. Transient overexpression of a constitutively active mutant of Rab3A reduces stimulus-induced secretion in neuroendocrine (33, 37) and in insulin-secreting cells (24). These observations have been taken as an indication that Rab3A is a negative Fig. 6. Effect of the Rab3 Mutants Deficient in Guanine Nu-

cleotide Binding on Insulin Secretion

HIT-T15 cells were transiently cotransfected with human proinsulin and with the N135I mutants of Rab3A and Rab3D.

Two days after transfection, the cells were incubated as described in Fig. 5. At the end of the incubation the cells were collected and analyzed by Western blotting using an antibody against the myc epitope (top panel). The amount of human C peptide secreted by the cells in response to the stimuli was measured by ELISA. Thelower panelshows the mean6^SEof three representative experiments. In control cells the mixture of nutrients and bombesin elicited a 4- to 6-fold increase in C-peptide release.

Fig. 7. Effect of Rab3A Q81L and Rab3A N135I on Exocy- tosis in Streptolysin-O Permeabilized HIT-T15 Cells

HIT-T15 cells were transiently cotransfected with human preproinsulin and with the vector alone (C), with Rab3A Q81L (Q81L), or with Rab3A N135I (N135I). Two days after transfection the cells were permeabilized with Streptolysin-O. The medium was then removed and the cells were incubated for 7 min in the presence of 0.1m^Mfree Ca²¹, 0.1m^MCa²¹plus 100 m^MGTPgS, or 10 m^MCa²¹. The amount of human C peptide secreted by the cells was measured by ELISA. The results represent the mean6^SEof three independent experiments.

Fig. 8. Effect of the GTPase-Deficient Mutants of Rab3 Iso- forms on Exocytosis in PC12 Cells

PC12 cells were transfected by electroporation with hGH and with the Q81L mutants of Rab3A (A), Rab3B (B), Rab3C (C), or Rab3D (D). Four days after transfection the cells were stimulated with depolarizing concentrations of K¹for 2 min.

The figure shows the amount of hGH released in response to the stimulus. The results correspond to the mean 6^SE of three independent experiments. K¹depolarization of control cells caused an increase of 4- to 6-fold in hGH release.

(9)

modulator of exocytosis (33, 37). In agreement with this model, Rab3A-deficient mice display an increased number of synaptic vesicle fusion events shortly after the arrival of the nerve impulse (38). An alternative interpretation of the results would be that the hydrolysis of GTP is a prerequisite for the fusion of secretory vesicles with the plasma membrane. In this case, the GTPase-deficient mutant of Rab3A associated with secretory vesicles would have a dominant-negative effect and would inhibit exocytosis. The latter view is favored by the results obtained in yeast with the corresponding GTPase-deficient mutant of Sec4 (39).

Rab3B has been suggested to have functional properties distinct from Rab3A. Thus, in pituitary cells, inhibition of Rab3B expression was found to attenuate Ca²¹-dependent exocytosis (40) and in PC12 cells stable expression of wild-type Rab3B or of Rab3B N135I has been reported to potentiate the efficiency of Ca²¹-evoked secretion and to markedly increase the accumulation of norepinephrine in secretory granules (17). Here we show that transient overexpression of wild-type Rab3B in HIT-T15 cells has no significant effect on stimulated secretion, and that Rab3B Q81L strongly inhibits exocytosis. Thus, the data are consistent with a similar role for Rab3A and Rab3B in insulin-secreting cells. Our findings cannot be attrib- uted to differences between insulin-secreting cells and neuroendocrine cells, since we demonstrate that transient expression of Rab3B Q81L diminishes stimulated secretion also in PC12 cells. Unfortunately, in HIT-T15 cells, Rab3B N135I was poorly expressed and it was, therefore, impossible to evaluate its effect on insulin release. A possible explanation for the low level of Rab3B N135I and Rab3C N135I detected in transfected cells is that the high dissociation rate for guanine nucleotides is affecting the turnover rate of the protein.

We have analyzed the effect of Rab3B mutants on exocytosis after short-term expression (2 days). In contrast, Weber et al. (17) analyzed the impact of Rab3B expression on secretion using stable cell lines.

After long-term expression of Rab3B, the stability of Rabphilin, a putative Rab3 effector, was found to decrease (17). Thus, it is possible that long-term expression of Rab3B leads to secondary alterations in the secretory phenotype that are not evident in transient transfection experiments.

The overexpression of a GTPase-deficient mutant of each of the four Rab3 isoforms tested in this study inhibited exocytosis. However, the decrease in nutrient-induced secretion observed in the presence of Rab3C was much less pronounced than the decrease caused by the other isoforms. Exchange of the carboxy-terminal domain of Rab3A with those of any of the other Rab3 isoforms did not alter the extent of inhibition measured after overexpression of the chimeric constructs. Substitution of the amino terminus of Rab3A with the corresponding domain of Rab3B or Rab3D was also without effect, but replacement of the amino terminus of Rab3A with the residues of Rab3C reduced the effect of the GTPase on exocytosis. The amino termini of Rab5 and of Rab2 have been shown to be required for the function of these GTPases and for the interaction with components of the transport machinery (41, 42). The variable domain at the carboxy terminus of Rab proteins appears to dictate the localization of the GTPases to specific cellular compart- Fig. 9. Schematic Representation of the Chimeric Con-

structs between Rab3A and Rab3C

The Rab3A/C chimera was generated by inserting the carboxy-terminal fragment of Rab3C (downstream to amino acid 179) in the corresponding region of Rab3A. To produce the amino-terminal Rab3C/A chimera, the first 19 amino acids of Rab3A were replaced by the corresponding residues of

Rab3C. Fig. 10. Effect on Exocytosis of Protein Chimeras between

Rab3A and Rab3C

HIT-T15 cells were transiently cotransfected with human proinsulin and with Rab3A Q81L (A), Rab3C Q81L (C), a Rab3A Q81L construct with the carboxy-terminus of Rab3C (A/C), and a Rab3A Q81L construct with the amino terminus of Rab3C (C/A). Two days after transfection the cells were incubated in the presence or in the absence of 10 mMglu- cose, 5 mMleucine, 5 mMglutamine, and 100 nMbombesin for 30 min. The cells were then collected, and the expression of the exogenous proteins was analyzed by Western blotting using an antibody against the myc epitope. Thetop panel shows the results of a representative experiment. The amount of human C peptide secreted by the cells in response to the stimuli was measured by ELISA. The lower panel shows the mean6^SEof three independent experiments. In control cells the mixture of nutrients and bombesin elicited a 4- to 6-fold increase in C peptide release.

(10)

ments (43, 44). Our results suggest that the amino terminus of the Rab3 isoforms may be involved in the interaction with an, as yet unidentified, component of the secretory machinery. Differences in the amino- terminal domain of the Rab3 isoforms may affect the affinity for this putative component and could determine the efficacy with which they regulate insulin secretion.

In conclusion, we have demonstrated that the secretory granules of pancreatic b-cells contain four Rab3 isoforms with similar roles in the regulation of insulin secretion. Further experiments are needed to determine the step(s) and the precise mechanism by which the GTPases of the Rab3 family control the exocytotic process.

MATERIALS AND METHODS Materials

The cDNA coding for human wild-type Rab3B was kindly provided by Dr. K. Kirk (University of Alabama at Birmingham, AL). Bovine wild-type Rab3C cDNA and polyclonal antibodies directed against Rab3A, Rab3B, and Rab3C were generously provided by Dr. A. Zahraoui (Curie Institute, UMR 144 CNRS, Paris, France). Plasmid pBRhins, derived from pBR322 and containing the human preproinsulin gene under the control of the cytomegalovirus promoter), was kindly provided by Dr.

J.-C. Irminger (University of Geneva). The generation of rabbit polyclonal antibodies against the amino-terminal region of Rab3D was described previously (14). The mouse monoclonal antibody specific for human c-myc(9E10) was produced from myeloma SP2/0; it was affinity purified from the hybrid- oma medium using sheep antimouse IgG antibodies coupled to Protein A Sepharose CL-4B (Sigma P-3391, Sigma Chem- ical Co., St. Louis, MO). A polyclonal rabbit antibody raised against thea-subunit of Na¹/K¹-ATPase was kindly supplied by Dr. E. Ferraille (University of Geneva).

The polyclonal antibody against BHKp23 was obtained from Dr. J. Gruenberg (University of Geneva); the polyclonal against calreticulin was provided by Dr. K.-H. Krause (Uni- versity Hospital, Geneva). The monoclonal antibody against synaptophysin was purchased from Sigma.

Cell Culture

The insulin-secreting cell lines HIT-T15 and INS-1 were cultured in RPMI 1640 medium supplemented with 10% FCS and other additions as described for each cell line (26, 45).

PC12 cells were cultured in DMEM supplemented with 6%

FCS and 6% horse serum.

Subcellular Fractionation

Subcellular fractionation of approximately 10⁸ INS-1 cells was performed as described (24). Briefly, a postnuclear supernatant obtained after disruption of the cells by nitrogen cavitation was loaded on a continuous sucrose density gradient (8 ml; 0.45–2Msucrose). After centrifugation for 18 h at 110,0003g, 16 fractions of 0.5 ml were collected from the top of the tube. The concentration of sucrose in the fractions was determined by measuring the refractive index of the solution. The amount of insulin present in the fractions was measured by RIA. The distribution throughout the gradient of BHKp23, calreticulin, Na¹/K¹ ATPase, synaptophysin, and

Rab3A/B/C and D was assessed by Western blotting (24) followed by densitometric scanning of the autoradiographic films.

Generation of DNA Constructs

The generation of myc-tagged human wild-type Rab3A and of the mutant at positions 81 (Q81L) and 135 (N135I) has been described previously (37). Myc-tagged wild-type Rab3A and the corresponding Q81L and N135I mutants were subcloned in the mammalian expression vector pcDNA3 (Invitro- gen, San Diego, CA). Wild-type Rab3B, Rab3C, Rab3D, and the Q81L mutant of Rab3D were subcloned into pcDNA3 containing the N-terminal myc epitope. To construct the Rab3A/C chimera containing the carboxy terminus of Rab3C, we generated aXhoI restriction site in Rab3A and Rab3C at the level of amino acid 179; the fragment of Rab3C downstream to theXhoI site was then excised and inserted in the correspondingXhoI site of Rab3A Q81L. A similar approach was used to produce the amino-terminal Rab3C/A chimera.

In this case, aXhoI site was introduced in the sequence of Rab3A and Rab3C at the level of amino acid 19; the fragment downstream to theXhoI site of Rab3A Q81L was inserted in the corresponding site of Rab3C.

Site-Directed Mutagenesis

The mutants of Rab3B and Rab3C were generated by site- directed mutagenesis according to the method of Kunkel (46). The mutations generated were confirmed by DNA se- quencing of the plasmids.

Transfection

For transient transfection experiments, HIT-T15 cells were seeded in 24-multiwell plates (4310⁵cells per well). After 3 days of culture, the cells were cotransfected using the li- popolyamine Transfectam (Promega) with 2.5mg of the vector encoding human preproinsulin and with 5mg of the plasmids containing the cDNAs under study (24, 47). Transient transfection of PC12 cells was performed by electroporation;

4310⁶cells were resuspended and electroporated in 400ml of serum-free DMEM in the presence of 40mg of a plasmid vector encoding hGH (Nichols Institute, San Juan Capistrano, CA) and 40 mg of pcDNA3 encoding the Rab3 proteins.

Immediately after transfection the cells were diluted in culture medium and seeded in 24-multiwell plates.

Secretion from Transfected HIT-T15 Cells

Forty-eight hours after transfection the cells were preincubated for 30 min in modified Krebs-Ringer bicarbonate buffer (24). They were then incubated for 30 min in Krebs-Ringer bicarbonate buffer in the presence or in the absence of 10 mM

glucose, 5 mMleucine, 5 mMglutamine, and 100 nMbomb- esin, a mixture known to strongly stimulate insulin secretion in these cells. Secretion from transfected cells was assessed by measuring the amount of human C peptide released into the medium during the incubation period by enzyme-linked immunosorbent assay (ELISA) (Dako Corp., Carpenteria, CA).

Secretion experiments with permeabilized cells were performed as described (48). Forty-eight hours after transfection the cells were permeabilized with Streptolysin-O for 8 min.

The medium was then removed, and exocytosis was triggered by adding 100m^MGTPgS or by increasing the free Ca²¹concentration from 0.1m^M(basal) to 10m^M.

Secretion from Transfected PC12 Cells

Four days after transfection the cells were preincubated during 30 min in 20 mMHEPES, pH 7.4, 128 mMNaCl, 5 mMKCl,

(11)

2.7 mMCaCl₂, 10 mMglucose, and 1 mMMgCl₂. The medium was then aspirated and the cells were stimulated for 2 min with the same buffer but containing 53 mMNaCl and 80 mM

KCl. Exocytosis from the subpopulation of transfected cells was determined by measuring by ELISA (Boeringer Mann- heim, Indianapolis, IN) the amount of hGH secreted into the medium during the incubation period.

Immunofluorescence

For immunofluorescence labeling, the cells were grown on glass coverslips coated with extracellular matrix produced by A431 epidermoid cells (49, 50). Subconfluent monolayers were fixed 30 min in a 4% paraformaldehyde-0.1Mphos- phate buffer solution, pH 7.4. After rinsing for 10 min in PBS supplemented with 0.5% BSA (PBS) and 10 mMNH₄Cl, the cultures were simultaneously exposed for 2 h at room temperature to a guinea pig polyclonal serum against porcine insulin (diluted 1:200) (51) and to a purified mouse monoclonal antibody against human c-myc (clone 9E10; diluted 1:600). After rinsing in PBS, the cultures were incubated again for 1 h at room temperature in the presence of both a rhodamine-labeled goat serum against guinea pig Ig (Cappel, Organon Teknika AG, Switzerland; diluted 1:200) and a fluorescein-conjugated sheep antimouse Ig (Biosys, Compie`gne, France; diluted 1:200). After careful rinsing, the coverslips were mounted with 0.02% paraphenylenediamine in PBS- glycerol (1:2, vol/vol).

Two independent experiments were run for quantitative analysis. In each experiment and for each of the four Rab3 isoforms, 10 fields were randomly photographed at the fixed magnification of 1003. Counts were performed on these photographs, which were projected on a screen at the final magnification of 3703. The total number of cells was evalu- ated by scoring the number of nuclei present per field. The number of insulin-containing cells was established by scoring those cells that showed a granular, rhodamine labeling of cytoplasm clearly above background level. The number of Rab3-transfected cells was established by scoring those cells that showed a fluorescein labeling of cytoplasm clearly above background level. Colocalization of the two immuno- reactivities was assessed by evaluating the insulin labeling of all Rab3-positive cells.

Two separate experiments were made for confocal microscopy (Leica Lasertechnik, Heidelberg, Germany model TCS NT) analysis, to discriminate intracellular granular staining from diffuse cytoplasmic or membrane staining. Excita- tion was obtained with an Argon-Krypton laser, with line set at 488 nm for fluorescein excitation and 568 nm for rhodamine excitation, and the emitted light was filtered through appropriate filters (BF 530/30 for fluorescein, LP 590 for rhodamine). Images of 5123512 pixels (RabA, -B, and -C) and 1024 3 1024 pixels (RabD) were taken with a 633 objective, NA 1.32. A typical pixel size was 35 nm (RabD, with a zoom of 2.3). For each field, digitized series of optical sections at different planes of focus were collected on the host computer; care was taken to use the full dynamic range of the photomultipliers by using a special look up table (glowover-glowunder, Leica); the sections were processed using Imaris software (Bitplane AG, Zurich, Switzerland) on a Silicon Graphics computer. No filters were applied, but background noise was reduced. To calculate colocalization between transfected Rabs and insulin granules, a correlogram of all intensities was generated, normalizing the intensities from 0 to 100%; all pixels whose intensities in both channels were simultaneously above the 40% level (representing ,10% of the total pixel number) were then plotted.

Acknowledgments

We are grateful to Dr. A. Zahraoui (Paris) for providing Rab3C cDNA and antibodies against Rab3A, Rab3B, and Rab3C, to

Dr. K. Kirk (Birmingham, AL) for providing the Rab3B cDNA, to Dr. Gruenberg (Geneva) for supplying the BHKp23 antibodies, and to Dr. Krause (Geneva) for supplying the antibody against calreticulin. We are also indebted to Miss D. Duhamel and Miss S. Lu¨thi for technical assistance.

Received June 1, 1998. Re-revision received October 9, 1998. Accepted October 13, 1998.

Address requests for reprints to: Dr. Romano Regazzi, Institut de Biologie Cellulaire et de Morphologie, Rue du Bugnon 9, 1005 Lausanne, Switzerland. E-mail: Romano.

Regazzi@ibcm.unil.ch

Supported by grants from the Juvenile Diabetes Founda- tion International (197124) PM, (196100) RR, the Swiss Na- tional Science Foundation (32–34086.95) PM, (32–32376.91 and 32–49755.96) CBW, (3100–050640.97) RR and the Eu- ropean Union (BMH4-CT96–1427).

REFERENCES

1. Wollheim CB, Lang J, Regazzi R 1996 The exocytotic process of insulin secretion and its regulation by Ca²¹ and G-proteins. Diabetes Rev 4:276–297

2. Ferro-Novick S, Jahn R 1994 Vesicle fusion from yeast to man. Nature 370:191–193

3. Rothman JE 1994 Mechanisms of intracellular protein transport. Nature 372:55–63

4. Jacobsson G, Bean AJ, Scheller RH, Juntti-Berggren L, Deeney JT, Berggren P-O, Meister B 1994 Identification of synaptic proteins and their isoform mRNAs in compartments of pancreatic endocrine cells. Proc Natl Acad Sci USA 91:12487–12491

5. Regazzi R, Wollheim CB, Lang J, Theler JM, Rossetto O, Montecucco C, Sadoul K, Weller U, Palmer M, Thorens B 1995 VAMP-2 and cellubrevin are expressed in pancreatic b-cells and are essential for Ca²¹- but not for GTPgS-induced insulin secretion. EMBO J 14:2723–

2730

6. Sadoul K, Lang J, Montecucco C, Weller U, Regazzi R, Catsicas S, Wollheim CB, Halban PA 1995 SNAP-25 is expressed in islets of Langerhans and is involved in insulin release. J Cell Biol 128:1019–1028

7. Wheeler MB, Sheu L, Ghai M, Bouquillon A, Grondin G, Weller U, Beaudoin AR, Bennett MK, Trimble WS, Gai- sano HY 1996 Characterization of SNARE protein expression inbcell lines and pancreatic islets. Endocrinol- ogy 137:1340–1348

8. Martin F, Moya F, Gutierrez LM, Reig JA, Soria B 1995 Role of syntaxin in mouse pancreatic beta cells. Diabe- tologia 38:860–863

9. Novick P, Zerial M 1997 The diversity of Rab proteins in vesicle transport. Curr Opin Cell Biol 9:496–504 10. Pevsner J, Hsu S-C, Braun JEA, Calakos N, Ting AE,

Bennett MK, Scheller RH 1994 Specificity and regulation of a synaptic vesicle docking complex. Neuron 13:353–

361

11. Schulze KL, Littleton JT, Salzberg A, Halachmi N, Stern M, Lev Z, Bellen HJ 1994 rop, aDrosophilahomolog of yeast Sec1and vertebrate n-Sec1/Munc-18 proteins, is a negative regulator of neurotransmitter release in vivo.

Neuron 13:1099–1108

12. Lupashin VV, Waters MG 1997 t-SNARE activation through transient interaction with a Rab-like guanosine triphosphatase. Science 276:1255–1258

13. Fischer von Mollard G, Stahl B, Li C, Su¨dhof TC, Jahn R 1994 Rab proteins in regulated exocytosis. Trends Bio- chem Sci 19:164–168

14. Baldini G, Scherer PE, Lodish HF 1995 Nonneuronal

(12)

expression of Rab3A: induction during adipogenesis and association with different intracellular membranes than Rab3D. Proc Natl Acad Sci USA 92:4284–4288 15. Martelli AM, Bareggi R, Baldini G, Scherer PE, Lodish HF,

Baldini G 1995 Diffuse vesicular distribution of Rab3D in the polarized neuroendocrine cell line AtT20. FEBS Lett 368:271–275

16. Lin C-G, Lin Y-C, Liu H-W, Kao LS 1997 Characterization of Rab3A, Rab3B and Rab3C: different biochemical properties and intracellular localization in bovine chromaffin cells. Biochem J 324:85–90

17. Weber E, Jilling T, Kirk KL 1996 Distinct functional properties of Rab3A and Rab3B in PC12 cells neuroendocrine cells. J Biol Chem 271:6963–6971

18. Geppert M, Bolshakov VY, Siegelbaum SA, Takei K, De Camilli P, Hammer RE, Su¨dhof TC 1994 The role of Rab3A in neurotransmitter release. Nature 369:493–497 19. Li C, Takei K, Geppert M, Daniell L, Stenius K, Chapman ER, Jahn R, De Camilli P, Su¨dhof TC 1994 Synaptic targeting of rabphilin-3A, a synaptic vesicle Ca²¹/phos- pholipid-binding protein, depends on rab3A/3C. Neuron 13:885–898

20. Wang Y, Okamoto M, Schmitz F, Hofmann K, Su¨dhof TC 1997 Rim is a putative Rab3 effector in regulating synaptic-vesicle fusion. Nature 388:593–598

21. Li G, Regazzi R, Balch WE, Wollheim CB 1993 Stimula- tion of insulin release from permeabilized HIT-T15 cells by a synthetic peptide corresponding to the effector domain of the small GTP-binding protein rab3. FEBS Lett 327:145–149

22. Olszewski S, Deeney JT, Schuppin GT, Williams KP, Corkey BE, Rhodes CJ 1994 Rab3A effector domain peptides induce insulin exocytosis via a specific interaction with a cytosolic protein doublet. J Biol Chem 269:27987–27991

23. Regazzi R, Sasaki T, Takahashi K, Jonas JC, Volker C, Stock JB, Takai Y, Wollheim CB 1995 Prenylcysteine analogs mimicking the C-terminus of GTP-binding proteins stimulate exocytosis from permeabilized HIT-T15 cells. Biochim Biophys Acta 1268:268–278

24. Regazzi R, Ravazzola M, Iezzi M, Lang J, Zahraoui A, Andereggen E, Morel P, Takai Y, Wollheim CB 1996 Expression, localization and functional role of small GTPases of the Rab3 family in insulin-secreting cells.

J Cell Sci 109:2265–2273

25. Reetz A, Solimena M, Matteoli M, Folli F, Takei K, De Camilli P 1991 GABA and pancreaticb-cells: colocalization of glutamic acid decarboxylase (GAD) and GABA with synaptic-like microvesicles suggests their role in GABA storage and secretion. EMBO J 10:1275–1284 26. Asfari M, Janjic D, Meda P, Li G, Halban PA, Wollheim CB

1992 Establishment of 2-mercaptoethanol-dependent differentiated insulin-secreting cell lines. Endocrinology 130:167–178

27. Rojo M, Pepperkok R, Emery G, Kellner R, Stang E, Parton RG, Gruenberg J 1997 Involvement of the trans- membrane protein p23 in biosynthetic protein transport.

J Cell Biol 139:1119–1135

28. Arber S, Krause KH, Caroni P 1992 S-Cyclophilin is re- tained intracellularly via a unique COOH-terminal sequence and colocalizes with the calcium storage protein calreticulin. J Cell Biol 116:113–125

29. Regazzi R, Kikuchi A, Takai Y, Wollheim CB 1992 The small GTP-binding proteins in the cytosol of insulin- secreting cells are complexed to GDP dissociation inhib- itor proteins. J Biol Chem 267:17512–17519

30. Bucci C, Parton RG, Mather IH, Stunnenberger H, Simons K, Hoflack B, Zerial M 1992 The small GTPase rab5 functions as a regulatory factor in the early endo- cytic pathway. Cell 70:715–728

31. Brondyk MK, McKiernan CJ, Burstein ES, Macara IG 1993 Mutants of Rab3A analogous to oncogenic Ras mutants. J Biol Chem 268:9410–9515

32. Ullrich S, Wollheim CB 1988 GTP-dependent inhibition of insulin secretion by epinephrine in permeabilized RINm5F cells. J Biol Chem 263:8615–8620

33. Holz R, Brondyk WH, Senter RA, Kuzion L, Macara IG 1994 Evidence for the involvement of Rab3A in Ca²¹- dependent exocytosis from adrenal chromaffin cells.

J Biol Chem 269:10229–10234

34. Weber E, Berta G, Tousson A, St. John P, Weaver M, Gopalokrishnan U, Jilling T, Sorscher EJ, Elton TS, Abra- hamson DR, Kirk K 1994 Expression and polarized targeting of a Rab3 isoform in epithelial cells. J Cell Biol 125:583–594

35. Baldini G, Baldini G, Wang G, Weber M, Zweyer M, Bareggi R, Witkin JW, Martelli AM 1998 Expression of Rab3D N135I inhibits regulated secretion of ACTH in AtT-20 cells. J Cell Biol 140:305–313

36. Chavez RA, Miller SG, H-P Moore 1996 A biosynthetic regulated secretory pathway in constitutive cells. J Cell Biol 133:1177–1191

37. Johannes L, Lledo P-M, Roa M, Vincent J-D, Henry J-P, Darchen F 1994 The GTPase Rab3a negatively controls calcium-dependent exocytosis in neuroendocrine cells.

EMBO J 13:2029–2037

38. Geppert M, Goda Y, Stevens CF, Su¨dhof TC 1997 The small GTP-binding protein Rab3A regulates a late step in synaptic vesicle fusion. Nature 387:810–814

39. Walworth NC, Brennwald P, Kabnecell AK, Garret M, Novick P 1992 Hydrolysis of GTP by Sec4 protein plays an important role in vesicular transport and is stimulated by a GTPase-activating protein inSaccharomyces cer- evisiae. Mol Cell Biol 12:17–28

40. Lledo PM, Vernier P, Vincent J-D, Mason WT, Zorec R 1993 Inhibition of Rab3B expression attenuates Ca²¹- dependent exocytosis in rat anterior pituitary cells. Na- ture 364:540–544

41. Steele-Mortimer O, Clague MJ, Huber LA, Chavrier P, Gruenberg J, Gorvel JP 1994 The N-terminal domain of a rab protein is involved in membrane-membrane recog- nition and/or fusion. EMBO J 13:34–41

42. Tisdale EJ, Balch WE 1996 Rab2 is essential for the maturation of pre-Golgi intermediates. J Biol Chem 271:29372–29379

43. Chavrier P, Gorvel JP, Stelzer E, Simons K, Gruenberg J, Zerial M 1991 Hypervariable C-terminal domain of rab proteins acts as a targeting signal. Nature 353:769–772 44. Stenmark H, Valencia A, Martinez O, Ullrich O, Goud B, Zerial M 1994 Distinct structural elements of rab5 define its functional specificity. EMBO J 13:575–583

45. Regazzi R, Li G, Deshusses J, Wollheim CB 1990 Stim- ulus-response coupling in insulin-secreting HIT cells.

J Biol Chem 265:15003–15009.

46. Kunkel TA 1985 Rapid and efficient site-specific mutagenesis without phenotypic selection. Proc Natl Acad Sci USA 82:488–492

47. Lang J, Nishimoto I, Okamoto T, Regazzi R, Kiraly C, Weller U, Wollheim CB 1995 Direct control of exocytosis by receptor-mediated activation of the heterotrimeric GTPases Gi and Go or by the expression of their active Gasubunits. EMBO J 14:3635–3644

48. Regazzi R, Sadoul K, Meda P, Kelly RB, Halban PA, Wollheim CB 1996 Mutational analysis of VAMP domains implicated in Ca²¹-induced insulin exocytosis. EMBO J 15:6951–6959

49. Wayner EA, Gil SG, Murphy GF, Wilke MS, Carter WG 1993 Epiligrin, a component of epithelial basement membranes, is an adhesive ligand for alpha 3 beta 1 positive T lymphocytes. J Cell Biol 121:1141–1152

50. Weitzman JB, Pasqualini R, Takada Y, Hemler ME 1993 The function and distinctive regulation of the integrin VLA-3 in cell adhesion, spreading, and homotypic cell aggregation. J Biol Chem 268:8651–8657

51. Wright PH, Makalu DR, Posey IJ 1968 Guinea-pig anti- insulin serum. Diabetes 17:513–516