Stimulation of insulin secretion by glucose and other secretagogues from pancreatic islet β-cells is mediated by multiple signaling pathways. Rac1 is a member of Rho family GTPases regulating cytoskeletal organization, and recent evidence also implicates Rac1 in exocytotic processes. Herein, we report that exposure of insulin-secreting (INS) cells to stimulatory glucose concentrations caused translocation of Rac1 from cytosol to the membrane fraction (including the plasmalemma), an indication of Rac1 activation. Furthermore, glucose stimulation increased Rac1 GTPase activity. Time course study indicates that such an effect is demonstrable only after 15 min stimulation with glucose. Expression of a dominant-negative Rac1 mutant (N17Rac1) abolished glucose-induced translocation of Rac1 and significantly inhibited insulin secretion stimulated by glucose and forskolin. This inhibitory effect on glucose-stimulated insulin secretion was more apparent in the late phase of secretion. However, N17Rac1 expression did not significantly affect insulin secretion induced by high K+. INS-1 cells expressing N17Rac1 also displayed significant morphological changes and disappearance of F-actin structures. Expression of wild-type Rac1 or a constitutively active Rac1 mutant (V12Rac1) did not significantly affect either the stimulated insulin secretion or basal release, suggesting that Rac1 activation is essential, but not sufficient, for evoking secretory process. These data suggest, for the first time, that Rac1 may be involved in glucose- and forskolin-stimulated insulin secretion, possibly at the level of recruitment of secretory granules through actin cytoskeletal network reorganization.
- G protein
- F-actin cytoskeleton
insulin secretion from pancreatic β-cells is highly regulated by complex signaling pathways upon stimulation by glucose and other fuels, and by multiple hormones and neurotransmitters (21, 33, 45). It is known that secretagogue-induced insulin secretion is mediated by a series of metabolic coupling factors (such as the ATP-to-ADP ratio), second messengers (Ca2+, cAMP), and regulators (e.g., protein kinases, inositol phospholipids, and GTP-binding proteins; see Refs. 10, 21, 33, 37, 53, 63). The exocytotic process is regulated at several stages, including transport of granules toward the plasma membrane, granule docking, priming, and finally fusing with the plasma membrane (58). It is well established that an increase of cytosolic free Ca2+ concentrations ([Ca2+]i) is one of the signals for exocytosis. However, this process also involves other proteins such as soluble N-ethylmaleimide-sensitive factor attachment protein receptors (SNAREs) and may be regulated by GTP and/or GTP-binding proteins (13, 25, 27, 33, 61). In many secretory systems, including insulin-secreting cells, GTPγS is able to induce secretion independent of Ca2+, suggesting that GTP/G proteins may exist as a parallel pathway for the regulated exocytosis (33, 47). Indeed, it has been reported that the GTP-binding protein rab3 stimulates release of neurotransmitters and hormones from permeabilized cells, including islet β-cells (15, 37).
The Rho family of small GTP-binding proteins is a subfamily of the Ras-related superfamily and contains >10 members in mammals, including Rho, Rac, and CDC42, which are the most well studied (6, 19, 50). Rho proteins are primarily involved in the regulation of actin cytoskeleton organization and are important for cell adhesion, migration, phagocytosis, cytokinesis, and other morphological changes, such as formation of ruffles and lamellipodial and filopodial extensions (5, 17, 19, 23, 59). In addition, there is evidence for the implication of Rho proteins and actin filaments in regulated exocytosis, although their actions seem quite complex (4, 19). The actin network beneath the plasma membrane is transiently depolymerized during exocytosis in mast cells and chromaffin cells (29, 54). Thus it appears that the cortical actin cytoskeletal web acts as a barrier to hamper the access of secretory granules to the plasma membrane, and a reorganization of F-actin in this region occurs during late steps of exocytosis. In pancreatic β cells, disrupting F-actin by cytochalasin B facilitated insulin release (28), in particular, the first phase of insulin secretion (39). On the other hand, maintenance of F-actin structure is also required for the secretory process. For instance, an increase of F-actin was observed when pancreatic β-cells were stimulated by glucose (57), and CCK-induced exocytosis in permeabilized pancreatic acinar cells was inhibited when the highly concentrated monomeric actin-binding protein β-thymosin was introduced (43). Moreover, disruption of F-actin filaments by using Clostridium botulinum C2 toxin preferably reduced the second phase of insulin secretion from islet β-cells (39). Thus the basic substructure of the actin cytoskeleton may also be essential for the normal recruitment of secretory granules to the plasma membrane. In addition, the actin network may provide contractile forces that expel the granule contents (60), since earlier studies have demonstrated that inhibition of myosin light chain kinase reduced insulin secretion from β-cells (39).
The members of Rho family proteins that may participate in the regulation of exocytosis through controlling reorganization of the cytoskeleton or other processes are still uncertain. It appears, however, that Rac proteins are particularly interesting in this context, as evidenced in several secretion systems, including mast cells, chromaffin cells, and neuronal cells (8, 11, 17, 30, 48). At least three Rac isoforms have been identified in mammals. Rac1 is ubiquitously expressed (18, 42), whereas Rac2 is hematopoietic specific and Rac3 is highly expressed in the brain (18, 56). It was found that constitutively active Rac enhanced regulated secretion, whereas dominant-negative Rac exhibited opposite effects in permeabilized mast cells (8, 48). Our previous study using clostridial toxins to specifically inactivate certain members of Rho proteins in pancreatic β-cells also implies that Rac and Cdc42 rather than Rho may be the candidate regulators implicated in stimulated insulin secretion (31).
In the present study, we provided evidence, for the first time, for the activation of Rac1 after glucose stimulation in insulin-secreting cells, as demonstrated by the translocation of Rac1 from the cytosol to the membranes (including plasmalemma). Importantly, insulin release stimulated by glucose alone or plus forskolin was significantly reduced in β-cells expressing dominant-negative Rac1. This inhibitory effect was more obvious on the sustained phase of insulin secretion, indicating that Rac1 might be involved in the recruitment of secretory granules by maintaining, at least in part, the F-actin structure in the interior of cytoplasm. Our results also suggest that cAMP and protein kinase A (PKA) may facilitate exocytosis of insulin via a mechanism dependent on the functional activation of Rac1.
MATERIALS AND METHODS
The plasmids containing human Rac1 mutants (pEXVV12rac and pEXVN17rac) were generously provided originally by Dr. Allan Hall (University College of London, UK). These two plasmids were constructed by cloning the dominant-negative form N17Rac1 (substitution of threonine-17 by asparagine) and the constitutively active form V12Rac1 (replacement of glycine-12 by valine) into vector pIREShyg1 (Clontech, Palo Alto, CA); the NH2-terminal of these Rac1 mutants was tagged with c-Myc as a mark that can be recognized by an anti-Myc antibody. A plasmid containing wild-type Rac1 was also generated in the similar way. Control (empty) vector was constructed by deleting the inserts from the above plasmids by digestion with 5′-BamH I and 3′-Not I restriction enzyme (Promega, Madison, WI), followed by blunt ligation using Klenow DNA polymerase I and T4 ligase (Promega). Plasmids were amplified in XL1 blue Escherichia coli and purified by using the NucleoBond Plasmid Kit (Clontech).
Cell culture and transfection.
INS-1 cells (passage 56–70) were grown in RPMI 1640 containing 10% FCS (GIBCO Invitrogen, Grand Island, NY), 50 μM 2-mercaptoethanol, and 1 mM pyruvate in culture flasks (Falcon; Becton-Dickson) at 5% CO2 (3). For transfection, cells seeded in six-well plates (Falcon) were cultured for 1 day to reach 80% confluence. Transfections were performed by using Superfect reagent according to the manufacturer's protocols (Qiagen, Hilden, Germany), and 1 μg of purified plasmid DNA was applied to each well. Control cells were transfected by incubation with an equivalent amount of empty vector DNA. Stably expressing (control, wild Rac1, and Rac1 mutant) cells were established by selecting transfected cells with 50 μg/ml hygromycin (Roche Diagnostics, Indianapolis, IN) for 4 wk and confirmed by Western blotting assay of cell lysates for Rac1 mutants (see below).
Subcellular fractions of INS-1 cells were isolated as previously described (35). All procedures were performed on ice. About 4.5 × 107 cells grown in 15-cm dishes were washed two times with cold PBS and scraped in 1 ml homogenization buffer (20 mM Tris·HCl, pH 7.4, 0.5 mM EDTA, 0.5 mM EGTA, 250 mM sucrose, and 1 mM dithiothreitol) containing the following protease inhibitors: leupeptin (10 μg/ml), aprotinin (4 μg/ml), pepstatin (2 μg/ml), and PMSF (100 μM). The cells were then disrupted by 10 strokes through a 27G needle. The cell homogenates were centrifuged at 900 g for 10 min to remove the nuclei and unbroken cells. Mitochondria-enriched, secretory granule-enriched, and microsomal fractions were isolated by centrifuging postnuclear supernatants at 5,500 g for 15 min (in Eppendoff 5417R), 25,000 g for 20 min, and 100,000 g for 60 min (in Beckman L-80), respectively. For the isolation of total membrane and soluble fractions, the postnuclear supernatants were centrifuged at 100,000 g for 60 min. The purity of isolated subcellular fractions from INS-1 cells was extensively assessed in our previous study (35) by analyzing the specific biochemical markers of organelles [cytochrome c for mitochondria, aryl sulfatase for lysosomes, insulin for secretory granules, NADPH reductase for endoplasmic reticulum (ER), and 5′-nucleotidase for the plasma membrane] for each of the fractions as well as by electron microscopy examination.
Isolation of the plasma membrane was carried out following a protocol by Hubbard et al. (26) that generates a highly purified plasma membrane fraction with only 4.6, 1.5, and 20% of contamination from mitochondria, lysosomes, and ER, respectively. In brief, the cell homogenates were centrifuged at 280 g for 5 min to remove the nuclei and unbroken cells. The pellet was resuspend in 1 ml medium A (0.25 M sucrose, 1 mM MgCl2, and 10 mM Tris·HCl, pH 7.4) and mixed well with 2 vol medium B (2 M sucrose, 1 mM MgCl2, and 10 mM Tris·HCl, pH 7.4). The mixture was overlaid with 0.5 ml medium A, and the tubes were centrifuged at 113,000 g for 1 h. A band at the interface containing the plasma membrane was collected and diluted to 2 ml with the homogenization buffer. The pellet was harvested after centrifugation at 3,000 g for 10 min. Protein concentrations in subcellular fractions were measured by using a Bio-Rad assay kit.
Detection of native Rac1, transfected wild-type Rac1, and Rac1 mutants by Western blotting.
Samples (20 μg protein each) harvested from subcellular fractionations were denatured in treatment buffer at 95°C for 5 min and separated on 15% polyacrylamide gels by electrophoresis. The separated proteins were electrotransferred to nitrocellulose membranes (Bio-Rad). After being blocked with 5% fat-free milk, blots were probed with either a monoclonal antibody against Rac1 (BD Biosciences) at 1:1,000 dilution or an anti-c-Myc (clone 9E10) monoclonal antibody (Roche Diagnostics) at a concentration of 2 μg/ml. The membranes were incubated with horseradish peroxidase-conjugated anti-mouse IgG antibodies (Santa Cruz) at a dilution of 1:1,000. Immunoreactive bands were visualized by the enhanced chemiluminescence system (Pierce Chemicals, Rockford, IL) and analyzed by densitometry (Bio-Rad ChemDoc).
For assessment of Rac1 translocation, cells cultured in 100-mm dishes for 3–4 days were washed two times with fresh Krebs-Ringer-HEPES buffer (KRB; see below for compositions). Subsequently, the cells were preincubated in KRB for 30 min at 37°C and then incubated with stimuli for indicated periods. Stimulation was stopped by aspirating the buffer and three washes with cold PBS. Cells were harvested for subcellular fractionation and Western blotting, as mentioned above.
Immunofluorescence staining of native Rac1 and Rac1 mutants.
INS-1 cells cultured on coverslips were washed two times in PBS (pH 7.4) and fixed for 10 min at room temperature in 3.7% paraformaldehyde in PBS. After two washes in PBS, the cells were permeabilized by a 10-min incubation in PBS containing 0.2% Triton X-100. After 30 min blocking in PBS containing 10% FCS, the cells were incubated in solution containing a monoclonal antibody against Rac1 (1:100 dilution in blocking buffer) or an anti-c-Myc monoclonal antibody (1:20 dilution) for 1 h at room temperature. The coverslips were washed six times in PBS and then incubated for 1 h with tetramethylrhodamine isothiocyanate (TRITC)-conjugated anti-mouse IgG (1:100 dilution; Sigma). Thereafter, coverslips were washed again six times in PBS and mounted on glass slides with mounting medium (Dako, Carpineria, CA). Samples were examined under a laser confocal microscopy (Olympus IX 70 fluoroview300).
Measurement of Rac1 GTPase activity.
Rac1 activity was determined by a pull-down assay using GST-PAK-CD (CD denotes to CRIB domain) as a probe that solely binds to active Rac1-GTP and Cdc42-GTP (40). INS-1 cells (∼1 × 107) were incubated with 2.8 or 15 mM glucose for 30 min. The cells were then washed with ice-cold PBS and incubated for 5 min on ice in 1 ml lysis buffer (50 mM Tris·HCl, pH 7.4, 30 mM MgCl2, 1% Triton X-100, 10% glycerol, 100 mM NaCl, 1 mM DTT, 1 μg/ml leupeptin, 1 μg/ml pepstatin, and 1 μg/ml aprotinin). Cells were harvested by lifter and centrifuged for 5 min at 21,000 g at 4°C. Aliquots (100 μl) were taken from the supernatant for detecting the total Rac1 amount. The remaining supernatants were incubated with 20 μg GST-PAK-CD fusion peptide (a generous gift from Dr. Edward Manser, Institute of Molecular and Cell Biology, Singapore) and 25 μl 50% glutathione-Sepharose 4B beads (Amersham) for 60 min at 4°C with gentle agitation. The mixture was centrifuged at 500 g for 5 min to sediment the matrix, followed by three washes with cold lysis buffer. The proteins were eluted by adding Lammili sample buffer and boiled for 5 min. The eluted samples, including those supernatant aliquots, were subjected to SDS-PAGE and Western blotting as described above.
Observation of cell morphology.
INS-1 cells were cultured in coverslip chambers (LAB-TEK; Nalge Nunc International, Naperville, IL) for 2–4 days. Living cell morphology was examined under differential interference contrast by laser confocal microscopy.
Assessment of F-actin filaments.
F-actin staining by phalloidin toxin was conducted as described previously (39). INS-1 cells were seeded on glass coverslips and cultured for 2 days. After two washes with PBS (pH 7.4), cells on coverslips were fixed with 3.7% paraformaldehyde in PBS for 10 min. Cells were washed two times again and then incubated with 330 nM rhodamine-phalloidin (Molecular Probes) and 300 ng/ml lysophosphatidylcholine (Sigma) for 20 min at room temperature. Subsequently, coverslips were washed two times with cold PBS and mounted on glass slides. The slides were examined by laser confocal microscopy.
Measurements of insulin secretion.
Insulin secretion was determined as described previously (3, 36). INS-1 cells were seeded in 24-well culture plates and cultured for 3–4 days. After two washes with KRB (containing in mM: 129 NaCl, 4.8 KCl, 1 CaCl2, 1.2 MgSO4, 1.2 KH2PO4, 5 NaHCO3, 2.8 glucose, and 10 HEPES, pH 7.4), the cells were preincubated for 30 min in 0.5 ml KRB at 37°C. Subsequently, the medium was replaced by 0.5 ml KRB containing 2.8 mM glucose (for basal secretion) or secretagogues, and the cells were incubated for the specified periods. The supernatants were removed for measurements of secreted insulin, and the attached cells were extracted by acid-ethanol for determination of insulin content. Insulin was assayed by an RIA kit (Linco Research, St. Charles, MO). The results of insulin secretion were expressed as percentages of insulin content to correct the possible influence of variance in the cell number between wells.
Nutrient metabolism, membrane potential, and intracellular free Ca2+ levels.
Nutrient metabolism was assessed by 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) test, which measures the intracellular reduction potential generated from both glycolysis and mitochondrial oxidation (36). For measuring membrane potential fluctuations, the fluorescent probe bisoxonol (a voltage-sensitive dye) was added to the cuvette (final concentration of 100 nM) containing ∼2 × 106 cells in suspension but without BSA. Fluorescent signal was recorded using a spectrofluorometer (Perkin-Elmer LS-50B) with excitation and emission wavelengths of 540 and 580 nm, respectively. The cells were allowed to equilibrate with the dye for ∼15 min to reach stable baseline before they were stimulated by test agents. For comparison and statistical analysis, the bisoxonol fluorescent signals were normalized by expression of results as the percentage of a near-maximal depolarization achieved by 40 mM K+ in each trace (36). [Ca2+]i levels were determined by the fluorescent probe fura 2, as described in detail previously (36).
Data were expressed as means ± SE and statistically analyzed by two-tail t-test.
Differential distribution of expressed Rac1 mutants compared with endogenous Rac1.
In control INS-1 cells, the endogenous Rac1 was mainly present in cytosol even though trace amounts were detected in the fractions enriched with mitochondria and secretory granules, as assessed by Western blotting (Fig. 1A, left). Immunofluorescence staining indicated that native Rac1 is distributed in whole cytoplasm (Fig. 1A, right).
An antiserum recognizing exclusively Myc-tagged proteins was employed for immunofluorescence staining of Rac1 mutants. No signal could be detected in control cells (additional data not shown). Although the staining in dominant-negative N17Rac1 transfected cells was observed primarily in the soluble compartment, considerable staining patterns were also detected around the cell periphery (Fig. 1B, left). The constitutively active V12Rac1 was more evenly distributed in the cells (Fig. 1B, right). Western blot analysis revealed that N17Rac1 was primarily detected in fractions enriched with mitochondria or insulin-containing secretory granules (Fig. 1C, lane 1), although V12Rac1 was distributed in all fractions, including the microsomal and cytosolic fractions (Fig. 1C, lane 3).
Glucose specifically stimulates translocation of Rac1 from cytosol to membranes in control but not in cells expressing the mutated Rac1.
When control INS-1 cells were stimulated by 15 mM glucose for 30 min, there was an increase of Rac1 in the membrane fraction, with a concomitant reduction in cytosol (Fig. 2B). The relative abundance of Rac1, expressed as its membrane-to-cytosolic ratio, was significantly increased (1.57-fold; from 0.19 ± 0.02 to 0.49 ± 0.04) after exposure to stimulatory concentrations of glucose (Fig. 2A). The cAMP-elevating agent forskolin (1 μM) did not significantly affect the degree of Rac1 translocation elicited by glucose. Such glucose-induced Rac1 translocation effect reached significance at 15 min (Fig. 2C). Further studies revealed that 15 mM glucose is able to specifically facilitate translocation of Rac1 into the plasmalemma fraction from the cytosol (Fig. 3). Importantly, activation of Rac1 by glucose was confirmed by the direct detection of Rac1 GTPase activity. Using a fusion peptide (GST-PAK-CD) that is solely bound to the activated Rac1 (GTP-Rac1) as a probe (40), we have been able to demonstrate that 15 mM glucose significantly increased GTP-bound Rac1 (i.e., elevation of Rac1 activity) in control INS-1 cells over 30 min stimulation (Fig. 4).
In contrast, stimulation by glucose plus forskolin for 30 min did not alter Rac1 distribution between cytosol and membranes in Rac1 mutant-transfected cells (Fig. 1C, lanes 2 and 4) when probed with an antibody that recognized solely the Myc-tagged Rac1 mutants. Interestingly, even though the basal levels of membrane-associated total Rac1 were higher by 32 and 62% in N17Rac1- and V12Rac1-transfected cells, respectively (assessed by an antibody that recognized both endogenous Rac1 and transfected Rac1 mutants), further stimulation of these cells by glucose failed to increase Rac1 association with the membranes (Fig. 2), including plasmalemma (Fig. 3). In addition, no apparent difference could be observed in the total Rac1 (both endogenous and mutated) content between control and the two Rac1 mutant-transfected cells (data not shown).
Expression of dominant-negative Rac1 significantly alters the morphology and causes disruption of F-actin filaments in INS-1 cells.
The control INS cells grew in monolayer on the culture surface with substantial lamellipodia (Fig. 5A). Remarkable changes in morphology occurred in N17Rac1-transfected cells, which had less lamellipodia [a known morphological marker of Rac function (19)] and tended to aggregate and stack (Fig. 5B). These cells rounded up, and their ability to adhere to the substratum was also weaker since they were more easily detached by trypsinization. Besides, the size of these cells was significantly smaller (diameter of 9.9 vs. 12.6 nm of control cells) in suspensions. By contrast, INS-1 cells expressing the V12Rac1 only revealed slight morphological changes, including more protrusions (Fig. 5C). In addition, these cells were more resistant to trypsinization. These observations indicate that active Rac1 may be important for INS-1 cells to maintain normal morphology.
The observed morphological changes might be the result of the interference with cytoskeleton organization in the transfected cells. Control INS-1 cells displayed strong F-actin distribution in cell periphery and also some filaments in the interior of cells, as assessed by rhodamine-phalloidin staining (Fig. 5D). By contrast, F-actin staining in N17Rac1-transfected cells almost completely disappeared (Fig. 5E). Despite some modest degree of reduction of F-actin staining in V12Rac1 transfected cells, the structure of F-actin filaments remained largely intact in these cells (Fig. 5F). These findings suggest that the organization of F-actin filaments in INS-1 cells requires functional Rac1.
Expressing dominant-negative N17Rac1 inhibits insulin secretion in INS-1 cells.
Stable expression of dominant-negative or -active Rac1 mutants did not alter either the insulin content in INS-1 cells (Table 1) or the basal insulin secretion rate (Fig. 6). Insulin secretion from control cells over 30 min was increased by 90% after exposure to 15 mM glucose and by >2.9-fold with simultaneous incubation with glucose and forskolin (Fig. 6). Expression of dominant-negative N17Rac1 significantly inhibited glucose-induced (>50%) and glucose- plus forskolin-induced (60%) insulin secretion (Fig. 6). However, insulin release induced by glucose alone or glucose plus forskolin was not significantly affected in cells expressing constitutively active V12Rac1. In addition, high-K+-stimulated secretion was not significantly affected in cells transfected with either of the Rac1 mutants (Fig. 6). Furthermore, stable expression of wild-type Rac1 (Fig. 7A) also did not alter secretagogue-induced insulin release (Fig. 7B), and morphology of these cells remained unchanged (data not shown), suggesting the specific N17Rac1 effects resulting from interference with Rac1. These data indicate that Rac1 may be involved in glucose- and forskolin-stimulated insulin secretion from INS-1 cells.
It is well known that glucose evokes a biphasic pattern of insulin secretion from islet β-cells, and each of the two phases are felt to be controlled by different mechanisms (7, 21, 53). To determine the possibly distinct roles of Rac1 in the two phases, we measured insulin secretion from the dominant-negative Rac1-transfected cells during two consecutive static stimulating periods [initial 15 min as the early secretion event and followed by a 20-min incubation reflecting the late (sustained) secretion event], an approach used previously to study the early and late-phase insulin secretion responses to glucose in isolated islets (44). The potentiating effect of forskolin on glucose-stimulated insulin secretion was greater (by 165%) during the late period than the early phase (by 101%) in the control cells (Fig. 8). In cells expressing N17Rac1, the early phase of insulin release stimulated by glucose plus forskolin was significantly inhibited (−42%), albeit without significant effect on the release elicited by glucose alone (Fig. 8A). Furthermore, the degree of insulin secretion induced by either glucose alone or plus forskolin during the late period was markedly inhibited by 63 and 49%, respectively (Fig. 8B). These results implicate that functional activation of Rac is essential for the late phase of insulin release, an effect that may be related to the Rac1 translocation and activation by glucose during this period (cf. Fig. 2C).
Dominant-negative Rac1-mediated inhibition of insulin secretion does not appear to involve nutrient metabolism, membrane potential, and [Ca2+]i increases in Rac1-mutated cells.
Incubation of control cells with 15 mM glucose for 30 min increased MTS reduction (reflecting metabolism) by about threefold over basal, an effect not significantly influenced by the expression of either Rac1 mutant (Table 1). In addition, there were no significant differences in either resting membrane potential or depolarization (induced by 15 mM glucose, 100 μM ATP, or submaximal KCl) between control cells and N17Rac1 transfected cells (Fig. 9, A-D, and Table 1). Furthermore, [Ca2+]i increments were also not affected by the expression of dominant-negative Rac1 when cells were stimulated by 15 mM glucose or 34 mM KCl (Fig. 9, E-H, and Table 1). These findings suggest that functional activation of Rac1 may not be requisite for the generation of proximal signals required for the exocytotic secretion of insulin in INS-1 cells elicited either by glucose or KCl.
Our data revealed, for the first time, that a physiological concentration of glucose is able to cause Rac1 translocation from the soluble to the membranous fraction in insulin-secreting INS-1 cells, a widely held proposal of Rac1 activation. Further direct evidence for Rac1 activation by glucose is the increase of Rac1 activity upon glucose stimulation. Expression of a dominant-negative Rac1 abolished glucose-induced Rac1 translocation and inhibited mainly the late phase of glucose and forskolin-stimulated insulin secretion. This action occurred without apparently affecting the secretory response to elevated [Ca2+]i per se, suggesting that the very distal steps in secretory machinery remain largely intact. The disappearance of F-actin filaments in these cells indicated that F-actin fibers may be necessary for the recruitment of secretory granules to the plasma membrane for exocytosis.
Earlier studies from our laboratory (31, 32) and by others (9) have provided indirect evidence for the involvement of the Rho subfamily of G proteins in physiological insulin secretion. For example, inhibition of requisite posttranslational modifications (i.e., farnesylation, carboxyl methylation, and fatty acylation) of these G proteins results in a marked reduction of glucose-stimulated insulin secretion from normal rat islets and clonal β-cell preparations (2, 38, 41). Furthermore, Clostridial toxins (e.g., toxin B and lethal toxin), which selectively glucosylate and inactivate Rho subfamily GTP-binding proteins (1), also produce similar effects (31). In addition, we have been able to demonstrate (32) that, under conditions of stimulated insulin secretion, glucose augmented the carboxyl methylation and membrane association of the Rho family of G proteins, including Cdc42. Together, these findings suggest the involvement of Rho GTPases in insulin secretion. Our current observations provide the direct evidence with regard to the identity of one of these G proteins as Rac1, since we have demonstrated that expression of the dominant-negative form of Rac1 results in marked reduction in glucose- and cAMP-mediated insulin secretion from INS cells. This finding is similar to the results by introducing dominant-negative Rac2 in permeabilized mast cells (8).
The dominant-negative N17Rac1 and constitutively active V12Rac1 are the two Rac mutants most widely used to study the Rac1 function. They have different mechanisms in the regulation of functional activation of Rac1 (14, 51). It is felt that N17Rac1 interferes with the activation of its normal counterpart by competing with the binding to guanine nucleotide exchange factors (GEFs) and thus blocking the functions of endogenous Rac1 (14). On the other hand, a mutation of Rac1 at glycine-12 (to valine-12) decreases its intrinsic and GAP-stimulated GTPase activity, allowing proteins to remain predominantly in the GTP-bound, active form (51). The mechanisms for the effects of V12Rac1 on cellular functions are less clear and complex. This mutant may directly activate certain Rac1-mediated pathways through interaction with putative effectors. However, it may also block the interaction of endogenous Rac1 with downstream targets by trapping Rac1 effectors in the cytosol and thus preventing their association with target membranes (14). Nonetheless, its ability to trigger some Rac1-mediated events may be dominant, as evidenced by the stimulation of exocytosis by introducing constitutively active Rac mutants in permeabilized mast cells (8, 48). However, such secretion-stimulatory action did not occur in wild Rac1- or V12Rac1-transfected INS cells, suggesting that activation of Rac1 may be necessary but insufficient to initiate and maintain stimulated insulin secretion.
An important observation in this study is that glucose stimulation was able to cause translocation of Rac1 from cytosol to membranes, including plasmalemma, in control INS-1 cells. Translocation of Rac1 to membranes occurs in other cell systems upon stimulation. A typical example is that, in neutrophils stimulated by N-formyl-methionyl-leucyl-phenylalanine or PMA, the cytosolic Rac1 and Rac2 were translocated to cell membranes, a process required for the activation of NADPH oxidase (12). Similar observations are also reported in ANG II-stimulated cardiomyocytes and vascular smooth muscle cells (34, 62), in serum-activated MDCK cells (20), and in platelet-derived growth factor-stimulated Swiss 3T3 fibroblasts (16). However, the distribution patterns of either dominant-negative or -active Rac1 mutants in our transfected cells were not changed when stimulated by secretagogues. Importantly, glucose was no longer able to induce endogenous Rac1 translocation in cells expressing either Rac1 mutant. This failure appears not to be the result of the impairment of glucose metabolism and other early steps in stimulus-secretion coupling (e.g., membrane depolarization and elevation of [Ca2+]i), rather it seems more likely the result of the interference with endogenous Rac1 by the mutants directly. The inhibitory effects of N17Rac1 could be mediated by the following two modes: preventing the activation of endogenous Rac1 by competitive binding to GEFs (14) and occupying the target sites of Rac1 on the membranes because of its dominant distribution in the membranes.
It is well known that glucose induces biphasic insulin secretion (7, 21, 53). Accumulating evidence favors the two-compartmental model (7, 53), in which the second phase of secretion requires the promotion/recruitment of secretory granules to the exocytotic sites, although the details of underlying mechanisms remain to be defined. Our studies revealed that only the late phase of glucose-stimulated insulin secretion was significantly reduced in N17Rac1-transfected INS-1 cells. Moreover, although both early and late insulin secretion events induced by forskolin plus glucose were attenuated in these cells, a larger effect occurred during the late period. These results suggest that Rac1 may play a greater role in the late phase of glucose-stimulated insulin secretion. That fact that glucose significantly increased Rac1 translocation to membranes only after 15 min of stimulation also supports this notion. The inhibitory effect on insulin secretion by N17Rac1 might be attributable to the interference with glucose-induced activation of endogenous Rac1, as discussed above.
Although an involvement of Rac proteins in the Ca2+-dependent exocytosis has been reported in neuronal cells (11, 30), we found that expression of the inactive Rac1 mutant did not significantly affect the insulin secretion stimulated by Ca2+ per se (through high-K+ depolarization) in INS-1 cells. Thus it seems that Rac1 does not engage in the very distal step of exocytosis in INS-1 cells. In addition, the preferred inhibition of the late secretion event by the expression of N17Rac1 may also explain its failure to significantly inhibit insulin secretion induced by high K+, since only a monophasic insulin release (equivalent to the early phase) is evoked under this stimulatory condition (55).
There is strong evidence that Rac is a key control element in the reorganization of actin cytoskeleton (19). Microinjection of V12Rac was capable of inducing lamellipodia and membrane ruffle as well as subsequent stress fiber formation, whereas injection of N17Rac abolished these effects in fibroblasts (52). Thus it is not surprising that the morphology of β-cells would be altered by interference with Rac function. The morphological changes in our Rac1 mutant-transfected cells were accompanied with the alterations of F-actin distribution. F-actin filaments almost completely disappeared in N17Rac1-transfected cells, and these cells were rounding up. Surprisingly, expression of constitutively active Rac1 mutant did not enhance the formation of actin cytoskeleton as in other cell types (52), but rather had a small opposite effect. Importantly, F-actin structure may be also involved in the transport of secretory granules, since loss of these F-actin filaments in N17Rac1-transfected cells displayed preferred inhibition of the late phase of stimulated insulin secretion. Our earlier study using Clostridium botulinum C2 toxin to disrupt F-actin structure also found a more severe inhibition of the second phase of insulin secretion in HIT-T15 cells and in isolated rat islets (39). Together, these findings suggest that the role of Rac1 in regulated insulin secretion may involve the regulation of actin cytoskeleton reorganization for secretory granule recruitment, which is required for the second phase of secretion (24, 39).
One should be aware of the possible nonspecific effects resulting from overexpression of a protein in transfection experiments. However, three lines of evidence suggested that such an effect was minimal or unlikely in our study. First, overexpression happens frequently in transient transfection. In our case, stable transfection was performed, and it did not produce overexpression, since the total amount of Rac1 (endogenous plus mutant) in the homogenates of Rac1 mutant-transfected cells was not different from that of control cells. Second, we only observed marked alterations of the insulin secretory response, cell morphology, and F-actin cytoskeleton in cells transfected with N17Rac1, but not V12Rac1, suggesting a specific effect of the former. Third, stable expression of wild-type Rac1 had no effect on morphology and secretagogue-induced insulin release in INS-1 cells.
Our data indicate that Rac1 may also be involved in the cAMP-modulated insulin secretory pathway. It is well known that PKA activation by cAMP can positively influence regulated secretion in many cell systems, including the islet β-cell. We observed that forskolin potentiated both early and late-phase insulin secretion but that this effect was more significant on the late secretion event. It has been reported that PKA can regulate Rac1-dependent organization of the actin cytoskeleton during the migration of carcinoma cells (46). It is thus likely that Rac1 inactivation may affect the PKA-mediated potentiation of insulin secretion through disruption of the actin cytoskeleton organization in our INS-1 cells. Indeed, the late phase of forskolin-potentiated insulin secretion was more heavily inhibited in N17Rac1-transfected cells, suggesting that forskolin-activated PKA may be important for the trafficking of secretory granules in β-cells (22, 49).
J. Li is a recipient of a Research Scholarship from the National University of Singapore. This work was supported by National Medical Research Council of Singapore Grants NMRC/0540/2001 and NMRC/0803/2003 (to G. D. Li) and a Merit Review Grant from the Department of Veterans Affairs (to A. Kowluru). A. Kowluru is also a recipient of a Research Career Scientist Award from the Veterans Affairs Medical Research Service.
We are very grateful to Dr. Edward Manser for providing the GST-PAK-CD fusion peptide and helping set up of the Rac GTPase assay. We thank Dr. Stewart A. Metz for critical reading and comments to this paper.
Portions of this work were presented at the 37th Annual Meeting of the European Association for the Study of Diabetes in Glasgow, UK, September 2001.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 by American Physiological Society