Molecular Basis of Insulin-stimulated GLUT4 Vesicle Trafficking

Among all the diverse actions of insulin, one of the most critical and intensively studied is its regulation of glucose homeostasis. In the postabsorptive state, insulin action in muscle and adipose tissue results in increased glucose uptake from the circulation, thereby maintaining plasma euglycemia and preventing hyperglycemia (1–3). Defects in this pathway result in insulin resistance, a condition in which excessive concentrations of insulin are required to reduce blood glucose levels and often precede the development of frank Type II diabetes (3, 4). Although it has been appreciated for almost two decades that this major action of insulin results from a redistribution of glucose transporter proteins from intracellular storage sites to the plasma membrane (5, 6), the cellular mechanisms responsible for these trafficking events and the defects associated with insulin resistance have remained largely enigmatic.

lead to GLUT4 translocation. At present, substantial evidence has documented that the activation and/or appropriate targeting of the Type I phosphatidylinositol (PI) 1 3-kinase is necessary for insulinstimulated GLUT4 translocation and glucose uptake (19). However, the specific downstream targets for PI 3-kinase have remained questionable with evidence both for and against protein kinase B, also known as Akt (20 -22). Alternatively, the atypical protein kinase C isoforms (protein kinase C and ) are downstream targets of PI 3-kinase and have also been implicated in the insulin stimulation of GLUT4 translocation (23,24,81).
In addition to PI 3-kinase-dependent signal transduction pathways, several studies have demonstrated the presence of PI 3-kinase-independent mechanisms of GLUT4 translocation. For example, osmotic shock, GTP␥S, and exercise/contraction in skeletal muscle induce GLUT4 translocation in the complete absence of PI 3-kinase activity (25)(26)(27)(28). These treatments can result in increased AMP levels, and it has been suggested that activation of the AMPdependent protein kinase stimulates glucose uptake (29). G protein-coupled receptors linked to G q have also been observed to induce GLUT4 translocation (30). Future studies will be necessary to ascertain the specific roles and relationships between protein kinase B, protein kinase C and , and/or other PI 3-kinase-dependent as well as independent signaling processes in GLUT4 translocation. In any case, whatever specific signal transduction pathway(s) are involved, they must directly impact on the GLUT4 intracellular storage sites and its subsequent trafficking to and fusion with the plasma membrane.

Identification of Intracellular GLUT4 Storage
Compartments It is well established that integral membrane proteins traffic to and from the plasma membrane through an endosomal recycling membrane system. Immunoelectron microscopy has demonstrated that the GLUT4 protein is predominantly localized in small vesicles and tubulovesicular structures, with relatively smaller amounts found in the trans-Golgi network, clathrin-coated vesicles, and endosomes (31)(32)(33)(34)(35). Subcellular fractionation and immunoabsorption studies have indicated a substantial co-distribution of GLUT4 with markers of several intracellular membrane compartments including GLUT1 and the transferrin and the insulin-like growth factor-2/mannose 6-phosphate receptors (9,10). However, several studies have observed the presence of a separate subpopulation of GLUT4-containing vesicles different from the constitutively recycling endosomal system. For example, GLUT4-containing vesicles are enriched in the v-SNARE protein, VAMP2, but not the related VAMP3/cellubrevin isoform, which is present in the constitutively recycling endosome population (36 -38). The presence of distinct vesicle populations may also account for the ability of insulin to stimulate 2-4-fold plasma membrane increases of several recycling proteins (i.e. GLUT1, transferrin, and insulin-like growth factor-2/mannose 6-phosphate receptors), whereas plasma membrane GLUT4 content can increase 10 -20-fold. Thus, it appears that this unique subpopulation of GLUT4-containing vesicles is primarily responsible for the majority of insulin-stimulated translocation.
Irrespective of which vesicle population is insulin-responsive, there must be specific mechanisms in place to localize GLUT4 to its appropriate intracellular storage sites. To address this issue, several laboratories have examined the subcellular localization of GLUT4/GLUT1 and GLUT4/transferrin receptor chimeric proteins. Steady-state analysis of the C-terminal region of GLUT4 identified a dileucine motif (SLL) that when mutated resulted in the plasma membrane accumulation of GLUT4 (39 -42). However, further mutational studies and kinetic analysis revealed that this motif was, in fact, required for efficient internalization (43,44). Thus the apparent lack of intracellular sequestration in these mutants resulted from an accumulation of GLUT4 at the plasma membrane because of a reduction in the rate of endocytosis. Even though a contiguous GLUT4 amino acid sequence responsible for intracellular sequestration or retention has not yet been identified, kinetic and morphological data strongly suggest the presence of an intracellular targeting signal within the GLUT4 cytoplasmic Cterminal domain.

Retention versus Synaptic Vesicle Model of GLUT4
Translocation There are two general models that can account for the insulinstimulated translocation of these intracellular GLUT4-containing vesicles to the plasma membrane. Schematic representations of the retention and synaptic models of GLUT4 vesicle trafficking are illustrated in Figs. 1 and 2. The retention model predicts that the specific sequences within GLUT4 or other co-localized vesicular proteins (see below) specifically target this vesicle population away from recycling endosomes ( Fig. 1). Insulin stimulation allows these sequestered vesicles to enter the constitutive recycling endosome pathway and thereby results in default trafficking to the plasma membrane. This model is consistent with the predominant localization of GLUT4 to small vesicles and tubulovesicular structures adjacent to endosomes typically underlying the plasma membrane (32,34,35). In addition, expression of GLUT4 in neuroendocrine cells and fibroblast cell lines results in the localization of GLUT4 to secretory granules and to endosomes (42,(45)(46)(47). These data are consistent with the presence of specific GLUT4 intracellular sequestration sequences that can function in heterologous cell systems.
A functional regulatory property of the retention model can be proposed based upon the recent observation that expression of the GLUT4 C-terminal domain results in the translocation of the endogenous GLUT4 protein to the plasma membrane (48). Similarly, the insulin-responsive aminopeptidase (IRAP), also known as vp165, co-localizes to biochemically and morphologically identical GLUT4 vesicles (49,50). The cytoplasmic N-terminal domain of IRAP/vp165 displays some homology with the GLUT4 C terminus, and introduction of the cytoplasmic domain of IRAP/vp165 also results in plasma membrane translocation of GLUT4 (51). This model also accounts for the apparent masking of the GLUT4 Cterminal domain, which becomes more immunoreactive following insulin stimulation (34,35). The simplest interpretation of these data is that insulin-responsive cells contain a saturable retention receptor, which is responsible for the intracellular sequestration of the GLUT4-containing vesicles. Disruption of this interaction by either insulin stimulation or through competition with an appro-priate peptide allows the GLUT4-containing vesicles to enter the recycling endosome system, thereby resulting in plasma membrane translocation.
Although several aspects of the retention model are quite appealing, there are several lines of evidence inconsistent with this model. First, a substantial proportion of the GLUT4-containing vesicles appears to be segregated away from recycling endosome markers, having characteristics analogous to small synaptic vesicles. This includes the co-localization of GLUT4 with VAMP2 and the fact that a substantial amount of the GLUT4 protein remains distinct from the recycling of the transferrin receptor (36,37). Furthermore, ablation of the transferrin receptor-containing endosomes does not prevent insulin-stimulated GLUT4 translocation. These data demonstrate that insulin-stimulated GLUT4 trafficking can occur independent of the recycling endosome system, at least the endosome population defined by the transferrin receptor. In either case, it is important to recognize that both the retention and synaptic vesicle models are not necessarily mutually exclusive. For example, insulin could induce the trafficking of multiple GLUT4 vesicle populations through different mechanisms. Alternatively, GLUT4 may be stored in synaptic-like vesicles that are sequestered because of GLUT4 and/or IRAP/vp165 retention signals.

SNARE Model of GLUT4 Trafficking
In contrast to GLUT4 translocation, substantial progress has been made in our understanding of synaptic vesicle trafficking in the regulation of neurotransmitter release from the presynaptic membrane (52)(53)(54). In this process, protein complexes in the vesicle compartment (v-SNAREs, for vesicle SNAP receptors) pair with their cognate receptor complexes at the target membrane (t-SNAREs, for target membrane SNAP receptors). Direct interactions between the t-SNARE and v-SNARE proteins in combination with the association of several accessory proteins (for example, ␣-SNAP, N-ethylmaleimide-sensitive factor, synaptophysin, and synaptotagmin) are responsible for determining trafficking specificity and formation of the core complex required for membrane fusion.
In adipocytes, the insulin-stimulated trafficking of GLUT4 vesicles has several features in common with the regulated exocytosis pathway of synaptic vesicles, including the presence of specialized vesicular proteins (Fig. 2). As described previously, the v-SNARE VAMP2 co-localizes with GLUT4 vesicles and, based on several criteria, appears to be essential for insulin-stimulated GLUT4 translocation. Although there are no specific endoproteases that cleave either syntaxin 4 or SNAP23/Syndet plasma membrane t-SNARE proteins, studies using blocking antibodies, dominantinterfering mutants, and/or peptide inhibitors have strongly impli- cated syntaxin 4 and SNAP23/Syndet as the t-SNARE proteins required for insulin-stimulated GLUT4 translocation (55)(56)(57)(58).
In the case of synaptic transmission there are also several modulators of v-and t-SNARE interactions. The mammalian Munc18 proteins, homologues of Saccharomyces cerevisiae n-Sec1/rbSec1, Caenorhabditis elegans unc18, and Drosophila melanogaster Rop proteins, bind with high affinity to their cognate plasma membrane syntaxins. Null mutations in the n-Sec1, unc18, or Rop genes dramatically reduce vesicle exocytosis, suggesting that these proteins are essential for normal v-and t-SNARE function. The Munc18a isoform is predominantly expressed in neurons, where it inhibits the association of VAMP1 and SNAP25 with syntaxin 1 (59). In contrast to Munc18a, Munc18b and Munc18c are expressed ubiquitously, but only Munc18c binds syntaxin 4 (60,61). In analogy to Munc18a, the binding of Munc18c to syntaxin 4 precludes the binding of VAMP2, and increased expression of Munc18c inhibits insulin-stimulated GLUT4 translocation (62,63). Because insulin treatment apparently reduces the binding of Munc18c to syntaxin 4, this suggests that GLUT4 translocation is inhibited by Munc18c in the basal state and that insulin stimulation may result in a removal of the Munc18c repression.
More recently, we have identified a novel protein (Synip) that specifically binds to syntaxin 4 and is uniquely expressed in muscle and adipose tissue. 2 Similar to Munc18c, the binding interaction of Synip with syntaxin 4 is regulated by insulin, which also induces a dissociation of the Synip-syntaxin 4 complex. However, in contrast to Munc18c, expression of the wild type Synip protein has no effect on insulin-stimulated GLUT4 or IRAP/vp165 translocation. Furthermore, expression of the C-terminal domain of Synip does not display an insulin-stimulated dissociation from syntaxin 4 but functions in a dominant-interfering manner by inhibiting both GLUT4 and IRAP/vp165 translocation. These data suggest that although Munc18c may have an inhibitory role, Synip functions as a positive regulator of insulin-stimulated GLUT4 translocation. Future studies will be necessary to determine the stoichiometry of these interactions and the relationship between Synip-syntaxin 4 and Munc18c-syntaxin 4 complexes.
In addition to the involvement of these SNARE proteins in GLUT4 translocation, the small GTP-binding protein Rab4 also appears to play an important functional role. Based upon the relationship between Rab3A function and synaptic transmission, insulin-responsive tissues were found to express several Rab isoforms. In particular, Rab4 was found to co-localize with the GLUT4-enriched microsome fraction (65). Furthermore, insulin stimulation resulted in a redistribution of Rab4 into the cytosolic fraction. More recent studies have also demonstrated that introduction of a Rab4 C-terminal peptide, expression of a C-terminal Rab4 deletion mutant, or high level expression of wild type Rab4 all result in an inhibition of insulin-stimulated GLUT4 translocation (66 -68). These data have been interpreted to indicate that Rab4 is an essential component in the GLUT4 translocation machinery.

Regulation of GLUT4 Endocytosis
Accompanying the paucity of information available about GLUT4 exocytosis, even less is known about the mechanism of plasma membrane GLUT4 retrieval (endocytosis). Kinetic analysis has indicated that insulin increases the rate of GLUT4 exocytosis approximately 10 -20-fold with a 2-3-fold decrease in the rate of endocytosis (10,69,70). Although insulin-stimulated exocytosis requires activation of the PI 3-kinase, inhibition of this activity has no effect on the rate of endocytosis (71). Electron microscopic analysis has demonstrated that in the basal state, the small amount of plasma membrane-associated GLUT4 is almost exclusively found in coated pit regions (32,33,72). However, following insulin stimulation only a fraction of the plasma membrane-localized GLUT4 is found in coated pits whereas the majority of GLUT4 is dispersed in non-coated plasma membrane regions (32,72). Although the mechanism responsible for this observation is not known, the insulinstimulated decrease in GLUT4 coated pit localization could result from either saturation of the targeting machinery or specific segregation of the GLUT4 protein. In any case, disruption of clathrin lattices by potassium depletion or acidification of the cytosol has clearly established a role for clathrin-mediated endocytosis in GLUT4 internalization (44,73).
One essential molecule implicated in endocytic coated vesicle formation is the GTPase dynamin (Fig. 3). In addition to its highly conserved N-terminal GTPase domain, dynamin contains a central pleckstrin homology domain and a C-terminal proline-rich region that negatively and positively regulate the dynamin self-assembly and GTPase activities, respectively (74,75). These domains can functionally interact with several signaling intermediates, including Grb2, inositol phospholipids, and amphiphysin (76,77). In this regard, insulin has been observed to induce the association of a dynamin-Grb2 complex with tyrosine-phosphorylated Shc and IRS1, which in turn modulates the rate of dynamin GTPase activity (78,79). Furthermore, insulin has been reported to induce the tyrosine phosphorylation of dynamin (79). Together, these data suggest that insulin may regulate the endocytic rate of constitutively recycling cell surface proteins such as GLUT4 through a dynamin-dependent mechanism.
Consistent with this hypothesis, expression of a dominant-interfering dynamin mutant (K44A/dynamin) completely prevented GLUT4 endocytosis without affecting insulin-stimulated GLUT4 translocation (16,17,80). In addition, microinjection of a dynamin peptide encompassing the amphiphysin binding sites prevents GLUT4 endocytosis and results in the accumulation of GLUT4 at the cell surface (64). These data provide compelling evidence that the internalization of GLUT4 occurs through a clathrin-coated pit-mediated dynamin-dependent endocytic pathway. These findings underscore the importance of dynamin function in GLUT4 endocytosis and suggest that the assembly and/or dynamin-dependent formation of coated vesicles may be a target for the insulin regulation of GLUT4 endocytosis.

Summary and Conclusions
Since the original observation nearly 20 years ago that insulin stimulates the translocation of glucose transport proteins from intracellular storage sites to the plasma membrane, we have considerably advanced our understanding of the insulin signal transduction pathway, the molecular/cellular regulation of glucose transport, and their relationships to whole body glucose metabolism. During this time, four functional facilitative glucose transporter isoforms have been identified, cloned, and characterized. We now know that the GLUT4 isoform is primarily responsive to insulin and accounts for the majority, if not all, of insulin-stimulated glucose transport in muscle and adipose tissue under normal physiological circumstances. In the basal state, this transporter is distributed between multiple intracellular membrane compartments, some of which overlap with the recycling endosome system FIG. 3. The internalization of GLUT4 occurs through a dynamindependent mechanism. Following the insulin-stimulated translocation of GLUT4 to the plasma membrane, the GLUT4 protein accumulates in clathrin-coated (black lines) invaginations of the plasma membrane. This presumably occurs through interaction of GLUT4 with the AP2 complex (red) and synaptotagmin (yellow lines). Dynamin (green collar) is recruited to these structures and is required for the formation of the intracellular clathrincoated vesicles. These vesicles rapidly lose their clathrin coat structure and are then available to either undergo another round of translocation and/or to repopulate the intracellular GLUT4 storage sites. and others which may be localized to unique vesicular compartments. Future studies are needed to unambiguously identify the intracellular retention signals present in GLUT4, the mechanism by which GLUT4 is targeted to these intracellular storage sites and the functional properties of these compartments. Although it is evident from this brief review that precisely defining the GLUT4 intracellular trafficking pathways continues to be a daunting task, substantial progress is being made on characterizing the specific vand t-SNARE proteins directly involved. Clearly, the next major advance will require the identification of the downstream signal transduction pathway that regulates the priming/docking and/or fusion of these GLUT4 vesicles with the plasma membrane. Progress on all these issues is continuing at an accelerating pace, and the future looks bright for the ultimate development of an integrated and uniform model that links insulin action at the molecular level with the physiological control of glucose homeostasis.