The glucose transport effector system
GLUT4 vesicle translocation and trafficking
Glucose transport proteins are the key functional units of the glucose transport effector system. Multiple glucose transporter genes have been identified [66,135] that encode a family of homologous proteins exhibiting different functional properties and marked differences in tissue-specific expression. Members of the GLUT gene family are delineated in Table 12.1. On the basis of hydrophobicity plots, GLUT proteins contain 12 α-helical transmembrane-spanning domains with both the NH2 and COOH terminals extending into the cytoplasm. This model forms six exofacial loops and five endofacial loops. Regarding GLUT1-4 isoforms (class I in Table 12.1), the five endofacial loops include a large hydrophilic central loop, while the first exofacial loop is the largest and contains an asparagine-linked glycosylation site (N57). GLUT4 glycosylation first occurs in the ER, and as GLUT4 travels on to the Golgi apparatus, additional glycan modification can occur in a multistep process, producing multiple versions of GLUT4, distinguishable from each other due to alternative side chain branching or differences in chain length. Alternatively N-glycosylated versions of GLUT4 may explain why GLUT4 differentially migrates during SDS-PAGE separation . A full glycosylation profile of GLUT4 has been shown to be very important for GLUT4 protein stability [136,137] and GLUT4 localization . Thus, GLUT proteins are intimately embedded in membranes, and the most highly conserved regions are the putative membrane-spanning domains that serve a common function, the creation of a pore for facilitative diffusion of monosaccharides.
The major insulin-responsive glucose transporter isoform is GLUT4, which is predominately expressed in insulin target tissues such as skeletal and cardiac muscle and adipose tissue. In the basal state, GLUT4 is predominantly localized intracellularly. The major mechanism by which insulin activates glucose transport activity is by recruiting intracellular GLUT4 to the plasma membrane, which results in facilitative diffusion of glucose into the cell where glucose is rapidly phosphorylated by hexokinase and metabolized. Importantly, the glucose transport step is rate-limiting for insulin-stimulated glucose uptake and metabolism in peripheral target tissues under physiologic conditions . Upon dissipation of the insulin signal, deactivation of glucose transport activity is the result of a net reverse translocation of GLUT4 back into the cell interior.
Based on subcellular membrane fractionation experiments, insulin-mediated translocation was initially conceptualized as a two-compartment process with recruitment of GLUT4 from Golgi-enriched microsomal membranes to the plasma membrane. However, these studies consistently indicated that only ∼50 – 60% of microsomal GLUT4 translocated to the plasma membrane fraction, suggesting that there were both insulin responsive and nonresponsive intracellular pools. It is now clear that GLUT4 trafficking involves multiple intracellular membrane compartments [66,105,139,140]. GLUT4 protein is synthesized in the endoplasmic reticulum and travels to the trans-Golgi network (TGN), where GLUT4 is then inserted into highly specialized vesicles. In basal cells, morphologic studies employing immunohistochemistry and electron microscopy have detected GLUT4 in the perinuclear TGN, in cytosolic tubulovesicular elements, and in subplasma membrane vesicles. In one study, disruption of the TGN using the antibiotic brefeldin A did not impair insulin-stimulated glucose transport activity or surface appearance of GLUT4, ruling out TGN-based GLUT4 as being part of insulin-stimulated glucose uptake . GLUT4 was shown to be largely recruited from small vesicles (GLUT4 storage vesicles, or “GSVs” ) located near the endofacial surface of the plasma membrane . Other morphologic data have implicated tubulovesicular structures as the source of the translocating GLUT4 pool . In addition to the GSV pool, GLUT4 also cycles through an early endosomal compartment in an endocytotic/exocytotic pathway (distinct from fluid-phase endocytosis) analogous to that described for other insulin-responsive proteins, such as the mannose-6-phosphate/insulin-like growth factor 2 receptor and the transferrin receptor. GLUT4 in the endosomal recycling system is clearly distinct from the intracellular, insulin-inducible GLUT4 storage compartment [143 – 147]. For example, ablation of the recycling compartment, accomplished via internalization of a transferrin-horseradish peroxidase conjugate, eliminates vesicles containing transferrin receptors and VAMP3/cellubrevin (endosomal markers) but leaves intact a large fraction of GLUT4-containing vesicles. There are two models for GLUT4 localization post-TGN: (1) the dynamic retention model states GSVs cycle between the specific, insulin-sensitive GSV compartment and the endosomal recycling pool, and (2) the static retention model which proposes GSVs are locked in place until insulin signals the release of the GSV for subsequent membrane translocation. Within both models, GSVs are sequestered intracellularly under basal conditions, and it is the GSV, not the endosomal GLUT4, that represents the major source for insulin-mediated exocytosis and translocation of GLUT4 to the plasma membrane. This unique GSV storage compartment can explain why insulin’s effect to augment cell surface proteins is much greater for GLUT4 (10 – 40-fold over basal) compared with IGF-1 and transferrin receptors that traffic only in the endosomal recycling pathway.
A model for GLUT4 traffic and translocation is shown in Figure 12.5 based on a consensus of current data. Following early endocytosis, GLUT4 is sorted from other receptor and channel proteins in the recycling pathway and sequestered in a unique storage compartment comprised of GSVs and tubulovesicular elements. Insulin recruits GLUT4 primarily by stimulating exocytotic fusion of GSVs with the plasma membrane resulting in exofacial exposure of functional glucose channels. In basal cells, constitutive cycling of GLUT4 between intracellular and cell surface compartments has been demonstrated which may reflect a proportion of GLUT4 remaining in the endosomal recycling pathway. Insulin increases numbers of plasma membrane GLUT4 primarily by augmenting net exocytosis and also exerts a smaller effect to retard endocytosis. Stimulation of exocytosis proceeds primarily from the GSV pool with a smaller contribution from the constitutive recycling pathway.
GLUT4 vesicle trafficking is directed by several factors, including targeting motifs intrinsic to GLUT4, posttranslational modifications of GLUT4, and interactions with multiple proteins known to regulate the vectorial flow of cellular vesicles and their cargo proteins. In terms of intrinsic signals, a di-leucine motif at positions 489 and 490 in the cytosolic C-terminal of GLUT4, and an aromatic-based motif (FQQI) in the NH2 terminus, function as internalization signals for endocytosis [139,148]. While these intrinsic motifs help direct internalization, the signals for intracellular retention in the inducible storage compartment have still not been identified with certainty. Interestingly, upon mutation of the core GLUT4 glycosylation site (within the first exofacial loop at the N terminus), GLUT4 has been shown to accumulate at the cell membrane , while at the same time insulin-mediated translocation of intracellular GLUT4 was impaired , suggesting GLUT4 glycosylation plays an important role in the trafficking of the GSV. Even so, others have produced contradictory evidence showing that GLUT4 glycosylation does not affect GLUT4 vesicular trafficking [136,137,149]. Some evidence has been obtained suggesting that an acidic cluster of amino acids (TELE LGP) downstream from the di-leucine motif is important for targeting GLUT4 to the inducible storage compartment . In other experiments, overproduction of the GLUT4 C-terminus resulted in partial GLUT4 translocation and activation of glucose transport, suggesting that this peptide is interacting with an intracellular retention protein . The COOH-terminus di-leucine motif is adjacent to a phosphorylation site (serine at 488) which, when mutated, can impair internalization when expressed in host cells. GLUT4 phosphorylation occurs at several residues within the large cytosolic loop between helices 6 and 7 and also at several sites within the C-terminal cytoplasmic tail. There is a lack of consensus over whether phosphorylation of GLUT4 is increased, decreased, or unaffected upon stimulation with insulin, since evidence has been shown to support all three of these possibilities . Thus, the role of GLUT4 phosphorylation in response to insulin is unclear.
In addition to phosphorylation and N-glycosylation, GLUT4 undergoes ubiquitination (the addition of ubiquitin, a 76-amino acid protein) at residues localized throughout the intracellular loops of the protein, and this modification is important for GLUT4 to properly localize to the GSV . GLUT4 also undergoes SUMOylation, which is the addition of the 101-amino acid protein SUMO-1, although the exact location of SUMOylation on GLUT4 is unknown. SUMOylation of GLUT4 is mediated by ubiquitin conjugating enzyme 9 (Ubc9), and was shown to considerably extend the half-life of GLUT4 protein expression, suggesting that this modification is crucial for GLUT4 protein stability [154,155]. Palmitoylation, the covalent attachment of fatty acids, of GLUT4 has also been observed , although the role of this posttranslational modification is unclear.
GSV traffic is controlled by multiple protein subclasses, also known to be involved in the regulated endocytosis/exocytosis of synaptic and neurosecretory vesicles [135,139,140,143 – 148]. Multiple trafficking proteins are actual constituents of the GLUT4 vesicle; GLUT4 vesicle cargo proteins include vesicle-associated membrane protein-2 (VAMP-2) or synaptobrevin, VAMP-3/cellulobrevin, syntaxin 4, 35-39 kDa secretory carrier-associated membrane proteins (SCAMPs), cellugyrin, acyl-CoA synthetase-1, and small molecular weight GTP binding proteins in the Rab and Arf subclasses. Of these proteins, VAMP-2, VAMP-3/cellubrevin, SCAMPs, and cellugyrin are known to be constituents of GLUT4-containing vesicles in the endosomal recycling pathway where they co-localize with IGF-1 and transferrin receptors. Of these, only VAMP-2 is present together with GLUT4 in the intracellular storage compartment. Several other constituents of the GSV are an aminopeptidase, referred to as the insulin-regulated aminopeptidase (IRAP) , low-density lipoprotein receptor-related protein 1 (LRP1), and sortilin, all of which contribute to promoting GLUT4 localization to the GSV [158 – 160]. These proteins translocate with GLUT4, while IRAP in particular has commonly been used to track the cellular itinerary of GSVs which is translocated along with GLUT4 to the plasma membrane.
The v-SNARE/t-SNARE model developed by Rothman and others explains how vesicles traffic to specific membrane compartments . In this model, shown in Figure 12.6, a docking and fusion reaction is initiated via the high affinity interaction between a ligand on the transport vesicle (v-SNARE) and a receptor in the target membrane (t-SNARE). In the case of GLUT4 vesicle translocation from the inducible storage compartment to the cell surface, evidence suggests that VAMP-2 functions as the v-SNARE and syntaxin 4 as the plasma membrane t-SNARE [135,139,140,143 – 148]. Other proteins facilitate and regulate this interaction. The N-ethylmaleimide sensitive factor (NSF) activates the v-SNARE for docking competency, and three proteins, the soluble NSF attachment 23 kDa protein (SNAP23), the mouse Unc homologue 18c (Munc-18c), and the syntaxin4-interacting protein (Synip), regulate docking efficiency at the level of the t-SNARE by facilitating or blocking access to syntaxin 4 . While SNAP23, Munc18c, and Synip form a complex with syntaxin 4 at the endofacial plasma membrane, their precise roles in modulating interaction between VAMP-2 and syntaxin 4 are not fully understood. Synip appears to block access of VAMP-2 to syntaxin 4 under basal conditions, and is released from the complex following insulin stimulation allowing docking of v-SNARE (VAMP-2) to the t-SNARE (syntaxin). The function of Munc-18c appears to be more complex with evidence that it can both inhibit (basal conditions) and facilitate (insulin-stimulated condition) VAMP-2-syntaxin4 interaction. The Rab family of monomeric GTPases also play an important role in vesicle traffic; different isoforms are targeted to specific organelle membranes and direct the vectorial flow of vesicle proteins from one compartment to another [163,164]. Rab GTPases also appear to catalyze the union of v- and t-SNARES in the docking process. Rab4 has been demonstrated to be associated with GLUT4 vesicles in basal cells, but is redistributed to cytosol in response to insulin where it associates with its GDP-association inhibitor . Furthermore, Rab4 has been shown to interact with syntaxin 4 in vitro, and overexpression of Rab4 or a COOH-terminal domain peptide, prevented insulin-stimulated GLUT4 translocation in adipocytes . Undoubtedly, there are additional factors participating in GLUT4 vesicle targeting, docking, and fusion in insulin target tissues that remain undiscovered.
The cellular cytoskeleton
There is increasing recognition that GLUT4 translocation involves the actin cytoskeleton [66,167 – 169]. One line of evidence supporting this hypothesis is that disruption of the actin cytoskeleton by cytochalasin D or latrunculin also inhibits insulin-mediated GLUT4 translocation [167,168]. In addition, insulin stimulates cytoskeletal rearrangement, in particular the cortical actin fibers subtending the plasma membrane, and regulates molecules that control actin polymerization. Insulin-stimulated actin reorganization is dependent upon members of the small G-proteins of the Rho family [66,115], with the protein Rac1 active in muscle and both TC10 and Cdc42 in adipose tissue. Insulin stimulates GTP loading of Rac1, TC10, and Cdc42 followed by activation of the nucleation-promoting factor (NPF) of the Wiskott – Aldrich syndrome protein (WASP) family member, N-WASP [169,170]. The WASP proteins are responsible for the binding and activation of the Arp2/3 complex. Arp2/3 then acts as the true actin nucleator by binding to an existing actin filament to create new actin branched networks. It is attractive to hypothesize that activation of TC10 via the CAP/Cbl/TC10 pathway could mediate regulation of the actin cytoskeleton in response to insulin, since TC10 and other Rho GTPases are generally known to affect biochemical modulators of cytoskeletal function . However, PI-3 kinase can also affect cortical actin polymerization and events such as membrane ruffling and filopodia formation; therefore, other insulin signaling pathways, like Rac1, could also contribute to this biologic effect. The exact role of the actin cytoskeleton has not yet been delineated; however, one proposal is that actin filaments serve as a scaffold to direct GLUT4 vesicle trafficking using actin-based motility as a motor for vesicle movement.
The microtubule cytoskeleton is also involved in GLUT4 vesicle trafficking [140,172,173]. The GSV translocate along microtubule tracks to arrive at the inducible storage compartment proximal to the plasma membrane [174,175], and it is not surprising that microtubule proteins such as dynein and kinesin have been co-purified with GLUT4. Disruption of microtubules using depolymerizing agents inhibits GLUT4 translocation and glucose transport stimulation, and expression of a dominant-interfering dynein mutant was shown to inhibit GLUT4 endocytosis in a process that may require Rab5 . This transport has been shown to be mediated by the cargo transport motor proteins of the kinesin family  including KIF5B  and KIF3 . Upon receiving the insulin signal, tether containing UBX domain for GLUT4 (TUG) was found to disassociate from GLUT4, which has been postulated to act as the release of a “handbrake mechanism,” allowing for GLUT4 to travel to the plasma membrane . Upon arrival at the interior surface of the cell, the GSV undergoes the “docking” phase of translocation, through association with the exocyst complex, which has been proposed to function in the targeting of the GSV to areas on the plasma membrane rich in the proteins involved in fusion of the GSV with the membrane [180,181]. The GSV then undergoes the transition to the docking/fusion phase, which is mediated by the SNARE complex.
With the advent of real-time total internal reflection fluorescence microscopy (TIRFM), which visualizes just the first 200 nm of the intracellular surface of the cell membrane, it was confirmed that microtubules form a network directly underneath the entire plasma membrane, which was proposed to allow GSVs to scan the cytoplasmic surface of the cell membrane during the highly mobile basal state . TIRFM has also shown that defective actin remodeling disrupted the exocytotic fusion step of the GSV, whereas the insulin-stimulated increase of GLUT4 vesicles proximal to the plasma membrane was not affected . TIRFM performed in living 3T3-L1 adipocytes proved that insulin-stimulated fusion of GSVs occurs proximal to microtubules at the plasma membrane. This study also captured, for the first time, video evidence in living adipocytes of the substantial increase of microtubule density and curvature that occurs at the plasma membrane during insulin stimulation. The authors did find long-range movements of GSVs along microtubules, although the GSVs rarely moved long distances before fusion, suggesting that long-range GSV trafficking is not involved in insulin-stimulated GLUT4 insertion into the membrane. Nocodazole treatment, for example, did not significantly decrease the number of GSV fusion events stimulated by insulin, leading the authors to conclude that microtubules are not vital for the insertion of GSVs into the plasma membrane, but rather play a more important role in site selection for delivery of GLUT4 prior to fusion . In line with this hypothesis, CLASP2, a microtubule associated protein that targets microtubules to the interior surface of the cell cortex, was found to undergo insulin stimulated co-localization with GLUT4 at the plasma membrane. This led the authors to propose that the distal end of microtubules undergo CLASP2 targeting to landing zones on the cell cortex, thereby situating microtubule-bound GSVs proximal to the plasma membrane . An attractive hypothesis is that GLUT4 begins exocytosis in a microtubular compartment, and then transfers to actin scaffolds that connect the microtubule cytoskeleton with the plasma membrane, possibly through the protein actinin-4 , which then positions GLUT4 vesicles for docking and membrane fusion. Upon endocytosis, GLUT4 traffics in an endosomal compartment regulated by actin filaments and eventually is sorted back to the microtubular inducible storage compartment. While this model is compelling, additional research is necessary to define the role of cytoskeletal elements in GLUT4 translocation. In any case, these considerations emphasize the complex nature of the glucose transport effector system, which functionally incorporates both vesicle traffic and cytoskeletal systems.