Glucose uptake into fat and muscle cells is achieved through the highly regulated intracellular trafficking of the insulin-responsive glucose transporter isoform, GLUT4. GLUT4 is localised to endosomes and tubulo-vesicular elements (TVE) clustered either in the trans-Golgi network (TGN) or in the cytoplasm, from where GLUT4 moves to the cell surface in response to insulin. The intracellular sequestration of GLUT4 in the absence of insulin is critical to this process. The aim of this project was to identify amino acid domains within GLUT4 which regulate its intracellular trafficking itinerary in 3T3-L1 adipocytes. To best achieve our aim, we utilised a retroviral expression system that enabled expression of recombinant GLUT4 constructs at levels comparable to, or below, endogenous GLUT4 levels. We have previously reported that overexpression of GLUT4 mutants in 3T3-L1 adipocytes results in saturation of targeting machinery (Marsh
et al., 1995). In light of this, low level expression of exogenous protein is desirable to best observe bona fide targeting defects due to mutagenesis of GLUT4 targeting domains. Previous studies have identified two cytosolic targeting motifs that regulate the intracellular sequestration of this protein: FQQI5-8 in the Nterminus and LL489-490 in the C-terminus. In the present study we show that a GLUT4 chimera in which the C-terminal 12 amino acids in GLUT4 have been replaced with the same region from human GLUT3 (HA-TAIL) is constitutively targeted to the plasma membrane when expressed in 3T3-L1 adipocytes. Using the Tfn-HRP endosomal ablation technique in 3T3-L1 adipocytes, we show that mutants in which this C-terminal domain has been disrupted are more sensitive to chemical ablation than wild-type GLUT4, indicating an increased association with recycling endosomal compartments. Immuno-fluorescence and immuno-EM
techniques reveal significant colocalisation between HA-GLUT4 (an exofacially HA tagged version of GLUT4) and Syntaxins 6 and 16, two TGN t-SNAREs, implying the TGN may play an important role in the intracellular sequestration of GLUT4. Confocal immunolocalisation studies reveal that C-terminal acidic cluster mutants, namely HA-TAIL and HAEXEY (a C-terminal acidic domain mutant), of GLUT4 do not colocalise with Syntaxin6.
In order to define the kinetic movement of GLUT4 between endosomes and the Syntaxin6-positive perinuclear compartment (PNC) we established an assay to track the intemalisation of epitope tagged GLUT4 (HA-GLUT4) from the cell surface. GLUT4 rapidly traversed the endosomal system (5-15 min) en route to a perinuclear location. This PNC did not correspond to recycling endosomes but rather overlapped with a Syntaxin6-positive compartment. GLUT4 trafficking from endosomes to the Syntaxin6-positive location was mediated via an
acidic targeting motif in the C-terminus of GLUT4, as a mutant lacking this motif (HA-EXEY) was retained in endosomes. Syntaxin6, together with the t-SNARE Syntaxinl6, has previously been shown to regulate traffic between endosomes and the TGN (Mallard et al, 2002). To determine whether GLUT4 follows a similar trafficking pathway, we studied resialylation of GLUT4 together with IRAP, a protein that displays almost identical trafficking properties to GLUT4. Following digestion of cells with neuraminidase we observed a significant decrease in the molecular weight of cell surface GLUT4 and IRAP, which was rapidly reversed (<60min) following incubation of cells at 37°C. The following are consistent with a role for Syntaxins 6 and 16 in GLUT4 trafficking: (1) Syntaxins 6 and 16 form a stable complex in adipocytes; (2) insulin stimulates their translocation to the cell surface; and (3) their expression is significantly increased during adipocyte differentiation. These data
suggest that the TGN together with the t-SNAREs Syntaxin6 and 16 may be involved in the biogenesis of insulin responsive GLUT4 vesicles in adipocytes.
We know that GLUT4 constitutively recycles between the cell surface and its intracellular storage compartment (Yang et al, 1992), which implies an important role for efficient internalisation in maintaining the sequestration of GLUT4 under non-stimulated conditions. GLUT4 is efficiently internalisated from the cell surface via a Tyrosine-based motif in the N-terminus (FQQI) and a dileucine motif in the C-terminus. We show that, contrary to previous reports, the Phenylalanine5 mutant of GLUT4 does not mistarget to the cell surface in basal adipocytes. In fact, we show by both subcellular fractionation and steady state immunofluorescence that at low levels of expression, this protein is targeted similarly to HA-GLUT4 in steady state adipocytes. The Phenylalanine-based
motif has been implicated in entry of GLUT4 into the GLUT4 storage compartment and it has been reported that mutagenesis of this motif results in increased GLUT4 degradation (Palacios et al., 2000). Using the retroviral expression system to express HA-F5A in adipocytes we found this protein to be both insulin-responsive and retrieved to the Syntaxin6-positive PNC. We suggest that the dominant role for the Phenylalanine-based motif of GLUT4 is related to endocytosis in insulin stimulated cells, as this mutant is significantly delayed at the cell surface under these conditions. In our hands Phenylalanine5 of GLUT4 has little involvement in inter-endosomal sorting events. In support of our previously published data, at low levels of expression, the diLeucine mutant of GLUT4 targets similarly to endogenous GLUT4. While we have not measured the exact kinetics of HA-LL endocytosis, our experiments reveal significant intracellular labelling at 5min
time points, similar to HA-GLUT4. While we cannot support a major role for the diLeucine motif in endocytosis of GLUT4 using this system, we did observe delayed endosome to TGN sorting as a result of this mutation. We suggest that the diLeucine motif plays a role in clustering GLUT4 in specialised domains of the early endosome, perhaps marked by a specialised coat, and subsequent sorting to a TGN sub-compartment enriched in Syntaxins 6 and 16.
Sandoval et al (2000) recently reported the involvement of basic residues, upstream of the diLeucine motif, as essential for efficient intracellular sorting and endocytosis of GLUT4. On the basis of homology between proteins that are known to traffic similarly to GLUT4 and the recent report from Sandoval et al (2000), we have proposed the existence of the "BaLLAc" motif The BaLLAc motif is comprised of a diLeucine (LL) motif flanked by basic residue/s (Ba)
N-terminal to the diLeucine motif and an acidic cluster of amino acids (Ac) C-terminal to the diLeucine motif We hypothesise that this motif regulates the sorting of this 'family' of proteins into the Syntaxin6/16 pathway from endosomes. In sharp contrast to this recent report (Sandoval et al, 2000), both fractionation analysis and immunofluorescence microscopy studies did not reveal a significant role for the diArginine motif in GLUT4 targeting. While the RR/LL double motif mutant of GLUT4 displayed delayed endocytosis at early time points of endocytosis, further incubation revealed significant accumulation of surface labelled recombinant transporter in the Syntaxin6/16- positive perinuclear compartment. This implies a supporting role for the diArginine motif of GLUT4 in targeting events, as it appears to cooperate with the diLeucine motif in regulating GLUT4 endocytosis and endosomal retention. In fact, the only BaLLAc motif mutations found to have a profound
effect on endosome to TGN trafficking of GLUT4 were those which incorporated the acidic motif We suggest that transient phosphorylation at Serine488 may modulate interaction of the BaLLAc motif with intracellular trafficking machinery. This may mediate successive sorting events of GLUT4 with adaptor protein complexes, which ultimately regulates the intracellular trafficking of GLUT4. We propose that the entry of GLUT4 into a cycle between endosomes and the TGN effectively excludes GLUT4 from the cell surface and that entry into this cycle facilitates the unique insulin-responsiveness of GLUT4. It is tantalising to think that the Syntaxin6/16 compartment described throughout these studies may in fact represent, or be a precursor to, the insulin-responsive compartment in 3T3-L1 adipocytes.