br Materials and methods br Results br

Materials and methods
Results
Discussion The glucose transporter GLUT1 can be acutely activated or inhibited in L929 fibroblast Aminophylline receptor by a wide variety of reagents or nutrient conditions [[9], [10], [11], [12],36,50]. This abrupt change in the activity of GLUT1 occurs within minutes and without a detectable change the concentration of GLUT1 in the cell membrane. In some cases of inhibition, the evidence suggests a direct interaction of the inhibitor with GLUT1 [36,50]. However, in most other cases of activation or inhibition the cellular mechanisms are not well understood. Previous work has suggested that the lipid environment of GLUT1 may play an important role in the transport activity of GLUT1 [[33], [34], [35]]. Kumar et al. found that activation of GLUT1 by glucose deprivation enhances targeting of GLUT1 to lipid rafts in 3T3-L1 adipocytes [33]. However, work by Rubin and Ismail-Beigi demonstrated that activation of GLUT1 by azide in Clone 9  cells and 3T3-L1 fibroblasts results in a redistribution of up to 50% of the GLUT1 out of the detergent resistant membrane fraction [34]. These later data were supported by Barnes et al. who reported a 3.5-fold increase in GLUT1 activity in Clone 9  cells following cholesterol depletion by MβCD treatment [35]. It became apparent early in the study that the GLUT1-containing low density domains in L929 cells are not the traditionally described lipid rafts [22,26]. First, the lipid raft marker CD44 does not isolate with the low density GLUT1 domains in L929 fibroblast cells, while GLUT1 and CD44 do isolate together in low density domains in HK2 cells. Second, Triton X-100, which is the traditional detergent used to isolate lipid rafts, completely collapses GLUT1 to high density fractions while maintaining caveolin-1 in the low density domains. This suggests that GLUT1 are in planar rafts rather than caveolae domains. Third, membrane cholesterol depletion by MβCD treatment does not affect the GLUT1 low density domains suggesting that cholesterol is not an important component of these domains. This conflicts with an earlier study in Clone 9  cells which did demonstrate a significant reduction of GLUT1 in the lipid raft fraction after treatment with MβCD [35]. It is likely that there are significant differences in the lipid environment for GLUT1 depending on cell type. We were unable to determine specific factors that either target or stabilize GLUT1 to these low density domains. It has been suggested that post-translational modifications, such as palmitoylation and glycosylation, are important in targeting proteins to lipid rafts [51,52] and cytoskeletal components are important to the stabilization of membrane regions [53]. However, neither inhibition of palmitoylation by 2-bromopalmitate nor inhibition of glycosylation by kifunensine altered the isolation of GLUT1 to low density domains. In addition, the disruption of either the actin cytoskeleton by Latruculin A or microtubules by Nocodazole had no effect on GLUT1 containing low density domains. These negative results, along with the other results from this study are summarized in Table 1. We were able, however, to demonstrate that both GLUT1-containing lipid microdomains and GLUT1 basal activity are somewhat dependent on sphingolipids. Our initial evidence for this was the observation that the sterol-based lipid, digitonin, partially maintains GLUT1 in low density domains, while the acyl chain detergents such as Triton X-100 and dodecyl-β-D-maltoside completely collapse GLUT1 to high density domains. Sterol-based lipids have been shown to create less disorder, maintain protein function and preferentially extract sphingolipid rich domains than the acyl-chain based detergents [48,54,55]. More conclusively, both inhibition of sphingolipid synthesis by myriocin and degradation by sphingomyelinase produced a shift in the density of the GLUT1 microdomains and decreased the basal activity of GLUT1. These results suggest that the GLUT1-containing low density domains are enriched in sphingolipids. This is consistent with recently described sphingolipid enriched class of low density domains [49] and match the model for the cell membranes of fibroblast cells described by Frisz et al. in which cholesterol is evenly distributed throughout the membrane and sphingolipids are variably clustered [46,47].