Cholesterol Rich Domains and Lipid Rafts

Originally, the cell membrane was conceptualized as an isotropic lipid bilayer, which allowed for free diffusion of all constituents therein. This model was metaphorically referred to as a fluid mosaic of freely and randomly diffusing proteins embedded in phospholipids [82]. Obviously, this model ignores the compartmentalization induced by the cortical actin (Fig. 4.3), which as mentioned above, may also affect the distribution of lipids [78]. In addition, it does not account for the fact that lipids are not equally distributed over the various cell membrane compartments, although these compartments all communicate with each other. Cholesterol, for instance, is enriched in the plasma membrane and in endosomal vesicles, but it is synthesized in the endoplasmic reticulum. It is also clear that the outer and inner leaflet of the membrane do not have the same lipid composition: for example, sphingolipids are enriched in the outer leaflet. These gradients must be actively maintained; this fact is generally best appreciated for phosphatidylserine, which is conspicuously absent from the outer leaflet unless cells undergo apoptosis. Thus, the asymmetric distribution of lipids is actively maintained. In addition, lipids have been proposed to segregate and self-organize based on their different miscibility. Because the saturated lipids of sphingomyelins are more prone to associate with cholesterol, membrane areas may arise that are enriched in these lipids and the platforms—termed lipid rafts—were proposed to attract specific proteins [83]. Cholesterol-rich domains may be further stabilized by the cytoskeleton (see above) and by caveolin [84]. These platforms (rafts and caveolae) have attracted much interest in the field of signal transduction, because many key signaling molecules are modified with lipid moieties (e.g., farnesyl or geranyl-geranoyl thioethers on the C-termini of RAS-like small G proteins or myris-toyl amide bonds on the N-termini of nonreceptor tyrosine kinases of the SRC family).

Because many GPCRs are palmitoylated (on a cysteine residue at the end of helix 8 in the C- t erminus) and several G protein a subunits are myris-toylated and palmitoylated, they were proposed to be attracted into lipid rafts. Accordingly, there is a long list of reports, which examined the role of lipid rafts and cholesterol-rich domains in signaling by GPCRs: most reports emphasize a model, where G protein activation preferentially takes place in lipid rafts/cholesterol-rich domains (for review, see References 85, 86). It is not clear, though, why this should be the case; the original reconstitution experiments (of purified receptors and purified G proteins) employed lipid mixtures that were devoid of cholesterol and found robust G protein activation [87] and transfer of the signal to the effector [ 88] . Similarly, when expressed in Escherichia coli and reconstituted with G proteins, GPCRs efficiently interact with their cognate G proteins [89-91], although the inner membrane of gramnegative bacteria is devoid of cholesterol.

Lipid rafts are thought to be too small to directly visualize them; their proposed size varies between 5 and 100 nm and is thus below the limit of optical resolution imposed by Abbe's limit. Similarly, caveolae are not present in all cells; the phenotype of mice genetically deficient in caveolin is very mild (for review, see Reference 84). Thus, the evidence for a role of these putative lipid rafts is only circumstantial. There have been two popular approaches used to examine the role of lipid rafts: (1) the (re)distribution of receptors and G proteins into detergent-resistant membranes and (2) the sensitivity of signaling events to cholesterol extraction. For (1), sucrose density gradient fraction-ation of cell membranes provide a means of resolving two surface membrane fractions. The lipid enriched light fraction is thought to contain the lipid rafts. Similarly, a proportion of membrane proteins cannot be extracted by Triton X-100, and this detergent-resistant fraction again is thought to represent lipid rafts. With (2), cells are pretreated with the polyene antibiotic filipin 3 and the caging compound methyl-P-cyclodextrin, which sequesters and depletes cholesterol, respectively, and thus disrupt lipid rafts/cholesterol-rich domains. The rationale underlying these approaches have been questioned [ 92] : typically cells are homogenized in ice-cold buffers and density gradient centrifugation is done at low temperature, which per se has a strong effect on the miscibility of lipids and thus on the diffusion of receptors [77]. Similarly, addition of Triton X-100 may per se affect the miscibility of lipids and thus induce segregation of lipids, which was initially not present [93]. Proteins may also be resistant to detergent extraction, because they are tightly bound to the cytoskeleton.

Extraction of cholesterol may have effects other than disrupting lipid rafts. In fact, like all other membrane proteins, GPCRs affect the shape of the membrane, because for any individual membrane, protein solvation of the hydro-phobic transmembrane segment may be more readily accomplished in one type of lipid. This will result in selective attraction of lipids to the anular protein lipid boundary [94, 95]. Cholesterol enhances the stability of rhodopsin but reduces the efficacy of G protein activation [96, 97] : Cholesterol may interfere with rhodopsin activation largely by steric hindrance: the protein expansion that is caused by transition of metarhodopsin I (MI) into the active form MII is impeded by dense packing of cholesterol, which renders the phos-pholipid bilayer more rigid. Rod outer segments are heterogeneous with respect to their cholesterol content. Disc membranes in the rod outer segments are continuously replenished from evaginations of the plasma membrane at the base of the outer segment. Within approximately 10 days, the disc membranes reach the top of the outer segment where they are engulfed by phagocytosis by the surrounding pigment epithelium. Whereas the ratio of protein to phospholipd remains constant during migration of the disc to the apical tip, cholesterol content falls. In newly formed disc membranes at the base of the outer segment, cholesterol accounts for 30% of the lipid mass; mature discs residing on the tip of the rod outer segment are depleted in cholesterol (with a ratio of phospholipid to cholesterol = 20:1; [64]). From a teleological perspective, this change in cholesterol is useful; a high level of cholesterol may not only stabilize rhodopsin, but it may also prevent its premature activation: photons may also activate those rhodopsin molecules that are still en route to the rod outer segment [98]. These effects of cholesterol do neither require the formation of specific cholesterol-enriched domains nor the presence of a cholesterol binding site on rhodopsin. However, an electron density compatible with cholesterol has been found to be trapped in both crystals of rhodopsin [99] and of the P2-adrenergic receptor [100], raising the possibility that cholesterol may also regulate receptor function by directly binding to the hydropho-bic core.

If GPCRs other than rhodopsin are examined, depletion of cholesterol does not uniformly affect signaling; in some instances, G protein activation is enhanced, while in others, it is abrogated (reviewed in Reference 85). In addition, in at least one receptor, the presence or absence of cholesterol, may affect signaling in a more subtle way than all-or-none: cholesterol depletion abrogates the ability of the A2A-adenosine receptor to activate Gs (the cognate G protein) and hence cAMP accumulation, but it does not interfere with G protein-independent recruitment of ARNO (the exchange factor for the small G protein ARF6) and the resulting stimulation of mitogen activated protein (MAP) kinase [ 68]. Because many GPCRs recruit more than one signaling cascade, it is likely that analogous effects will also be observed with other receptors. In the lipid raft model, one would be inclined to ascribe the differential effects of cholesterol extraction on the A2 A adenosine receptor to the segregation of signaling components, with G protein-dependent signaling contingent on cholesterol-rich domains and ARNO recruitment being independent of membrane compartmentalization. However, at the current stage, it is not possible to determine whether cholesterol extraction alters receptor-dependent G protein activation through disruption of lipid rafts or other cholesterol-rich domains by changing the ability of the lipid bilayer to accommodate conformational changes in an activated GPCR, or by removing cholesterol that is tightly bound to the GPCR.

This section has focused on factors extrinsic to the G protein cycle that may allow for organizing signaling components. However, it should be kept in mind that the reaction kinetics of the G protein cycle allow for self-organization to emerge without any need for additional factors: if a Gq - coupled receptor is allowed to interact with a G protein that is subject to rapid deactivation by the GAP activity residing in an RGS protein and in the phospholipase CP (the Gq-regulated) effector, a stable tightly associated complex emerges, which allows for very rapid cycling [101] .

0 0

Post a comment