Recent studies showing that detergent-resistant membrane fragments can be isolated from cells suggest that biological membranes are not always in a liquid-crystalline phase. Instead, sphingolipid and cholesterol-rich membranes such as plasma membranes appear to exist, at least partially, in the liquid-ordered phase or a phase with similar properties. Sphingolipid and cholesterol-rich domains may exist as phase-separated "rafts" in the membrane. We discuss the relationship between detergent-resistant membranes, rafts, caveolae, and low-density plasma membrane fragments. We also discuss possible functions of lipid rafts in membranes. Signal transduction through the high-affinity receptor for IgE on basophils, and possibly through related receptors on other hematopoietic cells, appears to be enhanced by association with rafts. Raft association may also aid in signaling through proteins anchored by glycosylphosphatidylinositol, particularly in hematopoietic cells and neurons. Rafts may also function in sorting and trafficking through the secretory and endocytic pathways.

It is well known that separate domains with different lipid compositions can exist in liposomes containing mixtures of different phospholipids. The question of whether cellular membranes contain similar lipid domains has intrigued workers for many years. One type of domain, sphingolipid and cholesterol-based structures called membrane rafts, has received much attention in the last few years. We will review the evidence that rafts exist in cells and focus on their structure, or the organization of raft lipids and proteins. Our discussion of function will focus on the role of rafts in signaling in hematopoietic cells, a particularly well developed area that has provided insights into raft organization in the membrane. Several reviews of rafts (1-4) and of related structures called caveolae (5-7) have appeared recently. Lipid Phase Behavior and Raft FormationSphingolipids differ from most biological phospholipids in containing long, largely saturated acyl chains. This allows them to readily pack tightly together, a property that gives sphingolipids much higher melting temperatures (T m ) 1 than membrane (glycero)phospholipids, which are rich in kinked unsaturated acyl chains. It is now clear that tight acyl chain packing is a key feature of raft lipid organization (3,8,9). In fact, the differential packing ability of sphingolipids and phospholipids probably leads to phase separation in the membrane. Thus, sphingolipid-rich rafts co-exist with phospholipid-rich domains that are in the familiar, loosely packed disordered state (variously abbreviated as L␣, l c , or l d ). Phase separation between lipids in different physical states, most often the l c and the solid-like gel phases, has been well characterized in model membranes. Indeed, the gel phase is the most familiar state in which acyl chains are highly ordered.However, because of the high concentration of cholesterol in the plasma membrane and other membranes in which rafts form, raft lipids do not exist in the gel phase. Cholesterol has important effects on phase behavior. It is well known that addition of cholesterol to a pure phospholipid bilayer abolishes the normal sharp thermal transition between gel and l c phases, giving the membrane properties intermediate between the two phases. This effect initially suggested that domains in ordered and disordered states cannot co-exist at high cholesterol levels. However, further work showed that a different kind of phase separation can occur in binary mixtures of individual phospholipids with cholesterol. In these mixtures, domains in an l c -like phase co-exist with domains in a new state, the liquid-ordered (l o ) phase. Acyl chains of lipids in the l o phase are extended and tightly packed, as in the gel phase, but have a high degree of lateral mobility (3).Rafts probably exist in the l o phase or a state with similar properties. In support of this model, detergent-insoluble membranes that can be isolated from cell lysates and are likely to be derived from rafts (discussed below) are in the l o phase (10, 11). M...

No abstract

This report describes a method suitable for determining the depth of a wide variety of fluorescent molecules embedded in membranes. The method involves determination of the parallax in the apparent location of fluorophores detected when quenching by phospholipids spin-labeled at two different depths is compared. By use of straightforward algebraic expressions, the method allows calculation of depth in angstroms. Furthermore, the analysis can be extended to quenching by energy-transfer acceptors or brominated probes under appropriate conditions. Application of the method to quenching of 7-nitro-2,1,3-benzoxadiazol-4-yl (NBD)-labeled lipids by spin-labeled lipids located at three different depths is demonstrated in model membranes. It is shown that the calculated depths of the NBD groups are self-consistent to the extent that they are the same no matter which two spin-labels have been used in a particular experiment. In addition, the calculated depth is independent of spin-label concentration in the membrane within +/- 1 A, ruling out major effects due to spin-label perturbation. The quenching experiments show that the location of the NBD group in head-group-labeled phosphatidylethanolamine is at the polar/hydrocarbon interface and that of an NBD label on the "tail" of cholesterol is deeply buried, as expected. Unexpectedly, NBD labels placed at the end of fatty acyl chains of phosphatidylcholines are also near the polar/hydrocarbon interface. Presumably, the polarity of the NBD group results in "looping" back to the surface of the NBD groups attached to flexible acyl chains.

Detergent-insoluble membrane fragments that are rich in sphingolipid and cholesterol can be isolated from both cell lysates and model membranes. We have proposed that these arise from membranes that are in the liquid-ordered phase both in vivo and in vitro [Schroeder et al. (1994) Proc. Natl. Acad. Sci. U.S.A. 91, 12130-12134]. In order to detect formation of the liquid-ordered phase while avoiding possible detergent artifacts, we have now used fluorescence quenching to examine the phase behavior of mixtures of phosphatidylcholines, sphingolipids, and cholesterol. Phase separation was found in binary mixtures of either dipalmitoylphosphatidylcholine (DPPC) or sphingomyelin (SM) and a nitroxide-labeled phosphatidylcholine (12SLPC). A DPPC- or SM-enriched solidlike gel phase coexisted with a 12SLPC-enriched liquid-disordered fluid phase at 23 degrees C. As expected, phase separation was not seen at low concentrations of DPPC or SM. Instead, only a uniform fluid phase was present. Including 33 mol % cholesterol in model membranes greatly promoted phase separation. Phase separation was seen at higher temperatures and/or at lower concentrations of DPPC or SM in the presence of cholesterol than in its absence. Mixtures of DPPC or SM and cholesterol are known to form the liquid-ordered phase. Therefore, the fact that phase separation was observed in the cholesterol-containing membranes shows that liquid-ordered and liquid-disordered phase domains coexist. At 37 degrees C, the SM-enriched liquid-ordered phase was first seen at a SM/PC ratio of close to 0.25, when SM made up 17% of the total lipid including cholesterol. (This is similar to or less than the SM concentration of the plasma membranes of mammalian cells.) Furthermore, the detergent insolubility of cholesterol-containing model membranes correlated well with the amount of liquid-ordered phase as detected by fluorescence quenching. Thus, the detergent-insoluble membranes isolated from cells are likely to exist in the liquid-ordered phase prior to detergent extraction. The promotion of liquid-ordered phase formation may be an important function of cholesterol and sphingolipids in cells and may be a major distinction between the cholesterol- and sphingolipid-rich plasma membrane and most other cellular membranes.

Proteins anchored by GPI are poorly solubilized from cell membranes by cold nonionic detergents because they associate with detergent-resistant membranes rich in cholesterol and sphingolipids. In this study, we demonstrated that cholesterol and sphingolipid-rich liposomes were incompletely solubilized by Triton X-100. GPI-anchored placental alkaline phosphatase incorporated in these liposomes was also not solubilized by cold Triton X-100. As sphingolipids have much higher melting temperatures (Tm) than cellular phospholipids, a property correlated with Tm might cause detergent inextractability. In support of this idea, we found that the low-T. lipid dioleoyl phosphatidylcholine (DOPC) was efficiently extracted from detergent-resistant liposomes by Triton X-100, whereas the high-Tm lipid dipalmitoyl phosphatidylcholine (DPPC) was not. The fluorescence polarization of liposome-incorporated diphenylhexatriene was measured to determine the "fluidity" of the detergent-resistant liposomes. We found that these liposomes were about as fluid as DPPC/cholesterol liposomes, which were present in the liquid-ordered phase, and much less fluid than DOPC or DOPC/cholesterol lposomes. These findings may explain the behavior of GPIanchored proteins, which often have saturated fatty acyl chains and should prefer a less-fluid membrane. Therefore, we propose that acyl chain interactions can influence the association of GPI-anchored proteins with detergent-resistant membrane lipids. The affinity of GPI-anchored proteins for a sphingolipid-rich membrane phase that is not in the liquid crystalline state may be important in determining their cellular localization.Membranes that contain GPI-anchored proteins have been isolated from cultured epithelial cells (1, 2), fibroblasts (2), and lymphocytes (3) on the basis of their surprising insolubility in nonionic detergents such as Triton X-100. These membranes are rich in sphingolipids and cholesterol (1) and contain several proteins in addition to those anchored by GPI (2,4,5). These include caveolin, a marker for 50-nm-diameter non-clathrin-coated plasma membrane invaginations called caveolae (6). This and other studies have led to the suggestion that caveolae can be isolated from cells by detergent extraction (7, 8).It is not known how the lipids and proteins in these membranes allow them to remain associated with each other in the presence of detergent. Bilayers consisting of phospholipids and glycosphingolipids contain compositionally distinct domains (reviewed in ref. 9). It has been suggested that the same interactions that cause glycolipids to form clusters may cause their detergent resistance (1, 10). There is evidence to suggest that hydrogen bonds between glycosphingolipids may help form membrane domains (11). Alternatively, or in addition, acyl chain interactions may be important in this process, as glycolipid-rich domains can be present in a gel-like state (9). We have tested these models by determining the detergent solubility of artificial liposomes, with or without ...

The insolubility of lipids in detergents is a useful method for probing the structure of biological membranes. Insolubility in detergents like Triton X-100 is observed in lipid bilayers that exist in physical states in which lipid packing is tight. The Triton X-100-insoluble lipid fraction obtained after detergent extraction of eukaryotic cells is composed of detergent-insoluble membranes rich in sphingolipids and cholesterol. These insoluble membranes appear to arise from sphingolipid- and cholesterol-rich membrane domains (rafts) in the tightly packed liquid ordered state. Because the degree of lipid insolubility depends on the stability of lipid-lipid interactions relative to lipid-detergent interactions, the quantitative relationship between rafts and detergent-insoluble membranes is complex, and can depend on lipid composition, detergent and temperature. Nevertheless, when used conservatively detergent insolubility is an invaluable tool for studying cellular rafts and characterizing their composition.

Detergent-insoluble membrane domains, enriched in saturated lipids and cholesterol, have been implicated in numerous biological functions. To understand how cholesterol promotes domain formation, the effect of various sterols and sterol derivatives on domain formation in mixtures of the saturated lipid dipalmitoylphosphatidylcholine (DPPC) and a fluorescence quenching analogue of an unsaturated lipid was compared. Quenching measurements demonstrated that several sterols (cholesterol, dihydrocholesterol, epicholesterol, and 25-hydroxycholesterol) promote formation of DPPC-enriched domains. Other sterols and sterol derivatives had little effect on domain formation (cholestane and lanosterol) or, surprisingly, strongly inhibit it (coprostanol, androstenol, cholesterol sulfate, and 4-cholestenone). The effect of sterols on domain formation was closely correlated with their effects on DPPC insolubility. Those sterols that promoted domain formation increased DPPC insolubility, whereas those sterols that inhibit domain formation decreased DPPC insolubility. The effects of sterols on the fluorescence polarization of diphenylhexatriene incorporated into DPPC-containing vesicles were also correlated with sterol structure. These experiments indicate that the effect of sterol on the ability of saturated lipids to form a tightly packed (i.e., tight in the sense that the lipids are closely packed with one another) and ordered state is the key to their effect on domain formation. Those sterols that promote tight packing of saturated lipids promote domain formation, while those sterols that inhibited tight packing of saturated lipids inhibited domain formation. The ability of some sterols to inhibit domain formation (i.e., act as "anti-cholesterols") should be a valuable tool for examining domain formation and properties in cells.