Detergent-insoluble glycosphingolipid/cholesterol-rich membrane domains,lipid rafts and caveolae (Review)

Nigel M. Hooper

School of Biochemistry and Molecular Biology, University ofLeeds, Leeds LS2 9JT, UK. e-mail: n.m.hooper@leeds.ac.uk

SummaryWithin the cell membrane glycosphingolipids and cholesterolcluster together in distinct domains or lipid rafts, along withglycosyl-phosphatidylinosito l (GPI)-anchored proteins in theouter leaflet and acylated proteins in the inner leaflet of thebilayer. These lipid rafts are characterized by insolubility indetergents such as Triton X-100 at 4 8 C. Studies on modelmembrane systems have shown that the clustering of glyco-sphingolipids and GPI-anchored proteins in lipid rafts is anintrinsic property of the acyl chains of these membranecomponents, and that detergent extraction does not artefactuallyinduce clustering. Cholesterol is not required for clustering inmodel membranes but does enhance this process. Singleparticle tracking, chemical cross-linking, fluorescence reso-nance energy transfer and immunofluorescence microscopyhave been used to directly visualize lipid rafts in membranes. Thesizes of the rafts observed in these studies range from 70 ±370 nm, and depletion of cellular cholesterol levels disrupts therafts. Caveolae, flask-shaped invaginations of the plasmamembrane, that contain the coat protein caveolin, are alsoenriched in cholesterol and glycosphingolipids . Although ca-veolae are also insoluble in Triton X-100, more selective isolationprocedures indicate that caveolae do not equate with detergent-insoluble lipid rafts. Numerous proteins involved in cell signal-ling have been identified in caveolae, suggesting that thesestructures may function as signal transduction centres. Deple-tion of membrane cholesterol with cholesterol binding drugs orby blocking cellular cholesterol biosynthesis disrupts theformation and function of both lipid rafts and caveolae,indicating that these membrane domains are involved in a rangeof biological processes.

Keywords: caveolae, cholesterol, detergent-insolubility, glyco-sphingolipids, membrane domains.

Abbreviations: CHO, Chinese Hamster Ovary; DIG, detergent-insoluble glycolipid-enriched domain; FRET, fluorescence resonanceenergy transfer; GPI, glycosyl-phosphatidylinositol; GSL, glyco-sphingolipid; MDCK, Madin-Darby canine kidney; NCAM, neural celladhesion molecule.

Introduction

In polarized cells, the plasma membrane lipids and proteinsare organized into distinct domains, for example, the apicaland basolateral domains of epithelial cells, the axon and cellbody of neurons, the sinusoidal and canalicular domains ofhepatocytes, and the head and tail of sperm (Keller andSimons 1997). These compositionally distinct plasma mem-brane domains are easily distinguishable morphologically.However, evidence is now accumulating that even inundifferentiated cells, lipids and proteins are organized indistinct domains within the plasma membrane. In particular,glycosphingolipids (GSLs), cholesterol and acylated proteins

appear to cluster together in distinct membrane domains orlipid rafts. Caveolae, flask-shaped invaginations of theplasma membrane, are also rich in cholesterol and GSLs,and have been implicated in cell signalling and the uptake ofsmall molecules. This article reviews the background to theexistence of lipid rafts, highlighting the biophysical andbiochemical studies that have been used to study suchstructures both in vitro and in vivo. The relationship betweenlipid rafts and caveolae is addressed, and the role ofcholesterol in the formation and function of these structuresdiscussed. In addition, the reliability of the criterion ofinsolubility in certain detergents as a means of isolating lipidrafts and identifying membrane components that are presentin them in vivo is assessed. Other recent reviews in this areainclude Brown and London (1998a, b), Rietveld and Simons(1998) and Zegers and Hoekstra (1998).

Lipid rafts

The apical membrane of polarized epithelial cells, such asMadin-Darby canine kidney (MDCK) cells, is enriched inGSLs and sphingomyelin. Studies on the delivery of newlysynthesized sphingolipids in MDCK cells revealed that theGSL, glucosylceramide, was preferentially transported to theapical membrane (Simons and van Meer 1988) (althoughlater studies have suggested that this may be artefactual,discussed in Weimbs et al. (1997)). To explain this directeddelivery to the apical domain Simons and Wadinger-Ness(1990) proposed that the GSLs cluster together within theexoplasmic leaflet of the Golgi membrane, from where theyare targeted, along with other lipids and proteins, to theplasma membrane. In support of this model, Brown andRose (1992) observed that the glycosyl-phosphatidylinositol(GPI) anchored protein alkaline phosphatase could berecovered from lysates of MDCK cells in a low density,detergent-insoluble fraction. This low density, detergent-insoluble fraction was enriched in GSLs and cholesterol.Pulse-chase labelling revealed that the detergent-insolubilityof alkaline phosphatase only appeared after transport to theGolgi complex, consistent with its association with GSLswithin this organelle.

Simons and colleagues have put forward a model for theorganization of GSL- and cholesterol-rich lipid rafts (figure 1)(Harder and Simons 1997, Simons and Ikonen 1997). Thesphingolipids (GSLs and sphingomyelin) in the outer leaflet ofthe mammalian cell membrane associate laterally with oneanother, probably through weak interactions between thecarbohydrate heads of the GSLs. The headgroups of thesphingolipids occupy larger areas in the plane of the outerleaflet of the lipid bilayer than their predominantly long andsaturatedacylchains, withthe spaces betweenthe acylchainsbeing filled by cholesterol molecules (figure 1) (Brown 1998).These close-packed GSL-cholesterol clusters behave asdomains within the exoplasmic leaflet of the bilayer, with the

Molecular Membrane Biology, 1999, 16, 145 ± 156

Molecular Membrane Biology ISSN 0968-7688 print/ISSN 1464-5203 online Ó 1999 Taylor & Francis Ltdhttp://www.tandf.co.uk /JNLS/mbc.htm

http://www.taylorandfrancis.com/JNLS/mbc.htm

Mol

Mem

br B

iol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Reg

ina

on 0

4/16

/13

For

pers

onal

use

onl

y.

regions between such domains occupied predominantly byunsaturated phospholipids. The nature of the phospholipidsoccupying the cytoplasmic leaflet of these GSL-cholesterol-rich domains is as yet unknown, but they may also be carryingmainlysaturated fatty acid chains in order to optimize packing.

Detergent-insoluble membrane domains

The use by Brown and Rose (1992) of insolubility in thedetergent Triton X-100, as the basis for isolating GSL-cholesterol-rich lipid rafts, was based on reports that bothGSLs (Hagman and Fishman 1982) and multiple GPI-anchored proteins (Hooper et al. 1987, Hooper and Turner1988a, b, Low 1989) were essentially insoluble in certainnon-ionic detergents such as Triton X-100. Previous to theseobservations, insolubility of a protein in Triton X-100 wasoften equated with association with the detergent-insolublecytoskeleton (Streuli et al. 1981, Kim and Campbell 1983,Davies et al. 1984, Hoessli and Rungger-Brandle 1985).From the seminal study of Brown and Rose (1992), acommon means of identifying proteins and lipids thatassociate with lipid rafts was developed in which thedetergent-insolubility at 4 8 C of the membrane componentof interest is examined. Because of their high lipid content,these detergent-insoluble complexes float to a low densityduring sucrose density gradient centrifugation (figure 2)(Brown and Rose 1992, Sargiacomo et al. 1993, Chang et al.1994, Lisanti et al. 1994b, 1995). The resulting low-density,detergent-insoluble membrane fraction is enriched not only in

cholesterol and GSLs but also in multiple GPI-anchoredproteins, which are localized to the outer leaflet of the bilayer(Chang et al. 1994, Lisanti et al. 1994b) (see figure 3), certaintransmembrane polypeptide anchored proteins such as thehaemagglutinin protein of influenza virus (Skibbens et al.1989) and intestinal epithelial sucrase-isomaltase (Danielsen1995), and numerous acylated (myristoylated and/or palmi-toylated) proteins which are localized to the cytoplasmicleaflet, such as the Src family tyrosine kinases (Lisanti et al.1994b), nitric oxide synthase (Shaul et al. 1996) andheterotrimeric G proteins (Chang et al. 1994, Lisanti et al.1994b) (figure 1).

Several names have been given to this low density,detergent-insoluble membrane fraction, including DIGs (de-tergent-insoluble glycolipid-enriched domain, which will beused here; Parton and Simons 1995), CHIFF (CHAPS-insoluble floating fraction; Xiao and Devreotes 1997), DIMs(detergent-insoluble membranes; Parkin et al. 1999), DRMs(detergent-resistant membranes; Melkonian et al. 1995),GEMs (glycolipid-enriched domains; Rodgers et al. 1994),LDTI (low-density Triton-insoluble fraction; Parolini et al.1996), TIFF (Triton-insoluble floating fraction; Kurzchalia etal. 1995) and TIM (Triton-insoluble membranes; Garcia-Cardena et al. 1996).

Detergent-extractionÐ artefact or not?

The technique of detergent-extraction to isolate lipid rafts hasgenerated much controversy. Do such domains, rich in

Figure 1. Model for the organization of lipid rafts and caveolae in the plasma membrane. The cholesterol-GSL-rich rafts segregate from the otherregions of the bilayer and possibly form an annulus around the neck of the caveola. Individual lipids and proteins may move between thecholesterol-GSL-rich domains and other regions of the bilayer, including the caveola. The lipid bilayer in cholesterol-GSL-rich domains isasymmetric, with sphingomyelin and GSLs enriched in the outer leaflet. Cholesterol is present in both leaflets and fills the spaces under theheadgroups of the GSLs. GPI-anchored proteins (alkaline phosphatase, Thy-1, etc) are attached to the outer leaflet of the bilayer, and acylatedproteins (Src family kinases, G proteins, etc.) are integrated into the cytoplasmic leaflet. Some polypeptide-anchored proteins (e.g. influenzavirus haemagglutinin and intestinal sucrase-isomaltase) are also associated with the cholesterol-GSL-rich domains. The flask-shapedinvagination of the plasma membrane or caveola is also enriched in GSLs and cholesterol and contains the coat protein caveolin/VIP21 whichforms a hairpin structure in the bilayer. Modified from Parton and Simons (1995) and Simons and Ikonen (1997). Reproduced with permission ofCurrent Biology from Hooper (1998).

146 N. M. HooperM

ol M

embr

Bio

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f R

egin

a on

04/

16/1

3Fo

r pe

rson

al u

se o

nly.

clustered GSLs, cholesterol, GPI-anchored proteins andacylated proteins, actually exist in the cell membrane priorto detergent-extraction? Or, does extraction of membraneswith detergent cause the more saturated and longer (andthus more hydrophobic) acyl chains of the GPI anchors,GSLs and acylated proteins to artefactually cluster togetherwith cholesterol; as the phospholipids, which generally haveshorter and more unsaturated (and thus less hydrophobic)acyl chains, are selectively removed? Studies on modelmembrane systems as well as more biophysical approachesto `visualize’ directly such domains in membranes arebeginning to provide answers to these questions.

Model membrane systems

In model membranes, physiological concentrations of cho-lesterol and sphingolipid induce formation of a liquid-orderedphase which has properties that are intermediate betweenthose of the fluid and gel phases (reviewed in Brown andLondon 1997, Brown 1998). The liquid-ordered phase ischaracterized by tight acyl chain packing and relativelyextended acyl chains. Liquid-ordered phase bilayers areinsoluble in Triton X-100. Model membranes with a similarlipid composition to that of cellular DIGs were also insolublein Triton X-100, as was the GPI-anchored alkaline phospha-tase inserted into such model membranes (Schroeder et al.1994), suggesting that the lipids themselves determine thedetergent insolubility of both the lipids and proteins in DIGs.However, the possibility still remained that detergent extrac-tion, as used to isolate DIGs, might alter the composition of

the domains or induce their formation from previouslyuniform lipid mixtures. In order to address this, Brown andcolleagues used a fluorescence quenching assay employingnitroxide-labelled lipids to detect phase separations in thepresence of cholesterol (Ahmed et al. 1997). In modelmembranes which had a similar lipid composition, inparticular in terms of their sphingomyelin composition, tothat of the plasma membrane, the formation of a liquid-ordered phase occurred. The presence of cholesterol, againat a concentration similar to that found in the plasmamembrane, promoted the formation of the liquid-orderedphase at 37 8 C. Critically, the detergent-insolubility ofcholesterol-containing model membranes closely correlatedwith the amount of liquid-ordered phase, as detected byfluorescence quenching. This led to the conclusion that thedetergent-insoluble membranes isolated from cells are likelyto exist in the liquid-ordered phase prior to detergentextraction, and that one of the more important roles ofcholesterol and sphingolipids in cell membranes may be toinduce formation of the liquid-ordered phase and, thus, theformation of lipid rafts (Ahmed et al. 1997).

In further experiments using lipid vesicles containingdipalmitoyl phosphatidylcholine and cholesterol in varyingratios, so as tohave cholesterol present in vesicles containingeither liquid-ordered or liquid-crystalline phases, the ability ofdetergent to promote the formation of the detergent-resistantliquid-ordered phase was examined (Schroeder et al. 1998).By monitoring the solubility of radiolabelled lipids in thevesicles, it was found that insoluble membranes were onlyobtained from lipid vesicles that contained lipids in the liquid-

Figure 2. Isolation of DIGs by solubilization in non-ionic detergent followed by bouyant density centrifugation. Cells/tissue are solubilized in non-ionic detergent (usually Triton X-100) at 4 8 C and adjusted to 40% (w/v) sucrose by the addition of an equal volume of 80% (w/v) sucrose. Thesamples are then layered under a 0 ± 30% (w/v) linear sucrose gradient and centrifuged overnight at 140 000 g. DIGs migrate up the sucrosegradient leaving the bulk of the solubilized membrane protein in the 40% (w/v) sucrose region at the base of the gradient. Detailed experimentalprotocols can be found in Parkin et al. (1997).

147Membrane domains, lipid rafts and caveolaeM

ol M

embr

Bio

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f R

egin

a on

04/

16/1

3Fo

r pe

rson

al u

se o

nly.

orderedphase before extraction, andthat the detergentdidnotcreate insoluble membranes where liquid-ordered domainsdid not exist previously. In a separate experiment, [3H]-

sphingomyelin or the GPI-anchored alkaline phosphatasewere inserted into fluid (non-detergent resistant) liposomesand then mixed with detergent-resistant liposomes prior to

Figure 3. Characterization of DIGs from human cerebral cortex. DIGs were prepared as shown in figure 2. The resulting sucrose gradients wereharvested in 0.5 ml fractions (fraction 0, insoluble pellet; fraction 1, base of gradient; fraction 9, top of gradient). (A) Absorbance of fractionsmeasured at 620 nm. The low-density, detergent-insoluble membrane domain in fractions 5 and 6 appears opaque and absorbs strongly at620 nm. (B) Distribution of protein in the sucrose gradient. The bulk of the protein is present in the 40% sucrose region of the gradient (fractions1 ± 3) which contains the solubilised proteins. (C) Distribution of the GPI-anchored alkaline phosphatase in the sucrose gradient. The activity ofthis enzyme is found exclusively in fractions 5 and 6. (D) Distribution of the transmembrane polypeptide-anchored aminopeptidase N in thesucrose gradient. This non-raft protein is found exclusively in fractions 1 and 2. Modified from Parkin et al. (1999).

148 N. M. HooperM

ol M

embr

Bio

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f R

egin

a on

04/

16/1

3Fo

r pe

rson

al u

se o

nly.

extraction with detergent. All of the [3H]sphingomyelin and allof the alkaline phosphatase were soluble in Triton X-100,indicatingthat these components did not j̀ump’ intodetergent-resistant domains during extraction (Schroeder et al. 1998).Thus, it wouldappear that detergent extraction of membranesdoes not induce significant artifacts in the organization of thelipids in lipid rafts, and that the lipids themselves areresponsible for their own clustering in the rafts.

Biophysical studies of membrane microdomains

Recently, several biophysical and biochemical approacheshave been employed in attempts to determine directlywhether lipid rafts exist in vivo (table 1). Single particletracking has been used to follow the movements of twocomponents of DIGs, the GPI-anchored Thy-1 protein andthe GSL GM1 on the surface of fibroblasts (Sheets et al.1997). Single particle tracking allows the movement on thesurface of a cell of individual molecules that have beenspecifically labelled with colloidal gold or fluorescent particlesto be measured with nanometer precision. Video-enhancedbrightfield microscopy was used to record the movement

(trajectories) of membrane components over a definedperiod of time. By single particle tracking, the movementsof both Thy-1 and GM1 could be categorized into four modesof lateral movement: (1) fast diffusion due to unobstructedBrownian motion within the lipid bilayer; (2) slow, anomalousdiffusion equivalent to movement through protein-rich do-mains; (3) coralled diffusion confined to 325 ± 370 nmdiameter regions; and (4) a fraction of the molecules thatwas essentially stationary on the 6.6 s time scale of theexperiment. Longer observations (60 s) showed that bothThy-1 and GM1 are transiently confined for 7 ± 9 s to regionsaveraging 260 ± 370 nm in diameter. Approximately 36% ofboth Thy-1 and GM1 undergo this confined diffusion. Incontrast, only 16% of fluorescein phosphatidylethanolamine,a phospholipid analogue which is not expected to be found inDIGs, displayed confined diffusion. Reducing the GSLexpression of the cells by ~ 40% with a glucosylceramidesynthase inhibitor (see table 2) caused the percentage oftrajectories exhibiting confinement and the size of theconfining domain for the GPI-anchored Thy-1 to be reduced~ 1.5-fold. Extraction of the cells with Triton X-100 left thefraction of molecules confined and the domain sizes for Thy-

Table 1. Summary of the techniques used to study GPI-anchored protein domains

Techniques Molecule studied

Size ofdomain(nm)

No. ofGPI-anchoredproteins perdomain

Effect of GSL/cholesteroldepletion Effect of detergent Reference

Single particletracking

Single particletracking

FRET

Cross-linking

Immunofluorescencemicroscopy

Thy-1 (GPI)GM1 (ganglioside)NCAM (GPI)

Folate receptor(GPI)

Growth hormone(GPI)

Alkalinephosphatease(GPI)

260± 370

300

<70

n.d.

n.d.

n.d.

n.d.

<50

15

n.d.

Reduces domainsize

n.d.

Disrupts domain

Disrupts domain

Disrupts domain

Domain sizeunchanged

n.d.

n.d.

Increases size ofdomain

n.d.

Sheets et al. (1997)

Simson et al. (1998)

Varma and Mayor(1998)

Friedrichson andKurzchalia (1998)

Harder et al. (1998)

n.d.=not determined.

Table 2. Agents that disrupt lipid rafts/caveolae

Action Compound Properties Reference

Inhibition of cellular cholesterolbiosynthesis

Compactin 3-Hydroxoy-3-methylglutaryl-coenzyme A reductase inhibitor

Rothberg et al. (1990)

Lovastatin (mevinolin) 3-Hydroxy-3-methylglutaryl-coenzyme A reductase inhibitor

Taraboulous et al. (1995)

Squalestatin Squalene synthase inhibitor Stulnig et al. (1997)Oxidation of cholesterol to

cholestenoneCholesterol oxidase Enzyme Smart et al. (1994b)

Cholesterol binding agents DigitoninFilipin

Methyl-b -cyclodextrin

NystatinSaponinStreptolysin O

Cardiac glycosidePolyene antibiotic

Glucopyranoside cyclic oligomers

Polyene antibioticSapogenin glycosideThiol-activated cytolysin

Rothberg et al. (1990)Rothberg et al. (1990),Schnitzer et al. (1994)Klein et al. (1995),Ilangumaran and Hoessli (1998)Rothberg et al (1990)Cerneus et al. (1993)Xie and Low (1995)

Inhibition of cellularsphingolipid biosynthesis

Fumonisin B1

D-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol.HCl

Mycotoxin, ceramide synthaseinhibitor

Glucosylceramide synthaseinhibitor

Stevens and Tang (1997)

Sheets et al. (1997)

149Membrane domains, lipid rafts and caveolaeM

ol M

embr

Bio

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f R

egin

a on

04/

16/1

3Fo

r pe

rson

al u

se o

nly.

1 and GM1 unchanged. This result is consistent with thepreferential association of GPI-anchored proteins with GSLsin distinct domains in the cell membrane. Furthermore, theresults suggest that the confining domains may be the in vivoequivalent of the detergent-insoluble membrane fractions.

Single particle tracking studies of the GPI-anchored ortransmembrane polypeptide-anchored isoforms of the humanneural cell adhesion molecule (NCAM) in fibroblasts andmuscle cells, revealed that ~ 30% of both isoforms experi-enced transient confinement for ~ 8 s within regions of~ 300 nm diameter (Simson et al. 1998). Thus, the samemechanism of confinement appears to apply to both the GPI-andthe polypeptide-anchored forms of NCAM. Diffusion of theprotein within the regions of ~ 300 nm diameter was anom-alous, consistent with movement through a dense field ofobstacles, which was likened to the movement of a ball in apinball machine. Simson et al. (1998) p. 297 concluded thatt̀he membrane appears as a mosaic containing regions thatpermit free diffusion as well as regions in which NCAM istransientlyconfined tosmallormore extendeddomains’whichmay well represent the lipid rafts.

Chemical cross-linking with a membrane-impermeablereagent with a spacer length of 1.14 nm has been usedrecently to investigate the clustering of GPI-anchored proteinson the surface of cells (Friedrichson and Kurzchalia 1998).Using this approach, a recombinant form of growth hormonewith the GPI-anchoring signal from decay-accelerating factorattachedtoits C-terminus was foundtobe clusteredindiscretedomains on the surface of MDCK cells. The clusters wereestimated to consist of at least 15 molecules of the GPI-anchoredprotein. This clusteredorganizationwas alsoseeninChinese Hamster Ovary (CHO) cells expressing the GPI-anchored folate receptor. This clustering was specific for theGPI-anchored form, as two transmembrane forms of growthhormone bearingthe same ectodomaindidnot formoligomersupon chemical cross-linking. Depletion of membrane choles-terol with methyl-b -cyclodextrin (see table 2) caused theclustering of the GPI-anchored proteins to break up, whereastreatmentof the cells withTriton X-114 substantially increasedthe size of the cross-linked complexes, possibly due tomerging of small microdomains.

Fluorescence resonance energy transfer (FRET) micro-scopy(MatkoandEddin1997) was usedtomeasure the extentof energy transfer between differently anchored isoforms ofthe folate receptor bound to a fluorescent analogue of folicacid, in terms of the dependence of fluorescence polarizationon fluorophore densities in membranes (Varma and Mayor1998). In CHO and Caco-2 cells, the extent of energy transferfor the GPI-anchored folate receptor isoform was density-independent, characteristic of organization in domains. It wasestimated that these domains are smaller than 70 nm andcontain fewer than 50 molecules of GPI-anchored proteins. Incontrast, the extent of energy transfer for the transmembrane-anchored folate receptor isoform was density-dependent,consistent with a random distribution in the membrane.Depletion of cell membrane cholesterol by treatment witheither saponin, methyl-b -cyclodextrin or compactin (see table2), disrupted the GPI-anchored protein containing domains,but had no effect on the energy transfer between thetransmembrane-anchored folate receptor isoform. Thus,

GPI-anchored proteins do appear to be organized in choles-terol-dependentsubmicron-sizeddomains on the cell surface.

Using immunofluorescence microscopy, the patching be-haviour of pairs of membrane components, as defined byinsolubility in Triton X-100, has been employed to investigatethe lipid domain structure of the plasma membrane (Harder etal. 1998). GPI-anchored alkaline phosphatase and Thy-1, thetransmembrane influenza virus haemagglutinin, and ganglio-side GM1, all components of lipid rafts, were cross-linkedusing antibodies or cholera toxin. The patches of these raftmarkers overlapped extensively in fibroblastoid BHK andJurkat T-lymphoma cells, whereas the patched raft compo-nents were sharply segregated from patches of the non-rafttransferrin receptor. Depletion of cellular cholesterol withmethyl- b -cyclodextrin disrupted the clusters of alkaline phos-phatase, indicating an involvement for cholesterol in thisprocess. Interestingly, patches of the GPI-anchored alkalinephosphatase accumulatedthe acylatedSrc-like proteinkinaseFyn which may be anchored in the cytoplasmic leaflet of lipidrafts (see figure 1). These observations led the authors tosuggest that coalescence of cross-linked raft elements ismediated by their common lipid environment, whereasseparation of raft and non-raft patches is caused by theimmiscibility of different lipid phases. These results areconsistent with the view that raft domains in the plasmamembrane are normally small and highly dispersed, but thatraft size can be modulated by oligomerization of raftcomponents.

Caveolae

Caveolae are 50 ± 100 nm flask-shaped invaginations asso-ciated with, or in the vicinity of, the plasma membrane thatwere first identified some 40 years ago in thin sectionelectron micrographs (Yamada 1955). By rapid-freeze, deep-etch electron microscopy or high resolution scanning micro-scopy a distinctive cytoplasmic coat material is visible,consisting of evenly spaced filaments that are arranged inconcentric whorls to form a `fingerprint-like’ structure.Caveolae are most abundant in simple squamous epithelia,fibroblasts, smooth muscle cells and adipocytes, but appearto be present in most cells (Bretscher and Whytock 1977,Forbes et al. 1979, Fan et al. 1983). These structures havebeen implicated in several cellular events, including signaltransduction (Sargiacomo et al. 1993, Chun et al. 1994,Schnitzer et al. 1995c) (see below), transcytosis (Simionescu1983), potocytosis (the internalization of small moleculesfrom the extracellular medium; (Rothberg et al. 1990,Anderson 1993, Smart et al. 1994a), and in regulatingcellular calcium (Fujimoto 1993, Isshiki et al. 1998).

Caveolae containGSLs (Tran et al. 1987), needcholesteroltofunction(Rothbergetal. 1990), andhave a characteristicallyhigh content of the 21 ± 24 kDa integral membrane proteincaveolin-1/VIP21 (Kurzchalia etal. 1992, 1994, Rothbergetal.1992). Molecular cloning has identified three distinct caveolingenes, caveolin-1/VIP21, caveolin-2, and caveolin-3/M-ca-veolin (for reviews see Parton 1996, Harder andSimons 1997,Okamotoetal. 1998). The caveolins forma hairpinstructure inthe membrane, with their N- and C-termini located on thecytosolic side of the membrane (see figure 1). After synthesis

150 N. M. HooperM

ol M

embr

Bio

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f R

egin

a on

04/

16/1

3Fo

r pe

rson

al u

se o

nly.

in the ER, the caveolin proteins interact with themselves toform homo- and hetero-oligomers which, upon transportthrough the Golgi, increase in size and become insoluble inTriton X-100 (Monieretal. 1995). Caveolin-1 binds cholesterol(Murata et al. 1995) and is multiply palmitoylated in its C-terminalregion(Dietzenetal. 1995, Monieretal. 1996), furthercausing the protein to associate with GSLs and cholesterol.The critical role of caveolin in the formation of caveolae wasdemonstrated when expression of caveolin-1 in lymphocytes,Fischer rat thyroid cells or Caco-2 cells, all of which normallylack caveolae, induced the formation of caveolae at the cell-surface (Fra et al. 1995, Lipardi et al. 1998, Vogel et al. 1998).

Like DIGs, caveolae are insoluble in certain non-ionicdetergents andhave a low-density as a consequence of a highlipid to protein ratio. These similarities originally led to theconclusion that DIGs equated with caveolae. However, thepioneering work of Schnitzer et al. (1995a, b, c) has clearlyshown that DIGs and caveolae are not identical structures.Using cationic colloidal silica to coat epithelial cell apicalmembranes insitu, theywere able toseparate morphologicallydistinguishable caveolae fromnon-invaginated plasma mem-brane DIGs. After in situ labelling, tissues were removed andhomogenized, and the apical membranes isolated by centrifu-gation though a Nycodenz density gradient because of theirartificially induced high density. Electron microscopy revealedthat the silica particles were excluded from caveolae, whichremained attached to the apical plasma membrane, whereasDIGslackingcaveolarmorphologywereeffectivelycoated.Themembrane fraction was subjected to detergent-solubilization,dislodging the caveolae which were then separated from thehigh-density silica coated membrane by sucrose densitygradient centrifugation. The isolated caveolae were enrichedin caveolin, ganglioside GM1, Ca2+-ATPase and the inositol1,4,5-trisphosphate receptor(Schnitzeretal. 1995b, c), as wellas proteins involved in vesicle budding, docking and fusion(vesicle-associated membrane protein, VAMP; N-ethylmalei-mide-sensitive fusion protein, NSF; soluble NSF attachmentprotein, SNAP; annexins II and VI; monomeric and trimericGTPases) (Schnitzer et al. 1995a). In contrast, the GPI-anchored proteins 5¢ -nucleotidase, carbonic anhydrase, andurokinase-plasminogen activator receptor were absent fromthe isolated caveolae, but found to be highly enriched in theDIGs. Immuno-gold electron microscopy revealed that GM1-enriched caveolae associated with an annular plasma mem-brane domainenriched inGPI-anchoredproteins (Schnitzeretal. 1995b). A variation of the method of Schnitzer andcolleagues to isolate caveolae following perfusion of rat lungwith cationic silica has been reported which, following densitygradient centrifugation through Nycodenz, used anti-caveolincoated magnetic beads, instead of detergent solubilization, toimmuno-isolate the caveolae (Stan et al. 1997).

It has been reported that GPI-anchored proteins only entercaveolae artefactually following either cross-linking withantibodies or detergent treatment (Mayor et al. 1994, Mayorand Maxfield 1995, Fujimoto 1996). However, patching ofmembrane components followed by electron microscopyrevealed that patches of the GPI-anchored alkaline phospha-tase were occasionally found in the vicinity of caveolarinvaginations but were mostly detected at smooth domainsof the plasma membrane, and that co-patching of haemagglu-

tinin and alkaline phosphatase did not cause these two raftcomponents to enter caveolae (Harder et al. 1998).

In addition to the proteins noted above, numerous othersignalling molecules have been localized to caveolae, includ-ing non-receptor tyrosine kinases (Fyn, Src, Ras, Yes), nitricoxide synthase, epidermal growth factor receptor, insulinreceptor, neurotrophin receptor, platelet-derived growthfactorreceptor, protein kinase C isoforms, phospholipase C,adenylyl cyclase, sphingomyelin, ceramide and phosphoino-sitides (Sargiocomoetal. 1993, Changetal. 1994, Lisantietal.1994b, Liu and Anderson 1995, Smart et al. 1995, Hope andPike 1996, Li et al. 1996, Mineo et al. 1996, Pike and Casey1996, Songetal. 1996, Huangetal. 1997, Liuetal. 1997, Wuetal. 1997). This has ledtothe hypothesis thatcaveolae are sitesfororganized cell surface signal transduction(Anderson 1993,Lisanti et al. 1994a, Okamoto et al. 1998).

Role of cholesterol in the formation and function of lipidrafts and caveolae

As noted above, both lipid rafts and caveolae are enriched incholesterol. Thus, altering the level of cholesterol in the cellmembrane has been used extensively to disrupt lipid raftsand caveolae, and to probe the role of these structures in avariety of biological processes, including clathrin-indepen-dent endocytosis (potocytosis) (Deckert et al. 1996, Smart etal. 1996a, Stevens and Tang 1997), transcytosis (Schnitzeret al. 1994), cholesterol transport (Smart et al. 1996b), signaltransduction (Stulnig et al. 1997), conversion of the prionprotein to the scrapie isoform (Taraboulos et al. 1995) andbacterial and viral internalization (Anderson et al. 1996,Baorto et al. 1997). Various agents have been described inthe literature which can be used to alter membranecholesterol levels (table 2) and it is clear from such studiesthat cholesterol is intimately involved in the formation andfunction of lipid rafts and caveolae.

But, whatexactly is the role of cholesterol in the formation oflipidrafts? Inphosphatidylcholine-cholesterol (9 : 1) liposomesonly 1% of the cholesterol (Schroeder et al. 1994) and only15% of the GPI-anchored membrane dipeptidase (Parkin,Turner and Hooper, unpublished) were resistant to detergentextraction, indicating that cholesterol on its own does notconfer detergent-insolubility. In liposomes containing a highsphingolipid concentration, [3H]sphingomyelin (Schroeder etal. 1998) or the GPI-anchored membrane dipeptidase (Parkin,Turner and Hooper, unpublished) were detergent-insoluble,even in the absence of cholesterol, consistent with a highsphingolipid concentration promoting a detergent-insolublephase and indicating that cholesterol is not required fordetergent-insolubility. At lower concentrations of sphingoli-pids, much less [3H]sphingomyelin was detergent-insoluble inliposomes lacking cholesterol than in cholesterol containingliposomes (Schroeder et al. 1998), indicating that cholesterolenhances the formation of a detergent-insoluble phase. Thisconclusion is supported by the observation that removal ofcholesterol, by saponin treatment, from liposomes containingthe GPI-anchored alkaline phosphatase, cholesterol and arelatively high concentration of sphingolipids (33 mol%) didnot affect the detergent-insolubility of the alkaline phospha-tase (Schroeder et al. 1998). On the basis of these and other

151Membrane domains, lipid rafts and caveolaeM

ol M

embr

Bio

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f R

egin

a on

04/

16/1

3Fo

r pe

rson

al u

se o

nly.

observations, Brown and colleagues have proposed that theplasma membrane of eukaryotic cells would contain enoughsphingolipids to form liquid-ordered, detergent-insoluble do-mains in the presence of cholesterol, but not enough to formordered domains in the absence of sterols (Schroeder et al.1998). This observation would go some way to explainingreports that depleting cellularcholesterol reduces the TritonX-100 insolubility of the GPI-anchored alkaline phosphatase(Cerneus et al. 1993, Hanada et al. 1995).

In a recent study (Ilangumaran and Hoessli 1998), it wasreported that treatment of lymphocytes with methyl-b -cyclodextrin, which removes cholesterol from membraneswithout binding to or inserting into the membrane, releasedsubstantial quantities of the GPI-anchored Thy-1, theganglioside GM1, and the acylated tyrosine kinases Lckand Fyn. The majority of the Thy-1 and GM1 released bymethyl- b -cyclodextrin were subsequently recovered in thelow density fractions on sucrose density gradient centrifu-gation. Previously, it has been reported that the cholesterolbinding agent streptolysin O released membrane vesiclesenriched in GPI-anchored proteins (Xie and Low 1995) anda depletion in cellular levels of the GPI-anchored prionprotein following methyl- b -cyclodextrin treatment of cellshas been noticed (Walmsley and Hooper, unpublished).Thus, it is not clear if some of these cholesterol bindingdrugs actually release whole cholesterol-GSL-GPI-anchoredprotein domains from the cell surface; an effect that willdrastically perturb the cell, as evidenced by the reduction incell volume and the release of the transmembranepolypeptide anchored protein CD45 (Ilangumaran andHoessli 1998). Further work, examining the effect ofmethyl- b -cyclodextrin extraction of cholesterol, from bothcell membranes and liposomes, is required to clarifywhether such treatment non-specifically releases GPI-anchored proteins, GSLs and other proteins from the cell-surface in relatively large membrane fragments.

Does detergent-insolubility equate with lipid raftlocalization?

The key biochemical tool used to study the composition oflipid rafts and caveolae is insolubility in the detergent TritonX-100 at 4 8 C. But, if a protein is insoluble in Triton X-100 atthis temperature, is it irrefutable evidence for its associationwith such structures? Before the discovery of GPI anchors,insolubility in Triton X-100 was taken often as an indication ofthe association of the protein with the cytoskeleton (seeabove and discussed in Low1989). Only when the detergent-solubility pattern of multiple GPI-anchored proteins wasexamined (Hooper and Turner 1988a, b), and proteinsreported to be detergent-insoluble were identified subse-quently as GPI anchored, was there evidence to indicate thatthe GPI anchor itself conveyed the property of detergent-insolubility, rather than a direct interaction of the protein withthe cytoskeleton. Studies with model membranes (seeabove) have now clearly shown that detergent-insolubility isan intrinsic property of the lipids in the GPI anchorassociating with GSLs. However, for non-GPI anchoredproteins, can the presence of the protein in DIGs followingsucrose density gradient centrifugation be taken to indicate

that in the cell membrane the protein associates with lipidrafts or caveolae?

In a critical commentary entitled `Guilt by insolubilityÐ doesa protein’s detergent insolubility reflect a caveolar location?’,Kurzchalia et al. (1995) questioned the use of detergent-insolubility as a marker for lipid rafts/caveolae in the absenceof other experimental evidence. Certainly, the presence of aprotein in DIGs needs to be interpreted with caution, as evenclathrin, which is known to be excluded from caveolae-likemembrane microdomains (Smart et al. 1995), has beenreported to be present in low-density, detergent-insolublefractions isolated fromratembryonic corticalneurons (Bouillotet al. 1996). However, by carefully monitoring the distributionof other detergent-soluble, non-raft marker proteins, it hasbeen shown clearly that the presence of clathrin in DIGs wasdue probably to contamination by non-raft proteins (Parkin etal. 1997). Suchcontaminationof DIGs bynon-raftproteins canreadily come about if the detergent:protein ratio is not highenough to effectivelysolubilize the membrane (Parkin, Turnerand Hooper, unpublished). In an attempt to overcome theproblem of detergent extraction causing non-raft proteins toco-isolate with resident lipid raft proteins and lipids, detergent-free methods have been developed for the isolation ofcaveolae-like fractions (Smart et al. 1995, Song et al. 1996).Again, however, these methods often result in contaminationof the caveolae-like fraction by non-caveolar proteins. It istherefore advisable to monitor the distribution not only ofknown lipid raft/caveolar markers (such as caveolin or flotillin(Bickel et al. 1997)), but also of known non-raft/caveolarmarkers (see figure 3).

For example, currently there is much controversy overwhethera sub-populationofthe Alzheimer’s amyloidprecursorprotein is present inlipidrafts. The amyloidprecursorprotein isa transmembrane polypeptide anchored protein that isproteolytically cleaved to generate the insoluble b -amyloidpeptide, which is then deposited in the extracellular senileplaques found in the brains of those individuals withAlzheimer’s disease (Price and Sisodia 1998). Some studieshave reported that a minor amount of the amyloid precursorproteinis presentinDIGs (Bouillotetal. 1996, Ikezuetal. 1998,Lee et al. 1998, Simons et al. 1998), while others have failed todetectit insuchfractions (Parkinetal. 1997, Tienarietal. 1997,Morishima-Kawashima and Ihara 1998). However, using therelatively crude technique of insolubility in Triton X-100, itcannot be ruled out that a minor proportion of the amyloidprecursor protein may reside in lipid rafts under certainphysiological/pathophysiological conditions, such as in-creased levels of membrane cholesterol (Simons et al. 1998).

Conclusions and future directions

There is a growing body of evidence to indicate that certainmembrane components do indeed exist in discrete domainsin the cell membrane in vivo. However, further work is stillrequired to determine if multiple GPI-anchored proteins arepresent in the same domains or if some are localized todifferent domains; to determine why only a subset ofpolypeptide-anchored proteins are present in such domainsand whether such non-acylated proteins are held in lipid raftsthrough specific aromatic amino acid-carbohydrate stacking

152 N. M. HooperM

ol M

embr

Bio

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f R

egin

a on

04/

16/1

3Fo

r pe

rson

al u

se o

nly.

interactions as seen in the formation of the purple membranecrystalline patches in Archaea (Weik et al. 1998); todetermine whether acylated proteins, such as the Src familykinases, are localized to the cytoplasmic leaflet of domains inwhich GPI-anchored proteins are localized to the outer leafletand how this may result in signal transduction across thebilayer (Brown 1993); and to determine if, and how, GPI-anchored proteins and other membrane components tran-siently move into caveolae under certain conditions. Simul-taneously tracking the movement of two different moleculeswith fluorescent particles having different colours (Sheets etal. 1997) or using the FRET technique in conjunction withappropriate protein chimeras tagged with green fluorescentprotein (Varma and Mayor 1998) may go some way toanswering these questions and revealing the moleculararrangement of many proteins and lipids in cell membranes.Also, methods other than gross insolubility in certaindetergents need to be developed to routinely distinguishbetween lipid rafts and caveolae and to reliably identifyproteins that reside in these structures in vivo, and the effectof cholesterol binding agents on such structures requirescareful re-examination. Undoubtedly though, the clustering ofcholesterol, GSLs, GPI-anchored proteins and certain othermembrane proteins in distinct domains within membranesfacilitates molecular interactions by bringing molecules intoclose physical association and/or by increasing the localconcentration of specific molecules that are required forspecialized cellular functions.

Acknowledgements

I gratefully acknowledge the financial support of the Wellcome Trustand the Medical Research Council, and thank Dr Ed Parkin forassistance with the figures.

References

Ahmed, S. N., Brown, D. A. and London, E., 1997, On the origin ofsphingolipid/cholesterol-rich detergent-insoluble cell membranes:physiological concentrations of cholesterol and sphingolipidinduce formation of a detergent-insoluble, liquid-ordered lipidphase in model membranes. Biochemistry, 36, 10944 ± 10953.

Anderson, H. A., Chen, Y. and Norkin, L. C., 1996, Bound simianvirus 40 translocates to caveolin-enriched membrane domains,and its entry is inhibited by drugs that selectively disrupt caveolae.Molecular Biology of the Cell, 7, 1825 ± 1834.

Anderson, R. G. W., 1993, Caveolae: where incoming and outgoingmessengers meet. Proceedings of the National Academy ofSciences USA, 90, 10909 ± 10913.

Baorto, D. M., Gao, Z., Malaviya, R., Dustin, M. L., van der Merwe,A., Lublin, D. M. and Abraham, S. N., 1997, Survival of FimH-expressing enterobacteria in macrophages relies on glycolipidtraffic. Nature, 389, 636 ± 639.

Bickel, P. E., Scherer, P. E., Schnitzer, J. A., Oh, P., Lisanti, M. P.and Lodish, H. F., 1997, Flotillin and epidermal surface antigendefine a new family of caveolae-associated integral membraneproteins. Journal of Biological Chemistry, 272, 13793 ± 13802.

Bouillot, C., Prochiantz, A., Rougon, G. and Allinquant, B., 1996,Axonal amyloid precursor protein expressed by neurons in vitro ispresent in a membrane fraction with caveolae-like properties.Journal of Biological Chemistry, 271, 7640 ± 7644.

Bretscher, M. and Whytock, S., 1977, Membrane-associatedvesicles in fibroblasts. Journal of Ultrastructural Research, 61,215 ± 217.

Brown, D. A., 1993, The tyrosine kinase connection: how GPI-anchored proteins activate T cells. Current Opinion in Immunol-ogy, 5, 349 ± 354.

Brown, D. A. and London, E., 1997, Structure of detergent-resistantmembrane domains: does phase separation occur in biologicalmembranes? Biochemical and Biophysical Research Communi-cations, 240, 1 ± 7.

Brown, D. A. and London, E., 1998a, Functions of lipid rafts inbiological membranes. Annual Review of Cell and DevelopmentalBiology, 14, 111 ± 136.

Brown, D. A. and London, E., 1998b, Structure and origin of orderedlipid domains in biological membranes. Journal of MembraneBiology, 164, 103 ± 114.

Brown, D. A. and Rose, J. K., 1992, Sorting of GPI-anchoredproteins to glycolipid-enriched membrane subdomains duringtransport to the apical cell surface. Cell, 68, 533 ± 544.

Brown, R. E., 1998, Sphingolipid organisation in biomembranes:what physical studies of model membranes reveal. Journal of CellScience, 111, 1 ± 9.

Cerneus, D. P., Ueffing, E., Posthuma, G., Strous, G. J. and van derEnde, A., 1993, Detergent insolubility of alkaline phosphataseduring biosynthetic transport and endocytosis. Role of cholesterol.Journal of Biological Chemistry, 268, 3150 ± 3155.

Chang, W.-J., Ying, Y., Rothberg, K. G., Hooper, N. M., Turner, A. J.,Gambliel, H. A., De Gunzburg, J., Mumby, S. M., Gilman, A. G.and Anderson, R. G. W., 1994, Purification and characterization ofsmooth muscle cell caveolae. Journal of Cell Biology, 126, 127 ±138.

Chun, M., Liyanage, U. K., Lisanti, M. P. and Lodish, H. F., 1994,Signal transduction of a G protein-coupled receptor in caveolae:colocalization of endothelin and its receptor with caveolin.Proceedings of the National Academy of Sciences USA, 91,11728 ± 11732.

Danielsen, E. M., 1995, Involvement of detergent-insoluble com-plexes in the intracellular transport of intestinal brush borderenzymes. Biochemistry, 34, 1596 ± 1605.

Davies, A. A., Wigglesworth, N. M., Allan, D., Owens, R. J. andCrumpton, M. J., 1984, Nonidet P-40 extraction of lymphocyteplasma membrane. Biochemical Journal, 219, 301 ± 308.

Deckert, M., Ticchioni, M. and Bernard, A., 1996, Endocytosis ofGPI-anchored proteins in human lymphocytes: role of glycolipid-based domains, actin cytoskeleton, and protein kinases. Journalof Cell Biology, 133, 791 ± 799.

Dietzen, D. J., Hastings, W. R. and Lublin, D. M., 1995, Caveolin ispalmitoylated on multiple cysteine residues. Palmitoylation is notnecessary for localization of caveolin to caveolae. Journal ofBiological Chemistry, 270, 6838 ± 6842.

Fan, J. Y., Carpentier, J. L., Vanobberghen, E., Grunfeld, C.,Gorden, P. and Orci, L., 1983, Morphological changes of the 3T3-L1 fibroblast plasma membrane upon differentiation to theadipocyte form. Journal of Cell Science, 61, 219 ± 230.

Forbes, M. S., Rennels, M. and Nelson, E., 1979, Caveolar systemsand sarcoplasmic reticulum in coronary smooth muscle cells.Journal of Ultrastructural Research, 67, 325 ± 339.

Fra, A. M., Williamson, E., Simons, K. and Parton, R. G., 1995, Denovo formation of caveolae in lymphocytes by expression ofVIP21-caveolin. Proceedings of the National Academy ofSciences USA, 92, 8655 ± 8659.

Friedrichson, T. and Kurzchalia, T. V., 1998, Microdomains of GPI-anchored proteins in living cells revealed by crosslinking. Nature,394, 802 ± 805.

Fujimoto, T., 1993, Calcium pump of the plasma membraneis localised in caveolae. Journal of Cell Biology, 120, 1147 ±1157.

Fujimoto, T., 1996, GPI-anchored proteins, glycosphingolipids, andsphingomyelin are sequestered to caveolae only after cross-linking. Journal of Histochemistry and Cytochemistry, 44, 929 ±941.

Garcia-Cardena, G., Oh, P., Liu, J., Schnitzer, J. E. and Sessa, W.C., 1996, Targeting of nitric oxide synthase to endothelial cellcaveolae via palmitoylation: implications for nitric oxide signaling.Proceedings of the National Academy of Sciences USA, 93,6448 ± 6453.

153Membrane domains, lipid rafts and caveolaeM

ol M

embr

Bio

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f R

egin

a on

04/

16/1

3Fo

r pe

rson

al u

se o

nly.

Hagman, J. and Fishman, P. H., 1982, Detergent extraction ofcholera toxin and gangliosides from cultured cells and isolatedmembranes. Biochimica et Biophysica Acta, 720, 181 ± 187.

Hanada, K., Nishijima, M., Akamatsu, Y. and Pagano, R. E., 1995,Both sphingolipids and cholesterol participate in the detergentinsolubility of alkaline phosphatase, a glycosylphosphatidylinosi-tol-anchored protein, in mammalian membranes. Journal ofBiological Chemistry, 270, 6254 ± 6260.

Harder, T. and Simons, K., 1997, Caveolae, DIGs, and the dynamicsof sphingolipid-cholesterol microdomains. Current Opinion in CellBiology, 9, 534 ± 542.

Harder, T., Scheiffele, P., Verkade, P. and Simons, K., 1998, Lipiddomain structure of the plasma membrane revealed by patching ofmembrane components. Journal of Cell Biology, 141, 929 ± 942.

Hoessli, D. and Rungger-Brandle, E., 1985, Association of specificcell-surface glycoproteins with a Triton X-100-resistant complex ofplasma membrane proteins isolated from T-lymphoma cells.Experimental Cell Research, 156, 239 ± 250.

Hooper, N. M., 1998, Do glycolipid microdomains really exist?Current Biology, 8, R114 ± R116.

Hooper, N. M. and Turner, A. J., 1988a, Ectoenzymes of thekidney microvillar membrane. Aminopeptidase P is anchored bya glycosyl-phosphatidylinositol moiety. FEBS Letters, 229, 340 ±344.

Hooper, N. M. and Turner, A. J., 1988b, Ectoenzymes of the kidneymicrovillar membrane. Differential solubilization by detergents canpredict a glycosyl-phosphatidylinositol membrane anchor. Bio-chemical Journal, 250, 865 ± 869.

Hooper, N. M., Low, M. G. and Turner, A. J., 1987, Renal dipeptidaseis one of the membrane proteins released by phosphatidylinositol-specific phospholipase C. Biochemical Journal, 244, 465 ± 469.

Hope, H. R. and Pike, L. J., 1996, Phosphoinositides andphosphoinositide-utilizing enzymes in detergent-insoluble lipiddomains. Molecular Biology of the Cell, 7, 843 ± 851.

Huang, C., Hepler, J. R., Chen, L. T., Gilman, A. G., Anderson, R. G.W. and Mumby, S. M., 1997, Organisation of G proteins andadenylyl cyclase at the plasma membrane. Molecular Biology ofthe Cell, 8, 2365 ± 2378.

Ikezu, T., Trapp, B. D., Song, K. S., Schlegel, A., Lisanti, M. P. andOkamoto, T., 1998, Caveolae, plasma membrane microdomainsfor a -secretase-mediated processing of the amyloid precursorprotein. Journal of Biological Chemistry, 273, 10485 ± 10495.

Ilangumaran, S. and Hoessli, D. C., 1998, Effects of cholesteroldepletion by cyclodextrin on the sphingolipid microdomains of theplasma membrane. Biochemical Journal, 335, 433 ± 440.

Isshiki, M., Ando, J., Korenaga, R., Kogo, H., Fujimoto, T., Fujita, T.and Kamiya, A., 1998, Endothelial Ca2+ waves preferentiallyoriginate at specific loci in caveolin-rich cell edges. Proceedings ofthe National Academy of Sciences USA, 95, 5009 ± 5014.

Keller, P. and Simons, K., 1997, Post-Golgi biosynthetic trafficking.Journal of Cell Science, 110, 3001 ± 3009.

Kim, H. S. and Campbell, B. J., 1983, Association of renaldipeptidase with the triton-insoluble fraction of kidney microvilli.Journal of Membrane Biology, 75, 115 ± 122.

Klein, U., Gimpl, G. and Fahrenholz, F., 1995, Alteration of themyometrial plasma membrane cholesterol content with b -cyclo-dextrin modulates the binding affinity of the oxytocin receptor.Biochemistry, 34, 13784 ± 13793.

Kurzchalia, T. V., Dupree, P. and Monier, S., 1994, VIP21-Caveolin,a protein of the trans-Golgi network and caveolae. FEBS Letters,346, 88 ± 91.

Kurzchalia, T., Dupree, P., Parton, R. G., Kellner, R., Virta, H.,Lehnert, M. and Simons, K., 1992, VIP21, a 21-kD membraneprotein is an integral component of trans-Golgi-network-derivedtransport vesicles. Journal of Cell Biology, 118, 1003 ± 1014.

Kurzchalia, T. V., Hartmann, E. and Dupree, P., 1995, Guilt byinsolubilityÐ does a protein’s detergent insolubility reflect acaveolar location? Trends in Cell Biology, 5, 187 ± 189.

Lee, S.-J., Liyanage, U., Bickel, P. E., Xia, W., Lansbury, P. T. J. andKosik, K. S., 1998, A detergent-insoluble membrane compartmentcontains Ab in vivo. Nature Medicine, 4, 730 ± 734.

Li, S., Seitz, R. and Lisanti, M. P., 1996, Phosphorylation of caveolinby Src tyrosine kinases. Journal of Biological Chemistry, 271,3863 ± 3868.

Lipardi, C., Mora, R., Colomer, V., Paladino, S., Nitsch, L.,Rodriguez-Boulan, E. and Zurzolo, C., 1998, Caveolin transfectionresults in caveolae formation but not apical sorting of glycosylpho-sphatidylinositol (GPI)-anchored proteins in epithelial cells. Jour-nal of Cell Biology, 140, 617 ± 626.

Lisanti, M. P., Scherer, P. E., Tang, Z. and Sargiacomo, M., 1994a,Caveolae, caveolin and caveolin-rich membrane domains: asignalling hypothesis. Trends in Cell Biology, 4, 231 ± 235.

Lisanti, M. P., Scherer, P. E., Viugirene, J., Tang, Z., Hermanowski-Vosatka, A., Tu, Y.-H., Cook, R. F. and Sargiacomo, M., 1994b,Characterization of caveolin-rich membrane domains isolatedfrom an endothelial-rich source: implications for human disease.Journal of Cell Biology, 126, 111 ± 126.

Lisanti, M. P., Tang, Z., Scherer, P. E. and Sargiacomo, M., 1995,Caveolae purification and glycosylphosphatidylinositol-linked pro-tein sorting in polarized epithelia. Methods in Enzymology, 250,655 ± 668.

Liu, J., Oh, P., Horner, T., Rogers, R. A. and Schnitzer, J. E., 1997,Organised endothelial cell surface signal transduction in caveolaedistinct from glycosylphosphatidylinositol-anchored protein micro-domains. Journal of Biological Chemistry, 272, 7211 ± 7222.

Liu, P. and Anderson, R. G. W., 1995, Compartmentalizedproduction of ceramide at the cell surface. Journal of BiologicalChemistry, 270, 27179 ± 27185.

Low, M. G., 1989, The glycosyl-phosphatidylinositol anchor of mem-brane proteins. Biochimica et Biophysica Acta, 988, 427 ± 454.

Matko, J. and Eddin, M., 1997, Energy transfer methods for detectingmolecular clusters on cell surfaces. Methods in Enzymology, 278,444 ± 462.

Mayor, S. and Maxfield, F. R., 1995, Insolubility and redistribution ofGPI-anchored proteins at the cell surface after detergent treat-ment. Molecular Biology of the Cell, 6, 929 ± 944.

Mayor, S., Rothberg, K. G. and Maxfield, F. R., 1994, Sequestrationof GPI-anchored proteins in caveolae triggered by cross-linking.Science, 264, 1948 ± 1951.

Melkonian, K. A., Chu, T., Tortorella, L. B. and Brown, D. A., 1995,Characterization of proteins in detergent-resistant membranecomplexes from Madin-Darby canine kidney epithelial cells.Biochemistry, 34, 16161 ± 16170.

Mineo, C., James, G. L., Smart, E. J. and Anderson, R. G. W., 1996,Localization of epidermal growth factor-stimulated Ras/Raf-1interaction to caveolae membrane. Journal of Biological Chem-istry, 271, 11930 ± 11935.

Monier, S., Dietzen, D. J., Hastings, W. R., Lublin, D. M. andKurzchalia, T. V., 1996, Oligomerisation of VIP21-caveolin in vitrois stabilised by long chain fatty acylation or cholesterol. FEBSLetters, 388, 143 ± 149.

Monier, S., Parton, R. G., Vogel, F., Behlke, J., Henske, A. andKurzchalia, T. V., 1995, VIP21-caveolin, a membrane proteinconstituent of the caveolar coat, oligomerizes in vivo and in vitro.Molecular Biology of the Cell, 6, 911 ± 927.

Morishima-Kawashima, M. and Ihara, Y., 1998, The presence ofamyloid b -protein in the detergent-insoluble membrane compart-ment of human neuroblastoma cells. Biochemistry, 37, 15247 ±15253.

Murata, M., Peranen, J., Schreiner, R., Wieland, F., Kurzchalia, T. V.and Simons, K., 1995, VIP21/caveolin is a cholesterol-bindingprotein. Proceedings of the National Academy of Sciences USA,92, 10339 ± 10343.

Okamoto, T., Schlegel, A., Scherer, P. E. and Lisanti, M. P., 1998,Caveolins, a family of scaffolding proteins for organising `pre-assembled signalling complexes’ at the plasma membrane.Journal of Biological Chemistry, 273, 5419 ± 5422.

Parkin, E. T., Hussain, I., Karran, E. H., Turner, A. J. and Hooper, N.M., 1999, Characterisation of detergent-insoluble complexescontaining the familial Alzheimer’s disease-associated presenilins.Journal of Neurochemistry, 72, in press.

154 N. M. HooperM

ol M

embr

Bio

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f R

egin

a on

04/

16/1

3Fo

r pe

rson

al u

se o

nly.

Parkin, E. T., Hussain, I., Turner, A. J. and Hooper, N. M., 1997, Theamyloid precursor protein is not enriched in caveolae-like,detergent-insoluble membrane microdomains. Journal of Neuro-chemistry, 69, 2179 ± 2188.

Parolini, I., Sargiacomo, M., Lisanti, M. P. and Peschle, C., 1996,Signal transduction and glycophosphatidylinositol-linked proteins(lyn, lck, CD4, CD45, G proteins, and CD55) selectively localize inTriton- insoluble plasma membrane domains of human leukemiccell lines and normal granulocytes. Blood, 87, 3783 ± 3794.

Parton, R. G., 1996, Caveolae and caveolins. Current Opinion in CellBiology, 8, 542 ± 548.

Parton, R. G. and Simons, K., 1995, Digging into caveolae. Science,269, 1398 ± 1399.

Pike, L. J. and Casey, L., 1996, Localization and turnover ofphosphatidylinositol 4,5-bisphosphate in caveolin-enriched mem-brane domains. Journal of Biological Chemistry, 271, 26453 ±26456.

Price, D. L. and Sisodia, S. S., 1998, Mutant genes in familialAlzheimer’s disease and transgenic models. Annual Review ofNeuroscience, 21, 479 ± 505.

Rietveld, A. and Simons, K., 1998, The differential miscibility of lipidsas the basis for the formation of functional membrane rafts.Biochimica et Biophysica Acta, 1376, 467 ± 479.

Rodgers, W., Crise, B. and Rose, J. K., 1994, Signals determiningprotein tyrosine kinase and glycosyl-phosphatidylinositol-an-chored protein targeting to a glycolipid-enriched membranefraction. Molecular and Cellular Biology, 14, 5384 ± 5391.

Rothberg, K. G., Heuser, J. E., Donzell, W. C., Ying, Y.-S., Glenney,J. R. and Anderson, R. G. W., 1992, Caveolin, a proteincomponent of caveolae membrane coats. Cell, 68, 673 ± 682.

Rothberg, K. G., Ying, Y.-S., Kamen, B. A. and Anderson, R. G. W.,1990, Cholesterol controls the clustering of the glycophospholipid-anchored membrane receptor for 5-methyl-tetrahydrofolate. Jour-nal of Cell Biology, 111, 2931 ± 2938.

Sargiacomo, M., Sudol, M., Tang, Z. and Lisanti, M. P., 1993, Signaltransducing molecules and glycosyl-phosphatidylinositol-linkedproteins form a caveolin-rich insoluble complex in MDCK cells.Journal of Cell Biology, 122, 789 ± 807.

Schnitzer, J. E., Liu, J. and Oh, P., 1995a, Endothelial caveolae havethe molecular transport machinery for vesicle budding, docking,and fusion including VAMP, NSF, SNAP, annexins, and GTPases.Journal of Biological Chemistry, 270, 14399 ± 14404.

Schnitzer, J. E., McIntosh, D. P., Dvorak, A. M., Liu, J. and Oh, P.,1995b, Separation of caveolae from associated microdomains ofGPI-anchored proteins. Science, 269, 1435 ± 1439.

Schnitzer, J. E., Oh, P., Jacobson, B. S. and Dvorak, A. M., 1995c,Caveolae from luminal plasmalemma of rat lung endothelium:microdomains enriched in caveolin, Ca2+,-ATPase, and inositoltrisphosphate receptor. Proceedings of the National Academy ofSciences USA, 92, 1759 ± 1763.

Schnitzer, J. E., Oh, P., Pinney, E. and Allard, J., 1994, Filipin-sensitive caveolae-mediated transport in endothelium: reducedtranscytosis, scavenger endocytosis, and capillary permeabilityof select macromolecules. Journal of Cell Biology, 127, 1217 ±1232.

Schroeder, R., London, E. and Brown, D., 1994, Interactionsbetween saturated acyl chains confer detergent resistance onlipids and glycosylphosphatidylinositol (GPI)-anchored proteins:GPI-anchored proteins in liposomes and cells show similarbehaviour. Proceedings of the National Academy of SciencesUSA, 91, 12130 ± 12134.

Schroeder, R. J., Ahmed, S. M., Zhu, Y., London, E. and Brown, D.A., 1998, Cholesterol and sphingolipid enhance the Triton X-100insolubility of glycosylphosphatidylinositol-anchored proteins bypromoting the formation of detergent-insoluble ordered membranedomains. Journal of Biological Chemistry, 273, 1150 ± 1157.

Shaul, P. W., Smart, E. J., Robinson, L. J., German, Z., Yuhanna, I.S., Ying, Y., Anderson, R. G. W. and Michel, T., 1996, Acylationtargets endothelial nitric-oxide synthase to plasmalemmal caveo-lae. Journal of Biological Chemistry, 271, 6518 ± 6522.

Sheets, E. D., Lee, G. M. S. R. and Jacobson, K., 1997, Transientconfinement of a glycosylphosphatidylinositol-anchored protein inthe plasma membrane. Biochemistry, 36, 12449 ± 12458.

Simionescu, N., 1983, Cellular aspects of transcapillary exchange.Physiological Reviews, 63, 1536 ± 1560.

Simons, K. and Ikonen, E., 1997, Functional rafts in cell membranes.Nature, 387, 569 ± 572.

Simons, K. and van Meer, G., 1988, Lipid sorting in epithelial cells.Biochemistry, 27, 6197 ± 6202.

Simons, K. and Wadinger-Ness, A., 1990, Polarised sorting inepithelia. Cell, 62, 207 ± 210.

Simons, M., Keller, P., De Strooper, B., Beyreuther, K., Dotti, C. G.and Simons, K., 1998, Cholesterol depletion inhibits the genera-tion of b -amyloid in hippocampal neurons. Proceedings of theNational Academy of Sciences USA, 95, 6460 ± 6464.

Simson, R., Yang, B., Moore, S. E., Doherty, P., Walsh, F. S. andJacobson, K. A., 1998, Structural mosaicism on the submicronscale in the plasma membrane. Biophysical Journal, 74, 297 ±308.

Skibbens, J. E., Roth, M. G. and Matlin, K. S., 1989, Differentialextractability of influenza virus hemagglutinin during intracellulartransport in polarised epithelial cells and nonpolar fibroblasts.Journal of Cell Biology, 108, 821 ± 832.

Smart, E. J., Foster, D. C., Ying, Y.-S., Kamen, B. A. and Anderson,R. G. W., 1994a, Protein kinase C activators inhibit receptor-mediated potocytosis by preventing internalisation of caveolae.Journal of Cell Biology, 124, 307 ± 313.

Smart, E. J., Mineo, C. and Anderson, R. G. W., 1996a, Clusteredfolate receptors deliver 5-methyltetrahydrofolate to cytoplasm ofMA104 cells. Journal of Cell Biology, 134, 1169 ± 1177.

Smart, E. J., Ying, Y.-S., Conrad, P. A. and Anderson, R. G. W.,1994b, Caveolin moves from caveolae to the Golgi apparatus inresponse to cholesterol oxidation. Journal of Cell Biology, 127,1185 ± 1197.

Smart, E. J., Ying, Y., Donzell, W. C. and Anderson, R. G. W.,1996b, A role for caveolin in transport of cholesterol fromendoplasmic reticulum to plasma membrane. Journal of BiologicalChemistry, 271, 29427 ± 29435.

Smart, E. J., Ying, Y., Mineo, C. and Anderson, R. G. W., 1995, Adetergent-free method for purifying caveolae membrane fromtissue culture cells. Proceedings of the National Academy ofSciences USA, 92, 10104 ± 10108.

Song, K. S., Li, S., Okamoto, T., Quilliam, L. A., Sargiacomo, M. andLisanti, M. P., 1996, Co-purification and direct interaction of Raswith caveolin, an integral membrane protein of caveolae micro-domains. Detergent-free purification of caveolae membranes.Journal of Biological Chemistry, 271, 9690 ± 9697.

Stan, R.-V., Roberts, W. G., Predescu, D., Ihida, K., Saucan, L.,Ghitescu, L. and Palade, G. E., 1997, Immunoisolation and partialcharacterisation of endothelial plasmalemmal vesicles (caveolae).Molecular Biology of the Cell, 8, 595 ± 605.

Stevens, V. L. and Tang, J., 1997, Fumonisin B1-induced sphingo-lipid depletion inhibits vitamin uptake via the glycosylphosphati-dylinositol-anchored folate receptor. Journal of BiologicalChemistry, 272, 18020 ± 18025.

Streuli, C. H., Patel, B. and Critchley, D. R., 1981, The cholera toxinreceptor ganglioside GM1 remains associated with Triton X-100cytoskeletons of BALB/c-3T3 cells. Experimental Cell Research,136, 247 ± 254.

Stulnig, T. M., Berger, M., Sigmund, T., Stockinger, H., Horejsi, V.and Waldhausl, W., 1997, Signal Transduction via glycosylphosphatidylinositol-anchored proteins in T Cells is inhibited bylowering cellular cholesterol. Journal of Biological Chemistry, 272,19242 ± 19247.

Taraboulos, A., Scott, M., Semenov, A., Avraham, D., Laszlo, L. andPrusiner, S. B., 1995, Cholesterol depletion and modification ofCOOH-terminal targeting sequence of the prion protein inhibitformation of the scrapie isoform. Journal of Cell Biology, 129,121 ± 132.

Tienari, P. J., Ida, N., Ikonen, E., Simons, M., Weidemann, A.,Multhaup, G., Masters, C. L., Dotti, C. G. and Beyreuther, K.,1997, Intracellular and secreted Alzheimer b -amyloid species aregenerated by distinct mechanisms in cultured hippocampalneurons. Proceedings of the National Academy of SciencesUSA, 94, 4125 ± 4130.

155Membrane domains, lipid rafts and caveolaeM

ol M

embr

Bio

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f R

egin

a on

04/

16/1

3Fo

r pe

rson

al u

se o

nly.

Tran, D. J., Carpentier, J. L., Sawano, F., Gorden, P. and Orci, L.,1987, Ligands internalised through coated or non-coated invagi-nations followa common intracellular pathway. Proceedings of theNational Academy of Sciences USA, 84, 7957 ± 7961.

Varma, R. and Mayor, S., 1998, GPI-anchored proteins areorganised in submicron domains at the cell surface. Nature,394, 798 ± 801.

Vogel, U., Sandvig, K. and van Deurs, B., 1998, Expression ofcaveolin-1 and polarised formation of invaginated caveolae inCaco-2 and MDCK II cells. Journal of Cell Science, 111, 825 ±832.

Weik, M., Patzelt, H., Zaccai, G. and Oesterhelt, D., 1998,Localisation of glycolipids in membranes by in vivo labelling andneutron diffraction. Molecular Cell, 1, 411 ± 419.

Weimbs, T., Low, S. H., Chapin, S. J. and Mostov, K. E., 1997,Apical targeting in polarised epithelial cells: there’s more afloatthan rafts. Trends in Cell Biology, 7, 393 ± 399.

Wu, C., Butz, S., Ying, Y.-S. and Anderson, R. G. W., 1997, Tyrosinekinase receptors concentrated in caveolae-like domains fromneuronal plasma membrane. Journal of Biological Chemistry, 272,3554 ± 3559.

Xiao, Z. and Devreotes, P. N., 1997, Identification of detergent-resistant plasma membrane microdomains in dictyostelium:enrichment of signal transduction proteins. Molecular Biology ofthe Cell, 8, 855 ± 869.

Xie, M. and Low, M. G., 1995, Streptolysin-O induces release ofglycosylphosphatidylinositol-anchored alkaline phosphatase fromROS cells by vesiculation independently of phospholipase action.Biochemical Journal, 305, 529 ± 537.

Yamada, E., 1955, The fine structure of the gall bladder epithelium ofthe mouse. Journal of Biophysical and Biochemical Cytology, 1,445 ± 458.

Zegers, M. M. P. and Hoekstra, D., 1998, Mechanisms andfunctional features of polarised membrane traffic in epithelial andhepatic cells. Biochemical Journal, 336, 257 ± 269.

Received 11 December 1998, and in revised form 20 January 1999.

156 N. M. HooperM

ol M

embr

Bio

l Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f R

egin

a on

04/

16/1

3Fo

r pe

rson

al u

se o

nly.