FEBS Letters 584 (2010) 1779–1786

journal homepage: www.FEBSLetters .org

Review

Transbilayer (flip-flop) lipid motion and lipid scrambling in membranes

F.-Xabier Contreras a, Lissete Sánchez-Magraner b, Alicia Alonso c, Félix M. Goñi c,*

a Heidelberg University Biochemistry Center, 69120 Heidelberg, Germanyb Department of Biotechnology, University of Verona, 37134 Verona, Italyc Unidad de Biofísica (Centro Mixto CSIC-UPV/EHU), Departamento de Bioquímica, Universidad del País Vasco, P.O. Box 644, 48080 Bilbao, Spain

a r t i c l e i n f o a b s t r a c t

Article history:Received 19 October 2009Accepted 18 December 2009Available online 30 December 2009

Edited by Wilhelm Just

Keywords:Transmembrane lipid motionLipid flip-flopBilayer scramblingCeramideDiacylglycerol

0014-5793/$36.00 � 2009 Federation of European Biodoi:10.1016/j.febslet.2009.12.049

* Corresponding author. Fax: +34 94 601 33 60.E-mail address: felix.goni@ehu.es (F.M. Goñi).

This paper reviews the current knowledge on the various mechanisms for transbilayer, or flip-flop,lipid motion in model and cell membranes, enzyme-assisted lipid transfer by flippases, floppasesand scramblases is briefly discussed, while non-catalyzed lipid flip-flop is reviewed in more detail.Transbilayer lipid motion may occur as a result of the insertion of foreign molecules (detergents, lip-ids, or even proteins) in one of the membrane leaflets. It may also be the result of the enzymatic gen-eration of lipids, e.g. diacylglycerol or ceramide, at one side of the membrane. Transbilayer motionrates decrease in the order diacylglycerol� ceramide� phospholipids. Ceramide, but not diacyl-glycerol, can induce transbilayer motion of other lipids, and bilayer scrambling. Transbilayer lipiddiffusion and bilayer scrambling are defined as two conceptually and mechanistically different pro-cesses. The mechanism of scrambling appears to be related to local instabilities caused by the non-lamellar ceramide molecule, or by other molecules that exhibit a relatively slow flip-flop rate, whenasymmetrically inserted or generated in one of the monolayers in a cell or model membrane.� 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction

As in all living structures, the molecules forming eukaryotic cellmembranes are in a non-equilibrium state. This is mainly due to acontinuous intracellular trafficking of molecules that control cellsurvival. In the last decades, a great deal of information has beengathered concerning membrane protein synthesis, transport, mat-uration and distribution into the target membrane [1]. However,the means by which membrane lipids are transported to their finaldestination after synthesis in the endoplasmic reticulum (ER) isstill to be clarified. Cellular membranes are composed of a doublelayer of lipids in which proteins are embedded. At least some bio-logical membranes, and particularly the plasma membrane (PM),are known to be asymmetric and composed of lipids that are het-erogeneously distributed in the different organelles. This specificmembrane lipid composition requires the transbilayer movementof lipids from one leaflet to the other. Phospholipid synthesis takesplace almost exclusively in the cytoplasmic leaflet of the ER, whichleads to an asymmetry in lipid composition. In order to maintainthe balance of lipid molecules in the ER, about one-half of the new-ly synthesized lipids are transferred to the other leaflet [2].

The PM is characterized by a strict lipid asymmetry in whichphosphatidylcholine (PC) and sphingomyelin (SM) are mainlylocalized in the extracellular leaflet, whereas phosphatidylethanol-

chemical Societies. Published by E

amine (PE) and phosphatidylserine (PS) are exclusively present inthe cytoplasmic leaflet. Maintenance of the lipid asymmetry inthe PM is an energy-dependent process that is important for a vari-ety of cell processes [3]. For example, PS exposure on the cell sur-face leaflet after certain stimuli has been associated with apoptosisand senescence [4,5]. It is thought that the recognition of exposedPS by macrophages is necessary for the engulfment of apoptoticcells [6]. It is also reported that PS exposure is observed in vascularcells in tumors but not in normal cells [7].

Transbilayer lipid motion has been dubbed flip-flop and it isunfortunate that this rather childish term has stayed in the scientificliterature, together with the flippase and floppase terms that desig-nate lipid translocating enzymes. Maintenance of membraneasymmetry in the resting state cannot be mediated by passive lipidflip-flop alone, since spontaneous translocation of phospholipidsacross the bilayer has been postulated to be extremely slow(10�15 s�1), on average a single lipid flip-flop event every 24 h[8,9]. The slow pace of this translocation is due to the high energybarrier imposed by the resistance to the passage of polar head groupsthrough the hydrophobic core but also to the additional resistanceimposed by the increased lateral tension developed in the receivingmonolayer upon insertion of the new phospholipids [10]. Therefore,asymmetric membranes cannot be maintained by passive flip-flops.Establishment and maintenance of lipid asymmetry involves severaldifferent membrane proteins. These lipid translocators include ATP-dependent flippases and floppases [10,11] and energy-independentscramblases [11]. The energy-dependent translocators have been

lsevier B.V. All rights reserved.

1780 F.-X. Contreras et al. / FEBS Letters 584 (2010) 1779–1786

described as lipid selective, while scramblases are non-selective,inducing non-specific transbilayer movement across the membrane[11]. We intend to briefly review the proteins catalyzing flip-flop,and deal in more detail with the less understood phenomenon oftransbilayer lipid motion induced by molecules that were not origi-nally present in the membrane.

2. Flippases and floppases

Flippases and floppases are ATP-dependent transbilayer lipidtranslocators. According to an extensively used, though not univer-sal, nomenclature they catalyze lipid transfer towards the inwardmonolayer (flippase) or towards the outward monolayer (floppas-es). Some authors however use the name floppase to denote anyenzyme catalyzing transbilayer lipid motion, in some cases evenin the absence of ATP requirements.

Flippases and floppases are required to equalize the number oflipids at both sides of the membrane when a new membrane/orga-nelle is being generated, lipid synthesis taking place only on oneside of the bilayer. Thus flippases and floppases help make themembrane more symmetric. Moreover in some cases they operatein the opposite way, generating lipid asymmetry, the latter givingrise in turn to lipid vesiculation. In general, this is a field in whichmany of even the basic aspects remain unknown [12]. Most pre-dicted flippases have not yet been characterized at the molecularlevel, and the few that have been identified by genetic methodshave not been biochemically characterized [13].

An important breakthrough in this area has been the involve-ment of the so-called P4-ATPases (type 4 P-type ATPases) in thephospholipid translocation from the exoplasmic to the cytoplasmicleaflet of cell membranes (most other P-type ATPases catalyze cat-ion transport across membranes). The b-subunit of P4-ATPasesshows similarities with the b and c subunits of Na+, K+-ATPases[12]. Members of the P4-ATPase group have been identified in Sac-charomyces cerevisiae, Caenorhabditis elegans, Arabidopsis thalianaand also in humans. As many as 14 P4-ATPases have been foundin our species [14]. ATP8B1 is a phosphatidylserine translocase thatis probably the most extensively studied mammalian P4-ATPase.This enzyme contributes to the maintenance of asymmetry in plas-ma membranes. Mutations in the ATP8B1 gene cause severe liverdisease, however the cellular function and biochemical activity ofmost mammalian P4 ATPases remains to be clarified. It should benoted though that no P4-ATPases have been identified in the ER,where most membrane lipids are synthesized thus the prime locusfor transbilayer lipid motion to occur. The subject of P4-ATPasesand their role as flippases in health and disease has been recentlyreviewed by Folmer et al. [15].

A particular case of lipid flipping in the endoplasmic reticu-lum is that of the glycosyldolichols. Dolichols are a group of avery long-chain unsaturated isoprene lipids. Protein N-glycosyla-tion occurs in the lumen of the ER, and the glycosyl units areprovided by glycosyldolichols synthesized on the cytoplasmicface or ER. Of course flipping of these glycolipids must occur,but the mechanism is unknown, except that specific flippasesare required [16].

Lipid flippases are also involved in a series of mechanisms thatlead to changes in the membrane physical properties. This aspectof the enzymes has been reviewed by Devaux et al. [10]. Phospho-lipid transfer may lead to local or extended cell shape changesthrough bending of the lipid bilayer. Bending occurs through theinbalance of lipids between the two monolayers, and sometimestension may be released through vesiculation, i.e. the budding ofnew, small vesicles. The newly-formed vesicles may in turn be in-volved in physiological events such as lipid and protein traffic.Muthusamy et al. [14] have recently described the involvementof yeast P4-ATPases in transport vesicle budding.

Floppases catalyze the exofacially oriented transport of lipidsin membranes. Even less is known of this group of enzymes thanof the above-mentioned flippases. The floppases known to dateare all members of the ABC (ATP-binding cassette) superfamily,a group of proteins that were originally related to multidrugresistance. The fact that they are rather non-specific transloca-tors of largely non-polar molecules explains their floppaseactivity [11,17]. In fact the best characterized floppase is themultidrug resistance protein ABCB1, or P-glycoprotein, but evenin this case it is doubtful that it translocates naturallong-chain lipids, as opposed to fluorescent short-chain ana-logues [18].

3. Scramblases: the human PLSCR1

Phospholipid scramblases (PLSCRs) constitute a group ofhomologous ATP-independent bidirectional lipid translocators thatare conserved in all eukaryotic organisms, from yeasts to humans.They are involved in the calcium-dependent transbilayer move-ment of phospholipids, which tends to eliminate their asymmetricdistribution across the membranes [19].

In humans, four related PLSCR genes have been identified,named as hPLSCR1–hPLSCR4. The predicted open reading framesencode proteins with 59% (hPLSCR2), 47% (hPLSCR3) and 46%(hPLSCR4) homology, respectively, to hPLSCR1 [20]. In the cellsPLSCR1 is located in the plasma membrane, PLSCR2 is found inthe nucleus, and PLSCR3 occurs in the mitochondrial membrane.While hPLSCR1, 3 and 4 are expressed in a variety of tissues includ-ing heart, kidney, pancreas, spleen, prostate, colon, the expressionof hPLSCR2 is restricted to testis. hPLSCR4 was not detected inperipheral blood lymphocytes, and hPLSCR1 and hPLSCR3 werenot detected in the brain [20,21].

The first described member and prototype of this family ishPLSCR1, a 37 kDa (318 aa) type II endofacial membrane proteinwith a single transmembrane segment near the C-terminus.hPLSCR1 remains inactive under normal conditions, but increasedcytosolic calcium concentrations (�1000-fold the basal level in re-sponse to cell injury, coagulation, cell activation or apoptosis) acti-vate the protein and induce a rapid transbilayer movement of PEand PS from the inner to the outer membrane leaflet destroyingthe bilayer asymmetry [19,20,22]. PLSCR1 is a palmitoylated pro-tein that partitions with flotillin-1 and EGFR into lipid rafts. EGFstimulates tyrosine phosphorylation of PLSCR1 and promotesinternalization of the protein [23]. In the absence of palmitoyla-tion, virtually all of the expressed hPLSCR1 localizes to the nucleus[20].

hPLSCR1 has a variety of important functional regions [24]. TheN-terminal region (aa 1–85) contains multiple PXXP and PPXY se-quences, which may bind proteins containing SH3 and WW do-mains, respectively [20]. The DNA binding domain (aa 86–118)enables the protein to interact with DNA. In the absence of palmi-toylation hPLSCR1 is imported into the nucleus in an importin a/b-,GTP-, and Ran-dependent manner and binds to a genomic DNAwith a higher affinity. This suggests a potential nuclear functionas a transcription factor within the nucleus for the unpalmitoylat-ed protein. One identified gene target of nuclear hPLSCR1 is theinositol-1,4,5-trisphosphate receptor type 1 gene (IP3R1), nuclearhPLSCR1 serving to increase both IP3R1 transcription and proteinexpression [25,26]. Recent studies suggest the interaction ofhPLSCR1 with topoisomerase II, a protein involved in numerouscellular processes and essential for cell division [27].

Another important region is the cystein-palmitoylation motif(184CCCPCC189). The protein is acylated by thioester linkages tothe sulfhydryl groups of one or more cysteines, and this regulatesits trafficking within the cell. This modification is essential for the

F.-X. Contreras et al. / FEBS Letters 584 (2010) 1779–1786 1781

hPLSCR1 scrambling activity providing a membrane anchorage forthe N-terminal region. Hydrolysis of the thioester-bound palmitoylgroups in hPLSCR1 purified from RBC markedly reduced the scram-bling activity [28,29], and the unmodified hPLSCR1 expressed andpurified from Escherichia coli had approximately 50% of the activityobserved for the endogenous protein purified from RBC [20]. Nu-clear localization of hPLSCR1 was also observed upon inductionof its de novo synthesis by interferon a (IFNa). These data suggestthat the post-translational acylation acts as a check-point deter-mining the localization of the protein in the cell and regulatingits normal function in the nucleus or incorporated to the mem-brane [28,29]. A recent structural model computed by homologymodeling suggests that palmitoylation may represent the principalmembrane anchorage for hPLSCR1 [30].

The nuclear localization of hPLSCR1 depends on cytosolic fac-tors and energy. The protein has a non-classical nuclear localiza-tion signal (NLS) (aa 257–266) which helps import the proteininto the nucleus by the importin a/b nuclear chaperones. ThisNLS is distinct from the prototypical NLS defined by a cluster of fivebasic residues, in this sequence only three out of nine amino acidsare basic (two lysines separated by two neutral amino acids andone histidine) and it is enriched in hydrophobic residues [25].hPLSCR1 interacts with importin-a with a high affinity, importin-a acting as an adaptor that recognizes and binds the scramblaseNLS sequence after association with importin-b. Then the complexis translocated through nuclear pores. The crystal structure ofimportin-a/scramblase/NLS complex has been determined at2.2 Å, the structure suggests that NLS binding is not accompaniedby a significant conformational change in importin-a, and that itsinteraction partially overlaps the classical NLS binding site [31].

The calcium binding motif (aa 273–284) with a putative EF-hand-like structure is essential for the scrambling activity ofhPLSCR1. This motif is known to bind calcium and other ions(e.g. Tb+3, La+3, Mn+2, Zn+2, Sr+2, Mg+2). In this motif residues inpositions 1, 3, 5, 7, 9 and 12 contribute to the octahedral coordina-tion of the calcium ion and substitution at any of these positionsreduces phospholipid scramblase function [32,33]. Calcium bind-ing induces protein oligomerization, either by cross-linkinghPLSCR1 monomers or by inducing a prominent conformationalchange that in turn promotes self-association. It remains to bedetermined how the conformational change serves to acceleratethe movement of membrane phospholipids between the innerand outer leaflets.

A recent structural model suggests that the calcium bindingmotif is not an EF-hand-like structure [30]. Studies of Sahu et al.suggest that the calcium binding motif present in scramblase rep-resents a new class of low-affinity calcium binding motifs, thescramblase-like calcium binding motif being also present in otherunrelated proteins known to exhibit changes in biochemicalparameters upon binding to calcium or other divalent ions [34].

hPLSCR1 structure is completed by the single predicted trans-membrane helix near the C-terminus (aa 291–309) with the in-side-to-outside orientation, that makes the hPLSCR1 a type IImembrane protein, and by a final short exoplasmic tail (aa 310–318) [21,24]. Bateman et al. provide an alternative structural mod-el in which the transmembrane helix is buried within the core ofthe hPLSCR1 structure, suggesting that this is not a transmembranehelix [30].

In addition to palmitoylation and calcium binding hPLSCR1undergoes a different post-translational modification that affectsprotein function, namely phosphorylation of Tyr and Thr residues,Tyr phosphorylation by c-Abl and c-Src tyrosine kinases at Tyr 69and 74 and Thr phosphorylation by PKCd at Thr 161 [21,23].

The actual cellular functions of hPLSCR1 remain largely unre-solved, and different studies suggest that in addition to its putativerole in mediating transbilayer movement of phospholipids,

hPLSCR1 may also function to regulate processes including signal-ing, cell differentiation, apoptosis, injury, cell proliferation andtranscription [25,26,35–37].

4. Transbilayer lipid movement induced by external agents

External addition of detergents to the plasma membrane of redblood cells at sub-solubilizing concentrations accelerates the slowtransbilayer motion of fluorescent-labeled PC analogs (NBD-PC)across the membrane [38]. Interestingly neutral, zwitterionic, an-ionic and cationic detergents containing different alkyl chainlengths (from C6 to C14) are able to accelerate the transfer of lipidsfrom the outer to the inner leaflet. It is important to mention thatthe ability of a detergent to induce this effect on the bilayer is com-pletely independent of its alkyl chain length. This detergent effectis likely to be induced by formation of transient hydrophobic struc-tural defects in the membrane barrier, which can be the result ofperturbation in the interfacial region of the bilayer associated withdetergent insertion. Moreover, the concentration of different deter-gents required to increase the constant rate of NBD-PC flip is verysimilar. Therefore the presence of a small neutral hydroxyl group, alarger sugar residue or an ionic group linked to detergent alkylchain is not essential to obtain this detergent accelerating effecton the transbilayer lipid motion. The membrane concentrationsof non-ionic detergents (i.e. Triton X-100) needed to induce an ob-servable flip of NBD-PC is around 10-fold lower than that observedwith other detergents. Reduction in the length of the oxyethylenechain is correlated with a decrease in effectiveness [38].

Our group has also described that alkanes promote the translo-cation of lipids between leaflets [39a]. Alkanes are often used in cellbiology as vehicles to incorporate hydrophobic molecules into cellmembranes. It is generally assumed that these molecules are inac-tive when used at low concentrations. However, under standardexperimental conditions, we observed that alkanes induce drasticchanges in the physical state of the membrane. Insertion of alkanesinto the outer leaflet is accompanied by a rapid transbilayer motionof lipids across the membrane. As described for detergents above,acceleration of lipid flip-flop may be induced by the transient for-mation of defects in the bilayer. Since alkanes stabilize and promotenon-lamellar structures in the membrane, a lipid mismatching de-fect can be created in the borders between lamellar and non-lamel-lar structures, facilitating lipid translocation within the bilayer.

Protein insertion into membranes may also cause flip-flop lipidmotion. Martin et al. [39b] studied the insertion of the so-called‘adenylate cyclase toxin’ from Bordetella pertussis into lipid vesi-cles. The toxin (1706 aa residues) targets cell membranes, in whichit breaks down the permeability barrier. The NBD-PC test, as men-tioned above, was used to demonstrate that as the toxin becameincorporated into the vesicles (and efflux of liposomal aqueouscontents took place) extensive flip-flop motions occurred in thebilayer. Similar observations have been made with proapoptoticBax-type proteins in liposomes [39c,d]. This may be a generalphenomenon for proteins that are not under native conditions partof the cell membrane, but instead become incorporated into itunder certain circumstances.

5. Spontaneous transbilayer diffusion of biological lipid secondmessengers

Important examples of lipid molecules that undergo a rapidtransbilayer motion, and can even induce flip-flop of other lipids,are the bioactive molecules diacylglycerol and ceramide. Althoughtheir average concentration in membranes is very low, it may in-crease locally by orders of magnitude as a result of cellular activa-tion by the appropriate agonist.

1782 F.-X. Contreras et al. / FEBS Letters 584 (2010) 1779–1786

5.1. Diacylglycerol

Diacylglycerol (DAG), a minor component of cellular membranes,plays major roles in biological processes. An important second mes-senger molecule, DAG is involved in cell signal transduction [40]. InDAG a fatty acyl chain is linked at position sn-2 via an ester linkage,however at position sn-1 the fatty acid can be linked via ester, etheror alkenyl ether bonds [41]. DAG can exist in at least 50 differentmolecular species, containing saturated, mono-unsaturated, di-unsaturated or poly-unsaturated fatty acids [42,43]. Combinationsof different fatty acids in position sn-1 and sn-2 create a variety of dif-ferent DAG species. It is still unclear which DAG species play a mostimportant role in cell signaling. Various studies link an increase inthe degree of unsaturation with a greater activating effect in the cell[44,45]. It is also still unclear to what extent the membrane structureis altered after formation of DAG by phospholipase C. Different bio-physical studies in model membranes have shown that the presenceof DAG causes a change from lamellar to hexagonal structure in thebilayer [46,47]. This transition to non-lamellar structure depends onthe amount of DAG in the membrane. In addition to the lamellar tonon-lamellar phase transition of lipids, DAG is also involved in fusionand fission processes [48,49].

Pagano and Longmuir explored the ability of fluorescent-labeledDAG to freely diffuse across the membrane [50]. In their studies theplasma membrane of Chinese hamster fibroblasts was labeled withfluorescent phosphatidic acid (NBD-PA). Upon dephosphorylation,the resulting product NBD-DAG was readily distributed into cyto-plasmic membranes. The labeling of intracellular membranes couldonly be due to a transbilayer movement of DAG within the plasmamembrane. Labeling of the cells directly with NBD-DAG or enzy-matic production of fluorescent DAG from NBD-PC upon phospho-lipase C addition led to a similar intracellular fluorescent DAGdistribution. To rule out the involvement of cytosolic proteins thatmay directly transport the lipid into the cytosolic membranes and apossible lateral diffusion of the NBD-DAG along cytoplasmic mem-branes in contact with or in close proximity of the plasma mem-brane, these authors observed in artificial lipid vesicles, mainlycomposed of DOPC, a direct transbilayer movement of NBD-DAG,but not NBD-PA. These results are in agreement with the transbilay-er movement of endogenous DAG described previously in humanerythrocytes [51]. Additional studies in vesicles using 13C NMRspectroscopy and other physical techniques show that the transbi-layer movement of DAG compared to phospholipids is extremelyfast. Predictions from the chemical shift of the carbonyl 1,2-DAGsignal in PC vesicles containing 20% DAG estimated a DAG transbi-layer movement rate of �63 s�1 (t1/2 � 11 ms at 38 �C) [52].Decreasing DAG concentration in the membrane accelerates the ex-change rate of the molecule between the leaflets.

Other studies using chemical tools estimate that the flip-flop ofDAG occurs in the range of seconds [53] to 1 min [51], in any caseseveral orders of magnitude faster than that observed for phospho-lipids. Compared to other hydrophobic molecules (i.e. alkanes,detergents or proteins) that induce the scrambling of other lipidsafter insertion in the membrane, it should be noted that DAG, whichcan freely diffuse through the membrane and change its physicalproperties, does not facilitate the transbilayer movement of otherlipids. Studies from our group have shown that increasing DAG levelsin the outer leaflet either via phospholipase C treatment or by exter-nal addition of DAG to preexisting vesicles does not induce transbi-layer translocation of fluorescent-labeled lipids within themembrane [54].

5.2. Ceramide

Ceramide represents the backbone structure of most sphingoli-pids and it also exists in free form, at low concentrations, in the cell

membranes. This small sphingolipid is produced in cell mem-branes via direct degradation of sphingomyelin (SM) or ganglio-sides. Recently much attention has focused on ceramide as apotent biological molecule involved in several biochemical and cel-lular responses to stress including, among others, apoptosis, senes-cence, cell cycle arrest and differentiation [55,56]. In the last twodecades the role of ceramide in apoptosis has been intensivelystudied. Results from different groups clearly show that ceramideis involved in apoptosis, in that ceramide accumulation regulatesthe execution phase of apoptosis involving mitochondria [55].

Recent data suggests that ceramide may act in changing thephysical state of the membrane. Various studies show that forma-tion of ceramide in the membrane leads to the segregation of thislipid into Cer-rich domains forming large platforms or rafts inwhich different cell surface receptors oligomerize (i.e. Fas, CD40).Formation of Cer-enriched domains has been intensively studiedin model membranes [57]. These rigid domains can be explainedby the strong inter-molecular interaction between ceramidemolecules, forming a compact membrane structure. London andco-workers show that formation of Cer-rigid domains induces thesegregation of cholesterol from rafts [58,59]. The authors observedthat ceramide and cholesterol, in liquid-ordered (Lo) mixtures,compete for SM-rich domains [60]. Additional studies describedthe requirement of at least 8 methylene groups in the ceramideN-acyl chain to segregate sterols from Lo domains (e.g. [61]). Thischolesterol displacement enables the formation of SM/Cer rigidgel domains [62]. The stability of SM/Cer interactions in domainshas been intensively studied using electron spin resonance (ESR)[63] and confocal fluorescence microscopy [64].

Another defining characteristic of ceramide is the ability to af-fect general membrane curvature and stability, probably becauseof its intrinsic negative curvature, that facilitates formation ofnon-lamellar structures [46,65]. It is thought that via the transientformation of non-bilayer intermediates, ceramide induces vesicleefflux [66], membrane fusion [67] and vesicle budding [68]. Sinceceramide production takes place in the extracellular leaflet, viaan exported sphingomyelinase, while proteins that interact withceramide are intracellular, several studies have emphasized theability of ceramide to diffuse through the membrane. For severalyears a controversy has existed concerning the ability of naturalceramide to freely diffuse across the membrane. Several groups re-port that ceramide, after formation, is strictly localized for a con-siderable time in the leaflet where it is synthesized [69]. Otherstudies postulate that the half-life for spontaneous bilayer transferof a BODIPY-Ceramide in PC vesicles is 22 min at 37 �C [70]. Com-pared to other lipid analogs, labeled with the same fluorophore, thespontaneous ceramide transfer across the membrane is faster thanthat measured for other lipids containing a more bulky polargroup. Taking into account that the presence of a bulky fluores-cence group, more polar than a fatty acid, can influence the rateof transbilayer lipid movement in the membrane, observation ofunlabeled-ceramide diffusion across the membrane has been nec-essary to understand the influence of the fluorophore on transbi-layer movement.

Recently, Devaux’s group investigated the transmembrane dif-fusion of different unlabelled ceramide species in giant unilamellarvesicles (GUVs) [71]. Ceramide was incorporated into the outerleaflet of GUVs and the transbilayer movement of ceramide inthe membrane was followed monitoring the GUV shape changes.Note that insertion of a small fraction (�1%) of exogenous lipidin one of the membrane lipid monolayers can lead to clearly visiblechanges in cell or vesicle shape [72]. Using this technique, the half-life of ceramide flip-flop was found to be 30 s at 37 �C, one order ofmagnitude lower than that observed with fluorescence analogs.Similar half-lives for ceramide diffusion were observed, usingspin-labeled ceramide analogues, in vesicles s1/2 = 1.1 min at

F.-X. Contreras et al. / FEBS Letters 584 (2010) 1779–1786 1783

20 �C) and human erythrocyte membranes s1/2 = 1.1 min at 20 �C).Interestingly short-chain ceramide, compared to long-chain cera-mide, can easily diffuse across erythrocyte membranes.

An additional study in GUVs containing liquid-ordered (Lo) andliquid-disordered (Ld) domains, after SM to ceramide conversionin situ by sphingomyelinase (SMase), shows a drastic decrease inLo domains followed by an increase in Cer-rich gel domains andthe generation of transitory nanoscale defects [73]. Formation ofstructural defects is the result of an increase in the lateral surfacetension induced by the enzymatic activity of SMase in one leaflet.This surface tension, produced by the asymmetric ceramide forma-tion in one leaflet, can in turn induce shape changes and transitorymembrane rupture. Our group has also reported that ceramide for-mation by SMase activity in one side of the membrane induces thetransbilayer lipid movement of other lipids in vesicle and erythro-cyte membranes [74]. SMase activity induces the flip-flop not onlyof fluorescence labeled phospholipids (PC, PE, PS), but also of nat-ural gangliosides (i.e. GM3) containing a bulky polar group. Gangli-oside diffusion across the membrane was investigated using anovel method in which vesicles containing neuraminidase trappedinside the liposomes were labeled on the outer leaflet with GM3.After SMase addition and SM hydrolysis, GM3 flops to the innerleaflet, to be hydrolyzed by neuraminidase (Fig. 1). During this pro-cedure no neuraminidase efflux from the vesicles was observed. Inthis experiment ceramide formation (and presumably transbilayermotion) facilitates the translocation of other lipids.

Fig. 1. Flip-flop of gangliosides induced by sphingomyelinase activity in largeunilamellar vesicles. (A) Sphingomyelin hydrolysis by sphingomyelinase. Averagevalues ±S.E. (n = 4). (B) GM3 ganglioside hydrolysis by entrapped neuraminidase.Average values ±S.E. (n = 5). Vesicle composition was SM:PE:Ch (2:1:1) + 10 mol%GM3 on the outer leaflet only [75].

In our opinion, SMase induces the transbilayer movement oflipids due, probably, to the intrinsic negative curvature of ceramideand subsequent ability to promote non-lamellar phase structuresin the membrane. The presence of a non-lamellar lipid such asCer in one of the membrane monolayers destabilizes locally the bi-layer. Cer could even induce transient non-lamellar structurescoexisting in a short time frame with lamellar structures in themembrane; defects transiently present between both structurescan facilitate the motion of lipids across the leaflets leading to amore symmetric lipid distribution. The transient membrane de-fects induced by ceramide formation in one leaflet have also beensuggested by other investigators [73]. It remains to be clarified ifthe transient nanoscale pores observed by Devaux and co-workersmostly at the interface between Lo and Ld domains, in the presenceof Cer-rich domains, and the corresponding lipid packaging defectsbetween both domains can facilitate the transbilayer movement ofother lipids in the membrane. Unfortunately observation of transi-tory non-lamellar states or more stable transient pores induced byceramide using electron microscopy has not yet been possible.

In a parallel study we investigated the ability of ceramide,added externally to pre-formed vesicles containing pyrene-la-belled PC (pyr-PC) in the outer leaflet, to promote this kind of lipidtranslocation [54,74]. Insertion of long-chain ceramide into theouter leaflet induces the redistribution of pyr-PC into both mono-layers resulting in an increase in monomer fluorescence intensityand a decrease in the excimer fluorescence intensity (Fig. 2). Inthe same study we demonstrated that short N-acyl chain cera-mides (C2 and C6) have a lower ability to trigger flip-flop betweenmonolayers. Short chain ceramides are used in cell biology to mi-mic some endogenous ceramide effects (i.e. apoptosis, cell differ-entiation), but are less potent than natural ceramides infacilitating the lamellar-to-inverse hexagonal phase transition oflipids [75]. This effect is related to the molecular geometry ofshort-chain ceramides. In the latter ceramides a short N-acyl-chainmismatches with the sphingosine backbone, leading to a smallerdiameter of its hydrophobic moiety section, compared to the diam-eter of the polar group section. Therefore short-chain ceramideshave a molecular shape that induces positive curvature of the mon-olayers, forming micelles that oppose the formation of invertedhexagonal phases in the bilayer [76]. In contrast, long-chain cera-mide has a shape that favors negative curvature and promotesthe formation of inverted hexagonal phases in the membrane. Itis important to stress that the low ability of short-chain ceramides

Fig. 2. Ceramide-induced transbilayer movement of pyr-PC in LUVs composed ofSM/PE/Ch. The original lipid concentration was 300 lM. The following additionswere made to 1 ml of the LUV suspension at time zero: 30 lM (final concentrationin buffer) egg ceramide in 2 ll ethanol (curve A), 2 ll ethanol (curve B), 1.6 Usphingomyelinase (curve C), and none (curve D). The data have been fitted tomonoexponential curves of the form y = a (1 � e�bx)c [54].

Fig. 3. Pore-mediated lipid flip-flop: (A) 0 ps, (B) 43.85 ns, (C) 118.9 ns, (D)122.4 ns, (E) 152.7 ns, (F) 204.65 ns, (G) 208.9 ns, and (H) 215 ns. Lipids (exceptfor the flip-flopped one) are not shown; water is shown in red and white, acylchains of the flip-flopped lipid are shown in yellow, and its choline and phosphategroups are shown in orange and green, respectively. (Taken from Gurtovenko, A.A.and Vattulainen, I. (2002) Molecular mechanism for lipid flip-flops. J. Phys. Chem. B.111, 13554–13559).

�C

opyr

igh

t�

2007

Am

eric

anC

hem

ical

Soci

ety

2007

1784 F.-X. Contreras et al. / FEBS Letters 584 (2010) 1779–1786

to promote lipid flip-flop in membranes is also connected withtheir less potent effect in ceramide-dependent signaling pathways(i.e. apoptosis or cell differentiation) [77]. Dihydroceramide, thesaturated ceramide counterpart and the precursor of ceramide dur-ing de novo synthesis, is unable to induce flip-flop motion. Thisinability is matched by the fact that the absence of the trans-D4

double bonds seems to transform the lipid into a less biologicallyactive molecule, incapable of reproducing most of the effects ob-served with the unsaturated ceramide, even though its uptakeand metabolism is similar [77]. To mention but one example, dihy-droceramide inhibits the apoptotic effect of ceramide in mitochon-dria [78].

6. Molecular dynamics simulations of lipid flip-flop

The mechanism of transmembrane diffusion of lipids in a pro-tein-free membrane is poorly understood. However, it is assumedthat the diffusion of a lipid in a membrane lacking proteins takesplace as a single molecular event in which nearby lipids play animportant role [79,80]. Perturbation of the membrane physicalproperties (i.e. protein insertion, ionic concentration imbalanceacross the membrane, enzymes) induces defects in the membranethat could allow the spontaneous formation of transitory waterpores, which would in turn facilitate the diffusion of lipids throughthe hydrophobic membrane core [81,82]. These nanoscale waterpores are the result of lipid mismatching within the membrane,and they could be related to the transient structural defects men-tioned in the previous section. As discussed above, ceramide pro-duction could induce the formation of such transitory waterpores as a result of the bilayer destabilization caused by the pres-ence of Cer in one of the monolayers.

Molecular dynamics simulations (MD) aid our understanding ofthe mechanism associated with lipid flip-flop [83]. MD simulationsthat mimic the physiological condition of the cell provide evidencethat formation of defects in the membrane can lead to the genera-tion of transient water pores. Once the water pore is created, thespontaneous translocation of lipids within the membrane can oc-cur (Fig. 3) [84]. Interestingly, the average time required for theflip-flop of a lipid across a water pore was found to be around60 ns. Together with the high number of flip-flop events observedduring the MD, these results suggest that translocation of a lipidacross the bilayer is a very fast process in which water pore forma-tion is the rate-limiting step. It has also been suggested that thetranslocation of one lipid mediated by this water pore can promotethe flip-flop of other lipids. In order to keep the pore stable, diffu-sion of one lipid to the opposite leaflet, at the edge of the pore, hasto be substituted by another lipid from the same leaflet. In theevent that this new lipid reaches the opposite leaflet, the substitu-tion process has to be repeated again. Moreover, from these simu-lations, the authors observed that the number of flip and flopevents were symmetric (similar number of flip and flop eventswere found). The ‘‘transient water pores” in the simulations byGurtovenko and Vattulainen [81–84] may have their physical cor-relates in the structural defects observed by López-Montero et al.[73], as suggested by the latter authors.

7. Flip-flop motion and bilayer scrambling

An overview of the results discussed above may clarify one as-pect of transbilayer lipid motion that has remained less clear in thepast, i.e. the difference between flip-flop, or transbilayer motion,and lipid scrambling in bilayers. The difference is both conceptualand mechanistic. Conceptually, transbilayer lipid motion refers tothe diffusion of single molecules, while lipid scrambling involvesnecessarily a population of lipid molecules. Mechanistically, a

single molecule may diffuse to the opposite bilayer spontaneously,as a result of thermal motion. This may be a relatively frequentevent, as with diacylglycerol, or a rare one, as with phospholipids.Transmembrane motion of a single lipid molecule may be enzyme-catalyzed. Lipid scrambling however is somewhat of a precipitousevent, which requires the building-up of a certain amount of en-ergy to occur.

The different effects of DAG and ceramide are rather clarifyingin this respect. Both molecules undergo spontaneous flip-flop mo-tion although at widely differing rates, about 1 min�1 for DAG butthree orders of magnitude slower for ceramide. DAG flip-flop is notaccompanied by bilayer scrambling, while both phenomena are de-tected with Cer.

One possible explanation, hinted at by Ruiz-Argüello et al. [67]years ago, is that because of the fast exchange DAG does not giverise to any significant membrane asymmetry, thus the process does

F.-X. Contreras et al. / FEBS Letters 584 (2010) 1779–1786 1785

not involve other than thermal energy and it is not energically cou-pled to any other event. Cer on the contrary has a much slower flip-flop rate. When Cer becomes incorporated in one of the membraneleaflets, a certain degree of membrane asymmetry occurs. Thisasymmetry is both chemical (ceramide concentration at both sidesof the membrane are different) and mechanical. See [85] for arecent review on ceramide-induced mechanical stress.

When ceramide, e.g. in organic solvent, is externally added toa membrane suspension, so that it gets incorporated into theouter monolayers, insertion of ‘‘foreign” ceramide molecules willlead to an imbalance of lateral tensions of the inner and outermonolayers, being higher in the outer monolayer. The same phe-nomenon, with an opposite sign, occurs when SM molecules inthe outer monolayers are hydrolyzed by a sphingomyelinase giv-ing rise to ceramide. In the latter case the outer leaflet lateraltension will decrease, because the area per molecule of ceramideis smaller than that of SM. Moreover, the propensity of ceramideto aggregate in the form of non-lamellar phases will also help indestabilizing the bilayer. In any case, the asymmetric distribu-tion of ceramide, which cannot be relaxed via rapid flip-flop, willgenerate a mechanical energy of bending, which will ultimatelylead to the local destabilization of the bilayer, and the resultinglipid scrambling. When ceramide is incorporated in the lipidmixture prior to liposome preparation, a symmetric distributionof ceramide molecules in the two monolayers is achieved, andno scrambling is observed under these conditions [54].

The same mechanism may explain the observed lipid scram-bling caused by the asymmetric insertion of proteins, or of somesurfactants, e.g. lysophosphatidylcholine, that are known to haveslow flip-flop rates. However, in the absence of specific flip-floprate values, it cannot be ascertained whether all detergents causescrambling via the same mechanism. In their case, the spontane-ous tendency to form lipid-detergent mixed micelles may providea specific mechanism for bilayer scrambling to occur.

Acknowledgements

This work was supported by grants from the Spanish Ministeriode Ciencia e Innovación No. BFU 2008-01637/BMC (A.A.), BFU2007-62062 (F.M.G.) and from the Basque Government (IT 461-07) (F.M.G.). F.X.C. is a postdoctoral fellow from Federation of Euro-pean Biochemical Societies. L.S.M. is supported by a postdoctoralgrant from the Basque Government.

References

[1] Schnell, D.J. and Hebert, D.N. (2003) Protein translocons: multifunctionalmediators of protein translocation across membranes. Cell 112, 491–505.

[2] Pomorski, T. and Menon, A.K. (2006) Lipid flippases and their biologicalfunctions. Cell. Mol. Life Sci. 63, 2908–2921.

[3] Seigneuret, M. and Devaux, P.F. (1984) ATP-dependent asymmetricdistribution of spin-labeled phospholipids in the erythrocyte membrane:relation to shape changes. Proc. Natl. Acad. Sci. USA 81, 3751–3755.

[4] Bevers, E.M., Comfurius, P. and Zwaal, R.F. (1996) Regulatory mechanisms inmaintenance and modulation of transmembrane lipid asymmetry:pathophysiological implications. Lupus 5, 480–487.

[5] Williamson, P. and Schlegel, R.A. (2002) Transbilayer phospholipid movementand the clearance of apoptotic cells. Biochim. Biophys. Acta 1585, 53–63.

[6] Fadok, V.A., Bratton, D.L., Rose, D.M., Pearson, A., Ezekewitz, R.A. and Henson,P.M. (2000) A receptor for phosphatidylserine-specific clearance of apoptoticcells. Nature 405, 85–90.

[7] Ran, S., Downes, A. and Thorpe, P.E. (2002) Increased exposure of anionicphospholipids on the surface of tumor blood vessels. Cancer Res. 62, 6132–6140.

[8] Kornberg, R.D. and McConnell, H.M. (1971) Inside–outside transitions ofphospholipids in vesicle membranes. Biochemistry 10, 1111–1120.

[9] Nakano, M., Fukuda, M., Kudo, T., Matsuzaki, N., Azuma, T., Sekine, K., Endo, H.and Handa, T. (2009) Flip-flop of phospholipids in vesicles: kinetic analysiswith time-resolved small-angle neutron scattering. J. Phys. Chem. B 113,6745–6748.

[10] Devaux, P.F., Herrmann, A., Ohlwein, N. and Kozlov, M.M. (2008) How lipidflippases can modulate membrane structure. Biochim. Biophys. Acta 1778,1591–1600.

[11] Daleke, D.L. (2003) Regulation of transbilayer plasma membrane phospholipidasymmetry. J. Lipid Res. 44, 233–242.

[12] Poulsen, L.R., López-Marqués, R.L. and Palmgren, M.G. (2008) Flippases: stillmore questions than answers. Cell. Mol. Life Sci. 65, 3119–3125.

[13] Sanyal, S. and Menon, A.K. (2009) Flipping lipids: why an what’s the reasonfor? ACS Chem. Biol. 4, 895–909.

[14] Muthusamy, B.P., Natarajan, P., Zhou, X. and Graham, T.R. (2009) Linkingphospholipid flippases to vesicle-mediated protein transport. Biochim.Biophys. Acta 1791, 612–619.

[15] Folmer, D.E., Elferink, R.P. and Paulusma, C.C. (2009) P4 ATPases – lipidflippases and their role in disease. Biochim. Biophys. Acta 1791, 628–635.

[16] Sanyal, S. and Menon, A.K. (2009) Specific transbilayer translocation ofdolichol-linked oligosaccharides by an endoplasmic reticulum flippase. Proc.Natl. Acad. Sci. USA 106, 767–772.

[17] Paulusma, C.C. and Oude Elferink, R.P. (2006) Diseases of intramembranouslipid transport. FEBS Lett. 580, 5500–5509.

[18] Halter, D., Neumann, S., van Dijk, S.M., Wolthoorn, J., de Mazière, A.M., Vieira,O.V., Mattjus, P., Klumperman, J., van Meer, G. and Sprong, H. (2007) Pre- andpost-Golgi translocation of glucosylceramide in glycosphingolipid synthesis. J.Cell Biol. 179, 101–115.

[19] Basse, F., Stout, J.G., Sims, P.J. and Wiedmer, T. (1996) Isolation of anerythrocyte membrane protein that mediates Ca2+-dependent transbilayermovement of phospholipid. J. Biol. Chem. 271, 17205–17210.

[20] Wiedmer, T., Zhou, Q., Kwoh, D.Y. and Sims, P.J. (2000) Identification of threenew members of the phospholipid scramblase gene family. Biochim. Biophys.Acta 1467, 244–253.

[21] Zhou, Q., Zhao, J., Stout, J.G., Luhm, R.A., Wiedmer, T. and Sims, P.J. (1997)Molecular cloning of human plasma membrane phospholipid scramblase. Aprotein mediating transbilayer movement of plasma membranephospholipids. J. Biol. Chem. 272, 18240–18244.

[22] Zhao, J., Zhou, Q., Wiedmer, T. and Sims, P.J. (1998) Level of expression ofphospholipid scramblase regulates induced movement of phosphatidylserineto the cell surface. J. Biol. Chem. 273, 6603–6606.

[23] Sun, J., Zhao, J., Schwart, M.A., Wang, J., Wiedmer, T. and Sims, P.J. (2001) c-Abltyrosine kinase binds and phosphorylates phospholipid scramblase 1. J. Biol.Chem. 276, 28984–28990.

[24] Sims, P.J. and Wiedmer, T. (2001) Unraveling the mysteries of phospholipidscrambling. Thromb. Haemost. 86, 266–275.

[25] Ben-Efraim, I., Zhou, Q., Wiedmer, T., Gerace, L. and Sims, P.J. (2004)Phospholipid scramblase 1 is imported into the nucleus by a receptor-mediated pathway and interacts with DNA. Biochemistry 43, 3518–3526.

[26] Zhou, Q., Ben-Efraim, I., Bigcas, J.L., Junqueira, D., Wiedmer, T. and Sims, P.J.(2005) Phospholipid scramblase 1 binds to the promoter region of the inositol1,4,5-triphosphate receptor type 1 gene to enhance its expression. J. Biol.Chem. 280, 35062–35068.

[27] Wyles, J.P., Wu, Z., Mirski, S. and Cole, S. (2007) Nuclear interactions oftopoisomerase II alpha and beta with phospholipid scramblase 1. NucleicAcids Res. 35, 4076–4085.

[28] Zhao, J., Zhou, Q., Wiedmer, T. and Sims, P.J. (1998) Palmitoylation ofphospholipid scramblase is required for normal function in promoting Ca2+-activated transbilayer movement of membrane phospholipids. Biochemistry37, 6361–6366.

[29] Wiedmer, T., Zhao, J., Nanjundan, M. and Sims, P.J. (2003) Palmitoylation ofphospholipid scramblase 1 controls its distribution between nucleus andplasma membrane. Biochemistry 42, 1227–1233.

[30] Bateman, A., Finn, R.D., Sims, P.J., Wiedmer, T., Biegert, A. and Soding, J. (2009)Phospholipid scramblases and Tubby-like proteins belong to a newsuperfamily of membrane tethered transcription factors. Bioinformatics 25,159–162.

[31] Chen, M.H., Ben-Efraim, I., Mitrousis, G., Walker-Kopp, N., Sims, P.J. andCingolani, G. (2005) Phospholipid scramblase 1 contains a nonclassical nuclearlocalization signal with unique binding site in importin alpha. J. Biol. Chem.280, 10599–10606.

[32] Zhou, Q., Sims, P.J. and Wiedmer, T. (1998) Identity of a conserved motifin phospholipid scramblase that is required for Ca2+-acceleratedtransbilayer movement of membrane phospholipids. Biochemistry 37, 2356–2360.

[33] Stout, J.G., Zhou, Q., Wiedmer, T. and Sims, P.J. (1998) Change in conformationof plasma membrane phospholipid scramblase induced by occupancy of itsCa2+ binding site. Biochemistry 37, 14860–14866.

[34] Sahu, S.K., Aradhyam, G.K. and Gummadi, S.N. (2009) Calcium binding studiesof peptides of human phospholipid scramblases 1 to 4 suggest thatscramblases are new class of calcium binding proteins in the cell. Biochim.Biophys. Acta 1790, 1274–1281.

[35] Sun, J., Nanjundan, M., Pike, L.J., Wiedmer, T. and Sims, P.J. (2002)Plasma membrane phospholipid scramblase 1 is enriched in lipid rafts andinteracts with the epidermal growth factor receptor. Biochemistry 41, 6338–6345.

[36] Amir-Moazami, O., Alexia, C., Charles, N., Launay, P., Monteiro, R.C. andBenhamou, M. (2008) Phospholipid scramblase 1 modulates a selected set ofIgE receptor-mediated mast cell responses through LAT-dependent pathway. J.Biol. Chem. 283, 25514–25523.

[37] Smrz, D., Lebduska, P., Draberova, L., Korb, J. and Draber, P. (2008) Engagementof phospholipid scramblase 1 in activated cells: implication forphosphatidylserine externalization and exocytosis. J. Biol. Chem. 283,10904–10918.

1786 F.-X. Contreras et al. / FEBS Letters 584 (2010) 1779–1786

[38] Pantaler, E., Kamp, D. and Haest, C.W. (2000) Acceleration of phospholipid flip-flop in the erythrocyte membrane by detergents differing in polar head groupand alkyl chain length. Biochim. Biophys. Acta 1509, 397–408.

[39] (a) Urbina, P., Alonso, A., Contreras, F.X., Goñi, F.M., López, D.J., Montes, L.R.and Sot, J. (2006) Alkanes are not innocuous vehicles for hydrophobic reagentsin membrane studies. Chem. Phys. Lipids 139, 107–114(b) Martin, C., Requero, M.A., Marin, J., Konopasek, I., Goñi, F.M., Sebo, P. andOstolaza, H. (2004) Membrane restructuring by Bordetella pertussis adenylatecyclase toxin, a member of the RTX toxin family. J. Bacteriol. 186, 3760–3765(c) Epand, R.F., Martinou, J.C., Montessuit, S. and Epand, R.M. (2003)Transbilayer lipid diffusion promoted by Bax: implications for apoptosis.Biochemistry 42, 14576–14582(d) Terrones, O., Antonsson, B., Yamaguchi, H., Wang, H.G., Liu, J., Herrmenn, A.and Basanez, G. (2004) Lipidic pore formation by the concerted action ofproapoptotic BAX and tBID. J. Biol. Chem. 279, 30081–30091.

[40] Goñi, F.M. and Alonso, A. (1999) Structure and functional properties ofdiacylglycerols in membranes. Prog. Lipid Res. 38, 1–48.

[41] Cook, H.W. (1991) Fatty acid desaturation and chain elongation in eukaryotesin: Biochemistry of Lipids, Lipoproteins and Membranes (Vance, D.E. andVance, J., Eds.), pp. 141–169, Benjamin/Cummings, Menlo Park, CA.

[42] Hodgkin, M.N., Pettitt, T.R., Martin, A., Michell, R.H., Pemberton, A.J. andWakelam, M.J. (1998) Diacylglycerols and phosphatidates: which molecularspecies are intracellular messengers? Trends Biochem. Sci. 23, 200–204.

[43] Wakelam, M.J. (1998) Diacylglycerol – when is it an intracellular messenger?Biochim. Biophys. Acta 1436, 117–126.

[44] Marignani, P.A., Epand, R.M. and Sebaldt, R.J. (1996) Acyl chain dependence ofdiacylglycerol activation of protein kinase C activity in vitro. Biochem.Biophys. Res. Commun. 225, 469–473.

[45] Schachter, J.B., Lester, D.S. and Alkon, D.L. (1996) Synergistic activation ofprotein kinase C by arachidonic acid and diacylglycerols in vitro: generation ofa stable membrane-bound, cofactor-independent state of protein kinase Cactivity. Biochim. Biophys. Acta 1291, 167–176.

[46] Veiga, M.P., Arrondo, J.L., Goñi, F.M. and Alonso, A. (1999) Ceramides inphospholipid membranes: effects on bilayer stability and transition tononlamellar phases. Biophys J. 76, 342–350.

[47] Epand, R.M. (1985) Diacylglycerols, lysolecithin, or hydrocarbons markedlyalter the bilayer to hexagonal phase transition temperature ofphosphatidylethanolamines. Biochemistry 24, 7092–7095.

[48] Basañez, G., Ruiz-Argüello, M.B., Alonso, A., Goñi, F.M., Karlsson, G. andEdwards, K. (1997) Morphological changes induced by phospholipase C and bysphingomyelinase on large unilamellar vesicles: a cryo-transmission electronmicroscopy study of liposome fusion. Biophys. J. 72, 2630–2637.

[49] Nieva, J.L., Alonso, A., Basáñez, G., Goñi, F.M., Gulik, A., Vargas, R. and Luzzati,V. (1995) Topological properties of two cubic phases of aphospholipid:cholesterol:diacylglycerol aqueous system and their possibleimplications in the phospholipase C-induced liposome fusion. FEBS Lett. 368,143–147.

[50] Pagano, R.E. and Longmuir, K.J. (1985) Phosphorylation, transbilayermovement, and facilitated intracellular transport of diacylglycerol areinvolved in the uptake of a fluorescent analog of phosphatidic acid bycultured fibroblasts. J. Biol. Chem. 260, 1909–1916.

[51] Allan, D., Thomas, P. and Michell, R.H. (1978) Rapid transbilayer diffusion of1,2-diacylglycerol and its relevance to control of membrane curvature. Nature276, 289–2190.

[52] Hamilton, J.A., Bhamidipati, S.P., Kodali, D.R. and Small, D.M. (1991) Theinterfacial conformation and transbilayer movement of diacylglycerols inphospholipid bilayers. J. Biol. Chem. 266, 1177–1186.

[53] Ganong, B.R. and Bell, R.M. (1984) Transmembrane movement ofphosphatidylglycerol and diacylglycerol sulfhydryl analogues. Biochemistry23, 4977–4983.

[54] Contreras, F.X., Basañez, G., Alonso, A., Herrmann, A. and Goñi, F.M. (2005)Asymmetric addition of ceramides but not dihydroceramides promotestransbilayer (flip-flop) lipid motion in membranes. Biophys. J. 88, 348–359.

[55] Hannun, Y.A. and Luberto, C. (2000) Ceramide in the eukaryotic stressresponse. Trends Cell Biol. 10, 73–80.

[56] Mathias, S., Peña, L.A. and Kolesnick, R.N. (1998) Signal transduction of stressvia ceramide. Biochem. J. 335, 465–480.

[57] Goñi, F.M. and Alonso, A. (1788) Effects of ceramide and other simplesphingolipids on membrane lateral structure. Biochim. Biophys. Acta, 169–177.

[58] Megha and London, E. (2004) Ceramide selectively displaces cholesterol fromordered lipid domains (rafts): implications for lipid raft structure andfunction. J. Biol. Chem. 279, 9997–10004.

[59] Megha, Sawatzki, P., Kolter, T., Bittman, R. and London, E. (2007) Effect ofceramide N-acyl chain and polar headgroup structure on the properties ofordered lipid domains (lipid rafts). Biochim. Biophys. Acta 1768, 2205–2212.

[60] Megha, Bakht, O. and London, E. (2006) Cholesterol precursors stabilizeordinary and ceramide-rich ordered lipid domains (lipid rafts) to different

degrees. Implications for the Bloch hypothesis and sterol biosynthesisdisorders. J. Biol. Chem. 281, 21903–21913.

[61] Nybond, S., Björkqvist, Y.J., Ramstedt, B. and Slotte, J.P. (2005) Acyl chainlength affects ceramide action on sterol/sphingomyelin-rich domains.Biochim. Biophys. Acta 1718, 61–66.

[62] Staneva, G., Chachaty, C., Wolf, C., Koumanov, K. and Quinn, P.J. (2008) The roleof sphingomyelin in regulating phase coexistence in complex lipid modelmembranes: competition between ceramide and cholesterol. Biochim.Biophys. Acta 1778, 2727–2739.

[63] Massey, J.B. (2001) Interaction of ceramides with phosphatidylcholine,sphingomyelin and sphingomyelin/cholesterol bilayers. Biochim. Biophys.Acta 1510, 167–184.

[64] Sot, J., Ibarguren, M., Busto, J.V., Montes, L.R., Goñi, F.M. and Alonso, A. (2009)Cholesterol displacement by ceramide in sphingomyelin-containing liquid-ordered domains, and generation of gel regions in giant lipidic vesicles. FEBSLett. 582, 3230–3236.

[65] Kolesnick, R.N., Goñi, F.M. and Alonso, A. (2000) Compartmentalization ofceramide signaling: physical foundations and biological effects. J. Cell. Physiol.184, 285–300.

[66] Montes, L.R., Ruiz-Argüello, M.B., Goñi, F.M. and Alonso, A. (2002) Membranerestructuring via ceramide results in enhanced solute efflux. J. Biol. Chem. 277,11788–11794.

[67] Ruiz-Argüello, M.B., Goñi, F.M. and Alonso, A. (1998) Vesicle membrane fusioninduced by the concerted activities of sphingomyelinase and phospholipase C.J. Biol. Chem. 273, 22977–22982.

[68] Holopainen, J.M., Lehtonen, J.Y. and Kinnunen, P.K. (1997) Lipid microdomainsin dimyristoylphosphatidylcholine-ceramide liposomes. Chem. Phys. Lipids88, 1–13.

[69] van Blitterswijk, W.J., van der Luit, A.H., Veldman, R.J., Verheij, M. and Borst, J.(2003) Ceramide: second messenger or modulator of membrane structure anddynamics? Biochem. J. 369, 199–211.

[70] Bai, J. and Pagano, R.E. (1997) Measurement of spontaneous transfer andtransbilayer movement of BODIPY-labeled lipids in lipid vesicles.Biochemistry 36, 8840–8848.

[71] López-Montero, I., Rodriguez, N., Cribier, S., Pohl, A., Vélez, M. and Devaux, P.F.(2005) Rapid transbilayer movement of ceramides in phospholipid vesiclesand in human erythrocytes. J. Biol. Chem. 280, 25811–25819.

[72] Sheetz, M.P. and Singer, S.J. (1974) Biological membranes as bilayer couples. Amolecular mechanism of drug–erythrocyte interactions. Proc. Natl. Acad. Sci.USA 71, 4457–4461.

[73] López-Montero, I., Vélez, M. and Devaux, P.F. (2007) Surface tension inducedby sphingomyelin to ceramide conversion in lipid membranes. Biochim.Biophys. Acta 1768, 553–561.

[74] Contreras, F.X., Villar, A.V., Alonso, A., Kolesnick, R.N. and Goñi, F.M. (2003)Sphingomyelinase activity causes transbilayer lipid translocation in modeland cell membranes. J. Biol. Chem. 278, 37169–37174.

[75] Sot, J., Goñi, F.M. and Alonso, A. (2005) Molecular associations and surface-active properties of short- and long-N-acyl chain ceramides. Biochim. Biophys.Acta 1711, 12–19.

[76] Sot, J., Aranda, F.J., Collado, M.I., Goñi, F.M. and Alonso, A. (2005) Differenteffects of long- and short-chain ceramides on the gel-fluid and lamellar-hexagonal transitions of phospholipids: a calorimetric, NMR, and X-raydiffraction study. Biophys. J., 3368–3380.

[77] Hannun, Y.A. (1996) Functions of ceramide in coordinating cellular responsesto stress. Science 274, 1855–1859.

[78] Stiban, J., Fistere, D. and Colombini, M. (2006) Dihydroceramide hindersceramide channel formation: implications on apoptosis. Apoptosis 11, 773–780.

[79] Pomorski, T., Hrafnsdóttir, S., Devaux, P.F. and van Meer, G. (2001) Lipiddistribution and transport across cellular membranes. Semin. Cell Dev. Biol.12, 139–148.

[80] Raggers, R.J., Pomorski, T., Holthuis, J.C., Kälin, N. and van Meer, G. (2000) Lipidtraffic: the ABC of transbilayer movement. Traffic 1, 226–234.

[81] Gurtovenko, A.A. and Vattulainen, I. (2007) Ion leakage through transientwater pores in protein-free lipid membranes driven by transmembrane ioniccharge imbalance. Biophys. J. 92, 1878–1890.

[82] Gurtovenko, A.A. (2005) Asymmetry of lipid bilayers induced by monovalentsalt: atomistic molecular-dynamics study. J. Chem. Phys. 122, 244902.

[83] Gurtovenko, A.A. and Vattulainen, I. (2007) Molecular mechanism for lipidflip-flops. J. Phys. Chem. B 111, 13554–13559.

[84] Gurtovenko, A.A. and Vattulainen, I. (2005) Pore formation coupled to iontransport through lipid membranes as induced by transmembrane ioniccharge imbalance. atomistic molecular dynamics study. J. Am. Chem. Soc. 127,17570–17571.

[85] López-Montero, I., Monroy, F., Vélez M. and Devaux, P.F. (2010) Ceramide:from lateral segregation to mechanical stress. Biochim. Biophys. Acta. [17 Dec.2009, Epub ahead of print].