Comparison of the liquid-ordered bilayer phases containing cholesterol or 7-dehydrocholesterol in modeling Smith-Lemli-Opitz syndrome.

The phase behavior of egg sphingomyelin (ESM) mixtures with cholesterol or 7-dehydrocholesterol (7-DHC) has been investigated by independent methods: fluorescence microscopy, X-ray diffraction, and electron spin resonance spectroscopy. In giant vesicles, cholesterol-enriched domains appeared as large and clearly delineated domains assigned to a liquid-ordered (Lo) phase. The domains containing 7-DHC were smaller and had more diffuse boundaries. Separation of a gel phase assigned by X-ray examination to pure sphingomyelin domains coexisting with sterol-enriched domains was observed at temperatures less than 38°C in binary mixtures containing 10-mol% sterol. At higher sterol concentrations, the coexistence of liquid-ordered and liquid-disordered phases was evidenced in the temperature range 20°–50°C. Calculated electron density profiles indicated the location of 7-DHC was more loosely defined than cholesterol, which is localized precisely at a particular depth along the bilayer normal. ESR spectra of spin-labeled fatty acid partitioned in the liquid-ordered component showed a similar, high degree of order for both sterols in the center of the bilayer, but it was higher in the coexisting disordered phase for 7-DHC. The differences detected in the models of the lipid membrane matrix are said to initiate the deleterious consequences of the Smith-Lemli-Opitz syndrome.

domain separation on the micrometer scale in giant unilamellar vesicles (GUV) comprised of ternary mixtures of sphingomyelin and sterol within a fl uid phospholipid matrix. Synchrotron X-ray diffraction (XRD) studies of binary mixtures of sphingomyelins and sterols have been performed at the molecular scale to investigate the association of major raft components over a temperature range spanning the gel-to-fl uid phase transition of the sphingomyelin. Finally, the dynamic properties of phases formed by sphingomyelin and sterol have been examined at a molecular and nanosecond scale using electron spin resonance (ESR) of spin-labeled probes to determine the differences between sterols in maintenance of molecular ordering. We demonstrate that subtle differences in the simple membrane models can be detected between cholesterol and its precursor 7-dehydrocholesterol. We conclude that the observed differences may explain the molecular basis of defective raft assembly in embryonic cells that results in Smith-Lemli-Opitz syndrome.

Lipids
Lipids were used without further purifi cation: egg yolk L-␣ -phosphatidylcholine (EPC), egg yolk sphingomyelin (ESM), cholesterol (CHOL), and 7-dehydrocholesterol (7-DHC) were obtained from Sigma (Sigma-Aldrich Chimie SARL, St. Quentin, France). The fl uorescent phospholipid analog Texas Red [1, 2-dihexadecanoyl-sn -glycero-3-phosphoethanolamine, triethylammonium salt (TR-DHPE)] was purchased from Invitrogen (Eugene, OR). The molar ratios of lipids in ternary mixtures of EPC/ESM/sterol, (45/45/10, 40/40/20, 34/33/33) were chosen for GUV studies for comparison with the lipid composition of detergent-resistant membrane fractions. For example, SM/ CHOL ratios reported in detergent-resistant membrane preparations vary widely from 1/1 to 2/1 ( 22 ). Inclusion of a fl uid lipid (EPC) is a technical requirement to observe partition of a fl uorescent probe between liquid-disordered (L d ) phase and liquidordered (L o ) phase in GUVs. For ESR and XRD studies, comparisons between the two sterols in binary mixtures was the more appropriate system to distinguish subtle differences in the association of the major components of membrane rafts in forming a liquid-ordered structure.

Preparation of giant unilamellar vesicles
The electroformation method developed by Angelova and Dimitrov ( 23 ) was used to form the vesicles. Binary and ternary mixtures of EPC, ESM, and sterol were prepared, in which all lipid ratios are given as molar. GUVs were formed in a temperature-controlled chamber. We followed the particular protocol for unilamellar giant heterogeneous vesicles formation described elsewhere ( 24 ). Electroformation temperature was chosen to be the maximal temperature at which a high yield of vesicles was consistently obtained. Although deposited lipids are uniformly mixed, each vesicle varied slightly in composition ( 25 ). This composition error is estimated to be ‫ف‬ 2%, which could be observed as different miscibility transition temperatures and brightness (mol% dye) between vesicles. The head-group-labeled lipid analog TR-DHPE is excluded from the more ordered phase for its steric hindrance and partitions predominantly in the disordered phase, which makes the ordered phase appear as a dark area within the bright fl uid vesicle membrane. feedback on the aberrant sterol production is also exerted by cholesterol supplied in excess ( 11,12 ).
There are currently two hypotheses to explain the cell defects of Smith-Lemli-Opitz syndrome (SLO) embryonic cells. The fi rst is that 7-DHC is not able to substitute for cholesterol as a component of cell membranes and, second, that the cleavage of the morphogenic protein, Sonic Hedgehog (SHH) is defective in the presence of 7-DHC. Both cholesterol and 7-DHC are known to augment cleavage of the N-terminal signaling peptide from SHH ( 13 ). Because the reaction to remove the signal peptide takes place in membrane "raft" domains that are enriched in cholesterol, the two explanations for SLO syndrome may be reconciled. It is possible that replacement of cholesterol by 7-DHC produces membrane domains that are unable to provide the necessary structural environment for posttranslational modifi cation of SHH ( 14 ).
There is some evidence that 7-DHC produces defective membrane rafts. Studies of the replacement of cholesterol with 7-DHC in preparations of hippocampal membranes in vitro failed to restore the binding activity of the serotonin 1A receptor despite the overall recovery of the membrane lipid order ( 15,16 ). Although the total sterol (7-DHC plus cholesterol) incorporation in rafts isolated from AY9944treated rat brains reaches levels similar to that of cholesterol incorporation in control preparations, the raft protein profi les are different ( 17 ), suggesting that the sterol composition changes the lipid arrangement that, in turn, infl uences protein sorting and defective signaling functions ( 14 ). Another difference has been reported in mast cells derived from Dhcr7 Ϫ / Ϫ mice accumulating 7-DHC in lipid rafts in vivo ( 18 ). 7-DHC was assumed to disrupt the raft stability in Dhcr7 Ϫ / Ϫ mice and, consequently, to produce a dysfunctional allergic response because of a loosening of the raft association of the regulatory Lyn kinase. Furthermore, the molecular simulation has shown that 7-DHC is found less effective compared with cholesterol in condensing the packing density of acyl chains ( 19 ). In model membrane studies, 7-DHC is known to partition into domains that were resistant to the solubilization by Triton X-100 and more stable at higher temperatures than cholesterol ( 20 ).
To understand the molecular mechanisms that underlie these observations, biophysical studies that can distinguish the subtle differences in membrane structure and stability when 7-DHC substitutes for cholesterol are required. The presence of only an additional double bond in the B-ring of the sterol nucleus contrasts with the profound consequences in embryos when the precursor replaces the endmetabolite. Cholesterol is an absolute requirement for a number of homeothermic animal species. K. Bloch postulated that sterol structure is gradually optimized during evolution by adding steps to the biosynthetic pathway to perform new functions ( 21 ). A concerted coevolution of sphingolipids and sterols in mammals is also postulated along with this sterol-only hypothesis.
The present study of model membrane systems was undertaken using three methods with complementary spatial and time scales to compare the properties of sterols. Fluorescence probe partition has been used to characterize represent the fractions of coexisting phases but likely have similar temperature dependence.

Phase behavior in GUVs
Domain separation of ordered and disordered phases in phospholipid/sterol mixtures was examined by fl uorescence microscopy of GUVs. Vesicles composed of EPC/ ESM/sterol in molar proportions 45/45/10 were examined at different temperatures in . Domain fusion typically occurs in less than 1 s, and the round shape of the dark domain is quickly restored after fusion. The separation of dark domains is reminiscent of liquid/liquid immiscibility between the liquid-ordered (dark area) domains and the liquid-disordered (bright area) domains previously described (32)(33)(34).
Domain patterns in GUV containing 7-DHC were different ( Fig. 1B ). Two types of dark domains were distinguished at 15°-20°C with crenellated borders or leaf-like domains [ Fig. 1B( b, c)]. Unlike cholesterol, in the presence of 7-DHC, the dark domains do not fuse but tend to cluster and form a "necklace" of small-size domains [ Fig.  1B( c)]. Note that dark domains assigned as liquid-ordered domains do not fuse together as do domains assigned to the liquid-ordered phase in the presence of cholesterol [ Fig. 1B( d, e)]. The properties of the 7-DHC domains appear similar to gel domains observed in the absence of GUVs comprised of binary mixtures of SM/sterols exhibited homogeneous appearance on the micron scale. Obviously, the gel/liquid-ordered phase separation that takes place on the submicron scale is not detectable at the actual resolution.

Video microscopy
The vesicles were observed using a Zeiss Axiovert 135 microscope, equipped with a 63× long working distance objective lens (LD Achroplan Ph2). The observations were recorded using a Zeiss AxioCam HSm CCD camera connected to an image recording and processing system (Axiovision, Carl Zeiss, Wien, Austria). The phase morphology transformations of the heterogeneous GUV membrane were followed in fl uorescence by the Zeiss fi lter set Ex/Em = 546/590 nm. As a function of time, fusion and alteration of the domain shape and size can be spotted by video microscopy.

Synchrotron X-ray diffraction methods
Samples for XRD examination were prepared by dissolving lipids in warm (45°C) chloroform/methanol (2:1, v/v) and mixing in the desired proportions (given as molar ratio). The organic solvent was subsequently evaporated in SpeedVac equipment (SPD111V, Savant, Thermo Fisher, Illkirch, France). The dry lipids were hydrated with an equal weight of water and heated repeatedly above 45°C, a temperature cycling which was found effi cient to fully hydrate a control of pure SM ( 26,27 ). Simultaneous small-and wide-angle X-ray scattering intensities were recorded during thermal scans between 20° and 50°C on beamline 2.1 of the Daresbury SRS as described elsewhere ( 27 ).

Analysis of X-ray diffraction data
The SAXS intensity profi les were analyzed using fi tting by Gaussian/Lorentzian symmetrical functions. The scattering intensity data from the fi rst fi ve orders of the Bragg refl ections from the multilamellar liposomes were used to construct electron density profi les ( 28 ). After correcting the raw data by subtracting the background scattering from both water and the sample cell, each of the Bragg peaks were fi tted by a Gaussian/Lorentzian (Voigt) distribution using PeakFit 4.12 (Systat Software). The protocol has been recently described in ( 27,29 ).

ESR spectroscopic methods
The ESR experiments have been performed at 9.4 GHz using the doxyl-16-stearic spin probe denoted hereafter as 16NS. In the whole temperature range of the experiments (12°-47°C), the reorientation of this probe is under the fast motional regime with correlation times << 2 ns ( Table 1 ). The experimental procedures and methods of determining ESR parameters from spectral fi ttings have been reported elsewhere ( 30,31 ).
In the present study and for most of the samples, the spectral fi ttings have been performed under the following assumptions: ( 1 ) One anisotropic (liquid crystalline) site; ( 2 ) two sites (anisotropic and isotropic); ( 3 ) two anisotropic sites. The most-and the less-ordered sites are denoted A and B, respectively.
The selection among these assumptions is based on the quality of the fi ts estimated from the standard deviation between the experimental and fi tted spectra, normalized with respect to the maximum amplitude of the former, as well as the identifi cation of the phases detected by X-ray diffraction.  plane. Notably, no "leaf-like" gel domains were detected in the presence of 20 mol% 7-DHC, even at temperatures of 4°C, indicating that the sterol was intermixed with sphingomyelin. The overall domain area in the presence of 20 mol% 7-DHC was roughly similar to the area of L o domain formed in the presence of cholesterol, however, the rate of fusion of 7-DHC domains remained relatively slow. Moreover, domains demonstrated plasticity during fusion compared with the fast, elastic behavior observed in cholesterol mixtures.
The higher molar proportion of sterol in ternary mixtures (33 mol%) resulted in the formation of dark domains at a temperature 3°C lower [ Fig. 3A, B(b)] than in the presence 20 mol%. Clusters of small-size domains were also decreased, and a single enlarged domain was occasionally observed at 33 mol% [ Fig. 3B( c)]. The domain pattern was more regularly rounded compared with mixtures containing less 7-DHC. As the dark area was enlarged with fewer separated domains, the observation was consistent with fusion. Nevertheless, fusion in the presence of 7-DHC proceeds at a slow rate, and observations have rarely captured fusion events.  Fig. 2A( b)] multiple, small L o domains emerge initially and fuse shortly after into few large domains [ Fig. 2A (c-e)]. The domains have a round shape owing to the "minimized tension" line energy at the rim of the two fl uid phases by optimizing the area/perimeter ratio. The mixtures containing 7-DHC show small dark domains that have an irregular round shape, and unlike cholesterol, they did not change signifi cantly with temperature [ Fig. 2B( b)]. Occasionally the clusters of two, three, or more quasiround small domains observed in the presence of 7-DHC remained connected by fi ne fi laments [ Fig. 2B( c-f)] and moved in concert in the membrane  temperatures ( Fig. 4B ). The "buffering" effect of cholesterol on the dimensions of the repeat unit as a function of temperature is explained by a compensatory disordering/ ordering infl uence on the gel/disordered state, respectively. However, this infl uence is less marked for 7-DHC compared with CHOL as a function of temperature ( Fig. 4B ). For example, at 20°C the mixture containing 50 mol% CHOL forms a structure with two distinct coexisting structures (d-spacing 6.02 and 6.15 nm). Such heterogeneity is hardly detected in a mixture containing 50 mol% 7-DHC (6.27 and 6.30 nm). Because both the hydration layer thickness and the bilayer thickness contribute to the total d-repeat, their values have to be judged after Fourier deconvolution of the transversal electron profi les ( Fig. 6 ). However, it is clear from the raw data that the two components in the diffractogram of the heterogeneous mixture show not only distinct d-spacing but also a different temperature variation in the presence of CHOL or 7-DHC ( Fig. 4B ).
In binary mixtures with a low proportion of sterol (10 mol%) [ Fig. 4B( a)], there is no variation of d-spacing in the lowest temperature range. When the temperature is increased, the variation is biphasic: the d-spacing increases and then decreases with heating. We have interpreted the increase observed at a temperature lower than the onset of the gel-to-liquid transition as the transition of an interdigitated gel phase of sphingomyelin to a noninterdigitated arrangement that was already described for pure sphingomyelin ( 31 ). A visual inspection of the SAXS profi les at high temperatures reveals coexisting phases. However, applying the procedure of deconvolution to partially resolved peaks as shown in Fig. 5 , the coexistence is detectable in the whole temperature range for sterol concentrations greater than 20 mol%.
Remarkably, in the presence of a suffi cient concentration of CHOL, a structure detected throughout the temperature range shows a constant d-repeat [ Fig. 4B( d)]. This temperature criterion has served previously to characterize the liquid-ordered phase identifi ed by ESR in sphingomyelin/cholesterol mixtures at concentrations of sterol greater than 17 mol% ( 31 ). The stability of the structure suggests a fi xed composition and temperatureindependent behavior of the molecular arrangement of cholesterol with SM. Notably the heterogeneity is detected by SAXS in sterol-enriched mixtures around mammalian body temperatures (37°C) and upward when the d-spacings of the coexisting components diverge; the L o component remains relatively unchanged while the d-spacing of the less-ordered component progressively decreases with increasing temperature ( Fig. 4B ).
To examine the localization of the sterol molecule within the coexisting structures and to measure the thickness of the hydration layer and the lipid bilayer in liposomes, electron density profi les perpendicular to the bilayer plane were calculated. Fourier syntheses were applied on the components separated after peak fi tting of the raw diffractogram ( Fig. 5 ) where peak 1 is assigned to the L o structure characterized by a longer, relatively temperature-independent d-spacing, and peak 2 to the L d The ability of sterols to form a L o phase was conveniently denoted as promoter or inhibitor by Beattie et al. ( 35 ). The authors showed that vesicles containing ''promoter'' sterols (such as cholesterol, ergosterol, 25-hydroxycholesterol, epicholesterol, and dihydrocholesterol) separated into coexisting liquid phases as the temperature lowered through the miscibility transition temperature. In contrast, vesicles containing the "inhibitor" sterols (such as androstenolone, cholestenone, and cholestane) showed coexisting gel and liquid phases. On this simple basis, the assignment of 7-DHC to one of these two categories is problematic as we have observed that 7-DHC can form micrometer-scale domains but at a slower rate compared with cholesterol. Thus the small-sized domains formed by 7-DHC are characterized by a slower rate of fusion to create large domains, and their boundaries are more rigid than CHOL domains.

Synchrotron X-ray diffraction studies
The phase behavior of ESM/sterol mixtures was examined by X-ray diffraction in multilamellar dispersions as a function of sterol structure and concentration. The study focused on the association of the two most prominent membrane raft components, SM and sterol. An overview of the small-angle X-ray scattering (SAXS) recordings is presented in Fig. 4A during the heating temperature ramp from 20° to 50°C. The fi rst two orders of Bragg refl ections of the lamellar repeat are shown enlarged in the SAXS diffractogram. The splitting of the peaks reveals the heterogeneity of the mixture. Differences in the thermotropic behavior between mixtures containing either CHOL ( Fig.  4A, left panel) or 7-DHC ( Fig. 4A, right panel) are also evidenced by distinct lamellar repeat spacings. As expected, increased proportions of sterols in the mixtures result in a decrease in the lamellar repeat spacing (d-spacing) at low profi le of the liquid-ordered as well as the liquid-disordered component of sterol-containing mixtures shows a shoulder or plateau of electron density between the dense phosphorus and oxygen-rich head-group position and the trough in the bilayer center ( 36 ). Such a pronounced structure, which is less ordered, has a shorter d-spacing, and is strongly temperature-dependent. In Fig. 6 the electron density profi les calculated for the liquid-ordered component (peak 1) is shown on the left, and the disordered component (peak 2) on the right. The electron density (WAXS) signal is illustrated in Fig.7 . In the presence of 7-DHC, an abundant gel component is detected at 10 mol% in agreement with GUV and SAXS observations ( Fig. 7D versus 7B in the presence of CHOL). More detail about the analysis of the WAXS intensity data is presented in the supplementary data.

ESR spectroscopic measurements
The motion and order in binary mixtures of ESM and sterols were investigated by ESR spectroscopy of a spin-labeled fatty acid with the nitroxide reporter group attached at position 16 of the acyl chain to probe the environment of the bilayer center. Order parameters and reorientation correlation times were derived from a spectral simulation analysis. The studies compared the properties of mixtures with different proportions of CHOL and 7-DHC.

ESM/sterol (83/17)
Order parameters and their assignments for these binary mixtures are presented in Fig. 8A . S zz shows a marked decrease between 30° and 40°C, which likely corresponds to the L ␤ ↔ L d transition occurring in the same temperature range for pure ESM. The temperature dependence of S zz has been analyzed as corresponding to a two-anisotropic sites system. It is observed that the values of the less-ordered B site below 30°C are aligned with the values of the moreordered A site above 40°C, suggesting that they belong to the same L o phase evidenced by the X-ray diffraction. Therefore L ␤ and L o phases coexist below 30°C, and L o and L d (full line) coexist above 40°C. At 32°C and 37°C there is possibly coexistence of three phases L ␤ , L o , and L d , but their respective contribution cannot be reliably obtained. At 32°C the S zz value assigned to L o is underestimated, owing possibly to an unresolved contribution of the L d phase. The disappearance of L ␤ above 37°C entails an enhancement of L o and L d fractions, which shows the same abundance at 42°C ( Fig. 8B ). Above this temperature, the L o fraction decays with a corresponding increase of the L d fraction.

ESM/sterol (66/34)
In the whole temperature range, the spectral fi ttings are consistent with the coexistence of L o and L d phases ( Fig. 8C ) in agreement with X-ray data. The L o fraction reported by the doxyl-16-stearic acid is nearly independent of temperature with 0.83 ± 0.026 and 0.81 ± 0.031 for ESM mixtures with CHOL and 7-DHC, respectively ( Fig. 8D ). In the L o phase, S zz is similar for CHOL and 7-DHC, but in the L d phase, S zz is larger above 25°C for 7-DHC than for CHOL ( Fig. 8C ). This difference suggests that the fraction of sterols that is present in the L d phase forms transient complexes with ESM ( 37 ) that are possibly distinct for the two sterols reported by ESR. For the stable assembly in L o , the infl uence is similar for the two sterols.

ESM codispersed with 41, 50, and 60 mol% sterol
When the sterol concentration is increased to 40 mol% or greater, the assumption of an anisotropic site and an isotropic one yields better ESR spectral fi ttings than two anisotropic sites observed at 17 and 34 mol% ( Fig. 8A-D ). break in the slope of electron density is consistently detected in cholesterol-containing liposomes, which refl ects the accurate alignment of the sterol nuclei at an accurate, precise depth in the bilayer. Electron density profi les for 7-DHC (dashed lines) have a smoother profi le compared with cholesterol in both L o and L d structures. The difference is interpreted as 7-DHC occupying a position in the bilayer depth that is less constrained than cholesterol.
Phase coexistence in binary mixtures of SM with sterol is also evidenced by heterogeneous X-ray scattering in the wide-angle region, a measurement sensitive to the molecular ordering on a short distance compared with SAXS. The distinction between gel and fl uid liquid-phase responses is shown in Fig. 7A for a control of pure SM (s = 1/d). The peak at d = 0.42 nm characterizing the gel phase disappears in the temperature interval coinciding with T m of ESM (37°-41°C). In binary SM/sterol mixtures, the scattering signal in the wide-angle region can also be deconvolved by two components. The ordered phase is assigned to a d-spacing of 0.426 nm, which corresponds to the most "compact" phase, and the disordered phase ( Fig. 7D, E ) to a broad band centered at about 0.445 nm, which indicates longer average intermolecular distances. The position of both wide-angle bands remained constant with different proportions of sterol but their relative proportions in the heterogeneous scattering signal change. The change in the proportion of the two components of the wide-angle X-ray scattering components of membrane lipids. Less attention has been given to sterol "acclimatization." In man, 19 enzymes are required after the cyclization of squalene, which leads to the fi nal cholesterol nucleus. It is remarkable that the penultimate structure in the pathway, 7-dehydrocholesterol, with only a difference of an extra double bond in the B-ring of the sterol nucleus, does not meet the stringent requirements needed to substitute for cholesterol in defi cited embryos with Smith-Lemli-Opitz syndrome.
Evidence that 7-DHC and cholesterol could not be interchanged had been obtained earlier for the restoration of the serotonin-binding capability of solubilized hippocampal receptor 1A in vitro ( 15,16 ). Desmosterol, a closely related structure with an extra double bond ( ⌬ 24) on the hydrocarbon side chain and also an immediate precursor of cholesterol in the Steinberg biosynthesis pathway, was also unable to substitute for cholesterol in the restoration of serotonin binding ( 38 ). Taken together, such biological studies have rarely received supporting evidence from detailed biophysical measurements to give credence to the concept proposed earlier by Bloch ( 39 ) that the "structural details of the cholesterol molecule and related sterols can be rationalized in terms of optimal hydrophobic interactions between the planar sterol ring system (and lateral side chain) and phospholipid acyl chains or ceramides" ( 21 ).
Following introduction of the so-called membrane raft hypothesis, which relies on the creation of sphingolipid/sterol At high concentration of sterol, the isotropic fraction remains constant at 0.04 ± 0.01 independently of temperature. The anisotropic site is assigned to the prominent L o phase evidenced by X-ray diffraction at high concentration of sterol. The isotropic fraction likely corresponds to the probe located around the boundaries of domains of the L o phase, undergoing exchange at a rate of 8 × 10 6 Ϫ 10 7 s Ϫ 1 with the probe incorporated in this phase. Fig. 8E-G shows only small differences between the S zz values corresponding to ESM mixtures with CHOL or 7-DHC. A comparison of spin-label motional parameters and activation energies derived from the spectroscopic measurements is presented in Table 1 . Activation energies of the CHOL mixtures tended to be greater than mixtures containing 7-DHC, but no trend could be clearly discerned from the data on motional parameters.

DISCUSSION
Many biological studies have demonstrated that the structure of sterol has evolved to meet the specifi c requirements for the maintenance of biological membrane function. The maintenance of the liquid-crystalline structure in homeothermic mammalian species is especially challenging compared with poïkilotherms. The homeoviscous adaptation of unicellular organisms and plants to temperature is mainly controlled by the activity of desaturases responsible for inserting cis -double bonds into fatty acyl insignifi cant as shown recently by examining the role of ceramide in quaternary mixtures in modulating the effect of sterol to counteract the separation of the L ␤ phase formed by the saturated sphingolipids at physiological temperatures ( 21,43 ). Furthermore, in another X-ray diffraction study of ternary mixtures composed of egg phosphatidylcholine/sphingomyelin/cholesterol, several quasi at-equilibrium lipid organizations were said to exist at different times in the same local patch of membrane under conditions leading to L o /L d phase separation ( 44 ). The authors state that "identifi cation of the conditions which can mimic heterogeneous dynamic membrane states in a lipid membrane model is a challenge. This is of particular importance as the lateral organization of lipid mixtures in fully equilibrated samples may differ from the arrangement found in quasi at-equilibrium conditions." These problems have been reviewed in the context of data obtained using a variety of biophysical techniques ( 45 ). In the present study, we have focused on a comparison of cholesterol with its precursor metabolite in simplifi ed binary mixtures, which enriched domains ( 40 ), a number of biophysical studies have emphasized the particular features of sterol structure in the separation of liquid-disordered from liquid-ordered domains ( 20,41 ). The structural requirements have been operationally established from differential detergent solubilization of the L o and L d phases and by fl uorescence quenching in stabilizing or disrupting lipid domains at different temperatures ( 42 ). The comparative studies have shown that the separation of a sterol-enriched domain formed with a saturated phospholipid, either SM or dipalmitoyl/phosphatidylcholine, can be counteracted by another sterol or steroid structure. This type of competition has possible consequences in SLO-affected fetus where the cholesterol of maternal origin and the endogenous 7-DHC are accumulating.
Binary and ternary lipid mixtures have been used for simple model membrane studies where the precision of biophysical methods can be fully exploited. This approach sacrifi ces the infl uence of minor lipid species of lipid on the L o /L d phase separation. This infl uence may not be  signal can be traced by peak deconvolution, demonstrating the coexistence of a compact ordered phase with loosely oriented acyl chains. 4 ) A remarkable result is that differences between sterols detected by WAXS and ESR in the nanometer range are translated into differences on the long distance scale documented by SAXS and GUV observations. The difference between 7-DHC-and cholesterol-containing liposomes detected by SAXS for intermediate sterol concentrations illustrates this point. While cholesterol at 40 mol% produces a clear-cut separation revealed by distinct Bragg's refl ections of the SAXS, 7-DHC produces only a shoulder of the diffraction peak. The difference between the two sterols is widened by changing the temperature. Taking into account the GUV observations showing the small-sized domains formed in the presence of 7-DHC, we assume that the cause of irregular arrays is a loose alignment of the small-sized domains formed in the adjacent liposome layers. The rigidity of domains created in mixtures containing 7-DHC can explain the different diffraction pattern compared with cholesterol. Nevertheless, the methods of SAXS peak fi tting and calculation of the transversal electron density profi le have been able to detect differences between sterols while the geometry of the SM/sterol assembly is almost the same. 5 ) Because it is can be inferred from WAXS and ESR that cholesterol and 7-DHC do not intercalate similarly to disrupt the gel phase composed of pure sphingomyelin, it is expected that imaging of domains at a low temperature (22°C) provided by the GUV observations will be different. Indeed, the domains of 7-DHC/sphingomyelin separated from the fl uid, brightly fl uorescent host phospholipid appear with crenulated boundaries and a small size. Their aspect and temperature behavior are more reminiscent of gel-like domains than they are of the large, round-shaped, rapidly merging domains that are observed in the presence of cholesterol and are typical of a liquid phase. The present data are in agreement with Langmuir monolayers data showing that the molecular area in 7-DHC/dipalmitoylphosphatidylcholine fi lm is larger than for mixtures containing cholesterol ( 46 ). Another monolayer study has reported no signifi cant difference between 7-DHC and cholesterol in mixtures with EPC ( 47 ). Therefore the monolayer studies with 7-DHC have concluded that the difference with cholesterol is likely found in the L o domains formed with the saturated lipids. This result is fully consistent with the present GUV and X-ray examination and is partially consistent with present ESR data showing differences in the less-ordered coexisting phase.
Besides the composition and biophysical properties of the raft membranes, the size of domains has also emerged as a factor infl uencing the effi ciency of the multiprotein signaling machinery ( 48 ). It was found that enrichment in 7-dehydrocholesterol causes raft instability as judged by detergent solubilization and methyl-cyclodextrin extractability. As a consequence, it was pointed out in the same study that the raft-altered structure resulted in a hyperallergic response of mast cells prepared from Dhcr7 Ϫ / Ϫ mice ( 19 ). More specifi cally, mast cells derived from 7-DHC reductase knockout mice show a high cytokine production can be readily related to earlier studies of sphingomyelin/ cholesterol binary mixtures ( 27,31 ).
The present study illustrates how subtle differences in the structure of the sterol nucleus may cause membrane lipid dysfunctions that "cascade" to create major consequences, such as SLO syndrome. The differences between cholesterol and 7-DHC can be summarized as follows: 1 ) At the molecular level, the electron density profi les of the L o and L d phases suggest a slightly different arrangement of cholesterol and 7-DHC relative to the lipid/water interface of the host sphingomyelin bilayer. The overall geometry of the assembly with SM is not signifi cantly altered in terms of the thickness of the bilayer and hydration layer, but an alteration in the electron density profi le of bilayers is observed. Bilayers containing 7-DHC do not show a pronounced plateau assigned to the electron-enriched position of the sterol nucleus as they do in the case of cholesterol ( 28. 36 ). 2 ) Because the sterol nucleus is rigid and localized superfi cially in the bilayer compared with the long hydrocarbon chains of SM, it is surprising that a slightly different positioning of the nucleus will have a consequence on the dynamics of spin-label motion in the center of the bilayer. However, ESR spectroscopy detects a subtle but signifi cant alteration in the molecular order parameter of 16NS in both of the coexisting phases of the heterogeneous binary mixture. In the bilayer center, the ordering infl uence of cholesterol is weak relative to positions closer to the lipid-water interface where the fl at and rigid sterol ring system is intercalated. Nevertheless, an increase in S zz for 16NS from 0.11 (pure L d at 41°C) to 0.30 was already detected in the L o phase applying similar ESR and spectral simulation methods ( 31 ). The present ESR observations show a stronger infl uence of 7-DHC in the bilayer center compared with cholesterol. We assume that the anchoring of the cholesterol nucleus at a more precise depth of the bilayer would limit its infl uence on regions around the bilayer center. Looser van der Waals interactions of the 7-DHC ring with the hydrocarbon chains of sphingomyelin can be expected to cause the averaging of the position of 7-DHC along the bilayer normal. It is also possible that the stabilization by 3-hydroxyl group of cholesterol, which participates in a precisely configured hydrogen-bond network formed with amino groups of the sphingosine moiety, is not permitted in the case of a loosely associated sterol such as 7-DHC. 3 ) Sterol-sphingomyelin interaction prevents the separation of the gel phase composed of the sphingomyelin molecular species at temperatures lower than their respective transitions. As observed from the WAXS measurements, the complete replacement of the gel phase is not seen until the proportion of either of the two sterols reaches 34 mol%. WAXS analysis indicates a greater abundance of a gel component in the presence of 7-DHC than cholesterol at 10 and 20 mol%. ESR methods cannot detect the difference evidenced by WAXS, and the binary mixture at 17 mol% 7-DHC and cholesterol is composed of equally abundant gel/fl uid fractions. Note that the partition of the 16NS spin label may obscure the actual abundance of the fractions. At higher temperatures and sterol concentration, the heterogeneity of WAXS