FTIR spectroscopic studies of lipid dynamics in phytosphingosine ceramide models of the stratum corneum lipid matrix

Chemistry and Physics of Lipids 134 (2005) 51–58

FTIR spectroscopic studies of lipid dynamics in phytosphingosine ceramide models of the stratum corneum lipid matrix Mark E. Rereka,∗ , Dina Van Wycka , Richard Mendelsohnb , David J. Moorea,1 a

International Specialty Products, 1361 Alps Road, Wayne, NJ 07040, USA b Rutgers University, Department of Chemistry, Newark, NJ 07102, USA

Received 5 October 2004; received in revised form 14 December 2004; accepted 14 December 2004 Available online 6 January 2005

Abstract IR spectroscopic studies are reported for N-stearyl-d-erythro-phytosphingosine (Cer NP) and N-stearyl-2-hydroxy-d-erythrophytosphingosine (Cer AP) in a hydrated model of the skin lipid barrier comprised of equimolar mixtures of each ceramide with cholesterol and d35 -stearic acid. Examination of the methylene stretching, rocking and bending modes reveal some rotational freedom and hexagonal packing in both the ceramide and stearic acid chains. Analysis of the acid carbonyl stretch and the ceramide Amide I modes show both shift to higher frequencies, indicating weaker hydrogen bonding, in the mixed systems compared to the pure materials. For both systems, the fatty acid chain disordering temperatures are significantly increased from those of the pure acids. The observed behaviors of these phytosphingosine ceramide systems are fundamentally different from the previously reported analogous sphingosine ceramide systems. The implications of these observations for lipid organization in the stratum corneum are briefly discussed. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Ceramides; Biophysics; Domains; Skin barrier; Spectroscopy

1. Introduction ∗

Corresponding author. Present address: Reheis, Inc, 235 Snyder Avenue, Berkeley Heights, NJ 07922, USA. Tel.: +1 908 219 5288; fax: +1 908 219 5288. E-mail addresses: [email protected] (M.E. Rerek), [email protected] (D.J. Moore). 1 Tel.: +1 9738724307; fax: +1 9736283401.

Human terrestrial life requires the presence of a skin permeability barrier that maintains water homeostasis by regulating water loss. This skin barrier function is accomplished by a remarkably thin layer (∼10–20 ␮m) of the outer epidermis; the stratum corneum (SC). A primary function of the viable epidermis is the genera-

0009-3084/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.chemphyslip.2004.12.002

52

M.E. Rerek et al. / Chemistry and Physics of Lipids 134 (2005) 51–58

tion of the SC with its attendant water barrier function. The SC also provides a significant barrier to a wide range of environmental chemical and biological toxins, as well as aiding in temperature regulation and UV protection. At the macroscopic level, the two major components of the SC are a highly cross-linked protein cell network and a continuous intercellular lamellar lipid matrix, organized materially in a manner analogous to a brick wall (Elias, 1991; Schaefer and Redelmeier, 1996). Various experimental approaches have clearly demonstrated that the water barrier function resides primarily in the hydrophobic lamellar lipid matrix of the SC (Wertz and van den Bergh, 1998; Williams and Elias, 1987). As a consequence of the SC’s critical physiological function, understanding barrier lipid composition and organization has become a necessary objective in skin research. The major SC lipid components are long chain free fatty acids, cholesterol, and ceramides (sphingosines and phytosphingosines), present in approximately equimolar concentrations (Norlen et al., 1998; Robson et al., 1994; Stewart and Downing, 1999). This unusual lipid composition for a biological membrane, particularly the lack of any phospholipids, is required to satisfy the unique physiological role of the SC. Recent biophysical reports have highlighted the complexity of lipid organization in the SC, leading to several models of SC lipid organization (Forslind, 1994; Bouwstra et al., 2000; Norlen, 2001). However, a detailed molecular description of SC lipid organization, based on biophysical spectroscopic measurements has only recently begun to emerge. Our laboratories are engaged in a series of studies aimed at describing intermolecular and intramolecular lipid organization in SC model membranes, ranging in complexity from pure ceramides to ternary component models of the SC lipid matrix (Moore et al., 1997a, 1997b; Moore and Rerek, 2000; Chen et al., 2000). Lipid organization and dynamics in the skin barrier are a consequence of chain conformational order and packing, as well as hydrogen bonding interactions. We have recently described major differences in intermolecular H-bonding, water penetration, and chain behavior between hydrated phytosphingosine and sphingosine ceramides (Rerek et al., 2001). Experimental evidence from a range of techniques (DSC, X-ray, Raman) indicates that phytosphingosine ceramide samples, in various mixtures and degrees of hydration, be-

have distinctly from the corresponding sphingosine ceramide systems (Ohta and Hatta, 2002; Raudenkolb et al., 2003a, 2003b). Particularly interesting is that phytosphingosine ceramides, in contrast to sphingosine ceramide systems, appear to form mixtures with stearic acid (Ohta and Hatta, 2002). This, and other studies, suggests that each ceramide class uniquely contributes to the cohesion to the SC lipid matrix (Glombitza and Muller-Goymann, 2002). The current study describes our Fourier transform infrared (FTIR) spectroscopy investigations of two phytosphingosine ceramides, Cer NP and Cer AP, in hydrated combinations with cholesterol and perdeuterated octadecanoic (stearic) acid. This FTIR spectroscopy study complements recently reported DSC and X-ray diffraction studies of anhydrous phytosphingosines by providing the addition molecular information inherent in IR spectroscopy measurements.

2. Experimental 2.1. Materials N-Stearoyl-d-erythro-phytosphingosine (Cer NP) and N-stearoyl-2-hydroxy-d-erythro-phytosphingosine (Cer AP), equivalent to human ceramides 3 and 7, respectively in the old nomenclature (as described in Stewart and Downing, 1999) were obtained from Cosmoferm (CT). The chemical structures of these ceramides are shown in Fig. 1. Deuterated stearic acid (d35 -stearic acid) was purchased from CDN Isotopes (Montreal, Canada) and cholesterol was purchased from Sigma (St. Louis, MO). 2.2. FTIR spectroscopy All samples containing were prepared by heating to 120 ◦ C (melting point for hydrated Cer NP) for 30 min in sealed culture tubes with 2 H2 O buffer at pH 5.5. Throughout the heating cycle, samples were periodically vortexed to ensure complete melting and mixing of all components. Following mixing and hydration, samples were sandwiched between heated IR windows and mounted in a temperature controlled transmission cell holder (Harrick Scientific, Ossining, NY). Fourier transform infrared (FTIR) spectra were collected on a Mattson spectrometer equipped with a

M.E. Rerek et al. / Chemistry and Physics of Lipids 134 (2005) 51–58

53

Fig. 1. Chemical structures of the ceramides used in this study and their abbreviations.

broadband mercury-cadmium-telluride (MCT) detector. Spectra were generated by co-addition of 256 interferograms collected at 2 cm−1 resolution, apodized with a triangular function, and fourier transformed with one level of zero filling. Spectra were routinely acquired at 2 or 4 ◦ C intervals from 16 to 116 ◦ C. Data were analyzed off-line with software written at the National Research Council of Canada and plotted with Sigma Plot® 2000 (SPSS, IL). In several previous papers and reviews we have discussed in detail the molecular information inherent in FTIR spectra of ordered lipid phases, and particularly in ceramide–fatty acid systems (Moore et al., 1997a; Mendelsohn and Moore, 1998, 2000). The reader is referred to these studies for further information on the underling spectroscopic techniques.

3. Results 3.1. Hydrated N-stearoyl-2-hydroxy-d-erythro-phytosphingosine (Cer AP)/d35 -stearic acid/ cholesterol Fig. 2 shows the thermotropic behavior of the conformationally sensitive CH2 symmetric stretching

Fig. 2. Thermotropic changes in Cer AP νsym CH2 frequency and d35 stearic acid νasym CD2 frequency for the hydrated (2 H2 O), equimolar Cer AP, d35 -stearic acid, cholesterol skin lipid model system. The data are superimposed in the lower panel.

mode (νsym CH2 ) and CD2 asymmetric stretching mode (νasym CD2 ) for a hydrated (2 H2 O), equimolar mixture of Cer AP, d35 -stearic acid and cholesterol. The initial frequencies of νsym CH2 and νasym CD2 appear at 2849.6 and 2193.8 cm−1 respectively. The fatty acid νasym CD2 frequency exhibits a monotonic one wavenumber increase between 15 and 70 ◦ C while the ceramide νsym CH2 increases only 0.2–0.3 cm−1 . Between 70 and 90 ◦ C both modes undergo a coordinated transition with a mid-point of 80 ◦ C. From 90 to 110◦ the rate of frequency increase is markedly reduced, after which both the stearic acid νasym CD2 and ceramide νsym CH2 frequencies exhibit a final sharp, coordinated transition. The initial frequencies of Cer AP ␳CH2 (rocking mode) and d35 -stearic acid δCD2 (scissoring mode) are 722.2 and 1088.6 cm−1 , respectively, indicative of hexagonal chain packing (data not shown). No differences in behavior were

54

M.E. Rerek et al. / Chemistry and Physics of Lipids 134 (2005) 51–58

Fig. 3. Temperature-induced changes in fatty acid carbonyl stretching and ceramide Amide I spectral region for the hydrated (2 H2 O), equimolar Cer AP, d35 -stearic acid, cholesterol skin lipid model system.

observed upon repeated heating and cooling of the samples. The behavior of the polar groups as a function of temperature for the hydrated equimolar Cer AP, d35 stearic acid, and cholesterol sample is shown in Fig. 3. Initially, two distinct Amide I peaks are observed at 1629.5 and 1637.1 cm−1 for Cer AP, and a single sharp d35 -stearic acid carbonyl stretch (νCO) is observed at 1709.2 cm−1 . Both Amide I and acid νCO are at higher frequencies than when hydrated alone. No changes are observed until the onset of chain disordering at high temperatures, whereupon νCO shifts to lower frequency and Amide I broadens considerably. Upon cooling, the initially observed sharp peaks are regenerated. 3.2. Hydrated N-stearoyl-d-erythro-phytosphingosine (Cer NP)/d35 -stearic acid/ cholesterol Fig. 4 shows the thermotropic behavior of the νsym CH2 and νasym CD2 bands for a hydrated (2 H2 O), equimolar mixture of Cer NP, d35 -stearic acid and cholesterol. The initial frequencies of νsym CH2 and

Fig. 4. Thermotropic changes in Cer NP νsym CH2 frequency and d35 stearic acid νasym CD2 frequency for the hydrated (2 H2 O), equimolar Cer NP, d35 -stearic acid, cholesterol skin lipid model system. The data are superimposed in the lower panel.

νasym CD2 are 2849.4 and 2194.8 cm−1 , respectively. Upon heating, a gradual, coordinated chain disordering process with Cer NP and d35 -stearic acid begins at about 40 ◦ C and continues beyond 110 ◦ C. The frequencies of the Cer NP rocking mode (␳CH2 ), 721.3 cm−1 , and the d35 -stearic acid scissoring mode (δCD2 ), 1088.6 cm−1 , are characteristic of hexagonal packing (data not shown). Head group spectra for the system are shown in Fig. 5. Initially, Cer NP shows two peaks in the Amide I region at 1611.2 and 1631.7 cm−1 . The d35 -stearic acid νCO band is broad and centered at 1704.0 cm−1 . As in the Cer AP ternary system, both Amide I and acid νCO are observed at significantly higher frequencies than when hydrated alone. Increasing temperature results in the loss of the band at 1631.7 cm−1 ; the other head group bands are hardly changed. Further heating results in a broad absorbance for Amide I centered at

M.E. Rerek et al. / Chemistry and Physics of Lipids 134 (2005) 51–58

55

terol samples are profoundly different from the corresponding sphingosine ceramide systems. The most important difference being that ternary phytosphingosine ceramide samples appear to form intimate, homogeneous, mixtures in which separate domains are not observed. Previously, an anhydrous, equimolar mixture of Cer NPA, cholesterol and stearic acid was found to form a molecular mixture, as observed by X-ray diffraction, in which distinct chain domains were not observed. (Ohta and Hatta, 2002). A not unexpected consequence of this is that pure hydrated phytosphingosine ceramide behaves quite differently from that in the ternary samples. 4.1. Chain behavior

Fig. 5. Temperature-induced changes in fatty acid carbonyl stretching and ceramide Amide I spectral region for the hydrated (2 H2 O), equimolar Cer NP, d35 -stearic acid, cholesterol skin lipid model system.

about 1630 cm−1 ; the acid νCO band continues to be little changed.

4. Discussion In previous reports we observed that sphingosine ceramides behaved similarly when hydrated at pH 5.5 (equal to SC pH) either individually or in equimolar samples with cholesterol and d35 -stearic acid (Moore et al., 1997a, 1997b; Moore and Rerek, 2000). This was consistent with the observation of separate sphingosine and fatty acid domains in the ternary samples. In further studies we reported significant differences in chain packing, chain conformational order and headgroup hydrogen bonding between hydrated sphingosine and phytosphingosine ceramides (Rerek et al., 2001). The current study completes this set of experiments by examining the molecular dynamics of hydrated phytosphingosine ceramides, stearic acid, and cholesterol samples buffered at pH 5.5. The data indicate that molecular organization and dynamics in the phytosphingosine ceramides, stearic acid, and choles-

It was previously reported that hydrated phytosphingosine ceramides are conformationally ordered at room temperature, but not as ordered as sphingosine ceramides (Rerek et al., 2001; Raudenkolb et al., 2003a). This behavior is maintained in equimolar mixtures with cholesterol and d35 -stearic acid. Both phytosphingosine ceramides Cer NP and Cer AP in equimolar systems with perdeuterated stearic acid and cholesterol undergo gradual and largely coordinated chain disordering processes over large temperature ranges (Figs. 2 and 4). The most striking feature of these plots are that disordering midpoint temperature (Tm ) of d35 stearic acid is significantly raised in both the Cer AP and Cer NP hydrated ternary systems from that of pure d35 -stearic acid. For the hydrated Cer AP system, the ceramide Tm is also raised. Only Cer NP displays the expected decrease in chain Tm , but is still close to the value of the hydrated individual system. This increase in the chain Tm in both systems is surprising given their hexagonal, rather than orthorhombic, chain packing. Although the phytosphingosine ceramide and stearic acid chains are conformationally ordered, i.e., fully extended all-trans chains, hexagonal packing permits greater rotational movement as there is more room between the chains (unit cell = 1.5 molecules) than with the orthorhombic packing (unit cell = 2 molecules) observed for pure d35 -stearic acid (hydrated and non-hydrated). The observation of chain disordering across a very broad temperature range in which both acid and ceramide chains act either together, or very nearly so, for both phytosphingosine ceramide ternary samples suggests all components are

56

M.E. Rerek et al. / Chemistry and Physics of Lipids 134 (2005) 51–58

mixed at the molecular level, at least upon disordering. It could be argued that in both systems, the d35 -stearic acid begins an individual thermal transition at about 40 ◦ C as it does in the sphingosine ceramide ternary systems. However, as indicated by an inflection point at about 50 ◦ C, the d35 -stearic acid process then tracks the ceramide chain process. Broad thermal transitions like these in lipid mixtures are often characteristic of extensive cholesterol mixing. We have observed that hydration is necessary to achieve these effects and will report on comparisons between the hydrated and nonhydrated ternary systems. Both phytosphingosine ceramide and fatty acid chains are hexagonally packed in these ternary hydrated samples. This is not unexpected for the phytosphingosine ceramide chains since hexagonal chain packing was previously reported in X-ray diffraction studies on the non-hydrated equimolar ternary Cer NP system and for hydrated and non-hydrated pure phytosphingosine ceramides (Rerek et al., 2001; Ohta and Hatta, 2002; Raudenkolb et al., 2003a, 2003b). Segregated domains of d35 -stearic acid would be expected to show orthorhombic packing, however this is not observed initially nor after repeated heating and cooling cycles. Previously, we had found that the d35 stearic acid domain sizes in the ternary sphingosine systems were at least 100 molecules (Moore et al., 1997a). To exclude the possibility of mixed d35 -stearic acid–ceramide orthorhombic chain packing (in which factor group splitting would not be observed) we examined the corresponding Cer AP, h35 -stearic acid and cholesterol system and found no evidence of band splitting. Therefore, we conclude that the stearic acid chains are not organized in an orthorhombic crystalline packing arrangement (data not shown). All acid and ceramide chains observed in these studies have hexagonal chain packing. 4.2. Head group behavior In both the 2 H2 O hydrated Cer AP/d35 -stearic acid/cholesterol and Cer NP/d35 -stearic acid/cholesterol systems, ceramide Amide I and acid νCO bands are observed at higher frequencies (∼15 cm−1 ) than for the pure components, indicating the hydrogen bonding has changed for these components when mixed in the ternary systems. A likely explanation is that hydrogen bonding is occurring between d35 -stearic acid and

either the Cer NP or Cer AP, depending on the system. It is interesting to note that when the analogous hydrated, equimolar sphingosine ceramide/d35 -stearic acid/cholesterol systems are initially prepared, similar hydrogen bonding changes are observed in the presence of hexagonal chain packing. However, in the case of the sphingosine ceramide systems, they undergo a thermal rearrangement in which some degree of phase separation occurs such that acid νCO and ceramide Amide I are very similar to the pure components (accompanied by a change to orthorhombic chain packing). Very little change in the head group region is observed for the hydrated phytospingosine ceramide ternary systems until high temperatures (Figs. 3 and 5). It was previously concluded that hydrogen bonding was the predominant effect in the condensed phase behavior of individual phytosphingosine ceramides (Rerek et al., 2001). It appears that this is also true in these equimolar ternary systems, although it is difficult to conclude whether chain packing drives hydrogen bonding or hydrogen bonding drives chain packing. However, it is not unreasonable to assume that d35 stearic acid with a single chain or cholesterol can intermix with both ceramide chains more effectively than the phytosphingosine ceramides can mix with themselves. Monolayer studies have shown that the headgroup area of phytosphingosine ceramides is several ˚ 2 ) larger than the corresponding square angstroms (A sphingosines (Dahlen and Pascher, 1979). This more open headgroup structure apparently allows much better intermolecular mixing with fatty acids and cholesterol in contrast to the sphingosine ceramides, which effectively form their own tightly packed phase. 4.3. Implications for lipid organization in stratum corneum With this study, we have now examined four of the most common ceramides found in human stratum corneum. While our work has focused on three component model systems, we believe the results provide direct relevance toward understanding molecular interactions in the stratum corneum lipid matrix, where ceramides are present with long chain fatty acids and cholesterol in a roughly equimolar mixture. Table 1 summarizes the major findings from our studies, from which several conclusions can be drawn. The first, as has been discussed in detail here, is that there

M.E. Rerek et al. / Chemistry and Physics of Lipids 134 (2005) 51–58

57

Table 1 Ceramide and fatty acid chain behavior in ternary models Attribute

Cer NS

Cer NP

Cer AP

Cer AS

Stearic acid chain packing Change in stearic acid chain packing? Stearic acid chain disordering Tm Ceramide chain packing Change in ceramide chain packing? Ceramide chain disordering Tm Stearic acid/ceramide chain disordering Tm Stearic acid νCO frequency Ceramide Amide I frequency

Orthorhombic No Decreased Orthorhombic No Decreased Separate Similar Similar

Hexagonal Yes Increased Hexagonal No Decreased Similar Increased Increased

Hexagonal Yes Increased Hexagonal No Increased Similar Increased Increased

Orthorhombic No Decreased Orthorhombic No Decreased Similar Increased Similar

Summary of the biophysical behavior of phytosphingosine and sphingosine ceramide hydrated, ternary lipid systems containing d35 -stearic acid and cholesterol. Comparative statements are against the single species, i.e., either pure ceramide or fatty acid.

is a fundamental difference between the behavior of sphingosine and phytosphingosine ceramide hydrated, ternary lipid systems. The second conclusion is that it is the ceramide component that drives the behavior of the ternary lipid systems. All four ceramides have the same chain packing in the ternary lipid systems as they do when hydrated individually. When the sphingosines, Cer NS and Cer AS, pack in orthorhombic subcells (forming distinct domains) orthorhombic packing is also observed for stearic acid. However, when the phytosphingosines, Cer NP and Cer AP, pack in the more open hexagonal structure then stearic acid also adopts hexagonal packing. The broad, coordinated stearic acid and ceramide chain disordering transitions suggest this looser stearic acid packing results from the incorporation of stearic acid into the open ceramide matrix along with the cholesterol. It follows that it is the chain behavior of these ceramides that drive the ultimate organization. For the phytosphingosines, hydrogen bonding is altered in the three component systems for both the stearic acid and the phytosphingosine ceramides in a way that can be interpreted as lower in bond strength. Yet the stearic acid chain disordering midpoint temperatures are greatly increased, as is the Cer AP chain disordering Tm . Although little change is seen in both the sphingosine chain packing and hydrogen bonding, the ceramide chain transition Tm is lowered, presumably by solubilization with fluid stearic acid. Given the water barrier function of the stratum corneum, it is not surprising that hydrophobic interactions drive the functionality. Finally, these results show that Cer NS, the most widely studied ceramide in model systems, is the least

similar to the other three ceramides in Table 1 (Moore et al., 1997a, 1997b; Moore and Rerek, 2000; Chen et al., 2000; Rerek et al., 2001; Velkova and Lafleur, 2002; Mimeault and Bonenfant, 2002). To date, it is the only ceramide we have studied, that does not undergo a coordinated disordering transition with stearic acid. It is also the system in which the stearic acid behaves most closely to when hydrated individually inferring, therefore, that Cer NS is the least interacting ceramide of those studied to date. Conclusions on general ceramide behavior based exclusively on observations from Cer NS model systems should be made with caution. These results are also consistent with the fundamental ideas in Forslind’s Domain Mosaic model of skin barrier lipid organization (Forslind, 1994). Sphingosine ceramides form distinct domains. Forslind postulated that permeation occurred between these domains where both edge defects and areas of more liquid crystalline behavior were found. Phytosphingosine ceramides could decrease the permeability through these regions by interaction with other lipid components in them to both reduce the defect structure and order more liquid domains, thus making the mosaic more coherent. This resulting mosaic produces a structure that is relatively water impermeable, cohesive, and yet mechanically flexible. This may also explain why there are higher levels of sphingosine ceramides compared to phytosphingosine ceramides in the SC. A greater level of phytosphingosine cermides might diminish the edge defects and liquid lipid grain boundaries that allow the slow rate of transport thereby making the SC almost impenetrable. This could greatly inhibit the skin’s ability to regulate temperature, maintain homeostasis, and communicate sensory signals.

58

M.E. Rerek et al. / Chemistry and Physics of Lipids 134 (2005) 51–58

References Bouwstra, J.A., Dubbelar, F.E.R., Gooris, G.S., Ponec, M., 2000. The lipid organisation in the skin barrier. Acta Derm. Venereol. Suppl. 208, 23–30. Chen, H.-C., Mendelsohn, R., Rerek, M.E., Moore, D.J., 2000. Fourier transform infrared spectroscopy and differential scanning calorimetry studies of fatty acid homogeneous ceramide 2. Biochim. Biophys. Acta 1468, 293–303. Dahlen, B., Pascher, I., 1979. Molecular arrangements in sphingolipids. thermotropic phase behaviour of tetracosanoylphytosphingosine. Chem. Phys. Lipids 24, 119–133. Elias, P.M., 1991. Epidermal barrier function: intercellular lamellar lipid structures, origin, composition and metabolism. J. Controll. Release 15, 199–208. Forslind, B., 1994. A domain mosaic model of the skin barrier. Acta Derm. Venereol. 74, 1–6. Glombitza, B., Muller-Goymann, C.C., 2002. Influence of different ceramides on the structure of in vitro model lipid systems of the stratum corneum lipid matrix. Chem. Phys. Lipids 117, 29–44. Mendelsohn, R., Moore, D.J., 1998. Vibrational spectroscopic studies of lipid domains in biomembranes and model systems. Chem. Phys. Lipids 96, 141–157. Mendelsohn, R., Moore, D.J., 2000. Infrared determination of conformational order and phase behavior in ceramides and stratum corneum models. In: Hannun, Y.A., Merrill, A.H. (Eds.), Methods in Enzymology: Sphingolipid Metabolism and Cell Signalling, 312. Academic Press, New York. Mimeault, M., Bonenfant, D., 2002. FTIR spectroscopic analyses of the temperature and pH influences on stratum corneum lipid phse behaviors and interactrions. Talanta 56, 395–405. Moore, D.J., Rerek, M.E., Mendelsohn, R., 1997a. Lipid domains and orthorhombic phases in model stratum corneum: evidence from Fourier transform infrared spectroscopy studies. Biochem. Biophys. Res. Commun. 231, 797–801. Moore, D.J., Rerek, M.E., Mendelsohn, R., 1997b. FTIR spectroscopy studies of the conformational order and phase behavior of ceramides. J. Phys. Chem. B 101, 8933–8940. Moore, D.J., Rerek, M.E., 2000. Insights into the molecular organization of lipids in the skin barrier from infrared spectroscopy

studies of stratum corneum lipid models. Acta Derm. Venereol. Suppl. 208, 16–22. Norlen, L., Nicander, I., Lundsjo, A., Cronholm, T., Forslind, B., 1998. A new HPLC-based method for the quantitative analysis of inner stratum corneum lipids with special reference to the free fatty acid fraction. Arch. Dermatol. Res. 290, 508–516. Norlen, L., 2001. Skin barrier structure and function: the single gel phase model. J. Invest. Dermatol. 117, 830–836. Ohta, N., Hatta, I., 2002. Interaction among molecules in mixtures of ceramide/stearic acid, ceramide/cholesterol and ceramide/stearic acid/cholesterol. Chem. Phys. Lipids 115, 93–105. Raudenkolb, S., Hubner, W., Rettig, W., Wartewig, S., Neubert, R.H.H., 2003a. Polymorphism of ceramide 3. Part 1: an investigation focused on the head group of Noctadecanoylphytosphingosine. Chem. Phys. Lipids 123, 9–17. Raudenkolb, S., Wartewig, S., Neubert, R.H.H., 2003b. Polymorphism of ceramide 3. Part 2: A vibrational spectroscopic and X-ray powder diffraction investigation of N-octadecanoyl phytosphingosine and the analogous specifically deuterated d(35) derivative. Chem. Phys. Lipids 124, 89–101. Rerek, M.E., Chen, H-C., Markovic, B., Van Wyck, D., Garidel, P., Mendelsohn, R., Moore, D.J., 2001. J. Phys. Chem. B 105, 9355–9362. Robson, K.J., Stewart, M.E., Michelsen, S., Lazo, N.D., Downing, D.T., 1994. 6-Hydroxy-4-sphingenine in human epidermal ceramides. J. Lipid Res. 35, 2060–2068. Schaefer, H., Redelmeier, T.E., 1996. Skin Barrier: Principles of Percutaneous Absorption, first ed. Karger, Basel. Stewart, M.E., Downing, D.T., 1999. A new 6-hydroxy-4sphingenine-containing ceramide in human skin. J. Lipid Res. 40, 1434–1439. Velkova, V., Lafleur, M., 2002. Influence of the lipid composition on the organization of skin lipid model mixtures: an infrared spectroscopy investigation. Chem. Phys. Lipids 117, 63–74. Wertz, P.W., van den Bergh, B., 1998. The physical, chemical and functional properties of lipids in the skin and other biological barriers Chem. Phys. Lipids 91, 85–96. Williams, M.L., Elias, P.M., 1987. The extracellular matrix of stratum corneum: role of lipids in normal and pathological function. Crit. Rev. Ther. Drug Carrier Syst. 3, 96–122.