HAIR LIPID COMPOSITION 5 at a flow rate of 1.0 ml/min with a gradient elution program, which included first holding (A) 100% for 15 min, second (B) 5%/min to (B) 100% (20 min), and finally holding of (B) 100% for 15 min. The content of sphingoids was calibrated based on area ratios of the peaks of interest to the IS (threo-dihydrosphingosine), resulting in CER content. Duplicate analyses were performed for each measurement. STATISTICAL ANALYSES Correlation analyses among the level of lipids and a principal component analysis (PCA) with a correlation matrix for hair lipid composition were performed using Excel Tahenryo-Kaiseki version 3.0 software (Esumi, Tokyo, Japan). RESULTS MOLECULAR SPECIES OF CERs IN HUMAN HAIR A total ion GC/MS chromatogram of silylated CERs in extractable lipids is shown in Figure 1, and those in integral lipids exhibited almost the same peak pattern. Peaks A-F in Figure 1 had distinctive fragments: (A) m/z = 313, 3 70, (B) m/z = 299, 313, 458, (C) m/z = 313,398, (D) m/z = 313,426, (E) m/z = 313,454, and (F) m/z = 313,480. It has been established that a mass spectrum of silylated CERs can be characterized by a C2-C3 fragmentation of its sphingoids amidified with its N-acyl moiety (19,25). Therefore, fragments of m/z = 311 or m/z = 313 and m/z = [M-311] or m/z = [M-313] provide sufficient information about their sphingoids such as sphingenine (sphingosine) or sphinganine (dihydrosphingosine) and about their N-acyl moieties. Accordingly, the peaks were identified as follows: (A) 2-N-palmitoylamino-octadecane-1,3-diol, (B) 2-N­ a-hydroxypalmitoylamino-octadecane-1,3-diol, ( C) 2-N-stearoylamino-octadecane-1,3- 15 B A C 20 F D E TlmMnn 2!5 A B C D E F �o � H �o �: �o � OH �OH �o � OH OH 0 � Figure 1. A GC/MS total ion chromatogram of silylated CERs from extractable lipids. A, B, C, D, E, and F show their molecular structure.

6 JOURNAL OF COSMETIC SCIENCE diol, (D) 2-N-eicosanoylamino-octadecane-1,3-diol, (E) 2-N-docosanoylamino­ octadecane-1,3-diol, and (F) 2-N-tetracosanoylamino-octadecane-1,3-diol. DETERMINATION OF THE HAIR LIPIDS BY CHROMATOGRAPHY The optimal conditions for extraction of hair lipids, excluding contaminated lipids, were examined for the relationship between hexane incubation time after twice shampooing hair fibers and the extracted lipid contents. When the square root of time was plotted against the relative extracted lipid content, a linear plotting was obtained during 5 to 30 min of incubation with hexane (Figure 2). In general, the following equation is used for the diffusion of a solute into a cylindrical fiber of infinite length (26): C/C00 = 4(Dt/(Tir2))112 where C/C00 represents relative solute uptake at a certain time (t), D is the diffusion coefficient, and r is the radius of the cylindrical fiber. According to this equation, if diffusion of the solute occurs, the relative solute uptake plotted against the square root of time should become linear because of the constant D, 'IT, and r. The observed linear line (Figure 2) implies the existence of diffusion of the lipids from the hair inside to the outside during the incubation. Therefore, in our study, a 5-min incubation with hexane was performed prior to the extraction of hair lipids to delete contaminated lipids. 0.4 0.0 0 30 min 5 min 10 min 20 min t f 2min t t . ............ ···· .... + ··········· 1 min + 1 2 3 4 5 Root square of hexane incubation time 6 min112 Figure 2. Effect of hexane incubation time after twice shampooing of hair fibers on extracted lipid contents. Each point represents the mean ± SD calculated in five hair samples for each hexane incubation time. Relative extracted lipid content was defined as a ratio of lipid contents extracted at a certain time (t) to that of total lipids exhaustively extracted.