HAIR PHOTODAMAGE 111 Fluorescence measurements were performed on an MPF-66 fluorescence spectrophotom- eter (Perkin Elmer Corp.) using the solid-sample accessory. The spectrophotometer is controlled with a Perkin Elmer 7000 computer running the PECLS software. Besides data acquisition, this software also allows limited mathematical data manipulation. Emission measurements were made with the excitation and the emission slits routinely set at nominal 1- and 5-nm band passes, respectively, to minimize photodamage during the measurements. The sample excitation wavelength was dictated by the experimental design. Typically, for measuring Trp emission, the excitation wavelength was set at 295 nm to avoid contribution from other aromatic residues. All spectra were measured in the ratio mode to correct for lamp intensity fluctuations. The spectra, however, were not corrected for the instrumental response. For strongly fluorescent samples, the emission was reduced using screen attenuators in the light path. In those experiments in which the fluorimeter was also used for initiating photodamage, the excitation slit was opened wide (10-nm band-pass) for the required time. The excitation slit was then cut down to 1 nm prior to measurement of the fluorescence emission. Typically, hair fibers were moistened with distilled water and finely chopped with a razor blade prior to mounting in the cell. Wet hair is easier to cut, does not build up a static charge during cutting, and makes good optical contact in the cell, which reduces unwanted light scattering during the fluorescence measurements. We have found that a typical run requires --50 mg of hair. A further increase in the amount of hair did not affect the emission intensity. Percent Trp damage was calculated as [(I o - It)/I o] * 100, where I t and I o represent, respectively, the Trp emission intensities at any time t and at zero time (no damage). The emission intensities have been assumed to be directly proportional to Trp concen- tration. To evaluate the role of water in hair Trp photodestruction, the hair samples were dried in a vacuum desiccator for four days prior to measurements. Also, the cell chamber in the fluorimeter contained a desiccant. As noted earlier, the spectral quality is quite poor when dry fibers are measured. This is primarily due to the scattering artifacts resulting from poor optical contact between individual fibers and between the quartz front plate and the sample. FTIR spectra were measured in an IR-PLAN TM microscope (Spectra Tech. Inc.) con- nected to an FTIR spectrometer (Model 1760-X, Perkin Elmer Corp.). Hair samples, --5 mm long, were prepared by flattening between two steel plates at 18 KPsi. The sample was then placed between 2-mm NaC1 windows of a !x-plan compression cell (Spectra Tech. Inc.). A small crystal of KBr was included with the sample to serve as a reference against which the sample was ratioed. All spectral measurements were made at ambient temperature and humidity, unless otherwise stated. RESULTS AND DISCUSSION PHOTODECOMPOSITION Initial data, suggesting that hair weathering results in damage to tryptophan, were obtained by comparing the fluorescence spectra of the sun-bleached tip ends (blond) and the unexposed root ends (brown) of hair fibers from a 7-year-old Caucasian boy. The

112 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS brown hair of this child, as of many dark-haired people, becomes streaked with blond on the surface in summer. We found that excitation of the pigmented (brown) hair at 295 nm resulted in a strong emission at •340 nm. This band, however, was absent in the blond hair spectrum, as shown in Figure 1. Although this finding was initially surprising, familiarity with protein fluorescence suggested that we may, indeed, be measuring Trp in hair under these experimental conditions. This paradigm, indeed, turned out to be correct when we compared the fluorescence emission from the brown hair (above) with authentic L-tryptophan powder under identical experimental condi- tions, as seen in Figure 2. The complete overlap of the two emission spectra convinc- ingly shows that excitation at this wavelength selectively probes the Trp in hair. It should be mentioned that the unpigmented (white) hair also showed this band, ruling out its association with melanin pigment. An additional feature of the spectrum of damaged hair is the presence of a weak band at •450 nm. A similar band has also been observed in wool (23) and is probably due to a product of Trp photoxidation. We have also simulated hair Trp photodamage in a fluorimeter. In this experiment the irradiation wavelength could be precisely selected by using the monochromator of the instrument. Also, since the sample cell remains in the same position at all times, the experimental artifacts associated with reproducibly filling and positioning the cell are eliminated. Piedmont hair was exposed to 295-nm light for various times (see Exper- imental section for details). The emission spectra, recorded after each irradiation step, are presented in Figure 3. A gradual lowering of the emission intensity in the 338-360- nm range is related to a decrease in the concentration of Trp as a result of photode- 310 3•0 4i0 4•0 5i0 560 Wavelength (nm) Figure 1. Fluorescence emission spectra of virgin brown (T.J.) and naturally weathered (blond) hair from the same source. The excitation wavelength was 295 nm and the excitation and emission slits were both 5 nm. Notice the loss of --340 nm band upon hair weathering.