656 JOURNAL OF COSMETIC SCIENCE together to form coiled-coil dimers, intermediate filaments, and macrofibrils (2). To achieve a rigid hair structure, they form a strong compact network by both intra- and intermolecular interactions such as disulfide bonds, hydrogen bonds, electrostatic salt bonds, and amide bonds (3). Of major interest in hair science are the protein cross-links, such as disulfide bonds, which lead to the formation of covalent linkages between intermolecular amino acid residues and hence determine which regions within peptides and proteins are involved in the formation of protein networks. In addition to these native protein cross-links in hair, artificial cross-links, including lanthionine and lysinoalanine, can also be induced, mostly on insult such as heat, UV, or the use of alkaline products (4). Networks of protein–protein cross-links underpin all the key physicomechanical properties of mammalian fibers and strongly influence how the fiber responds to treatments and environmental insults. Understanding and controlling these networks is therefore critical to enhance our understanding of fiber science generally and develop new fiber treatments that enhance properties. However, due to the large degree of cross-linking present in mammalian fibers, analysis of these covalent bonds has been very difficult. Although some technologies have successfully led to the detection of cross-links (3), most techniques involve total protein destruction, making mapping where these cross-links are within and between proteins extremely difficult. In this proof-of-concept study, mass spectrometry–based proteomics strategies are used to unravel the details of cross-link location between trichocyte keratin proteins in the hair shaft. This work focuses on two cross-links known to exist in hair proteins and that are used as a proxy for the degree of damage in hair proteins: lanthionine and lysinoalanine. Both cross-links can be the result of degraded disulfide bonds to form dehydroalanine, or alternatively, of the elimination of water or hydrogen sulfide from serine and cysteine, respectively, to form dehydroalanine. This dehydroalanine intermediate will react with either lysine to form lysinoalanine or with cysteine to from lanthionine. Although mass spectrometry has been used before to elucidate protein cross-links in a food or medical context (5,6), this work, to our knowledge, is the first to decipher cross-links in the complex hair matrix with the aim of finding mass spectrometric evidence of the existence of cross-links within hair fibers and mapping their exact location within the protein in situ. METHODS SAMPLE PREPARATION Blended Caucasian virgin hair tresses were obtained from International Hair Importers & Products, Inc. (Glendale, NY, United States). Full-length fibers were cut into snippets of approximately 5 mm in length. For amino acid analysis, two types of samples were prepared: first, 10 mg of hair snippets were weighted, and second, hair protein extracts were prepared by using 100 mg of hair snippets, which were dissolved in 5 mL of an extraction buffer (7 M urea, 2 M thiourea, 0.05 M Tris, and 50 mM dithiothreitol (DTT), pH 7.5) at room temperature for 18 h using vigorous shaking. For proteomics analysis, 10 mg of hair snippets were dissolved in 1 mL of an extraction buffer (7 M urea, 2 M thiourea, 0.05 M Tris, and 50 mM DTT, pH 7.5) at room temperature for 18 h using vigorous shaking.

657 MAPPING PROTEIN CROSS-LINKS IN HUMAN HAIR PROTEIN DIGESTION After the protein extraction, an acetone precipitation was carried out to remove the extraction buffer components by adding 5 mL of ice-cold acetone to 1 mL of the sample. The samples were stored at –20°C for 48 h before they were centrifuged at 10,000 g and 4°C for 10 min. The protein pellet was washed twice with ice-cold acetone, air-dried, and resuspended in 200 μL of 10 mM ammonium bicarbonate. The protein concentration per sample was determined by Direct Detect (Millipore, MA, United States) according to manufacturer’s instructions. Next, 100 μg of each sample was aliquoted, and proteins were digested by sequencing grade trypsin (Promega) at a 1:50 ratio (trypsin:protein) for 18 h at 37°C at 300 rpm in a thermomixer. Next, the tryptic digest peptide mixtures were stored at –20°C until further analysis. AMINO ACID ANALYSIS The hair snippets and the tryptic digests were hydrolyzed with 6 N HCl at 110°C for 24 h. Amino acid analysis was performed using a Dionex HPLC system (Dionex UltiMate 3000 with Fluorescence Detector, Thermo Scientific, San Jose, CA, United States). Free amino acid residues and standards (containing lanthionine and lysinoalanine at a predetermined concentration) were derivatized with an AccQ-Tag reagent (Waters Corporation, Milford, MA, United States) and separated using a Thermo Accucore XL C18 column (4.6 mm i.d. × 250 mm, 4 μm particles) (Thermo Scientific, San Jose, CA, United States) at 37°C using a flow rate of 1.0 mL/min. Eluent A was AccQ-Tag solvent A, and eluent B was 100% acetonitrile. An excitation wavelength of 250 nm and emission wavelength of 395 nm were used for the quantitative analysis using a fluorescence detector (Dionex Ultimate FLD 3000, Thermo Scientific, San Jose, CA, United States). LIQUID CHROMATOGRAPHY–TANDEM MASS SPECTROMETRY To detect the cross-links via mass spectrometry, the peptide mixtures were separated on an Ultimate 3000 RSLCnano system (Thermo Scientific, San Jose, CA, United States) using a C18 PepMap100 nano-Trap column (200 μm i.d. × 2 cm) (Thermo Scientific, San Jose, CA, United States) connected to a ProntoSIL C18AQ analytical column (100 μm i.d. × 15 cm, 3 μm particle size, 200 A pore size) (nanoLCMS Solutions, Gold River, CA, United States). For each sample 1 μL (equal to 1 μg) was loaded on the trap column at 3 μL/min. The trap column was then switched in line with the analytical column. Reverse phase liquid chromatography (LC) separation was performed by a linear gradient of mobile phase B (0.1% formic acid in 100% acetonitrile) from 2% to 45% in 60 min, followed by a steep increase to 98% mobile phase B in 6 min, held at 95% mobile phase B for 2 min, returned to 2% mobile phase B over 5 min, and then re-equilibrated at 2% mobile phase B for 15 min resulting in a total run of 88 min used at a flow rate of 1,000 nL/min. The LC was coupled online to mass spectrometry via a Bruker CaptiveSpray ion source (Bruker Daltonics, Bremen, Germany) equipped with a nanobooster device and operated at 1,400 V. Data were acquired with a Q-TOF Impact II (Bruker Daltonics, Bremen, Germany) mass spectrometer in a dynamic data-dependent auto-MS/MS mode where a full scan spectrum (150–2,200 m/z, 2 Hz) was followed by a dynamic inclusion of collision-induced dissociation