OA p C E a b c d e f g h A R R A A K P R D P D A 1 m i t c p e v 1Journal of Cultural Heritage 16 (2015) 896–903 Available online at ScienceDirect www.sciencedirect.com riginal article pplication of redox proteomics to the study of oxidative degradation roducts in archaeological wool aroline Solazzoa,b,∗, Stefan Clerensb, Jeffrey E. Plowmanb, Julie Wilsonc,d, lizabeth E. Peacocke,f, Jolon M. Dyerb,g,h BioArCh, Biology (S Block), Wentworth Way, University of York, York YO10 5DD, United Kingdom AgResearch, Proteins and Biomaterials, Lincoln Research Centre, Private Bag 4749, Christchurch NZ 8140, New Zealand Department of Mathematics, University of York, York YO10 5YW, United Kingdom Department of Chemistry, University of York, York YO10 5YW, United Kingdom NTNU Museum, Norwegian University of Science and Technology, NTNU 7491 Trondheim, Norway Department of Conservation, University of Gothenburg, SE-405 30 Gothenburg, Sweden Biomolecular Interaction Centre, School of Biological Sciences, University of Canterbury, Private Bag 4800, Christchurch 8140, New Zealand Riddet Institute, Massey University, PB 11 222, Palmerston North 4442, New Zealand a r t i c l e i n f o rticle history: eceived 19 November 2014 eceived in revised form 9 February 2015 ccepted 10 February 2015 vailable online 7 March 2015 eywords: hoto-oxidation edox proteomics eamidation roteins yes lum mordant a b s t r a c t Most archaeological and historical textiles (clothing, tapestries, blankets, carpets, etc.) present traces of UV-induced damage when exposed to light during their lifetime. Yellowing of the fibres, fading of the dyes and loss of physical properties, such as tensile strength are the typical indicators of photodegradation. Natural fibres made of proteins, such as wool and silk are particularly sensitive to UV damage. Photo- oxidative damage is caused by the accumulation of chemical modification at the amino acid residue level that lead to a range of oxidation products, including chromophores responsible for changes in coloration, as well as to the breaking of peptide bonds in the protein backbone. Amino acid residues with aromatic side-chain groups are particularly sensitive to photo-oxidation and breakthroughs have been made in recent years in the field of protein science to identify the photoproducts and locate them within proteins. This study explores new methodologies using redox proteomics-based strategies to assess the extent of photodamage in ancient wool textiles, by identifying modifications occurring at the molecular level. Using a scoring system to determine the level of oxidation in amino acids with aromatic side-chains (tryptophan, tyrosine, histidine and phenylalanine), we compare the effects of dyes and mordants on fibres after UV ageing, and assess the extent of oxidation on the different proteins composing the wool fibres. We determine that dyes and mordants have the capability of slowing down photo-oxidation during ageing. We also assess the effect of UV irradiation on deamidation, a modification targeting glutamine and asparagine, as it is a common marker of ageing in ancient proteins. © 2015 Elsevier Masson SAS. All rights reserved.. Research aims Important breakthroughs using redox proteomics have been ade in recent years to identify the coloured photoproducts formed n proteins exposed to UV light (chromophores) [1]. A redox pro- eomics approach is based around characterisation of the complex ascade of oxidation and reduction events occurring at the protein rimary level [2]. Proteomics is underpinned by mass spectrom- try, with electrospray ionization mass spectrometry (ESI-MS) ∗ Corresponding author. Smithsonian’s Museum Conservation Institute, 4210 Sil- er Hill Road, Suitland MD 20746, United States of America. Tel.: +1 301 238 1284. E-mail address: solazzo.c@gmail.com (C. Solazzo). http://dx.doi.org/10.1016/j.culher.2015.02.006 296-2074/© 2015 Elsevier Masson SAS. All rights reserved.and matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) the predominant modes used [3]. A redox proteomic approach consisting of: • digestion with a proteolytic enzyme to produce peptides; • separation of the peptides with appropriate liquid chromatogra- phy; • tandem mass spectrometric peptide fragmentation; • targeted bioinformatic evaluation for key redox products can provide detailed identification and location of modifications throughout the wool proteome. In this study, these relatively new tools are applied to investigate changes occurring at the amino acid level in keratins when wool has ltural b l i u i e a t a p i 2 t P [ a s w i o i a h p i fi p o o ( t a dC. Solazzo et al. / Journal of Cu een exposed to UV light. Changes in wool after exposure to UV ight for up to 48 h were evaluated by mass spectrometry. Chem- cal changes (oxidation and deamidation) were observed in both ndyed and dyed samples. Fabrics buried for more than eight years n North-European experimental burial sites [4] offered a basis for valuating possible increases in oxidation in wool by simulating rchaeological biodegradation. Finally, we tracked these modifica- ions in actual archaeological finds from medieval sites in the UK nd Iceland to determine whether it was possible to identify known roducts typically associated with photo-oxidation in archaeolog- cal samples after hundreds of years of burial. . Introduction Exposure to natural light is one of the major factors that con- ribute to fibres’ fragility and loss of coloration in ancient textiles. hotodegradation of wool by UV light is associated with yellowing 5,6], fading of the dyes [7–9] and loss of physical properties such s tensile strength [10]. Not only the length and intensity of expo- ure to UV light produce long-term damage, the treatment of fibres ith dyes and mordants can also influence photodegradation pos- tively or negatively [5,11], either by improving the photostability f the wool or by increasing the level of phototendering resulting n loss of strength and flexibility [12]. The presence of trace met- ls (for example iron and copper) that increase the production of ydroxyl radicals also influence wool photostability and accelerate hotoyellowing [13,14]. Photoyellowing results from exposure to high-energy UV light n the 320–400 nm (UVA) and 280–320 nm (UVB) range, while ltering of UVA and UVB can lead to the competing process of hotobleaching dominating (blue light 400–460 nm) [5]. Photo- xidation is initiated via radical species that react with atmospheric xygen and produce peroxide radicals. Reactive oxygen species ROS) attack both amino acid residue side-chains and the pro- ein backbone itself. When dyed, photochemical reactions are lso transferred to the dyes as UV radiation is absorbed by the ye molecules. Dye chromophores are destroyed, resulting in dye Fig. 1. Pathways of oxidation showing the resulting produ Heritage 16 (2015) 896–903 897 fading (photofading). Dyes can be affected by UV light or by photo- products of the substrate itself that react with the dyes. In addition, photodegradation leads to peptide chain scission, as well as cleav- age of disulphide bridges, while it might inversely contribute to the formation of cross-links that will increase tensile strength but may also result in increased brittleness. Redox proteomic-based evaluation of wool has found that oxi- dation products of phenylalanine, tryptophan and tyrosine are mainly responsible for the discoloration of wool due to the sus- ceptibility of aromatic side-chains to oxidation [3,15–17]. The level of oxidation was calculated according to a classification system of the products resulting from the oxidation of the aromatic amino acids, while deamidation was also calculated in key peptides (see below). Supplementary Table S1 summarises the samples and the associated experiments. 2.1. Calculation of the oxidation score MS/MS data were obtained by nanoLC-ESI-MS/MS to locate the induced oxidative modifications in the aged, buried modern and archaeological fibres. Based on reported photomodifications to aro- matic amino acid residues [3,15,17], single and double oxidation on aromatic residues (tyrosine, tryptophan, histidine and pheny- lalanine), quinone and hydroxyquinone (oxidized tyrosine) and kynurenine and hydroxykynurenine (oxidized tryptophan) were chosen as variable modifications (Fig. 1). The degree of oxida- tive degradation for each modified peptide has been evaluated by assigning a score to each individual observed oxidative mod- ification within the peptide based on the relative level of the modification within this oxidative cascade. Scores were assigned as 1 for those modifications classified as single oxidation, 2 for double oxidation, 3 for quinone and kynurenine formation and 4 for hydroxyquinone and hydroxykynurenine formation. The score given to each modification reflects the relative level of modification with respect to the native residue; with initial oxi- dation products being further modified themselves in a cascade of degradation (Fig. 1). Quinone, hydroxyquinone, kynurenine and cts of oxidation of (a) tryptophan, and (b) tyrosine. 898 C. Solazzo et al. / Journal of Cultural Heritage 16 (2015) 896–903 Table 1 Deamidation markers identified by MALDI-TOF-MS: sequence, theoretical [M+H]+, modifications, and deamidation sites [18]. Sequence [M+H]+ Modifications Deamidation sites QNQEYQVLLDVR 1487.74 Gln->pyro-Glu N-term Q or Glu->pyro-Glu N-term Q Q1-N2-Q3-Q6 QNQEYQVLLDVR 1504.77 Q1-N2-Q3-Q6 h a 2 m r b r f l t p d a 3 3 w 4 f a i b d d w s W t t a d f fi ( o a o t 3 ( U ( b r p a sLNVEVDAAPTVDLNR 1625.84 TVNALEVELQAQHNLR 1834.97 ydroxykynurenine are the most important modifications as they re the chromophores primarily responsible for yellowing. .2. Calculation of deamidation Deamidation rates were calculated for selected peptides com- only observed by peptide mass fingerprinting and previously eported (Table 1) [18]. Peptide mass fingerprints were obtained y MALDI-TOF-MS. The deamidation of glutamine or asparagine esults in a mass shift of + 0.984 Da so that the isotope distributions or the deamidated and non-deamidated states of a peptide over- ap. However, comparison with the theoretical distribution allows he percentage of each to be determined [19]. We refer to the ercentage of non-deamidated peptide as %Gln–Asn. The level of eamidation was calculated on peptides previously characterised nd described in Solazzo et al. [18]. . Experimental .1. Dyeing and mordanting Raw wool from Icelandic sheep was obtained from the Honor- ood flocks (Cefn Llanfair Llanfair Road, Llandysul Ceredigion SA44 RB). Alum (AlK(SO4)2·12H2O), madder and weld were obtained rom the mulberry dyer shop (http://www.mulberrydyer.co.uk/) nd mordanting and dyeing was done according to traditional dye- ng recipes using natural dyes [20]. Scoured fibres were mordanted by simmering for 1 h in a water ath with alum (10% in weight of the fibre weight, pH ∼ 3), let to cool own in the bath overnight, removed and rinsed. The fibres were ried and stored in the dark and at room temperature. For dyeing ith weld, the dyestuff (50% in weight of the fibre weight) was immered in water for 45 min The solution was cooled and strained. etted mordanted fibres were dyed by simmering for 30 min in he dye bath (pH ∼ 8) then cooled down overnight, removed from he bath and rinsed in warm water, dried and stored in the dark t room temperature. The fibres took a bright yellow colour. For yeing with madder, the dyestuff was first washed in boiled water or 3 min and the solution discarded. Dyestuff (50% in weight of the bre weight to use) and wetted fibres were added to the same bath pH ∼ 8) and heated up to about 40–50 ◦C for 1 h, then cooled down vernight, removed from the bath and rinsed in warm water, dried nd stored in the dark at room temperature. Mordanted fibres took n a red colour while non-mordanted wool was dyed a dark orange one. .2. UV ageing Snippets of wool were spread across a glass plate 22 cm × 19 cm) as a thin open web and irradiated in a Luzchem V chamber (LZC4-14, Luzchem, Ontario, Canada), using UVB Luzchem LZC-UVB) narrow bandwidth lamps–spectral distri- ution 281–315 nm, dose 35.33 Wm−2. Samples were turned egularly to allow even irradiation as much as possible and sam- ling was done at 0, 24, and 48 h, with additional samples taken t 3, 6, 12 and 18 h for the undyed samples. After irradiation the amples were stored in the dark at room temperature prior toN2-N14 N3-Q10-Q12-N14 further handling. Note for reference that 3 h of accelerated ageing utilised in this study is approximately equivalent to 1.5 years actual ageing in sunshine based on average UV exposure levels in the United Kingdom. 3.3. Experimental burial Between 1998 and 2006, samples of fulled twill (vadmel) fab- rics (undyed or dyed with madder and weld) were buried in bogs at “Land of Legends Lejre” (Denmark) and Rørmyra, Sør-Trondelag County (Norway), while a second series of burial experiments was initiated in 2002 in the harbour sediment at Marstrand on the coast of West Sweden [4]. For the bog study, samples were placed together with excavated soil into perforated PVC plastic pipes (16 mm diameter) and the modules buried in hand-drilled boreholes 0.5 to 1 m deep. For the Marstrand study, samples were placed together with sediment in perforated trays and buried at a depth of 0.5 m in the harbour bottom. Control samples and sam- ples retrieved from burial were kept in darkness in climate-control stores at the NTNU Museum, Norwegian University of Science and Technology in Trondheim, Norway (Table S1). They were subse- quently kept in drawers, away from sunlight until analysis. 3.4. Medieval samples Nine samples (9–13th c.) are from the textiles finds from the excavations conducted at 16–22 Coppergate in York in 1979–1981 [21] and seven (10–11th c.) from the excavations conducted at 6–8 Pavement on the site of the Lloyds Bank in York in 1972–1973 [22]. One sample was obtained from a 13th c. site on Queen Street, Quay- side in Newcastle upon Tyne [23]. The final eight samples (13–16th c.) come from an archaeological high status farm site at Reykholt, Borgarfjörður in Iceland [24,25]. 3.5. Sample preparation for mass spectrometry Samples of up to 10 mg (10 mg for the UV aged samples, 2 to 10 mg for the experimentally buried samples [4] and between 1–3 mg for the archaeological samples [18]) were ground in liq- uid nitrogen unless too small to be handled. The samples were solubilised by overnight shaking in a solution of 8 M urea, 50 mM Tris and 50 mM TCEP at pH 8.3. An aliquot of the supernatant was alkylated with 150 mM IAA and vortexed for 4 h in the dark. This was followed by 24 h dialysis with 100 mM AB on 3500 MWCO Slide A Lyzer® Mini Dialysis units from Thermo Scientific (two changes). About 25 g of samples were digested with 0.5 g of trypsin, overnight at 37 ◦C. All samples were then dried down and re-solubilised in 10 L of 0.1% TFA. A 1 L aliquot was used for MALDI-TOF-MS analysis and a diluted aliquot (1:40) used for nanoLC-ESI-MS/MS. 3.6. Peptide mass fingerprinting by MALDI-TOF-MSA matrix solution was prepared by diluting 0.1 mg of CHCA (-cyano-4-hydroxycinnamic acid) in 97/3 (acetone/0.1% TFA) and 1 L was applied onto an AnchorChipTM target (Bruker) and allowed to dry. A 1 L aliquot of analytical solution was applied and C. Solazzo et al. / Journal of Cultural Heritage 16 (2015) 896–903 899 0 5 10 15 20 25 30 0 3 6 12 18 24 48 0 24 48 0 24 48 0 24 48 Sc or e of o xi da tio n Time (hours) IU IM IAM IAW a) 0 5 10 15 20 25 30 0 48 0 48 0 48 0 48 Sc or e of o xi da tio n IU IM IAM IAW b) F by GP p r-dye t T t T s r t o d ( f ( 3 n s w p w C a 5 0 a t s t w a t i b m l b cig. 2. Oxidation scores in UV aged samples in (a) single LC–MS/MS run and (b) roteins KAPs. IU = undyed (white); IM = Madder-dyed (orange); IAM = alum/Madde hen removed after one minute and 1 L of washing buffer (0.1% FA) added. The residual droplet was removed and 1 L of recrys- allisation solution (0.1 mg of CHCA in 6/3/1 (ethanol/acetone/0.1% FA)) applied. The plate was loaded in an UltraflexTM III mass pectrometer (Bruker), and analyses were carried out in positive eflector mode using a Nd:YAG laser operating at 355 nm. Spec- ra were acquired using flexControl 3.0 (Bruker) on a mass range f 700–4000 Da with an accumulation of 500 shots on the stan- ards and 1000 shots on the samples. The calibration standard Bruker) was prepared according to the manufacturer’s instructions or instrument calibration and consisted of angiotensin I, ACTH clip 1–17), ACTH clip (18–39) and ACTH clip (7–38) peptides. .7. Protein analysis by nanoLC-ESI-MS/MS Protein separation was carried out on an Ultimate nanoflow anoLC equipped with Famos autosampler and Switchos column witching module (LC-Packings, The Netherlands). A 10 L sample as loaded on a C18 precolumn (Varian Microsorb 300 m ID, 5 m articles, 300 A˚ pore size) at a flow rate of 8 L/min. The precolumn as then switched in line with the analytical column (Microsorb 18, 20 cm, 75 m ID, 5 m particles, 300 A˚ pore size), and eluted t a flow rate of 150 nL/min, with a gradient from 2% to 55% B in 0 min. Solvent A was HPLC-grade H2O (Fisher Scientific, USA) with .2% formic acid, solvent B was LC–MS grade ACN with 0.2% formic cid. Using a stainless steel nanospray needle (Proxeon, Denmark), he column outlet was directly connected to a Q-STAR Pulsar i mass pectrometer (Applied Biosystems, USA) which was programmed o acquire MS/MS traces of 1+, 2+, 3+, 4+ and 5+ peptides. MS data as acquired from m/z 350–1200 and MS/MS from 40–1600 m/z ccumulating three cycles over 1.3 s duration each. One run was conducted for all UV-irradiated samples from 0 o 48 h. In addition, the control samples (0 h) and the 48 h UV- rradiated samples were analysed by Gas Phase Fractionation (GPF); y defining smaller mass-to-charge ranges, this method allows ultiple analyses to be performed and increases the number of ower abundance peptides identified. Four runs were performed etween m/z 400–550, 550–680, 680–785 and 785–1200, then ombined into a single file.F or Gas Phase Fractionation. Bottom = keratins. Top (in grey) = keratin-associated d (red); IAW = alum/weld-dyed (yellow). 3.8. Bioinformatic analysis Mascot Daemon (Matrix Science, UK) was used to extract peak lists from the LC–MS/MS data files. The peak lists from all m/z segments of each sample were concatenated and imported in Protein-Scape v2.1 (Bruker Daltonics). Subsequently, Mascot was used to search for matches with known Ovis aries sequences, using an in-house database compiled and curated by AgResearch, NZ. Parameters were set as followed: no enzyme, peptide mass toler- ance (MS) of 150 ppm, fragment mass tolerance (MS/MS) of 0.4 Da, carboxymethylation of cysteine as a fixed modification and acetyl (N-term), carbamyl (N-term), deamidated (NQ), Gln- > pyro-Glu (N- term Q), methyl (DE) and oxidation (M) as variable modifications. For oxidative modifications, single and double oxidation of H, W, F and Y was allowed, as well as kynurenine and hydroxykynurenine, quinone and hydroxyquinone. The peptide score cut-off was set at 30. 4. Results and discussion 4.1. Oxidation and deamidation in UV-irradiated samples After a short exposure to UV irradiation, undyed wool became yellow. This happened as quickly as 3 h after exposure. For dyed wool, discoloration was slower and irradiation was conducted for up to 48 h. The scores calculated for each sample are shown in Fig. 2 for the keratins and the keratin-associated proteins KAPs (in grey). Peptides observed with oxidative modifications are given Table S2. The oxidation scores, for both types of analysis conducted (single LC–MS/MS in Fig. 2a and Gas Phase Fractionation GPF in Fig. 2b) showed that the total oxidation score before irradiation was higher in the dyed samples than the undyed wool. Background oxidation has been observed in wool samples before irradiation [16] but com- parison between undyed and dyed samples indicate that the dyeing and mordanting treatments induced an increase in damage prior to accelerated UV ageing. This is consistent with the higher level in 4-hydroxybenzoic, a product of the oxidation of amino acids with aromatic side-chains, observed in aged alum-mordanted wool compared to unmordanted wool [26]. Evaluation of 900 C. Solazzo et al. / Journal of Cultural Heritage 16 (2015) 896–903 Table 2 Possible forms of oxidation taken by peptide DVEEWYIR, and presence in the UV-irradiated samples. Peptide Mr Modifications Name IU IM IAM IAW R.DVEEWYIR.Q 1108.52 W Tryptophan 0-3-6-12-18-24-48 0-24-48 0-24-48 0-48 R.DVEEWYIR.Q 1124.51 W + O Hydroxytryptophan 3-6-12-18 R.DVEEWYIR.Q 1140.51 W + 2O Dihydroxytryptophan 3-6-12-18-24-48 48 0-24-48 24-48 R.DVEEWYIR.Q 1112.51 W + O–C Kynurenine 12-24-48 24-48 24-48 24-48 on in h f w ( t F QFig. 3. Peptide DVEEWYIR with kynurenine modificati ydrothermal modification in wool has further revealed the ormation of oxidative products of tyrosine and tryptophan, which ere associated with the production of reactive oxygen species ROS) [27]. The higher levels of oxidation in the dyed samples prior o irradiation can then be attributed to the in-solution heating 50 60 70 80 90 100 0 12 24 36 48 % G ln -A sn Time in hours IU IM IAM IAW a) m/z 1487 50 60 70 80 90 100 0 12 24 36 48 % G ln -A sn Time in hours IU IM IAM IAW c) m/z 162 5 ig. 4. Percentage Gln–Asn values over time for UV-irradiated wool (on a 0–100% sca NQEYQVLLDVR, m/z 1487.74; (b) QNQEYQVLLDVR, m/z 1504.77; (c) LNVEVDAAPTVDLNthe undyed sample after 48 h of UV irradiation (IU48). of wool during the dyeing treatment rather than to the dyes or mordants themselves. Following irradiation, the analyses showed the highest increase in the oxidation score for the undyed sample (up to five times the initial score after 48 h). The increase was less significant for the dyed samples, consistent with lower oxidation 50 60 70 80 90 100 0 12 24 36 48 % G ln -A sn Time in hours IU IM IAM IAW b) m/z 150 4 50 60 70 80 90 100 0 12 24 36 48 % G ln -A sn Time in hours IU IM IAM IAW d) m/z 183 5 le, with 100% representing no deamidation and 0% complete deamidation); (a) R, m/z 1625.84; (d) TVNALEVELQAQHNLR, m/z 1834.97. C. Solazzo et al. / Journal of Cultural Heritage 16 (2015) 896–903 901 Fig. 5. Peptide m/z 1834.97 obtained by peptide mass fingerprinting and processed using mMass [28] (files for calculation of deamidation were processed using a specially- designed computed algorithm [19]) in undyed wool at (a) 0 h and (b) 48 h, in alum-mordanted madder-dyed wool at (c) 0 h and (d) 48 h, and in alum-mordanted weld-dyed wool at (e) 0 h and (f) 48 h. Table 3 Oxidation scores in control (year 0) and experimentally buried samples (years 1, 2, 4, 7 or 8). U = Undyed; M = Madder-dyed; W = Weld-dyed. L = Lejre (Denmark); R = Rørmyra (Norway); Mu = Marstrand (Sweden) uncovered (sewn into an open mesh nylon envelope); Mc = Marstrand (Sweden) covered (sewn into an open mesh nylon envelope and enclosed in an additional envelope constructed of a non-woven geotextile fabric) [4]. Samples U M W LU LM LW RU RM RW Year 0 0 0 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 1 2 4 8 Score 1 0 1 0 1 0 0 0 0 1 0 2 1 0 0 0 0 3 3 2 0 0 0 1 3 1 1 Samples U M W MuU MuM MuW McU McM McW Year 0 0 0 1 2 3 7 1 2 3 7 1 2 3 7 1 2 3 8 1 2 7 8 1 2 3 7 Score 1 0 1 0 0 1 1 0 0 0 0 0 2 0 0 2 0 0 0 0 0 0 0 2 1 2 0 902 C. Solazzo et al. / Journal of Cultural Heritage 16 (2015) 896–903 Table 4 Oxidation scores summarised for the archaeological samples. Coppergate (York) Pavement (York) 3959 4060b 4066 4067 4070 4073 4076 4077 4078 4082 4085 4087a 4089 4091 4093 4094 Score W 0 0 4 0 6 7 0 0 0 7 9 5 2 3 9 7 Score Y 4 9 15 5 9 7 9 10 6 4 8 9 6 7 14 0 Score F 1 3 1 2 1 0 0 2 1 2 3 1 1 2 1 1 Score H 0 0 2 0 0 0 1 0 0 0 0 0 0 0 0 0 Total score WYFH 5 12 22 7 16 14 10 12 7 13 20 15 9 12 24 8 New Reykholt (Iceland) 4544 2896 2897 2901 2902 2906 3962 3966 4120 Score W 9 0 0 4 8 4 2 2 2 Score Y 13 0 4 0 0 4 3 0 0 Score F 3 1 1 1 1 3 1 0 1 Score H 0 0 0 0 0 1 0 0 0 N l s F s d o t i e t a I a a u f t p t r t h t a o m t ( 0 s s i u d w d 1 d c 4 a pTotal score WYFH 25 1 5 5 ew: Newcastle. evels observed in other studies by amino acid analysis [11] and uggesting a level of protection to photo-oxidation in dyed wool. urthermore, oxidation in both mordanted weld and madder-dyed amples was primarily observed in the keratins while the unmor- anted samples had a large contribution to oxidation from the KAPs. This scoring approach represents a relative evaluation of levels f oxidative damage rather than a quantitative measure of oxida- ion. However, specific markers of oxidation have previously been dentified [16]. Peptide DVEEWYIR in particular is oxidised at sev- ral levels: single, double and kynurenine-type oxidation on the ryptophan [27,28]. Table 2 shows the products of oxidation cre- ted after UV irradiation in modern wool for peptide DVEEWYIR. n the undyed sample, single and double oxidations are detected fter only 3 h. Single oxidation is found in the undyed sample only nd up to 18 h, while the double oxidation is found up to 48 h in the ndyed and dyed samples. The kynurenine modification (Fig. 3) is ound after 12 h of UV exposure in the undyed samples and 24 h in he dyed samples. As would be expected, the cascade of oxidation is rogressively observed in the undyed samples: the gradual evolu- ion from single oxidation to kynurenine in undyed wool therefore epresents a good basis to assess UV photodegradation in historical extiles exposed in the past or modern times to light. In addition to oxidation, the mordanting and dyeing treatments ave been shown to induce deamidation at specific residues (a post- ranslational modification that converts glutamine into glutamic cid, and asparagine into aspartic acid), adding a +1 Da to the mass f the peptide. Of the four selected peptides in Table 1, peptides at /z 1834.97 and m/z 1625.84 were shown to have the fastest and he slowest rates of deamidation [18]. Fig. 4 shows the %Gln–Asn on a 0–100% scale, with 100% representing no deamidation and % complete deamidation) for all four peptides and all irradiated amples. In contrast with the photo-oxidation scores, the undyed ample remained undeamidated for all peptides and up to 48 h of rradiation. The unmordanted madder-dyed sample also remained ndeamidated for up to 24 h, but showed a large increase in deami- ation for all peptides by the 48-h point. Alum-mordanted samples ere deamidated as early as 24 h, and for the weld sample, deami- ation was observed in the control sample for all but peptides m/z 487.74. Fig. 5 shows the changes in the isotopic envelope due to eamidation for peptide m/z 1834.97 in madder and weld samples ompared to the unchanged profile in the undyed sample. .2. Oxidation and deamidation in experimentally buried and rchaeological samples Since oxidation of the aromatic residues is likely primarily a re-burial event triggered by the production of reactive oxidative9 12 6 2 3 species under UV irradiation, the amount of photo-oxidation calculated from the photoproducts detected, although not quan- titative, can provide significant information on the use of a textile during its lifespan. To verify that the burial environment does not significantly influence oxidation, modern buried samples were analysed and the results given in Table 3 show that samples buried for up to 8 years have few or no oxidized peptides. The experimen- tal burial control samples were kept in the dark in a controlled environment for 13 years before analysis and have oxidation scores of 0 for madder and 1 for undyed and weld. The buried samples gen- erally display no observable oxidation either, with the maximum score of 3 found in some of the Rørmyra samples (undyed and weld). Table 4 summarises the scores reported for each relevant amino acid (W, Y, F and H) and the total score in the archaeological sam- ples. The observed peptides with oxidative modifications are given in Table S3. The scores are variable, with scores between 1 to 12 for the Reykholt samples, 5 to 22 for Coppergate, 8 to 24 for Pave- ment and 25 for the Newcastle sample. These scores mainly reflect the oxidation on the keratins rather than the KAPs, as KAPs are less-resistant amorphous proteins that are usually degraded first in buried samples. Less oxidation was generally observed in the Reykholt samples, potentially indicating that these textiles suffered less damage from photo-oxidation. In contrast, the highest levels of deamidation were reported for the Reykholt samples [18] (in Supplementary information Fig. S1), indicating that, in archaeolog- ically buried samples, deamidation would be influenced more by the burial environment than pre-burial conditions. In the experi- mentally buried samples, the total time of burial (up to seven or eight years) was too short for any significant deamidation to be observed, although greater deamidation due to the dyeing treat- ment was evident in comparison to undyed wool. 5. Conclusions Traditionally, the chemical and biodegradation of textiles has been investigated with techniques that focus on the overall changes of the material. For example, with FT–IR, changes in the bands’ shape and intensity of functional groups either indicate damage or the creation of products of oxidation (for example, cysteic acid from cysteine). More specifically amino acids analysis can reveal patterns of oxidative degradation through the loss of certain amino acids, such as tyrosine and the formation of products, such as cysteic acid. But these methods are not precise enough to give details of the ori- gin or location of the oxidative reactions. Methodologies based on mass spectrometric analysis of ancient proteins are complementary and have the potential to determine particular types of degradation pathways and the conditions that might be required to generate ltural t t a i t t t o O o d t o a b h T u a a m o [ m d A g b g h f t A A A f 2 R [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ gen and hydroxyl radical-mediated oxidative pathways, Photochem. Photobiol.C. Solazzo et al. / Journal of Cu hem. We have demonstrated here how oxidative modifications at he amino acid residue level can be assessed qualitatively, offering new analytical approach that should be considered when study- ng questions of fibre treatment and degradation in archaeological extiles. However using the scoring system alone, it is only possible o compare samples from similar contexts. Ancient samples, in par- icular from the ground, suffer protein damage, and possibly loss f oxidised peptides. KAPs are for instance easily lost in old wool. ther factors that could influence background oxidation is the type f wool, processing of the fibres from scouring to dyeing, type of yes, or exposure to light before the product is finished, some of hem difficult to assess in archaeological samples. Therefore, the xidation scores could not be directly compared between modern nd ancient samples. Specific markers of degradation have however een identified. As with deamidation [18], peptides are likely to ave different rates of oxidation, which will have to be determined. his redox proteomics approach should be further developed by sing quantitative methods (for example iTRAQ labelling, which llows peptides from different samples to be compared in the same nalysis by binding them with specific isobaric mass tags). Such ethods have been used to compare photo-oxidation in samples riginating from similar contexts and comparable conditions [17] 29]. This approach would be particularly useful to the study of useum textiles, for which exposure to light and history are well ocumented. cknowledgments This research was supported by a Marie Curie International Out- oing Fellowship (FP7-PEOPLE-IOF-GA-2009-236425, THREADs) etween the University of York, UK and AgResearch, NZ. We are rateful to Elizabeth E. Peacock (NTNU University Museum, Trond- eim, Norway) for offering the experimentally buried samples or analysis, and Penelope Walton Rogers (Anglo-Saxon Labora- ory, York, UK) for making available the archaeological samples. ll experiments were conducted in the proteomics facilities at the gResearch Lincoln Research Centre. ppendix A. Supplementary data Supplementary data associated with this article can be ound, in the online version, at http://dx.doi.org/10.1016/j.culher. 015.02.006. eferences [1] J.M. Dyer, Protein photo-oxidative damage-consequences, characterization and control, in: L.N. Collignon, C.B. Normand (Eds.), Photobiology: principles, appli- cations and effects, Nova Science Publishers, Inc, Hauppauge, New York, USA, 2010, pp. 91–113. [2] J.M. Dyer, C. Solazzo, Evaluating the degradation of silk fabrics and silk-based biomaterials, in: P. Aramwit (Ed.), Silk: Properties, Production and Uses, Nova Science publishers, Bangkok, Thailand, 2012, pp. 117–134. [3] J.M. Dyer, S.D. Bringans, W.G. Bryson, Characterisation of photo-oxidation prod- ucts within photoyellowed wool proteins: tryptophan and tyrosine derived chromophores, Photochem. Photobiol. Sci. 5 (2006) 698–706. [4] C. Solazzo, J.M. Dyer, S. Clerens, J.E. Plowman, E.E. Peacock, M.J. Collins, Pro- teomic evaluation of the biodegradation of wool fabrics in experimental burials, Int. Biodeter. Biodegr. 80 (2013) 48–59. [ Heritage 16 (2015) 896–903 903 [5] K.R. Millington, Photoyellowing of wool. Part 1: factors affecting photoyellow- ing and experimental techniques, Color Technol. 122 (2006) 169–186. [6] K.R. Millington, Photoyellowing of wool. Part 2: photoyellowing mechanisms and methods of prevention, Color Technol. 122 (2006) 301–316. [7] E.S.B. Ferreira, A.N. Hulme, H. McNab, A. Quye, The natural constituents of historical textile dyes, Chem. Soc. Rev. 33 (2004) 329–336. [8] M.P. Colombini, A. Andreotti, C. Baraldi, I. Degano, J.J. Łucejko, Colour fading in textiles: a model study on the decomposition of natural dyes, Microchem. J. 85 (2007) 174–182. [9] K. Yoshizumi, P.C. Crews, Characteristics of fading of wool cloth dyed with selected natural dyestuffs on the basis of solar radiant energy, Dyes Pigments 58 (2003) 197–204. 10] G.J. Smith, New trends in photobiology (invited review) photodegradation of keratin and other structural proteins, J. Photochem. Photobiol. B 27 (1995) 187–198. 11] I. Vanden Berghe, Towards an early warning system for oxidative degradation of protein fibres in historical tapestries by means of calibrated amino acid analysis, J. Archaeol. Sci. 39 (2012) 1349–1359. 12] I.J. Miller, G.J. Smith, Protection against phototendering of wool by metal salts and mordanted dyes, J. Soc. Dyers Colour. 111 (1995) 103–106. 13] A.L. King, K.R. Millington, Trace metals can affect hydroxyl radical pro- duction and yellowing of photo-irradiated wool, J. Text. I. 104 (2013) 648–654. 14] K.R. Millington, L.J. Kirschenbaum, Detection of hydroxyl radicals in photoir- radiated wool, cotton, nylon and polyester fabrics using a fluorescent probe, Color Technol. 118 (2002) 6–14. 15] J.M. Dyer, S.D. Bringans, W.G. Bryson, Determination of photo-oxidation prod- ucts within photoyellowed bleached wool proteins, Photochem. Photobiol. 82 (2006) 551–557. 16] J.M. Dyer, J.E. Plowman, G.L. Krsinic, S. Deb-Choudhury, H. Koehn, K.R. Millington, S. Clerens, Proteomic evaluation and location of UVB- induced photo-oxidation in wool, J. Photochem. Photobiol. B 98 (2010) 118–127. 17] A. Grosvenor, J. Morton, J. Dyer, Profiling of residue-level photo-oxidative dam- age in peptides, Amino Acids 39 (2010) 285–296. 18] C. Solazzo, J. Wilson, J.M. Dyer, S. Clerens, J.E. Plowman, I. von Holstein, P. Wal- ton Rogers, E.E. Peacock, M. Collins, Modeling deamidation in sheep -keratin peptides and application to archaeological wool textiles, Anal. Chem. 86 (2014) 567–575. 19] J. Wilson, N.L. van Doorn, M.J. Collins, Assessing the extent of bone degradation using glutamine deamidation in collagen, Anal. Chem. 84 (2012) 9041–9048. 20] J. Dean, Colours from Nature. A Dyer’s Handbook, Search Press, Tunbridge Wells, 2009. 21] P. Walton, Textiles, Cordage and Raw Fibre from 16-22 Coppergate. The Archae- ology of York, London, Council for British Archaeology and York Archaeological Trust, London, 1989. 22] J. Hedges, M.L. Ryder, P. Walton Rogers, A. Muthesius, in: A. MacGregor (Ed.), Textiles, York Archaeological Trust by the Council for British Archaeology, Lon- don, 1982, pp. 102–136. 23] P. Walton Rogers, Caulking, cordage and textiles, in: C. O’Brien, L. Bown, S. Dixon, R. Nicholson (Eds.), The Origins of the Newcastle Quayside: Excavations at Queen Street and Dog Bank, Society of Antiquaries of Newcastle upon Tyne, Newcastle upon Tyne, 1988, pp. 78–92. 24] G. Sveinbjarnardóttir, E. Erlendsson, K. Vickers, T. McGovern, K. Milek, K. Edwards, I. Simpson, G. Cook, The palaeoecology of a high status Icelandic farm, Environ. Archaeol. 12 (2007) 187–206. 25] P. Walton Rogers, Textiles, pelts and related organic finds, in: G. Sveinb- jarnardóttir (Ed.), Archaeological investigations at a high status farm in western Iceland, IS: Þjóðminjasafn Íslands/Snorrastofa, Reykjavík, 2012, pp. 197–217. 26] M. Dilillo, A. Restivo, I. Degano, E. Ribechini, M.P. Colombini, GC/MS investi- gations of the total lipid fraction of wool: a new approach for modelling the ageing processes induced by iron-gallic dyestuffs on historical and archaeolog- ical textiles, Microchem. J. 118 (2015) 131–140. 27] A.J. Grosvenor, J.D. Morton, J.M. Dyer, Proteomic characterisation of hydrother- mal redox damage, J. Agric. Food Chem. 91 (2011) 2806–2813. 28] J.E. Plowman, S. Deb-Choudhury, A.J. Grosvenor, J.M. Dyer, Protein oxidation: identification and utilisation of molecular markers to differentiate singlet oxy-Sci. 12 (2013) 1960–1967. 29] A.J. Grosvenor, J.D. Morton, J.M. Dyer, Isobaric labelling approach to the tracking and relative quantitation of peptide damage at the primary structural level, J. Agric. Food Chem. 58 (2010) 12672–12677.