- Email: [email protected]
Molecular evidence of use of hide glue in 4th millennium BC Europe
Journal of Archaeological Science 63 (2015) 65e71
Contents lists available at ScienceDirect
Journal of Archaeological Science journal homepage: http://www.elsevier.com/locate/jas
Molecular evidence of use of hide glue in 4th millennium BC Europe Niels Bleicher a, Christian Kelstrup b, Jesper V. Olsen b, Enrico Cappellini c, * a
City of Zürich, Office for Urbanism, Competence Centre for Underwater Archaeology and Dendrochronology, Switzerland Novo Nordisk Foundation Center for Protein Research, Faculty of Health Sciences, University of Copenhagen, Blegdamsvej 3B, 2200 Copenhagen, Denmark c Centre for GeoGenetics, Natural History Museum of Denmark, University of Copenhagen, Øster Voldgade 5-7, 1350 Copenhagen, Denmark b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 February 2015 Received in revised form 6 August 2015 Accepted 16 August 2015 Available online 18 August 2015
A well-preserved bow, dated by dendrochronology to 3176e3153 BC, was found at the waterlogged ra” in Zurich (Switzerland). The surface of the bow, made of yew (Taxus Neolithic site “Parkhaus Ope baccata), was decorated with bark strips from a different, broad-leaved, tree species. In order to investigate whether the bark decoration was fixed to the bow with hide or fish glue, mass spectrometry (MS)based ancient protein sequencing was attempted to detect possible traces of collagen residues. The sequences retrieved, in particular collagen type 3 (COL3A1), indicate that most probably skin, and possibly other slaughtering by-products, were used as the initial materials to produce hide glue. Amongst the candidate animal species that the glue could have originated from, cattle and domestic ovicaprids were confidently identified. This is, to the best of our knowledge, the oldest evidence of the use of animalbased glue in Europe. It demonstrates that in the late 4th millennium BC human communities, aside from benefitting from more commonplace primary and secondary products, also exploited domestic animals to extract a high value-added biochemical. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Ancient proteins Collagen Animal glue Domestic animals Neolithic Europe
1. Introduction Protein residues are probably the most abundant class of ancient biomolecules preserved in bones and soft tissues from archaeological and palaeontological contexts (Di Lullo et al., 2002; Lynnerup, 2007; Prockop and Kivirikko, 1995). Moreover, due to their chemical and mechanical properties, proteinaceous materials have also been used to satisfy primary human needs, including sheltering (Crone, 2007), transportation, clothing and food production (Hollemeyer et al., 2008; Kirby et al., 2013; Pinhasi et al., 2010). Production of adhesives, although less intuitive and more technologically refined, represented another way to use abundant animal proteins. In particular, hide glue, can be relatively easily extracted from animal skin, bone and other slaughtering by-products immersed in hot water. During this process, known as gelatinization, high temperature hydrolyses, denatures and solubilises collagen (Kuckova et al., 2009, 167). Collagen is a heterotrimer composed of three amino acid chains in a triple-helix structure. The transformation of collagen into gelatin is interpreted as the disintegration of the helical structures into
* Corresponding author. E-mail address: [email protected] (E. Cappellini). http://dx.doi.org/10.1016/j.jas.2015.08.012 0305-4403/© 2015 Elsevier Ltd. All rights reserved.
random coils (Chen et al., 2014). The use of adhesives is one of the most ancient human technological solutions for producing elaborate tools with superior mechanical properties, overcoming the technical limitations associated with products derived from a single component (Koller et al., 2001). Therefore, by allowing the production of objects made out of different parts and materials, the introduction and use of natural adhesives represents one of the most important breakthroughs in object manufacturing. Although easy to produce, even in rudimentary conditions, hide glue presents remarkably interesting properties, thus making it the adhesive of choice for many applications, until the relatively recent introduction of glues produced by organic synthesis (e.g. Brauns, 1858). It is well documented that birch bark tar was used as early as the Middle Palaeolithic (Koller et al., 2001; Grünberg, 2002; Mazza et al., 2006; Wadley, 2010). Other adhesives, such as conifer resins and their heated derivatives, have been identified as adhesives in later periods (Regert, 2004). Conifer resin has been documented on a limited number of archaeological finds from the 7th Millennium BC in Canada (Helwig et al., 2008), and on a 5000 yearold copper age arrow point (Strnad, 1990). Together with beeswax, other substances, such as pistacia resins and plant oils, have also been found on archaeological objects from Bronze and Iron Age in Europe (Regert and Rolando, 2002). Similarly, the most ancient
66
N. Bleicher et al. / Journal of Archaeological Science 63 (2015) 65e71
archaeological traces of a more elusive type of glue, based on collagen, originate from a site in Israel, radiocarbon-dated between 8310 and 8110 years before present (Nissenbaum, 1997). In Egypt, where glue was identified as an ingredient of plaster from the third millennium BC, an ingot of pure hide glue was also found at Deir-elBahari dating to the late second millennium BC (Lucas, 1934). To the best of our knowledge, however, no occurrence of animal glue has been documented in Neolithic Europe. Analytical methods for detection of proteinaceous residues in archaeological and artistic material initially relied on gas or liquid chromatographic methods (GC or HPLC) coupled to mass spectrometry (MS) (Schneider and Kenndler, 2001; Chiavari et al., 1998). These approaches only allow amino acid identification and tentative reconstruction of the proteins they possibly originated from. Immunological methods, namely Enzyme-Linked Immuno Sorbent Assay (ELISA), and ancient DNA sequencing were also occasionally reported for the identification of collagen and hide glue in artworks (Cartechini et al., 2010; Albertini et al., 2011). However, none of these approaches allow for protein sequencing. This limit was overcome with the recent introduction of mass spectrometrybased methods for ancient protein identification (Ostrom et al., 2000) and sequencing (Nielsen-Marsh et al., 2002). The use of liquid chromatography, coupled to tandem mass spectrometry (LCMS/MS), allowed species identification of biological samples, such as bone and skin, from archaeological contexts (Buckley et al., 2009; Brandt et al., 2014). The same approach was also applied to successfully identify protein binders in historical paintings (Tokarski et al., 2006), as well as food residues in pottery (Solazzo et al., 2008). Finally, a proteomics strategy based on LC-MS/MS was also used to identify the origin species of glues used in gilt from the 18th century (Dallongeville et al., 2011). In 2009, during the archaeological excavation of the Neolithic ra” in Zurich (Switzerland), settlement at the site “Parkhaus-Ope Fig. 1a and b, the discovery of an object made of yew (Taxus baccata), suggested the possible use of hide glue, Fig. 1c. The finding, dated by dendrochronology to 3176e3153 BC (Bleicher and Burger, in press.), was identified as the symbolic representation of a hunting bow. The preservation of wood was excellent due to the anaerobic conditions in the waterlogged deposit encasing the bow (Madigan et al., 1997; Bleicher and Schubert, 2015). The bark on its surface was shaped in silvery-reddish strips showing large lenticels, a feature absent in yew. Although a more detailed anatomical characterisation was not possible, these traits were sufficient to confidently assign the bark to a broadleaved species and tentatively identify the origin of the bark to a young cherry tree. The straight edges and regular distribution of the strips along all the length of the bow, Fig. 1c-inset, strongly indicated that they were intentionally glued as a decoration on the surface of the bow. In the Swiss Neolithic site of Cham Eslen, birch tar was utilised to glue birch bark to an axe handle (Gross-Klee and Hochuli, 2002). For the new find from Zurich, though, the use of birch tar can be ruled out since no dark residue was detected under the bark. Although in the Near East the use of collagen-based adhesives is recorded at much earlier times (Nissenbaum, 1997), so far no Neolithic European evidence of animal-based glue has been found. The earliest trace of adoption of hide glue in Europe is possibly a mention by Aristoteles (Anheuser, 2001). If confirmed by molecular evidence, the case we describe would represent the most ancient occurrence of hide glue in Europe, and provide the first positive evidence for its use in this area predating the written sources by several millennia. We set out to test whether hide glue or, taking into account that the prehistoric settlement was situated on the lakeshore, fish glue was used to affix the bark to the body of the bow. To achieve this,
we attempted LC-MS/MS proteomics analysis of organic residues extracted from wood flakes collected under the bark. 2. Materials and methods Immediately after unearthing the object, small wood flakes from two areas underneath two different bark strips, lifted from the body of the bow, were carefully removed using a razor blade. Since the sampled area had been covered by the bark for over 5000 years, chances of contamination before or during sample collection can be considered extremely limited. A negative control prepared and analysed following exactly the same procedure adopted for the archaeological samples, except for the initial addition of ancient material, was processed together with the samples. Each of the two samples, approximately 15 mg, was then transferred in a protein LoBind 1.5 mL tube (Eppendorf, Germany). Each sample was processed independently as follows. Each sample was re-suspended in 500 mL of 100 mM ammonium bicarbonate solution at pH 8.00 and was briefly washed by mechanical shaking at room temperature to remove soil residues and other debris, prior to pelleting by centrifugation at 14,000 g for 5 min. The supernatant was discarded and this step was repeated two times in total. The pellet was resuspended in 1 mL 1.2 M HCl and incubated at 4 C for 24 h to gently hydrolyse collagen. This approach was chosen with the aim of maximising recovery of collagen altered by prolonged exposure to protein crosslinkers present in soil, such as clay, humic and fulvic acids, as well as wood lignin. After centrifugation at 20,000 g for 20 min, the supernatant was removed and stored, while the pellet was re-suspended in 200 mL 50 mM ammonium bicarbonate pH 8.00 and incubated at 70 C for 24 h, to solubilise collagen. The pH was checked using pH indicator strips and adjusted to 8.00 with diluted ammonium hydroxide. The sample was then pelleted by centrifugation at 14,000 g for 10 min and the supernatant was collected and reduced by incubation for 1 h at 60 C with 5 mM (final concentration) of dithiothreitol (VWR-BDH, England), dissolved immediately before use. The reduced cysteines were then alkylated by incubating in the dark, at room temperature, for 45 min with 15 mM (final concentration) chloroacetamide (SigmaeAldrich, Denmark), dissolved immediately before use. The pH was then checked using pH-indicator strips, adjusted to 8.00 with diluted ammonium hydroxide and digestion was started by adding 4 mL of 0.5 mg/mL sequencing grade trypsin solution (Promega, Nacka Sweden) and incubating at 37 C overnight. The following morning, 2 mL of fresh trypsin were added and digestion was extended for 6 additional hours. Digestion was quenched with 10% trifluoroacetic acid to a final concentration of 0.2e0.8%, as necessary to reach pH < 2.00. After centrifugation at 14,000 g for 10 min, to precipitate any eventual insoluble residue, the tryptic peptides in the supernatant were immobilised on C18 stage tips as previously described (Cappellini et al., 2012). Peptide mixtures were analysed by online nanoflow reversedphase C18 liquid chromatography tandem mass spectrometry (LC-MS/MS), as described previously (Cappellini et al., 2012). Briefly, the LC-MS system consisted of an EASY-nLCTM system (Proxeon Biosystems, Odense, Denmark) connected to the LTQOrbitrap Velos (Thermo Electron, Bremen, Germany) through a nano-electrospray ion source. An aliquot of 5 uL of each peptide sample was auto-sampled onto and directly separated in a 15 cm analytical column (75 mm inner diameter) with a 90 min gradient from 5% to 30% acetonitrile in 0.5% acetic acid at a flow rate of 250 nL/min. The effluent from the HPLC was directly electrosprayed into the mass spectrometer by applying 2.2 kV through a platinumbased liquid-junction. The LTQ-Orbitrap Velos instrument was operated in datadependent mode to automatically switch between full scan MS
N. Bleicher et al. / Journal of Archaeological Science 63 (2015) 65e71
67
ra archaeological site (A and B) where the wood bow was found in a waterlogged deposit (C). Inset: detail of the bark surface. Fig. 1. Positioning of the Zürich-Parkhaus Ope
and MS/MS acquisition. Instrument control was through Tune 2.6.0 and Xcalibur 2.1 Survey, full scan MS spectra (from m/z 300e2000) were analysed in the orbitrap detector with resolution R ¼ 30,000 at m/z 400 (after accumulation to a ‘target value’ of 1e6 in the linear ion trap). The ten most intense peptide ions with charge states 2 were sequentially isolated to a target value of 5e4 and fragmented in octopole collision cell by Higher-energy Collisional Dissociation (HCD) with a normalized collision energy setting of 40%. The resulting fragments were detected in the Orbitrap system with resolution R ¼ 7500. The ion selection threshold was 5000 counts and the maximum allowed ion accumulation times were 500 ms for full scans and 250 ms for HCD. Standard mass spectrometric conditions for all experiments were: spray voltage, 2.2 kV; no sheath and auxiliary gas flow; heated capillary temperature, 275 C; predictive automatic gain control (pAGC) enabled, and an Slens RF level of 65%. For all full scan measurements with the Orbitrap detector a lock-mass ion from ambient air (m/z 445.120024) was used as an internal calibrant. Raw files generated during spectra acquisition were searched on a workstation using the MaxQuant (MQ) algorithm, v. 1.4.1.2 (Cox
and Mann, 2008), and the Andromeda peptide search engine (Cox et al., 2011) initially against a target/reverse custom-made list of all the mammalian Alpha-1 type I, Alpha-2 type I and Alpha-1 type III collagen sequences publicly available in UniProt and nrNCBI, as well as the complete list of collagen sequences available in NCBI RefSeq after restriction to the Actinopterygii taxon. Subsequently, the raw files were also searched against the target/reverse protein list in reference proteomes or extended proteins lists available in NCBI RefSeq for the following species: Homo sapiens (human, reference proteome -including isoforms-downloaded from UniProtKB on Dec. 31 2012), Bos taurus (cow, reference proteome Oct. 24 2012), Equus caballus (horse, reference proteome downloaded on April 23 2014), Ovis aries (sheep, 22,752 accessions, downloaded on Jan. 8, 2013), Capra hircus (goat, 31,461 accessions, downloaded on Dec. 2, 2013), Sus scrofa (pig, 33,148 accessions, downloaded on Jan. 26 2013), Oryctolagus cuniculus (rabbit, 21,150 accessions, downloaded July 22 2014) and Oncorhynchus mykiss (Rainbow trout, 3839 accessions, downloaded July 22 2014). In every search, spectra were also matched against the common contaminants, such as keratin and porcine trypsin sequences, downloaded from
68
N. Bleicher et al. / Journal of Archaeological Science 63 (2015) 65e71
Uniprot. Trypsin was selected as the proteolytic enzyme and two missed cleavages were allowed. Oxidation (M and P), deamidation (N and Q), acetylation (K), Q / pyro-Glu (N-term Q), and E / pyro-Glu (Nterm E) were selected as variable amino-acid modifications. Carbamidomethylation (C) was selected as fixed modification. Default values were used for precursor (6 ppm) and fragment (20 ppm) ions mass tolerance. False-discovery rate was set at 1% and minimum values for peptide-score and peptide sequence length were set at 70 and 7 respectively. The amount of random matches was evaluated performing MaxQuant search against reversed sequences. For the identification of species-diagnostic peptides, possible environmental contaminants, such as actin, tubulin, keratins and keratin-associated proteins, as well as proteins commonly used in mass spectrometry facilities as standards or calibrants and proteins highly conserved, such as histones, were excluded from further investigation. A minimum of two unique peptides was required for confidently identifying proteins. The Blastp algorithm (Camacho et al., 2009) was used to identify protein-unique peptides by aligning the peptides identified using MaxQuant against the entire nrNCBI protein database, as previously described (Brandt et al., 2014). 2.1. Data deposition note The mass spectrometry proteomics data were deposited in the ProteomeXchange Consortium (http://proteomecentral. proteomexchange.org) via the PRIDE partner repository (Vizcaino et al., 2014) with the data set identifier PXD00172700. 3. Results The analysis of the negative control sample, blank, did not identify any protein residue apart from the ordinary contaminants, such as keratins and porcine trypsin. Two spectral files were generated, one from each protein extract fraction. A total of 25,608 MS/MS spectra were acquired, of which 4134 (14.4%) could be identified. 101 of these provide confident identification of five proteins (relying on at least two unique sequence peptides), based on the assignment of 27 razor þ unique peptides, Tables 1 and 2. The term “razor” designates those peptide sequences matching with more than one protein group and assigned to the one with the largest number of total identified peptides. This approach follows the parsimony principle, to infer a list of the lowest possible number of proteins that can account for the observed peptides (Nesvizhskii and Aebersold, 2005). Four out of five sequences were identified as collagen proteins. Although most of the collagen peptides identified are extremely common among many animal species, some of them can be used as a marker to narrow down the animal species used as collagen source, Table 2 (Brandt et al., 2014). Based on the sequences
available in public protein databases at the time of data analysis, two of the peptides identified can be uniquely assigned to cattle (B. taurus) and one to domestic ovicaprids, i. e. sheep and goat (O. aries and C. hircus), Table 2 and Fig. 2a and b. Peptides were considered diagnostic when, after BLAST searching (Camacho et al., 2009) against the entire nrNCBI protein database, they were assigned to a single species, or to a limited number of species among which only one can be considered plausible. These considerations were based on the nature of the samples, such as geographic origin and distribution. For example, peptides equally present in cattle (B. taurus), water buffalo (Bubalus bubalis) and yak (Bos mutus) were considered diagnostic for cattle. Alternative animal sources for the peptides here assigned to domestic species cannot be completely excluded. This is due to the limited availability in public databases of protein sequences belonging to other non-domesticated ovicaprids, likely present in the same epoch and geographic area as the bow. Nevertheless, the co-occurrence of ovine and bovine bone remains within the archaeological context from which the bow was also retrieved from supports our species assignments. Identification of collagen alpha1(III), Fig. 2c, supports the hypothesis that the adhesive was produced from skin, as this molecule, together with collagen alpha-1 and -2(I), is particularly abundant in this organ. Although Collagen alpha-1(III) is also present in other organs, such as the lung and the vascular system, their use for glue preparation can be considered less probable arguably due to their primary use as a food source. No evidence of hare or fish collagen could be detected, providing no support for the use of glue originating from these animals. Finally two peptides confidently assigned to beta haemoglobin were also identified. Although their sequences, widely conserved among many species, do not provide any further information about the animal sources used to produce the hide glue, their identification suggests that traces of blood were present in the glue.
4. Discussion The results demonstrate how analytical approaches for protein sequencing have improved to the point of allowing unexpected application in waterlogged archaeological contexts, even from temperate environments and a wide range of epochs (Cappellini et al., 2010). Col1A1 and Col1A2 are ubiquitous in all the collagen-bearing tissues, while Coll3A1 occurs more abundantly in soft connective tissues and skin. Consequently, the collagen-based animal glue we detected was extracted, most probably, from ovine and cattle skin, as well as from other slaughtering byproducts. Identification of haemoglobin is compatible with the possibility that blood traces, originally associated with the materials used to produce hide glue, were accidentally co-extracted during the gelatinization process. Alternatively, deliberate inclusion of blood
Table 1 Identified proteins. N.
Protein name
All matching peptidesa
Razor þ unique peptides
Total seq. cov. [%]b
Unique þ razor sequence coverage [%]
Seq. length
Matched spectra
Species of origin
1 2 3 4 5
Collagen_alpha1(I) Collagen_alpha-2(I) Collagen_alpha-2(I) Collagen alpha-1(III) Haemoglobin sub. beta
10 8 9 4 2
10 2 9 4 2
11.1 8 9.3 3.9 15.6
11.1 2.1 9.3 3.9 15.6
1313 1364 1364 1467 147
42 22 26 8 3
mammals bovine sheep/goat ovine/bovine mammals
a
Including non-unique peptides. Based on all matching peptides. No peptide allowed discrimination between sheep and goat. Collagen_alpha1(I) and haemoglobin were identified only by peptides with very limited taxonomic variability among mammals. b
N. Bleicher et al. / Journal of Archaeological Science 63 (2015) 65e71
69
Table 2 Identified peptides. Protein N.a
Sequence
Length (aa)
Mass
Charge
MaxQuant score
Total matched spectra
1 1 1 1 1 1 1 1 1 1
GEPGPAGLPGPPGER GETGPAGPAGPIGPVGAR GETGPAGRPGEVGPPGPPGPAGEK GETGPSGPAGPTGAR GNSGEPGAPGSK GPSGPQGPSGPPGPK GSAGPPGATGFPGAAGR GSEGPQGVR GVQGPPGPAGPR GVVGLPGQR
15 18 24 15 12 15 17 9 12 9
1386.69 1559.81 2167.07 1310.62 1056.48 1315.65 1426.70 885.43 1088.57 881.51
2 2 3 2 2 2 2 2 2 2
88.42 135.72 108.57 144.80 93.35 187.16 113.48 111.17 132.25 119.29
4 3 1 5 1 7 2 1 11 7
2 2 2 2 2 2 2 2
AGVMGPAGSR GDGGPPGATGFPGAAGR GEAGPAGPAGPAGPR GSTGEIGPAGPPGPPGLR IGQPGAVGPAGIR PGPIGPAGAR SGETGASGPPGFVGEK VGAPGPAGAR
10 17 15 18 13 10 16 10
901.44 1440.67 1260.62 1615.83 1191.67 891.49 1475.69 851.46
2 2 2 2 2 2 2 2
93.33 114.63 135.65 118.21 81.92 84.08 72.58 117.81
1 6 4 4 4 1 1 1
3 3 3 3 3 3 3 3 3
AGVMGPAGSR GAPGAVGAPGPAGANGDR GDGGPPGATGFPGAAGR GEAGPAGPAGPAGPR GSTGEIGPAGPPGPPGLR PGPIGPAGAR TGEPGAAGPPGFVGEK TGQPGAVGPAGIR VGAPGPAGAR
10 18 17 15 18 10 16 13 10
901.44 1490.72 1440.67 1260.62 1615.83 891.49 1469.72 1179.64 851.46
2 2 2 2 2 2 2 2 2
93.33 105.90 114.63 135.65 118.21 84.08 82.35 112.84 117.81
1 1 6 4 4 1 2 6 1
4 4 4 4
GEVGPAGSPGSSGAPGQR GPAGANGLPGEK GPPGAGGPPGPR GPPGPPGTNGAPGQR
18 12 12 15
1566.74 1066.54 1015.52 1358.67
2 2 2 2
148.52 88.02 164.97 147.26
1 1 2 4
5 5
LLVVYPWTQR VNVDEVGGEALGR
10 13
1274 1314
2 2
104.22 99.94
1 2
Species indicator
References
bovine
Pep. A (Buckley '09)
bovine
Brandt et al., 2014
sheep/goat sheep/goat/pig
Brandt et al., 2014 Pep. A (Buckley '09)
ovine/bovine
References report previously identified peptides are cited. a Protein numbering according to Table 1.
in the glue formulation could represent an intriguing example of a highly refined technological understanding about the wood binding properties of different biological materials. Indeed, until just a few decades ago, blood-based glues had been extensively employed for industrial production of plywood (NPCS Board of Consultants and Engineers (2007)). Identification of collagen alpha-1(III), reveals that most probably skin, and possibly other slaughtering by-products, were used as source material for animal glue production. The case we present, as well as others available in literature (Kirby et al., 2013; Brandt et al., 2014), contribute to illustrate how ancient protein sequencing can provide conclusive and, unlike aDNA, tissue-specific ancient biomolecular evidence (Cappellini et al., 2014). Until recently, however, ancient protein sequencing has only been applied to archaeological animal macroscopic remains, such as bone and skin, either in their natural form (Buckley et al., 2009; Maixner et al., 2013), or processed to produce artefacts (Kirby et al., 2013; Brandt et al., 2014). Ancient protein residues interpreted as food remains were previously reported using immunological methods (Craig et al., 2000), and MS-based protein sequencing (Buckley et al., 2013; Solazzo et al., 2008). However this is the first time, to our knowledge, where proteomics methods can be used to confidently reconstruct manufacturing practices in prehistoric contexts. The application of this approach to characterise protein-based adhesives in archaeological contexts has been extremely limited. Only
Buckley et al. (2013) suggested that milk residues detected within archaeological vessels could represent traces of a sealant, and not necessarily traces of food. It can be anticipated that future research will make the recovery of proteinaceous glue traces from archaeological findings routine, thus allowing for a better reconstruction of the until recently invisible technological expertise and even ritual practices utilised by early ancient human communities. 5. Conclusions LC-MS/MS-based ancient protein sequencing demonstrates that, as early as the 4th millennium BC, farmers in the Zurich area performed rudimentary chemical extractions to produce the oldest so far known animal-based adhesive in Europe. Hide glue was most probably extracted from skins, as well as other connective tissues rich in collagen of cattle and domestic ovicaprids. Results show that the community adopting these procedures used livestock to extract a high value-added biochemical, besides generation of more commonplace primary (meat, hide, bone) and secondary products (milk, wool, traction) associated with farming. The possibility to relay on such an advanced, for the time, technology enabled the construction of elaborate objects with superior mechanical properties. This is the first time, as far as we know, where proteomics methods were used to confidently reconstruct manufacturing practices in prehistoric contexts. As very recently reported by Villa
70
N. Bleicher et al. / Journal of Archaeological Science 63 (2015) 65e71
Fig. 2. Examples of tandem MS/MS spectra supporting identification of species-diagnostic collagen peptides and collagen alpha-1(III). (A) MS/MS spectrum, confidently assigned to amino acid sequence SGETGASGPPGFVGEK, diagnostic for cattle; (B) MS/MS spectrum, confidently assigned to amino acid sequence TGEPGAAGPPGFVGEK diagnostic for domestic ovicaprids; (C) MS/MS spectrum, confidently assigned to amino acid sequence GPPGAGGPPGPR, diagnostic for ovine and bovine collagen alpha-1(III).
N. Bleicher et al. / Journal of Archaeological Science 63 (2015) 65e71
et al. (2015), systematic integration of GC-MS-based analysis of lipid/terpenoid fractions with proteomic studies will provide a much more comprehensive insight about ancient technological processes adopted by human and, possibly, their behaviours. Acknowledgements EC was supported by the Danish Council for Independent Research, grant n. 10-081390 (awarded to Prof. MTP Gilbert). The authors would like to thank Anna Katerina Fotakis for the technical help. References Albertini, E., Raggi, L., Vagnini, M., Sassolini, A., Achilli, A., Marconi, G., Cartechini, L., Veronesi, F., Falcinelli, M., Brunetti, B.G., Miliani, C., 2011. Tracing the biological origin of animal glues used in paintings through mitochondrial DNA analysis. Anal. Bioanal. Chem. 399 (9), 2987e2995. €tter Anheuser, K., 2001. Historische Klebstoffe und ihre Identifizierung. Arbeitsbla für Restaur. 19, 247e264. Bleicher, N., Burger, M., 2015. Pfahlfeldanalyse und Dendroarch€ aologie. In: ra. Eine neolithische Bleicher, N., Harb, Chr (Eds.), Zürich-Parkhaus Ope €ologie, Feuchtboden fundstelle. Band 1. Befunde, Schichten und Dendroarcha vol. 48. Monographien der Kantonsarch€ aologie Zürich, Zürich, Egg (in press). Bleicher, N., Schubert, C., 2015. Why are they still there? A model of accumulation and decay of organic prehistoric cultural deposits. J. Archaeol. Sci. 61, 277e286. Brandt, L.Ø., Schmidt, A.L., Mannering, U., Sarret, M., Kelstrup, C.D., Olsen, J.V., Cappellini, E., Hart, J.P., 2014. Species identification of archaeological skin objects from Danish Bogs: comparison between mass spectrometry-based peptide sequencing and microscopy-based methods. PLoS ONE 9 (9), e106875. Brauns, 1858. Über Kitt und Leim, nach Haslett und Hackley. Polytech. Cent. 9 (6), 89e91. Buckley, M., Collins, M., Thomas-Oates, J., Wilson, J.C., 2009. Species identification by analysis of bone collagen using matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry. Rapid Commun. Mass Spectron. 23, 3843e3854. Buckley, M., Melton, N.D., Montgomery, J., 2013. Proteomics analysis of ancient food vessel stitching reveals >4000-year-old milk protein. Rapid Commun. Mass Spectrom. 27 (4), 531e538. Camacho, C., Coulouris, G., Avagyan, V., Ma, N., Papadopoulos, J.e.a, 2009. BLASTþ: architecture and applications. BMC Bioinforma. 10, 421. Cappellini, E., Gilbert, M.T.P., Geuna, F., Fiorentino, G., Hall, A., Thomas-Oates, J., Ashton, P.D., Ashford, D.A., Arthur, P., Campos, P.F., Kool, J., Willerslev, E., Collins, M.J., 2010. A multidisciplinary study of archaeological grape seeds. Naturwissenschaften 97 (2), 205e217. Cappellini, E., Jensen, L.J., Szklarczyk, D., Ginolhac, A., da Fonseca, R.A.R., Stafford, T.W., Holen, S.R., Collins, M.J., Orlando, L., Willerslev, E., Gilbert, M.T.P., Olsen, J.V., 2012. Proteomic analysis of a pleistocene mammoth femur reveals more than one hundred ancient bone proteins. J. Proteome Res. 11 (2), 917e926. Cappellini, E., Collins, M.J., Gilbert, M.T.P., 2014. Unlocking ancient protein palimpsests. Science 343 (6177), 1320e1322. Cartechini, L., Vagnini, M., Palmieri, M., Pitzurra, L., Mello, T., Mazurek, J., Chiari, G., 2010. Immunodetection of proteins in ancient paint media. Acc. Chem. Res. 43 (6), 867e876. Chen, L., Ma, L., Zhou, M., Liu, Y., Zhang, Y., 2014. Effects of pressure on gelatinization of collagen and properties of extracted gelatins. Food Hydrocoll. 36, 316e322. Chiavari, G., Gandini, N., Russo, P., Fabbri, D., 1998. Characterisation of standard tempera painting layers containing proteinaceous binders by pyrolysis (/methylation)-gas chromatography-mass spectrometry. Chromatographia 47 (7e8), 420e426. Cox, J., Mann, M., 2008. MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat. Biotechnol. 26, 1367e1372. Cox, J., Neuhauser, N., Michalski, A., Scheltema, R.A., Olsen, J.V., et al., 2011. Andromeda: a peptide search engine integrated into the MaxQuant environment. J. Proteome Res. 10, 1794e1805. Craig, O., Mulville, J., Pearson, M.P., Sokol, R., Gelsthorpe, K., Stacey, R., Collins, M., 2000. Archaeology: detecting milk proteins in ancient pots. Nature 408 (6810), 312. Crone, P., 2007. Quraysh and the Roman army:making sense of the Meccan leather trade. Bull. SOAS 70 (1), 63e88. Dallongeville, S., Koperska, M., Garnier, N., Reille-Taillefert, G., Rolando, C., Tokarski, C., 2011. Identification of animal glue species in artworks using proteomics: application to a 18th century gilt sample. Anal. Chem. 83 (24), 9431e9437. Di Lullo, G.A., Sweeney, S.M., Korkko, J., Ala-Kokko, L., San Antonio, J.D., 2002.
71
Mapping the ligand-binding sites and disease-associated mutations on the most abundant protein in the human, type I collagen. J. Biol. Chem. 277 (6), 4223e4231. Gross-Klee, E., Hochuli, St, 2002. Die jungsteinzeitliche Doppelaxt von Cham-Eslen. Gesamtbericht über einen einzigartigen Fund aus dem Zugersee. Tugium 18, 69e101. Grünberg, J., 2002. Middle Palaeolithic birch-bark pitch. Antiquity 76, 15e16. Helwig, K., Monahan, V., Poulin, J., 2008. The identification of hafting adhesive on a slotted antler point from a southwest Yukon ice patch. Am. Antiq. 73 (2), 279e288. Hollemeyer, K., Altmeyer, W., Heinzle, E., Pitra, C., 2008. Species identification of Oetzi's clothing with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry based on peptide pattern similarities of hair digests. Rapid Commun. Mass Spectrom. 22 (18), 2751e2767. Kirby, D.P., Buckley, M., Promise, E., Trauger, S.A., Holdcraft, T.R., 2013. Identification of collagen-based materials in cultural heritage. Analyst 138 (17), 4849e4858. Koller, J., Baumer, U., Mania, D., 2001. High-tech in the middle Palaeolithic: Neandertal-manufactured pitch identified. Eur. J. Archaeol. 4, 385e397. Kuckova, S., Hynek, R., Kodicek, M., 2009. MALDI-MS applied to the analysis of protein paint binders. In: Colombini, M.P., Modugno, F. (Eds.), Organic Mass Spectrometry in Art and Archaeology. John Wiley & Sons, Chichester, pp. 165e187. Lucas, A., 1934. Ancient Egyptian Materials and Industries. Edward Arnold Publishers Ltd., p. 447 Lynnerup, N., 2007. Mummies. Year book of Physical Anthropology, vol. 50, pp. 162e190. Madigan, M.T., Martinko, J.M., Parker, J., 1997. Brock's Biology of Microorganisms. Prentice Hall College, p. 986. Maixner, F., Overath, T., Linke, D., Janko, M., Guerriero, G., van den Berg, B.H.J., Stade, B., Leidinger, P., Backes, C., Jaremek, M., Kneissl, B., Meder, B., Franke, A., Egarter-Vigl, E., Meese, E., Schwarz, A., Tholey, A., Zink, A., Keller, A., 2013. Paleoproteomic study of the Iceman's brain tissue. Cell. Mol. Life Sci. 70 (19), 3709e3722. Mazza, P.P.A., Martini, F., Sala, B., Magi, M., Colombini, M.P., Giachi, G., Landucci, F., Lemorini, C., Modugno, F., Ribechini, E., 2006. A new Palaeolithic discovery: tarhafted stone tools in a European Mid-Pleistocene bone-bearing bed. J. Archaeol. Sci. 33 (9), 1310e1318. Nesvizhskii, A.I., Aebersold, R., 2005. Interpretation of shotgun proteomic data: the protein inference problem. Mol. Cell. Proteomics 4 (10), 1419e1440. Nielsen-Marsh, C.M., Ostrom, P.H., Gandhi, H., Shapiro, B., Cooper, A., Hauschka, P.V., Collins, M.J., 2002. Sequence preservation of osteocalcin protein and mitochondrial DNA in bison bones older than 55 ka. Geology 30 (12), 1099e1102. Nissenbaum, A., 1997. 8000 years Collagen from Nahal Hemar Cave. Archaeol. Nat. Sci. 5, 5e9. NPCS Board of Consultants and Engineers, 2007. The Complete Book on Adhesives, Glues &. Resins Technology. Asia Pacific Business Press, Delhi, p. 696. Ostrom, P.H., Schall, M., Gandhi, H., Shen, T.-L., Hauschka, P.V., Strahler, J.R., Gage, D.A., 2000. New strategies for characterizing ancient proteins using matrix-assisted laser desorption ionization mass spectrometry. Geochim. Cosmochim. Acta 64 (6), 1043e1050. Prockop, J.D., Kivirikko, K.I., 1995. Collagens: molecular biology, diseases, and potentials for therapy. Annu. Rev. Biochem. 64 (1), 403e434. Pinhasi, R., Gasparian, B., Areshian, G., Zardaryan, D., Smith, A., Bar-Oz, G., Higham, T., Petraglia, M.D., 2010. First direct evidence of Chalcolithic Footwear from the near Eastern Highlands. PLoS ONE 5 (6), e10984. Regert, M., 2004. Investigating the history of prehistoric glues by gas chromatography-mass spectrometry. J. Sep. Sci. 27 (3), 244e254. Regert, M., Rolando, C., 2002. Identification of archaeological adhesives using direct inlet electron ionization mass spectrometry. Anal. Chem. 74 (5), 965e975. Schneider, U., Kenndler, E., 2001. Identification of plant and animal glues in museum objects by GC-MS, after catalytic hydrolysis of the proteins by the use of a cation exchanger, with simultaneous separation from the carbohydrates. Fresenius J. Anal. Chem. 371 (1), 81e87. Solazzo, C., Fitzhugh, W.W., Rolando, C., Tokarski, C., 2008. Identification of protein remains in archaeological potsherds by proteomics. Anal. Chem. 80 (12), 4590e4597. nne ho artefaktu z Medve dí Strnad, M., 1990. Chemick a analýza tmelu z koste u Ru jeskyne zína. Slov. Archeol. 38 (1), 39e41. , C., 2006. Identification of proteins in Tokarski, C., Martin, E., Rolando, C., Cren-Olive renaissance paintings by proteomics. Anal. Chem. 78 (5), 1494e1502. Vizcaino, J., Deutsch, E., Wang, R., Csordas, A., Reisinger, F., et al., 2014. ProteomeXchange provides globally co-ordinated proteomics data submission and dissemination. Nat. Biotechnol. 30, 223e226. Villa, P., Pollarolo, L., Degano, I., Birolo, L., Pasero, M., Biagioni, C., Douka, K., Vinciguerra, R., Lucejko, J.J., Wadley, L., Bicho, N., 2015. A milk and ochre paint mixture used 49,000 years ago at Sibudu, South Africa. PLoS ONE 10 (6), e0131273. Wadley, L., 2010. Compoundadhesive manufacture as a behavioral proxy for complex cognition in the middle stone age. Curr. Anthropol. 51 (1), 111e119.
Contents lists available at ScienceDirect
Journal of Archaeological Science journal homepage: http://www.elsevier.com/locate/jas
Molecular evidence of use of hide glue in 4th millennium BC Europe Niels Bleicher a, Christian Kelstrup b, Jesper V. Olsen b, Enrico Cappellini c, * a
City of Zürich, Office for Urbanism, Competence Centre for Underwater Archaeology and Dendrochronology, Switzerland Novo Nordisk Foundation Center for Protein Research, Faculty of Health Sciences, University of Copenhagen, Blegdamsvej 3B, 2200 Copenhagen, Denmark c Centre for GeoGenetics, Natural History Museum of Denmark, University of Copenhagen, Øster Voldgade 5-7, 1350 Copenhagen, Denmark b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 February 2015 Received in revised form 6 August 2015 Accepted 16 August 2015 Available online 18 August 2015
A well-preserved bow, dated by dendrochronology to 3176e3153 BC, was found at the waterlogged ra” in Zurich (Switzerland). The surface of the bow, made of yew (Taxus Neolithic site “Parkhaus Ope baccata), was decorated with bark strips from a different, broad-leaved, tree species. In order to investigate whether the bark decoration was fixed to the bow with hide or fish glue, mass spectrometry (MS)based ancient protein sequencing was attempted to detect possible traces of collagen residues. The sequences retrieved, in particular collagen type 3 (COL3A1), indicate that most probably skin, and possibly other slaughtering by-products, were used as the initial materials to produce hide glue. Amongst the candidate animal species that the glue could have originated from, cattle and domestic ovicaprids were confidently identified. This is, to the best of our knowledge, the oldest evidence of the use of animalbased glue in Europe. It demonstrates that in the late 4th millennium BC human communities, aside from benefitting from more commonplace primary and secondary products, also exploited domestic animals to extract a high value-added biochemical. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Ancient proteins Collagen Animal glue Domestic animals Neolithic Europe
1. Introduction Protein residues are probably the most abundant class of ancient biomolecules preserved in bones and soft tissues from archaeological and palaeontological contexts (Di Lullo et al., 2002; Lynnerup, 2007; Prockop and Kivirikko, 1995). Moreover, due to their chemical and mechanical properties, proteinaceous materials have also been used to satisfy primary human needs, including sheltering (Crone, 2007), transportation, clothing and food production (Hollemeyer et al., 2008; Kirby et al., 2013; Pinhasi et al., 2010). Production of adhesives, although less intuitive and more technologically refined, represented another way to use abundant animal proteins. In particular, hide glue, can be relatively easily extracted from animal skin, bone and other slaughtering by-products immersed in hot water. During this process, known as gelatinization, high temperature hydrolyses, denatures and solubilises collagen (Kuckova et al., 2009, 167). Collagen is a heterotrimer composed of three amino acid chains in a triple-helix structure. The transformation of collagen into gelatin is interpreted as the disintegration of the helical structures into
* Corresponding author. E-mail address: [email protected] (E. Cappellini). http://dx.doi.org/10.1016/j.jas.2015.08.012 0305-4403/© 2015 Elsevier Ltd. All rights reserved.
random coils (Chen et al., 2014). The use of adhesives is one of the most ancient human technological solutions for producing elaborate tools with superior mechanical properties, overcoming the technical limitations associated with products derived from a single component (Koller et al., 2001). Therefore, by allowing the production of objects made out of different parts and materials, the introduction and use of natural adhesives represents one of the most important breakthroughs in object manufacturing. Although easy to produce, even in rudimentary conditions, hide glue presents remarkably interesting properties, thus making it the adhesive of choice for many applications, until the relatively recent introduction of glues produced by organic synthesis (e.g. Brauns, 1858). It is well documented that birch bark tar was used as early as the Middle Palaeolithic (Koller et al., 2001; Grünberg, 2002; Mazza et al., 2006; Wadley, 2010). Other adhesives, such as conifer resins and their heated derivatives, have been identified as adhesives in later periods (Regert, 2004). Conifer resin has been documented on a limited number of archaeological finds from the 7th Millennium BC in Canada (Helwig et al., 2008), and on a 5000 yearold copper age arrow point (Strnad, 1990). Together with beeswax, other substances, such as pistacia resins and plant oils, have also been found on archaeological objects from Bronze and Iron Age in Europe (Regert and Rolando, 2002). Similarly, the most ancient
66
N. Bleicher et al. / Journal of Archaeological Science 63 (2015) 65e71
archaeological traces of a more elusive type of glue, based on collagen, originate from a site in Israel, radiocarbon-dated between 8310 and 8110 years before present (Nissenbaum, 1997). In Egypt, where glue was identified as an ingredient of plaster from the third millennium BC, an ingot of pure hide glue was also found at Deir-elBahari dating to the late second millennium BC (Lucas, 1934). To the best of our knowledge, however, no occurrence of animal glue has been documented in Neolithic Europe. Analytical methods for detection of proteinaceous residues in archaeological and artistic material initially relied on gas or liquid chromatographic methods (GC or HPLC) coupled to mass spectrometry (MS) (Schneider and Kenndler, 2001; Chiavari et al., 1998). These approaches only allow amino acid identification and tentative reconstruction of the proteins they possibly originated from. Immunological methods, namely Enzyme-Linked Immuno Sorbent Assay (ELISA), and ancient DNA sequencing were also occasionally reported for the identification of collagen and hide glue in artworks (Cartechini et al., 2010; Albertini et al., 2011). However, none of these approaches allow for protein sequencing. This limit was overcome with the recent introduction of mass spectrometrybased methods for ancient protein identification (Ostrom et al., 2000) and sequencing (Nielsen-Marsh et al., 2002). The use of liquid chromatography, coupled to tandem mass spectrometry (LCMS/MS), allowed species identification of biological samples, such as bone and skin, from archaeological contexts (Buckley et al., 2009; Brandt et al., 2014). The same approach was also applied to successfully identify protein binders in historical paintings (Tokarski et al., 2006), as well as food residues in pottery (Solazzo et al., 2008). Finally, a proteomics strategy based on LC-MS/MS was also used to identify the origin species of glues used in gilt from the 18th century (Dallongeville et al., 2011). In 2009, during the archaeological excavation of the Neolithic ra” in Zurich (Switzerland), settlement at the site “Parkhaus-Ope Fig. 1a and b, the discovery of an object made of yew (Taxus baccata), suggested the possible use of hide glue, Fig. 1c. The finding, dated by dendrochronology to 3176e3153 BC (Bleicher and Burger, in press.), was identified as the symbolic representation of a hunting bow. The preservation of wood was excellent due to the anaerobic conditions in the waterlogged deposit encasing the bow (Madigan et al., 1997; Bleicher and Schubert, 2015). The bark on its surface was shaped in silvery-reddish strips showing large lenticels, a feature absent in yew. Although a more detailed anatomical characterisation was not possible, these traits were sufficient to confidently assign the bark to a broadleaved species and tentatively identify the origin of the bark to a young cherry tree. The straight edges and regular distribution of the strips along all the length of the bow, Fig. 1c-inset, strongly indicated that they were intentionally glued as a decoration on the surface of the bow. In the Swiss Neolithic site of Cham Eslen, birch tar was utilised to glue birch bark to an axe handle (Gross-Klee and Hochuli, 2002). For the new find from Zurich, though, the use of birch tar can be ruled out since no dark residue was detected under the bark. Although in the Near East the use of collagen-based adhesives is recorded at much earlier times (Nissenbaum, 1997), so far no Neolithic European evidence of animal-based glue has been found. The earliest trace of adoption of hide glue in Europe is possibly a mention by Aristoteles (Anheuser, 2001). If confirmed by molecular evidence, the case we describe would represent the most ancient occurrence of hide glue in Europe, and provide the first positive evidence for its use in this area predating the written sources by several millennia. We set out to test whether hide glue or, taking into account that the prehistoric settlement was situated on the lakeshore, fish glue was used to affix the bark to the body of the bow. To achieve this,
we attempted LC-MS/MS proteomics analysis of organic residues extracted from wood flakes collected under the bark. 2. Materials and methods Immediately after unearthing the object, small wood flakes from two areas underneath two different bark strips, lifted from the body of the bow, were carefully removed using a razor blade. Since the sampled area had been covered by the bark for over 5000 years, chances of contamination before or during sample collection can be considered extremely limited. A negative control prepared and analysed following exactly the same procedure adopted for the archaeological samples, except for the initial addition of ancient material, was processed together with the samples. Each of the two samples, approximately 15 mg, was then transferred in a protein LoBind 1.5 mL tube (Eppendorf, Germany). Each sample was processed independently as follows. Each sample was re-suspended in 500 mL of 100 mM ammonium bicarbonate solution at pH 8.00 and was briefly washed by mechanical shaking at room temperature to remove soil residues and other debris, prior to pelleting by centrifugation at 14,000 g for 5 min. The supernatant was discarded and this step was repeated two times in total. The pellet was resuspended in 1 mL 1.2 M HCl and incubated at 4 C for 24 h to gently hydrolyse collagen. This approach was chosen with the aim of maximising recovery of collagen altered by prolonged exposure to protein crosslinkers present in soil, such as clay, humic and fulvic acids, as well as wood lignin. After centrifugation at 20,000 g for 20 min, the supernatant was removed and stored, while the pellet was re-suspended in 200 mL 50 mM ammonium bicarbonate pH 8.00 and incubated at 70 C for 24 h, to solubilise collagen. The pH was checked using pH indicator strips and adjusted to 8.00 with diluted ammonium hydroxide. The sample was then pelleted by centrifugation at 14,000 g for 10 min and the supernatant was collected and reduced by incubation for 1 h at 60 C with 5 mM (final concentration) of dithiothreitol (VWR-BDH, England), dissolved immediately before use. The reduced cysteines were then alkylated by incubating in the dark, at room temperature, for 45 min with 15 mM (final concentration) chloroacetamide (SigmaeAldrich, Denmark), dissolved immediately before use. The pH was then checked using pH-indicator strips, adjusted to 8.00 with diluted ammonium hydroxide and digestion was started by adding 4 mL of 0.5 mg/mL sequencing grade trypsin solution (Promega, Nacka Sweden) and incubating at 37 C overnight. The following morning, 2 mL of fresh trypsin were added and digestion was extended for 6 additional hours. Digestion was quenched with 10% trifluoroacetic acid to a final concentration of 0.2e0.8%, as necessary to reach pH < 2.00. After centrifugation at 14,000 g for 10 min, to precipitate any eventual insoluble residue, the tryptic peptides in the supernatant were immobilised on C18 stage tips as previously described (Cappellini et al., 2012). Peptide mixtures were analysed by online nanoflow reversedphase C18 liquid chromatography tandem mass spectrometry (LC-MS/MS), as described previously (Cappellini et al., 2012). Briefly, the LC-MS system consisted of an EASY-nLCTM system (Proxeon Biosystems, Odense, Denmark) connected to the LTQOrbitrap Velos (Thermo Electron, Bremen, Germany) through a nano-electrospray ion source. An aliquot of 5 uL of each peptide sample was auto-sampled onto and directly separated in a 15 cm analytical column (75 mm inner diameter) with a 90 min gradient from 5% to 30% acetonitrile in 0.5% acetic acid at a flow rate of 250 nL/min. The effluent from the HPLC was directly electrosprayed into the mass spectrometer by applying 2.2 kV through a platinumbased liquid-junction. The LTQ-Orbitrap Velos instrument was operated in datadependent mode to automatically switch between full scan MS
N. Bleicher et al. / Journal of Archaeological Science 63 (2015) 65e71
67
ra archaeological site (A and B) where the wood bow was found in a waterlogged deposit (C). Inset: detail of the bark surface. Fig. 1. Positioning of the Zürich-Parkhaus Ope
and MS/MS acquisition. Instrument control was through Tune 2.6.0 and Xcalibur 2.1 Survey, full scan MS spectra (from m/z 300e2000) were analysed in the orbitrap detector with resolution R ¼ 30,000 at m/z 400 (after accumulation to a ‘target value’ of 1e6 in the linear ion trap). The ten most intense peptide ions with charge states 2 were sequentially isolated to a target value of 5e4 and fragmented in octopole collision cell by Higher-energy Collisional Dissociation (HCD) with a normalized collision energy setting of 40%. The resulting fragments were detected in the Orbitrap system with resolution R ¼ 7500. The ion selection threshold was 5000 counts and the maximum allowed ion accumulation times were 500 ms for full scans and 250 ms for HCD. Standard mass spectrometric conditions for all experiments were: spray voltage, 2.2 kV; no sheath and auxiliary gas flow; heated capillary temperature, 275 C; predictive automatic gain control (pAGC) enabled, and an Slens RF level of 65%. For all full scan measurements with the Orbitrap detector a lock-mass ion from ambient air (m/z 445.120024) was used as an internal calibrant. Raw files generated during spectra acquisition were searched on a workstation using the MaxQuant (MQ) algorithm, v. 1.4.1.2 (Cox
and Mann, 2008), and the Andromeda peptide search engine (Cox et al., 2011) initially against a target/reverse custom-made list of all the mammalian Alpha-1 type I, Alpha-2 type I and Alpha-1 type III collagen sequences publicly available in UniProt and nrNCBI, as well as the complete list of collagen sequences available in NCBI RefSeq after restriction to the Actinopterygii taxon. Subsequently, the raw files were also searched against the target/reverse protein list in reference proteomes or extended proteins lists available in NCBI RefSeq for the following species: Homo sapiens (human, reference proteome -including isoforms-downloaded from UniProtKB on Dec. 31 2012), Bos taurus (cow, reference proteome Oct. 24 2012), Equus caballus (horse, reference proteome downloaded on April 23 2014), Ovis aries (sheep, 22,752 accessions, downloaded on Jan. 8, 2013), Capra hircus (goat, 31,461 accessions, downloaded on Dec. 2, 2013), Sus scrofa (pig, 33,148 accessions, downloaded on Jan. 26 2013), Oryctolagus cuniculus (rabbit, 21,150 accessions, downloaded July 22 2014) and Oncorhynchus mykiss (Rainbow trout, 3839 accessions, downloaded July 22 2014). In every search, spectra were also matched against the common contaminants, such as keratin and porcine trypsin sequences, downloaded from
68
N. Bleicher et al. / Journal of Archaeological Science 63 (2015) 65e71
Uniprot. Trypsin was selected as the proteolytic enzyme and two missed cleavages were allowed. Oxidation (M and P), deamidation (N and Q), acetylation (K), Q / pyro-Glu (N-term Q), and E / pyro-Glu (Nterm E) were selected as variable amino-acid modifications. Carbamidomethylation (C) was selected as fixed modification. Default values were used for precursor (6 ppm) and fragment (20 ppm) ions mass tolerance. False-discovery rate was set at 1% and minimum values for peptide-score and peptide sequence length were set at 70 and 7 respectively. The amount of random matches was evaluated performing MaxQuant search against reversed sequences. For the identification of species-diagnostic peptides, possible environmental contaminants, such as actin, tubulin, keratins and keratin-associated proteins, as well as proteins commonly used in mass spectrometry facilities as standards or calibrants and proteins highly conserved, such as histones, were excluded from further investigation. A minimum of two unique peptides was required for confidently identifying proteins. The Blastp algorithm (Camacho et al., 2009) was used to identify protein-unique peptides by aligning the peptides identified using MaxQuant against the entire nrNCBI protein database, as previously described (Brandt et al., 2014). 2.1. Data deposition note The mass spectrometry proteomics data were deposited in the ProteomeXchange Consortium (http://proteomecentral. proteomexchange.org) via the PRIDE partner repository (Vizcaino et al., 2014) with the data set identifier PXD00172700. 3. Results The analysis of the negative control sample, blank, did not identify any protein residue apart from the ordinary contaminants, such as keratins and porcine trypsin. Two spectral files were generated, one from each protein extract fraction. A total of 25,608 MS/MS spectra were acquired, of which 4134 (14.4%) could be identified. 101 of these provide confident identification of five proteins (relying on at least two unique sequence peptides), based on the assignment of 27 razor þ unique peptides, Tables 1 and 2. The term “razor” designates those peptide sequences matching with more than one protein group and assigned to the one with the largest number of total identified peptides. This approach follows the parsimony principle, to infer a list of the lowest possible number of proteins that can account for the observed peptides (Nesvizhskii and Aebersold, 2005). Four out of five sequences were identified as collagen proteins. Although most of the collagen peptides identified are extremely common among many animal species, some of them can be used as a marker to narrow down the animal species used as collagen source, Table 2 (Brandt et al., 2014). Based on the sequences
available in public protein databases at the time of data analysis, two of the peptides identified can be uniquely assigned to cattle (B. taurus) and one to domestic ovicaprids, i. e. sheep and goat (O. aries and C. hircus), Table 2 and Fig. 2a and b. Peptides were considered diagnostic when, after BLAST searching (Camacho et al., 2009) against the entire nrNCBI protein database, they were assigned to a single species, or to a limited number of species among which only one can be considered plausible. These considerations were based on the nature of the samples, such as geographic origin and distribution. For example, peptides equally present in cattle (B. taurus), water buffalo (Bubalus bubalis) and yak (Bos mutus) were considered diagnostic for cattle. Alternative animal sources for the peptides here assigned to domestic species cannot be completely excluded. This is due to the limited availability in public databases of protein sequences belonging to other non-domesticated ovicaprids, likely present in the same epoch and geographic area as the bow. Nevertheless, the co-occurrence of ovine and bovine bone remains within the archaeological context from which the bow was also retrieved from supports our species assignments. Identification of collagen alpha1(III), Fig. 2c, supports the hypothesis that the adhesive was produced from skin, as this molecule, together with collagen alpha-1 and -2(I), is particularly abundant in this organ. Although Collagen alpha-1(III) is also present in other organs, such as the lung and the vascular system, their use for glue preparation can be considered less probable arguably due to their primary use as a food source. No evidence of hare or fish collagen could be detected, providing no support for the use of glue originating from these animals. Finally two peptides confidently assigned to beta haemoglobin were also identified. Although their sequences, widely conserved among many species, do not provide any further information about the animal sources used to produce the hide glue, their identification suggests that traces of blood were present in the glue.
4. Discussion The results demonstrate how analytical approaches for protein sequencing have improved to the point of allowing unexpected application in waterlogged archaeological contexts, even from temperate environments and a wide range of epochs (Cappellini et al., 2010). Col1A1 and Col1A2 are ubiquitous in all the collagen-bearing tissues, while Coll3A1 occurs more abundantly in soft connective tissues and skin. Consequently, the collagen-based animal glue we detected was extracted, most probably, from ovine and cattle skin, as well as from other slaughtering byproducts. Identification of haemoglobin is compatible with the possibility that blood traces, originally associated with the materials used to produce hide glue, were accidentally co-extracted during the gelatinization process. Alternatively, deliberate inclusion of blood
Table 1 Identified proteins. N.
Protein name
All matching peptidesa
Razor þ unique peptides
Total seq. cov. [%]b
Unique þ razor sequence coverage [%]
Seq. length
Matched spectra
Species of origin
1 2 3 4 5
Collagen_alpha1(I) Collagen_alpha-2(I) Collagen_alpha-2(I) Collagen alpha-1(III) Haemoglobin sub. beta
10 8 9 4 2
10 2 9 4 2
11.1 8 9.3 3.9 15.6
11.1 2.1 9.3 3.9 15.6
1313 1364 1364 1467 147
42 22 26 8 3
mammals bovine sheep/goat ovine/bovine mammals
a
Including non-unique peptides. Based on all matching peptides. No peptide allowed discrimination between sheep and goat. Collagen_alpha1(I) and haemoglobin were identified only by peptides with very limited taxonomic variability among mammals. b
N. Bleicher et al. / Journal of Archaeological Science 63 (2015) 65e71
69
Table 2 Identified peptides. Protein N.a
Sequence
Length (aa)
Mass
Charge
MaxQuant score
Total matched spectra
1 1 1 1 1 1 1 1 1 1
GEPGPAGLPGPPGER GETGPAGPAGPIGPVGAR GETGPAGRPGEVGPPGPPGPAGEK GETGPSGPAGPTGAR GNSGEPGAPGSK GPSGPQGPSGPPGPK GSAGPPGATGFPGAAGR GSEGPQGVR GVQGPPGPAGPR GVVGLPGQR
15 18 24 15 12 15 17 9 12 9
1386.69 1559.81 2167.07 1310.62 1056.48 1315.65 1426.70 885.43 1088.57 881.51
2 2 3 2 2 2 2 2 2 2
88.42 135.72 108.57 144.80 93.35 187.16 113.48 111.17 132.25 119.29
4 3 1 5 1 7 2 1 11 7
2 2 2 2 2 2 2 2
AGVMGPAGSR GDGGPPGATGFPGAAGR GEAGPAGPAGPAGPR GSTGEIGPAGPPGPPGLR IGQPGAVGPAGIR PGPIGPAGAR SGETGASGPPGFVGEK VGAPGPAGAR
10 17 15 18 13 10 16 10
901.44 1440.67 1260.62 1615.83 1191.67 891.49 1475.69 851.46
2 2 2 2 2 2 2 2
93.33 114.63 135.65 118.21 81.92 84.08 72.58 117.81
1 6 4 4 4 1 1 1
3 3 3 3 3 3 3 3 3
AGVMGPAGSR GAPGAVGAPGPAGANGDR GDGGPPGATGFPGAAGR GEAGPAGPAGPAGPR GSTGEIGPAGPPGPPGLR PGPIGPAGAR TGEPGAAGPPGFVGEK TGQPGAVGPAGIR VGAPGPAGAR
10 18 17 15 18 10 16 13 10
901.44 1490.72 1440.67 1260.62 1615.83 891.49 1469.72 1179.64 851.46
2 2 2 2 2 2 2 2 2
93.33 105.90 114.63 135.65 118.21 84.08 82.35 112.84 117.81
1 1 6 4 4 1 2 6 1
4 4 4 4
GEVGPAGSPGSSGAPGQR GPAGANGLPGEK GPPGAGGPPGPR GPPGPPGTNGAPGQR
18 12 12 15
1566.74 1066.54 1015.52 1358.67
2 2 2 2
148.52 88.02 164.97 147.26
1 1 2 4
5 5
LLVVYPWTQR VNVDEVGGEALGR
10 13
1274 1314
2 2
104.22 99.94
1 2
Species indicator
References
bovine
Pep. A (Buckley '09)
bovine
Brandt et al., 2014
sheep/goat sheep/goat/pig
Brandt et al., 2014 Pep. A (Buckley '09)
ovine/bovine
References report previously identified peptides are cited. a Protein numbering according to Table 1.
in the glue formulation could represent an intriguing example of a highly refined technological understanding about the wood binding properties of different biological materials. Indeed, until just a few decades ago, blood-based glues had been extensively employed for industrial production of plywood (NPCS Board of Consultants and Engineers (2007)). Identification of collagen alpha-1(III), reveals that most probably skin, and possibly other slaughtering by-products, were used as source material for animal glue production. The case we present, as well as others available in literature (Kirby et al., 2013; Brandt et al., 2014), contribute to illustrate how ancient protein sequencing can provide conclusive and, unlike aDNA, tissue-specific ancient biomolecular evidence (Cappellini et al., 2014). Until recently, however, ancient protein sequencing has only been applied to archaeological animal macroscopic remains, such as bone and skin, either in their natural form (Buckley et al., 2009; Maixner et al., 2013), or processed to produce artefacts (Kirby et al., 2013; Brandt et al., 2014). Ancient protein residues interpreted as food remains were previously reported using immunological methods (Craig et al., 2000), and MS-based protein sequencing (Buckley et al., 2013; Solazzo et al., 2008). However this is the first time, to our knowledge, where proteomics methods can be used to confidently reconstruct manufacturing practices in prehistoric contexts. The application of this approach to characterise protein-based adhesives in archaeological contexts has been extremely limited. Only
Buckley et al. (2013) suggested that milk residues detected within archaeological vessels could represent traces of a sealant, and not necessarily traces of food. It can be anticipated that future research will make the recovery of proteinaceous glue traces from archaeological findings routine, thus allowing for a better reconstruction of the until recently invisible technological expertise and even ritual practices utilised by early ancient human communities. 5. Conclusions LC-MS/MS-based ancient protein sequencing demonstrates that, as early as the 4th millennium BC, farmers in the Zurich area performed rudimentary chemical extractions to produce the oldest so far known animal-based adhesive in Europe. Hide glue was most probably extracted from skins, as well as other connective tissues rich in collagen of cattle and domestic ovicaprids. Results show that the community adopting these procedures used livestock to extract a high value-added biochemical, besides generation of more commonplace primary (meat, hide, bone) and secondary products (milk, wool, traction) associated with farming. The possibility to relay on such an advanced, for the time, technology enabled the construction of elaborate objects with superior mechanical properties. This is the first time, as far as we know, where proteomics methods were used to confidently reconstruct manufacturing practices in prehistoric contexts. As very recently reported by Villa
70
N. Bleicher et al. / Journal of Archaeological Science 63 (2015) 65e71
Fig. 2. Examples of tandem MS/MS spectra supporting identification of species-diagnostic collagen peptides and collagen alpha-1(III). (A) MS/MS spectrum, confidently assigned to amino acid sequence SGETGASGPPGFVGEK, diagnostic for cattle; (B) MS/MS spectrum, confidently assigned to amino acid sequence TGEPGAAGPPGFVGEK diagnostic for domestic ovicaprids; (C) MS/MS spectrum, confidently assigned to amino acid sequence GPPGAGGPPGPR, diagnostic for ovine and bovine collagen alpha-1(III).
N. Bleicher et al. / Journal of Archaeological Science 63 (2015) 65e71
et al. (2015), systematic integration of GC-MS-based analysis of lipid/terpenoid fractions with proteomic studies will provide a much more comprehensive insight about ancient technological processes adopted by human and, possibly, their behaviours. Acknowledgements EC was supported by the Danish Council for Independent Research, grant n. 10-081390 (awarded to Prof. MTP Gilbert). The authors would like to thank Anna Katerina Fotakis for the technical help. References Albertini, E., Raggi, L., Vagnini, M., Sassolini, A., Achilli, A., Marconi, G., Cartechini, L., Veronesi, F., Falcinelli, M., Brunetti, B.G., Miliani, C., 2011. Tracing the biological origin of animal glues used in paintings through mitochondrial DNA analysis. Anal. Bioanal. Chem. 399 (9), 2987e2995. €tter Anheuser, K., 2001. Historische Klebstoffe und ihre Identifizierung. Arbeitsbla für Restaur. 19, 247e264. Bleicher, N., Burger, M., 2015. Pfahlfeldanalyse und Dendroarch€ aologie. In: ra. Eine neolithische Bleicher, N., Harb, Chr (Eds.), Zürich-Parkhaus Ope €ologie, Feuchtboden fundstelle. Band 1. Befunde, Schichten und Dendroarcha vol. 48. Monographien der Kantonsarch€ aologie Zürich, Zürich, Egg (in press). Bleicher, N., Schubert, C., 2015. Why are they still there? A model of accumulation and decay of organic prehistoric cultural deposits. J. Archaeol. Sci. 61, 277e286. Brandt, L.Ø., Schmidt, A.L., Mannering, U., Sarret, M., Kelstrup, C.D., Olsen, J.V., Cappellini, E., Hart, J.P., 2014. Species identification of archaeological skin objects from Danish Bogs: comparison between mass spectrometry-based peptide sequencing and microscopy-based methods. PLoS ONE 9 (9), e106875. Brauns, 1858. Über Kitt und Leim, nach Haslett und Hackley. Polytech. Cent. 9 (6), 89e91. Buckley, M., Collins, M., Thomas-Oates, J., Wilson, J.C., 2009. Species identification by analysis of bone collagen using matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry. Rapid Commun. Mass Spectron. 23, 3843e3854. Buckley, M., Melton, N.D., Montgomery, J., 2013. Proteomics analysis of ancient food vessel stitching reveals >4000-year-old milk protein. Rapid Commun. Mass Spectrom. 27 (4), 531e538. Camacho, C., Coulouris, G., Avagyan, V., Ma, N., Papadopoulos, J.e.a, 2009. BLASTþ: architecture and applications. BMC Bioinforma. 10, 421. Cappellini, E., Gilbert, M.T.P., Geuna, F., Fiorentino, G., Hall, A., Thomas-Oates, J., Ashton, P.D., Ashford, D.A., Arthur, P., Campos, P.F., Kool, J., Willerslev, E., Collins, M.J., 2010. A multidisciplinary study of archaeological grape seeds. Naturwissenschaften 97 (2), 205e217. Cappellini, E., Jensen, L.J., Szklarczyk, D., Ginolhac, A., da Fonseca, R.A.R., Stafford, T.W., Holen, S.R., Collins, M.J., Orlando, L., Willerslev, E., Gilbert, M.T.P., Olsen, J.V., 2012. Proteomic analysis of a pleistocene mammoth femur reveals more than one hundred ancient bone proteins. J. Proteome Res. 11 (2), 917e926. Cappellini, E., Collins, M.J., Gilbert, M.T.P., 2014. Unlocking ancient protein palimpsests. Science 343 (6177), 1320e1322. Cartechini, L., Vagnini, M., Palmieri, M., Pitzurra, L., Mello, T., Mazurek, J., Chiari, G., 2010. Immunodetection of proteins in ancient paint media. Acc. Chem. Res. 43 (6), 867e876. Chen, L., Ma, L., Zhou, M., Liu, Y., Zhang, Y., 2014. Effects of pressure on gelatinization of collagen and properties of extracted gelatins. Food Hydrocoll. 36, 316e322. Chiavari, G., Gandini, N., Russo, P., Fabbri, D., 1998. Characterisation of standard tempera painting layers containing proteinaceous binders by pyrolysis (/methylation)-gas chromatography-mass spectrometry. Chromatographia 47 (7e8), 420e426. Cox, J., Mann, M., 2008. MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat. Biotechnol. 26, 1367e1372. Cox, J., Neuhauser, N., Michalski, A., Scheltema, R.A., Olsen, J.V., et al., 2011. Andromeda: a peptide search engine integrated into the MaxQuant environment. J. Proteome Res. 10, 1794e1805. Craig, O., Mulville, J., Pearson, M.P., Sokol, R., Gelsthorpe, K., Stacey, R., Collins, M., 2000. Archaeology: detecting milk proteins in ancient pots. Nature 408 (6810), 312. Crone, P., 2007. Quraysh and the Roman army:making sense of the Meccan leather trade. Bull. SOAS 70 (1), 63e88. Dallongeville, S., Koperska, M., Garnier, N., Reille-Taillefert, G., Rolando, C., Tokarski, C., 2011. Identification of animal glue species in artworks using proteomics: application to a 18th century gilt sample. Anal. Chem. 83 (24), 9431e9437. Di Lullo, G.A., Sweeney, S.M., Korkko, J., Ala-Kokko, L., San Antonio, J.D., 2002.
71
Mapping the ligand-binding sites and disease-associated mutations on the most abundant protein in the human, type I collagen. J. Biol. Chem. 277 (6), 4223e4231. Gross-Klee, E., Hochuli, St, 2002. Die jungsteinzeitliche Doppelaxt von Cham-Eslen. Gesamtbericht über einen einzigartigen Fund aus dem Zugersee. Tugium 18, 69e101. Grünberg, J., 2002. Middle Palaeolithic birch-bark pitch. Antiquity 76, 15e16. Helwig, K., Monahan, V., Poulin, J., 2008. The identification of hafting adhesive on a slotted antler point from a southwest Yukon ice patch. Am. Antiq. 73 (2), 279e288. Hollemeyer, K., Altmeyer, W., Heinzle, E., Pitra, C., 2008. Species identification of Oetzi's clothing with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry based on peptide pattern similarities of hair digests. Rapid Commun. Mass Spectrom. 22 (18), 2751e2767. Kirby, D.P., Buckley, M., Promise, E., Trauger, S.A., Holdcraft, T.R., 2013. Identification of collagen-based materials in cultural heritage. Analyst 138 (17), 4849e4858. Koller, J., Baumer, U., Mania, D., 2001. High-tech in the middle Palaeolithic: Neandertal-manufactured pitch identified. Eur. J. Archaeol. 4, 385e397. Kuckova, S., Hynek, R., Kodicek, M., 2009. MALDI-MS applied to the analysis of protein paint binders. In: Colombini, M.P., Modugno, F. (Eds.), Organic Mass Spectrometry in Art and Archaeology. John Wiley & Sons, Chichester, pp. 165e187. Lucas, A., 1934. Ancient Egyptian Materials and Industries. Edward Arnold Publishers Ltd., p. 447 Lynnerup, N., 2007. Mummies. Year book of Physical Anthropology, vol. 50, pp. 162e190. Madigan, M.T., Martinko, J.M., Parker, J., 1997. Brock's Biology of Microorganisms. Prentice Hall College, p. 986. Maixner, F., Overath, T., Linke, D., Janko, M., Guerriero, G., van den Berg, B.H.J., Stade, B., Leidinger, P., Backes, C., Jaremek, M., Kneissl, B., Meder, B., Franke, A., Egarter-Vigl, E., Meese, E., Schwarz, A., Tholey, A., Zink, A., Keller, A., 2013. Paleoproteomic study of the Iceman's brain tissue. Cell. Mol. Life Sci. 70 (19), 3709e3722. Mazza, P.P.A., Martini, F., Sala, B., Magi, M., Colombini, M.P., Giachi, G., Landucci, F., Lemorini, C., Modugno, F., Ribechini, E., 2006. A new Palaeolithic discovery: tarhafted stone tools in a European Mid-Pleistocene bone-bearing bed. J. Archaeol. Sci. 33 (9), 1310e1318. Nesvizhskii, A.I., Aebersold, R., 2005. Interpretation of shotgun proteomic data: the protein inference problem. Mol. Cell. Proteomics 4 (10), 1419e1440. Nielsen-Marsh, C.M., Ostrom, P.H., Gandhi, H., Shapiro, B., Cooper, A., Hauschka, P.V., Collins, M.J., 2002. Sequence preservation of osteocalcin protein and mitochondrial DNA in bison bones older than 55 ka. Geology 30 (12), 1099e1102. Nissenbaum, A., 1997. 8000 years Collagen from Nahal Hemar Cave. Archaeol. Nat. Sci. 5, 5e9. NPCS Board of Consultants and Engineers, 2007. The Complete Book on Adhesives, Glues &. Resins Technology. Asia Pacific Business Press, Delhi, p. 696. Ostrom, P.H., Schall, M., Gandhi, H., Shen, T.-L., Hauschka, P.V., Strahler, J.R., Gage, D.A., 2000. New strategies for characterizing ancient proteins using matrix-assisted laser desorption ionization mass spectrometry. Geochim. Cosmochim. Acta 64 (6), 1043e1050. Prockop, J.D., Kivirikko, K.I., 1995. Collagens: molecular biology, diseases, and potentials for therapy. Annu. Rev. Biochem. 64 (1), 403e434. Pinhasi, R., Gasparian, B., Areshian, G., Zardaryan, D., Smith, A., Bar-Oz, G., Higham, T., Petraglia, M.D., 2010. First direct evidence of Chalcolithic Footwear from the near Eastern Highlands. PLoS ONE 5 (6), e10984. Regert, M., 2004. Investigating the history of prehistoric glues by gas chromatography-mass spectrometry. J. Sep. Sci. 27 (3), 244e254. Regert, M., Rolando, C., 2002. Identification of archaeological adhesives using direct inlet electron ionization mass spectrometry. Anal. Chem. 74 (5), 965e975. Schneider, U., Kenndler, E., 2001. Identification of plant and animal glues in museum objects by GC-MS, after catalytic hydrolysis of the proteins by the use of a cation exchanger, with simultaneous separation from the carbohydrates. Fresenius J. Anal. Chem. 371 (1), 81e87. Solazzo, C., Fitzhugh, W.W., Rolando, C., Tokarski, C., 2008. Identification of protein remains in archaeological potsherds by proteomics. Anal. Chem. 80 (12), 4590e4597. nne ho artefaktu z Medve dí Strnad, M., 1990. Chemick a analýza tmelu z koste u Ru jeskyne zína. Slov. Archeol. 38 (1), 39e41. , C., 2006. Identification of proteins in Tokarski, C., Martin, E., Rolando, C., Cren-Olive renaissance paintings by proteomics. Anal. Chem. 78 (5), 1494e1502. Vizcaino, J., Deutsch, E., Wang, R., Csordas, A., Reisinger, F., et al., 2014. ProteomeXchange provides globally co-ordinated proteomics data submission and dissemination. Nat. Biotechnol. 30, 223e226. Villa, P., Pollarolo, L., Degano, I., Birolo, L., Pasero, M., Biagioni, C., Douka, K., Vinciguerra, R., Lucejko, J.J., Wadley, L., Bicho, N., 2015. A milk and ochre paint mixture used 49,000 years ago at Sibudu, South Africa. PLoS ONE 10 (6), e0131273. Wadley, L., 2010. Compoundadhesive manufacture as a behavioral proxy for complex cognition in the middle stone age. Curr. Anthropol. 51 (1), 111e119.