Direct evidence of milk consumption from ancient human dental calculus

Direct evidence of milk consumption from ancient human dental calculusC. Warinner1, 2 n1, J. Hendy3 n1, C. Speller3, E. Cappellini4, R. Fischer5, C. Trachsel6, J. Arneborg7, 8, N. Lynnerup9, O. E. Craig3, D. M. Swallow10, A. Fotakis4, 11, R. J. Christensen11, J. V. Olsen11, A. Liebert10, N. Montalva10, 12, S. Fiddyment3, S. Charlton3, M. Mackie3, A. Canci13, A. Bouwman2, F. Rhli2, M. T. P. Gilbert4, 14 M. J. Collins3Scientific Reports 4, Article number: 7104 (2014)doi:10.1038/srep07104Download CitationArchaeologyBiomarkersProteomicsZoologyAbstractMilk is a major food of global economic importance, and its consumption is regarded as a classic example of gene culture evolution. Humans have exploited animal milk as a food resource for at least 8500 years, but the origins, spread, and scale of dairying remain poorly understood. Indirect lines of evidence, such as lipid isotopic ratios of pottery residues, faunal mortality profiles, and lactase persistence allele frequencies, provide a partial picture of this process; however, in order to understand how, where, and when humans consumed milk products, it is necessary to link evidence of consumption directly to individuals and their dairy livestock. Here we report the first direct evidence of milk consumption, the whey protein lactoglobulin (BLG), preserved in human dental calculus from the Bronze Age (ca. 3000 BCE) to the present day. Using protein tandem mass spectrometry, we demonstrate that BLG is a species specific biomarker of dairy consumption, and we identify individuals consuming cattle, sheep, and goat milk products in the archaeological record. We then apply this method to human dental calculus from Greenland's medieval Norse colonies, and report a decline of this biomarker leading up to the abandonment of the Norse Greenland colonies in the 15th century CE.IntroductionMilk is a major nutritional resource. In addition to being a source of clean liquid (milk is 80 90% water), milk solids contain approximately 25 55% sugar (lactose), 25 45% fat, and 5 35% protein (caseins and whey proteins), as well as calcium, potassium, and B vitamins1. Adoption of animal milk consumption by humans typically requires behavioral adaptations, such as culturing and curdling techniques, to remove or reduce the lactose content of milk in order to make dairy products digestible after infancy. Additionally, populations with long pastoralist traditions in Europe and India, East Africa, and the Arabian peninsula have also independently evolved lactase persistence (LP),van cleef and arpels necklace sale replica, a genetic adaptation in the regulation of the lactase gene (LCT) that allows continued adult digestion of milk2,3 (Fig. 1). LP is hailed as one of the clearest examples of gene culture co evolution in humans4, yet many fundamental aspects of its evolution remain unknown5,6,7,8,9 and the socioeconomic context and scale of prehistoric and historic dairying are only poorly understood. Helena). Pie charts for each region are scaled by sample size and indicate the proportion of individuals from each region testing positive for milk BLG peptides (black) in dental calculus. A pooled sample of five individuals from Norway testing positive for BLG is shown in gray indicating the uncertainty of the number of BLG+ individuals. Interpolated contour map of lactase persistence frequencies were generated from allele frequencies of all 5 known LP causal alleles ( 13907G, 13910T, 13915G, 14009G and 14010C) in present day populations in Europe, Africa, and northern Southwest Asia. The map was generated with the R statspat package32 using published data available as of March 2014 (see Methods section). Data points are shown as dots, and interpolation may be inaccurate where there are few data points.The ability to directly identify milk consumption patterns in past populations would thus advance understanding of human dietary ecology, evolution, and cultural agency. However, previous attempts to directly measure milk consumption using bone calcium isotopes have proven unsuccessful10. Other milk biomarkers offer only indirect lines of evidence. For example, isotopic inference of milk lipids from pottery residues11 is the most widely used approach to identify dairying, but this method cannot discriminate species of origin, and reuse of communal vessels and exploitation of cervid adipose tissue12 both pose further challenges to interpretation. Milk proteins have also been recovered from food residues13,14,15, but such finds are exceptional and rare, and animal bone evidence for the exploitation of secondary products is limited by the availability of large, well preserved assemblages to provide interpretable mortality profiles16,17.To address these problems we turned to dental calculus, a mineralized form of dental plaque that serves as a long term reservoir of dietary biomolecules and microfossils18. Nearly ubiquitous in archaeological populations and sourced directly from the oral cavity, dental calculus presents a unique opportunity to access primary evidence of ancient diets at an individual level. In a recent study18, we identified the milk whey protein lactoglobulin (BLG) in the dental calculus of a modern Swiss dental patient using tandem mass spectrometry. This protein was also recently identified within preserved kefir cheese curds associated with the mummified remains of Bronze Age Xiaohe pastoralists (ca. 1980 1450 BC) in Xinjiang, China15. BLG is a lipocalin within the calycin superfamily of proteins19, and it is the dominant whey (milk serum) protein in ruminant milk, making up 11% of the total milk proteins and 50% of the whey proteins20. BLG offers many advantages as a milk biomarker19,20 including: 1) humans do not produce BLG and therefore the presence of the protein in dental calculus excludes a host origin; 2) BLG is present only in milk and thus it is a specific biomarker for this fluid; 3) BLG is more resistant to enzymatic degradation and microbial proteolysis than other milk proteins21; 4) BLG lacks close bacterial orthologs, making it readily identifiable against a background of bacterial proteins; 5) over half of the amino acid residues in BLG are variable among traditional dairy livestock, allowing genus and species discrimination between cattle, buffalo, sheep, goat, horse, donkey, and reindeer, among others (Supplementary Fig. S1); 6) BLG is the dominant protein in the whey fraction of milk and partitions with lactose during dairy processing, thereby making BLG a superior proxy for lactose than caseins or milk fats, which separate from the lactose rich whey fraction during cheese and butter production20; 7) and finally, because BLG is identified directly from protein sequence data, uncertainties arising from indirect detection methods, such as isotopic analysis, are minimized.This study seeks to employ shotgun protein analysis by liquid chromatography tandem mass spectrometry to identify dairy consumption in the archaeological record. In this study, we analyzed 92 archaeological dental calculus samples selected from regions with (Europe and northern Southwest Asia) and without (Central West Africa) long standing dairying traditions for the presence of BLG (Fig. 1; Table 1). After establishing that BLG can be detected in the dental calculus of archaeological populations from dairy consuming regions, we then applied this approach to 6 additional medieval dental calculus specimens from the Norse Greenland sites Brattahli (Qassiarsuk) and Sandnes (Kilaarsarfik). These sites date to a period during which a major dietary shift from ruminant dairy to marine resources has been previously hypothesized on the basis of bone stable isotope and zooarchaeological evidence22,23. We confirm that a decline in BLG observed at these sites is consistent with a dietary shift, and specifically diminished access to dairy products, leading up to the abandonment of the Norse Greenland colonies in the 15th century CE.Table 1: Proteomic results of archaeological dental calculus samples analyzed in this studyFull size tableResultsLactase Persistence in Present Day PopulationsTo assist in the selection of archaeological samples for analysis, we generated an updated interpolated contour map of present day LP frequency in Europe, northern Southwest Asia, and Africa (Fig. 1) using recently published LP genotype data24,25. We then selected archaeological dental calculus specimens from sites located within regions with high (Northern Europe), moderate (Central Europe), low (northern Southwest Asia), and very low (Central West Africa) present day LP frequencies.BLG in Europe and northern Southwest AsiaSeventy four dental calculus samples were selected from 20 sites in Northern Europe (Britain, Denmark, and Norway; n = 40), Central Europe (Germany,replica clover necklace van cleef inspired, Hungary,van cleef and arpel necklace replica, and Italy; n = 38), and northern Southwest Asia (Armenia and Russia, n = 6) dating from the Bronze Age (ca. 3000 BCE) through the 19th century CE (Table 1; Supplementary Table S1). Approximately one quarter (25.7%) of the Eurasian dental calculus samples tested positive for BLG peptides (Fig. 1). In total, 229 spectra (representing 37 unique peptide sequences) from the Eurasian dataset were assigned to BLG (Table 1; Supplementary Table S2), resulting in a reconstruction of 72% of the protein (Fig. 2). For each of the 19 Eurasian dental calculus samples that tested positive for BLG, the consensus BLG sequence could be assigned to ruminants of the Pecora infraorder of Artiodactyla, and 18 samples contained bovid specific (Bovidae) peptides. Among these samples, 4 samples contained cattle specific (Bos sp.) peptides, 3 samples contained sheep specific (Ovis sp.) peptides, 2 samples contained goat specific (Capra sp.) peptides, and 3 samples contained BLG peptides from multiple ruminant species (Supplementary Fig. S1; Supplementary Table S2).Figure 2: Protein coverage of lactoglobulin identified within Eurasian archaeological dental calculus.(a) Human dental calculus from the British Anglo Saxon site of Norton on Tees (sample NEM18, ca. 6th century CE) found to contain seven lactoglobulin peptides. (b) Three dimensional structure of bovine lactoglobulin protein, rendered from PDB 3NPO using VMD v.1.9.136. The mapped locations of all BLG peptide sequences identified by tandem mass spectrometry within archaeological dental calculus are shown in red, resulting in a coverage of 72% of the reconstructed consensus BLG protein.Absence of BLG in Central West AfricaEighteen dental calculus samples were selected from a 19th century cemetery on the island of St. Helena, located approximately 2,000km west of Angola in the southern Atlantic Ocean. The cemetery contains the remains of Central West Africans26 originating from a region with traditionally very low or no milk consumption (Fig. 1; Supplementary Table S1). As expected, BLG peptides were not identified in any of the West African samples (Fig. 1; Table 1).BLG in Norse GreenlandRecent isotopic and faunal evidence suggests that Greenland Norse settlements shifted from an economy initially based on dairy to one increasingly reliant on marine mammals after the onset of the Little Ice Age ca. 1250 CE23. To test this hypothesis, we analyzed dental calculus from individuals buried at Tjodhildes Church, an early cemetery (ca. 985 1250 CE) at the Eastern settlement landnm site of Brattahli (Qassiarsuk) established by Erik the Red, the founder of the Norse Greenland colonies23,van cleef arpels alhambra necklace replica, and at Sandnes (Kilaarsarfik), a high status farm and church in the Western Settlement that continued to be in use until the abandonment of the settlement in the 15th century CE. The individuals analyzed from Tjodhildes Church (n = 2) exhibited strong evidence of dairy consumption (Fig. 3), with a total of 38 spectra matching BLG peptides (12 unique peptides including one Bos specific sequence). Because these two individuals had not been previously analyzed isotopically, we then performed carbon and nitrogen stable isotope analysis on bone collagen extracted from these individuals and confirmed that their isotopic values are consistent with a terrestrial diet (Figure 3). Analysis of the Sandnes individuals (n = 4) revealed that only one individual showed even weak evidence of milk consumption, as evidenced by a single spectrum (Fig. 3). This individual was previously determined to have consumed a primarily terrestrial diet on the basis of isotopic evidence23. The remaining BLG negative individuals exhibited bone stable isotope values consistent with increasing marine resource consumption (Figure 3).Figure 3: BLG pattern in dental calculus is consistent with bone collagen stable isotope evidence of a decline of the dairy economy in Norse Greenland with the onset of the Little Ice Age (ca.1250 CE).(a) Total spectra matching BLG peptides recovered from dental calculus samples from the earlier Tjodhildes Church at 29a Brattahli in the Eastern Settlement (individuals KAL1064 and KAL1052) and from the later V51 Sandnes site in the Western Settlement (individuals KAL934, KAL933, KAL936, and KAL930). (b) Bone collagen carbon and nitrogen stable isotope values measured from burials at Tjodhildes Church (black) and Sandnes (white)22,23, showing a major dietary shift toward marine resources at the later Sandnes site. Isotopic values for individuals also analyzed for dental calculus BLG peptides are represented by triangles, and from left to right on the x axis are: KAL1064, KAL1052, KAL934, KAL933, KAL936, and KAL930. Isotopic data for KAL1064 and KAL1052 were measured in this study.DiscussionOur results show that the milk protein BLG preserves in archaeological dental calculus and can be identified in specimens dating back to at least the Bronze Age (ca. 3000 BCE) in Europe and northern Southwest Asia. Moreover, we demonstrate that BLG is a species specific milk biomarker that allows cattle, sheep, and goat dairy product consumption to be distinguished. As expected, no BLG was detected in specimens from Central West Africa, where dairy consumption was historically very low or absent and current LP frequencies are very low.In addition to establishing BLG as a biomarker of milk consumption in the archaeological record, the broad survey of dental calculus BLG conducted in this study reveals previously uncharacterized temporal and geographical complexities in dairy consumption. In Central Europe, for example, it is intriguing given the high prevalence of LP in both modern day (Fig. 1) and medieval27 German populations that no BLG peptides were detected in Bronze Age or medieval German samples (0/9 individuals) (Table 1). This stands in contrast to the northern Italian and Hungarian samples where BLG peptides were detected at a relatively high frequency (5/11 individuals) (Table 1; Supplementary Table S2). Using this approach, it now becomes possible to explore specific cultural, social, and environmental factors influencing past dairy economies at both a population and an individual level.To explore a specific dairy economy in greater detail, we applied this approach to Norse Greenland to examine the hypothesis that this population underwent a dramatic dietary shift in response to environmental change during the Little Ice Age23. The medieval Greenland Norse economy was primarily based on animal husbandry and especially dairying. The short growing season and cold climate of Greenland precluded the successful establishment of agriculture but was sufficient for ruminant pastoralism. Like other Scandinavian populations, the medieval Norse utilized nearly all by products of dairy processing, including BLG rich whey28,29.Our dental calculus BLG results confirm that dairy products were consumed by individuals buried in the early Norse cemetery at Tjodhildes Church (ca. 985 1250 CE), and bone collagen stable isotopic values from the same individuals are consistent with a terrestrial diet. By contrast, BLG was very low or absent in the dental calculus of individuals buried at the Sandnam site, which was occupied until ca. 1430 CE, and bone stable isotope values from these individuals indicate a dietary shift toward marine resources.Diminished access to dairy products and a collapse of dairy herds would have had a strongly negative effect on the Norse economy, removing not only a major source of storable nutrition but also impeding the ability to preserve other perishable foods, such as meats, thereby exacerbating food instability especially during the winter months. Our findings support the hypothesis that climate change and the consequent decline of dairy herds contributed to the decline and ultimate abandonment of the Norse Greenland colonies in the 15th century CE.In this study, we identify the protein lactoglobulin in archaeological dental calculus and demonstrate that it is a species specific milk biomarker and an indicator of dairy consumption in the archaeological record. Nearly ubiquitous and obtained directly from the oral cavity of individuals, dental calculus provides a novel approach to detecting patterns of milk consumption and the dietary variables driving recent natural selection in humans. Latitude and longitude of the data points were taken as near as possible to the collection sites where these were known. Where country alone was known these were estimated using major cities. The contour map was constructed in 'R' (v.3.1.0, 2014 04 10, "Spring Dance") using the spatstat package32 and included weighting for sample size. Interpolation smoothing was conducted at the lowest non overflowing bandwidth (value of sigma) allowable from the heterogeneous data available. Interpolation may be inaccurate where there are few data points, and it should be noted that neighboring populations with different ancestry and life style, in Africa particularly, sometimes have very different allele frequencies.Samples and MS/MS analysisDental samples (n = 98) were obtained from diverse historic human populations in Eurasia, Africa, and Greenland dating from the Bronze Age to the present (Supplementary Table S1). Dental calculus was removed using a dental scaler and stored in sterile 2.0mL tubes until further analysis. Tryptic peptides were extracted from decalcified dental calculus using a filter aided sample preparation (FASP) protocol modified for degraded samples33 according to previously published protocols18. The extracted peptides were then analyzed using shotgun protein tandem mass spectrometry (MS/MS) to detect the presence of lactoglobulin. MS/MS analysis of samples were performed at three independent laboratories in Switzerland, the UK, and Denmark:Functional Genomics Center Zurich at the University of Zurich and Swiss Federal Institute of TechnologySamples from Greenland and Germany (Z1, Z2, Z27, Z46, and Z28) were analyzed by tandem mass spectrometry at the Functional Genomics Centre Zrich (FGCZ) using an LTQ Orbitrap VELOS mass spectrometer (Thermo Fischer Scientific, Bremen, Germany) coupled to an Eksigent NanoLC Ultra 1D plus HPLC system (Eksigent Technologies, Dublin (CA), USA). Solvent composition at the two channels was 0.2% formic acid, 1% acetonitrile for channel A and 0.2% formic acid, 100% acetonitrile for channel B. Peptides were loaded on a self made tip column (75m 80mm) packed with reverse phase C18 material (AQ, 3m 200, Bischoff GmbH, Leonberg, Germany) and eluted with a flow rate of 250nl per min by a gradient from 0.8% to 4.8% of B in 2min, 35% B at 57min, 48% B at 60min, 97% at 65min. Full scan MS spectra (3001700m/z) were acquired in the Orbitrap with a resolution of 30000 at 400m/z after accumulation to a target value of 1,000,000. Higher energy collision induced dissociation (HCD) MS/MS spectra were recorded in data dependent manner in the Orbitrap with a resolution of 7500 at 400m/z after accumulation to a target value of 100, 000. Precursors were isolated from the ten most intense signals above a threshold of 500 arbitrary units with an isolation wind