MALDI-TOF MS spectra of archaeological whale bone specimens from Atlantic Europe

Whale bones are regularly found during archaeological excavations. Identification of these specimens to taxonomic levels is problematic due to their fragmented state. This lack of taxonomic resolution limits understanding of the past spatiotemporal distributions of whale populations and reconstructions of early whaling activities. To overcome this challenge, we performed Zooarchaeology by Mass-Spectrometry on an unprecedented selection of 719 archaeological and palaeontological specimens of probable whale bone from Atlantic European contexts, from the Middle to Late Neolithic (c.3500–2500 BCE) to the eighteenth century CE.

The results show high numbers of Balaenidae (most likely North Atlantic right whale (Eubalaena glacialis)) and grey whale (Eschrichtius robustus) specimens, two species no longer present in the eastern North Atlantic. Many of these specimens derive from contexts associated with the known medieval whaling cultures of the Basques, northern Spaniards, Normans, Flemish, Frisians, Anglo-Saxons, and Scandinavians. This association raises the likelihood that pre-industrial whaling impacted these taxa, contributing to their extinction and extirpation respectively. Much lower numbers of other large whale taxa were identified, suggesting that it was once abundant and accessible whales that suffered the greatest long-term impact. The pattern of natural abundance leading to over-exploitation, well-documented for other taxa, is thus applicable to early whaling.

Methods
Whale bone samples were taken using a ©Dremel rotary tool removing a small piece of bone weighing up to c.500 mg. For 474 specimens, collagen was extracted using a modified Longin (1971) method as detailed in Seiler et al. (2019), with the addition of a lipid extraction step and the use of a higher acid concentration, at the National Laboratory for Age Determination, Norwegian University of Science and Technology, Norway.

Initially, the samples were crushed into small pieces and cleaned in an ultrasonic bath with 18.2 MΩ-cm ultrapure water (type 1) (three times five minutes). The samples were then ultrasonicated for 15 minutes in dichloromethane and methanol (2:1). This step was repeated three or more times until the solution was clear. Following this, the material was demineralized overnight using 2.44 M HCl (50 ml of solution per 100 mg of bone) in glass tubes within a vacuum desiccator kept at room temperature. The samples were then washed with ultrapure water until a pH of 3 to 4 and 4 ml of 0.5 % NaOH was added for 2–4 hours at room temperature to dissolve humic acids. After washing with ultrapure water till pH<10, 5 ml of 1.22 M HCl was added for 1–2 hours to remove atmospheric CO2. The sample was then washed with ultrapure water till pH=3±0.2 and hydrolysed to gelatine at 70 °C overnight. Finally, the gelatine was filtered through a prebaked (900 °C) quartz filter (Merck Millipore, AQFA04700, 99.998 % capture for 0.3 μm particles) and the filtrate freeze-dried. The final collagen was taken to the Henry Wellcome Laboratory for Biomolecular Archaeology, University of Cambridge, UK, for ZooMS analysis. 

Collagen extraction failed for 25 specimens and when sufficient original bone material remained a subsample was taken directly to the Henry Wellcome Laboratory. Additionally, another 237 specimens were analysed by ZooMS without prior collagen extraction. For these 262 (25+237) specimens, approximately 30 mg subsamples were taken and processed at the Henry Wellcome Laboratory for Biomolecular Archaeology. The bones were demineralised in 0.6 M hydrochloric acid at 4°C for two weeks. Each sample was then centrifuged, the hydrochloric acid discarded, and the retained material rinsed three times with 200 μl of 50 mMol ammonium bicarbonate (AmBic) pH 8.0 solution, before being gelatinised in 100 μl of AmBic solution at 65˚C for one hour. For the 449 collagen samples prepared at NTNU (474 samples, minus the 25 which failed), 0.1 mg of collagen was also placed in AmBic solution at 65˚C for one hour.

For both the 262 and 449 sample groups, the gelatinised collagen, still in AmBic, was incubated with 0.4 μg of trypsin at 37˚C overnight, and subsequently acidified with 0.1% trifluoroacetic acid (TFA). The collagen was then purified using a 100 μl C18 resin ZipTip® pipette tip (EMD Millipore) with conditioning and eluting solutions composed of 50% acetonitrile and 0.1% TFA and washing solution composed of 0.1% TFA; the samples were eluted in a volume of 50 μl. Equal amounts of the collagen extract and α-cyano-hydroxycinnamic acid matrix solution (1% in conditioning solution) were mixed (1 μl each) and spotted onto a 384 spot MALDI target plate. Each sample was externally calibrated against an adjacent spot containing a mixture of six peptides (des-Arg1-bradykinin m/z = 904.681, angiotensin I m/z = 1295.685, Glu1-fibrinopeptide B m/z = 1750.677, ACTH (1–17 clip) m/z = 2093.086, ACTH (18–39 clip) m/z = 2465.198 and ACTH (7–38 clip) m/z = 3657.929). Samples were spotted in triplicate, and run on a Bruker ultraflex III MALDI TOF/TOF mass spectrometer with a Nd:YAG smart beam laser. A SNAP averaging algorithm was used to obtain monoisotopic masses (C 4.9384, N 1.3577, O 1.4773, S 0.0417, H 7.7583). Averaged spectra were created from the replicates for each specimen using mMass software (Strohalm et al., 2008), and then visually compared to published m/z markers for mammals, as presented in Buckley et al. (2009), Kirby et al. (2013), Buckley et al. (2014) and Hufthammer et al. (2018). An additional eight specimens were previously analysed at BioArch at the University of York following the protocol outlined in Rodrigues et al. (2018) and the results incorporated into this study.

For additional information see manuscript.

考古发掘中经常发现鲸骨遗骸。但由于标本保存破碎，对其进行分类学鉴定存在较大难度。这一分类分辨率的缺失，限制了学界对历史时期鲸类种群时空分布的认知，也阻碍了早期捕鲸活动的重建研究。为解决这一难题，我们针对欧洲大西洋沿岸遗址中出土的、年代从中新石器时代至晚新石器时代（约公元前3500年至公元前2500年）直至公元18世纪的719件疑似鲸骨的考古及古生物标本，开展了前所未有的大规模动物考古质谱分析（Zooarchaeology by Mass-Spectrometry, ZooMS）。

分析结果显示，标本中须鲸科（Balaenidae，大概率为北大西洋露脊鲸*Eubalaena glacialis*）和灰鲸（*Eschrichtius robustus*）的数量占比极高，而这两个物种如今已不再出没于东北大西洋海域。其中多数标本出自与巴斯克、西班牙北部、诺曼、佛兰德斯、弗里西、盎格鲁-撒克逊及斯堪的纳维亚地区已知中世纪捕鲸文化相关的遗址层位。这一关联提示，前工业化时期的捕鲸活动可能对这些鲸类类群造成了冲击，分别促成了它们的区域性消亡与全球性灭绝。其余大型鲸类类群的鉴定数量则少得多，这表明曾大量存在且易于捕获的鲸类才是长期受影响最严重的类群。这种“自然丰度驱动过度开发”的模式——已在其他类群中得到充分证实——同样适用于早期捕鲸活动。

研究方法
本研究使用得美乐（Dremel）旋转工具采集少量骨样，单份样品重量不超过约500毫克。针对其中474件标本，我们在挪威科技大学年代测定国家实验室采用改良的Longin（1971）法提取胶原，具体操作参照Seiler等人（2019）的研究，并新增了脂质提取步骤，同时提高了酸溶液浓度。

初始步骤为将样品粉碎为小块，使用电阻率18.2 MΩ·cm的超纯水（Ⅰ型水）在超声清洗仪中清洗三次，每次5分钟。随后将样品置于二氯甲烷与甲醇混合液（体积比2:1）中超声处理15分钟，该步骤重复三次以上直至清洗液澄清。之后将样品置于玻璃管中，在室温真空干燥器内用2.44 M盐酸（每100毫克骨样使用50毫升溶液）进行脱矿过夜。接着用超纯水冲洗样品至pH值为3~4，再加入4毫升0.5%氢氧化钠溶液，于室温下静置2~4小时以溶解腐殖酸。再用超纯水冲洗至pH值＞10，随后加入5毫升1.22 M盐酸静置1~2小时以去除大气二氧化碳。之后用超纯水冲洗至pH值为3±0.2，将样品置于70℃下水解为明胶过夜。最后将明胶溶液通过预烘烤（900℃）的石英滤膜（默克密理博，AQFA04700，对0.3μm颗粒捕获率达99.998%）过滤，滤液经冷冻干燥得到最终胶原提取物。将该胶原提取物送至英国剑桥大学亨利·韦尔科姆生物分子考古实验室进行ZooMS分析。

有25件标本的胶原提取失败，若剩余足够的原始骨样，则直接取子样本送至亨利·韦尔科姆实验室。另有237件标本未预先进行胶原提取，直接开展ZooMS分析。针对这262件（25件提取失败+237件未提取）标本，取约30毫克子样本在亨利·韦尔科姆实验室进行处理：将骨样置于0.6 M盐酸中4℃脱矿两周，随后离心弃去盐酸，用200μl 50 mM碳酸氢铵（AmBic，pH 8.0）溶液冲洗留存的样品三次，再将样品置于100μl碳酸氢铵溶液中65℃水浴1小时以胶凝。对于在挪威科技大学制备的449份胶原样品（474份总样品减去25份提取失败的样品），同样取0.1毫克胶原置于碳酸氢铵溶液中65℃水浴1小时。

上述262份与449份样品组，均将仍处于碳酸氢铵溶液中的胶原样与0.4μg胰蛋白酶于37℃下孵育过夜，随后用0.1%三氟乙酸（TFA）酸化。之后使用100μl C18树脂ZipTip®移液吸头（EMD密理博）纯化胶原提取物：活化与洗脱液为50%乙腈+0.1% TFA，清洗液为0.1% TFA，最终以50μl体积洗脱样品。将等体积的胶原提取物与α-氰基-4-羟基肉桂酸基质溶液（1%溶于活化液中）混合（各取1μl），点样至384孔基质辅助激光解吸电离飞行时间/串联飞行时间（MALDI TOF/TOF）靶板上。每个样品以相邻的6肽混合点作为外标（校准肽分别为：去精氨酸1-缓激肽m/z=904.681、血管紧张素Ⅰm/z=1295.685、Glu1-纤维蛋白肽B m/z=1750.677、促肾上腺皮质激素（1–17片段）m/z=2093.086、促肾上腺皮质激素（18–39片段）m/z=2465.198、促肾上腺皮质激素（7–38片段）m/z=3657.929）。每个样品重复点样三次，使用搭载Nd:YAG智能光束激光器的布鲁克ultraflex III型MALDI TOF/TOF质谱仪进行检测。采用SNAP平均算法获取单同位素质量（C 4.9384、N 1.3577、O 1.4773、S 0.0417、H 7.7583）。使用mMass软件（Strohalm等人，2008）对每个标本的重复谱图进行平均，随后与已发表的哺乳动物m/z标记物（引自Buckley等人，2009；Kirby等人，2013；Buckley等人，2014；Hufthammer等人，2018）进行目视比对。另有8件标本此前按照Rodrigues等人（2018）的方案在约克大学生物考古实验室完成分析，其结果已纳入本研究。

更多详细信息请参阅研究手稿。