Potential taxonomic implications of endogenous ancient proteins in vertebrate fossils from the Paleogene Ergilin Dzo Formation, Mongolia

Inaba, H,. Chiba, K., Saneyoshi, M., Miyaji, T., Kawakami, A., Eda, M., Tanaka, M., Kobayashi, Y., Tsogtbaatar, Kh., Tsujigiwa, H.

Society of Vertebrate Paleontology 84th annual meeting

Society of Vertebrate Paleontology

Minneapolis, Minnesota, USA

Proteomic techniques have been increasingly applied in paleontological research, facilitating taxonomic assessments of extinct taxa through molecular biology. To date, taxonomically-informative ancient protein sequences have been recovered only from fossil materials up to approximately 6.5 Ma. The endogenous nature of recovered proteins in older fossils and their taxonomic implications remain subjects of debate. In this study, improvements in histological staining and extraction techniques are employed to assess the preservation of endogenous collagen in late Eocene vertebrate fossils, and the resulting taxonomic implications are discussed. A partial rhinocerotoid mandible, an anthracotheriid astragalus, and a testudinid carapace fragment were sampled from the upper Eocene (approximately 33.9 Ma) Ergilin Dzo Formation in Mongolia. Initially, the histological staining for assessing collagen presence in extant samples was conducted, and all samples were positive. Subsequently, protocols optimized for collagen extraction in modern samples were performed with SDSPAGE. The extracts were separated into four bands in all samples, but the bands were absent in the negative control. The molecular weights of the bands from the fossil samples appeared to correspond to those of α-1, α-2, β, and γ chains of type-I collagen. Next, we employed mass spectrometry on some bands derived from each fossil sample. De novo sequencing of the spectra data using PEAKS recovered partial, relatively short sequences. Nevertheless, the recovered sequences exhibited repetitive occurrences of glycine in every three residues, characteristic of collagen. Additionally, database searches for spectra with common contaminants and type-I collagen sequences of extant vertebrates were iteratively employed using Mascot. The results indicated that the data of rhinocerotoid and anthracotheriid fossils matched most closely with perissodactyls and artiodactyls, respectively, in the laurasiatherian database. The spectral data of the testudinid fossil matched peptides in the type-I collagen of extant testudines in the archelosaurian database. Our findings indicate the presence of taxonomically-informative endogenous collagen in the late Eocene vertebrate fossils examined here, surpassing previous studies by a magnitude of five. This study represents methodological improvements, thereby extending the temporal range of molecular phylogenetic analyses on fossil vertebrates deeper into the fossil record.