hide
Free keywords:
-
Abstract:
Collagen type I sequence variation has recently been used for the identification of archaeological bone fragments (ZooMS) and for phylogenetic analysis of extinct species. However, full collagen type I amino acid sequences are only available for a small number of relevant species. Proteomics provides a powerful way to characterize such amino acid sequences for species of interest. This requires a formal way of sequence authentication other than the phylogenetic position of the obtained sequence. Acceptance of previously uncharacterized amino acid sequences is commonly based on ion-score cut-offs and MASCOT peptide/protein scores. For collagen type I, this can be problematic as it consists of a large proportion of (hydroxylated) prolines. Incomplete fragmentation series, commonly associated with proline residues, could lead to false alanine<>serine substitutions next to or near proline residues.
Here, a large comparative dataset (+40 species) derived from high-coverage genome sequences is used to characterize amino acid substitution patterns among placental mammals. These patterns include col1a1 and col1a2 chain organization, average collagen substitution rates and frequencies of specific amino acid substitutions, and are seemingly irrespective of phylogenetic position of individual taxa.
We performed MS/MS sequencing using a combination of complementary digestion and platforms (Orbi-trap and maXis HD) to obtain de novo collagen type I sequences for modern and Pleistocene samples; obtaining roughly 80-90% sequence coverage for both col1α1 and col1α2 chains. De novo sequences are then authenticated by reference to the comparative genomic dataset, demonstrating the feasibility of de novo proteomics on archaeological and Pleistocene bone specimens.
