Isolation and characterisation of naturally occurring cross-linked collagen peptides using mass spectrometry
Catherine Gilbert1, Katell Bathany1, Benoit Bertrand2, Valery Hedouin2, Caroline Tokarski1
1 CBMN, UMR CNRS 5248, Proteome Platform, University of Bordeaux, F-33000 Bordeaux, France
2 Univ. Lille, CHU Lille, ULR 7367 – UTML&A – Forensic Taphonomy & Anatomy, F-59000 Lille, France
Collagen, the most abundant extra-cellular matrix protein in the body, significantly contributes to the high mechanical strength of many tissues, due to enzymatically controlled cross-linking of the collagen network. This can be a key biomarker in many biological processes, such as cancer, gestation, and osteogenesis imperfecta. We are interested in the analysis of collagen cross-links to identify biomarkers to determine the post-mortem interval, focusing on the structural analysis of collagen cross-linking in forensic bones using optimized proteomic workflows.
To characterise cross-linked collagen peptides with MS, there are significant challenges, primarily due to the low abundance and structural heterogeneity of naturally occurring cross-links, but also due to the lack of bioinformatic support for the proteomic analysis of the mature trivalent cross-links in collagen. We present a multi-angle analysis strategy to characterise the different stages of collagen cross-linking. This include the analysis of post-translational modifications (PTMs) that are known to be associated with cross-linking, the analysis of divalent cross-links using currently available bioinformatic tools, and the targeted, multi-stage enrichment of trivalent cross-links.
For the optimal extraction of cross-linked collagen peptides, an optimised protocol has been developed efficiently solubilising collagenous tissue, through a multi-stage extraction protocol targeting insoluble collagen using chemical digestion.
For the analysis of divalent, immature cross-links (eg. Ketonorleucines), a targeted reduction was performed in order to stabilise the cross-link structure before proteomic sample preparation, followed by fractionation of the resulting digest. These fractions were then analysed using nanoLC-MS/MS (Orbitrap, Thermo Fisher Scientific), data were searched for cross-links using XlinkX (Thermo Fischer Scientific), and cross-link identifications were manually confirmed and correlated with PTM identification to identify potential cross-link sites.
We focus on the structural analysis of trivalent, mature cross-links (eg. pyridinolines); as these are more likely to persist in forensic samples, in comparison to the chemically unstable immature cross-links. To isolate naturally occurring, trivalent, cross-linked peptides, a multi-chromatographic separation was used to enrich cross-linked collagen peptides based on fluorescence activity. These fractions were analysed using nanoLC-MS/MS (Orbitrap), with targeted middle-down proteomics methodologies combined with MS3 fragmentation methods to fully characterise isolated peptides. With the low concentration of naturally-occurring cross-linked peptides in biological tissues, as well as the structural heterogeneity of the cross-links and of the highly modified collagen protein itself, a multi-level enrichment procedure was necessary to confidently identify and characterise such structures using targeted MS analysis.
Event Timeslots (1)