Brachylophosaurus canadensis, Sternberg, 1953

Michael Buckley, Stacey Warwood, Bart van Dongen, Andrew C. Kitchener & Phillip L. Manning, 2017, A fossil protein chimera; difficulties in discriminating dinosaur peptide sequences from modern cross-contamination, Proceedings of the Royal Society, Series B 284, pp. 544-544 : 544-

(b) Brachylophosaurus canadensis collagen sequences

Following the initial 2007 report [ 5]̗ the same team reported similar collagen peptide sequence matches from a hadrosaurine dinosaur̗ an approximately 78 Ma Brachylophosaurus canadensis (MOR 2598; table 1 ) [ 14]. However̗ although it had already been suggested that standards be set in place̗ like those for the field of ancient DNA̗ this second study once again aimed to rely on an immunological approach as the main line of support̗ despite the ability to record chemical decay within proteins through PTMs such as oxidations and deamidations [ 15 – 17]̗ or even a range of others identified by the same team in ancient moa [ 18].

Following Schweitzer et al. [ 14]̗ there have been no further published attempts to verify the endogeneity of either published samples of purported dinosaur collagen sequences from other research groups̗ despite the lack of potential means to clarify the extent of decay within the proteins̗ of which we would expect substantial alteration [ 7]; members of the same team subsequently went on to report even more exceptional peptide matches to soft-tissue structures̗ in which they interestingly did report on the levels of deamidation and made clear attempts to separate modern from fossil material during the laboratory process [ 19]. The published record to date could be considered to lean in favour of endogeneity̗ with Peterson et al. [ 20] arguing against the microbial biofilm interpretation̗ suggesting that the crystallization of microbial biofilms on decomposing organic matter within vertebrate bone in early taphonomic stages may contribute to the preservation of primary soft tissues deeper in the bone structure [ 14].

A subsequent study mapping the molecular locations of the matched collagen peptides from both dinosaurs also implied that it was functionally significant regions of the collagen fibrils that were matched [ 21]. Although it was suggested that this non-random distribution could support the hypothesis that the peptides are produced from the extinct organisms̗ while also suggesting a chemical mechanism for survival̗ it does not rule out cross-contamination in which the same ‘mechanism for survival’ could equally apply to enhanced likelihood of contaminant peptides. More recently̗ a second collagen-based study has been published that placed further emphasis on the cleaning of the instrumentation used in addition to separate laboratories for extant and fossil material [ 22]̗ presenting an overlapping set of peptides. Intriguingly̗ these do not include the peptide sequence found as unique to both dinosaurs ( table 2). As a result̗ the phylogenetic analysis of this latest extraction places the Brachylophosaurus as sister-group to alligators as well [ 22]̗ clearly highlighting concern regarding the limitations of the study to date. They do̗ however̗ all match with peptides from alligator type 1 collagen̗ a species concurrently analysed in their previous works as modern reference material [ 5 ̗ 14] even if not necessarily contamination caused at the time of the most recent sampling.

Given that the only reports that appear to favour the most recent studies cannot rule out cross-contamination̗ we set out to test whether or not the reported set of unique collagen peptides (i.e. [ 5 ̗ 14]̗ excluding [ 22] as not containing unique peptides) could simply reflect cross-sample contamination from the modern reference material used; in this case̗ ostrich ( S. camelus ) bone (alligator was also used in the latter study̗ but not evaluated here in determining the unique dinosaur peptide because it was not used in the earlier study). In this study̗ we aimed to investigate the differences between sequences from ostrich bone collagen and those reported for both T. rex ( MOR 1125 ) and B. canadensis (MOR 2598).

Tables

Table 2. Peptide matches to Brachylophosaurus collagen from Schroeter et al. [22] and the taxa they can be found in by BLAST search in parentheses, with an emphasis on the potential contaminants. Underlined residues indicate post-translational modification (oxidation of P). The emboldened peptide is reported as having two hydroxylated prolines even though we routinely observe this peptide with only one (at P3), with a nearby A–S transition identified previously as being problematic to distinguish; they acknowledge in the electronic supplementary material that it could be either.

peptide sequence	protein	taxa
*GSAGPPGATGFPGAAGR	COL1A1	Tyrannosaurus rex , chicken, mammals (and others, including alligator)
*GATGAPGIAGAPGFPGAR	COL1A1	Tyrannosaurus rex , chicken, alligator and amphibians (and other reptiles and birds)
GFPGADGIAGPK (GFPGADGIsGPK)	COL1A1	ostrich and others (*alligator)
GFPGLPGPSGEPGK	COL1A1	alligator and ostrich (and others, ranging from fish to mammals)
GQAGVMGFPGPK	COL1A1	alligator and ostrich (and others)
EGPVGFPGADGR	COL1A2	alligator and ostrich (and others, including reptiles, birds and mammals)
GATGLPGVAGAPGLPGPR	COL1A2	alligator (and rodents)
GEPGNIGFPGPK	COL1A2	alligator and ostrich (and others, including birds and mammals)

*Note that these were the two peptides observed in both analyses [ 5].