Abstract
The study of faecal samples to reconstruct the diets and habitats of extinct megafauna has traditionally relied on pollen and macrofossil analysis. DNA metabarcoding has emerged as a valuable tool to complement and refine these proxies. While published studies have compared the results of these three proxies for sediments, this comparison is currently lacking for permafrost preserved mammal faeces. Moreover, most metabarcoding studies have focused on a single plant-specific DNA marker region. In this study, we target both the commonly used chloroplast trnL P6 loop as well as nuclear ribosomal ITS (nrITS). The latter can increase taxonomic resolution of plant identifications but requires DNA to be relatively well preserved because of the target length (∼300–500 bp). We compare DNA results to pollen and macrofossil analyses from permafrost and ice-preserved faeces of Pleistocene and Holocene megafauna. Samples include woolly mammoth, horse, steppe bison as well as Holocene and extant caribou. Most plant identifications were found using DNA, likely because the studied faeces contained many vegetative remains that could not be identified using macrofossils or pollen. Several taxa were, however, identified to lower taxonomic levels uniquely with macrofossil and pollen analysis. The nrITS marker provides species level taxonomic resolution for commonly encountered plant families that are hard to distinguish using the other proxies (e.g. Asteraceae, Cyperaceae and Poaceae). Integrating the results from all proxies, we are able to accurately reconstruct known diets and habitats of the extant caribou. Applying this approach to the extinct mammals, we find that the Holocene horse and steppe bison were not strict grazers but mixed feeders living in a marshy wetland environment. The mammoths showed highly varying diets from different non-analogous habitats. This confirms the presence of a mosaic of habitats in the Pleistocene ‘mammoth steppe’ that mammoths could fully exploit due to their flexibility in food choice.
Highlights
Northern Canada were connected, forming a dry and largely treeless landmass known as Beringia (Hopkins, 1959; Hopkins et al, 1982)
We found most plant species in faecal samples uniquely using DNA, some of which abundantly so. This could be because of the large number of vegetative plant remains in the faeces which have become unidentifiable for macrofossil analysis due to masticatory and digestive processes
We further show that relatively long nuclear ribosomal ITS (nrITS) fragments can be amplified from faecal samples as old as 28,610 14C BP and that these help to increase species resolution for many plant families (e.g. Asteraceae, Cyperaceae and Poaceae) as well as mosses that could not be retrieved using trnL
Summary
Northern Canada were connected, forming a dry and largely treeless landmass known as Beringia (Hopkins, 1959; Hopkins et al, 1982). Recent studies have changed the view of the mammoth steppe vegetation into a more heterogeneous mosaic of different habitats This mosaic consisted of areas rich in shrubs combined with permanent moist areas and productive grasslands (Chytrý et al, 2019; Lozhkin et al, 2019; Zazula et al, 2006). Several studies on lake sediments have shown that instead of replacing traditional methods, DNA metabarcoding acts as a complementary proxy by revealing both additional taxa and providing increased taxonomic resolution Most studies of ancient DNA from sediments have relied either on the P6 loop of the chloroplast trnL (UAA) intron or the rbcL gene, and both give good taxonomic resolution for some plant taxa but limited for others (Sønstebø et al, 2010; Taberlet et al, 2006). NrITS has the advantage of being able to provide a higher taxonomic resolution, which in turn can give better insight into the palaeoenvironmental conditions represented by the taxa in a sample
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
