Ancient Cave Bear DNA Successfully Sequenced

The Pleistocene Perspective

In a thrilling leap backward for science, the new approach to sequencing ancient genomic DNA pushes the age limit on what we can sequence fully beyond 40,000 years. Previous studies of similarly aged specimens have been limited to mitochondrial DNA, which helps in defining the branches on the tree of life but doesn't address the actual changes in function that set them apart. Besides more than doubling the timescale on which scientists can now study the evolution of whole genomes, the opening of the Pleistocene era to functional genomic inquiry promises crucial information about how we humans evolved. This time window includes early modern humans such as Cro-Magnons, and Neandertals, commonly considered our closest ancestor species. If the techniques used here to sequence the cave bear can be applied to the remains of Pleistocene hominids, we will learn much about the descent of man and what differentiates modern and ancient hominids.

Judging by the success of a recent effort to sequence genomic DNA from Pleistocene cave bears, the door is now wide open to the sequencing of extinct species from that era. Researchers from JGI, Lawrence Berkeley National Laboratory, the Max Planck Institute for Evolutionary Anthropology, and Lawrence Livermore National Laboratory have obtained informative DNA sequences from cave bear bones more than 40,000 years old. Their work offers strong evidence of the feasibility of eventually sequencing Pleistocene hominids (e.g., Neandertals).

Cave bear tooth (CB1) and bone (CB2) used to make libraries for sequencing.

The collaborators developed a new metagenomic approach to sequencing ancient samples. They began by extracting DNA from a cave bear tooth found in Ochsenhalt Cave, Austria, and from a long bone found in Gamssulzen Cave, Austria. The DNA was not amplified (which would favor modern contaminant sequences over the scarce ancient ones), and the only enzymatic treatment before ligation was blunt-end repair. Each sample was used to generate one metagenomic library, containing DNA from all organisms in that environment. The work was carried out in a laboratory into which no modern carnivore DNA had ever been introduced.

The team compared sequences from the cave bear metagenomic libraries to sequences in GenBank and to the whole dog genome. Sequences with BLAST hits to the dog genome (averaging 92.4% and 92.3% identity for the two libraries) were considered candidate cave bear sequences. The researchers made primers for 124 of the candidate sequences, using these to successfully amplify and sequence 116 orthologous sequences from modern brown bears. All 116 of these showed 97% identity or better with the cave bear sequences. In total, the cave bear sequences accounted for 1.1% of the sequence obtained from the tooth sample and 5.8% of that from the long-bone sample. Other sources of DNA that matched the sample sequences included bacteria and archaea, eukaryotes such as plants and fungi, and a tiny proportion of modern human (only 6 of 14, 027 sequenced clones). As is typical for metagenomic samples, most of the sequences did not match any sequences in public databases.

Predicted sources fo the two DNA libraries, CB1 and CB2, as determined by BLAST comparisons against GenBank sequences and against environmental sequences. "Other" indicates viral or plasmid-derived DNAs.

In all, 26,861 base pairs of cave bear genomic sequence were obtained. Based on the percentage of cave bear DNA identified from the long-bone library, the researchers estimated that, if sequenced completely, that library could yield 10x coverage of the cave bear genome.

Phylogenetic relationships between modern and ancient bears, determined by maximum-likelihood estimation, excluding two heavily damaged cave bear clones.

To discover how useful the identified cave bear sequences might be in phylogenetic studies, the scientists generated and compared orthologous sequences from cave bear and modern brown, black, and polar bears. A maximum likelihood analysis of the sequences yielded phylogenies that substantively match those previously obtained by comparing mitochondrial DNA sequences for the same species. Besides providing further evidence that the sequences identified as cave bear sequences really are such, this analysis demonstrated that the sequence data from such a study can be used to infer phylogenetic relationships.

BLAST hits from cave bear sequences to the dog genome and hits to modern human sequences differed significantly in insert length and in the percentage of the insert covered. This, together with the relatively small proportion of contamination by human DNA, bodes well for future attempts to sequence early hominids. Thus, the metagenomic approach is very promising for future studies of ancient specimens and likely to aid in identifying important functional differences between ancient and modern species.

Authors

J.P. Noonan and E.M. Rubin (JGI and Lawrence Berkeley National Laboratory); M. Hofreiter, N. Rohland, J. Krause, and S. Paabo (Max Planck Institute for Evolutionary Anthropology); D. Smith (JGI); J.R. Priest (Lawrence Berkeley National Laboratory); J. Rabeder (University of Vienna); and J.C. Detter (JGI and Lawrence Livermore National Laboratory).

Publication

"Genomic Sequencing of Pleistocene Cave Bears," Science Express, 02 June 2005, doi: 10.1126/science.1113485.

Funding

This research was funded by the U.S. Department of Energy Office of Biological and Environmental Research.