BIOL 606 Home

The Use of Ancient DNA to Infer Phylogenies

Lecture by Greg Wilson

Rapporteur: Chris Kyle

Molecular data has been used successfully to infer phylogenies of many species. This raises the question: how useful is such data in elucidating phylogenies of extant and ancient species? The first challenge is to extract ancient DNA (aDNA) in order to infer a phylogeny. Successful extraction of such material may be hampered by several factors including the decomposition of macromolecules through the mechanisms of hydrolysis and oxidation. The hydrolysis of DNA can lead to depurination of G/A residues whereas oxidation may lead to fragile DNA strands susceptible to breakage due to weakened T/A ribose bonds. Hence, the rate of DNA damage is strongly correlated to the presence of water and oxygen as well as temperature. The possibility that aDNA might still be intact was illustrated by the pollination of seeds found in ancient tombs. Infact, some of the first attempts at extracting aDNA were from Egyptian mummies (approx. 2500yra) found in tombs where large amounts of DNA were preserved under ideal dry and oxygen-free conditions. The advent of PCR has opened up new possibilities for the analysis of very low concentrations of DNA which could potentially be found in older specimens and perhaps those found in less than ideal conditions. Amplification of DNA by PCR may, however, lead to serious problems such as erroneous DNA sequences being created because of mistakes by the polymerase at damaged DNA sites and an increase in errors due to very low concentrations of original DNA template. These problems are relatively well documented from studies claiming to have amplified dinosaur DNA showing over 10% sequence divergence between bones of what may or may not be an individual specimen and which resulted in a sequence more closely related to humans than birds. This result demonstrated that contamination is a huge factor when dealing with such low concentrations of DNA and that it must be dealt with by performing numerous controls before claims of aDNA amplification can be made. It was thought that amber fossils may be a good source of aDNA as they show remarkably well preserved tissues, probably due to the desiccating effect of resins. One experiment which attempted to amplify DNA from an amber-fossil using the 18s rRNA region of mtDNA produced unexpected and unreproducable results. Problems in this case might be due to the permeability of resins to gases, hence oxidative decomposition.

These initial attempts at amplifying aDNA have lead to experiments which have tried to determine if useable aDNA exists in fossils. Levels of amino acid racemization were found to be good indicators of DNA damage which potentially lead to errors in amplification or no amplification at all. Protein biosynthesis takes place by levrorotatory (L) entantiomers which become dextrarotatory (D) as a cell dies and decomposes. A ratio of 0.08 D to L entantiomers was found to be the maximum level relating to DNA damage which will result in successful amplification. This ratio is roughly correlated to a time of about 100, 000 yra for specimens under ideal conditions.

The study by Krings et al, 1997 attempts to amplify Neandertal DNA from a fossil preserved under ideal conditions to determine the phylogenetic relationship of this hominid to modern humans. This relationship is thought to fall under one of the following three hypotheses: The Out of Africa Theory where humans would not be direct descendants of Neandertals; The Replacement with Hybridization Theory where some hybridization may have occurred between the two species before the demise of Neandertals, and finally the Multiregional Hypothesis where Neandertals evolved directly into modern humans. Fossil evidence from Western Europe and the Middle East are most closely in agreement with the Out of Africa theory whereas Eastern European fossils suggest that the other two hypotheses may be the case.

This experiment was very thorough, using many negative controls to detect contamination and performing racemization tests to determine if the Neandertal DNA was indeed amplifiable. After these conditions were met cloning techniques were used to uncover any errors in the amplified sequence. These errors would be due to polymerase incorporation errors at damaged DNA sites. New primers were also designed in an attempt to determine if the sequence under analysis was not simply a nuclear DNA sequence insert of mtDNA and the result of contamination. All of these experiments strongly reinforced the conclusion that ancient Neandertal DNA was being amplified. Further cloning experiments were then used for adjacent amplification products resulting in a 379bp region of the hypervariable mtDNA sequence. The majority sequence from cloning results was taken as the true sequence. Comparison of the resulting majority sequence to that of known human variation revealed that the variation in this sequence was outside that which would be expected from a modern human segment. It was also found that using the chimp/human time of divergence that Neandertals were not the direct ancestors of humans.

These results demonstrate that aDNA can be extracted from fossils of up to 100,000 yra if found in a suitable environment (dry, oxygen deprived, and cold) to minimize DNA damage. These results also show that aDNA can be useful in determining phylogenies of extant and ancient species although more than one sample would, ofcourse, be preferred for the purposes of comparison.


Disscussants: Colin Reynolds and Yazdan Kievany

Rapporteur: Chris Kyle

The discussion begins by asking how Neandertals and humans coexisted; be it sympatrically, allopatrically, or otherwise. What isolation mechanisms would have led to no mtDNA being exchanged between the two groups? The conclusion drawn was that there may have been some genetic exchange between the two and that may not be reflected by mtDNA which is usually maternally inherited. This question was then expanded to ask if there was some hybridization between humans and Neandertals, and if human mtDNA was present in Neandertals, would this not simply be disregarded as contamination? The suggestion was that this would, indeed, be a difficult case to deal with under these circumstances. To determine if hybridization is a potential problem, a much larger sample size would be needed to undercover it, especially depending on the direction of the flow of genetic information and hence which maternal line would be revealed in Neandertals. Simply determining the time the two groups diverged could also be a good indicator of the potential for hybridization.

Another possibility which could lead to a clear demonstration of how Neandertals and humans were related would be to have fossils from a similar time period where the two groups were known to coexist. Neandertals and humans have been discovered in Israel where the fossils have been preserved in dry caves, but not cool temperatures other than during the last glaciation. Amplification of DNA from these fossils could potentially give conclusive evidence of this relationship and the origins of modern humans.

It is always difficult to know how much phylogenetic information can be drawn from one individual. Could this individual be unique, and therefore not representative of a population? Similar fossils exist, but DNA has not been successfully extracted from them. The likelihood of finding similar fossils which are as well preserved are rather slim.

The example of black bears, which contain mtDNA haplotypes known to be some 500, 000 yrs divergent from one another, was brought up to illustrate the point that even though a great deal of genetic variation exists between the Neandertal sequence and current human mtDNA sequences it does not necessarily mean they are not the same species. Hence the criticism of how some information was presented in the paper. It was suggested that the divergence between Neandertals and humans could have been better illustrated by showing which mutations in the Neandertal sequence related to the variable sites of current human haplotypes (i.e. show the unique mutations). This would clearly show that the sequences are totally divergent and that Neandertals are clearly not the precursors to modern humans. It was then suggested that this was not illustrated, along with a more detailed phylogenetic analysis, because there were too few unique differences in the Neandertal sequence to demonstrate that is was obviously unrelated to humans. Or was this simply an oversight by the authors?

Experimental controls are necessary to test for the insertion of aDNA into nuclear DNA sequences in such cases, but in the discussion it was mentioned that rather than redesigning primers one might simply try a Southern blot to determine if more than one area of DNA is amplified by a specific set of primers.

Cautionary notes were made by Krings et al. (1997) about the destructive sampling of such rare specimens. This was also mentioned in relation to amber-fossils, some of which are relatively plentiful, and may show the greatest potential for amplifying sequences from greater than 100,000 ya fossils. It was questioned if the proteins persist in these samples (as shown by the excellent preservation of tissues and the levels of racemization) and the DNA is simply oxidized over time due to the permeability of the resins which make up amber. There has been very limited reproducibility of results from such samples, whereas in studies on mammoths the results have been consistent.

Literature Cited:

Krings, M., A. Stone, R.W. Schmitz, H. Krainitzki, M. Stoneking, and S. Paabo. 1997. Neandertal DNA Sequences and the Origin of Modern Humans. Cell 90: 19-30.

BIOL 606 Home