Researchers led by a group at the MRC Laboratory of Molecular Biology in Cambridge, UK, have engineered DNA polymerases that are better than standard Taq polymerase at PCR-amplifying ancient, damaged DNA.
The new polymerases, derived from genes from Thermus aquaticus, T. thermophilus, and T. flavus, might help to prepare damaged DNA samples for sequencing that conventional Taq polymerase has trouble amplifying. Such DNA is found in archeological, paleontological, forensic, and certain clinical samples.
For example, the enzymes might be “very valuable” for samples where PCR reactions with conventional polymerases failed, and researchers could not obtain any sequence data as a result, according to Stephan Schuster, a professor at Pennsylvania State University who co-authored a paper in late 2005 on sequencing DNA from a 27,000-year-old mammoth using 454’s platform (see GenomeWeb Daily News, In Sequence’s sister publication, 1/3/2006).
“What they have done is very, very impressive,” said Tom Gilbert, who has been working on ancient DNA as a postdoctoral fellow in the lab of Eske Willerslev at the University of Copenhagen, in an e-mail message. Gilbert said the researchers have “demonstrated that one can evolve polymerases to deal with damage” and laid “the groundwork for future experiments” to develop the enzymes further.
The Cambridge researchers, led by Phil Holliger, used a technique called molecular breeding followed by selection in an emulsion to generate polymerases that can amplify DNA with mismatches, process misaligned primer-template duplexes, and bypass certain types of damages, in particular abasic sites and hydantoins. They published their study, a collaboration with Svante Pääbo’s group at the Max Planck Institute for Molecular Anthropology in Leipzig, Germany, last week in Nature Biotechnology.
A previous study by Pääbo, whose group is sequencing the Neandertal genome with 454 Life Sciences, had shown that one type of damage in the DNA template, hydantoins, makes a PCR reaction particularly likely to fail, so “this seemed to be a particularly relevant type of damage” to focus on, Holliger told In Sequence last week.
The enzymes his group developed “turned out to be quite good, and certainly much better than the polymerases [that] are normally used for ancient DNA, at bypassing this type of damage,” he said.
Like Taq polymerase, Holliger’s polymerases were “very active” in PCR and were “more effective at getting sequences out of ancient DNA samples,” specifically, from two samples of cave bear DNA that are 47,000 and 60,000 years old, respectively, Holliger said.
While the enzymes can be used to amplify template DNA for sequencing with any kind of technology — both standard Sanger sequencing and next-generation technologies — they might be especially suitable for next-generation technologies that use emulsion PCR.
The reason is that the selection process to generate the enzymes involves emulsion PCR, Holliger said, “so the polymerases are pre-adapted to work in emulsions.” Both 454 and Applied Biosystems’ SOLiD system use ePCR to amplify the template DNA.
Since the enzymes amplify ancient DNA more efficiently than standard polymerases, they might also help with ancient samples that are contaminated with modern DNA, which usually takes over rapidly in a PCR reaction because it is amplified so much better. However, Holliger’s group has not tested that.
“Modern DNA amplifies so much better [that] even miniscule contaminations can be a huge issue,” Holliger said, particularly in projects involving ancient human DNA or DNA similar to human like the Neandertal’s.
According to Holliger, the long-term goal is to apply the new polymerases in ancient DNA projects in Pääbo’s lab.
Besides paleontological samples, the enzymes might also be used to analyze archaeological samples. “If it is verified to be true, it would be extremely useful,” said Angelique Corthals, a lecturer in biomedical and forensic studies at the University of Manchester, in an e-mail message.
However, Corthals, who has been studying mummies from Egypt, cautioned that the new enzymes might not be able to tackle certain types of damage in those samples.
“It is possible that the chemical actions of natron and resin, combined with the high temperature and sometimes humidity in which the bodies have been preserved, would effectively create lesions that are different and wider ranging than those addressed by [the new enzymes],” she said.
Other types of damages that the new enzymes might not be able to deal with are crosslinks and highly fragmented DNA, Gilbert said.
Also, in certain clinical samples, chemical or radiation treatments may have damaged the DNA beyond the new enzymes’ ability to work, according to Holliger. However, researchers could create specialized polymerases for these applications, he suggested.
“One could envisage that one would tailor them for a particular application, where they are particularly good at bypassing the type of damage one would expect in this particular setting.”
“Our approach shows a general strategy how one would generate polymerases [that] can do this,” he said. “One could envisage that one would tailor them for a particular application, where they are particularly good at bypassing the type of damage one would expect in this particular setting.”
Penn State’s Schuster said that the new polymerases might not be as useful for DNA samples that are well preserved, like the mammoth samples taken from permafrost that he has been working with. This is the case because the enzymes are likely to have a larger error rate than standard PCR polymerases, “so we might actually decrease the sequence quality that we would get.” However, for samples in which standard PCR fails, “people would rather have an error-prone sequence than no sequence,” he added.
Gilbert agreed that the high error rate is the “biggest problem” of the new enzymes, which he also said “incorporate bases into lesions that do not necessarily pair with the opposite strand.”
Holliger acknowledged that the new polymerases had a three- to four-fold lower fidelity than Taq polymerase in ancient DNA PCR. However, he said this “is not a problem, as the correct ancient sequence is easily reconstructed by sequencing multiple clones.”
Gilbert also pointed out that the Cambridge researchers only compared the new enzymes to regular Taq polymerase in their article, and not to other, more sensitive and more accurate polymerases like the Platinum Taq Hifidelity polymerase from Invitrogen that his lab has been using.
Holliger’s group is currently working on a “number of bottlenecks” that, if solved, might further improve the polymerases’ ability to amplify ancient DNA. Such bottlenecks include DNA crosslinks, which “present a particularly thorny problem” for a polymerase, he said.
Also, biological decay often releases polymerase inhibitors, he said, preventing the enzymes from amplifying the DNA in the test tube. “If we could evolve polymerases that are resistant to this, I think we would once again take a big step forward, because the sample preparation causes big losses in DNA,” Holliger said.
The best approach in the future might be to use a blend of different types of polymerases, he said, each optimized for different purposes. “Some would be good to bypass specific types of lesions, some more tolerant to inhibitors,” he explained.
Holliger already has collaborations to apply the enzymes to different types of damaged DNA samples, and is making the polymerases available to academic researchers. He has several patents or patent applications on both the polymerases and the polymerase selection technology, which he said the MRC exclusively licensed to Domantis, a company developing therapeutic antibodies that GlaxoSmithKline agreed to acquire last year.
“I hope these will be used widely, and hopefully will be useful as well,” he said. “They really represent a step towards a goal.”