by Pbo, Svante
Chapter 9
Nuclear Tests
___________________
Henrik’s work on the X chromosome had shown that the patterns of similarities and differences seen in the mitochondrial DNA of humans and apes were extendable to at least one part of the nuclear genome. Whether we would ever be able to study nuclear DNA from Neanderthals or be forever limited to their mitochondrial genome was not at all apparent. In my darker moments, I thought we were going to be stuck with mtDNA’s blurry, one-eyed view of human history. Certainly, if one disregarded results from animals and plants embedded in amber, dinosaurs, and other fanciful “antediluvian” studies (which I did), nobody had yet succeeded in retrieving any nuclear DNA from ancient remains. But in my more considered moments, I felt that we should give it a try.
It was at this point that Alex Greenwood, a diminutive but determined new postdoc from the United States, arrived in the lab. I told him about our hopes for retrieving nuclear DNA from Neanderthals, noting that it was a high-risk project but also a very important one. He was eager to take on the challenge.
I suggested a “brute-force” approach. My plan was to test samples of many bones to find those with the most mtDNA and then extract DNA from yet larger samples in an attempt to retrieve any nuclear DNA. This approach meant that we could not perform our initial experiments with the uncertain technique on Neanderthal remains; they were too scarce and valuable to use when the risk of failure was so high. Instead we resorted to animal bones, which were both considerably more abundant and less valuable to paleontologists. The cave-bear bones I had brought back from the dark basement of the Quaternary Institute in Zagreb now came in handy. They had been found in Vindija Cave, a limestone cave that had also produced some Neanderthal remains that contained mtDNA. So if we were able to retrieve nuclear DNA from the cave bears, we could hope to do the same with the Neanderthals.
Alex began by extracting DNA from the Croatian cave-bear bones, which were between 30,000 and 40,000 years old, and checked to see if they contained any bear-like mtDNA. Many of them did. He then took the extracts that seemed to contain the most mtDNA and tried to amplify short fragments of nuclear DNA. This failed. He was frustrated, and I was dismayed but not surprised. The problem he faced was a familiar one to me: because each cell in a living animal contains hundreds of mitochondrial genomes but only two nuclear genomes, any particular piece of nuclear DNA was present in 100- or 1,000-fold fewer copies in the extracts than any particular piece of mitochondrial DNA. So even if some nuclear DNA was present in minute amounts, the chances of amplifying it were a 100- or 1,000-fold lower.
One obvious way to overcome this problem was to simply use more bone. Alex made extracts of ever larger amounts of cave-bear bone and tried amplifying ever shorter pieces of nuclear DNA using primers flanking nucleotides where bears were known to differ from humans. That would enable him to discriminate between ancient bear DNA and contaminating human DNA. But in these mega-extracts, nothing could be amplified—not even bear mitochondrial DNA. He got no products at all.
After several weeks of repeated failures with multiple bones, we realized it was impossible to make useful DNA extracts from such large amounts of bone material. This was not because the bones contained nothing to amplify but because the extracts contained something that inhibited the enzyme used for the PCR; it became inactive and no amplification at all took place. We struggled to purify the unknown inhibitor away from the DNA in the extract but failed. We diluted the extracts in small steps until they started working again for the amplification of mtDNA. Then, at that dilution, we tried the nuclear amplification. It always failed. I tried to remain upbeat but as the months passed, Alex became more and more frustrated and anxious about whether he would ever produce any results that would justify a paper. We began to wonder if after a bear’s death the nuclear DNA might be degraded by enzymes leaking through the nuclear membrane of the decaying cells. Perhaps the DNA in mitochondria, having a double membrane, would have been better protected, making the mtDNA more likely to survive until the tissue dried out, froze, or was otherwise protected from enzymatic attack. This possibility made me wonder whether it would be possible to find nuclear DNA in ancient bones at all, even if we could overcome the inhibition of the PCR. I was slowly becoming as frustrated as Alex.
Thwarted by the cave bears, and wondering whether the conditions in the cave may simply have been too unfavorable to preserve nuclear DNA, we decided to switch to material that we expected to show the very best preservation—permafrost remains of mammoths from Siberia and Alaska. These had been frozen ever since the animals died and freezing, of course, will slow down and even stop both bacterial growth and many chemical reactions, including those that degrade DNA over time. We also knew, from Matthias Höss’s work, that mammoths from the Siberian permafrost tended to contain large amounts of mtDNA. Of course, no Neanderthals had ever been found in the permafrost—so switching to mammoths meant taking a step away from my ultimate goal. But we needed to know whether nuclear DNA could survive over tens of thousands of years. If we found no nuclear DNA in the frozen remains of mammoths, then we could forget about finding it in Neanderthal bones preserved under much less ideal conditions.
Fortunately I had systematically collected ancient bones from different museums over the past few years, so Alex could immediately try remains of several mammoths. He found one mammoth tooth that contained particularly large amounts of mtDNA. It had been pulled out of the frozen ground when the Alaska Highway, extending from northeastern British Columbia to near Fairbanks, was built in great haste during World War II and stored ever since in a huge box in the American Museum of Natural History. To make the search for DNA a bit easier, we carefully targeted a segment of the nuclear genome containing part of the gene known as 28S rDNA, which encodes an RNA molecule that is part of the ribosome, a structure that synthesizes proteins in cells. For our purposes, this gene had the great advantage of existing in a few hundred copies per cell. It should thus have been about as abundant as mitochondrial DNA in the extracts, assuming that the nuclear DNA had not been degraded more than mitochondrial DNA after death. To my delight and profound relief, Alex could amplify the ribosomal gene. He sequenced clones of the mammoth PCR products and reconstructed the gene’s sequence using the overlapping-segments approach we had established when studying Neanderthal mtDNA. He then wanted to compare the sequence to those from African and Asian elephants, the closest living relatives of mammoths. I had been so paranoid about contamination that, until Alex had the mammoth results, I had forbidden him or anyone else to work on elephants. But now, working outside our clean room, Alex used the same primers he had used for the mammoth work to amplify and sequence the 28S rDNA fragment from an African and an Asian elephant. The mammoth sequences were identical to those of the Asian elephant but differed at two positions from the African elephant version, suggesting that mammoths were more closely related to Asian than to African elephants. But comparing mammoths to living elephants hadn’t been the point of the exercise: finding ancient nuclear DNA had been. To clinch it, we sent a bit of the tooth off for carbon dating. When the 14,000-year-old date came back, I felt satisfied for the first time in months. It was now official. These were the first nuclear DNA sequences ever determined from the late Pleistocene.
Encouraged by these results, Alex designed primers for amplifying two short pieces of a fragment of the von Willebrand factor gene, only one copy of which exists in the elephant genome. This gene, abbreviated vWF, encodes a blood protein that helps platelets stick to wounded blood vessels. We focused on it because others had already sequenced it from elephants (and many other extant mammals), so if we managed to determine a sequence from the mammoth, we could directly compare it to those present-day sequences. I could hardly believe my eyes when, during our weekly lab meeting, Alex showed pictures of bands in a gel that suggested he was able to amplify these gene fragments from the mammoth. He repeated the experiment twice, using an independently prepared
extract from the same mammoth bone. Among the many clones he sequenced, he saw errors in individual molecules, presumably due either to chemical damage in the old DNA or to the DNA polymerase’s addition of the incorrect nucleotide during the PCR cycles (see Figure 9.1). At one position, however, he saw an interesting pattern. He had sequenced a total of thirty clones from three independent PCR amplifications. At one position, fifteen of those clones carried a C, fourteen carried a T, and one carried an A. The single A, we assumed, was an error caused by the DNA polymerase, but the others represented something that made my heart beat a bit faster. This particular spot in the sequence was clearly what geneticists call a heterozygous position, or a single nucleotide polymorphism (SNP for short), a place where the two copies of the gene that this mammoth had received from its mama and papa mammoth differed. What we saw was the first heterozygous position or SNP from the Ice Age. This was, if you like, the essence of genetics—a nuclear gene that has two variants in a population. Things were looking up. If we could see both versions of this mammoth gene, then all parts of the genome must be potentially accessible. It should thus be possible, at least in theory, to retrieve any genetic information we wanted from a species that went extinct many thousands of years ago. To drive home this point, Alex amplified pieces of two more single-copy genes: one encoding a protein regulating the release of neurotransmitters in the brain and one encoding a protein that binds vitamin A and is produced by the rods and cones in the eyes. He was successful in both cases.
Since we had struggled so long to retrieve nuclear DNA, Alex’s mammoth results were very welcome indeed and for several days I was very happy about them. But of course I wasn’t all that interested in mammoths. I was interested in Neanderthals, and I was painfully aware that there were no Neanderthals in permafrost. I urged Alex to go back and try the cave-bear remains from Vindija again, to see if he could retrieve nuclear DNA from remains that were not frozen. He analyzed mitochondrial DNA from several Croatian cave bears and identified one bone that seemed to contain a lot of it. We carbon-dated it and found it to be 33,000 years old, and thus roughly contemporaneous with the Neanderthals. Alex concentrated on this bone. He tried the ribosomal RNA gene that occurs in many copies in the genome. He obtained small amounts of an amplification product. Reconstructing its sequence from clones, he found that the sequence was indeed identical to that in present-day bears.
Figure 9.1. Cloned DNA sequences from three amplifications of a nuclear gene fragment from a 14,000-year-old mammoth. The arrow points to the first heterozygous position or SNP ever observed from the Late Pleistocene. From A. D. Greenwood et al., “Nuclear DNA sequences from Late Pleistocene megafauna,” Molecular Biology and Evolution 16, 1466–1473 (1999).
This was a success, but one with a dark side. It was so hard to amplify this multicopy-gene fragment that any attempt to get single-copy genes such as the vWF gene he studied in the mammoth seemed doomed to fail. Alex tried it anyway, of course, but as expected he was unsuccessful. So, after all the excitement over the mammoth results, I was—secretly—deeply disappointed by these experiments. We had demonstrated that nuclear DNA could survive over tens of thousands of years in permafrost, but only traces of very common nuclear DNA sequences could be found in the bones of cave bears. There was an enormous difference between permafrost and limestone caves.
In 1999, we published Alex’s findings in what I considered to be a beautiful paper, though it was subsequently largely overlooked. {40} It demonstrated that nuclear DNA survives in remains found in permafrost and that even heterozygous positions, where the two chromosomes in an individual differ in DNA sequence, can be determined. We were optimistic about the prospects for genetic research in the permafrost, and we noted at the end of the paper that
A plethora of faunal remains exist in permafrost deposits and other cold environments. The fact that such remains can yield not only mtDNA but also single-copy nuclear DNA sequences in a substantial amount of cases opens up the possibilities of using nuclear loci in phylogenetic and population genetic studies and of studying genes determining phenotypic traits.
Eventually others would take up this line of work, although not for another five to ten years. Worse, barring the discovery of a Neanderthal in permafrost, it seemed we might never see the whole genome of a Neanderthal.
Chapter 10
Going Nuclear
____________________
In the lab, I went about the business of supervising the experiments that pushed the work slowly but reliably forward. But whenever I was confined to a small seat during a long airplane ride, or to a darkened lecture hall during a seemingly irrelevant presentation at a conference, I came back to my greatest frustration: our inability to retrieve nuclear DNA from Neanderthals. I felt that it had to be there, even if the PCR could not retrieve it. We simply had to come up with a better way to find it.
A fresh attempt in this direction was made by Hendrik Poinar. Worn down by his fruitless quest for DNA from fauna and flora encased millions of years ago in amber, he had decided to move on to more promising endeavors. Fortunately for us I had just spent time at some boring conference lectures, where I got to thinking about the work we’d done on retrieving DNA from animal droppings. One of those we’d studied was the extinct American ground sloth, an Ice Age animal. The giant sloths had left behind large amounts of droppings, which archaeologists dressed up with the fancy name of coprolites. In fact, in some caves in places like Nevada, the entire floor, to some depth, is largely made up of old ground-sloth feces. In a paper in Science in 1998, Hendrik had already shown that mitochondrial DNA was preserved in such material and we described the retrieval of plant DNA from a single sloth bolus, showing how it could be used to reconstruct the ingredients of the meal the sloth had ingested shortly before its death 20,000 years ago.{41} This success suggested that lots of DNA, even nuclear DNA, was preserved in ancient fecal remains. I suggested that Hendrik try to find it.
Hendrik had begun the search with a chemical trick we had developed the year before. Back in 1985, when I was analyzing the mummies from Berlin, I had noticed that almost all of the extracts contained a component that produced blue fluorescence in UV light, and that if an extract fluoresced blue, it never yielded DNA. I did not know what this component was, but the observation was painfully memorable because of the disappointment I felt when I saw the blue instead of the pink glow I was hoping for. As I learned more about the chemistry that may have gone on in dead tissues over thousands of years, I came across the phenomenon known as the Maillard reaction, a chemical reaction much studied by the food industry. As it happened, my mother was a food chemist, so she sent me lots of literature on it. The Maillard reaction occurs when common forms of sugar are heated or persist at less-hot temperatures for a long time. They then form chemical cross-links with amino groups found in proteins and in DNA, resulting in large, tangled molecular complexes. The Maillard reaction occurs in many forms of cooking, and side products of the Maillard reaction result in the pleasant smell and color of freshly baked bread. But most interesting to me was that Maillard products give off a blue fluorescence in UV light. I thought that this might have been what was going on with the Egyptian mummies. I associated this reaction not only with the blue fluorescence of the mummy extracts but also (perhaps incorrectly) with their brown color and their characteristic smell, which is sweet and not unpleasant. And I wondered if the reason I couldn’t extract DNA from them was that the Maillard reaction bonded it to other molecules.
There was a way to find out. In 1996, a paper in Nature described a chemical reagent—N-phenacylthiazolium bromide, abbreviated PTB—that could break down the complexes formed by the Maillard reaction.{42} PTB, when added to baked bread, would turn it into dough again (albeit, surely, not dough that anyone would be tempted to put back in the oven). Since PTB could not be bought commercially, Hendrik synthesized it in the lab. When we added PTB to extracts from ancient samples of cave bears and Neanderthals, it indeed sometimes resulted in
better amplification. And when Hendrik added PTB to extracts from the 20,000-year-old Nevada coprolites, he was able to amplify fragments of the vWF gene that Alex had partially sequenced from mammoths as well as fragments from two other nuclear genes, all to my great surprise. We published this work in July 2003,{43} finally demonstrating that the nuclear genome could be preserved even when remains were not frozen.
Encouraged by those results, I felt that there was now every reason to persist in our attempts to retrieve nuclear DNA from the bones of cave bears—again, with PTB. But sadly, this time the chemical trick didn’t help. Actually, it turned out that the Nevada coprolites were a rare exception where PTB could turn failure into success. The coprolites did, however, confirm my feeling that the nuclear DNA was there, and that we simply needed new techniques to find it.
To get ideas about new techniques I took to consulting as many people as possible about ways to sequence small amounts of DNA. One of the people subjected to my questioning was the Swedish biochemist Mathias Uhlén, a creative inventor and biotech entrepreneur. Mathias combines an apparently unlimited energy and a childlike enthusiasm for new ideas with a knack for collecting creative people around him and transmitting his enthusiasm to them. I always felt energized after my encounters with him. One of the many creative people around Mathias was Pål Nyrén. Ten years earlier, he had conceived of a new technique for DNA sequencing he had developed in spite of widespread skepticism. Mathias had realized the potential of Pål’s idea. He also saw that it was timely to think about new ways to sequence DNA: we were still using the approach invented by Fred Sanger in the UK, which, in 1980, had earned him his second Nobel Prize in chemistry.
