by Pbo, Svante
By the mid-1990s, the field of ancient DNA studies had somewhat stabilized. Many researchers had come to realize what was possible and what was not. Zoological collections of skins and other parts of animals that had been dried shortly after death could be used routinely for DNA extraction, as we had shown by our investigations of kangaroo rats, back in Berkeley. Studies on pocket gophers, rabbits, and many other animals followed, and several of the big zoological museums established molecular labs in the 1990s devoted to the DNA study both of their old collections and of new samples collected specifically for that purpose. The Smithsonian Institution in Washington, DC, and the Natural History Museum in London were among the first to do this, and they were followed by others. Similarly, forensic scientists analyzed DNA that they could now extract and amplify from evidence collected many years earlier. This led to courts overturning the convictions of wrongfully imprisoned people and to new strides, based on genetic evidence, in identifying remains and apprehending criminals. The despondency of my first years in Munich, when I and my group had struggled with contamination and other methodological issues while others published nonsensical millions-of-years-old DNA sequences in Science and Nature, was now replaced by a sense of satisfaction that it had all been worthwhile. The field had become established. It was time to return to the old challenge: human remains.
As noted, there are plenty of ways that modern human DNA can contaminate an experiment. The curator in London had demonstrated one very obvious way to me when he put his tongue to the sloth bone, but dust, bad reagents, and much else also pose problems. Human history was, for me, the ultimate goal. The question was whether, despite the impediments, we could find a way forward.
Oliva Handt devoted herself to this quest. Oliva is a warm, almost motherly person who tended to be overcritical of her own work. I felt that this was an excellent trait for the job she was about to undertake. She had to deal with the same issues as Matthias did in his sloth work, but in addition she had to worry about the odd dust particle that might land in a test tube where the extract of an ancient human bone was kept. If no such dust particle had landed in the blank extract performed in parallel with the bone extract, then it might be hard or impossible to tell whether or not the sequence she determined was from the bone or from a contaminating dust particle. For this reason, we decided that Oliva would work on Native American remains, whose mtDNA contained certain variants not found in Europeans. Although I much disliked doing experiments where only results in accordance with pre-expectations would be credible, this seemed to be one of the few ways in which we could reliably work out the methods for retrieving ancient human DNA sequences. So Oliva started working on skeletons and mummified human remains from the American Southwest that were around 600 years old. As she was toiling away on this, repeating her extractions over and over to test the reproducibility of her results, there came an opportunity that was too good to pass up.
In September 1991, two German hikers had found the mummified body of a man in the Alps in the Oetztal near Hauslabjoch, on the border between Austria and Italy. They, and the authorities they contacted, first believed it to be a modern corpse, perhaps a war victim or an unfortunate hiker who had been lost in a snowstorm. But after the man’s body was removed from the ice, what was left of his clothes and equipment made it plain that he was not a recent soldier or hiker; rather, he died on the Alpine pass about 5,300 years earlier, during the Copper Age. From the news media, I learned that the Austrian and Italian governments were each claiming that the mummy had been found on their territory. There were also disputes between the discoverers and government officials about finders’ fees, and difficulties with the people in the pathology department of the University of Innsbruck, in Austria, who kept the frozen corpse and jealously guarded him against outsiders. In short, it seemed to be a legal and general mess. I was therefore surprised when, in 1993, I was approached by a professor from Innsbruck who asked if we wanted to analyze the DNA of the Ice Man—or Oetzi, as he had come to be called, after the valley in which he was found. We expected that a body that had been frozen continuously for more than 5,000 years would be much better preserved than any mummies from Egypt or bones from North America. We decided to give it a try.
Oliva and I traveled to Innsbruck, where the pathologists removed eight small samples from Oetzi’s left hip, which had been damaged when the body (as yet unrecognized as being ancient and unique) was freed from the Alpine ice by means of a sledgehammer. Back in Munich, Oliva set about extracting and amplifying mtDNA. We were all excited when she obtained nice PCR products, but when she sequenced them, the sequences were uninterpretable. At many positions, there seemed to be several different nucleotides present. To sort this out, she went back to the old cloning approach I had used in Uppsala: she cloned each PCR product and then sequenced several clones. Since each clone came from a single original DNA fragment that had been amplified by the PCR, she could see whether all the original DNA fragments carried the same nucleotide sequence as would be expected if they came from a single individual or, alternatively, whether they carried different sequences and therefore came from different individuals. The latter turned out to be the case, and in fact different samples gave different mixtures of sequences. This was maddening. Most, if not all, of the mtDNA must have come from people who had handled the Ice Man since he was discovered. How were we to determine whether a sequence came from the Ice Man or not? After all, in evolutionary terms, he had not lived all that long ago, so he would undoubtedly have a version of mtDNA similar or identical to those found in Europeans today, and many Europeans had apparently come into contact with him since his discovery.
Fortunately, two of the samples we had obtained in Innsbruck were large enough for us to remove the surface tissue and extract material from the inside, untouched part of the sample, in hopes that whatever contamination was present would primarily be on the surface. This helped, but only to a degree. Oliva found six positions where she saw a mixture of sequences suggesting that the number of different mtDNA variants was smaller, coming from perhaps three or four individuals. But the sequences did not neatly group into three or four classes of identical sequences. Oliva found that the variants at the six positions were scrambled among the molecules, especially when she looked at positions far removed from one another. This must be the result of the “jumping PCR” I had described in Berkeley, whereby instead of copying a single continuous piece of DNA, the polymerase stitches fragments of DNA together into new combinations. Could this scrambled mixture of DNA sequences be disentangled so that we could decide which, if any, of these sequences was from the Ice Man?
We argued that the jumping phenomenon should occur primarily in attempts to amplify longer pieces of DNA rather than shorter ones because shorter pieces would be more likely to be preserved in an intact form in the tissues while longer ones were likelier to be stitched-together molecules or contaminants. So Oliva did a PCR of extremely short pieces. This helped. Whenever she amplified pieces shorter than about 150 nucleotides, not only did she not get scrambled sequences but almost all of her clones carried the same sequence. The picture was becoming clearer. Our extracts contained one mtDNA sequence that was there in large quantities but degraded into short pieces. They also contained mtDNA sequences from two or more additional people that were less frequent and present in larger pieces. We suggested that the abundant and more degraded DNA was likely to have come from the Ice Man, whereas the other DNAs, which were less abundant but also less degraded, were likely to have come from modern individuals who had contaminated the Ice Man.
By amplifying each of the short pieces at least twice, cloning them, and sequencing several clones from each amplification product, Oliva could eventually reconstruct the mtDNA sequence that the Ice Man was likely to have carried when alive. The overlapping fragments she generated determined a sequence of a bit more than 300 nucleotides. Only two substitutions distinguished it from a commonly used reference sequence for modern European mtDNA, a
nd the identical sequence is not uncommon in Europe today. This was not terribly unexpected. From the perspective of someone hoping to live for 80 or 90 years, 5,300 years is a long time, representing about 250 generations. From an evolutionary perspective, however, this is a short time. Unless major disasters, such as epidemics, kill much of a population or major population replacements occur, not much would change in our genes in 250 generations. Indeed, my lab mates and I had anticipated that at most one mutation would have occurred in the segment we studied since the Copper Age.
But before we could publish our results, we faced one additional hurdle: our rule, borne out of frustration with the many unreliable results published in the field, that important or unexpected results should be replicated in a second laboratory. The sequence we had determined from the Ice Man was not biologically unexpected but it certainly would attract attention, so this was an opportunity to show how things should be done. We decided to send one of our unused tissue samples to Oxford, where Bryan Sykes, a geneticist who had left his earlier career in connective-tissue diseases to work on mtDNA variation in humans and in ancient DNA, was eager to help. Sykes’s student extracted and amplified a piece of the sequence we had determined and reported the sequence back to us. It was identical to Oliva’s, and we described our findings in a paper in Science.{32}{33} Although regarded as a success at the time, to my mind this experience primarily showed how difficult it was to work with ancient human remains. The Ice Man had been frozen and could be expected to be unusually well preserved, and moreover had been found only two years earlier, so that not too many people had had the chance to contaminate it, yet we had found a mixture of different sequences that had been difficult to sort out. We had succeeded only thanks to Oliva’s patience and perseverance and our inference of what was likely to be the correct sequence, which necessarily relied on assumptions about the different populations of molecules in the tissue. Any study of recent human evolution, where one would need to study populations of many individuals, probably all preserved as skeletons, seemed too daunting to contemplate.
On the bright side, we had gained an enormous amount of experience with human sample materials and a better appreciation of the difficulties involved. To profit from this, Oliva returned to the Native American remains. As we expected, it wasn’t easy. My friend Ryk Ward arranged for us to get ten samples of mummies that were about 600 years old from Arizona in the American Southwest. As one might imagine, the outcome was comparable to that of the Ice Man analysis. From nine of the individuals, Oliva either could not amplify anything at all or found sequences so mixed up that it was impossible for her to determine whether any of them was endogenous to the individual. In only one case was she able to do short amplifications, and by sequencing many clones from repeated amplifications she showed that this sample contained relatively many molecules and that they came from an mtDNA that resembled mtDNA sequences found among modern Native Americans. Somewhat frustrated, we wrote, in the summary of a 1996 paper describing Oliva’s work, that “these results show that more experimental work than is often applied is necessary to ensure that DNA sequences amplified from ancient human remains are authentic.”{34} This, obviously, was also an implicit criticism of much of the work that was being performed by others on ancient human remains.
Despite all of Oliva’s efforts, I now decided to abandon all work on ancient human remains. Other laboratories continued to publish results, but I felt that much of what appeared was unreliable. The situation was deeply frustrating.{35} In 1986, I had left what looked like the start of a promising career in medical research because I wanted to introduce new and accurate methods of studying human history in Egypt and elsewhere. By 1996, I had been able to establish reliable methods that turned zoological museums into veritable gene banks and made it possible to study mammoths, ground sloths, ancestral horses, and other animals from the last Ice Age. This was all well and good, but it was not where my heart was, and I worried that I would turn into a zoologist against my will.
I did not torture myself daily with these thoughts, but again and again, when I reflected on what I might do in the future, I felt frustrated. What I wanted to do was to illuminate human history, but it seemed next to impossible to study ancient humans because, in most cases, their DNA could not be distinguished from that of living people. But after a while I realized that I could perhaps do something of even greater relevance to understanding human history than the study of DNA from Bronze Age people or Egyptian mummies. Perhaps I could study people of another sort, people who had been in Europe long before the Ice Man—the Neanderthals.
Turning to Neanderthals might seem strange, given that I had just sworn off ancient humans. But of crucial importance for me was that they could be expected to have DNA sequences that were recognizably different from present-day humans. This was not just because they lived more than 30,000 years ago but also because they’d had a long history different from ours. Some paleontologists estimated that we shared a common ancestor with them at least 300,000 years ago, and some said they were a different species. Anatomically, the Neanderthals looked strikingly different from present-day humans and also from early modern humans living elsewhere in Europe at approximately the same times. Yet Neanderthals are the closest evolutionary relatives of all contemporary humans. Studying how we differed genetically from our closest relatives would potentially allow us to find out what changes set apart the ancestors of present-day humans from all other organisms on the planet. In essence, we would be studying perhaps the most fundamental part of human history—the biological origin of fully modern humans, the direct ancestors of all people alive today. Such research might also tell us exactly how Neanderthals were related to us. Neanderthal DNA seemed like the coolest thing imaginable to me. And by sheer luck I was in Germany, where Neander Valley is situated, where the first Neanderthal had been found, the so-called type specimen used to define Neanderthals. I wanted desperately to get in touch with the museum in Bonn where the Neanderthal type specimen was housed. I had no idea whether its curators would be reluctant to give me a sample. This type specimen was, after all, what some (perhaps in an attempt to forget certain aspects of twentieth-century German history) called “the most well-known German.” It was something of an unofficial national treasure.
I worried for months over how to go about this. I knew all too well how tricky it could be to work with museum curators, who were entrusted with the difficult task of preserving valuable specimens for future generations while at the same time facilitating research. I had found, in some instances, that they considered their primary role to be the exertion of power, refusing access to specimens even when the possible gain of knowledge seemed to far outweigh the value of preserving a small piece of a bone. If such curators were approached in the wrong way, they could say no and then, for all-too-familiar reasons of human pride, find it difficult to go back on their word. While fretting about this, I received one day, in the most remarkable of convergences, a phone call from Bonn. The caller was Ralf Schmitz, a young archaeologist who, together with the Bonn museum curator, was responsible for the Neanderthal type specimen. He asked if I remembered an exchange we had had a few years back.
He reminded me that in 1992 he had asked what the chances of success were if one were to try to obtain DNA from a Neanderthal. This conversation had slipped out of my mind, being one out of many with archaeologists and museum curators. Now I remembered. At the time I had not known what to answer. My immediate, and slightly delinquent, impulse had been to suggest that the chances were good, so that they might readily part with a Neanderthal bone. But almost as quickly I had realized that honesty was the best way forward. After some hesitation, I had said that in my opinion we might have a 5 percent chance of success. Ralf had thanked me and I had heard nothing more from him since.
Now, almost four years later, Ralf was on the phone and said that, yes, we would be allowed to have a piece of the Neanderthal from Neander Valley. As it turned out (Ralf later tol
d me), others had approached the museum to request samples, saying they were almost certain to get usable DNA out of the specimen. The museum authorities had then prudently decided to get the opinion of another lab and had asked Ralf to contact me. Not only our track record but also our apparent honesty in suggesting that the chances were slim had convinced Ralf and the museum that we would be their best partners. They were, as it turned out, the exact opposite of the obstructionist museum curators I had been fretting about. I was delighted.
What followed were weeks of discussions with the museum about how much bone material we would get, and from which part of the skeleton. In total, there was about half of a skeleton of what seemed to be a male individual. Our experience had shown us that the best chances of success were with compact bone—for example, a part of the shaft of an arm or leg bone, or the root of a tooth, rather than thin bones with a large marrow cavity, such as ribs. Eventually we agreed on a piece of the right upper arm, from a part where the shaft had no ridges or other features of interest to paleontologists, who study how muscles had attached to the bone. It also became clear that we would not be allowed to remove the sample ourselves. Ralf and a colleague came to see us in Munich, and we gave them a sterile saw, protective clothing, sterile gloves, and containers in which to store the sample—and off they went. In the end, it was probably fortunate that I was not allowed to put the saw to the archetypal Neanderthal myself. I would probably have been too intimidated by this iconic fossil and would have cut off a very small piece, perhaps too small for success. When we received the sample, we were impressed by the size of what they had removed—3.5 grams of what looked like very well-preserved whitish bone (see Figure 5.3). Ralf reported that when they sawed through the bone, a distinct smell of burnt bone spread through the room. This, we believed, was a good sign; it had to mean that collagen, the protein that makes up the matrix of bone, had been preserved. It was with awe and trepidation that I approached my graduate student Matthias Krings, who had spent more than a year on fruitless attempts to extract DNA from Egyptian mummies—with the plastic bags containing the piece of the Neanderthal type specimen and asked him to apply our latest and best methods to it.
