by Bryan Sykes
I still have a vivid memory of the moment when I first saw the cover of the April 18, 1985, issue of Nature.1 On it, in full color, was the fabulously decorated sarcophagus of an Egyptian mummy and underneath it the caption “Mummy DNA cloned.” Inside, a relatively unknown Swedish scientist, Svante Paabo, claimed to have extracted and cloned human DNA from one of twenty-three Egyptian mummies from museums in Sweden and Germany. The successful extraction was from the skin of a one-year-old child from the Egyptian Museum in Berlin, which was still in Communist East Germany at the time. What was remarkable about this article was not so much the detail of the result itself, but the sheer gall of the attempt. To geneticists like myself, the thought that DNA could survive longer than a few minutes in the open air was unimaginable. Although not strictly the first extraction of ancient DNA (that was published six months earlier, when DNA was recovered from a museum specimen of the quagga, an extinct relative of the zebra),2 the effect of the mummy finding was galvanizing. Few people had ever heard of a quagga, but everyone knew about Egyptian mummies especially after the triumphant touring exhibition The Treasures of Tutankhamun in the 1970s, and again, as Tutenkhamun and the World of the Pharoahs, in 2011.
The visions opened up by the recovery of ancient DNA were completely mesmerizing, and well off the beaten track. In 1985, in my late thirties, I was enjoying my research in medical genetics at a very exciting time, truly a golden age, but, surrounded as I was by well-meaning scientists and physicians, I also had a hankering to do something of no conceivable practical value. I can’t really explain why that was, but ancient DNA certainly fitted the bill and, besides, it took me back to my childhood love of fossils, dinosaurs, and grand museums. Before long I had teamed up with Oxford archaeologist Robert Hedges to see whether we could get any DNA from ancient human bone. There was a genuine connection with my mainstream research at the time, which was investigating the genetic causes of inherited diseases affecting the skeleton, so it did not raise too many eyebrows when chunks of bone started to arrive in the lab. Robert and I reasoned that Egyptian mummies, artificially preserved as they were in the strongly alkaline salt natron, might be a special case. Untreated bones and teeth were far more abundant in archaeological sites, so if this new science was to be genuinely useful and widely applicable it was vital to work out ways of getting DNA from this material. We succeeded after three of years of trying, and published the results in Nature in 1989.3
In the intervening two decades between then and now I think it is fair to say that progress in ancient DNA research has not been entirely smooth. By far the greatest problem has been contamination of the material by minute amounts of modern DNA, usually from excavators, curators, or laboratory staff. DNA does survive for thousands of years but it gets fragmented in the process. DNA that is damaged in this way is much harder to recover than youthful, modern DNA—so much so that even a few molecules of the modern stuff will swamp the damaged original DNA, and you end up with your own or a colleague’s DNA sequence. That’s disappointing enough, but not so bad if you recognize it as such. The real problem comes when you don’t recognize it and interpret the results as being genuinely old.
That happened a great deal in the early days, and I even heard Svante Paabo once say in a seminar that the DNA he got from the Egyptian mummy may well have been his own.4 At first none of us appreciated the extent of the problem. I vividly remember a talk given by a deer biologist who showed a slide of his DNA extraction equipment that he had set up right underneath the head of a white-tailed deer mounted on the wall. No surprise then, that he found deer DNA in the ancient samples he was analyzing. There were some real shockers in the first few years, like dinosaur DNA recovered from a lump of coal and insect DNA from fifty-million-year-old insects trapped in amber. There were certainly many mistaken claims that have never been verified, but they did lead to Michael Crichton’s novel Jurassic Park and the unforgettable 1993 Steven Spielberg movie of the same name. The almost credible basis of this wonderful piece of science fiction was the recovery of dinosaur DNA from the blood in the stomach of a mosquito trapped in amber that had bitten one of the extinct reptiles. I was asked to the London premiere of the film to comment on it for the evening news on TV. Interviewed outside the film theater in Leicester Square after the screening, I emphasized that even if it were possible to recover genuine dinosaur DNA, putting it all back together and creating a living, breathing Tyrannosaurus rex was way beyond current technology.
The next day the lab was invaded by TV news crews who wanted me to repeat the same words for their own bulletins. I eventually got bored with this and arranged for one of my young son’s small model dinosaurs to be dragged on a piece of fishing line across the lab bench behind me, just as I put on my most sober expression and announced that I could reassure the public that no cloning of this kind was taking place anywhere in the world. The film crew saw the rubber creature scuttle across the bench and, to their credit, kept it in. The clip was shown that evening, complete with baby dinosaur, without additional comment. Somewhere in the archives the footage rests to this day. But beneath the exuberance, and carelessness, of the first few years, enough results were being independently verified to show that DNA in bones and teeth really did survive for thousands of years. Even if not completely intact, it could be sufficiently undamaged to be extracted and analyzed.
Techniques have improved a lot in the last few years, and early in 2010, the same Svante Paabo, now director of the Max Planck Institute for Evolutionary Anthropology in Leipzig, published the draft genome sequence of a forty-thousand-year-old Neanderthal.5 Three months earlier a research group from the University of Copenhagen announced the almost complete genome sequence of a four-thousand-year-old Paleo-Eskimo from hair preserved in the Greenland permafrost.6 As a reflection of the scale of the endeavor needed to do this work, there are fifty-six authors on the Neanderthal paper and fifty-two on the Greenland article. On the original Egyptian mummy paper there was just one author, Svante Paabo himself, while on our 1989 Nature article there were only three. Things are now definitely looking up for ancient DNA after a shaky start, but while some technical aspects have definitely improved, others have stayed stubbornly the same.
The press coverage of ancient DNA can always be counted on to be enthusiastic, if not positively melodramatic. The first recovery of DNA was greeted by a memorable headline that gives us a flavor: “US Scientists Clone Dinosaurs to Fight On after Nuclear War,”7 while the latest announcements from Leipzig once again resurrected the notion that Neanderthals could be cloned.8
While complete genomes, which include all of our twenty-three thousand or so genes, are almost routinely sequenced these days, the early work on genetic reconstructions of the past used a much simpler and rather unusual piece of DNA. This is mitochondrial DNA, or mDNA for short. Unlike most of our DNA, which is confined to the cell nucleus, mDNA is found inside small particles, mitochondria, which live in the fluid, called the cytoplasm, that surrounds the cell nucleus but is still contained within the cell membrane. Mitochondria are thus an integral part of every cell, and their job is to convert glucose into energy using oxygen, leading to their sobriquet as “the powerhouses of the cell.” Mitochondria were once free-living bacteria that, hundreds of millions of years ago, were kidnapped and put to work by cells. They have lived there ever since, becoming over time completely dependent on their captors. Mitochondria have their own DNA, which, as a legacy of their bacterial origins, is circular rather than linear like our nuclear DNA.
By a great stroke of luck mDNA has several special qualities that are almost perfectly suited for exploring human evolution. For a start, there is lots of it in a cell, far more than its counterpart in the cell nucleus. For every nuclear gene there are hundreds or even thousands more copies in the cell’s mitochondria. That made mDNA the natural choice for ancient DNA work, and it was for this reason that we were first introduced to each other. There was always going to be precious little DNA left in an ancient sample
, so it made sense to go for mDNA to have the best chance of success. Even these days, when techniques for recovering DNA are far better than they were, there are still plenty of ancient samples where mDNA is the only DNA you can get out. I always check that I am able to extract mDNA from an ancient sample before embarking on recovering other genes.
While its cellular abundance was reason enough for me to choose mDNA for ancient work, it also comes with two other great advantages. The first of these is that mDNA changes far faster over time than the cell’s other DNA in the nucleus. However, the fast rate of mutational change in mDNA is only relative and, to most of us, appears mind-numbingly sluggish. On average mDNA changes at the rate of one mutation every twenty thousand years. Though that is undeniably slow by everyday standards, it is still twenty times faster than the mutation rate of nuclear DNA. Very conveniently, as we shall see, the mDNA mutation rate fits in very well with the timescale of human evolution. The second great advantage of mDNA is that, unlike most nuclear DNA, it is inherited down only one line of ancestors. While human eggs are crammed with a hundred thousand mitochondria, sperm have hardly any, and none at all that survive past fertilization. This means that the mitochondria in a newborn baby have come exclusively from his or her mother. This fact makes the inheritance pattern of mDNA simplicity itself. Every man, woman, and child’s mDNA is inherited from his or her mother, who gets it from her mother, who inherits it from her mother, and so on back in time. The direct matrilineal line stretches back hundreds, thousands, tens of thousands of years. All that time the mDNA stays exactly the same, save for any changes introduced by mutation once in every twenty thousand years or so. There is another tremendous advantage of mDNA when it comes to looking at ancestral origins, and especially those of indigenous people, but we will look into that a little later.
Mitochondrial DNA, like its counterpart in the nucleus, is made up of a series of four slightly different chemicals, called nucleotide bases, strung together in a particular sequence. Four different-colored beads on a string is a useful and workable metaphor here. It is the order of these colored beads that is all-important. However, unlike necklaces that might stretch to a couple of hundred beads, DNA is immensely long. Even mitochondrial DNA is extensive by familiar standards. Its circle of DNA contains more than sixteen thousand of these metaphorical beads, but this is minuscule compared to the entire human genome, which is made up of three thousand million of them. I will explain more about how these bases direct the intricate working of our cells later, but for now the important detail is their precise sequence, which we can read using a variety of techniques. For mDNA we are concerned only with the sequence of its sixteen thousand or so bases and, most important, the fact that this sequence can and does change through the process of mutation. In fact we can narrow down our interest to a stretch of four hundred bases that change, or mutate, more rapidly than the rest because mutations in this segment, called the control region, have no biological consequences. Changes outside the control region can interfere with the efficient working of the mitochondria in such a way that the mutated mDNA is eliminated and not passed on to future generations.
If two people have the same mDNA sequence, then they must have inherited it from a shared maternal ancestor. It is this simple concept that underlies the way mDNA is used to reconstruct the past. Even when the two people concerned live in completely different parts of the world, if their mDNA sequences match, then there must be an ancestral connection between them, even if they are not aware of it. Imagine your own line of maternal ancestors going back through your mother, her mother, and so on back into the past. Now imagine someone else who has a matching mDNA sequence. If you were able to trace his or her maternal ancestral line and your own far enough back in time, they would meet in one woman, your shared maternal ancestor. There is no need for paper records to confirm this ancestral connection. It is there in your DNA.
In this scenario, where two people have matching mDNA sequences, their next question is going to be this: How long ago did the woman, from whom we are both descended through the maternal line, actually live? This question introduces another key element, that of time. While we might occasionally be able to give a precise answer, through establishing a rigorous genealogical connection back to a common ancestor, most of the time this will be impossible. Nevertheless we can arrive at some sort of answer by using the fact that, although mDNA is very stable, it does change slowly over time through the process of mutation.
In the control region these mutations happen when mDNA is copied as cells and their mitochondria divide, which they are doing all the time. Returning to the bead necklace analogy, just occasionally the copying is inexact and a different-colored bead is substituted in the copy. This is a mutation, and as far as we know it is an entirely random event. The only mutations that concern us here are those that occur in the female germ line—that is, in the cells which become eggs. They are the only ones that get passed on to the next generation. All the other mDNA mutations that happen in our other body cells, like muscle and blood, are of no consequence as they will not be passed on and will disappear when we die. In men even germ-line mDNA mutations are irrelevant, since the few that get into sperm cells are destroyed in the fertilized egg and never make it to the next generation.
From a variety of measures, we have a pretty good idea of the mDNA mutation rate, and it is about one every twenty thousand years. If the average generation time for women over the course of human evolution was twenty years, which is not unreasonable, then this is the equivalent of roughly a thousand generations. In other words, in one of every thousand female births, the mDNA of the child will differ by one mutation from the mDNA of her mother. Notice that I haven’t counted the changes between mothers and sons because, although they will also occur, they don’t matter because the son’s mDNA is not going anywhere.
To illustrate how we can use this for genetic dating, let us imagine two people who have mDNA sequences that do not precisely match but differ by one mutation. In fact we can use an actual example. The mDNA of the last Russian czar, Nicholas II, which was recovered from his remains after they were excavated in 1991, differs from mine at only one position. Our shared maternal ancestor must therefore have lived sufficiently long ago for one mutation to have accumulated in one of the two lines of descent leading from that shared ancestor to the czar and myself. Using the average mDNA mutation rate, we can assume that the czar and I are separated by a total of a thousand generations from our common maternal ancestor. That is five hundred generations along one line and five hundred along the other. So, by this reckoning, the woman from whom we are both maternally descended lived five hundred generations ago. If we take twenty years as the generation time, then she lived ten thousand years ago. I have to emphasize here that this is an average figure. Our common maternal ancestor may have lived more recently, or longer ago than a thousand generations. This is all down to the random nature of mutation and makes genetic dating of this sort very approximate for individual cases. In this respect it is much less exact than the carbon dating that we covered earlier. But it can be very useful when large numbers of people are compared. And that is how mDNA first came to be used to explore the way in which humans have moved around the planet over the past quarter of a million years.
Regardless of where they were working in the world, as soon as scientists began to look at their results from mDNA analysis they noticed that, far from being uniformly spread, individual DNA sequences tended to fall into distinct clusters. Within the clusters the sequences were clearly related to one another. To give you an example, one of the clusters found among Native Americans, as we shall see shortly, features a mutation at position 16,111. The number 16,111 refers to its place in the sequence around the mDNA circle, starting at an agreed point. The mDNA control region that concerns us here extends between positions 16,000 and 16,400, thus I will drop the 16,000 prefix from now on so that 16,111 becomes 111, and so forth. The mutations are changes relative to an inte
rnationally agreed reference sequence, which is that of the very first mDNA to be sequenced in 1981. The DNA base at 111 in the Native Americans is different from the base at position 111 in the reference sequence. In fact no Native American mDNA so far sequenced has only the mutation at 111, but many people have 111 plus one or more other changes. Likewise in Europe, many people share a mutation at position 224 in their mDNA when compared to the reference sequence. In Polynesia almost everyone has a mutation at position 247, on so on.
In the early years of my research with mDNA, I got to know these numbers very well indeed, and it was an exciting time trying to make sense of the different genetic clusters that they defined. In Europe, and among Americans with European roots on their mothers’ side, there are seven such clusters and I eventually realized that this had to mean that each cluster has just one matrilineal ancestor. From this it followed by an irrefutable logic that there were only seven women, seven “clan mothers,” from whom almost all native Europeans are maternally descended. In time these women became the heroines of my first book, The Seven Daughters of Eve.
3
The First Americans
Leaping Coho salmon.
Until the invention of agriculture, which over the last ten thousand years has happened at various times in different parts of the world, our ancestors were on the move most of the time. They needed to be, as their main food supply derived from the herds of game that migrated with the seasons. Bison, mammoth, and wild horse were all on the menu, and on the move, and our ancestors had no choice but to follow. They scavenged the carcasses from the kills of lion and hyena, or killed their own by ambush or by stampeding the herds over cliffs. If there were encampments, they were transient and seasonal. Only in a very few parts of the world were there anything resembling permanent settlements, and one of these was the Pacific Northwest of America. The reason for this luxury of permanence, from which all sorts of benefits flowed, is summed up in one word: salmon.
