Lone Survivors

by Chris Stringer

  Charles Darwin and his contemporaries had no real knowledge of the mechanisms behind the inheritance of bodily characters, and their predominant ideas were of blending traits between the two parents—and in Darwin’s case, that each cell in the body gave out gemmules, which agglomerated to reconstitute individuals of similar structure in the next generation. As is well known, while Darwin was writing on such matters, the monk and scientist Gregor Mendel was conducting experiments on heredity in Brno (Czech Republic), using peas and bees. He realized that much of inheritance was particulate rather than blended, and that characteristics (often in several alternative states) were inherited following certain rules. Mendel’s work was largely overlooked for another thirty-five years but was rediscovered around 1900, sixteen years after his death, by which time the units of inheritance were known as genes.

  Half a century after that recognition, the structure and role of deoxyribonucleic acid (DNA) in the makeup of genes was discovered, and the modern science of genetics began to take off. It was realized that the ability of DNA to replicate itself resided in its unique twisted ladder of paired bases. The chemical base adenine (A) was always paired with thymine (T) across the strands of DNA, and cytosine (C) with guanine (G). Thus when the ladder splits for replication, each half can form a template from which its whole structure can be re-created. DNA research has become increasingly important for anthropology in studies of the evolution of primates and their present social structures, and for humans in terms of our affinities to the other primates, to population relationships today, and to reconstructions of our evolutionary history.

  Nowadays, our close kinship to the African apes is well established—something that would undoubtedly have pleased Darwin and his close ally Thomas Henry Huxley. But before the impact of genetic studies, it was normal practice for anthropologists to argue that although we were undoubtedly related to the great apes biologically, our special “human” features, such as walking upright, having a large brain, making tools, and speaking, fundamentally set us aside from them. This meant that we were justified in classifying humans as a separate zoological family (the hominids) as distinct from the apes (pongids). Moreover, it was believed that our special features must have taken a very long time to evolve, so many anthropologists favored the idea that our lineage split from that of the apes more than 15 million years ago.

  This view has been swept aside in the last twenty years by a wealth of genetic data that suggest that the chimpanzees (common and bonobo) only differ from us in about 2 percent of their genetic material. The actual figure given varies because experts differ in the way they count up the data; for example, whether considering total DNA sequences, including regions of DNA that do not seem to be functional and can be duplicated many times over, or by comparing sequences that can be shown to be precise equivalents to each other, or by restricting the calculation to “functional” or coding DNA regions. Regardless, the differences are comparable to those found between closely related mammals such as African and Indian elephants, horse and zebra, or jackal and wolf. Such similarity implies that there must be a close evolutionary relationship, and calibration (age estimates) using the fossil record and the genetic distances involved suggests that our line of evolution and that of the chimpanzees may only have separated about 6 million years ago. This view began to gain ascendancy about thirty years ago, following the pioneering work of Allan Wilson and Vince Sarich, who conducted studies using genetic differences in the protein albumin, showing that the Asian orangutan was less closely related to us than were the African great apes. This close relationship is now often recognized by admitting chimps (and less consistently the gorilla as well) into the hominid family, along with us and our immediate extinct relatives.

  Large-scale comparative sweeps of the genomes of humans and chimpanzees show that the vast majority of the 3 billion or so “letters” of our genetic codes are shared, but the rare stretches of distinct DNA are beginning to yield information of great evolutionary interest. Some are clearly related to the various past epidemics to which we and our ape relatives have been exposed, for example, in conferring resistance to retroviruses like HIV, but others can be related to physical changes. For example, a group of 118 bases known as human accelerated region 1 (HAR1) is virtually identical in animals as different as chickens and chimps, with only two coding differences, but humans have accumulated eighteen further mutations. Experiments showed that this DNA sequence is important in building the structure and connections of the cerebral cortex, the wrinkled outermost layer of the brain that is so important for human intelligence (see chapter 8). Many other genes involved in the growth of the brain as a whole, such as ASPM, CDK5RAP2, CENPJ, and MCPH1 (microcephalin), also show accelerated change compared with chimps, and we will return to the last of these—microcephalin—shortly.

  Interestingly, many of the DNA sequence differences that are accentuated between us and our closest living relatives are not concerned with direct changes in, say, the structure of a protein or enzyme. Instead, insertions of what are called transposable elements affect portions of the genetic code by acting as switches in turning functional genes on and off. If the direct products of the DNA can be compared with the ingredients in a recipe, these equally important switches in regulatory genes can be seen as altering the instructions for exactly how the meal is to be cooked, which will produce different results (for example, chimps or humans) from a similar recipe (our DNA coding). Thus the human accelerated region 2 (HAR2 or HACNS1) drives gene activity in building the structure of the wrist and hand bones before birth, and it is likely that these novel DNA changes in humans contribute to our distinctive hands and their greater dexterity, compared with those of chimps and gorillas.

  As well as comparing our DNA with that of our closest living relatives, the chimpanzees, we can infer an increasing amount of information about our evolutionary past from the DNA of living humans, since each of us carries an ancestral record locked up in our genes, something much more detailed than a set of parish records, and one that goes much farther back in time. Because DNA is repeatedly copied, especially when it is passed on from parents to their children, copying mistakes are made, and if the changes are not greatly disadvantageous or lethal, these mutations are then also copied. Thus they can accumulate through time and allow us to follow particular lines of genetic evolution, and to estimate the time involved in their accumulation.

  For our purposes, there are three kinds of DNA that can be studied. The first type is called autosomal DNA. This DNA makes up the chromosomes contained within the nucleus of our body cells, but excluding the special case of the male-related Y-chromosome, which we will come to later in this chapter. It contains the blueprints for most of our bodily structures, and we inherit a combination of it, with our parents making contributions of about 50 percent each. Autosomal DNA also contains many long segments of so-called junk DNA, which do not code for features such as eye color or blood group type. These segments nevertheless get copied, along with the coding DNA, and mutate through time too. Despite their “junk” nickname, some are known to operate as genetic switches, and they can give us valuable information on evolutionary relationships. In fact these sequences are generally more useful in evolutionary and population studies because they may not have been so affected by the distorting consequences of selection, which is strongest on functional DNA—that is, containing genetic code (although junk DNA can be affected when it is structurally linked to functional DNA that is under selection).

  The second type of DNA lies on the Y-chromosome, which determines the male sex in humans. Normal females have twenty-three pairs of chromosomes, including a pair of X-chromosomes, whereas normal males only have twenty-two pairs, plus an X-chromosome (inherited from the mother) and a Y-chromosome (inherited from the father). The DNA on this chromosome can be used to study evolutionary histories in males only, without the complication of inheritance from two parents that comes with the study of autosomal DNA—rather like the continui
ty given by male surnames in many societies.

  The third type of DNA is the now-famous mitochondrial DNA (mtDNA), which is found outside the nucleus of cells and which is inherited through females only. Although this last type of DNA has attracted the most attention in the media and popular science—because it gives such a clear signal of ancestry—analysis of the more extensive autosomal DNA and its products (most of our bodies’ vital constituents such as organs, proteins, enzymes, antigens) has a much longer history in evolutionary studies. For example, a study of ape and human blood proteins led to the first suggestion of a later divergence between humans and African apes, compared to Asian apes.

  As its name suggests, mitochondrial DNA is found in the mitochondria. These little bodies are the power stations of the cells, turning nutrients into usable energy for the cells to do their work. Their DNA is passed on in the egg of the mother when it becomes the first cell of her child, and little or no DNA from the father’s sperm seems to be incorporated at fertilization. This means that mtDNA essentially tracks evolution through females only (mothers to daughters), since a son’s mtDNA will not be passed on to his children. The molecule of mtDNA is shaped in a loop and consists of about 16,000 base pairs. Only some of these are functional—that is, contain genetic code to produce specific proteins such as cytochrome—and the rest of the mtDNA is therefore much more prone to mutation. Thus mtDNA generally changes at a much faster rate than nuclear DNA, making it ideal for studying recent events and short-term evolution. As mentioned in the introduction, prior to the recovery of Neanderthal DNA, the biggest single impact of genetic data on research on human evolution came in 1987, with the publication of Cann, Stoneking, and Wilson’s study of mtDNA variation in modern humans. I described how the work came under heavy attack, especially from disgruntled multiregionalists, but the increasingly detailed analyses carried out since then have shown that the 1987 conclusions were essentially correct, even if they were somewhat overinterpreted.

  Some calculations now place the last common mtDNA ancestor (Eve) at less than 150,000 years old, and it is clear that across the whole human species today, our mtDNA varies far less than is the case in great ape species. This has led to the idea that a recent bottleneck—a drastic drop in population—pruned the variation previously found in the modern human line. However, in line with the alternative nickname for Eve—“lucky mother”—some geneticists have explained that this pattern could have occurred purely by chance if just one woman from those ancient times was lucky enough to have a fertile chain of female offspring through to the present. Thus all other mothers from that time ended up unluckily (in terms of the continuity of their mtDNA) having no surviving children, or just boys, or daughters who failed to provide the necessary ongoing chain of fecund females. From that perspective Eve was not a special female and did not necessarily live in special times, but she gained her unique status retrospectively, through her mtDNA’s good fortune. And we should remember that while this female gave rise to all mtDNA variants known in humans today, many other individuals have contributed their Y and autosomal DNA to succeeding generations. MtDNA is important because it provides such a clear signal of ancestry and descent, but it is effectively inherited as a unit like a single gene, and all our gene variants have their own history, converging back (coalescing) to ancestral genes at various times in the past. Some of those genes have developed very recently, within smaller or larger segments of the modern human population, some go farther back to our common ancestor with the Neanderthals, and some stretch back to our common ancestor with the apes and beyond. There is, too, another potential complication in assessing the evolution of mtDNA. Although many of the distribution patterns we see today appear to be the result of chance, or of historic events such as migrations of females, it is apparent that mtDNA, which does contain some functional genes, can also be subject to the effects of selection.

  The famous mtDNA tree published in 1987.

  The changes in living humans compared to Eve’s reconstructed mtDNA genome sequence average about fifty substitutions, and the different mtDNA types of modern humans have been divided into haplogroups, clusters that share changes in their DNA sequences, and that have descended from female ancestors who first expressed those mutations. The most ancient haplogroup, named L, was the ancestral one derived directly from Eve and is found across the majority of Africa’s populations today. L can in turn be divided into subgroups L0–L3, in terms of their order of branching. The most ancient of these, L0, is found in southern and eastern Africa, with its oldest branches among the Khoisan hunter-gatherers of southern Africa. L1 is mainly found in central and western Africa, including so-called pygmy populations of the central equatorial forest, while L2 is the most common in Africa, at about 25 percent, mainly in the west and southeast. The youngest of these major haplogroups, L3, is common across sub-Saharan Africa, especially in the widespread Bantu-speaking populations, and is thought to have originated in eastern Africa. This makes good sense in evolutionary terms, because this region and L3 were probably the main source of the populations who moved out of Africa and founded the non-African haplogroups M and N, which are found across the rest of the world.

  Of course population movements, particularly in the last millennium, have translocated many lineages far from their places of origin, and a large industry has grown up to help people trace their ancient ancestry through mtDNA. This has proved controversial because your mtDNA ancestry is only a small part of your total genetic ancestry, but along with Y-chromosome DNA for males, mtDNA is very easy to sample, sequence, and track. Yet even when it is tracked back successfully, the results are only as good as the comparative data that are used to “relocate” people (or at least those small bits of their DNA) to their original homelands, and many parts of the world, including Africa, are still poorly sampled for DNA. We do know that African mtDNA contains the most ancient lineages and the greatest diversity for modern humans, consistent with Africa being both our place of origin and the region with the largest ancient population size, which was thus able to conserve that diversity.

  Mitochondrial DNA has been widely used to gauge ancient population sizes, though estimates of these from genetic data are fraught with difficulties, one of which is that calculations generally provide an effective population size—in essence, the size of the breeding population. For mtDNA, this is the estimated size of the pool of “mothers,” while the actual population size (including breeding males and individuals either too young or too old to be involved in breeding) would obviously be much larger. However, many estimates of ancient population size, whether from mtDNA, Y-DNA, X-DNA, or other autosomal DNA, are startlingly low when we consider the billions of humans on Earth today. The long-term effective size of the ancestral population for modern humans might have been only about 10,000 breeding individuals, while the effective size of the female population, judged from surviving mtDNA, is sometimes estimated at less than 5,000!

  If such numbers are a true reflection of the original population size in Africa, humans were present only in numbers comparable to those of gorillas and chimpanzees, species that inhabit relatively small parts of the African continent today. Our ancestors cannot have been widespread across the continent, let alone spread far outside of it, but were probably concentrated in pockets, and those pockets would have been vulnerable to extinction. Using three complete human genomes, the geneticists Chad Huff, Lynn Jorde, and their colleagues made comparisons that reached even deeper back in time to suggest that human population numbers a million years ago (the time of Homo erectus) were somewhat larger, closer to 20,000 breeding individuals, but even this size could hardly have spread across a continent as large as Africa.

  Mitochondrial DNA can also be used to track population growth, and some studies suggest that while haplogroups L0 and L1 grew steadily in their early history, L2 expanded only quite recently, while L3 grew rapidly about 70,000 years ago. In mtDNA terms, as we have seen, the latter group was ancestral to lineages M a
nd N found outside of Africa, so that expansion might well have spilled over into western Asia and hence to the rest of the world.

  MtDNA has been used to calibrate events in human evolution, as we saw from the original calculation of Eve’s antiquity of about 200,000 years, and from the estimate of the expansion of haplogroup L3 at about 70,000 years, but as with population size estimates such calculations are reliant on several assumptions and can only be approximate. For example, most calibrations are based on the assumption that we split from our closest living relatives, chimpanzees, about 6 million years ago. The number of substitutions in our mtDNA compared with that of chimpanzees is then compared with the number of substitutions determined for other events, such as our split from Neanderthals or our exit from Africa. The ratio of substitutions found is then converted into a “date,” along a 6-million-year time scale. However, when substitution rates are determined in very recently diverged human mtDNA, such as in historic populations on islands, or family studies where there are unusual mitochondrial diseases, the rates are much faster than the rate found when comparing our mtDNA with that of chimps. Scientists have argued that “purifying selection” removes many disadvantageous mtDNA mutations through time, thus explaining the rate discrepancy between the short-term and long-term evolutionary events. But when we attempt to calibrate relatively recent events in human evolution, such as the date for Eve or our exit from Africa, should the slow (long-term) rate be used, as it most often is, or should a faster rate be applied?