References

Recent advances in DNA technology have added a new dimension to the search for our origins. In addition to studying the bones, teeth and artifacts of prehistoric people, archeologists can now also analyze the DNA of past humans. With a technique known as the polymerase chain reaction, or PCR, scientists can extract a tiny scrap of DNA from a fossil bone and make millions of copies of it. That provides enough material to enable one to compare the nucleotide sequences of ancient human DNA with comparable sequences in living humans.

The bones of several humans from the Paleolithic era, predating the end of the last Ice Age, have now yielded DNA sequences. These include some Neanderthals, one early modern Australian contemporaneous with Neanderthals, and some modern humans who date to the Paleolithic but who lived after Neanderthals had vanished.

	Cave-bear remains such as this skull occur at many European sites, often in Neanderthal caves. A cave-bear bone from one such cave in Croatia yielded a nuclear DNA sequence, raising hopes we may one day sequence Neanderthal nuclear DNA. (Courtesy of Vojtech Turek, Palaeontological Department, National Museum, Prague, Czech Republic. Skull is from the Palaeontological Collection of the National Museum.)

The early modern Australian may date back 62,000 years before the present (B.P.), making him the oldest human to yield a DNA sequence. The most ancient Neanderthal to furnish a DNA sequence, from a site in Croatia, dates to over 42,000 B.P.

So far, scientists have only been able to analyze sequences of mitochondrial DNA (mtDNA) from human bones of Paleolithic age (O'Rourke, et al, 2000). Nuclear DNA has been taken from human remains this ancient, but no one has yet been able to reconstruct human nuclear DNA sequences this old. Hope for such sequences exists, however, since cave-bear remains some 33,000 years old have yielded a nuclear DNA sequence (Greenwood, et al, 1999). It is easier to study mtDNA from ancient bones, because many mitochondria, and therefore many copies of the mtDNA molecule, exist in a cell, compared to only a single copy of the nuclear genome.

One problem with mtDNA, however, is that it functions as a single genetic locus, inherited as a single gene, compared to the multitude of genes present in nuclear DNA. Studies of single loci are more likely to be affected by random genetic fluctuations than studies that examine numerous loci. They are therefore less likely to accurately render the overall genetic patterns that characterize a population. Another problem with mtDNA is its strictly maternal inheritance. Because of marriage customs governing the transfer of women between groups, genetic patterns drawn from mtDNA studies can differ from those that emerge from nuclear DNA analyses. These factors could yield misleading results in ancient mtDNA studies (Williams, et al, 2002).

Ancient mtDNA sequences taken from the bones of Neanderthals and Paleolithic-era modern humans have all come from a small portion of the mtDNA molecule, just over 1,100 base pairs long, called the control region. This region, which does not code for any proteins, is sometimes also called the displacement loop, or D-loop, after a loop-like structure that occurs within it. Researchers have found that the control-region sequences of Paleolithic modern humans who lived after the Neanderthals vanished are similar to those of living people. The control-region sequences of Neanderthals, however, differ from those of living humans.

A researcher works on a limb bone from Germany's famous Feldhofer Neanderthal. This Neanderthal's mtDNA may support excluding Neanderthals from our ancestry. (Courtesy of the Rheinisches Landesmuseum, Bonn, Germany.)

The sequence taken from the early modern Australian who lived during the Neanderthal era also diverges from our own, though not to the same degree as the Neanderthal sequences.

The first Neanderthal bone to yield a mtDNA sequence for analysis was a humerus, or upper arm bone, from the original Neanderthal type specimen found in the Feldhofer Cave in Germany's Neander Valley. This individual, as well as a second Neanderthal from the site whose mtDNA has also been studied but not yet described, both date to just under 40,000 B.P. (Schmitz, et al, 2002). Researchers examined two different control-region sequences from the original Feldhofer Neanderthal and compared them with the same sequences in hundreds of living humans from around the world. They discovered large differences between the Neanderthal's mtDNA and our own (Krings, et al, 1997, 1999).

In their most recent study, the scientists found an average of 35.3 differences between the Neanderthal's sequences and those of living humans, compared to only 10.9 differences between present-day sequences, casting doubt on the view that Neanderthals contributed to our mitochondrial gene pool. They concluded that the mtDNA lineage of Neanderthals branched off prior to the common ancestor of all present-day human lineages, probably around 465,000 B.P. Still, the Feldhofer Neanderthal was not necessarily from a separate species, because the difference between its mtDNA and ours was less than that between certain chimpanzee subspecies. Also, mtDNA differences between certain present-day individuals are actually greater than some of those between the Feldhofer Neanderthal and living people (Wolpoff, 1999: 759).

If Neanderthals were among our ancestors, one might expect them to have been most closely related to living Europeans, who now occupy their territory. But the Feldhofer Neanderthal's mtDNA was no closer to that of living Europeans than to the mtDNA of living people from any other region. This is also true for all other Neanderthals whose mtDNA has been studied. Some scholars think, however, that over time, with sufficient migration and gene flow, any Neanderthal DNA sequences that might have lasted to the present would have become evenly distributed in low frequencies across all regions (Relethford, 2001).

The Feldhofer results find support from an analysis of mtDNA from a Neanderthal bone, radiocarbon dated to more than 42,000 B.P., found in Middle Paleolithic level G3 at the Vindija cave site in Croatia (Krings, et al, 2000). Only 9 differences occur between the mtDNA sequences of the Vindija and Feldhofer Neanderthals. But the mtDNA of both these Neanderthals show roughly 35 differences from the same sequences in living people.

Further support for the distinctiveness of Neanderthal mtDNA may come from mtDNA taken from the rib of an infant, identified as a Neanderthal, buried 29,000 years ago in the Mezmaiskaya Cave in the northern Caucasus region of Russia (see Central and Eastern Europe web page). The mtDNA sequence of the Mezmaiskaya infant shows only 12 differences from the same sequence in the Feldhofer Neanderthal, but presents 22 sequence differences from the mtDNA of present-day humans (Ovchinnikov, et al, 2000).

	The rate at which genetic differences have accumulated between chimpanzees and humans is often employed as a molecular clock. Researchers use that clock to estimate when Neanderthals might have diverged from the main human line. (Photo by Curt Busse. Provided by Curt Busse's Gombe Chimpanzee Photo Gallery, under the terms of the Design Science License.)

The scientists studying this infant's mtDNA estimate that the mtDNA lineages of Neanderthals and modern humans diverged early, between 365,000 and 853,000 B.P.

These divergence times are calculated based on a presumed rate of change in the mtDNA control region over time, a so-called "molecular clock." The rate of this clock is derived from the differences that have accumulated between the control-region sequences of humans and chimpanzees since the human and chimpanzee lineages diverged. Neanderthal mtDNA researchers have assumed these lines split no earlier than 5 million years ago, but recent fossil discoveries have now pushed the date back to before 6 million years ago (Brunet, et al, 2002). As a result, molecular-clock estimates of when Neanderthals branched off from the ancestors of living humans would seemingly need to be pushed back in time as well.

In truth, however, it may not be possible to accurately track the pattern and pace of human evolution with molecular clocks. One problem is that when comparing lineages within a species, rather than looking at two different species, any interbreeding between those lineages will disrupt the clock. Neanderthal mtDNA researchers have tended to assume that Neanderthals and early modern humans did not interbreed, but it is quite possible they did. Even without interbreeding, genetic differences between lineages often do not accumulate in clock-like way.

The control region especially falls short as a molecular clock because some of its nucleotide sites mutate slowly, while others, often termed "hotspots," change rapidly (Stoneking, 2000). Some of these hotspots generate "somatic" mutations that last no longer than a lifetime. These mutations, which are rather common in the aged (Nekhaeva, et al, 2002), arise in the mitochondria of various bodily tissues during an individual's life and therefore do not pass to the next generation. Scientists have tried to compensate for the differing mutation rates within the control region in various ways (Heyer, et al, 2001), but these various methods have yielded divergent results. Doubts about the rate of change in the control region have in fact led some scientists to stop using control-region sequences to reconstruct human genetic history (Ingman, et al, 2000: 708).

Another problem is that researchers come up with different mutation rates for the control region depending on whether they derive the rate indirectly from the present-day genetic differences between two lineages or calculate it more directly from ancient mtDNA. The more direct method, which has not yet been applied to human data, compares sequences from living individuals with those taken from the skeletal remains of individuals that lived at various times in the past. An initial test of this method using samples from ancient and living penguins found a mutation rate for the control region 2 to 7 times faster than what had previously been assumed from indirect methods (Lambert, et al, 2002).

Even if scientists could agree on a mutation rate for the control region, that rate would not be the only factor influencing how control-region sequences have changed over time. Natural selection can alter the rate of change at a genetic locus by affecting the rate at which mutations are either eliminated from, or spread within, the gene pool. Demographic factors such as bottlenecks or expansions in population size can exert similar effects. The control region does indeed appear to have been affected by such forces over the course of human evolution (Jorde, et al, 1995).

If natural selection has affected the control region, then Neanderthals could have contributed to our ancestry, either by interbreeding with early modern humans or by evolving directly into modern people, and yet still have had mtDNA sequences quite unlike our own. Some time in the past, selection for a favorable mitochondrial genotype may have caused that

MtDNA from this Ice-Age skeleton from Wales suggests that modern humans living just after Neanderthals had vanished were already genetically like us. (Courtesy of the Oxford University Museum of Natural History, Oxford, England.)

genotype to spread across the globe, eliminating much of the earlier mtDNA diversity (Adcock, et al, 2001a: 541). MtDNA sequences from Neanderthal remains predating that change would differ from ours, even if Neanderthals were among our ancestors. Our mtDNA would also differ from that of any modern humans who lived prior to this selective event.

It may also be that Neanderthals were among our ancestors but were too unsuccessful reproductively to pass on many of their genes. Rigors of the last Ice Age may at times have caused Neanderthal groups to reproduce at such low rates that they could only keep up adequate breeding sizes by mating with immigrants from more fecund regions (Enflo, et al, 2001). The genes of these incoming, probably anatomically modern, people would soon have supplanted most, if not all, of the distinctly Neanderthal genes in the population.

This type of hybridization, in which a reproductively successful species or subspecies interbreeds with a closely related but less prolific species until the less populous species no longer exists in its original form, is increasingly being recognized as a cause of extinctions (Levin, 2002). It is therefore quite possible that Neanderthals disappeared by interbreeding with early modern humans, rather than by being outcompeted or exterminated by them.

To better assess the extent to which Neanderthals might have contributed to our ancestry, we need to compare their mtDNA to that of early modern humans who coexisted with them. Control-region sequences taken from modern human remains of late Paleolithic age from England and Australia show that prehistoric modern humans had mtDNA sequences matching those of living humans by at least 12,500 B.P. and perhaps as early as 15,000 B.P. (Adcock, et al, 2001a; Sykes, 2000). A bone from an earlier modern human, buried at the Paviland Cave in Wales 26,000 years ago, just one or two thousand years after Neanderthals had vanished, has also yielded a control-region sequence matching those of present-day people. But it is uncertain whether this sequence is genuinely ancient or might instead have come from a living person who handled the bone (Sykes, 2000).

If the mtDNA of the Paviland individual indeed matched that of present-day humans, then it becomes significant that the mtDNA of the Mezmaiskaya Neanderthal, who lived just three thousand years earlier, did not. Such a stark genetic difference between a modern human and a Neanderthal who were nearly contemporaries, and who both lived in Europe, would seem to suggest that Neanderthals did in fact occupy an evolutionary path separate from our own.

On the other hand, some researchers think the Mezmaiskaya infant was actually not a Neanderthal, but an early modern human (Hawks and Wolpoff, 2001). If true, then the divergence between its mtDNA and our own would suggest that early modern humans and Neanderthals were both genetically unlike us. This would support the view that Neanderthals, as well as modern humans, were ancestral to living people and that the differences between Neanderthal mtDNA and our own simply reflect changes that have occurred in the mitochondrial genome since Neanderthal times.

	DNA from an early modern human from Germany's Vogelherd cave seemingly differs from Neanderthal DNA, but it is not yet clear in what way. (Courtesy of the Town of Niederstotzingen, Germany.)

Because of the uncertainties surrounding the Paviland and Mezmaiskaya data, more evidence is needed to determine how Neanderthals might fit into our ancestry. The view that Neanderthals and early modern humans were genetically dissimilar finds some support from a study of DNA from an early modern human from Germany who coexisted with Neanderthals, although this study did not look at any actual DNA sequences.

In the study, which dealt with extracts of nuclear DNA, researchers hybridized DNA from two European Neanderthals with DNA from the humerus of a 35,000-year-old anatomically modern human from the Stetten (Vogelherd) cave site. These DNA samples were also hybridized with additional, control samples of DNA, including from a living human (Scholz, et al, 2000). The two Neanderthal samples always yielded hybridization signals that were similar to each other, but which differed from the signal of the Stetten early modern human by at least a factor of two. These results led the researchers to conclude that Neanderthals and early modern Europeans belonged to separate species. Hybridization studies may not be sensitive or reliable enough, however, to allow one to make that type of determination (Geigl, 2001).

To date, the best evidence on the genetic makeup of early modern humans comes from an mtDNA control-region sequence taken from an early modern man buried perhaps as early as 62,000 years ago at Lake Mungo in New South Wales, Australia. This individual, known as Lake Mungo 3, clearly lived during the Neanderthal era, but unfortunately resided far outside Neanderthal territory. This individual's mtDNA indicates that early modern humans did differ genetically from Neanderthals, but not to the extent that people had thought, and that they also differed genetically from us.

The mtDNA of Lake Mungo 3 has been found to belong to a primitive mtDNA lineage that predates the common ancestor of all present-day mtDNA lineages (Adcock, et al, 2001a). The presence of such non-modern mtDNA in this early modern man threatens the view that the mtDNA variant ancestral to all present-day mtDNA sequences arose in, and was spread by, early modern immigrants from Africa who replaced the indigenous archaic peoples they met. If Lake Mungo 3 descended from early modern Africans who possessed this root mtDNA type, and who were the ancestors of us all, why is his mtDNA sequence more primitive than this supposed ancestral sequence?

Lake Mungo, now dry, yielded remains of one of Australia's earliest inhabitants. This early modern human's mtDNA differs from that of living humans. (From the Max Preece web site. Courtesy of June Preece. Photo © 1998 Max Preece.)

The Lake Mungo researchers suggest that some time after modern humans such as Lake Mungo 3 were already present in Australia, an mtDNA variant arose that was particularly favored by natural selection. That variant, ancestral to all present-day mtDNA sequences, proliferated throughout the Old World because of that selective advantage, not because it was spread by a migrating population. Under this scenario, the earliest modern humans of Australia and Eurasia could have evolved from local archaic populations. On the other hand, they might still have arisen from early modern African immigrants if some of those immigrants, perhaps ancestors of Lake Mungo 3, had unexpectedly primitive mtDNA sequences, which later vanished.

The Lake Mungo evidence has not been accepted by all scholars. Some think the Lake Mungo 3 mtDNA is not a genuinely ancient sample but a recent contaminant (Cooper, et al, 2001). They doubt that DNA would survive 62,000 years in the relatively warm environment of Lake Mungo. Although mtDNA sequences have been recovered from animal bones dating back at least 90,000 years (Loreille, et al, 2001: table S1) and perhaps 150,000 years or more (Jones, 2001: 116), these remains came from either cave or permafrost contexts where low temperatures helped preserve the DNA. Similarly, the Neanderthal bones that have yielded mtDNA sequences have all come from relatively cool European caves (Smith, et al, 2001). Temperature is not the only factor, however, that influences how long DNA can last (Ovchinnikov, et al, 2001).

Researchers who doubt the Lake Mungo findings also claim that even if the mtDNA sequence truly did come from Lake Mungo 3, the sequence is not as primitive as has been proposed and can be accommodated into the range of variation of present-day mtDNA. The scientists who extracted and studied the Lake Mungo 3 mtDNA reject this claim and also dismiss the charge that the mtDNA sequence is not genuinely ancient (Adcock, et al, 2001b). To satisfy their critics, however, they are reportedly now reanalyzing the mtDNA of Lake Mungo 3 and are allowing three independent labs to analyze it as well (D'Agnese, 2002: 57).

If primitive mtDNA is in fact present in this early modern Australian, and also in what may be an early modern human from Mezmaiskaya, Russia, it becomes harder to argue that the differences between Neanderthal mtDNA and our own resulted from a genetic splitting between separate Neanderthal and modern lineages. If both Neanderthals and early modern humans had mtDNA more primitive than ours, why should Neanderthals be singled out as a divergent human type to be excluded from our ancestry? On the other hand, if the mtDNA from the 26,000-year-old modern human from Paviland Cave in Wales is genuinely ancient, its close similarity to present-day mtDNA would appear to justify bumping Neanderthals off our evolutionary path.

Although the Lake Mungo 3 evidence narrows the genetic gap between Neanderthals and modern humans, it does not support the idea that Neanderthals contributed as much to the mitochondrial gene pool of present-day humans as early modern humans did. The mtDNA sequences of the undisputed Neanderthals from Feldhofer and Vindija show 35 differences, on average, from those of present-day humans. In contrast, the mtDNA sequence of Lake Mungo 3 typically shows only about 12 differences from the mtDNA sequences of living humans.

The greater similarity of the mtDNA of Lake Mungo 3 to our own, compared to that of the Feldhofer and Vindija Neanderthals, cannot be attributed to changes in the mitochondrial genome over time. Lake Mungo 3 is reportedly more ancient than these Neanderthals, although he might actually be contemporaneous with them, if his true age turns out to be just 40-45,000 years, as some people think (Bowler and Magee, 2000; Gillespie and Roberts, 2000). In any case, Lake Mungo 3 clearly did not come from a population more recent, and for that reason more genetically evolved, than the Feldhofer and Vindija Neanderthal populations. This means that true genetic differences existed between the Neanderthals and their early modern contemporaries in Australia, and the mtDNA of the early modern Australians was more like our own.

Despite the mtDNA differences between Neanderthals and early modern humans, the nuclear DNA of living people suggests we may nevertheless have inherited at least some Neanderthal genes. Some evidence indicates that the mutation that causes most cases of cystic fibrosis, called delta F508, may have originated in Neanderthals (Wolpoff, 1999: 759-760). Delta F508 is clearly ancient. It and other common cystic fibrosis mutations are associated with unique genetic patterns rare among living groups. It is also possible, however, that these mutations arose not in Neanderthals, but in early modern humans whose genetic traces have now mostly vanished (Mateu, et al, 2002).

Back to Top