Unlocking the Past

by Martin Jones

  For most of her short life, carbon and nitrogen isotopes in the maiden’s hair revealed a familiar peasant diet; not much meat, more like root crops, vegetables, and quinoa. It may have been a mobile childhood; fluctuations in the stable hydrogen and oxygen isotopes within her hair suggesting a seasonal fluctuation in her water source, perhaps alternating between the valley and the high mountain. In her final year however, this changed.

  The length of hair closest to the root displayed different levels of enrichment for both carbon and nitrogen; she was consuming a lot more meat and maize. From the lipids in that length of her hair, it was evident she was also consuming a plant of a more particular kind.

  Tucked into a mouth is a ‘quid’ of coca leaves, a plant still used to ease life at high altitudes. However, this was more than a one-off final palliative. A coca-derived lipid called benzoylecgonine was present along stretches of her hair at unusually high quantities. By taking samples from along the full length of her hair, it transpired that her mind had been in a significantly altered state for several months leading up to her death. The maize may have contributed to that altered state; the precise balance of different lipids in her hair is consistent with consumption of alcohol alongside the coca. She was so very high (in mental state as well as in altitude) her dreamy death may have arisen from hyperthermia alone.

  The whole narrative is consistent with a documented Capacocha rite, which entailed sacrifice of young children to the fertility goddess Pachamama, in anticipation of abundant harvests the following year. Their mitochondrial DNA indicated that the two younger children accompanying her were of different families, but that a third isotope measurement, this time of sulphur, showed that their dietary profiles converged (most likely in the capital city of Cuzco) six months before death, some time after the maiden had been fattened up with meaty food, and not long before they would need to set off on their final trek to the volcano’s summit and ultimate sacrifice.

  To best appease the goddess, the sacrificial children needed to be physically perfect and healthy. The rich meaty diet of the last six months would no doubt have helped that, but the maiden nonetheless was not entirely well, something revealed from further molecular research, conducted on swabs from the maiden’s excellently preserved lips and mouth. Those swabs retained DNA traces of a familiar bacterium Mycobacterium, not in conclusive evidence of a tuberculosis-type infection, but interesting in the light of proteins recovered from the swab.

  Angelique Corthals employed an approach called shotgun proteomics which analyses proteins in the context of an extremely well understood genome, in this case the human genome. Rather in the same way that DNA methods have moved in the direction of elucidating whole genomes, shotgun proteomics uses high performance liquid chromatography and mass spectrometry to assemble and quantify a total protein suite. The suite in the maiden’s mouth contrasted with that in mouth of one of the accompanying children, and the differences were consistent with a pronounced inflammatory/immune response. This may connect with respiratory infection from the Mycobacterium DNA, found in the maiden’s swab, but not that of the accompanying child.

  These two examples, one on the scale of continental communities over millennia, the other the last few years of a teenager’s life, both show various ways in which interferences from the DNA blueprint can be extended, to gene expression, physiology, function and ecology, by venturing beyond DNA to explore the other molecules.

  9

  friends and relations

  individuals

  There was always a hope, ever since it became clear that ancient biomolecules survived, that they might lead us to rediscover some of prehistory’s individuals. We might find out something about them, who their families were, and with whom they made contact for the exchange of goods and ideas, for the establishment of social connections, and for marriage partners. This was in part because of the particular powers of DNA, the molecule that was attracting the most attention, and in part because of the possibility of working with very tiny traces indeed. It was, at least in theory, feasible to trace an individual’s path through life by typing the DNA from shed hairs or flecks of skin. Even a fingerprint might leave a mark of its maker behind. Just as molecular archaeology was gaining momentum, so those same possibilities were revolutionizing forensic science. Individuals were indeed being identified from minimal traces at the crime scene–a speck of blood, a semen stain, or even a discarded cigarette butt that had been in contact with the wrongdoer’s lips. From these tiny traces, the polymerase chain reaction was able to bulk up sufficient amplified DNA for it to be typed and linked to a specific individual.

  A method for tracking down individuals from their DNA was already in place before PCR came on the scene. It is what has become known as ‘genetic fingerprinting’. What happens is that selected restriction enzymes are used to break up the entire DNA blueprint into strands of different lengths, according to where the enzymes bind. That multitude of created strands is then allowed to migrate along an electrophoresis gel to generate a characteristic sequence of bands, the ‘fingerprint’. As most of the genome is identical in most of us, the great majority of bands will also be identical. However, the various hot-spots along the genome will lead to strands of varying lengths and thus minute differences will be encountered between individuals, according to how distantly or closely related they are. The entire fingerprint is unique to the individual concerned.

  This remarkable method was unfortunately not transferable to the archaeological record, for one simple reason. As soon as the cell dies and the active DNA repair mechanisms cease, the DNA itself begins to fragment. In a period of a few years it becomes broken up into the very short lengths that are the stuff of ancient DNA science. The restriction enzymes used in fingerprinting will only add to a process of fragmentation that is already heavily advanced. As with all ancient DNA projects, we have to work on variation that is sufficiently localized to be retained within those fragmented strands.

  In its original form, genetic fingerprinting was also limited in its forensic applications by the quantities needed, which were often in excess of what the criminal left at the scene of the crime. This constraint disappeared with the advent of PCR, opening the way to the acquisition of a fingerprint from minute traces. A second modification of the method opened the door to identifying past individuals from traces of their ancient DNA. This entailed a shift of attention from the whole corpus of DNA within the cell, to short, informative strands of DNA instead. It was an approach that Erika Hagelberg took with some considerable success.

  The opportunity to develop such an approach arose when she turned her attention to two criminal cases that were very much more recent than traditional archaeological remains, but not so fresh that modern DNA methods could be applied. One of these cases involved a murder victim, a fifteen-year-old girl whose decomposed remains were unearthed eight years after her death in 1981. The other involved a man who had died in a swimming accident two years previously. It was suspected that he might have been involved with murder, but as perpetrator rather than victim. It was thought that his true identity was the notorious ‘Angel of Death’ of Auschwitz, Dr Joseph Mengele. In both these cases there was an uncertainty about identity, and DNA science was brought in to resolve that uncertainty. However, even though the bones examined were from bodies that had only been dead for a few years, the human DNA within them proved to be heavily fragmented, apparently more so than some of the better conserved prehistoric bodies. Although the DNA under examination was ‘younger’ than some of the DNA still active within the bodies of the scientists doing the work, in experimental terms it was ‘ancient’.

  The approach that Hagelberg took was to search for the most highly variable hot-spots she could find within the genome. They would need to be sufficiently compact to show up within the short fragments to which their DNA had been reduced. Furthermore, they would need to be even ‘hotter’, in evolutionary terms, than the regions used to explore migrations and agricultur
e, as here Hagelberg was looking at specific individuals rather than broader lineages. The obvious choice was a feature known as a ‘microsatellite’. This is a short strand of highly repetitive DNA, whose length is extremely variable. The repetition is often composed of just two or three bases, repeated in tandem any number of times. When the repeat unit gets larger, and the overall length goes much beyond 150 base-pairs, the term used is ‘mini-satellite’, but the principle is the same. In both the studies, Erika Hagelberg turned her attention to microsatellites made up of variable repeats of a cytosine-adenine pair. Such microsatellites are found on a number of different human chromosomes, and their variation might allow a kind of fingerprint that would hold true even in fragmented DNA.

  In each of the two cases, she needed access to tissue from close living relatives. The bereaved parents of the girl thought to be a murder victim provided blood samples, as did Joseph Mengele’s wife and son. From the samples, Hagelberg and her colleagues built up a picture of microsatellite variation among two parents and their offspring in each of the two suspected family groups. If either identification were correct, then the range of microsatellite variation in the offspring should be neatly attributable to an even inheritance from the two parents. In both family groups this was found to be the case. In England, the bereaved parents were now able to bury their daughter, and on the other side of the Atlantic in South America, Auschwitz’s Angel of Death had been identified.

  This approach has generated a rich new strand of forensic science. There are, tragically, a vast number of ‘missing persons’, unceremoniously dumped in the context of totalitarianism and genocide around the world. In shallow graves especially, the transformation from intact ‘modern’ to fragmented ‘ancient’ DNA takes only a few years and, for many, ancient DNA techniques may offer the only means of identifying lost relatives. Erika Hagelberg and others have continued with this forensic project, at the same time exploring extending the approach to the less harrowing identifications from the more distant past.

  the death of the romanovs

  One renowned massacre, on the boundaries of living memory, took place in 1918, in the Urals of central Russia at Ekaterinburg. What we know of events derives from a dossier compiled over the following year by one Nikolai Sokolov. His dossier records the events of the night of 16 July, or shortly thereafter. The last Tsar of the Russian Empire, Nicholas II, was taken down to a cellar of the house that was serving as his makeshift prison. Accompanying him was his Tsarina, their five children, their doctor and three servants. Down in the cellar, they were shot and mutilated. A plan to dispose of them down a mine-shaft went wrong, and the bodies ended up in a hastily dug pit. The backfilled hollow was driven over to flatten it, and drenched with sulphuric acid in an attempt to erase the evidence. Seventy-two years later, twenty miles outside Ekaterinburg, two amateur historians recovered nine badly damaged skeletons from a shallow grave.

  A consortium of English and Russian scientists was invited to investigate the discovery. The problem facing the international team, which included Erika Hagelberg, was at least one step more complex than that of the more recent bodies of Mengele and the young murder victim. The treatment of the bodies was clearly unfavourable to molecular conservation, but that had been true of the younger remains as well. In all cases, the DNA would be heavily fragmented. An added difficulty in identifying the massacred bodies of the Romanovs was that the very purpose of the killing was to extinguish the family line. There were no spouses and offspring to provide samples. However, the Romanovs belonged to one of the best documented families in the world, and the possibility of tracking down distant surviving relatives remained open. As this would involve tracking through several generations, a different approach was required from that used with Mengele and the murder victim. At the remove of only one generation, nuclear microsatellites would remain sufficiently intact to link parent and child. In the context of tracking over several generations, those same patterns begin to become confused by the several episodes of sexual recombination of DNA sequences. This does not render them useless, and the research team did manage to establish from them that five of the skeletons conformed to the pattern expected from a family group. But to get a sure result, what they needed was to follow a ‘haplotype’ pattern that would remain much more intact between generations. For this they used the sequence that has underpinned so many ancient DNA projects, a hypervariable segment of the mitochondrial control region.

  What they were after were control region haplotypes that would remain faithful across the different maternal lineages that converged on the Tsar’s family. If these led to living relatives, then the identity of the bones from the shallow grave might be established. The two maternal lines they tracked were the ones running through the Tsar and his Tsarina respectively. The Tsar’s mother was Dagmar, daughter of Louise of Hesse-Cassel. Her mitochondria would be borne by the Tsar, but not by his children, who would carry the Tsarina’s mitochondria instead. The continuation of the Tsar’s mitochondrial genome would only occur through his maternally linked female relatives. He had a sister and an aunt sharing his mitochondrial genome, and both these women had living descendants following the maternal line. The great-great-grandson of his aunt and the great-granddaughter of his sister are both alive today, and each agreed to provide blood samples.

  As for the Tsarina, her mother was Princess Alice, the second daughter of Queen Victoria. The Tsarina had a sister, Princess Victoria of Hesse, and all these women naturally possessed the same mitochondrial genome. That genome was in turn shared by Princess Victoria’s daughter, Alice of Battenburg, and her son Philip, who went on to marry another of Queen Victoria’s descendants, the reigning British Queen, Elizabeth. Philip, now Duke of Edinburgh, was Chancellor of Cambridge University when Erika Hagelberg was working there. He too agreed to supply a blood sample.

  Having carefully targeted the appropriate descendants, always following the maternal line, the team had managed to identify one mitochondrial haplotype that would match that of the Tsar, and another mitochondrial haplotype that would match that of the Tsarina and all her children. If Sokolov’s account of the massacre was accurate, and the two amateur historians had actually recovered the bodies of the imperial family, then this would be the pattern: among the bones, those of one adult male would have carried the first of the two haplo-types, that of the Tsar, and those of one adult female, four younger females and one younger male would have carried the second, that of the Tsarina. We would also expect four other adult males (the servants and doctor) to match neither. That would constitute a direct translation of Sokolov’s account into mitochondrial genetics. With that model result in mind, the team set to work recording the hypervariable sections of the control region, as preserved in the recently unearthed bones.

  What they recovered was a striking, but incomplete, match. First of all, there were four adult skeletons matching neither haplotype–these could correspond to the three servants and the doctor. Second, the Duke of Edinburgh’s haplotype matched that of the adult woman and of the younger skeletons, which numbered not five but three. It seemed that two of the children were absent from the shallow grave. The haplotype of the remaining adult male almost matched that of the living descendants of Louise of Hesse-Cassel and, by inference, the Tsar. At one point along the Tsar’s sequence the nucleotide was mixed, a fascinating reminder that these DNA regions are studied because evolution along them is so rapid. In addition to providing compelling evidence that the Romanov family had been rediscovered, this slight difference within individuals separated by four generations has left a graphic biochemical trace of evolution in action.

  breaking the link with the present

  The Ekaterinburg bones demonstrated how, through ancient DNA analysis, the study of kin relations could be taken back in time to a poorly conserved archaeological deposit, albeit a rather recent one. It remained, however, a rather special case. We rarely have such an extensive family tree with which to reach back from the living
descendants to the excavated bones of their presumed ancestors. That is rare enough in burials 100 years old, let alone those from conventional archaeological deposits.

  One research group that from the outset has been tackling the possibility of approaching kin relationships through ancient DNA is the Gottingen laboratory of Susanne Hummel and Bernd Herrmann. They too have been attempting to make sense of archaeological burial groups by amplifying microsatellites and other fast evolving units. In 1993, archaeologists were able to provide the Gottingen team with a family group that could be analysed in as fine detail as the Romanov group. They found it in an ancient church in Lower Bavaria at Reichersdorf. Here eight skeletons had been excavated from beneath the chancel floor. Set into the surrounding chancel walls were the memorial stones that provided Julia Gerstenberger at the Gottingen lab with the template she needed. The stones recorded seven male members of an aristocratic line, the Earls of Konigsfeld. That seemed close enough to look for a match with the eight burials. The first of the named earls was Hanns Christoph, who died in 1546. The last was Georg Josef, who died two centuries later in 1749. The remaining five were the intervening earls in the patrilineage. The church clearly served as a sepulchre for this male aristocratic line, and it seemed quite feasible that the excavated skeletons might be of the earls themselves. Even prior to the team testing this through ancient DNA analysis, there were hints from examination of the skeletons that the match would not be 100 percent. At least one of the bodies had both the skeletal form, and fragments of the dress, of an adult woman, and there were no stones in the chancel commemorating her.

  The Gottingen lab had also been using ancient DNA as a means of sexing ancient skeletons. There is a gene involved in tooth development, the amelogenin gene, which is carried by both the X and the Y chromosome. The gene is of a slightly different length in the two chromosomes. When an appropriate sequence from the gene is amplified and banded on an electrophoresis gel, a female (XX) will generate a single band, and a male (XY) a double band. On the basis of this difference, Gerstenberger established that the skeleton, as suspected, was indeed genetically female. What is more, she was not the only female who had been laid to rest within this apparently male shrine. In all, the amelogenin tests indicated that only six of the skeletons were male, leaving two who were female. If the skeletons beneath the chancel really did correspond to the names recorded on the wall, then one of the earls was missing, and there were two female interlopers in the shrine.