by Bryan Sykes
The day after I got back, I unwrapped the samples. There was blood oozing everywhere. The glass tubes had smashed – but fortunately, not all of them. Twenty were still intact, and I got on with sequencing their mitochondrial DNA. Nowadays, DNA sequencing is done automatically in extremely expensive machines, but in the early 1990s it was a manual operation which involved tagging fragments of DNA with mildly radioactive isotopes and separating them in an electric field. There was a point at the end of the long process when the X-ray film which recorded the pattern of radioactive bands that revealed the sequence slowly issued from the developing machine. It was always a tense moment. Will there be a good set of bands? Will there be any bands at all? If the bands are too weak or absent altogether, then something has gone wrong and it’s back to the laboratory bench for another three days.
This time, with the first ten of the twenty samples, everything had worked. Drawn across the X-ray film were four wide columns of dark bands, like bar codes, where the weak radioactivity had blackened the photographic emulsion. The four columns were each divided into ten tracks, one for each sample. Each of the four columns read the sequence of one base, so by putting them together the complete sequence could be worked out. I had arranged it this way, with the ten samples side by side, so that it was easy to see where the differences between individuals were. That was what I really wanted to focus on – the differences between people, rather than the similarities. A straight line across all ten tracks meant that all ten samples were identical at that position; a line with interruptions meant that some samples were different at that position.
In the lab we had sequenced ourselves and a few friends, mostly European, and typically there would be a couple of dozen lines in each batch of ten samples that showed these tell-tale interruptions. When the Rarotongan film came sliding out of the developer there were bands all right, but there was not a single interruption. They were all exactly the same. Had I made a mistake? Had I inadvertently combined the samples somewhere along the line? I needed to develop the second film from samples 11–20 to find out. When this came out it looked at first as if I really had made a mistake. Another whole film of straight lines. But then I saw one track, one individual, that was different from all the rest. Very different. And three other tracks had a single interrupted line. So they hadn’t been mixed. They were real results. I realized at once that they were stunning, and that before very long I would have the answer to the origin of the Polynesians.
Reading through the sequences more carefully and comparing them with the European reference sequence, I saw that the major sequence shared by sixteen of the twenty Polynesians was different at four positions: 189, 217, 247 and 261. The similar sequence shared by three individuals on the second film differed from this only in that they did not have the variant at 247. Otherwise their mitochondrial DNA was identical; they had to be very closely related to the first sixteen. But the twentieth sample was completely different. It had nine variants from the reference sequence along the control region, none of which was shared with the main Rarotongan cluster. Since the blood samples had come from the outpatient clinic in Avarua, there was no guarantee that they were from native Rarotongans, and so I assumed that this unusual sequence was from a tourist or a visitor from another part of the world. Since very few mitochondrial sequences had been published in 1991, there was no telling where on the globe this might be.
I concentrated on the main result – the astonishing similarity of nineteen of the twenty samples. This had to be the mitochondrial DNA of the original Polynesians. All we had to do to solve the Polynesian question was to look in both south-east Asia and South America for comparisons. If we found DNA matches in Chile or Peru, or even in coastal North America, then Heyerdahl was right. If we found them in south-east Asia, he was wrong. If we didn’t find a match in either region, then everyone was wrong. Whichever turned out to be true, one thing was certain: we were going to settle, once and for all, the debate that had raged for over two hundred years. I started to plan my next trip.
You might be asking yourself at this point: ‘Surely if it were as easy as that, blood groups would have given the answer long ago?’ It isn’t as if the blood groups of Polynesia had never been studied; indeed, the first results from Samoa in central Polynesia had been published in 1924, only five years after the Lancet paper by the Herschfelds which first introduced the potential of blood grouping in anthropology. The south Pacific, as I was fast learning, had been a popular place for scientific fieldwork for a long time. However, while they formed a plank in the argument in favour of a south-east Asian origin, decades of work on blood groups and other classical genetic systems had still not produced a definite answer to the puzzle, first because the variations are not definitive, and second because the evolutionary relationships between the types are not known. For example, Polynesians, native South Americans and south-east Asians all have a high frequency of blood group O. Polynesians also have quite a high frequency of blood group A, which is virtually absent in South America. But they also have a low frequency of blood group B, which is quite common in south-east Asia. So what can you make of all that? Which theory do these data support? Advocates of the Asian origin would argue that the extreme rarity of blood group A in native South Americans means that the Polynesian blood group A couldn’t have come from South America. Supporters of the South American case could legitimately respond by saying, as Arthur Mourant suggested in 1976, that the Polynesian blood group A came originally not from Asia but from Europeans through intermarriage over the last three hundred years. And anyway, where’s all the blood group B that should have come from Asia? Add to all this uncertainty the fact that, ultimately, all native Americans trace their origins to Asia through the settlers who crossed the Bering land bridge thousands of years before, and you have a complete mess. Blood group O could have reached Polynesia either directly from Asia or via the Americas. There is no way of knowing. With only three blood group genes – A, B, and O – certainty remains out of reach.
Other classical genetic markers are more variable, none more so than the one that controls the tissue type system important in organ transplantation. Just as blood needs to be cross-matched before a transfusion to avoid a fatal immune reaction, so you must match tissue types between donor and recipient when transplanting organs like heart, kidneys or bone marrow. You don’t hear of people waiting for a blood transfusion because they can’t find a match, but it is a sadly familiar story to hear of patients waiting for months or even years for a suitable heart or kidney donor, often dying before one is found. This is because while there are only four blood groups (A, B, AB and O) there are scores of different tissue types.
I must admit here and now to a serious personal weakness. I have a complete mental block when confronted with the bewildering variety of tissue types. Some of my best friends are cellular immunologists who live, work and breathe tissue types, and the Institute where I work is packed with them. Yet something switches off in my brain when they start describing the various types. All of them begin with the three letters HLA. Then numbers and letters are tacked on to the end: HLA–DRB1, HLA–DPB2, HLA–B27 and so forth. Time and again I go to seminars which kick off with a slide showing a table of this horrendous alphanumeric mélange. For years I concentrated, thinking it would sink in eventually if I tried hard. After all, I have to teach this stuff in my genetics classes. But to no avail. I reluctantly conclude that I am genetically incapable of understanding tissue types beyond knowing that there are an awful lot of them. Which, fortunately, is all you need to know as well. Since there are lots of them, and there are quite a lot of data from Polynesia, South America and south-east Asia, it is relatively easy to track them; and sure enough, most of the tissue type connections are between Polynesia and Asia. But not all. A type called HLA–Bw48 is very rare everywhere except among Polynesians, Inuit and native North Americans. However, though there is certainly plenty of variation, the evolutionary connection between the different types was
not known. So, for example, you couldn’t tell whether HLA–Bw48, the type found also in North America, was closely related to other Polynesian types or not. Compare that to the situation with the mitochondrial DNA from Rarotonga. We know that there are three types; we also know that two of them are very closely related to each other, while the third is not. That, as we will see, is an enormous help. We can search other lands not only for the Polynesian mitochondrial types themselves, but for others that are closely related to them as well.
By the time I had planned the return trip, and persuaded the Royal Society to pay for it – after all, they had paid for Cook’s first voyage to Tahiti, as I pointed out in my application – data from native North and South Americans produced by other researchers had begun to circulate. Just as there was one cluster in the Rarotongan sample (if we include the two closely related types in a single cluster and forget about the single sequence from the ‘tourist’), so there were four main clusters in the Americas. Three of these had quite different mitochondrial DNA sequences; the fourth was rather similar to the main Rarotongan sequence of 189, 217, 247, 261, but with variants at positions 189 and 217 only. This looked very interesting. Moreover, both the native American and Rarotongan DNAs shared another unique feature. At the opposite side of the mitochondrial DNA circle from the control region that we had sequenced, a small piece of DNA, only nine bases long, was missing. This definitely increased the chances that the American and Polynesian types were related. Things were looking up for Heyerdahl.
I had heard that in Hawaii Rebecca Cann, one of the authors with Allan Wilson of the original 1987 paper on mitochondrial DNA and human evolution, was studying the DNA of native Hawaiians. This is hard work because, unlike in Rarotonga, there are very few of them left. Two hundred years of immigration, mainly from Asia and America, have reduced the native Hawaiians to a fringe population, many of them living a marginal existence – an all too familiar legacy of colonialism. However, schemes have recently been introduced by which special grants and scholarships are awarded to those who can prove they are of native Hawaiian ancestry. One way of proving this ancestry is through DNA testing; so there was an extra incentive to find out about the mitochondrial genetics of the native Hawaiians.
On my return visit to Rarotonga I arranged to call on Becky Cann in Hawaii, where we sat down in her lab with her postgraduate student, Koji Lum, to compare results. It didn’t take long to discover that we had both found the same major Polynesian type, with the deletion and the same control region variants. This was very exciting, and confirmed the link between the people of Hawaii and those of Rarotonga, three thousand miles to the south. Already I was imagining the enormous ocean distance that separated the two island groups, and the fantastic voyages that must have carried these genes across the sea. Even though it was not unexpected, given the wealth of evidence from the days of Captain Cook onwards that connected all the Polynesians to a common ancestry, just seeing the proof was thrilling. Reluctantly, Becky left to prepare for a seminar, leaving Koji and me in the office sharing our admiration for the voyages of the Polynesians that had carried these genes to Rarotonga and Hawaii.
What followed was one of those rare moments in science when something is revealed that has never been seen before. I was about to pack away my data when I remembered the unusual Rarotongan sequence that I had interpreted as belonging to a tourist and more or less forgotten about. I turned to Koji and asked him if he had ever seen anything like it in native Hawaiians. He agreed to have a look and unpacked his own sheets of results. There was one that stood out from the rest. I laid out my sheet, rather like a roll of wallpaper – this was in the days before laptops – on which the Rarotongan sequences were drawn out, and soon located the unusual sequence. At first Koji’s and my sequences looked completely different; then we realized that we had been reading them from opposite ends. I turned mine around, and began to go through the strange Rarotongan sequence. I read from the left-hand side. The first variant was at position 144.
‘Do you have anything with 144?’ I asked.
‘Yes,’ said Koji.
I carried on four more bases to 148. ‘Anything at 148?’
‘Yes, in the same sample,’ he replied.
I could feel the thrill of discovery tingling up my spine. I carried on. ‘223?’
‘Yes.’
‘241?’
‘Yes.’
I accelerated. ‘293?’
‘Yes.’
‘362?’
‘Yes.’
They were identical. We both looked up at the same time. Our eyes met and two huge, silent smiles shone out from our faces. This was not the DNA of a tourist at all. Discounting the remote possibility that I had accidentally collected a blood sample from a native Hawaiian on holiday in Rarotonga, this had to be a second genuine Polynesian DNA type that had reached into the Pacific as far as the Cook Islands and Hawaii. But where had it come from? It would take another six months to find out.
I flew down to Rarotonga, more determined than ever that we would solve the mystery surrounding the origins of the Polynesians. When I got there, Malcolm, my host from my first visit, arranged for me to meet the man who ran the Prime Minister’s office. In most countries this would be quite impossible, but in Rarotonga it was accomplished at Malcolm’s Christmas party on the beach. It was fortunate that I met Tere Tangiiti and arranged an appointment early on in the proceedings; because my abiding memory of that party was not of making a crucial diplomatic contact, but of the colour blue: the colour of Curaçao which, mixed with champagne, makes the cocktail Blue Lagoon. Blue Lagoon, seafood omelettes and my digestion don’t mix. I was soon to discover the interesting scientific fact that whatever it is they use to colour Curaçao, it is not destroyed in the human stomach. Ten years later I still feel sick at the sight of it.
I needed to get the permission of the cabinet and the co-operation of George Koteka at the health department to collect a substantial DNA sample from Rarotonga and the other islands. I met the cabinet in the Prime Minister’s office above the post office, and they could not have been more helpful. Within a few weeks I had collected five hundred samples from Rarotonga, Atiu, Aitutaki, Mangaia, Pukapuka, Rakahangha, Manihiki and even from the tiny atoll of Palmerston (population sixty-six). I packed them carefully in ice and took them back to Oxford.
7
THE GREATEST VOYAGERS
The Institute of Molecular Medicine, where my laboratory is based, is built around the pioneering work of its first director, Professor Sir David Weatherall. His research over the past twenty-five years has been focused on the inherited diseases of the blood, in particular those involving the main component of red blood cells – haemoglobin. These diseases are not particularly common in northern latitudes, but have a quite devastating effect on public health in parts of Africa, Asia and Mediterranean Europe. The main diseases, sickle cell anaemia in Africa south of the Sahara and thalassaemia in Asia and Europe, kill hundreds of thousands of children every year. The causes of all this misery are tiny mutations in the haemoglobin genes, which very slightly alter the oxygen-carrying properties of the red blood cells. In sickle cell anaemia, the usually circular red blood cells visibly change shape, as the name implies, and can no longer slide past each other in the narrowest of blood vessels. This clogs up the flow of blood to vital tissues. In thalassaemia the haemoglobin itself forms clumps inside the red blood cells, which are then destroyed in the spleen. Either way, the anaemias can be fatal if left untreated; and still the only effective remedy is repeated blood transfusions which – quite apart from the side-effects these cause by overloading the body with iron – are beyond the public health budgets in most of the affected regions.
Why do these diseases occur in some places and not in others? The answer is – malaria. Sickle cell anaemia and thalassaemia are found principally in parts of the world where malaria is, or has been, endemic. Both diseases, in order to develop, require a double dose of the mutant haemoglobin gene, one from
each parent. Many inherited diseases follow the same pattern; among Europeans the most familiar is cystic fibrosis, where the parents are carriers with one copy each of the mutant gene but no symptoms of the disease. For a reason that even now is not entirely clear, the parasite that causes malaria finds it difficult to infect the red blood cells of sickle cell anaemia and thalassaemia carriers, who as a consequence become at least partially resistant to the disease. Over many generations, this resistance leads to a spread of the haemoglobin mutations in the malarial regions through the forces of natural selection. However, while the mutations are good for carriers, they can be devastating for their children, because some of the offspring of two carrier parents get the double dose of haemoglobin mutants and develop the potentially fatal anaemias. This cruel balance of carrier advantage and offspring elimination keeps the haemoglobin mutants at a high frequency wherever malaria is found. Malaria does not cause these diseases directly, but does so indirectly by allowing, indeed encouraging, the haemoglobin mutations – which are the real cause – to survive and prosper. So, even if you eliminate malaria, you do not at once eliminate these diseases. In Mediterranean Europe – Sardinia, Italy, Greece, Cyprus and Turkey – programmes to eradicate the mosquitoes which carry the malarial parasite have virtually eliminated malaria – but not thalassaemia. Tens of thousands of people still carry the haemoglobin mutations, and only an entirely different programme, built around the genetic testing of prospective parents to see if they are carriers, is reducing the incidence of the disease.
