A century of human genetics
This lecture by Sir Alec Jeffreys was originally delivered at the Leicester Medical Society Bicentenary in 2000 and is reproduced here with the kind permission of Sir Peter Bell, former Society president.
A century of human genetics: Exploring variation and mutation in the human genome
Professor Sir Alec Jeffreys, HonFRCP FRCPath FMedSci FLS FRS
Origins
Human genetics in its modern sense can be traced to the beginning of the 20th Century, and to just two key individuals. The first was the Austrian physician Karl Landsteiner who in 1900 discovered the ABO blood group system, the first example of a variable human characteristic that was inherited according to the simple rules of Mendel. The second was the English physician Sir Archibald Garrod whose work at St Bartholomew's Hospital on the inherited disorder Alkaptonuria provided, in 1902, the first example of a human inherited disease which similarly showed Mendelian inheritance.
These two discoveries really defined the subsequent two great pathways of research in human genetics that continue to this day. The first is the study of normal genetic variation in humans, initially through blood group variation and subsequently in other classes of protein using biochemical techniques. These studies provide fundamental information on the genetic variability of the human species and on the diversification of mankind. The second great theme has been understanding how inherited defects in our chromosomes can cause inherited disease. Both approaches were however severely limited by the lack of large numbers of genetic markers, and only fully blossomed in the wake of the DNA revolution.
The rise of genomics
The announcement by Watson and Crick in 1953 of the structure of DNA signalled the beginning of a molecular revolution that is still with us. Recombinant DNA technology, first developed in 1973, allowed complex genomes to be manipulated, dissected and sequenced. The first human gene was isolated in 1977 and today we now have the final draft of the entire human genome sequence covering most of the three thousand million base pairs that constitute the human book of life.
This new science of genomics, exploring gene and chromosome organisation, however, tells us nothing about genetics, defined as the study of inherited variation. Genomic technologies have proved critical in exploring variation between individuals in our precise DNA sequences and it is this theme that has fascinated me ever since I first moved to Leicester in 1977.
Inherited variation in human DNA
Back in 1977, the only tools we had available for exploring human DNA variability were probes which enabled us to detect, by nucleic acid hybridisation, specific regions of DNA in human chromosomes, and restriction enzymes that allowed us to cut DNA at predetermined sites.
By combining these two approaches, we showed that humans do indeed show variation from person to person in the way that restriction enzymes cut their DNA and that these variations in the genetic material were inherited in a simple Mendelian fashion.
These so-called restriction fragment length polymorphisms allowed us to estimate for the first time just how variable human DNA was from person to person; the answer, at the molecular level, was not very great, with perhaps one base pair in every three hundred showing polymorphic variation between people. However, at the level of the whole genome, the variation was immense with perhaps ten million different positions in our chromosomes showing variation.
The problem of limited numbers of genetic markers was therefore, in principle at least, solved. The drawback was that these single nucleotide polymorphisms were at the time rather difficult to detect and assay, and were in any event genetically not very informative.
We therefore started in the early 1980s a quest for far more variable regions in human DNA that could provide highly informative markers for the mapping of human chromosomes.
Purely by accident, we stumbled across one of the first examples of a human minisatellite, a region of DNA consisting of 30 or so base pairs repeated over and over again for tens or hundreds of times. Not only were some of these minisatellites extraordinarily variable between people, with variation arising through inherited differences in the number of repeats, but they also seemed to share a similar DNA sequence.
We therefore reasoned that it should be possible to detect, by nucleic acid hybridisation, many minisatellites simultaneously by making use of this shared sequence motif. This approach worked remarkably well, not only displaying large numbers of minisatellites, but also providing for the first time a truly individual-specific pattern of human DNA. Thus was DNA fingerprinting born.
Forensic DNA - an unexpected diversion
It was blindingly obvious that DNA fingerprinting not only had generated large numbers of highly informative genetic markers for medical genetic research, but also had accidentally solved another major problem in human genetics, namely the issue of biological identification and the establishment of family relationships in forensic and legal medicine.
Within months of developing the first DNA fingerprint, we had applied this approach to resolving an immigration case and shortly afterwards a paternity dispute, the first cases of their kind ever to be tackled using molecular genetics. Using both DNA fingerprinting and the more refined DNA profiling technology, we carried out our first forensic investigation in 1986 in the Enderby murder case. It was this case that triggered the application of molecular genetics to criminal investigations worldwide, with DNA profiling remaining the standard technology well into the 1990s.
The next revolution came with the development of the polymerase chain reaction (PCR) invented by Kary Mullis, Henry Erlich and others. This exquisitely elegant and sensitive technology enabled minute amounts of DNA to be copied over and over again in the test tube, allowing traces of biological evidence found for example at the scene of crime to be amplified and typed. It also allowed the development of highly variable markers much shorter than minisatellites that could be easily amplified and typed.
We used these short markers, called microsatellites, in 1989 in an investigation of skeletal remains exhumed from a gravesite in Brazil and believed to be those of Dr Josef Mengele, the notorious Auschwitz concentration camp doctor. By comparing bone DNA profiles with living relatives of Mengele, it was possible to establish that the remains were indeed those of the war criminal, allowing a major investigation that had proceeded for decades to be brought to a conclusion.
Subsequently, microsatellite typing has been greatly refined and largely automated, allowing in 1995 the development in Britain of the first national criminal DNA intelligence database which currently holds over two million DNA profiles, mainly from convicted criminals but also from unsolved casework. This database is proving to be extraordinarily effective at identifying repeat offenders and in establishing links between different unsolved crimes.
Indeed, it has been so successful that momentum is now gathering for the possible databasing of the entire UK population. Such a global database would unquestionably prove to be an enormously powerful tool in the fight against crime, but does raise major social, legal and ethical issues that have yet to be adequately addressed.
On the origins of DNA variation
The science of DNA forensics having been largely solved, we turned our attention to more fundamental issues of human DNA variation, in particular trying to understand the mechanisms which result in heritable changes in our DNA.
There are two key processes at work here. The first is mutation whereby spontaneous, or possibly environmentally-induced, alterations in our DNA become established and transmitted to subsequent generations; unfortunately, mutation occurs over most of our DNA at a frequency so low as to be undetectable in human families.
The second process is recombination or crossing over, whereby maternal and paternal (homologous) chromosomes pair up at meiosis and exchange genetic material before being separated into individual sperm or eggs. Recombination is fundamentally important in reshuffling patterns of variation in human chromosomes, greatly increasing the genetic diversity of mankind and exposing novel combinations of genetic markers to the forces of natural selection.
Again, our knowledge of processes of recombination in human chromosomes is severely limited by the very low rate, at the molecular level, at which this process occurs. Our challenge therefore has been to develop new approaches to the very high resolution analysis of mutation and recombination in an attempt to study the evolution of human DNA in real time.
Minisatellites have proved to be a superb system for developing these technologies. These regions of DNA are not only extremely variable but can also mutate (change their number of repeats) at a phenomenal rate, with in some cases as high as one in ten sperm or eggs carrying a new mutation, compared with perhaps one in one million for mutation in conventional genes. Thus it is possible to see minisatellites mutating in families.
Even better, it is possible to use PCR to detect new mutations directly in sperm DNA, enabling for the first time unlimited numbers of germline mutants to be recovered from any appropriate man. In addition, we can analyse the internal structures of minisatellites in great detail, and by comparing the structures before and after mutation, we have shown, rather surprisingly, that minisatellite mutation is in fact driven by the process of crossing over at meiosis; this shows that for this class of DNA at least, mutation and recombination are not different and distinct processes.
In most cases minisatellites mutate by the copies (alleles) on homologous chromosomes pairing, followed by information being copied and transferred from one allele into the other, resulting in a recombinant mutant but without crossover as such. However, using new very powerful single sperm techniques it has been possible to study true crossovers at very high resolution.
This has shown that minisatellites are located at the boundaries of intense crossover hotspots, with current evidence suggesting that it is these hotspots that drive instability at minisatellites, and not vice versa. This in turn suggests that minisatellites are hotspot parasites that have succeeded in engaging the recombination machinery to propagate themselves.
The discovery of crossover hotspots near minisatellites raises major issues about the distribution of crossover events along human chromosomes, long thought to be fairly random. To see whether these hotspots are the rules rather than the exceptions in human chromosomes, we have recently commenced an analysis of a region within the Major Histocompatibility Complex, a giant complex of genes intimately involved in the function of the immune system.
This region has long been studied in families, in particular in relation to tissue matching in transplantation surgery, and these family studies have provided provisional evidence that crossovers may be significantly non-randomly distributed in this region of the human genome. We have therefore used very high-resolution sperm analysis to begin to dissect the molecular distribution of crossovers in this region.
All the evidence to date suggests that the highly localised hotspots associated with minisatellites are not the exception in the human genome, but are instead the rule. This in turn implies that crossovers in human chromosomes are tightly regulated, resulting in heavy clustering into highly localised regions of intense activity.
We are now extending these findings to other parts of the human genome, as well as trying to understand the rules that govern why hotspots exist and how they are distributed in human chromosomes. This is not just an academic quest; the clustering of crossovers into hotspots will profoundly influence the way in which variation is distributed along human chromosomes, with major implications for the search for genes involved in human inherited disease.
Environmental mutagenesis
It has long been known that environmental agents such as ionising radiation can induce inherited mutation in the DNA of experimental organisms. Surprisingly, there is no evidence for such a phenomenon in human populations, in for example individuals exposed to radiation following the Hiroshima and Nagasaki atom bombs.
The problem lies in the very low mutation rates associated with traditional germline monitoring systems, which require the monitoring of hundreds of thousands, or indeed millions, of families to be able to detect induced inherited mutation. My colleague Dr Yuri Dubrova and I have therefore used highly unstable minisatellites as a novel approach to mutation monitoring, in particular to see whether they respond to environmental agents such as ionising radiation.
Exposure experiments are of course very difficult to perform in humans. We have therefore focused our attention on populations accidentally exposed to radioactivity. In a major study on the Chernobyl disaster, we have shown that populations in Belarus living in contaminated regions show an unusually high mutation rate in their children, which appears to correlate with the degree of radioactive fallout in their environment.
This provides the first direct evidence for radiation-induced inherited mutation in humans, but must be treated with considerable caution since there may be other confounding environmental factors responsible for shifting mutation rates.
However, Dr Dubrova has observed a similar phenomenon in populations exposed to radioactive fallout in the vicinity of the Soviet nuclear weapons testing facility at Semipalatinsk in the former Soviet Union of Kazakhstan. Two studies on the Chernobyl liquidators, the men sent in to clean up the power station following the disaster, have produced conflicting results, one showing evidence for radiation-induced mutants appearing in their offspring and the other showing no effect.
Similarly, minisatellite mutation studies on Japanese atom bomb survivors and their families have shown no evidence for radiation-induced mutation, nor is there any evidence for radiation inducing sperm mutation in men undergoing lower abdominal radiotherapy. Clearly, much more work is needed to see whether human minisatellites do respond to radiation in the remarkable fashion seen for repeat DNA in mice. If radiation-induced mutation in humans is verified, then it clearly has major implications both for understanding the interaction between radioactivity and the germline and for radiobiological protection.
The future
With the human genome sequence largely completed, human genetics is poised to move rapidly in two major directions.
First, functional genomics, namely an attempt to understand the functions of the tens of thousands of genes present in our chromosomes. The second challenge is to gain a far more global picture of diversity in the human genome and to use this to understand our evolutionary origins and to begin to dissect the many and varied processes of mutation, recombination and natural selection that drive variability into human populations.
In the latter respect, we have made our first few steps using minisatellites as model systems. The challenge for the future is to extend this work to other classes of rarer but potentially more biologically relevant inherited mutation.
Finally, I would like to thank all my colleagues both in my research group and in the Department of Genetics at the University of Leicester, together with many other friends both within and outside the University, who have contributed so much to all our research efforts.