Meeting the matchmaker

Nearly 20 years ago, Sir Alec Jeffreys made a discovery that would lead to the development of DNA fingerprinting, one of the most powerful tools available for identifying criminal suspects or for establishing personal identification. Jonathan Cox went to meet Jeffreys to find out more.

Exceptionally for a scientist of his age - he is in his mid-fifties - Sir Alec Jeffreys spends at least a third of his time at the bench, partly to keep up with the technology, but also because he enjoys it. It should not have been a surprise, then, to find him rooting around in a small box of samples from the freezer when I arrived to interview him. Even so, it took me a moment to realise who I was looking at.

Jeffreys’s office is a modest room set in the corner of a lab on the ground floor of the Genetics department at Leicester University, UK. A silver Royal Society of Chemistry plaque by the entrance to the building proudly announces that it was here that DNA fingerprinting was discovered, almost 20 years ago.

On a chair in his office sat a biography of Josef Mengele, Auschwitz’s Angel of Death. Jeffreys had authenticated Mengele’s remains in the early 1990s, an experience he admits was ’spooky’. I also spotted on his noticeboard a stegosaurian histogram from his 1991 Nature paper on minisatellite variant repeat PCR1 [polymerase chain reaction], a very clever development of the DNA fingerprinting method. (’Too bloody clever,’ Jeffreys later told me. ’The forensics community never really got their heads round it.’)

And, of course, fixed to the noticeboard were several panels of DNA fingerprints. One of these, a grey plastic sheet with a handful of black cloudy caterpillars lined up side-by-side, turned out to be the X-ray film from the very first DNA fingerprinting experiment. Although sold in auction, this film has never been published. Jeffreys described it, not without fondness, as ’pretty rubbish’.

Five minutes standing in the darkroom
I wanted to ask Jeffreys about the possibility of establishing a national DNA database and whether DNA fingerprinting had any drawbacks. But first I asked how he discovered fingerprinting, and whether there was there a specific moment when he thought, ’My God, this is it’.

Before answering, Jeffreys made the point that the discovery was ’a complete accident from beginning to end.’ In the late 1970s, he was hunting for genetic markers that would help map genes on human chromosomes and aid the diagnosis of disease. He began by looking at restriction fragment length polymorphisms (RFLPs) - alterations in the DNA sequence that lead to the appearance or disappearance of recognition sites for DNA chopping enzymes - but these were not very variable. (With improvements in molecular biology techniques, he has since returned to RFLPs, although they are now called single nucleotide polymorphisms [SNPs, pronounced ’snips’]). In 1980, molecular geneticists Arlene Wyman and Ray White, working in Massachusetts, US, had by chance discovered a highly variable region in human DNA. Others, also by chance, found similar regions of high variability (’hypervariable regions’), and deduced that they arose from tandemly repeated sections of DNA (see box, p53). Jeffreys appreciated that the variability of this tandemly repeated DNA would make it much more valuable as a genetic marker than RFLPs. ’The really key observation’ said Jeffreys, ’was that there seemed to be a short DNA sequence that was shared by these [hypervariable regions].’

His group made a probe consisting of this short sequence repeated over and over again, and asked themselves a very simple question: would it hybridise to anything in human DNA?

Jeffreys developed and fixed the X-ray film with the answer to this question on Monday 17 September 1984: ’Five past nine - I remember it extremely clearly’. He saw ’a very messy looking picture. My first reaction looking at this was "Oh, this is a complicated mess, forget it". But then within a minute the penny dropped. You could see, even though the quality was poor, highly variable patterns. And on that very first Southern blot [the X-ray film on which the fingerprints appear] we had a family group - a mother/father/child trio - and we could see that mum and dad were clearly different, and the child was a sort of composite of bits of mum and bits of dad . . . so we could see not only the potential for individual identification, but also establishing family relationships.’

As well as human DNA, Jeffreys had probed a veritable menagerie of DNA from other animals on that first Southern blot - mouse DNA, rat DNA, seal DNA, baboon DNA and cow DNA - together with some tobacco plant DNA for good measure.

’To our amazement we could see these patterns which appeared to be variable right across the board. So we could see the potential for a whole range of non-human testing as well . . . it really was a eureka moment. It was five minutes standing in the darkroom that literally changed my life - I mean it just went off in a totally different direction.’

About this boy
When he charged out of the darkroom and saw his technician, Victoria Wilson, they started very quickly writing down lists of things they could do with their new discovery (that would come to be known as DNA fingerprinting). ’We could see things like forensics, paternity testing, testing twins to see whether they are identical or not, possibly following successful transplantation - for example bone marrow transplants to see whether they’ve taken or not.’ But the application that really launched DNA fingerprinting was spotted when he went home that evening and sat down with his wife, Sue. She pointed out that he had forgotten immigration testing. When he heard this, Jeffreys said that three things crossed his mind: ’First, interesting; second, it’s not going to happen; third, rather scary, because as soon as you’re into immigration, suddenly it’s not straightforward . . . this has got a very strong political dimension.’

The press picked up on his first paper on DNA fingerprinting2, ’in particular The Guardian did a very nice write-up on it3.’ A lawyer, Sheona York, working at the UK’s Hammersmith and Fulham Community Law Centre, was shown the article, which had been spotted by a colleague. York wrote to Jeffreys asking for his help in an immigration dispute over a young Ghanian boy. The dispute was whether a person claiming to be his mother really was the boy’s mother. Blood group testing had shown that he was related in some way to her, but could not pin down the precise relationship.

Jeffreys admits it was a ’formidably tricky case. We didn’t have the father, there was some doubt about who the actual father was anyway, and we couldn’t get hold of the sisters [of the alleged mother] in Ghana. All we had was the mother and her three undisputed kids. So we took this on and I thought, there’s no way we’re going to get anything out of this . . . and I was just absolutely gobsmacked when the results came in. I personally did the analysis and it was just blindingly obvious that every single character in this boy’s DNA pattern . . . you could either see it in mum, or in one of the undisputed children as a paternal characteristic; there wasn’t a single thing amiss about this boy, everything matched.’ In other words, she was his mother.

A bit like fingerprints
The term ’DNA fingerprinting’ arose from a talk Jeffreys gave at Oxford University, UK, shortly after his discovery. After the talk, a friend, Nick Proudfoot, came up to him and said that his Southern blots looked like fingerprints. This comment led Jeffreys to coin the phrase that is now so familiar.

It was clear to Jeffreys from the outset that the original way of generating DNA fingerprints was not going to find common currency in forensic labs. Too much DNA was required, the patterns were too complicated, and it was too open to challenge in a court of law. Instead, rather than look at many different hypervariable regions at once with a universal probe, forensics labs would interrogate about half a dozen different hypervariable regions one at a time with separate, region-specific probes, to build up a profile of the individual concerned - hence the term DNA profiling. This modified method was used by Jeffreys and his team in the first ever murder investigation to employ DNA-based techniques. The results showed that the police had arrested the wrong man, not the first time DNA profiling has sprung a surprise. DNA profiling was also used in the O J Simpson case, and is effectively still in use today, although all the laborious individual interrogations can now be done in a single step, thanks to the polymerase chain reaction (see box, below).

Scene of the crime science
I asked Jeffreys whether DNA profiling had any drawbacks.

’Not drawbacks. The one significant limitation is that the technology is still quite cumbersome and still quite slow. If you’re a crime scene officer and you come across a blood stain at the scene of the crime, what you really want to do is test it at the scene of the crime, particularly if you’ve got a very good database, and do comparisons.’ He mused about a little hand-held device that could be taken out to the crime scene, rather than have the crime scene brought to the lab, as happens now.

So, how long does it take to acquire a single profile? ’Going flat out, I think you would probably get it in four to five hours from receipt of the sample. A good example of that was [the capture of] Saddam Hussein. The period from him being discovered in a hole in the ground to the Iraqi authorities saying, ’Yes, DNA testing has confirmed this is Saddam Hussein’ was 16 hours.’ When asked where the reference DNA for the test would have come from, Jeffreys replied: ’I would be amazed if they didn’t have DNA from his two sons.’

Returning to the original question, he suggested that in theory the 4-5 hour acquisition time could be whittled down to 5-10 minutes, which seemed very impressive to me. He agreed that it was, but pointed out that even this improved acquisition time precluded the use of DNA profiles in smart cards to replace personal identification numbers. He said that people would get pretty fed-up waiting around for 10 minutes at the supermarket checkout while they tried to match the DNA profile from saliva (obtained by ’a quick lick on a pad’) with the profile on the smart card. However, he concluded that it was not impossible to reduce the acquisition time to a matter of seconds with new developments.

A national DNA database
The National Criminal Intelligence DNA database has been running for nine years and, according to Jeffreys, ’it works brilliantly. There are over two million people on it - most of the habitual offenders in the UK have now been databased - talk to any policeman and they’ll say it’s one of the best tools they’ve ever had in the fight against crime - probably the best tool they’ve ever had. But the problem is that round about half, probably a little more, of all crime scenes where you have DNA evidence, that DNA evidence doesn’t match anybody in the database. So the offender has no record. And so the question is, how do we get around that?

’There are two broad approaches,’ Jeffreys continued. ’The first is to try and tease out some information on the physical appearance of the person from the crime scene DNA. So it could be gender, ethnic origin, hair colour, eye colour, facial features and so on and so forth. The problem with that approach is that we’re barely at the starting line of understanding how genetic variation at the level of DNA influences things like facial features. So we’ve got nothing to test. And it might well be that even if that problem were solved, the interaction between variation in DNA and facial features was so complex that there would be no simple diagnostic test to come out of it. The other concern I have is that that would give the police the ability to look at markers which were actually rather important to you as an individual. I’d be uncomfortable about that. It’s an issue of genetic privacy.

’The alternative, which is scientifically more robust and will work, is simply to throw everybody on a database. I would be in favour of that, with a number of provisos and with very stringent safe-guards. The first is that the database shouldn’t be held by the police. That sends out entirely the wrong signals. I would see instead a database being held by a separate agency - equivalent to good old Somerset House [where the UK’s General Register Office was housed until 1970] - where they would hold your DNA profile as essentially a certificate of identity. And that would be of use to you as an individual. For example, if you were walking home and were knocked down by a car and hadn’t got any ID on you, and you were taken to hospital . . . who are you? With a national DNA database, you can do a quick DNA test: that’s who you are. Abandoned babies . . . lost babies . . . there’s a whole host of applications in addition to the straight criminal investigation. The key question is how you interface the two databases, the criminal one with the national one. And that would need to be thought through very carefully. So I’m not recommending that the police be given carte blanche to trawl through everybody. There have to be very secure safeguards.’

Jeffreys stressed that only DNA profiles (never actual DNA) should be stored on the database, that only 15 markers should ever be analysed, and that these markers should be chosen such that they were not associated with disease. He felt that there was a risk that ’some hypothetical future malign government . . . could plough into the national database [of actual DNA] under a GATTACA-like scenario to find bits of DNA that control lifespan and all the rest of it. If you hold only the profiles, that’s not an issue.’

Living for tomorrow
Jeffreys’s current work involves analysing large quantities of human semen, not for forensic purposes - he took the view back in the early 1990s that the basic science of DNA identification had been solved - but to answer some fundamental questions about how DNA changes on transmission between generations. ’The fact that we can analyse DNA in a single sperm [and pan through up to a million sperm in one day] gives us a tremendous route into doing that.’ Semen is obtained from volunteers from his department and from surplus-to-requirement samples from local fertility clinics.

This current project stems from an exciting feature that appeared in DNA fingerprints from the beginning. In general, the bands in an individual’s DNA fingerprint arise either from their mother or their father. Occasionally, however, a band arises that cannot be traced to either parent. A parentless band is in fact the result of a mutation in the hypervariable region, causing it to expand or contract in size. Unlike RFLPs, this type of mutation is quite common - ’even within small-sized families you can occasionally see mutations popping up’.

Jeffreys and his team asked themselves why these hypervariable regions were so volatile. The answer, it seems, is that they occur at the same sites at which chromosomal transactions take place during sperm cell production, the very same transactions that lead to variation in human DNA.

Is this a coincidence? Jeffreys believes it is: the hypervariable DNA is merely a passenger, or, as he later put it, a parasite. Thus a mutation arises in a hypervariable region when it latches onto the machinery that orchestrates the chromosomal interchanges, and goes ’haywire’.

Addressing this problem has led Jeffreys to explore how the sites for these interchanges are distributed along the chromosome. Intriguingly, he says that ’the traditional picture of [chromosomal exchange points] being randomly scattered down human chromosomes is absolutely incorrect. All the evidence we have now [from single sperm analysis] says that [exchange] is a highly targeted process . . . almost all events land at what we call hotspots.’

Even more intriguingly, Jeffreys says, ’the segments of DNA between hotspots are behaving almost as if they are not even in a sexual organism . . . they’re not being scrambled up by [exchange]. So a given region will have just a single ancestor going way back, not multiple ancestors through [exchange].’

Now the questions for Jeffreys are far broader. For instance, he would like to know how the hot spot distributions and exchange processes impact on patterns of human DNA diversity.

Curiously, Jeffreys has not received a Nobel Prize for discovering DNA fingerprinting. One possible reason for this is the commercial side of the work. Almost immediately after Jeffreys realised the potential of DNA fingerprinting, he went, via the MRC [Medical Research Council], who funded part of his work, to the British Technology Group, who promptly sent a patent agent to Leicester ’to draw up some extremely good specifications’. At least half a dozen patents have come out of Jeffreys’s fingerprinting work; most of them are owned by the Lister Institute of Preventive Medicine, Hertfordshire, UK. (Jeffreys was a Lister Institute Research Fellow when he discovered DNA fingerprinting. He generously cites the additional time this fellowship allowed him in the lab as being pivotal to the discovery.) But in any case, Nobel Prize or no Nobel Prize, one hopes that Jeffreys continues to ’mess with mother Nature [in order] to find out some of her secrets’.

Acknowledgements

Thanks to Yogesh Prasad of Applied Biosystems for supplying DNA profiling data, Pooja Kumar for comments on the manuscript and of course Alec Jeffreys for agreeing to be interviewed.