Interview The Eureka Moment: An Interview with Sir Alec Jeffreys

Jane Gitschier* Department of Medicine and Pediatrics, University of California San Francisco, San Francisco, California, United States of America

In 1984, while tracking the veins of globin gene evolution and panning the for hypervariable linkage markers, Sir Alec Jeffreys accidentally struck gold—he discovered a way to identify any human being by a DNA ‘‘fingerprint’’. To use Jeffreys’ words, he has been ‘‘branded’’ by DNA fingerprint- ing, but he delights in its application and the hook it provides for public curiosity about science. Like Jeffreys himself, I wanted to dig below the surface of this discovery as well as that of another genetic nugget—the intervening sequence—found as a post-doctoral fellow seven years earlier. On the heels of my interview with Adrian Bird (published in the October issue of PLoS ), I made my way to Jeffreys through another branch of the British Rail system. When I arrived at his building on the leafy campus about 45 minutes early for our appoint- ment, his assistant suggested I get a cup of coffee while Jeffreys finished his experi- ment. I certainly wouldn’t have needed one. Jeffreys (Image 1) is an animated speaker, with a resonant voice and a rapid delivery of succinct clauses strung together in run-on sentences. His story could have cut through anyone’s jet lag. Gitschier: I didn’t realize that you still work in the laboratory. Jeffreys: I certainly do! Gitschier: Tell me about the experi- ment you were just doing. Jeffreys: Right, well, we won’t go into the gory details. Copy number variation [CNV] in the human genome is a real hot topic at the moment. Gitschier: The kind of variations people are looking for in association with autism and psychiatric diseases. Image 1. Sir Alec Jeffreys. Jeffreys: That’s exactly right. It’s a doi:10.1371/journal.pgen.1000765.g001 common phenomenon, and we’ve actu- ally known that for decades. What we’re around the place in forensics, and return- novo copy number variation in the fetal c- doing is going back to some of the ing back to my first love. The experiment globin genes at the single molecule level in absolutely classic examples of CNV. I’m doing at the moment is looking at de both somatic and germline DNA. These are in my favorite gene family— the globin genes—and that’s where I cut Citation: Gitschier J (2009) The Eureka Moment: An Interview with Sir Alec Jeffreys. PLoS Genet 5(12): my scientific teeth. e1000765. doi:10.1371/journal.pgen.1000765 Gitschier: We’re going to be coming Published December 11, 2009 back to that! Copyright: ß 2009 Jane Gitschier. This is an open-access article distributed under the terms of the Creative Jeffreys: Right. So, what I’ve done in Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, my scientific career is this gigantic circle, provided the original author and source are credited. starting off in globin genes, going all * E-mail: [email protected]

PLoS Genetics | www.plosgenetics.org 1 December 2009 | Volume 5 | Issue 12 | e1000765 All of this comes out of my work on Jeffreys: Right. And the penny During that experiment, we had to recombination hotspots. And the general dropped then, that this was going to be develop methods for monitoring purifica- feeling was that recombination hotspots the way forward. tion, and the only way we could see to do function at meiosis—they drive allelic I get to Amsterdam, and Piet said, you that was to use Ed Southern’s blotting recombination, and they may well drive canworkonthisifyoulike,butyoumight technique, which at that point was only a ectopic recombination. also like to have a chat with this guy Dick year or two old. Gitschier: Define ‘‘ectopic.’’ Flavell, he’s got a collaboration with So, as we purified the DNA we could Jeffreys: The term ectopic originally Charlie Weissman in Zurich, on trying monitor the fractions just by running them came from yeast and it applied there to a to isolate a mammalian gene. And I out on an agarose gel, doing the Southern situation in which you have a sequence thought—whoa! That’s sounds really blot and then hybridizing with an appro- repeated, say here and there, so that they exciting. The idea of the project was to priate complementary probe. And that not can undergo unequal crossover and cause get to a single-copy gene. No one had ever only worked, but we could actually see the duplication and deletion. ‘‘Ectopic’’ re- done that in a mammalian system. The fragment of DNA we were trying to purify combination means it’s ‘‘out of place.’’ only one we could possibly do, we felt, in the starting EcoRI digest of genomic As of yesterday, I found there is copy was either rabbit a-orrabbitb-globin, DNA. number instability not just in the germline, because the mRNA had been purified. Gitschier: Hadn’t he shown that but in somatic DNA. That largely rules Thegeneisolationwouldbebyphysical before? out meiosis and meiotic recombination purification. Jeffreys: No, Ed was desperately hotspots. Even in the germline it is quite Gitschier: No cloning? trying to get this going. I know Ed clear that the substantial proportion, Jeffreys: Well, cloning came in right at extremely well, and there was a bit of possibly the great majority of rearrange- the end. It simply wasn’t around at the discomfort on my part thinking that we ments, are again pre-meiotic, arising time. It was by hybridization enrichment had trampled on his patch. On the other during germ cell development. We’re with prodigious quantities of DNA [from hand, that is what we needed to do. trying to drill down below the applied rabbit liver]. The experiment was to cut it Having got the ability to detect down to genetics [looking for variation associated up with EcoRI restriction enzyme. Re- the single gene level, we thought we should with disease], to some of the fundamental member, this is back in the days when you see if we could make a restriction map mechanisms, to understand the dynamics couldn’t just buy enzymes off the shelf, you around the gene, which is what we did. of rearrangements in the human genome. had to make them. Gitschier: Were there EcoRI sites in So, if you want to put a simple summary Then denature the DNA and hybridize it the cDNA? on what this lab is about, it’s about human to globin mRNA. This was a two-pronged Jeffreys: No. The cDNA had been DNA diversity and the processes that attack. In Amsterdam we were going to use cloned by Tom Maniatis, and we pretty generate it. the mRNA to pull out the complementary quickly moved over to using his rabbit b- Gitschier: OK! Now let’s get to the strand, heavily enriched, and in Zurich, globin cDNA that he very generously first question on my list, which indeed is Charlie Weissman had managed to make a provided to act as a probe for monitoring. about globin. It’s about the period of your cDNA so he could pull out the other strand, We just wanted to check that everything post-doc in Amsterdam. Why did you go and the idea was to purify our complemen- was OK. And we built up a restriction there and why work with Flavell? tary strands and then meet somewhere in the map around it [on genomic DNA via Jeffreys: OK. I did my D. Phil. at middle to hybridize the two stands back Southern blotting]. University on human somatic cell together. Then, because this was an EcoRI We then discovered that there was an genetics. Then went to a Biochemical fragment, we could then pop it into a vector EcoRI site right smack bang in the middle Society meeting and chatted with a chap that we hoped someone was about to of the gene! [That meant] our enrichment named , a very senior scientist, develop. experiment was a total disaster, because who at the end said, if you are interested in Gitschier: How were you selecting the we would have purified one end of the doing a post-doc with me, just let me mRNA? gene in Amsterdam, and in Zurich, they know. Jeffreys: We were selecting by attach- would have purified the other end of the And I thought, that’s great, 9cause I ing mercury to the RNA and then gene, and to put them together, there wanted to get out of Oxford, and Holland I capturing it on a thiol column. would be nothing. The flop of the really fancied because the language wasn’t Gitschier: That’s a dangerous millennium that was! going to be a problem; everyone speaks experiment. But, the question then was, what the English. So, I got myself an EMBO fellow- Jeffreys: Oh, the whole thing was hell is the EcoRI site doing in the middle ship to work with Piet on yeast tRNA genes. horrendous. We were using radioactive of the gene? And then we started to do In 1975, the door was clearly opening mercury. more and more fine-mapping and it was on molecular genetics, before that, it Gitschier: But hold on. Since there clear there was a huge gap in the gene. wasn’t worth talking about. was no reason to suspect that there were I remember sitting down with my Dutch Gitschier: Expand on that statement. intervening sequences, what is the point of technician, saying we’ve got the restriction Jeffreys: I remember very clearly. going after the gene? mapping data, let’s try putting all this There was a colleague of mine at Oxford Jeffreys: Nobody had ever seen a together. And I knew it was just nuts, but I called David Finnegan and we’re waiting mammalian gene. No one had any idea thought we could solve it if we just put an in the lunch queue and he wanted to go off of what it would look like. extra dollop of DNA inside the gene. All of to the States, and I said what’s the project, Gitschier: So, the idea was to get this was done without reference to Phil and he said the idea is to take Drosophila something bigger than the mRNA itself. Sharp and Rich Roberts’s work with DNA and to try to stick bits of that into Jeffreys: Yes, that’s right. To look at adeno [which was happening at the same lambda phage. the flanking regions. Basic academic time]. I knew instinctively that this was Gitschier: With David Hogness? curiosity. something pretty exciting. And then, Dick

PLoS Genetics | www.plosgenetics.org 2 December 2009 | Volume 5 | Issue 12 | e1000765 was over for, I think, a Cold Spring So, that was our first quest. By early moving out, taking DNA with it and Harbor Meeting, and everything started 1978 we had picked up our first RFLP, a creating this length variation. But, I read falling together. About the same time, Phil rare variant in a single individual. Again, their interpretation of transposition and I Leder managed to clone in lambda the these were in the globin gene clusters, just felt not so sure about that. So, we then mouse b-globin gene and showed by because again, these were the only genes started redoubling our efforts and still electron analysis that there for which probes existed at that time. getting nowhere at all. was additional sequence inside the gene. Really excited, but we got pipped to the Then Graeme Bell described the se- But the trouble there was it had been post because Kan and Dozy published quence of the human insulin gene and cloned in E. coli and perhaps it was an their RFLP and the association with sickle right next door to it was a —a insertion sequence. And then [there was] cell disease. highly variable tandem repeat region. And Chambon’s ovalbumin gene story. Gitschier: I think they just bumped then Doug Higgs in the a-globin region. Looking back on it, basically in 1977, into that discovery. Gitschier: What approach were you introns were going to be discovered. Full Jeffreys: What we had done was to do using to try to find these variable stop. The technology had arrived to the a fairly systematic survey for RFLPs in the ? point where the discovery was inevitable. I b-gene cluster. Jeffreys: It was primarily physical think all of us in the field were grateful that Gitschier: What made you think that enrichments. These sequences might have we just happened to be at the right place at there would be variation in restriction sites unusually fast reannealing kinetics, so you the right time. among people? could do a COT approach. Or, since these When it was time to leave Amsterdam, Jeffreys: I can’t remember. It seemed sequences might be quite long but consist- one possibility was to do a post-doc with fairly obvious at the time. I knew enough ed of repeats over and over again they Ed Southern up in . He’s a human genetics to know that there must would tend to be resistant to restriction great guy and the stuff he was doing was be a significant amount of variation in enzymes, so, if you took a load of common fantastic. He’s one of my heroes. We are DNA sequence. I’d been brought up in the cutting restriction enzymes, you would actually quite similar. We like fiddling days of serology and biochemical genetics, leave these things intact. around with things. He gave this wonder- enzyme polymorphisms, and we knew that We were still getting nowhere. But ful quote a few years ago that he misses the that was sampling only a tiny proportion meanwhile [in a separate project], we days when he could get at the data before of all diversity in the genome. So, if there is were doing some globin gene family the computer did. diversity, then it will be agnostic with evolution work. We thought, OK there is But at the same time, I thought I’d like to respect to restriction sites, so if you luck a missing gene in the story, and that is try running my own lab, and out of the blue out, you’ll find a polymorphism that hits a myoglobin. Could we get the myoglobin came a phone call from this guy called Bob restriction site and that makes it assayable. gene out and see how it fitted in to the Pritchard who founded this Department [of Having come up with these RFLPs, we hemoglobin gene family as a very diverged Genetics] in the early 60s. He said, ‘‘Would then got fed up with them, cause everyone member of that family? you be interesting in coming for an inter- was doing it. So, we then started thinking So, this is really the start of the DNA view?’’ I said, ‘‘Where is it?’’ He said, that surely in this enormous human fingerprint story, because we got the ‘‘Leicester.’’ And I said, ‘‘That will be fine.’’ genome, there must be bits of DNA that human myoblobin gene and found a I put the phone down and I said, are more variable than these RFLPs, and minisatellite inside the intron. ‘‘Where the hell is Leicester?’’ All these we thought intuitively that the right place Gitschier: How did you find that? Dutch people were running around trying to look was tandem repeat DNA. I’ve been Jeffreys: By sequencing. It wasn’t to find a map of Europe. brought up in the school of satellite , variable between people, but I realized I Gitschier: Pre-internet. which was the only class of DNA you had seen this sequence somewhere else. Jeffreys: Pre-everything! These were could purify going back to the old cesium So, I went back and looked at the a-globin the days if you wanted a sequence you had chloride density gradient days. The satel- and the insulin minisatellites, and you to get out a typewriter and type it in. lite DNAs incidentally show a lot of could see this sort of vague suggestion that So, I visited Leicester and I immediately variability in copy number. there might be some sort of shared fell in love with the department. I came as I felt intuitively that if you had local sequence in there. So, we then took that a temporary lecturer, and I’m still here 32 tandem repeat sequences on a smaller myoglobin minisatellite and hybridized it years later, so it says something about the scale in the genome, they’d have potential to a human lambda library and lo and environment. I love it here. variation as well. The hypothesis was that behold a number of clones lit up. We then So, the question then was, what was I there may be bits of DNA with repeats, started systematically isolating those going to do? It was clear that carrying on maybe 10 or 20 bases long repeated 10 or clones, showed that they contained minis- with the intron work was not going to be 20 times, so we started all kinds of crazy atellites and some of them were pretty viable. Suddenly everybody was moving experiments trying to physically purify variable loci. into the field—evolution of introns, mech- these bits of DNA from the human Gitschier: So, you were checking this anisms of splicing, etc. I thought, take your genome. on a Southern blot? education in human genetics and your Then in 1980 Arlene Wyman and Ray Jeffreys: Southern blot and character- new-fangled molecular biology and stick White described the first hypervariable izing by sequencing. And, as we were them together. If you can pick up specific locus, so I thought WOW they do exist! building up the repeat sequences from the bits of human DNA, then you should be But their interpretation was one of trans- clones coming out of the library, the able to scan for variation. Variation that position. Why? Because they came from a shared sequence motif, the minisatellite affects a restriction enzyme site will transposable element background. So, core, became more and more obvious. It manifest as what is now called, I think quite reasonably, they were thinking, OK was a short sequence, about 15 bases long, very uglily, an RFLP [restriction fragment it’s hypervariable because we’ve got a embedded within the repeats of the length polymorphism]. transposable element that is moving in and minisatellites. It was almost as if this was

PLoS Genetics | www.plosgenetics.org 3 December 2009 | Volume 5 | Issue 12 | e1000765 some kind of sequence driving this tandem That was my golden moment. Without Jeffreys: We knew that DNA finger- repetition. But, more important, it could DNA, he could have been deported. prints were too insensitive for forensic give you a much more effective generic Gitschier: That set of events must casework. So, we simply took out the way of getting minisatellites out of the have built up momentum for you and your minisatellite core probe, we went back into genome, because rather than using this lab. our libraries of DNA and cloned out the crummy myoglobin probe, you take a Jeffreys: Oh, it did. We hadn’t most variable single locus probes, each of probe that consists of just this core realized how many thousands of other which gave a simple but highly variable sequence repeated over and over again. people were trapped in these disputes! two-band pattern. We knew that was the So, we took that and hybridized it to a So, the next thing was a complete way forward. Southern blot, which happened to have avalanche of letters and phone calls; Gitschier: OK, back to the case. [DNA from] the lab technician and her people were turning up at my home! Jeffreys: The forensic samples arrived, mom and dad. We got this fuzzy splodgy Gitschier: What did you do? and I have to say that was a chilling mess, but the DNA fingerprint was Jeffreys: Well, I nearly had a nervous moment. An ordinary academic and absolutely obvious. We got a pattern like breakdown, but I kept going. It was an suddenly you’ve got murder samples in a fuzzy bar code. These patterns were insane two years, 1985–1987, before the front of you. I remember my blood individual specific, and seemed to be thing went commercial. We were the only literally running cold at that point. inherited within the family. That was a lab providing any testing at all. We put the first probe on, and the real eureka moment, because we were Gitschier: And then there is the local prime suspect wasn’t a match [with the suddenly onto something completely new, double rape/murder case in a village near semen sample from the second murder]! which was DNA-based identification. Leicester. I read ‘‘The Blooding.’’ Suddenly we were into the world of Recall, the driver for this experiment Jeffreys: It’s a good book. It’s exclusion, and how many probes do you was medical genetics. You needed these accurate. need for that? One. The result was so improved markers for facilitating construc- Gitschier: Ah, you’ve answered my wacky, so totally out of keeping from what tion of linkage maps of the human genome question. It depicted you as this chain- the police were expecting to see. We and helping in linkage analysis of inherited smoking guy in a black jumper [sweater]. thought better do another one [probe]. disease. This thing would have been useful What did you think about that? The results were totally astonishing, totally were it just a single location in the Jeffreys: Well, let me tell you a little genome, but the fact that there were story. overturned what the police had got fixed multiple copies of the repeat sequence in The author was , an in their minds about the guilt of this prime the genome gave it a new meaning, in ex-LA cop who happened to read about suspect. He was released. terms of DNA identification. the story in his dentist’s office in Hippocrates The police said, OK we now believe all When I talked about it in a Department magazine. He thought this is brilliant, and this DNA testing, let’s go and pan the seminar, and then speculated about what he took the plane straight over here and entire local community and see if we can we could use this for, like catching rapists interviewed all sorts of people, myself flush out the true murderer. That was all from semen—about a third of the audi- included. My secretary had written ‘‘in- done by Home Office forensic scientists, ence fell over laughing. It sounds bizarre terview with Rambo’’ on my calendar. No who at that point had our DNA finger- now because it’s so blindingly obvious that idea who he was. And I was very cautious. printing in place. But of those 5000 you can use DNA for this, but believe me, He arrived, and we did not get along samples, only 500 were DNA fingerprint- back in the 80s it was simply not there. terribly well, talking at cross-purposes. He ed. The others were all excluded by The only reason I came up with the idea wanted to dig as deep as he possibly [biochemical] testing. of DNA-based identification was that it could—that was his job as an author—and The upshot of that was that the true just hit you in the face! my instinct was to keep stum. perpetrator was flushed out, and the rest is So, within the first day, we saw It was an extraordinary case. We were history. identification, we could foresee forensic approached in 1986 by the police. They Gitschier: Have you been in any other analysis if DNA survived in forensic said, we’ve got these terrible double rape/ books? specimens, zygosity testing in twins, pater- murder cases, we have a prime suspect Jeffreys: I’ve certainly turned up in all nity testing, and immigration disputes. Just who has confessed to the second murder. sorts of science-y books. DNA fingerprint- like drawing up a shopping list—if we We’ve heard about this DNA fingerprint- ing is now part of the curriculum for kids could get this technology improved, what ing and could you use this technology, not age 14–15 in the UK. it could be applied to. to confirm his guilt with respect to the So, I’ve achieved that sort of rare status Gitschier: I clearly remember that second murder, we know that, but to have of science reaching out to the public and Nature paper [1985] involving the immigra- a look at the first murder and see if we can being understood by school kids. And tion dispute that you helped to settle, the case tie him in. literally every 2 or 3 days I get an email, where a boy was threatened with deportation So, I said we’ll do this, but I explained mainly from the States, from school kids because the immigration authorities alleged at the outset that we wouldn’t be using saying, ‘‘I’ve got to do a project on a he wasn’t the biological son. DNA fingerprinting, but we’d use this famous scientist, so I’ve chosen you,’’ and Jeffreys: That was the first DNA case derived technology DNA profiling, which I love that. I always respond. tackled anywhere in the world, and it is we thought would be much more appro- It’s great because if you think you are still my favorite case because I was there at priate. And we said, ‘‘Don’t hold your doing even the tiniest bit to switch people the tribunal where they dropped the case breath. No one has ever attempted this on to science, and this DNA stuff is against the boy, when the mother was before.’’ great—OJ Simpson, the Romanovs, it’s told—and just the look in that mother’s Gitschier: Tell me about profiling, got everybody. If you can’t hook people eyes! She had been fighting the case for what it means, and why you used it into science with that story, give up. two years. instead.

PLoS Genetics | www.plosgenetics.org 4 December 2009 | Volume 5 | Issue 12 | e1000765