Sex determination, how an embryo decides to become male or female, is such a fundamental process that it comes as a surprise that we had no clue how it worked until a little over half a century ago; in 1959, in a great example of scientific serendipity, the beginnings of the answer were born out of the boredom of a 24-year-old researcher in Edinburgh, Pat Jacobs.

Pat’s expertise was in cytogenetics, the analysis of chromosomes, and in the wake of the 1950s nuclear arms race, she was hired by a small MRC-funded group to compare the chromosomes from patients with radiation-induced leukaemias with those with normal leukaemia. It was a laudable project, but unfortunately, Pat couldn’t find a single example of a radiation-induced leukaemia and had to mark time making a larger and larger control sample of normal leukaemic chromosome spreads. Therefore, when her boss Michael Court Brown introduced her to John Strong, an endocrinologist who wanted some chromosome spreads performed from one of his patients, she jumped at the chance to break the monotony. Strong’s patient had been classified as suffering from Klinefelter’s syndrome, a sex-reversal disorder then held to result from a female with two X chromosomes somehow becoming a male. However, Pat noticed something rather odd when she looked at the chromosome spreads; there were definitely two X chromosomes, but instead of the normal human count of 46 chromosomes, there were 47. Even more weirdly, the interloper was small, and to Pat’s practised eye, looked an awful lot like a Y chromosome. Because the Y chromosome was popularly held to be a piece of functionless junk, this seemed odd, but analysis of other Klinefelter’s patients yielded the same result—all had two X chromosomes and an additional Y.

In the 30-odd years between the Y being fingered as the source of the Testis Determining Factor, or TDF (developed testes being the physiological criteria for maleness), and the discovery of what exactly on the Y was responsible, science had to cover a lot of ground. In 1959, genetics, the study of how characteristics are inherited, had accumulated a venerable history, but the gap between the statistically derived maps of linked genetic loci, based on the frequency with which two loci were inherited together or separately, and the physical reality of the chromosomes, where the genes corresponding to the loci were actually located, was enormous. How did one superimpose the abstract linkage maps onto the DNA? Those working in the field of human genetics had an additional problem, because they were unable to set up the breeding experiments that were the lifeblood of genetic research in lower organisms, and could only make very basic linkage maps based on a few family pedigrees.

Guido Pontecorvo.

(Photograph courtesy of CRUK London Research Institute Archives.)

Sorieul, Cornefert, and Barski’s method for somatic cell fusion was based on tissue culture. They reported that if they mixed two different sorts of cells together, the resulting mixed cultures contained cells with characteristics of both parental types. With Boris Ephrussi, Sorieul showed in 1962 that the cells had in effect “mated”—their chromosomes had mixed together and then resegregated to create offspring carrying a genetic contribution from each parent. Most excitingly for geneticists, in 1967, Mary Weiss and Howard Green found that if mouse cells were mixed with human cells, the battle of which chromosomes were retained by the hybrid interspecies offspring was almost invariably won by the mouse chromosomes.

With the advent of somatic cell hybrids, as they were called, it was possible to derive cell lines whose chromosome complement included just one human chromosome, with all the rest being mouse. These could then be used to map loci to a particular chromosome, as long as one had an assay for the locus of interest. Panels of cell lines carrying different human chromosomes could be screened for the presence or absence of a human gene product, and hence its parental gene. Experiments were relatively simple: Set up your assay, whatever it was, on the 20 or so cell lines, each with their single human chromosome, and see which line comes up positive. Check which chromosome is in that cell line, and bingo! You’ve mapped another gene. Combined with the fact that breeding was no longer necessary, experiments took days rather than years, and in the decade following, somatic cell hybrids were used to map a few hundred genes with known and assayable gene products onto specific human chromosomes, including the X.

After this and other equally misguided attempts to find a gene, any gene, on the Y, the more enlightened amongst the field realised that to identify the genes on the Y chromosome and to find which of them was TDF, was clearly going to need a different technique, one that was independent of what the genes encoded; in other words, the genes would have to be mapped simply by their presence on the Y, not what they did. And, at the end of the 1970s, a suitable mapping technique appeared, provided, inevitably, by the fertile minds of molecular biology.

Following on from the idea of using RFLPs as a way of distinguishing between different genomes, the one extra conceptual step taken by Bodmer and Botstein and their colleagues was to realise that RFLP technology could also be used within a single genome to create an ordered map. By comparing the restriction enzyme patterns made by digesting all of the clones in a DNA library made from just one genome, shared patterns, indicating overlaps between different clones, could be detected. Rather like piecing together a panoramic photograph from individual overlapping shots, clones could then be put in order. This simple but clever idea was the seed from which the Human Genome Project grew, culminating in the publication in 2001 of a working draft of the complete DNA sequence of the human genome.

But what of the Y? The advent of positional cloning techniques fired the starting gun in the real race to find TDF, and the runners and riders were an interesting bunch. Two in particular seemed to be worth betting on: in London at the ICRF, Peter Goodfellow, a former Bodmer graduate student recently returned from Stanford, had moved onto the Fifth Floor and was looking for a project with which to make his name; and in Boston, David Page, a newly qualified MD and protégé of Bodmer’s RFLP rival David Botstein, was just starting his own lab at the newly formed Whitehead Institute.

Goodfellow and Page are an interesting pair, about as different in personality and appearance as could be imagined, although their origins have some similarities: Page was born in rural Pennsylvania, Goodfellow in the depths of the East Anglian Fens, and both were the first in their families to go to university, Page to Swarthmore, and Goodfellow to Bristol. After that, things begin to diverge.

Peter’s arrival at ICRF was something of a shock. Scruffily dressed, rake thin, trailing a cohort of lab members whose dress sense showcased the “punk meets new romantic” vibe of the time, he was a bit of a novelty to some of the crustier members of staff, to put it mildly. The wild-looking new lab head was in addition highly driven, mostly by an intense curiosity regarding science, but also by a fierce ambition to succeed. Combined with what Peter describes as a slight social autism, this drive meant that he was a ruthless adversary in a scientific argument, taking no prisoners and never tolerating mediocrity. However, these sometimes intimidating qualities masked a person entirely lacking in malice, with an almost palpable delight in having found something he loved on which to spend his working life. His scientific persona, combined with amongst other things his penchant for poetry (good) and football (perhaps less good, given his allegiance to the Arsenal team of the 1980s), made Peter a complex, brilliant man who was a delight to his friends and a red rag to his enemies.

Peter Goodfellow, ca. 1987. (Photograph courtesy of Viesturs Simanis.)

Back in Boston, David Page was having a bit of a crisis of self-confidence. His position as a Research Fellow of the Whitehead Institute was entirely novel, as was the Institute itself. Jack Whitehead, its benefactor, had got together with Nobelist David Baltimore with the intention of producing a kind of artists’ colony for scientists. They wanted to attract the best possible people and give them the resources and intellectual freedom to be as wildly creative as they wished. They were clearly onto a winner because they managed to recruit the five founding members of the Whitehead from the top echelons of molecular biology and genetics. Plunged into this boiling intellectual soup, Page, a newly qualified MD with only a couple of years’ lab experience, was simultaneously exhilarated by the exalted company and terrified that they would dismiss him as an intellectual pygmy. To his further alarm, his expectations of being a kind of glorified postdoctoral fellow with Rudolf Jaenisch, a world leader in the new field of making genetically modified mice, were confounded when it became clear that he was expected to be working independently. Page had to perform a hasty mental regrouping and decided, rather like Peter Goodfellow a few years before, to stick with what he knew and was good at. In Botstein’s lab, he had started a collaboration with Finnish geneticist Albert de la Chapelle—who in 1964 had described the first cases of sex-reversed XX males—to explore how RFLP probes could be used to study such disorders. It seemed natural to carry on with the collaboration and, in keeping with the ethos of the Whitehead, to aim high. Page set about finding a research assistant with a capacity for infinite hard work; he too, had decided to clone the Testis Determining Factor.

In 1986, the two labs published the first deletion map of part of the Y chromosome, showing that TDF had to be in a region of the short arm of the Y that they called Interval 1. Interval 1 fulfilled all of the criteria above—it was there in all of the XX males with additional Y DNA, and furthermore, it was within a larger region missing in an XY female. Interval 1, however, was not small; it still covered ∼3 Mb.

Peter Goodfellow’s lab had also been mapping busily, but using different methods. Rather than relying on patient material, which was very hard to come by in London for various annoying nonscientific reasons related to the egos of the clinicians involved, they had gone for a combination of traditional and cutting-edge molecular biology techniques, based around the MIC2Y gene. MIC2Y and its partner MIC2X lie in the pseudo-autosomal regions of the sex chromosomes, meaning that MIC2Y was at the tip of Yp, in the region next door to where TDF would be found. By meiotic mapping, looking at how often pseudo-autosomal genes were able to switch between the X and Y chromosomes during the chromosomal do-se-do of meiotic recombination, the Goodfellow lab realised that MIC2Y hardly ever switched, in contrast to the rest of their pseudo-autosomal-specific markers, which changed chromosomes with happy abandon. This reluctance to switch placed MIC2Y very close to the boundary between the pseudo-autosomal region and the Y-specific region next door, and therefore made it a very good point of departure from which to start mapping the rest of Yp, and TDF.

Chromosome walking in those days was extraordinarily tedious. You began with a probe a few thousand base pairs long from the known anchor point on the chromosome, which in Goodfellow’s case was a clone from MIC2Y, and in Page’s the first clone he’d ever isolated in Botstein’s lab, which flanked the other end of Interval 1, towards the centromere. Using a library of Y-specific sequences, in which each library clone contained at most 30,000 bp of Y DNA, you’d hybridise the anchor probe to the library. Any clones lit up by the probe were picked and grown up to amplify the DNA they contained, and then the DNA was restriction-digested along with the anchor clone. You kept looking until you found a clone that had some digested bands in common with your anchor, and then, you had an overlap. After that, you started the process all over again, using as the new probe the region of your new clone that was farthest from the anchor clone. If you were lucky, you did this 50–100 times and ended up with a beautifully detailed panoramic photo of your region of interest.

The Page and Goodfellow labs were not lucky. During their previous low-resolution mapping adventures, it had become very clear that the Y chromosome was going to be a complete pain to analyse: Almost all of the probes isolated, instead of recognising one nice, clean band on a Southern blot corresponding to one locus, saw multiple bands; the Y was covered in repetitive sequences, some of which were also found on autosomal (nonsex chromosome) and X-chromosome DNA, and some of which cropped up several times on the Y itself. This was a mapping nightmare; if the overlap clone that you found was full of repetitive sequences, it was useless, because it could be from anywhere. In Peter’s words: “We spent two years doing the longest walk …: Clone, subclone, pull out a probe, find something repetitive, go back and start again, chunk by chunk by chunk. It took four to five people two years … pretty brutal work.”

In Page’s lab, things were moving equally slowly, perhaps more so, because at that point the Page laboratory comprised just two people: David himself and his research assistant, Laura Brown, with occasional help from Harvard undergraduates with research project assignments (their Y genomic library was made by one of these, Jonathan Pollack, now a professor at Stanford). They evolved a system in which Laura worked the day shift and David worked nights, so the lab was a 24-hour operation. (Fortunately for his future happiness, David’s wife-to-be was a dermatology resident in Toronto at that point, and thus remained blissfully ignorant of her boyfriend’s Stakhanovite tendencies until she moved to Boston some time later.) David and Laura’s task was made no easier by the fact that although David knew where to start, he didn’t know which direction to take, and therefore had to try both ways. His probe was in a region of the Y with homology to the X, and only when he started to find sequences not hybridising to the X would he know that he was going the right way, into Yp. In the summer of 1986, he came into the lab for his night shift and realised from looking at Laura’s data that they had finally walked into the Y-specific sequences, and knew which way to go. Before heading off for a rare holiday the next morning, he left a note for Laura, picking up on that summer’s hottest MTV hit from Aerosmith/RunDMC; in an unlikely collision of heavy metal, rap, and science, Laura came in the next day to find a piece of paper with “WALK THIS WAY!” written on it.

By late 1986, Page and Laura Brown had narrowed down the part of Interval 1 carrying TDF to Interval 1A, adjoining the pseudo-autosomal tip of Yp; by studying more sex-reversed XX male DNA samples, they had found that this was the smallest part of the Y chromosome that could still specify maleness. Crucially, they also had a very interesting XY female DNA sample, WHT1013. WHT1013 had the misfortune to have had her father’s Y chromosome accidentally mixed up with autosomal chromosome 22 (formally, a reciprocal Y;22 translocation), meaning that she had apparently acquired the whole of Yp except for Interval 1A2, a subregion of Interval 1A, and a region next door to it, Interval 1B, which was known from the XX males not to be necessary for maleness. Interval 1A2 was 140,000 bp long, and all of the data suggested that TDF had to be in it somewhere.

The point at which you know you have something really important, but you still have to do the confirmatory experiments and then get the paper published before you get scooped, is just excruciating. Nothing can ever be done fast enough, nothing can allay the gnawing anxiety that even as you slog away, trying to be careful but trying to be quick, someone else is doing the same work, is maybe a little closer than you.… You don’t sleep properly, you don’t do anything other than go to work, you neglect your nearest and dearest, and the whole time, a little kernel of panic in your gut nags away like a physical pain. Therefore, it was hardly a surprise that the Page lab went into overdrive. Fortunately for David, the spring of 1987 had seen the arrival of his first tranche of postdocs, all attracted by his growing reputation in the field. They all set to work with a vengeance, blotting, mapping, sequencing, and checking that everything was right, that there was no possibility of error. And they got there. In November, they sent a paper to Cell entitled, “The Sex-Determining Region of the Human Y Chromosome Encodes a Finger Protein,” and on Christmas Eve, 1987, the paper was published.

After such an event, it takes a while for a lab to pick itself up, dust itself down, and think what to do next, and the Goodfellow lab was no exception. Peter spent the next few months half-heartedly doing experiments, reading a lot of poetry, and drinking too much coffee. His postdoc Paul Goodfellow did manage to cheer him up slightly by pointing out to him that there were an awful lot of people in the world who would like to be failures like them, sitting in a well-funded lab full of highly talented scientists, but moving on to the next thing was always going to be hard. Matters were made no easier by the world’s press; with alarming eagerness, they had latched onto the story of how one gene can make the difference between becoming male or staying female, meaning that Peter had to dole out quote after quote saying how important the work was. Even worse, scenting a great human interest story, the press reframed the whole story as a big transatlantic fight to the finish, won by the quiet American, lost by the flamboyant Englishman, completely ignoring the tedious reality that science and, indeed, life is never that simple. Early 1988 was not a good time.

Curiously, David Page was not having quite as much fun as he should have been either. With the announcement of the cloning of ZFY, he had been catapulted from being a medium-successful scientist with a growing reputation amongst his immediate circle into a media phenomenon. He was on the NBC News, the front page of The New York Times, and many other papers worldwide, and even made it into a local paper in the Caribbean, much to the bemusement of his postdoc Lizzy Fisher, taking a well-earned Christmas holiday there in the hope that she could keep ZFY out of her life for a short while. He didn’t like it at all. The nature of the attention meant that some of his scientific colleagues, whose opinions of him had been fairly neutral up to this point, felt obliged to become more extreme; although his close colleagues were delighted at his success, many others prone to jealousy began to snipe at him.

Back in the relative sanity of the laboratories, investigators in the sex determination field were reformatting their ideas to take into account Page’s new information and beginning to climb the next experimental mountain. Contrary to what the world’s press had decided, ZFY was still only a candidate for the Testis Determining Factor; in human genetics, a gene cloned by reverse genetics can only be definitively assigned to a condition when it has passed all reasonable scientific tests, and ZFY now had to take those tests. The first, that it should be in the sex-determining part of the Y chromosome, was clearly fine, and that it appeared to be the only gene in a rather empty area of DNA was also in its favour. Whilst labs around the world got hold of ZFY probes and tested them on their own sex-reversed patient samples to make sure that the equations ZFY = male, no ZFY = female held in all cases, other experiments also got under way, but this time not in humans. Sex determination as a problem of developmental biology had long been of interest, and, unsurprisingly, the mouse model people were very interested in checking out ZFY in their favourite furry organism.

Based on the idea that TDF in humans and mice was likely to be conserved, Robin and Peter decided in 1985 at the start of their collaboration that a good strategy to clone TDF might be to try to express the human gene in mice, and hope to change female mice into males. Unfortunately, the idea was so far ahead of its time that it failed; to get the mystery gene into mice involved putting enormous chunks of the human Y chromosome into ES cells, selecting the cells that were expressing the MIC2Y antigen, and then making mice with the MIC2Y-positive cells. Because virtually every single technique in the process had to be made up as they went along, there were just too many obstacles to success.

Undaunted, Robin got talking to Liz Robertson, another transgenic/ES cell pioneer, and came up with another idea, to mutate mouse Tdy (by convention, and as not detailed in the previous chapter, gene names in humans are all capitalised, and those in mice appear as proper nouns; to further confuse matters, “TDF” is referred to as “Tdy” in mice) by hopping retroviral DNA into mouse ES cell genomes. If the retrovirus hopped into the Tdy coding sequence, it would disrupt the ability of the gene to encode Tdy protein, and therefore any mouse made from the mutant ES cell would be female irrespective of its genetic makeup. Slightly amazingly, this idea actually worked, giving rise to an XY female mouse that was even able to breed, albeit with huge difficulty. Unfortunately, when the DNA from this animal was analysed, Robin and Liz were unable to find any trace of retroviral DNA, which they needed as a flag to mark the position of the Tdy gene, so they were left in the teeth-gnashingly frustrating position of having mutated the Tdy gene but being no further on towards finding out what it was.

What Robin’s lab found regarding mouse Zfy was rather unexpected. There were actually four, not two, copies of Zfy in the mouse genome, two on the Y, one on the X, and one on an autosome, making it very hard to see how the gene could be functioning as a Y-chromosome-specific sex-determining factor, unless the prevailing ideas regarding how sex was determined were radically wrong. Because this was actually perfectly possible, given how little was known regarding how Zfy might work, this did not constitute solid evidence for the prosecution, but the lab’s next finding was a bit more worrying for Zfy supporters. However hard the lab tried, they were unable to find any expression at all of Zfy2 in foetal gonads, only in adult testes, and although Zfy1 was detected, it was only on in the germ cells, which play no part in testis determination; in the crucial somatic cells of the foetal gonad, there was no sign of either gene. The clincher consigning Zfy being mouse Tdy to the bin was that in Robin and Liz’s Tdy ^m1 sex-reversed XY females, there was no change in any of the four ZFY-like genes—they were all there, and on at normal levels. There was absolutely no way they could be responsible for maleness.

These findings were so startling that although they got their first inkling of what was going on quite quickly after the Page Cell paper was published, Robin’s lab sat on the data for some considerable time, in order to make absolutely sure that they were correct. In the meantime, however, other doubts were starting to surface regarding ZFY. One of the earliest came from an unlikely source: a marsupial genetics lab in Australia, run by Jenny Graves. Page and Goodfellow both contacted her independently to get her to probe the marsupial genome, to check that marsupial Zfy was also on the Y chromosome. To everyone’s surprise, it was not; it lay on an autosome, which was very unexpected, because maleness in marsupials is also specified by the presence of a Y chromosome. Although the discovery made it onto the cover of Nature in late 1988, in an unlikely collaboration between all three labs, it was still not conclusive proof, because marsupial sex determination could quite feasibly just be a bit different from the mammalian mechanism. At this point, Page, certainly, was not terribly worried.

Peter contacted Marc Fellous, an old friend who had been a postdoc with Walter Bodmer when Peter was a student, and Fellous sent him his patient DNA samples. The four test samples were positive for the Y pseudo-autosomal boundary region, which Peter’s lab had recently cloned, meaning that the sex reversal was likely caused by an abnormal exchange between the X and Y chromosomes and should therefore be due to the presence of TDF, and hence ZFY. However, as Burgoyne had foreseen, all tests to find ZFY came up negative, and further mapping of exactly which part of Yp was present in the patients showed conclusively that all four of the chromosomal breakpoints were far closer to the pseudo-autosomal tip, ∼60,000 bp away from the boundary, and nowhere near Interval 1A2, where ZFY was located. ZFY could not be TDF.

ZFY could not be TDF. Peter describes the lab’s reaction to this incredible, mind-blowing news as “pretty excited,” which may be a little bit of an understatement. Combined with the Lovell-Badge data, it meant the end of ZFY’s short time in the limelight. The back-to-back papers describing the downfall of ZFY were published in the 21 December 1989 issue of Nature, two years after Page’s Cell paper; working on sex determination seems to bring with it the hazard of ruined Christmases for one’s competitors.

In July 1990, the two labs published their findings in two articles in Nature. Together, they had constructed an almost watertight case. The Goodfellow lab had done the human genetics, the positional cloning, and the sequencing, and had further shown that SRY was conserved in the males of multiple species, and in humans was only on in testes. Lovell-Badge’s lab had cloned the equivalent mouse gene, shown that it was male specific in mice, and that, crucially, it was expressed during development of the gonads in the cells required for testis formation.

Although SRY was universally agreed to be an excellent candidate for the elusive TDF, there was still one final test to be performed: An experiment was needed to show that by itself, SRY could cause sex reversal. Robin, the transgenic mouse expert, set about introducing Sry into female mice and seeing if it would be enough to turn them into males. The technical side of this was pretty much a doddle for Robin’s lab, but unfortunately, there had been an outbreak of Mouse Hepatitis Virus in the NIMR animal house (Robin had moved to NIMR from University College in 1988), and in order to eradicate it, there was a complete ban on breeding for some months. Robin and his lab kicked their heels and waited out the delay and, finally, managed to make five mice carrying multiple copies of artificially introduced transgenic Sry. One of these mice, very happily for all, had two X chromosomes and no Y, but was very clearly a boy, with all the right bits both inside and out. Randy, as he was christened, was very keen on mating with females but was infertile.

Photographing Randy. Robin Lovell-Badge is kneeling, centre.

(Photograph by Jérôme Collignon, courtesy of Robin Lovell-Badge.)

Robin and Peter wrote up the paper and sent it off to Nature, where it was duly accepted and scheduled for publication. Much to their delight, Robin’s photo of Randy climbing on a bar with one of his normal male companions, proudly showing off his very male private parts, was accepted as the front cover image for the 9 May 1991 issue of the magazine.

Randy. (Figure used with permission of the Medical Research Council/Photo Researchers.)

The paths of the three main protagonists of this story have diverged considerably in the 20-something years since Randy won the day for SRY. In the first months after ZFY’s downfall, David Page had to consider what to do next, and after much thought, decided, against the massed advice of his friends and colleagues, to carry on with the Y chromosome. His tenacity, superb bench skills, and infinite appetite for wringing the last drop of information from a complicated data set seemed to be perfectly suited to navigating the convolutions and complexity of the Y, and he refused to believe that the Y contained nothing more of interest. He was right: In 1992, his lab published a paper detailing the complete cloning of the Y chromosome, making it one of the first two chromosomes to be cloned, and signaling the renaissance of the Y. He collaborated in the sequencing of the Y chromosome in the late 1990s, finally resolving all of the hideous repeat sequences that had dogged the early chromosome walks and showing that they were there for a reason; the Y chromosome, uniquely amongst its fellows, uses the massive repeats to recombine with itself, maintaining genetic diversity. Since that time, Page has shown that the Y carries genes of paramount importance in spermatogenesis, and hence male fertility, work of great clinical significance. He is now Director of the Whitehead Institute, still in the same laboratory he first occupied as a 28-year-old novice. Looking back on the ZFY story now, Page is almost nostalgic. At the time, he says, it seemed perfectly fine to be working like a dog, maintaining his position at the front of the field by sheer determination and lack of sleep, and fighting intellectual battles with Peter Goodfellow at their regular conference encounters, because if you love something enough, that is what you do. The press attention and the trauma of the very public downfall of ZFY were horrible, but, despite all of his subsequent successes (and there have been many), the exhilarating rollercoaster of those days persists as the most exciting time of his career.

And Robin Lovell-Badge? Robin’s encounter with the media following Randy’s appearance has morphed into a long-term relationship, and he is now a highly respected commentator and communicator on science issues, especially in the moral minefield of stem cell research. In the lab, he still works at the bench when he can, and his record of publishing high-impact work in high-profile journals is undiminished. In a recent twist to the story of SRY, the long-held dogma that femaleness was a default state, and to be male one simply had to activate SRY and SOX9, was overturned in a 2009 Cell paper, in which Robin and collaborators showed in mice that the autosomal gene FoxL2 specifies femaleness, and when present, overrides Sox9 and prevents ovaries changing into testes. Loss of FoxL2 in adult female mice up-regulates Sox9, causing reprogramming of some ovarian cell types to those found in testes. Remarkably, as in life, the Sox9/FoxL2 story shows that maleness and femaleness appear to be established by what some might view as a balancing act, and others as a war.

Web Resources