by Matthew Cobb
Maurice’s dry English tone did not permit enthusiasm as he stated that the picture showed much more detail than previous pictures and could, in fact, be considered as arising from a crystalline substance. … Suddenly I was excited about chemistry. Before Maurice’s talk I had worried about the possibility that the gene might be fantastically irregular. Now, however, I knew that genes could crystallize; hence they must have a regular structure that could be solved in a straightforward fashion. Immediately I began to wonder whether it would be possible for me to join Wilkins in working on DNA.
At the time, Watson was based in Copenhagen, rather unsuccessfully studying phage duplication, having obtained his PhD on phage genetics with Salvador Luria a year earlier, at the remarkably young age of 22 years.24 Watson had become obsessed with how genes copied themselves, and Wilkins’s data seemed to suggest a way of attacking the problem. On his return to Copenhagen, Watson read Pauling’s slew of articles on the α-helix and became determined to use X-ray crystallography to study genes. After some squabbles with his funding agency, in autumn 1951 Watson moved to Cambridge and began work in Max Perutz’s group, which had world-leading expertise in X-ray crystallography. In Cambridge he shared an office with Francis Crick.25 One of the great scientific partnerships began. Watson described his new colleague in a letter to Delbrück:
The most interesting member in the group is a research student named Francis Crick. … He is no doubt the brightest person I have ever worked with and the nearest approach to Pauling I’ve ever seen. … He never stops talking or thinking and since I spend much of my time in his house (he has a very charming French wife who is an excellent cook) I find myself in a state of suspended stimulation.26
Crick recalled he was ‘electrified’ by meeting Watson. ‘It was remarkable’, he said, because they had the same focus on understanding gene structure but they had entirely different skills – phage genetics and crystallography, respectively. In 1947, Crick had described his desire to unravel ‘the chemical physics of biology’.27 He was not initially particularly interested in genes or DNA – like most people, he assumed that the genetic material would be a protein. In 1950 he had teased Wilkins for his work on DNA, telling him, ‘What you ought to do is get yourself a good protein.’28
*
During the summer of 1951, Wilkins became increasingly convinced that the X-ray diffraction data showed that DNA had a helical structure. When Wilkins first presented this idea, Franklin’s response was to tell him to stop working on X-ray data from DNA and ‘go back to your microscopes’. Wilkins, who was the assistant director of the laboratory, was incredulous that a postdoctoral researcher should speak to him in such a fashion, but typically said nothing.29 Matters got worse when another member of the King’s group, Alex Stokes, found mathematical support for Wilkins’s intuition. When Wilkins told Franklin of their success, she responded furiously, saying, ‘How dare you interpret my data for me?’30
To overcome the tension between Wilkins and Franklin, Randall adopted the simplest solution: the two researchers would be kept apart. There was a ready justification for this. Franklin had shown that DNA came in two distinct forms, named A and B, depending on the degree of humidity. The A form, which could be seen under drier conditions, gave precise but highly complex images; the B form, which occurred at high humidity levels, was more blurry and less enticing. X-ray crystallography requires precise measurements from the diffraction images; if the photo is blurred, it will be impossible to come to an accurate description. Randall decided that Franklin should study the A form DNA using Signer’s samples, while Wilkins would investigate the B form using Chargaff’s less pure DNA. Randall could perhaps have resolved everything by being frank with the pair of them; but he was not, and matters got no better.
Shortly afterwards, on 21 November 1951, the King’s group organised a small meeting on DNA at which they all presented their latest findings. Jim Watson was in the audience and when Franklin spoke, his attention was torn between her results and idle musings about her looks. Franklin showed DNA images she had made with the lab’s new fine-focus X-ray tube, which had a very narrow beam, and she described the two forms, A and B, highlighting the fact that the A form produced clearer images, showing ‘evidence for spiral structure’. The next day, Watson and Crick excitedly discussed the details that Watson could recall – as was his habit, the cocky young American had taken no notes. Crick became convinced that only a few structures would fit Franklin’s data and within a week they had a model for DNA. This was accompanied by a rather pompous eight-page ‘memorandum’ by Crick that outlined their strategy, which was above all ‘to incorporate the minimum number of experimental facts’.31 This was quite appropriate, as neither man had done a single experiment on DNA structure.
The first Watson and Crick model of DNA was a triple helix – there were three intertwined phosphate-sugar strands in the centre, with the bases sticking out like fingers. In triumph the pair invited Wilkins, Gosling and Franklin to come to Cambridge to view their creation. Franklin took one look at the model and dismissed it. Her X-ray data that had so entranced Watson had clearly shown that the phosphate-sugar groups were on the outside, not the inside, whereas the magnesium ions that held Watson and Crick’s triple helix together would be unable to fulfil this function because they would be surrounded by water molecules. If such a structure ever existed, it would instantly fly apart. All this had been explained in her talk at the King’s meeting, but Watson had not fully understood what she was saying and had not written anything down. Watson and Crick’s first venture into model-building ended in embarrassing failure.
Watson’s failure to pay attention was even more significant than he realised. According to Franklin’s notes, when she spoke at the November meeting she described the shape of the ‘unit cell’ (the shape of each molecule) of the DNA crystal as ‘monoclinic’. This crystallographic jargon meant that the molecule would show rotational symmetry, and that if there were chains of molecules wrapped around each other in the structure, they must run in opposite directions. This turned out to be a vital insight into the structure of DNA, but Watson did not understand enough crystallography to grasp its significance. Crick did, in an instant, when he eventually learned of it fifteen months later.
When Randall heard about the Cambridge duo’s attempt to muscle in on the structure of DNA, he furiously asked the head of the Cavendish Laboratory, Sir Lawrence Bragg, to tell the two upstarts to leave DNA alone. Bragg had a low opinion of Crick and probably no opinion of Watson, who was beneath his notice, so he willingly banned them from doing any further work on DNA. Crick returned to his PhD on haemoglobin structure, and Watson began studying the nucleic acids in the tobacco mosaic virus, learning elementary X-ray crystallography. The fiasco also reinforced Franklin’s prejudices against building speculative models. The data had to lead to the model, she felt. Watson and Crick, high on mathematics and fixated with Pauling the alpha-helix male, had been utterly confident that logic and ‘the minimum number of experimental facts’ would lead them to the answer. Instead it had led to ridicule.
Despite Franklin’s conviction that the results would speak for themselves, her data were confusing because she was looking at the precise and detailed images produced by the A form of DNA. She apparently assumed they were so sharp because the A form was an array of crystals that were all oriented in the same direction. In fact, the A form is made up of small crystalline blocks in which the crystals within a block have the same orientation, but where different blocks have different orientations, producing an image that is both clear and complex.32 Deducing the structure of DNA from the A form image was going to be extremely difficult. When Crick eventually saw the A form data, in 1954, he told Wilkins:
This is the first time I have had an opportunity for a detailed study of the picture of Structure A, and I must say I am glad I didn’t see it earlier, as it would have worried me considerably.33
Eventually, putting too much faith in the sensitivity of
her apparatus, Franklin concluded that the A form was not helical and sent round a spoof death notice, edged with black, to members of the laboratory:
It is with great regret that we have to announce the death, on Friday 18th July, 1952, of D.N.A. HELIX (crystalline).
By this time Franklin had already decided she had had enough of the terrible atmosphere at King’s; she agreed with Randall that at the end of the year she would move to Birkbeck College and would abandon her studies of DNA.
Despite the triple helix fiasco, Watson and Crick did not stop thinking about the structure of DNA. Crick asked a colleague to work out the chemical bonds that could exist between the bases and was delighted when he was told that A would bind with T, and C with G. In a flash, Crick realised that this provided the clue to gene replication, through what he called complementary replication. If A on one molecule bound with T on another, you would get a kind of mirror image of the original DNA; if the same process was then repeated with that ‘mirror’, a new strand of DNA, identical to the original one, would have been created. If there were two molecules of DNA bound together at the outset, replication would be even more straightforward – simply copy each strand and you would duplicate the original molecule.34
At the end of May 1952, Chargaff visited Cambridge and had a meal with Watson and Crick at which they talked about DNA. It did not go well. Chargaff – a notoriously prickly character – poured scorn on them because of their ignorance of chemistry and of his work. He later recalled that his first impression was ‘far from favorable; it was not improved by the many farcical elements that enlivened the ensuing conversation … So far as I could make out, they wanted, unencumbered by any knowledge of the chemistry involved, to fit DNA into a helix. The main reason seemed to be Pauling’s alpha-helix model of a protein.’35
Despite his evident irritation, Chargaff told Watson and Crick of the enigma of the apparent 1:1 ratios of A:T and C:G. As Crick later recalled:
Well, the effect was electric … I suddenly realized, by God, if you have complementary replication, you can expect to get one-to-one ratios.36
In a rare foray into the laboratory, Crick spent the next week trying to get bases to pair spontaneously in the test tube. It did not work, and Crick’s flash of insight led nowhere, for the moment.
*
King’s and Cambridge were not the only places where scientists were trying to understand the structure of DNA. On 28 May and 1 June 1951, Elwyn Beighton in Bill Astbury’s laboratory in Leeds took some of Chargaff’s DNA and made several X-ray images. They had the classic X-shape which we now know reveals the presence of a helix. Astbury was not impressed, and did not encourage Beighton to continue his work; he was not able to interpret the images correctly because Crick had not yet published his papers that described the diffraction pattern produced by a helix.37 Astbury may have felt that the material was less pure than the extracts he had worked with before the war, or he may have merely been frustrated that the image seemed too simple at a time when all the lines of argument were suggesting that DNA was the genetic material, the carrier of specificity. At around this time, the Medical Research Council (MRC) rejected Astbury’s proposal to create a new department, and the arrival on the scene of the well-funded group at King’s College may also have discouraged him from pursuing the structure of DNA. Whatever the case, the Leeds images became a historical dead-end, an enigmatic curiosity, and Astbury’s direct involvement in the determination of the structure of DNA was over.38
At about the same time, Edward Ronwin of the University of California produced a model of DNA. This was superficially similar to Astbury’s 1947 model – it had the phosphate-sugar backbone in the centre, with the bases fanning out. However, Ronwin had made some basic biochemical errors and his model contained too much phosphorus. Linus Pauling was outraged, and wrote a letter to the editor of the Journal of the American Chemical Society criticising the ‘foolishness’ of publishing Ronwin’s model, and complaining about ‘the irresponsible publication of unsupported hypotheses.’39
More seriously, in the first half of 1952, John Rowen at the National Cancer Institute in Maryland studied the molecule using light-scattering electron microscopy and viscometry. In 1953 he published an article describing its configuration as ‘intermediate between a rod and a coil’ before concluding that one of ‘its most striking properties is its tendency to spiral, twist and intertwine with neighbouring molecules’.40
*
Beginning in the second half of August 1952, Franklin took X-ray photos of her DNA samples every day, using a heavy disc-shaped metal camera that was about the width of an orange. Her PhD student, Raymond Gosling, recalled that she bounced ideas off him – they had what he called ‘pretty hot discussions’ in which she played the role of devil’s advocate and which he found enormously stimulating.41 She never had that kind of discussion with Wilkins. Together with Gosling, Franklin began to calculate the Patterson function, the difficult mathematical procedure used by Sven Furberg at Birkbeck. This involved projecting the X-ray photos in a dark room, measuring the position and intensity of the various blobs on the pictures and then spending hours doing complex calculations. In November 1952, Franklin summarised the data she had obtained with Gosling as part of a brief report by Randall for the Biophysics Committee of the MRC. There was nothing in Franklin’s few paragraphs that had not been presented at the King’s symposium the previous year, but this time the data – including the different sizes of the repeating patterns in the A and B forms and above all the dimensions of the monoclinic unit cell, were written out clearly and slightly more precisely.42 In the middle of December, members of the Biophysics Committee made an official tour of the King’s lab and were each given a copy of the informal document. One of the visitors was Max Perutz from Cambridge.
*
At the end of 1952, Pauling finally stirred. Even though he still assumed that proteins explained genetic specificity, Pauling had been snuffling around DNA for some time – a year earlier he had cheekily written to Randall asking to see their data; Randall had given him the brush-off.43 In Watson’s mind, Pauling had become a bogeyman, a powerful competitor who had the ability to steal the prize, should he so desire. At the end of the year, just as Watson feared, Pauling submitted an article to the US journal Proceedings of the National Academy of Sciences, based entirely on measurements from Astbury and Bell’s data from 1938.44 But when the manuscript arrived in Cambridge at the end of January, Watson and Crick were relieved to see that the structure was just as wrong as their own first attempt. It, too, was composed of three intertwined helices; it, too, had the bases on the outside; and to their amazement the way in which the model was built meant that the molecule was not an acid at all. As Watson later put it, ‘a giant had forgotten elementary college chemistry’. Finally, the model did not explain the nature of the gene or its essential functions: replication and specificity. It was biologically mute.
A few days after reading Pauling’s manuscript, Watson went to King’s, where he had a brief squabble with Franklin over her apparent refusal to accept that DNA was helical. He then saw Wilkins, who took him to his office and showed him a photo of the B form that Raymond Gosling had given him a few days earlier as the pair worked on Gosling’s thesis; with Franklin’s departure imminent, Gosling was once again being supervised by Wilkins. This photo – ‘photo 51’– had been taken in May 1952 by Gosling, but it had not been studied and had lain in a drawer for months.45 Watson was stunned – the image was so much simpler than any he had previously seen. As he put it: ‘The instant I saw the picture my mouth fell open and my pulse began to race.’46 In the centre was an X shape; Watson was a crystallographic novice, but from his discussions with Crick, who had been working on the crystallographic interpretation of helical molecular structures, he knew that the X could only come from a helix.
The significance of photo 51 in the identification of the double helix structure of DNA has often been overstated, mainly because of the weight gi
ven to it in Watson’s own world-famous account, The Double Helix. In reality, the insight given by the photo was extremely limited. Everyone at King’s – even Franklin – now accepted that the B form was helical, and without any more precise details of the measurements of the molecule, all that happened was that Watson’s preconception was confirmed – DNA had a helical structure. Furthermore, there was nothing underhand going on – Watson was shown the photo in perfectly legitimate circumstances. Raymond Gosling was absolutely certain: ‘Maurice had a perfect right to that information’, he said later. Whether Wilkins used that information wisely is another matter. He later regretted his action – ‘I had been rather foolish to show it to Jim’, he wrote in his memoirs.47 Despite the excitement that Watson felt, all the main issues, such as the number of strands and above all the precise chemical organisation of the molecule, remained a mystery – a glance at photo 51 could not shed any light on those details. The decisive information, which was unwittingly provided by Franklin herself, came from another source.
Pauling’s foray into DNA structure led Bragg to lift his injunction against Watson and Crick working on a model of DNA – he was not going to allow Pauling to repeat the coup of the keratin α-helix. Wilkins reluctantly agreed. There followed a rapidly evolving frenzy of dead ends, brilliant insights and chance encounters as Watson and Crick worked furiously to solve the problem. Most decisively, Max Perutz showed Crick the MRC report that included Franklin’s brief summary of her data from fifteen months earlier. Although it contained nothing that Watson should not have noted in November 1951, it now provided Crick with the information he needed. By chance it chimed completely with what he had been working on for months: the type of monoclinic unit cell found in DNA was also present in the horse haemoglobin he had been studying. This meant that DNA was in two parts, each matching the other. As Crick later recalled, ‘the chains must come in pairs rather than three in a molecule, and one must run up and the other down.’48 Then another occupant of the Watson and Crick office, Jerry Donohue, pointed out that Watson was using the wrong forms of the bases when he was trying to build the model. With the correct structures, adenine bound with thymine, and cytosine bound with guanine, using hydrogen bonds, as Gulland had reported six years earlier.