A Sidelight on Watson and Crick:
(A Corollary to Murphy's Law: Seekers/Finders >> 1)
The elucidation of the double helix structure of DNA by Watson and Crick has taken on mythic dimensions. The authors of the myths, non-scientists and (even) scientists like Watson himself, at the time ignorant of the history of DNA research, have presented a sometimes self-serving scenario akin to the primitive myths of creation in which the world—or the DNA structure—is derived from a formless void. Of course, it wasn't like that.
Science today is a cumulative enterprise rising step by step on a staircase of four centuries of sustained investigation and analysis. Each step has a background and a context. Indeed, the quick acceptance of the double helix structure derived from the fact that it made coherent and even plausible so many prior observations from genetics and DNA biochemistry.
As a participant in DNA research throughout this period, I had a close-up and somewhat bemused view of the ongoing research. As is the norm in science, several lines of research, including my own, were converging on the structure of DNA. The concept then of the double helix was, in itself, not so large a step—but it was a step that brought biology onto a new plateau with wide horizons and numerous paths to explore.
By the early 1950s, several laboratories were intensively engaged in the study of DNA. Arthur Kornberg's laboratory, then at Washington University and later at Stanford, sought to unravel the biochemistry of
DNA synthesis. Paul Doty's laboratory at Harvard applied powerful tools to the analysis of the peculiar physical chemistry of DNA. Alfred Mirsky at Rockefeller Institute had measured the DNA content of cell nuclei of varied species while Erwin Chargaff had shown that the purine and pyrimidine composition of these DNAs always displayed a curious regularity: the moles of adenine equaled those of thymine and the moles of guanine equaled those of cytosine. In London, Franklin and Wilkins sought to improve on the earlier X-ray diffraction studies of DNA by William T. Astbury.
I too, by a somewhat circuitous route, had become interested in aspects of DNA structure and sequence. In the course of my earlier studies, intended initially solely to provide the mononucleotides I needed for irradiation studies, I realized I had also opened an avenue into questions of nucleotide sequence and DNA structure. For in the course of the mononucleotide preparation, we at one stage reduced the DNA to a mixture of small polymers of an average length of four nucleotides. If we could fractionate this mixture we would be in a wholly unexplored realm, since until then no one had ever isolated, identified, or purified oligonucleotides (short strings of nucleotides) of any size or variety.
We found that we could indeed separate the digest into dinucleotides, trinucleotides, and larger fractions. And further, by careful choice of conditions, we could fractionate all of the dinucleotides according to their particular nucleotide composition. The ultraviolet absorption of each dinucleotide suggested its composition. This could be verified by degrading each to its mononucleotides and fractionating these.
This work proceeded well and brought us, for the first time, a small step into the world of DNA sequence—the order of nucleotides in DNA. By midsummer 1952, we knew that:
(1) The unusual nucleotide 5-methylcytidylic acid was to be found only in one isomeric form of one dinucleotide, which meant that it always preceded a guanylic in the DNA chain (this has proven to be general in the DNA of mammals and the presence of methylcydylic acid is now believed to be involved in the control of genetic expression);
(2) All of the other two-nucleotide combinations could be found in reproducible but varying proportions.
This latter was quite significant. One possible explanation of the Chargaff regularities (moles adenine = moles thymine and moles guanine = moles cytosine) had been that perhaps adenylic was always fol-
lowed by (or preceded by) thymidylic, and guanylic similarly by cytadylic. However, this clearly was not the case. The Chargaff regularities demanded another explanation.
I discussed this question with Fritz Schlenk, then professor of microbiology at Ames. The simplest explanation I could propose was that there must be two DNA chains, related in some way, so that wherever there was as an adenylic in one chain, there was thymidylic in the other, and likewise for guanylic and cytidylic plus methylcytidylic. This hypothesis had an obvious consequence, but unfortunately I was not then in a position to test it. If it were correct, then for every dinucleotide sequence, say CA, there should be an equimolar amount of the complementary sequence GT from the other chain (or TG if the two chains should have opposite orientation).
But our dinucleotides amounted to only one-sixth of the partial digest, and we could not then sequence the larger oligonucleotides. Nor could the dinucleotides present be considered a random sample as we were ignorant of possible preferences in the enzymatic digestion. Indeed, biochemical verification of this concept had to await the development of the "nearest neighbor" technique by Arthur Kornberg in his later studies of DNA synthesis.
At that time, I did not think in terms of helical structures and while I thought of the DNA as genes, I could not suggest any good reason for the presence of two such chains other than vague ideas about possibly coding for two proteins with some oddly defined relationship. My thoughts were oriented toward gene structure and expression, rather than toward gene reproduction and mutation, and so I missed the true significance of the complementary chains (so obvious when they are paired). This oversight was the result of an education that emphasized biochemistry and biophysics rather than inheritance and variation.
Shortly after I began my fellowship at Caltech in January 1953, I presented our dinucleotide data in a biochemistry seminar. I was somewhat surprised at the intense interest shown. I mentioned the two-chain notion, but, remarkably, the focus of attention was on the quantitative relationships among the dinucleotides, even though these accounted for but a fraction of the digest. I soon appreciated that the reason for this interest lay in prior discussions at Caltech about possible coding schemes to relate DNA sequence to protein sequence. Ours was the first quantitative data about any aspect of nucleotide sequence.
In February 1953, Linus Pauling announced a seminar at which he
would present the structure of DNA as deduced from X-ray diffraction data. As Pauling had earlier discovered the basic structures of proteins from X-ray diffraction data, his seminar was eagerly awaited. He put forward a triple helical structure for DNA with the phosphate groups on the inside and the rings of the nucleotides facing out. This structure was met with much skepticism. Many of the chemists present felt that the charged phosphate groups could not possibly be located in such close juxtaposition in the interior of the molecule.
Max Delbrück and I went to see Pauling to suggest possible modifications. Max was concerned that the chains of the helix should be paranemic (i.e., combable) rather than plectonemic (i.e., intertwined) so that they could be separated without the complication of unwinding—a very real problem for which the cell has had to devise special mechanisms. I wanted to propose that Pauling examine a two-strand helix to incorporate the Chargaff regularities and my notion of two complementary strands.
Pauling, however, could not be persuaded. He felt that his model was the only solution to the X-ray data and the known density of DNA. Unfortunately, as it later developed, he was using the old X-ray data and density data of Astbury, which had been obtained with a poor preparation of DNA that had been significantly denatured. Pauling had no access to the newer X-ray diffraction data of Franklin and Wilkins, which undoubtedly would have quickly set him on the right track.
The following month, Delbrück received a letter from Watson setting forth the Watson-Crick model. This immediately resolved many of the concerns raised about the Pauling model, which was abruptly discarded. The Watson-Crick model was then presented with immediate acceptance at the Cold Spring Harbor Conference in 1953. The convincing evidence for the biological, as opposed to the chemical, reality of the Watson-Crick model came from the Meselson-Stahl experiments at Caltech in 1957 demonstrating the separation and conservation of each of the DNA strands on replication.
The only substantial, if temporary, challenge to the Watson-Crick model came with my later discovery of the single-stranded DNA of the bacteriophage, f X. If Watson and Crick were right, how could this DNA reproduce or even code? The conundrum was subsequently resolved with the demonstration that, within the infected cell, the single-stranded viral DNA was quickly converted to a double-stranded replicative form, which replicated as such and which coded for proteins.
Later in the infective cycle, viral-coded proteins produced the single strands of DNA for progeny viral particles, using one strand of the double-stranded replicative form DNA as a template.
The Watson-Crick structure, by providing a firm biochemical basis for genetic phenomena, led biology into a new era, but it did not emerge unprecedented from a void. Mine was but one approach. Other lines of research in the laboratories of Kornberg, Doty, and Franklin and Wilkins were converging on the double helix. The double helix was out there waiting to be revealed. By 1950, after a century of biochemistry and fifty years of modern genetics, it was but a short step further in the unknown.
Science is, in fact, a collective enterprise. To outsiders, however, Watson and Crick by virtue of their great discovery and their strong personalities have made it seem a more personal adventure.
Jim Watson is the stuff of which People magazine is made. Brilliant, arrogant, verbally crude, with a skewed, off-center personality, fond of publicity, he sees the world in black and white with little gray in between. In his career, he has consistently demonstrated excellent scientific judgment and has been a superb director of the Cold Spring Harbor Laboratory. Crick is an interesting complement—highly articulate, sometimes glib, skilled in debate, sophisticated, and seemingly self-confident, even arrogant, yet harboring a persistent personal reserve.
Watson's book The Double Helix was published in 1968. Seemingly an exercise in candor, it portrays him as a total opportunist, devoid of any sense of scientific community, heedless in his passion to "beat Pauling" in the race to the "golden prize." Of course, Pauling never knew he was in a race. As a testament, it conveys an impression of science as just another cutthroat competition, akin to Wall Street or even the political arena. Shortly after the book appeared, I was awarded the California Scientist of the Year prize by the California Museum of Science and Industry for our research in collaboration with Arthur Kornberg at Stanford. At the inevitable news conference, reporters—with Watson's account fresh in their minds—could not comprehend the genuine, and really quite routine, collaboration between our two laboratories in the common effort to advance scientific knowledge.