Answer in Summary
The essential features of the structure of DNA proposed by Watson and Crick in their 1953 Nature paper were
- a helix with the sugar–phosphate backbone at the outside
- two anti-parallel strands
- hydrogen bonds between A-T and G-C base pairs
at the centre of the helix, holding the two strands together
The double-stranded anti-parallel helical aspects of the model were based on or confirmed by interpretations of X-ray diffraction photos taken by Rosalind Franklin, primarily of fibres of the B-form of DNA, which she had discovered. The specific base-pairing was based on model-building following chemical principles, with X-ray diffraction serving only to indicate a general constant diameter to the structure and to confirm or refine the repeat distances of the components. This latter was the biologically most important aspect of the Watson and Crick structure which “suggest[ed] a possible copying mechanism for the genetic material”.
More Detailed Answer
I start by providing a broad bibliography relevant to my account. However what I write is based mainly on the 2012 annotated version of Watson’s The Double Helix, and Aaron Klug’s 2004 J.Mol.Biol. article, The Discovery of the DNA Double Helix. Both contain facsimiles of important documents.
Time Line and Bibliography
1950 Chargaff paper including ratios of bases in different DNAs: Experientia 6, 201–240
1951 Pauling et al. Protein α-helix structure: Proc.Natl.Acad.Sci. 37, 205–211
1951 Rosalind Franklin moves to MRC unit in King’s College London
1951 James Watson moves to Cambridge Cavendish Lab.
1951 (early December) Debacle of presentation of initial triple-helical model of DNA
1952 Chargaff meeting with Watson and Crick (recounted in his book, Heraclitian Fire, Rockerfeller University Press, New York, 1978)
1953 (February) Pauling incorrect triple-helical structure of DNA: Proc.Natl.Acad.Sci. 39, 84–97
1953 (March 14) Franklin moves to Birkbeck College and writes up her work in an unpublished typescript dated March 17.
1953 (March 18) Watson/Crick Nature manuscript received at King’s.
1953 Three Nature papers published, in April 25 issue: Watson and Crick; Wilkins, Stokes and Wilson; Franklin and Gosling.
1953 (May) Watson and Crick paper with illustration of proposed DNA base-pairing. Nature 171, 964–967
1958 Franklin’s death, 16 April.
1962 Crick, Watson and Wilkins Noble Prize in Physiology or Medicine.
1968 JD Watson The Double Helix (Atheneum Press, New York) — His account of events. An Annotated and Illustrated version of 2012 (ed Gann and Witkowski, Simon and Schuster) contains interesting facsimiles.
1968 A. Klug article Rosalind Franklin and the Discovery of the Structure of DNA: Nature 219, 808–810/843–844
1974 A. Klug update on the previous article in light of subsequently discovered unpublished draft manuscript Nature 248, 787–788
1975 Publication of Anne Sayre’s biography of Franklin — Rosalind Franklin and DNA (W. W. Norton & Company)
1980 First crystal structure of B-DNA confirming structure including the base pairing Nature 287, 755–758
2002 Publication of Brenda Maddox biography of Franklin — Rosalind Franklin: The Dark Lady of DNA (HarperCollins)
2004 A. Klug, extensive article The Discovery of the DNA Double Helix. J.Mol.Biol. 335, 3–26
2023 Article by Cobb and Comfort, reanalysing Franklin’s contribution and criticisms of Watson, Crick and Wilkins in the Sayle and Maddox biographies: Nature 616, 657–660
What was the importance of the structure of DNA?
It seems highly likely that the determination of the structure of DNA would have received the Nobel Prize, whatever its biological implications. However its importance was heralded in the second line of the original paper “This structure has novel features which are of considerable biological interest”, and this is echoed in the
Nobel award citation:
The Nobel Prize in Physiology or Medicine 1962 was awarded jointly to
Francis Harry Compton Crick, James Dewey Watson and Maurice Hugh
Frederick Wilkins “for their discoveries concerning the molecular
structure of nucleic acids and its significance for information
transfer in living material” [My emboldening]
Of the three features of the structure of DNA listed in the summary, the aspect that embodies this significance is the A-T, G-C base-pairing between the two DNA strands that “suggests a possible copying mechanism for the genetic material”.
The lack of appreciation of this latter point in some quarters may be due to the fact that the chemical details of the base-pairing were not illustrated in the original April paper:
but left until their May paper:
Nature of the information obtained from X-ray analysis of fibres
(This is dealt with in a more authoratative and comprehensive fashion in the Klug article.) For molecules of identical structure that can be made to form suitable crystals it is possible to use X-ray diffraction (with isomorphous replacement) to determine the position of every atom other than hydrogen in the structure. Work by Kendrew in the Cavendish laboratory in Cambridge produced the first such structures for the globular protein, myoglobin, in 1958. For extended macromolecules that can be made to form semi-crystalline fibres, X-ray diffraction can reveal only general features such as symmetry, atomic density and distances apart of repeating components. It is possible to perform mathematical analysis to predict the pattern a particular structure such as a helix would produce and to determine its pitch. Both Crick and Stokes (Wilkins’ coworker) had made important contributions to the theory underlying this. However the precise positions of the nucleic acid bases in DNA could only be postulated from building models based on chemical principles. Such model structures were not provable, but could be supported or eliminated by calculating the diffraction pattern they would produce, and seeing whether it was consistent with that observed. (This same situation held for the amino acids in Pauling’s 1951 structure of the α-helix of keratin.) The occurrence of different bases in the DNA strand is irregular, and it was only because of the purine¬–pyrimidine base pairing and the similarities in size of the two purines (A and G) and of the two pyrimidines (T and C) that the DNA fibres had a uniform diameter that gave precise X-ray diffraction. Whereas, Pauling’s α-helix was observed in detail as a component of myoglobin seven years after was proposed, it took almost 30 years for the base-paired structure proposed by Watson and Crick to be formally confirmed using crystals of a specific synthetic 12 base-pair oligonucleotide.
Watson and Crick, their approach and the development of their model
Francis Crick was trained as a physicist and was in Perutz’s group in the Cavendish laboratory in Cambridge, working on protein structures and the theoretical interpretation of X-diffraction patterns that could arise from them. Watson was a biologist working in Luria’s phage genetics lab in the US, who had come to Europe to learn more about the chemistry of DNA and ended up in the Cambridge lab. He had previously encountered Maurice Wilkins briefly at an international meeting, and was greatly influenced by his photos of X-ray diffraction of DNA fibres that indicated to him that DNA had a regular solvable structure. In Cambridge Watson established a strong intellectual and personal relationship with Crick (who, incidentally, knew Wilkins) and persuaded him to apply his knowledge and intellect to the problem of the structure of DNA. Their approach to modelling the structure of DNA is set out in a memorandum Crick wrote in late 1951:
“We have tried in this approach to incorporate the minimum number of
experimental facts, although certain results have suggested ideas to
us. Among these may include the possible helical nature of the
structure, the determination of the unit cell, the number of residues
per lattice point, and the water content.” [p.85 Annotated and Illustrated Edition of Watson’s The Double Helix]
The ‘possibility’ of a helical structure was, in fact, a basic assumption, and Erwin Chargaff who met the pair in 1952 writes in his book, Heraclitian Fire “So far as I could make out, they wanted, unencumbered by any knowledge of the chemistry involved, to fit DNA into a helix.”
Nevertheless, their models were built in relation to information they had from (predominantly) Franklin’s studies in London. They built an initial model in late 1951, in the context of Watson’s misremembering or misunderstanding of a seminar presented by Franklin in mid-November on the A-form of DNA. (The A- and B- forms of DNA are explained in the section on Franklin, below). This model was a triple helix with a phosphate core, which was humiliatingly rebuffed by Franklin, one key point being that the X-ray photographs were only consistent with the dense phosphate groups being at the exterior of the structure. Watson and Crick were then ordered by Lawrence Bragg, the head of the Cambridge unit, to stop building models of DNA. However Watson persuaded Bragg to allow them to resume in January 1953. He had presented the evidence from Franklin’s Photo 51 (below) that the B-form was helical, and Bragg presumably made the decision in the light of knowledge that Pauling (his old rival) was engaged on the problem but that the King’s group divided and were not engaged in model building to counter him. (The Cambridge and King’s groups had details of Pauling’s incorrect proposal for a DNA triple-helical structure before publication.)
Watson and Crick’s model building was facilitated by two pieces of information obtained by Franklin and communicated to them indirectly.
As just mentioned, one was an X-ray photo (Photo 51) of the B-form of DNA, which Franklin had obtained in mid 1952, but gave to Wilkins (through the intermediary of Gosling) before she left King’s College for Birkbeck in 1953. Her reason for doing this would appear to be that she did not have time to finish experimental and theoretical analysis of the B-form, that at Birkbeck she was changing her topic to tobacco mosaic virus, and the data belonged to the project (just as did the DNA that Wilkins was obliged to hand over to her initially). The B-form had a simpler pattern than the A-form, and the particular photo — which Wilkins showed to Watson “some time” in January 1953 — indicated clearly that the structure was helical with a 3.4Å repeat, consistent with the stacking of the bases in a direction perpendicular to the helical axis.
The photo did not indicate how many strands were in the putative helix, but other considerations available from the outset suggested either two or three. Watson decided to concentrate on a two-stranded model, not because of any evidence in favour of this, but because of his perception of repeated “two-ness in biological systems”. This rather arbitrary choice was subsequently found to be correct when in the second week of February 1953 Perutz showed him and Crick part of a progress report on the work of the King’s group that Perutz had received as member of the review committee. Whether or not it was unethical of Perutz to do so, it was he, not Watson or Crick, who took this initiative. When shown the report, it was obvious to Crick (because of research experience the others lacked) that seemingly mundane information about the space group of the crystals of the A-form had important implications. This was that the symmetry of the C2 space group indicated any helix would consist of two antiparallel strands. Watson and Crick were able to continue building models of a helix with two strands, now anti-parallel, focusing on a sugar–phosphate backbone positioned on the outside of the helix because of the long-standing data to that effect. To me the irony of this is that it was no secret — Franklin had included details of the space group at her seminar in King’s in November 1951, but Crick was not at the seminar and nobody else (neither Franklin, Wilkins nor Watson) appreciated its significance. (This point is only brought out in an editorial footnote in the annotated edition of The Double Helix, although Klug deals with it at greater length.)
The reason Crick and Watson had been reluctant to place the bases at the interior from the outset was that the X-ray photos indicated a regular structure, yet it appeared that DNA contained an irregular sequence of purines and pyrimidines which differ in size. (This, presumably, also applied to Pauling whose publication of an incorrect triple helical DNA structure with a central phosphate core was part of the stimulus for the resumption of model building.) Having satisfied themselves of the overall orientation of the sugar phosphate backbone the were then able to address this problem of the bases. The clue of Chargaff’s findings of the equivalent abundance of A and T, and G and C, together with the help of chemical colleagues in the choice of the enol tautomer of the bases, finally led to the AT/GC hydrogen-bonded base-pairing postulate, which produced similarly sized occupancy in the centre of the structure. Adjustment to the model was then made in the light of repeat distances indicated in photos taken at different orientations.
The role of Maurice Wilkins
Maurice Wilkins initiated the X-ray diffraction studies on fibres he prepared in 1950 with calf thymus DNA obtained from Rudolf Signer in Switzerland. In these fibres the DNA was in the A-form. He worked together with the theoretician, A.R.Stokes, on interpreting diffraction photographs, especially in relation to possible helical structures for DNA. The complexity of the diffraction patterns of the A-form prevented Wilkins and Stokes from concluding whether or not it was helical, although they were able to argue in their Nature paper that the diffraction patterns of Franklin’s B-form were consistent with Watson and Crick’s helical model.
For reasons that are not clear — discussed by Neidle, 2023 — JT Randall, the head of the group in London, without informing Wilkins, promised to hand over the complete project — including materials — to Franklin, who was supported by her own fellowship (initially obtained to work on protein rather than DNA). This led to extremely bad relations between Wilkins and Franklin, and complete lack of scientific co-operation between them. Wilkins was obliged to do less consequential work with sperm heads until Franklin left for Bernal’s lab.
Rosalind Franklin
As, mentioned above, Franklin took over Wilkin’s work (and his student, Gosling) and his A-DNA. She was extremely gifted, technically as well as intellectually, and discovered by preparing fibres at different humidities that a the structure changed to a different form — B-DNA — at higher humidities. This gave much simpler diffraction patterns, which her notebooks make it clear she understood were consistent with a helical structure. However she felt that analysis of the more complex A-structure was both necessary for a valid structure and would provide more information the analysis. She was categorically opposed to Crick and Watson’s model-building approach, feeling that when analysis was complete the structure would reveal itself automatically.
It should already have been made clear that Franklin’s X-ray diffraction observations were the basis of the structure for the backbone of DNA in Watson and Crick’s model: the external position of the phosphates, its helical nature, the two antiparallel strands and the specific dimensions of the helix. As regards the bases, the measurements indicated their repeat distance bases, but nothing more.
Appendix I: Rosalind Franklin’s own interpretation of her data
Although not directly relevant to the question posed here, there has been considerable interest in the extent to which Franklin had understood the implications of her experimental data, and indeed a question on this topic has been asked previously. I have provided a brief answer to that question, referring forward to this appendix, which can be read in the a much broader context.
The fact that Franklin never published models based upon her interpretation of her data — much less discussed her ideas with Watson and Crick — has led to conflicting speculation on the matter. However there is definitive information in the account given by Aaron Klug, who worked with Franklin at Birkbeck, and was given possession of her notebooks and other papers (now in the library of Churchill College, Cambridge) after her death. Klug is of the opinion that her focus on the A-form, although understandable, was the factor that obstructed her interpretation. It did not provide additional structural information that was useful to her, and she found it difficult to reconcile it with the structure of the B-form, for which a helical structure was a possibility.
In particular, she was misled by the diffraction pattern of one anomalous fibre preparation (Klug refers to it as “a mechanical accident”) which was inconsistent with a helical structure. This is presumably what gave rise to her ill-considered obituary notice for helical DNA (parenthetically “crystalline”, i.e. A-form) that she and Gosling circulated in King’s in July 1952, and is perhaps the basis of the view that she was dogmatically anti-helical throughout.
Her views, however, are clearly stated in a typescript (which contains the essentials of the April Nature paper by her and Gosling) she wrote shortly after leaving King’s, and a couple of days before receiving a copy of Watson and Crick’s proposal. To quote Klug:
“In Franklin’s draft, it is deduced that the phosphate groups of the
backbone lie, as she had long thought, on the outside of the two
co-axial helical strands whose geometrical configuration is specified,
with the bases arranged on the inside. The two strands are separated
by 13 Å (three-eighths of the helix pitch in the axial direction). But
the draft shows she had not yet grasped that the two chains in B also
ran antiparallel as in the A form.”
There is nothing in the draft about the bases, but Klug writes that:
“Her notebooks show that for fitting the bases into the centre of a
double helix, she had already formed the notion of the
interchangeability of the two purine bases with each other, and also
of the two pyrimidines. She also knew the correct tautomeric forms of
at least three of the four bases, and was aware of Chargaff’s base
ratios.”
The relevant conclusions of the Franklin/Gosling Nature paper from their own data are:
“Thus, while we do not attempt to offer a complete interpretation of
the fibre-diagram of structure B, we may state the following
conclusions. The structure is probably helical. The phosphate groups
lie on the outside of the structural unit, on a helix of diameter
about 20 A. The structural unit probably consists of two co-axial
molecules which are not equally spaced along the fibre axis, their
mutual displacement being such as to account for the variation of
observed intensities of the innermost maxima on the layer lines ; if
one molecule is displaced from the other by about three-eighths of the
fibre-axis period, this would account for the absence of the fourth
layer line maxima and the weakness of the sixth. Thus our general
ideas are not inconsistent with the model proposed by Watson and
Crick in the preceding communication.”
Note the distinction between the specific conclusions drawn from their data, and the final sentence limiting their ideas to not being inconsistent with Watson and Crick’s model.
Appendix II: Erwin Chargaff’s observations
The A–T, G–C base-pairing in the Watson–Crick model was not, as already mentioned, derived from any X-ray diffraction data, but informed by Chargaff’s observations of the equivalence of these bases in DNA from a range of sources. It should be emphasized that the implications of Chargaff’s observations were by no means obvious, either to him or anyone else (they were first published in 1950). Their relevance to both Watson and Franklin became apparent only because of a specific context — the accommodation of different sized bases in the centre of a helix with a fixed diameter. If one made the leap to hydrogen-bonding between purine and pyrimidine, identifying the atoms involved still required knowing that only a single tautomer existed for each base, and what that tautomer was. The non-trivial nature of this is illustrated by the fact that the original Watson–Crick model had only two base-pairs for G–C, rather than the three subsequently shown to be present.
