Letter from Francis Crick to Rosalind Franklin (and notes on two papers by Franklin and Raymond Gosling)
Crick here discussed the X-ray crystallographic evidence offered by Rosalind Franklin and her doctoral student Gosling on
the structure of DNA extracted from calf thymus, specifically on the distinction between the crystalline A form and the paracrystalline,
heavily hydrated, longer and thinner B form of DNA, the form which DNA takes before replication and the form that yielded
the clearest X-ray evidence for a helical structure. In addition, Crick commented on Franklin's evidence that the position
of the sugar-phosphate backbone on the outside of the DNA molecule, where the phosphate atoms can react with hydrogen and
become ionized, that is, acquire the electrical charge that turns nucleic acid into an acid.
A final, critical piece of evidence Crick noted was Franklin's determination of the space group of DNA crystals, face-centered
monoclinic. Crystallographers group crystals in 230 different space groups according to their symmetry, that is, according
to the shape of their unit cell. As Crick realized, crystals in a face-centered monoclinic space group have a dyadic structure,
meaning that one half of a structure matches the other half in reverse. Which is to say that if the structure is rotated
by 180 degrees, it comes back into symmetry with itself. This crucial evidence, which Franklin herself as well as other researchers
in her laboratory failed to appreciate, pointed to a molecule made of two parallel chains running in opposite directions.
Number of Image Pages:
6 (410,707 Bytes)
1953-06-05 (June 5, 1953)
Original Repository: Wellcome Library for the History and Understanding of Medicine. Francis Harry Compton Crick Papers
I am returning the two papers which you so kindly lent to Jim, which as you can imagine I read with very great interest.
I am sorry not to have sent them before, but Jim and I had to write a paper before he left for the States, and this kept us
very busy. I am enclosing a few comments on the papers and some longer remarks on salt, on Riley and Oster's results,
and on calculating structure factors. What a problem it is!.
With best wishes,
F. H. C. Crick.
Notes On Two Papers By Franklin and Gosling
p. 5 I am not really clear about fibres which only give Structure B. How do the water content, intensities and the equatorial
spacing change with different R.H.?
p. 7 Comparison with Riley Oster. I take it this is Column 4 of their Table 1, or is it Column 5?
p. 9 I agree that the phosphates must be accessible. The general evidence suggesting they are on the "outside" seems
plausible, but less compelling.
p.12 I had not realized that calf thymus was the only material to give Structure A. Is this still true? Is it perhaps due
to the superior method of preparation used for calf thymus?
I am not quite clear how one can be certain that the unit cell is truly face-centered monoclinic, and not really triclinic,
with two angles 90 degrees. The point is important because if the unit cell is strictly C2 one must have the DNA chains in
pairs, running in opposite directions.
p.11 I take it that "12.14 A" is a slip.
p.11 Jerry Donohue is worried because in Fig. 3 there is so little vector density at x = 40 A, z = o. A more natural choice
would have been x = 38, z = o. As you do got list your observed equatorial spacings, it is difficult to see how things would
fit (Incidentally 22 x 3 = 38).. Is the effect perhaps due to the negative contributions from the peaks near x = 44, z =
p.15 As you know, we have never believed the anti-helical implications of the last paragraph, because of polybenzyl glutamate.
Would you call an 11-fold axis a high degree of symmetry?
Calculation of Intensities
Jim tells me that you claimed that "the water had only a general lowering effect, and that the Na+ was negligible".
This is not strictly correct. For calculating structure factors for the longer spacings it is convenient to allow for the
water by taking the average electron density of water as the "zero" of electron density. Thus any group which has
the same electron density as water makes no contribution. For each group its average electron density must be calculated,
and that of water subtracted, before its contribution can be given its proper weight. Thus, if the electron density of a
base were, say, 1.2 times that of water, only a fraction (0.2/1.2) of its electrons would be counted. On the other hand the
Na+ ion, due to strong electrostrictive, probably has a very small (or negative?) partial specific volume, and thus almost
all of its electrons will count. Thus the effect of one Na+ may be about the same as that of a base. It is this necessity
for allowing for the water which makes structure factor calculations difficult.
Riley and Oster
There seems little doubt that at least some of R. and O's long spacings are genuine, and it is a pity that in studying
Structure B the backstop in your experiment was placed so that it would hide any such spacings, though one can see you wanted
It is interesting to note that the spacings in your Paper 1, plate 6, are almost exactly in the ratio of 3140 and 4040 suggested
by R. and O. If it were not for this complication of a longer spacing, one could use their results to obtain the number of
chains in Structure B. Moreover the changes of the intensities of the equatorial reflexions as the spacings increase would
be very informative. I don't feel much progress can be made with Structure B until the long-spacing position is cleared
up. I surmise the micelle structure is caused by the 10-fold screw axis of the fibre trying to give a hexagonal pack and
not quite making it.
I notice that your plate 5, paper 1, shows "spots". Are these due to salt in the specimen, or to some other cause?
Oster and Riley had a NaCl ring in one of their speciments, and also got a sharp 3.25 A reflexion which looks fishy to me.
The reason I ask is as follows. The equatorial spacings for Structure B appear to show one main spot (neglecting doubling
for the moment). If the structure is pseudo-hexagonal this is likely to be either + of -. If +, it means a great lump at
the origin of the C projection. This could only be phosphates near the centre. If -, it means a lot of material between
the helices (unless your suggestion of helices in the trigonal position is correct; even then I am not clear if it would give
the right answer). Part of this interhelical material could be the Na+, but this seems hardly enough. However, if some extra
salt were there this might be sufficient. Moreover, it would explain a lot of the puzzling density discrepencies which are
difficult to account model. Of course, you add distilled water when making the fibre, but it is quite certain that, in the
specimens which give X-ray photos, you have only the same amount of Na+ (or other icons) as Po4-? The dilemna is a very real
one because in effect one is deducing one "radius" from the general reflexions and another from the equatorials
This suggests that there is something non-helical in the structure. This is likely to be where adjacent helices interact.
A structure which is cylindrical contributes nothing to the general reflexions but alters the equatorials, and the most likely
thing to do this is the ions in the water.