Thank you for your letter of 7th January. I have heard from Wyckoff who seems to think there will be little difficulty in
getting our visa interview rather earlier. I shall write to the Embassy as soon as I have cleared Michael with the Divorce
Court. Naturally I did not expect the travelling money before we got the visa. I will let you know later on how things can
most easily be arranged.
As to the apartment, I suggest you relax first the price, and go higher, say up to $120. If there are still difficulties
I think we should not worry too much about the neighbourhood. We certainly don't want to buy furniture, and we really
need three bedrooms if this is at all possible. It is really very kind of you to do this for us.
I was very pleased to hear about your various forms of ribonuclease. The work in Bernal's laboratory has suffered, in
my view, due to their excessive concentration on one form. Have you been able to derive any information from studying the
changes due to different salt concentrations?
As to L-keratin, I think Pauling's model is on the right lines, but probably for my reasons. A tilt of about 9 degrees
is certainly indicated, not 10 degrees as I suggested. The simple formula 5.4 is correct for a tilted helix, but not for
a coiled helix, as can be seen either from the exact theory which I have worked out, or from symmetry considerations. I have
written more fully about this to Lang, and he might show you the letter.
I was interested to hear about Dr. Luzzati, both from you and from Perutz. Some time ago I looked into the question of putting
the haemoglobin data on an absolute scale. The question is complicated by doubts as to how many of the atoms in the crystal
are ordered, and by the fact that the observed data only goes out to 2.8A degrees. I considered the integra1 Sp2dv. This
can be predicted, given sufficient resolution, from the chemical formula - since one knows the radial distance of all neighbours
within about 2.7A degrees of all the atoms irrespective of how the molecule folds up. One has to calculate the temperature
factor from the data, and it is difficult to do this unambiguously unless the recorded intensities go a good way out. I came
to the conclusion that one should always put the intensities on an absolute scale by direct measurement. Theory might then
be able to show what fraction of the atoms is disordered.
As to the L-helix in globular proteins, I don't think anybody here now believes that haemoglobin consists mainly of long
straight parallel rods, since such a model is clearly incomparable with the Patterson. The key question is whether it consists
mainly of 4-helices, not necessary parallel. This is plausible, since it may be made up of four non-parallel myoglobin-type
units, and moreover packing considerations show that L-helices of the same sense probably prefer to close-pack at about 20
degrees from parallel, for the reasons given in my Nature letter.
I have just drafted a paper in which I show that the total intensity in and around the 10 degree region is crudely comparable
with such a model. I use the relation
after applying a heavy artificial temperature factor to ensure convergence. This approach can be generalized by repeated
differentiation, and I shall be interested to see if Luzzati is working on similar lines.
I don't think we shall be able to come to any definite conclusion here until Kendrew has the 3 dimensional Patterson of
myoglobin, which we believe to be distinctly more "roddy" than haemoglobin.