Klug here related results from his X-ray diffraction studies of chromatin, the name given to the chromosomal material when
extracted from the nucleus of cells in higher organisms (eukaryotic cells). Chromatin consists of DNA and an equal weight
of associated, basic proteins called histones that serve as structural scaffolding. Klug examined the structural organization
of chromatin in order to gain an understanding of how DNA is folded into the tight structure seen in a chromosome, how genes
are separated by function, and how their expression is controlled.
Nucleosomes, also mentioned in the letter, are distinct complexes of histone and DNA in eukaryotic cells (cells with clearly-defined
nuclei). Under the electron microscope they appear as bead-like bodies on a string of DNA.
Klug won the 1982 Nobel Prize in chemistry "for his development of crystallographic electron microscopy and his structural
elucidation of biologically important nucleic acid-protein complexes," chromatin and nucleosomes in particular.
NOTE: The margins of the second page are cut off in the original photocopied document.
Item is a photocopy.
Number of Image Pages:
3 (334,970 Bytes)
1976-09-28 (September 28, 1976)
Original Repository: Wellcome Library for the History and Understanding of Medicine. Francis Harry Compton Crick Papers
Thank you for your letter of 17 September, with the enclosed internal memo. The latter is very clear and it is good to have
your ideas written down.
I believe I may now have an explanation of the packing in the crystals which relates to other information and ideas on chromatin
structure. I would be very glad to have your opinion soon on this as I am considering mentioning these ideas in the paper
on crystals (so far not beyond the draft stage!).
The unit cell dimensions are a = 110 A, b = 193 A, c = 340 A, and all odd reflections are missing on the principal reciprocal
lattice lines, suggesting a halving in projection onto these lines. My interpretation is as follows. Columns of nucleosomes
parallel to the c axis are hexagonally packed to give the pseudo-hexagonal cell of the ab plane. Neighbouring columns are
staggered by half a nucleosome to account for the halving of the 00l reflections. There are three nucleosome cores in the
repeat distance of 340 A, successive particles being rotated through 120 degrees.
How does this relate to other ideas? Along a natural nucleofilament, the successive nucleosomes must be related by translation
only so that they can form a solenoid in which, as you have pointed out, all nucleosomes must be equivalent, relative to
the helical axis. Consider an ideal solenoid which has the minimum amount of DNA, i.e. just sufficient to make two turns
of helix, without the additional loop which you postulate as an excrescence looping around the H1 on the outside of the nucleofilament.
The number of base pairs to two turns is say 170 - 180 (we need to know the minimum repeat size in different species - physarum
is 171). Now what happens in a Zachau reconstitution which gives a repeat of about 140 base pairs? Presumably the nucleosomes
pack more tightly into the "screwed down" position and there will now be 140/170 x 2 turns per bead. We know that
the DNA does not have to be continuous for nucleosome association, so one assumes that the packing in the columns in the crystals
is the same as in Zachau's reconstituted material. If there were 1 2/3 turns of DNA per 140 base pair nucleosome core,
and the packing was such as to make the DNA geometrically continuous, then the 120 degrees screw along the 340 x axis would
come out naturally.
If this picture is right, then the inter-nucleosome spacing in the screwed down position is 113 A but, in the natural position
in a nucleofilament, this spacing would be increased by a factor of 6/5, i.e. to 130 A. If one takes the Sperling and Tardieu
determination of mass per unit length, indeed one finds this is the spacing of nucleosomes. Moreover, Chambon has always
quoted the diameter of a nucleosome (by which I think he means inter-nucleosome spacing) in the SV40 complex as 125 - 135
A. If this is correct, then the pitch of the DNA helix would be 130/2 = 65 A, a value not inconsistent with the apparent
helical character of the a axis photograph of the crystals (at face value here the pitch would be more like 340/6 = 58 A,
but this is dominated by the sampling). Some of the curves Linda Sperling obtained from nucleosomes in Paris show peaks more
in the 60 A region than the 55, the latter will of course be the second order of the spacing between turns of the solenoid
or columns of the crystal.
This picture is of course over-simplified. From looking at the X-ray photographs, it seems likely that the ideal structure
described above is perturbed. Moreover it would be neater if the nucleosome particles contained 150 base pairs, i.e. 1 6/9
turns of a helix kinked a la Sobell. This would also give a threefold character in projection onto the ab plane. The minimum
nucleofilament structure containing two turns would then have 180 base pairs. Is this all too neat?
Everybody is now back after holidays and conferences, but we miss Daniela, who will be at home with her baby for three to
six months, and Barbara has been ill, so we have not made much progress in finding out just what the nature of the histone
degradation is that makes the big crystals. I am pretty convinced that the packing in the microcrystals (i.e. the hexagonal
layers of which we had e.m.s and powder photographs) which are made by intact core particles, is different from that of the
large crystals. The simplest explanation is that neighbouring columns pack in a register to give a sheet structure, but the
packing isn't as neat so that the crystals don't grow as long.
More news. Uli Laemmli came and, by his new technique, produced some very long 300 A fibres which John Finch has looked at
in e.m. and by X-rays. The X-rays don't show anything new because, I think, the solenoids are twisted and need straightening.
One assumes that in the nuclei, solenoids are not perfect, but take up various modifications as the gene is packed. However,
the material is certainly a great advance, and we hope to be able to get a good estimate of the mass per unit length (we may
find how to straighten it out in solution) and also use it for making oriented specimens for X-ray work. The material is
fairly stable because one doesn't need to add magnesium: it seems to contain a fair number of non-histone proteins and
probably represents the most intact extracted material anybody has so far obtained. We are not advertising our work on this
new material because Uli Laemmli wants a breathing space to study the physical chemistry, and, likewise, we need to get the
feel of the material which will take a few months, but I think we do have a way in now to higher order structures in a more
However the number of things that needs to be done is still enormous. We need to know the mass per unit length of the Zachau
reconstituted material, the distance between nucleosomes in this and in natural nucleofilaments and so on, so that even dealing
with the ordered structures alone is going to keep our hands full.
Len is going to write up his stuff, but he is trying to obtain some good gel patterns, all taken on the same specimen, for
By the way, what do you think of this idea? It is quite striking that the DNAase I digestion pattern of the degraded nucleosome
cores is so similar to that of the particles with intact histones. On the assumption that the amino acids come from the N-terminus,
this would mean that the so called tails of the histones are not vital to packing the DNA onto the core - this will be done
by other interactions. What then are the tails for? Presumably to enable nucleosomes to pack side by side in the contacts
necessary to form solenoids. This would explain why we get a different packing in the crystals when the tails are removed.
Clearly to prove the idea will be difficult, but I am thinking of ways of testing this point.
Enjoy sunny California. Here it has been raining almost every day and the grass is green and high.