ULTIMA - An Approach by Rabinovich & Shakked (72)
ULTIMA is presented as a computer program which, inter alia, can be used in
a manner so as to solve the structures of crystallised oligonucleotides ab
initio.
While this very obviously represents excellent current practice, some
detailed scrutiny might be warranted in the light of earlier Parts in this
series.
First, the process seems to have been influenced by knowledge of the double
helix model, whose conformation seems to have informed the distances apart
chosen by the researchers for the phosphate, sugar and base components (72,
page
197). Partly refined (double helical) structures were the result, an
outcome perhaps predetermined by the conformational library.
Next, during the refinement process, the partly refined structures from
ULTIMA were compared with idealised structures for high polymer DNA
published by Arnott & Hukins (73) which are informed by prior knowledge and
acceptance of the
double helix. Rabinovich & Shakked seem to have selected the ULTIMA
structure having the best fit to the Arnott & Hukins model and then refined
it further using CORELS (Part 3 of 6).
As with many other studies, a further, more general difficulty may relate to
the choice of unit cell dimensions, when there is a choice, since the cell
dimensions in (72) accord well with those chosen earlier by other workers
who knew they had to accommodate the Watson-Crick double helix in their unit
cells. The choice of unit cell dimensions, where there is discretion, will
determine, of course, the axial scales in Patterson and other maps and may
force algorithms to move towards a double-helical diameter to fill the
chosen unit cell.
The literature contains many examples of x-ray diffraction from true
oligonucleotide crystals where the reported patterns of systematic absences,
sometimes plus weak presences, would have allowed the original researchers
to have chosen a smaller unit cell, for example, 74 (A-DNA), 75 (B-DNA), 76
(B-DNA), 77 (Z-DNA) & 78, into which reduced unit cells a double helix would
not fit.
Systematic weak presences are supposed to arise from distortions of the
(large) unit cell which diminish the intensities otherwise to be expected.
However, their very existence could arise from distortions of a small unit
cell from whose diffraction patterns they would otherwise not have been
observed at all.
Crystallographers rarely discuss in detail the patterns of systematic
absence plus any systematic weak presences to be seen in their structure
factors, and the Brookhaven oligonucleotide structure factor files on
deposit often reveal patterns which would seem to warrant further close
analysis.
For the DNA double helix, such is "the intensity of (our) conviction that
(the) hypothesis is true", to use Sir Peter Medawar's phrase, that we may be
able to agree with John Stuart Mill:
"When has there been a dominion which has not appeared natural to them that
possess it?"
---------------------------------------------------------------------------
72 A New Approach to Structure Determination of Large
Molecules by Multi-dimensional Search Methods; D. Rabinovich & Z. Shakked;
Acta Cryst. Vol A40 (1984) 195 - 200
73 S. Arnott & D. Hukins; Biochem. Biophys. Res. Commun. Vol
47 (1972) 1504 - 1509
74 Molecular Structure of d(GTGTACAC) in the spermine-bound
tetragonal and the spermine-free hexagonal crystal forms; Sanjeev Jain, PhD
Thesis; University of Wisconsin at Madison (1990)
75 Double helix conformation, groove dimensions & ligand binding
potential of a G/C stretch in B-DNA; Udo Heinemann, Claudia Alings & Manju
Bansal; The EMBO J Vol 11 (1992) 1931 - 1939
76 Unusual helical packing in crystals of DNA bearing a mutation hot
spot; Y Timsit et al.; Nature Vol 341 (1989) 459 - 462
77 A Salt-induced Conformational Change in Crystals of the
Synthetic Tetramer d(CGCG); H.R. Drew, R.E. Dickerson & K. Itakura; J Mol
Biol Vol 125 (1978) 535 - 543
78 The Crystal & Molecular Structure of a Calcium Salt of
Guanylyl-3',5'-Cytidine (GpC); B. Hingerty et al.; Acta Cryst. B32 (1976)
2998 - 3013
------------------------------------------------------------
Clive Delmonte
