DNA Structure: An Age of Refinement Part 2 of 6

Clive Delmonte clived at ndirect.co.uk
Tue Jan 26 08:15:02 EST 1999

DNA Structure Refinement using Constrained & Restrained Least Squares

We saw that in Part 1 that a natural desire to maximise the information that
could be extracted from sparse diffraction data from DNA and oligonucleotide
studies inevitably has led to the incorporation of double-helical features
into the LALS algorithm.   Structures refined using this algorithm have
informed the subsequent work in very many studies, often being used as
models based upon which new structures are refined.

A similar rationale seems to underpin the mode of application of the CORELS
approach to nucleic acid structure refinement (62, 63):

"We have used this method for idealizing a nucleic acid model..."   (62,
page 801), and

"The structure of this fragment (d(GGTATACC)) of DNA was determined ...  It
incorporated..... calculations assuming models of both B and A types of
DNA."  (63, page 229)

It was this structure which was then refined with CORELS.   We also see in
this latter quotation, as in the work of other research groups, the
assumption that the
crystallisation of short oligonucleotides gives rise to the same structure
in vitro as that formed by proteins when they synthesise and assemble high
polymer DNA in vivo, even though this proposition has never been proved.

In the years since 1953, a growing body of experimental results, illustrated
by the twenty DNA Structure Puzzles I selected for citation recently in
these newsgroups, has become impossible to explain using the DNA double
helical paradigm.

These days biophysics and molecular biology is increasingly constrained by a
model of duplex DNA which no longer offers helpful structural insights in a
wide range of situations and many researchers find themselves reporting
results in the literature without being able to offer any structural or
mechanistic insights to explain those results.  A wide-ranging review of
such literature reports up to 1990 has been published (64).

We find ourselves increasingly in a situation which calls to mind a remark
of Sir Lawrence Bragg:

"The important thing in Science is not so much to obtain new facts, as to
obtain new ways of thinking about them."

Oligonucleotide structures deduced by assuming duplex DNA to double helical
at any stage in the process are therefore likely to be compromised and can
only be used with caution, if at all, to explain experimental results in the
wider field outside oligonucleotide crystallography.

An entirely new possibility for duplex DNA has been described (64) and
visualised experimentally in STM (65, Figure 3b)

In An Age of Refinement Part 3 we shall consider the algorithm NUCLSQ.
62       A Structure-Factor Least-Squares Refinement Procedure for
Macromolecular Structures using Constrained and Restrained Parameters; J.L.
Sussman, S.R. Holbrook, G.M. Church & Sung-Hou Kim; Acta Cryst. A33 (1977)

AND NUCLEIC ACIDS; J.L. Sussman; pages 206-237 in Methods & applications in
crystallographic computing; Ed. S.R. Hall & T. Ashida; Clarendon Press,
Oxford, 1984 (ISBN 0-19-855190-8)

64        "Towards a New Structural Molecular Biology" by Clive Delmonte,
ISBN 0 9512276 0 2  (1991)

65        Scanning Tunnelling Microscopy of Nucleic Acids: G. Lee et al.;
SCIENCE Vol 244 (1989) 475 - 477
Clive Delmonte

More information about the Xtal-log mailing list