DNA Structure: Puzzle Number 12

Harry Powell hrp1000 at cus.cam.ac.uk
Tue Dec 15 05:11:53 EST 1998


On Tue, 15 Dec 1998, clive delmonte wrote:

> These days, the structure of oligonucleotides is deduced from diffraction
> from true crystals and purports to show double helical structures.
> 
true (in part), but...

> But there is a slight snag with this process.  There are no oligonucleotide
> structural solutions based upon direct methods on their own, because direct
> methods require data at a very high resolution, about 0.1 nm, 

Tosh. Although "direct methods" as used (and understood) by small molecule
crystallographers require data to that resolution, there is a basic flaw
to this, namely that you don't need direct methods or molecular
replacement to solve oligonucleotide structures. You can use the heavy
atom method in one of its modern guises, e.g. MAD. 

There are several oligonucleotide structures in the literature which have
been solved using this method without resorting to molecular replacement
and its requirement for a model.

check:

*>TI: MAD phasing strategies explored with a brominated oligonucleotide
      crystal at 1.65 angstrom resolution
  AU: Peterson_MR, Harrop_SJ, McSweeney_SM, Leonard_GA, Thompson_AW,
      Hunter_WN, Helliwell_JR
  JN: JOURNAL OF SYNCHROTRON RADIATION, 1996, Vol.3, No.Pt1, pp.24-34

> and the
> highest oligonucleotide resolution to date is about 0.15 nm.  

->TI: The crystal structure of a parallel-stranded guanine tetraplex at
      0.95 angstrom resolution 
  AU: Phillips_K, Dauter_Z, Murchie_AIH, Lilley_DMJ, Luisi_B
  JN: JOURNAL OF MOLECULAR BIOLOGY, 1997, Vol.273, No.1, pp.171-182

->TI: The bi-loop, a new general four-stranded DNA motif
  AU: Salisbury_SA, Wilson_SE, Powell_HR, Kennard_O, Lubini_P, Sheldrick_GM,
      Escaja_N, Alazzouzi_E, Grandas_A, Pedroso_E
  JN: PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES
      OF AMERICA, 1997, Vol.94, No.11, pp.5515-5518

(this last structure was refined with data to ~1.1A, but had 67% complete
date to 0.98A. It was solved with phases derived from the position of an
adventitious Ba++, so we didn't even use MAD...)

> All the
> oligonucleotide studies reported to date involve feeding the double helix,
> directly or indirectly, into the computational algorithms along with the raw
> data.

Most, maybe, but not all.

I could be _very_ picky about the term "raw data". I don't know any
crystallographers anywhere who use raw, unprocessed, data in their
structure solution or refinement, but I'm prepared to be
corrected ;-)

Harry
-- 
Dr Harry Powell, MRC Laboratory of Molecular Biology, MRC Centre, Hills
Road, Cambridge, CB2 2QH





More information about the Xtal-log mailing list