[buster-discuss] Weird B-factors distribution in protein/DNA structure

[Miguel Ortiz Lombardía]

Dear Buster users and developers,

I'm analysing a protein/DNA structure refined with autobuster and have
trouble interpreting the distribution of the B factors: indeed, the
solvent atoms have a smaller average temperature factor than the
macromolecules.

Some information:

The space group is P3221 (confirmed by Zanuda) with cell dimensions 98,
98, 74. In some crystals the data can be processed in a cell with the
c-axis doubled (~ 150) Resolution is 2.4 A.

The structure contains two proteins (combined MW 20 kDa, but only 15 kDa
actually seen in the density) and a DNA double-helix (MW 9 kDa). The
solvent content is ~68.5% (VM 3.5)

The Self-Rotation function (SRF) shows some characteristics of 622,
perhaps due to the internal DNA symetry given that the DNA is nearly
perfectly perpendicular to the three-fold axis. No significant peak in
the native Patterson map.

The best data set, used for refinement, is consistently indexed with the
"short" c-axis in both autoPROC and xdsme. The first models after
molecular replacement showed alternate positive/negative residual
densities interweaved with the DNA bases. We thought at first that it
could come from a non-identified long c-axis. Attempts to reprocess the
data to force the long c-axis does not change this problem, which
eventually is solved once the model gets better and, more importantly,
TLS refinement (buster TLSbasic macro, one TLS group per chain) is used.
Current R/Rfree factors are 0.212/0.235.

All that could seem more or less all right, especially considering that
~25% of the protein mass is not seen in the density. However, the
residual map is quite noisy, especially around the DNA, I can find fewer
well-ordered water molecules that what I would expect at this resolution
(~ 60) and I have very high total temperature factors for some zones of
the protein (peaking at 180). Average B factors are ~90 for proteins,
~80 for DNA and ~70 for water.

I wonder if I'm missing something very basic here... or just worrying
for not that much. I would appreciate anyone's comments.

Thanks!
Cheers,

-- 
Miguel Ortiz Lombardía

Architecture et Fonction des Macromolécules Biologiques (UMR7257)
CNRS, Aix-Marseille Université
Case 932, 163 Avenue de Luminy, 13288 Marseille cedex 9, France
Tel: +33(0) 491 82 86 44
Fax: +33(0) 491 26 67 20
mailto:miguel.ortiz-lombardia at afmb.univ-mrs.fr
http://www.afmb.univ-mrs.fr/Miguel-Ortiz-Lombardia