***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***


Hi All,

We recently crystallized a protein fragment that turned out be a straight
66-residue alpha helix.  The space group appears to be I4(1)22.  The crystals
diffract to ~2.3Å.  The native data set seems to be of good quality (very
redundant - 20, strong I/sigma, overall Chi square ~1.5, linear Rmerge 5-6%
etc).  The structure is solved by Se-Met sad (Solve and DM).  Because of the
large solvent fraction (70%), the experimental map is very good and building a
single continuous helix based on this map is unlikely generating major model
errors.  But after trying a series of refinement strategies in CNS, the
Rfree/Rw stuck at around 38%/36%.  If we switch to a lower space group (C2, a
subspace group of the I4(1)22), we can get better refinement  with a 4 fold NCS
(Rfree/Rw: 34.7/33.3%).  The helices in the crystal align approximately along
the same direction.  As a result, our diffraction images usually have very
strong diffusive X-shaped crossing streaks running around 8-10Å and 4-5Å. 
Although these streaks do not seem to affect data integration and reduction, we
apparently do have more overloads and high Chi square in low resolution bins. 
The refined Rfree also shows high value in low resolution range (50Å-10Å, 37%),
goes down a little bit around 7Å (33%), and spikes up to 40% again around 5-6Å,
and then goes down to 24% around 4Å and gradually rise to 47% in high
resolution bins (2.3-2.4Å).  I am wondering if the strong streaks due to the
parallel helices in the crystals are causing this puzzling behavior (to us) of
refinement.  Any ideas on how to solve this problem?

Thanks in advance for you help!

Lin
--
Lin Chen, Ph.D.
Associate Professor
Department of Chemistry and Biochemistry
University of Colorado at Boulder
Boulder, CO 80309-0215
Phone: 303-735-0071
Fax: 303-735-1347
e-mail: [EMAIL PROTECTED]
Web page: http://keres.colorado.edu/


Reply via email to