Thursday, 20 October 2011

Anomalous patterson not consistent with systematic extinctions

From: Marta Ferraroni
Date: 2011/10/3
Dear all,

We collected some Xenon derivatives of a protein (an heterotetramer
α2β2)that seems to crystallize in P21 since the 0k0 reflections with
k=2n+1 are not present. However in the anomalous Patterson we found
strong peaks in the section v=0 yet none in the Harker section v=1/2.
Furthermore we weren't able to solve the structure both in P2 and P21
using these derivatives with the most commonly used programs.

The cell is 89 125 90 90 102 90 so a is approximately equal to c that
could permit pseudomerohedral twinning albeit the tests (Padilla-Yeates
and Britton) estimate a fraction of twinning around 0.05.

Data cannot be scaled as C orthorhombic even if the data reduction
programs (XDS and imosflm) assign a higher score to the related
orthorhombic cell with dimensions 112 139 124 90.00 90.00 90.00. In the
native Patterson map there are not strong peaks. According to the
Matthews coefficient the asymmetric unit contains 2 heterotetramers and
the self rotation function may indicate a 222 non crystallographic
symmetry with one two-fold axis perpendicular to the crystallographic
one (see figure attached).

Our questions are:

1) why the anomalous Patterson is not consistent with the space group?

2) Is there the possibility that the NCS could hamper the determination
of the correct space group and eventually determine a lower estimate of
the twinning fraction?

thanks in advance for your help

Regards,


Marta Ferraroni
Dept. of Chemistry
University of Florence
Italy


----------
From: Eleanor Dodson
Further Qs.

Do you have a noncryst translation parallel to the b axis (ctruncate will list any such translation..)

If the b shift is 0.5 then the 0k0 "absences" will be present whether the spacegroup is P2 or P21.

How many Xe sites do you expect? If there is only one then phasing is more difficult in monoclinic SGs - you have to break the centrosymmetry of the heavy atom distribution.

Do you have native data without Xe?

I always check for peaks which are consistent in the isomorphous and anomalous difference pattersons.

The NCS symmetry will doubtless make the exptl phasing more complicated, but it should help you with averaging later!
Eleanor

----------
From: Jacob Keller
Why not just go to P1, then build up the symmetry? Is completeness low?

JPK
--
*******************************************
Jacob Pearson Keller
Northwestern University
Medical Scientist Training Program
*******************************************

No comments:

Post a Comment