>In a previous article, scavetta at XTREME.UCR.EDU ("Robert D Scavetta") wrote:
>>I'm trying to refine several Ca+2 binding proteins that have been crystallized
>>in the presence of either La+3 or Lu+3. The parameter files for these ions
>>give me the following values for eps and sigma (these values were someones
>>guess, and not a very good one). I know these are wrong!!
and matte at skyfox.usask.ca added:
>I am trying to do a similar thing with XPLOR, in refining terbium
>ions bound to a protein.
>Is using the electrostatic function and the CHARge
>specification in XPLOR a better way to handle this ? Any comments
>would be welcome.
Surely in the case of a lanthanide the metal position should be well-
determined by the electron density alone? Just because XPLOR is capable
of imposing restraints on cation geometry does not make it a desirable
thing to do. Unless you have some specific reason for imposing geometry
restraints I would say the best thing to do is simply to turn off all
energy terms involving the lanthanide during XPLOR refinement.
If you feel that this is causing the liganding atoms from the protein
to repel each other during refinement then you can turn off the
energy terms corresponding to interactions among that set of atoms
If the geometry of the metal site is of particular importance to you
then you may wish to go farther yet; not only would you remove any
restraints from the positional parameters for the metal, you would
model and refine anisotropic thermal parameters for this one atom.
Just for completeness I should mention that if you are modeling a metal
center at this level of detail you also need to be careful that the
f' and f" components of the atomic scattering factors are correctly
entered and that you have properly handled the F+/F- pairs of reflections
during data processing.
Ethan A Merritt
merritt at u.washington.edu