Re: [ccp4bb]: ligands in a 2-fold axis

[Dirk Kostrewa <dirk.kostrewa@psi.ch>]

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

On Tuesday 03 December 2002 09:49, Eleanor J. Dodson wrote:
> ***  For details on how to be removed from this list visit the  ***
> ***          CCP4 home page http://www.ccp4.ac.uk         ***
>
> Rongsheng Jin wrote:
> > ***  For details on how to be removed from this list visit the  ***
> > ***          CCP4 home page http://www.ccp4.ac.uk         ***
> >
> > Hi there,
> >
> > My protein crystallized in the P21212 space group. It forms a homo-dimer
> > around the crystallographic 2-fold axis. There is one small molecule
> > ligand binds to each protomer. The interesting thing is that the 2-fold
> > axis goes through the middle of the ligand molecules, although the
> > ligand itself is not perfectly 2-fold symmetric. It ends up with two
> > ligands overlap with each other while the 2-fold axis goes in the middle.
> > Does anybody know how to do the refinement (CNS?) without reducing to the
> > p21 space group?
> >
> > Thanks,
> >
> > Rongsheng
>
>  First - if this is happened your spacegroup is not P21212 - it is P21
> with an NCS axis relating the two protein molecules..
>
>  You must first process the data with P2/m symmetry -
> once you have merged h k l and -h -k l you have lost the information
> about the ligand orientation. (lets hope you collected enough redundant
> data to cover the larger asym unit!!)
>
>  Then impose NCS restraints on the protein and look for the ligand in
> the diff map..
>
> Eleanor

I think, this is not so clear! Crystallographic symmetry is never perfectly 
obeyed by macromolecules. We only make more or less safe approximations about 
perfect mathematical symmetry based on our intensity statistics. We all know 
of completely disordered side chains or two mutually exclusice conformations 
that can't obeye perfect symmetry and that we usually model with partial 
occupancies in the higher space group symmetry. The reason, why we do not 
process all data and treat all models in P1 is that the true crystal might 
have a close enough approximation to the mathematical symmetry, allowing us 
to refine an almost perfectly averaged molecule with fewer number of 
parameter, especially if we take the relatively poor data quality of protein 
crystals into account.  Maybe, I'm completely wrong, but I think in this case 
it could either be a combination of crystallographic and non-crystallographic 
symmetry as Eleanor suggests, but it could also be that it can be modelled in 
the higher space group symmetry as half of the ligands binding in one and the 
other half in the other conformation, especially given that this ligand has 
an internal pseudo-twofold symmetry (I don't think that molecules "know" 
about symmetry - they just form solid states with the lowest free energy 
state which could well be degenerate). The underlying question is - when does 
a deviation of perfect crystallographic symmetry justifies a treatment in a 
lower symmetry and when can we approximate it as having higher symmetry with 
a few "multiple" conformations?

I'm looking forward to a stimulating debate!

Best regards,

Dirk Kostrewa.

-- 
****************************************
Dirk Kostrewa
Paul Scherrer Institut
Life Sciences, OSRA/007
CH-5232 Villigen PSI, Switzerland
E-mail: dirk.kostrewa@psi.ch
Phone: +41-56-310-4722
Fax: +41-56-310-4556
WWW: http://www.sb.psi.ch
****************************************