You know what, decarboxylation could be an issue. But that'll explain the B-factors only for the side chains and some of the atoms in the ligand.
I don't see in what other orientation/conformation the ligand can fit the density, not clash with itself or with the protein, lie on a 2-fold symmetry axis, and have two enantiomers in each asymmetric unit. The atoms that don't have density are terminal atoms. When the density is an average of four different species on a symmetry axis, that some atoms of a 16 atom-ligand stick out of density is frustrating but not entirely surprising, I think. What do you all think? The protein was co-crystallized with a racemic mixture of the ligand and I haven't been able to find any evidence that one binds better than the other, or one binds exclusively, not the other. On Thu, Jan 28, 2010 at 11:32 AM, Clemens Vonrhein < vonrh...@globalphasing.com> wrote: > Dear Sangeetha, > > before worrying about the B-factor column in your PDB file I would be > worrying about atoms not having density. This should tell you that > your model (atoms and/or parametrisation) isn't correct. > > One possible reason for those high-B side-chain atoms could be a > sequence shift: sometimes the difference between neighbouring residues > could be very small in terms of atoms and constellation. Just putting > the B-factors into a numerical range you are happy with doesn't make > your model better - it still doesn't fit your density. > > Same for your ligand: if you have one atom inside density with a B of > 40 and the next atom bonded to that one sticks out with a B-factor of > 200 ... are you sure your ligand is actually in there and has bound in > that orientation? > > So I would say these high B-factors are a good thing: they show you > where your model is poor. But then: you already see those atoms > sticking out of density, so you know where you still have to improve > your model, right? > > There could also be some issue with radiation damage (are those high-B > atoms on the ligand maybe I or Br?) ... decarboxylation on side-chains > etc etc. > > Just some ideas ... > > Cheers > > Clemens > > On Thu, Jan 28, 2010 at 11:04:52AM -0500, Sangeetha Vedula wrote: > > Dear all, > > > > I am refining a crystal structure with two enantiomers of the ligand > lying > > on a two-fold crystallographic axis (making the density an average of 4 > > orientations/optical identity). The ligand fits pretty well over all in > the > > density, but some atoms stick out of density. While B-factors for the > atoms > > that are within the density, the B-factors are in the range of 20-40. For > > the ones sticking out, the B-factors are 200-400. Which doesn't make > > chemical sense, as clearly, one atom isn't jumping around so much more > than > > the others. > > > > Is there any way I can force refmac to make the B-factors more > reasonable? I > > also notice that some side chain atoms that stick out of density have > > horrendous B-factors. > > > > Thanks in advance, > > > > Sangeetha. > > -- > > *************************************************************** > * Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com > * > * Global Phasing Ltd. > * Sheraton House, Castle Park > * Cambridge CB3 0AX, UK > *-------------------------------------------------------------- > * BUSTER Development Group (http://www.globalphasing.com) > *************************************************************** >
