Dear Ed,
Have you considered the possibility that it could be a water
stepping in to fill the void created by partial decarboxylation of the
glutamate? That could be easily modelled, refined, and tested for its
ability to flatten the difference map.
Gerard.
--
On Wed, May 03, 2017 at 02:29:11PM -0400, Edward A. Berry wrote:
> Thanks to all who replied. Yes, in some of the other cases density between Cd
> and Cg is conspicuously weak. From preliminary tests it looks like dual
> conformations, in some cases together with correlated waters, will account
> for the density adequately. No Zebras here!
> eab
>
> On 05/03/2017 04:26 AM, Matthew Merski wrote:
> >Have you tried just a double conformation of the Glu? Its hard to tell in
> >your pictures but it looks like there might be less than perfect density for
> >the CG-CD bond? If you just try adding a second conf of the Glu that might
> >work too (and perhaps be a horse rather than a zebra making your hoofprints?)
> >
> >Matthew Merski
> >Crystallochemistry Laboratory
> >Univ. of Warsaw
> >
> >On Wed, May 3, 2017 at 6:29 AM, Edward A. Berry <ber...@upstate.edu
> ><mailto:ber...@upstate.edu>> wrote:
> >
> > I'm finishing up refinement of a 1.8A structure (R's 0.17, 0.20) , and
> > among the largest peaks in the difference map are small spherical blobs
> > that seem to be attached (1.46 A here) to carboxylate O's (Figures). Are
> > these likely artifacts? If not, how can I interpret/model them? One idea is
> > that the acid has reacted with peroxide from the PEG to make the
> > (hydro)peroxy-acid. I don't know how stable that would be, and I don't see
> > any peroxyglutamate in Ligand Depot or HIC-Up. Another guess would be acid
> > hydroxamate but I don't know how that would be generated. Methyl ester
> > seems to be ruled out by the proximity of the two water molecules (2.45 and
> > 2.48 A here) suggesting the mystery atom is an H-bond acceptor or donor.
> > However since the occupancy seems to be < 1, the waters may be there only
> > when the atom is not.
> > I guess another possibility is there is a lot of motion in the plane of
> > the carboxylate (up and down here) which cannot be modeled by my isotropic
> > B-factors. In some cases the green blobs appear on both sides of the
> > carboxylate (but that could also be alternate conformations of
> > peroxyglutamate).
> >
> > The difference map (mFo-DFc, green) is contoured at 3 sigma (.06
> > e-/A^3). The difference peak is 5.4 sigma (0.1 e/A3).
> > The 2mFo-DFc map is contoured at 1.5 sigma (0.1 e/A3). 2mFo-DFc density
> > extends to the difference peak if I contour down at 0.64 sigma (0.04 e/A3,
> > third figure).
> >
> > If I put an O atom there, link it with plenty of slack, and refine
> > occupancy, it goes to 1.54 A from the carboxylate O and refines to
> > occupancy 0.35, B-factor 15 (carboxylate O is 30). Now it is reached by
> > 2mFo-DFc density at 1.5 sigma (0.1 e/A3).
> > Any suggestions would be welcome.
> > eab
