iodine scattering factors
csquire at CCU1.AUCKLAND.AC.NZ
Thu Nov 20 15:38:12 EST 1997
I have an interesting problem with my refinemnt of a DNA oligonucleotide
with a ligand binding in the minor groove.
The ligand contains a phenyl ring with an iodine substituent. Uisng the
X-PLOR library scattering factor for iodine results in very high temp.
factors for the iodine (80-90!) and all the atoms of the ring (60-70).
The temp factors for the rest of the ligand remain fine (20-30).
The phenyl ring fits nicely into a nice planar blob of e- density and the
phenyl temp. factors should not be this high. The ligand is not disordered
and the iodine fits a nice spherical blob of density.
If I change the scattering factor for the iodine to that for a lighter
element then the temperature factors not only for the iodine but for
all the phenyl atoms behave. The phosphorous atoms on the DNA have far larger
peaks in the elctron density than the iodine has. This is consistent in all
the crystals I have tried (6 in total). The data is at 2.4 ang. resolution
and the density from the iodine "smears" into the density from the phenyl
ring. I also have structures at 1.6 ang. resolution where this problem is not
encountered (although the phosphorous atoms of the DNA still have greater
density than the iodines) and the iodines are discrete blobs of density with no
"smearing" of the density into the phenyl ring.
So what is going on? How does X-PLOR deal with a situation where it is
expecting more density from the iodine than is observed?
In the 1.6 ang. res. data is the program recognising that the iodine is a
discrete sphere of density and thus limiting the temperature factor?
In the 2.4 ang. res. structure is the program trying to fit the iodine into
the phenyl ring density and thus disrupting not only the iodine temp.
factors but also the phenyl carbon temp. factors?
If the iodine density is a great deal less than the DNA phosphorous density
(as I have observed) and the ligand appears to be ordered and full
occupancy can I be justified in "fudging" the scattering terms for the iodine
in the ligand?
This sounds like a very dodgy idea to me so if anybody has
any other ideas I would greatly appreciate their input.
University of Auckland
More information about the X-plor