Hello All,
A grad student here is working on a structure, actually a series of
structures, that have in interesting feature that we are not sure how
to handle. She has taken a know protein and linked the N and C
terminal ends varying the linkage from being direct to a 10 residue
linker. The resulting crystals are in the same space group and
essentially same cell as the wild type.
When the student processed and solved the structures they appeared
to refine nicely when in P32 with 2 of the non linked molecules being
in the asymmetric unit. The wild type is in P3221 so running
Pointless we determine the the most probable space group really was
P3221 and not P32.
Now the problem. By going to the higher space group we drop down to
one non linked molecule in the asymmetric unit. If the linker is put
in fully, then the crystallographic symmetry produces what would
appear to be a cyclic protein, obviously not possible. What seems to
be going on is the some percentage of the unit cells have the linked
protein going in one direction in the cell and the rest of the cells
have it going in the other direction, since the structure as a whole
is just an average of all the unit cells and the protein does not
have a preferred orientation. When the occupancy of the terminal
end residues is dropped to 0.5, the linkage shows up beautifully.
The problem that is now cropping up is how does one create the
linkage so that the system can be refined and displayed properly? To
head off any suggestions that the linkages are not there, gels of the
protein show the linkage is there.
Any suggestions on how to describe this are welcome.
Cheers,
Len
Leonard M. Thomas Ph.D. New Address as of
November 2008:
Director, Macromolecular Crystallography Laboratory University of
Oklahoma
Howard Hughes Medical Institute Department of
Chemistry and Biochemistry
California Institute of Technology
620 Parrington Oval
Division of Biology
Norman OK, 73019
1200 E. California Blvd. MC 114-96
Pasadena, CA 91125
626-395-2453
[EMAIL PROTECTED]
http://www.br.caltech.edu/cmclab
