We are trying to solve the structure of a zinc finger by NMR. Using
CNS, we have put in the zinc as an extra residue and created new angles in
the .param file e.g. angles between the zinc atom and each of the ligating
cysteines. These angles were originally set out thus
angles s zn+2 s 50.0 109 (force constant is 50)
angles ch2e s zn+2 50.0 107
angles s zn+2 nr 50.0 107 (angle made by sulfur, zinc and nitrogen ligand)
angles cr1e nr zn+2 50.0 120
angles crh nr zn+2 50.0 120
These angles are based on the zinc and the 4 ligating residues being
arranged in a tetrahedral manner. Calculating structures, we found that
these angles were consistently violated by approx 10 degrees. However,
checking crystal structures of zinc fingers, we discovered that these
angles tended to be vary (approx 5 to 7 degrees) from the perfectly
tetrahedral arrangement above.
In an effort to get more realistic angles for our structure, we
lowered the force constants for these angles to 0. This lowered the angle
energy and the overall energy of our structures significantly.
Unfortunately, these new strucutres were no longer accepted structures in
CNS. Changing the force constant to 1, 5 and 10 did not help. Accepted
structures were being generated previously when the force
constant was 50.
Can someone tell us what the criteria for an accepted (as opposed
to trial) structures for CNS/XPLOR are and why our lower energy structures
are not accepted. Is it possible then to have a force constant of 0 for an
angle (ie not restrain it). Thanks very much.
Chu Kong Liew
Department of Biochemistry