[phenixbb] disulfide bonds in phenix.refine

Allister.Crow at uea.ac.uk Allister.Crow at uea.ac.uk
Fri Jan 18 08:31:14 PST 2008

Hello all,

I am currently trying to refine a structure (with phenix.refine) that
contains a number of disulfide bonds.

Many of these disulfides appear to be partly broken such that the default
restraints given to these disulfides by phenix.refine cause the cysteine
sulfur atoms to be placed too close to one another.  This results in
difference density in the fo-fc map.  I want to try to model these
partially broken states a bit more accurately, either by modelling two
alternative conformations (i.e. one cysteine pair in a disulfide and the
other pair as free thiols) or by allowing the disulfide bond distance to

Currently refinement drives the S-S bond length close to dictionary values
regardless of how far apart I model the cysteines in my input pdb file and
whether the density implies a bridge or not.

My two questions are:

1) How do I specifically prevent a disulfide bond from being defined
between two sulfurs that are close together (I tried manipulating the
"disulfide_distance_cutoff" variable but this doesn't seem to stop a
disulfide being formed).

2)  How can I soften the restraint on the disulfide bond length for
specific cysteine residues.

Thanks for your time.

Dr. Allister Crow

More information about the phenixbb mailing list