Disulfide bonds restraints aren't respected

Dear discussion board,

I'm trying to get a homology model of a GPCR using I-TASSER.

I know for sure that there is a disulfide bond between two cysteine residues (161 and 173), but the prediction server doesn't seem to respect the restraints I give him (attached in this post).

I've tried various methods to try to force the two sulfur atoms together:

  • Setting the distance constraint of the two sulfur atoms to 2 Å (jobID: S451021), getting a final distance of about 4 Å
  • Setting the distance constraint of the two sulfur atoms to 1 Å (jobID: S451412), hoping that this would push them closer together
  • Adding a CONTACT constrain between the two residues (jobID: S451418) as suggested by this post in this forum, but to no avail.

Am I missing something?

Thank you in advance,

Marco