generating homology models of reduced disulfide bonds
Hello,
I recently obtained the xray structure of my wildtype protein which I am using as my template. This protein has a redox active disulfide (an other structural disulfides). I would like to generate homology models in which the redox active disulfide bond is broken (the model would represent the cysteines in a reduced environment).
I am new to modeler, but what I have done is edited the template PDB (wildtype) to remove the active site disulfide (SSbond) and then generated a homology model using the modified wt PDB as the template. Is this the correct approach to generating a reduced active site model?
If this is not an acceptable approach what would you suggest?
Thank you for your time, Stephanie Schaefer
University of Delaware Biochemistry PhD Student 302-831-2658
On 2/1/12 2:27 PM, Stephanie Schaefer wrote: > I am new to modeler, but what I have done is edited the template PDB > (wildtype) to remove the active site disulfide (SSbond) and then > generated a homology model using the modified wt PDB as the template. > Is this the correct approach to generating a reduced active site > model?
Modeller doesn't use the SSBOND entries in the PDB file - it looks at the geometry instead: http://salilab.org/modeller/9.10/manual/node179.html
So modifying the SSBOND entries won't remove the disulfide bridges. Disulfide bridges are added automatically during homology modeling. If you really want to turn off this behavior, you can override the automodel.default_patches() method (which normally calls patch_ss_templates, above) with a script like the following:
from modeller import * from modeller.automodel import *
class MyModel(automodel): def default_patches(self, aln): # do nothing, not even disulfide patching pass
then continue as normal but use the MyModel class rather than automodel in your script (e.g. a = MyModel(env, alnfile=..., knowns=..., sequence=...)).
I should warn you however that Modeller builds models that structurally resemble the template (this is by construction - it's how template-based modeling works). So your model is likely to have a similar backbone to the template, and even though you're not forcing the disulfide bonds, the two cysteines will probably still end up being close.
Ben Webb, Modeller Caretaker
participants (2)
-
Modeller Caretaker
-
Stephanie Schaefer