I want to model the tetrahedral transition state of an enzyme where serine OG attacks the peptide backbone C atom. In this case the bond order of the backbone C=O bond changes to C-O(H) and the C-O(H) is not in plane anymore.
So far I wrote this patch file and added a hydrogen to the O atom in the input model which is named H01. The patch file is used since I get an error when I don't include it.
ADD_ATOM H01 Hpol HC 0.33
ADD_BOND H01 O
SET_ICOOR H01 180 70.5 1 O C CA
SET_ICOOR O 120.999985 70.000000 1.231015 C CA UPPER
SET_ATOM_TYPE O OH
SET_MM_ATOM_TYPE O OH1
However, after running flexpep, the O atom is still in plane with the peptide bond. What else would be needed obtain the 109° angle between O-C-CA?
To model the covalent bond and specific angles between the serine OG and the peptide atoms I am using geometric constraints with a high force constant (100000) to ensure that the transition state is modeled as accurate as possible. However there are still deviations beyond the allowed range. Would it make sense to additionally add AtomPair and Angle constraints or can the force constant be increased even further?