Hi,
I am currently working on an enzyme with zinc octahedral coordination. I want to calculate accurate partial charges for the enzymatic pocket in order to enhance docking. we will not be simulating bond formation or breaking. How can I possibly calculate those partial charges and are there parameters other than partial charges i should calculate to enhance docking ?
I did a DFT calculation only for the 6 Zn coordinating residues, minimized the enzyme first, selected the residues, ran DFT using Discovery studio Dmol3 (BLYP, Auto multiplicity, DNP basis set,), COSMOS solvation by chloroform, obtained ESP charges, hirshfeld, mulliken charges. for ex ESP charge for Zn was +1.28 instead of +2 in the forcefield. Is this procedure robust enough to parameterize the forcefield ?
Chloroform has a dielectric constant of 4 (most close to the internal environment of the protein) since the zinc is located in the pocket and its not embedded deep enough, is this correct ?
The enzyme code is 3W0T
Your procedure looks valid, but does not necessarily lead to better results.
Getting partial charges from QM for the active/binding site seems like a good idea and you defnitely will get more realistic partial charges (remembering, that partial charge is an artificial, abstract idea). The problems is, that your new partial charges have no guarantee of resulting in better binding poses in docking. The docking programs scoring functions are usually parametrized to work best with a certain set of partial charges (every docking program should list the preferred method for obtaining partial charges) and there is a good chance that more realistic partial charge would lead to worse docking poses.
In your case I'd first see which partial charges were used in the program you use for docking. Then, I'd take a few known cases (known ligand-protein complexes), redo the partial charges in your way and see if this approach improves the docking results. If so, you'll get a good validation of your methodology.
I have few ligands with good activity on bioassay, the docking score, however, doesnt sort them accurately (very weak correlation between IC50 and the docking score). I will dock those ligands on the QM parameterized receptor and check if they have better correlation. Is there any software that will dock using QM approach directly ?
- Badges
- Science topic
- Similar topics
- Biological Science
- Biochemistry
- Enzymology
- Enzymes