I am trying to find a way to examine whether the electrostatics are affected by a mutation 9 angstrom away of my proteins QM region . The mutation did not change any charge . The active site geometry of the mutant and wild type is almost the same . The first idea was to calculate charge density . I am using charmm36 for the simulations but I cannot find a way to calculate in that way the potential inside charmm.
Any suggestions on what I could calculate to prove that the electrostatics are not affected by the mutation ?
Thank you, to be honest I have already thought of that but the problem is that I do not know whether that is correct cause I want the electrostatics inside the protein not on the surface. The active site (QM region) is only in the A chain of my trimer ... I want to have a charge distribution with reference to volume for example.
I appreciate your answer
Regards, Ioanna
So you have (at least) two simulations, one with WT and one with mutated protein. And now you want to know if simulation of the mutated protein exhibits different charge distribution in the active site. Is this correct? Is all this analysis supposed to be time resolved? Surely the charge distribution will fluctuate quite a bit if your system is reasonably dynamic. Or are you interested only in mean changes?
I have at leat 60 trajectories for each system 500fs each .I did QM/MM molecular dynamics (squantum module, using PM3) simulations using CHARMM . I mutated two residues a GLU to ASP and a LEU to ALA so there is no change in 'properties' meaning charge or polarization etc , the new residues are just 'smaller'. Those mutations are 9 angstroms away from the QM region (cutoff 12). The geometry in the QM region is not really affected between the two systems. I know that there is change in partial charges and in electrostatics as those are dynamic systems , but I want to prove that the change is not significant. It seems that I cannot find a way to calculate electostatic potential or charge distribution inside CHARMM. I am trying to find inside the source code how charges, partial and Mulliken are being calculated...
Thank you for your time and suggestions, unfortunately this command is not working with the module squantum that I use but with an older one - quantum-.
So you did these simulations already and you did not write out atomic charges during the runs? In this case, your best bet is probably apbs indeed. Why would it matter that binding site and mutated residues are within the protein and not on the surface? Just restrict yourself to the relevant residues. Start by focusing on the mutated residues and estimate their electrostatic potential at the active site before and after mutation, subtract the apbs maps from one another and see what remains. If your intuition is correct, it is probably nothing.
The Zeta potential is an experimentally accessible electrokinetic property of the entire protein as a solute in an electrolyte solution. It is tenuously connected to some average electrostatic surface potential of the solvated protein, but is not a local internal property of the protein and therefore seems irrelevant to the initially posed question.
Apparently , there is no straightforward way to calculate the electrostatic potential inside my protein and especially of the qm region . Charmm forum and documentation is a mess and no matter how I tried I cannot even get the mullijen charges right . It is really weird though that I cannot find anyone who has done it . People who do enzymatic reactions and qm / mm simulations , especially when they do a mutation aren't they supposed to be able to prove whether electrostatics i
I would really propose you to use the APBS tool and pdb2pqr. Another method could be the gromacs tools g_potential/g_dielectric/g_dipoles. But for that you have to convert your structure/trajectory into a pdb + gromacs topology.
http://www.poissonboltzmann.org
- Badges
- Science topic