I have a protein, where one of the reaction intermediate is covalently attached with a residue inside the active site. I need to parametrize this part of the protein. I used standard RESP calculations to get the charges and for the covalent bond I took the help of GAFF parameters. Is it a right way of dealing with this type of problem?
Hello Tarak,
In general, your approach sounds reasonable, but it is hard to say without more details. I assume that you obtained the partial atomic charges by using a molecule that capture the ligand and covalent fragment portion, in a conformation that resembles the active and bound conformation. A more rigorous way would to be to take a collection of possible conformations and determine an averaged charge set - but this requires much more time to do.
If you are using gaff parameters for modeling you ligand, then borrowing one to describe the covalent bond between the ligand and protein is reasonable too. However, usually when a bond is missing, so are angle and torsion parameters. You will need to be careful with how you define the atom types in the force field so that you can merge gaff and parmXX (e.g. c3-CT).
You are welcomed to use Wolf2Pack (www.wolf2pack.com) to do some parametrization or to validate the parameters that you choose. It has both gaff, parm99SB, and glycam force fields available. Depending on the atom types that compose your covalent bond, you might find a representative(s) molecule already present within the database. I also assume that you know about the online R.E.D. Server for computing partial atomic charges - if not, you should also check it out.
Best regards,
Karl
Hello Tarak,
In general, your approach sounds reasonable, but it is hard to say without more details. I assume that you obtained the partial atomic charges by using a molecule that capture the ligand and covalent fragment portion, in a conformation that resembles the active and bound conformation. A more rigorous way would to be to take a collection of possible conformations and determine an averaged charge set - but this requires much more time to do.
If you are using gaff parameters for modeling you ligand, then borrowing one to describe the covalent bond between the ligand and protein is reasonable too. However, usually when a bond is missing, so are angle and torsion parameters. You will need to be careful with how you define the atom types in the force field so that you can merge gaff and parmXX (e.g. c3-CT).
You are welcomed to use Wolf2Pack (www.wolf2pack.com) to do some parametrization or to validate the parameters that you choose. It has both gaff, parm99SB, and glycam force fields available. Depending on the atom types that compose your covalent bond, you might find a representative(s) molecule already present within the database. I also assume that you know about the online R.E.D. Server for computing partial atomic charges - if not, you should also check it out.
Best regards,
Karl
In addition to what Karl said, this (somewhat dated) tutorial might give you some ideas:
http://ambermd.org/tutorials/advanced/tutorial1/section1.htm
Thanks Dr. Kirschner, Dr. Christofferson, and Dr. Çelebi-Ölçüm for valuable suggestions and directing me towards a nice protocol.
- Badges
- Science topic