[gmx-users] Help: A ligand covalently bond to protein

[Mark Abraham]

Tanping Li wrote:
> Dear all,
> 
> I searched the mailling list, and still can't find a
> way to set up my system: a ligand covalently bond to
> protein.
> 
> I don't know if there is a easier way to do this. What
> I can think of is:
> 
> 
> 1) Add a bond I like in the specbond.dat file;
> 2) Add a block for ligand in the rtp file;
> 3) run pdb2gmx, gromacs will treat protein and ligand
> a one chain since it can see the bond in the
> specbond.dat.
> 
> 
> I just wonder if there is a eaiser way to add a ligand
> to rtp file? When I run Dundee PRODRG2, I get a .itp
> file. Is there some way to change this into rtp
> format? It is not trival to write a rtp manually.
> Really appretiate your help.

There's no automated way to convert an .itp file (which is a format 
intended for a whole molecule in a topology) to an entry in an .rtp file 
(which is a format intended for a residue in a heteropolymer). There are 
lots of similarities, however, since both are describing the same kinds 
of information. If you read chapter 5 of the manual carefully, and 
consider the existing contents of an .rtp file, you should be able to 
adapt the .itp suitably.

You should get your non-ligand-bound structure working properly first!

The reason there's no automated way to do any of this is that you are 
changing the environment for which the residue and ligand parameters 
were originally chosen, and there's no automated way to make sensible 
new decisions in general cases. That's what human judgement is for. 
Check out http://wiki.gromacs.org/index.php/Parameterization and 
http://wiki.gromacs.org/index.php/Exotic_Species, and be prepared to 
work hard!

Any charges produced by PRODRG won't necessarily be any good, because 
when making the covalent bond, charge groups (see manual) will likely 
break, and the environment next to the protein is very different from 
bulk solvent.

Mark