ADL: Modelling water

Garrett M. Morris garrett at
Wed Sep 8 15:04:52 PDT 2004

Hi Smita,

On Sep 8, 2004, at 7:53 am, Bhatia, Smita wrote:

> Hi everyone,
> The ligand binding site in my protein has 2 water molecules that could 
> be
> interacting with the ligand. How can I use autodock to model these 
> waters?

You can treat the waters as part of the protein, then compute a set of 
grid maps around the protein + water combination, but first you must 
add their hydrogens.  Placement of the hydrogens is particularly 
important for Thr, Ser and Tyr hydroxyls, as well as waters, of course. 
  Some PDB structures could also form better hydrogen bond networks if 
an Asn or Gln side chain's terminal O=C-NH2 groups are flipped 
(presumably because the electron density map couldn't unambiguously 
indicate which heteroatom is which).

A related point to consider is the protonation state of all your 
ionisable side chains: this will be determined by the pH and salt 
concentrations, and you should look into pKa calculations: there is a 
nice introduction at:

along with a list of software 

You will need to place the hydrogens in a reasonable place. There are 
at least 4 options:

(1) Use AutoDockTools to add hydrogens.

(2) You can use InsightII's Discover, AMBER, CHARMM, GROMOS, VMD or any 
other molecular mechanics/dynamics package to minimize the energy of 
the system, but remember to fix the heavy atom positions and just 
optimise the positions of the hydrogens.

(3) Alternatively, you could use WHAT-IF's WHAT_CHECK.

R.W.W. Hooft, G. Vriend, C. Sander, E.E. Abola, "Errors in protein 
structures."  Nature, 381, 272-272 (1996).

(4) A fourth option is you could use the PDB2PQR web server through 
APBS portal. This can optimize the protein hydrogen network (but can 
significantly increase computational time); and if you do this, you can 
choose to optimize the water hydrogen network.  See:

Dolinsky TJ, Nielsen JE, McCammon JA, Baker NA. "PDB2PQR: an automated 
pipeline for the setup, execution, and analysis of Poisson-Boltzmann 
electrostatics calculations." Nucleic Acids Research 32 W665-W667 

Remember we treat the protein with the United Atom model, which means 
we keep polar hydrogens; the non-polar hydrogens (attached to carbons) 
can be "merged" using AutoDockTools.  Merging means we add the partial 
charge from the hydrogen to that of the carbon it is bonded to, then 
delete the hydrogen.

I hope this helps,

Best wishes,


> I also think that Autogrid will have to be changed so it would 
> consider an
> oxygen bound to 2 hydrogens as a potential bond acceptor, allowing for
> correct treatment of water molecules. Can anyone suggest how to do 
> this?

AutoGrid can properly handle waters in the protein.

> Thanks a lot,
> Smita
> **************************************
> Smita Bhatia
> Institute for Biological Sciences
> National research Council
> 100-Sussex Drive, Room # 3031
> Ottawa, ON  K1A 0R6
> Phone: (613)990-0855
> Email : Smita.Bhatia at 
> <mailto:Smita.Bhatia at>
> **************************************
> ________________________________________________
> --- ADL: AutoDock List  --- 
> ---
Dr Garrett M. Morris, MA, DPhil

The Scripps Research Institute,       tel: (858) 784-2292
Dept. Molecular Biology,  MB-5,       fax: (858) 784-2860
10550  North Torrey Pines Road,       email: garrett at
La Jolla,  CA 92037-1000,  USA.

More information about the autodock mailing list