40.1. Checking ligand fit to a receptor

Grid may be used after a novel ligand has been designed and fitted to its receptor. The ligand-receptor system will probably have been modelled and relaxed by molecular mechanics. A Grid map is then prepared using a water Probe with close-spaced grid points (eg: NPLA=5). The whole ligand-receptor complex is the Target, but directives TOPX... BOTY are used to restrict the computations to the region near the ligand itself.

The Grid map is then searched for places at which water molecules have NOT been displaced by the ligand from the site. These remaining waters should all be making appropriate hydrogen-bonding contacts to the atoms of the ligand, or to the receptor, or to neighbouring water molecules. However, it often happens that water molecules could be trapped in hydrophobic regions between the ligand and receptor molecules, at places where they cannot make appropriate hydrogen bonds. It is not easy to detect this possibility without using Grid.

These trapped waters are unfavourable and tend to destabilise ligand binding. The design of the ligand molecule should therefore be modified, so that the unfavourably trapped water molecules will be displaced when the new modified ligand binds.

40.2. Organised water

The representation of water as a HETATM ' O2 ' is particularly appropriate when the displacement of bound and organised water is being studied. The Type of a water ' O2 ' is automatically adjusted by Programme GRIN, in order to take account of the hydrogen bonding interactions which the water is already making. The Types of the other Target atoms (ie: their JTYPE values) and the Types of the other ' O2 ' waters are also adjusted, so that a complete water shell can be constructed round the Target, in which all the hydrogen bonds will be mutually compatible.

40.3. The design of mutant proteins

Consider a Protein of known three-dimensional structure with a glutamate residue at a certain position in the sequence. It is decided that a mutant might be prepared, with a lysine instead of the glutamate. However the question arises: "what will be the conformation of the lysine side chain".

The following procedure can give some insight about the answers to this question. The two side chains are:

  GLU:       -CH2.CH2.COO           with an anionic carboxy group and
  LYS:       -CH2.CH2.CH2.CH2.NH3.  with a cationic NH3+ group, 

and the first step when using Programme GRID is to model the glycine analogue; i.e. replace the -CH2.CH2.COO side chain of the glutamate residue with a hydrogen atom in the PDB file. This is straightforward, and the position of the hydrogen is uniquely defined.

Two separate GRID runs are then made on the glycine mutant, using a -CH2- and an NH3+ Probe. The NH3+ map is used to decide where the end of the lysine side chain might bind favourably, and the -CH2- map is used to show how the methylene chain of the lysine might be arranged. The probable positions of the lysine side chain can then be modelled and compared. Finally a GRID map with a water Probe can be used in order to see if any water molecules would be trapped between the methylene chain and the adjoining surfaces of the main protein.