You can obtain the complex's structure directly from the Brookhaven Protein Data Bank, using PDBBrowse. Note: some accession codes have been superseded, as indicated in the superseding PDB file by the new PDB record "SPRSDE", so it is a good idea to search with the ID set to the 3-letters of the desired accession code. Use the "Send entire text" option in the returned page, and then use the browser's "File:Save As..." option to save the PDB file, using the "Format for Saved Document:" as "Text".
If you have your browser set up to launch RasMol (which you should!), you can also take a look at the protein. Try using the "Colours:Chain" option, to see what the different chains are and where any ligands or heteroatoms are.
getready
"
to extract an information file, with the extension ".info", and split the PDB file by chain ID, as in this example:
% getready 4tmn.pdbThis script uses Mike Pique's Unix scripts
pdbsplitchains
, pdbwaters
and
pdbdewater
. The getready
script splits each chain
found into A_wat.pdb
and A_nwa.pdb
, and the
non-waters into
A_het.pdb
and A_atm.pdb
files.
You need to check that the files called A_het.pdb
,
B_het.pdb
, etc. do not contain more than just the ligand: sometimes they
also contain cofactors and metal ions. If this is the case, edit out the
the different parts and treat them appropriately. If there are cofactors
and metal ions, then in almost all cases, the resulting PDBQ files need
to be appended on to the protein PDBQ.
BIOPOLYMER BROOKHAVEN IN M1 \ /tmp_mnt/nfs/eve/export/eve6/gmm/desolv/dir/A_atm.pdb NO COLOR TYPE M1 BIOPOLYMER ADDH M1(*) ESSENTIAL_ONLY BIOPOLYMER FIX_END_GROUPS M1 CHARGED BIOPOLYMER LOAD CHARGES M1(*) KOLL_UNIThen save the protein using Sybyl "mol2" format.
mol2cnv
"
to convert into PDBQ format, for example:
% mol2cnv A_atm.mol2This script actually converts from Sybyl mol2 to PDBQ, sorts on residue number, sums partial atomic charges in residues and checks for non-integer residue charges, and writes out REMARKS informing the user of the total charges and any errors.
BIOPOLYMER BROOKHAVEN IN M1 \ /tmp_mnt/nfs/eve/export/eve6/gmm/desolv/dir/ligand.pdb NO COLOR TYPE M1
At this point you may have a collection of atoms that are unbonded. This can be fixed either manually, or automatically using "Build/Edit:Crystals..." on M1 and using the "CRYSIN command" "CONNECT" on all atoms, using the "NO_SYMMETRY_SEARCH" and the "DISTANCE_CRITERION" with the default bond distances. This should fix most bonds; if you use a maximum distance of 1.80Å; instead of 1.65Å; this may may form spurious bonds; delete any such erroneous bonds.
Once all the atoms are properly bonded, it will be necessary to define the atom types for most if not all atoms. Use "Build/Edit:Modify:Atom...:ONLY_TYPE" on all the set of "UNK_ATOMS", or unknown atoms; as each atom is highlighted, pick or type in the correct atom type.
FILLVALENCE M1(*) H MODIFY ATOM NAME M1((<H>+<H.spc>+<H.t3p>)) SEQUENTIAL_AUTO CHARGE "M1" COMPUTE GASTEIGER CHARGE "M1" DISPLAY LABEL ID M1(*) UNDISPLAY M1((<H>+<H.spc>+<H.t3p>))Finally, save the ligand as a Sybyl "mol2" formatted file, then run the script "
deftors
" to define any
torsions in the ligand. Keep the ligand visible in Sybyl, so you can see the
atom serial IDs, to help in the dialogue with AutoTors. Use deftors on the ligand mol2 file written out by Sybyl:
% deftors lig.mol2This will launch AutoTors, and you will be prompted as usual for the root atoms and which torsions to allow to rotate.