Preparing the protein

Revision as of 02:54, 5 December 2009 by Mysinger (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Preparing the protein

Items which are prefixed with 'AH' are relevant for docking HEIs to amidohydrolases and can safely be ignored for most metal-free proteins.

Modifying the PDB file

  • prepare rec.pdb by removing all lines that do not commence with 'ATOM', the chain column, all columns to the right of the z-coordinate and the TER statements.
  • treat all selenomethionines (MSE) as methionines (MET) by replacing the selenium atom (SE¤) with sulphur (¤SD). Be careful about the correct alignment!
  • atom enumeration does not matter, so don't bother to renumber after any of the following steps. Unique numbers are a good idea, presumably.
  • select the protonation states of HIS residues to be either δ- (rename residue to HID), ε- (rename residue to HIE) or doubly protonated (rename residue to HIP). HIS on the surface should be HIP. HIS residues coordinating the metal ions should have their protons pointing away from the ions. Base your decision on the immediate environment of the HIS residue: are there potential hydrogen bonds that can be formed?; are there charged residues close by?; would a certain protonation lead to clashes with other residues?; etc.
  • AH: the carboxylated LYS of subtype I is CYK, but this is not tolerated by startdockblaster5 , so store and delete the 3 surplus atoms and call the residue LYS.
  • AH: the more buried metal ion is ZB (charge 1.4), the other one ZA (charge 1.3). Atom names are right-aligned!

Running startdockblaster5

  • generate the file xtal-lig.pdb , which should only contain atoms of the MMFF. Prepare it in the same way as above: remove the chain column, all columns to the right of the z-coordinate and the TER statements.
  • generate the files .only_spheres and – in case you would like the matching spheres to be based on the heavy atoms in xtal-lig.pdb.useligsph and write `on' to the latter. Be careful not to add blank lines at the end, this will not be understood by . In any case, the entry in .useligsph will be ignored by .
  • on sgehead (or, as of dock67, on any machine), run startdockblaster5 to set up the data structure and copy all relevant files. It is a good idea to use csh and to source .login beforehand.
  • if startdockblaster5 doesn't finish for any obvious reason and with no clear error message, or rec.crg has very funny hydrogen placements, make sure that you have no non-printing characters in rec.pdb or xtal-lig.pdb. Do that by running your file through pc2unix rec.pdb. Check that your file is clean by looking at it with od -c rec.pdb | less . The only character with a backslash should be \n — you should see no \t, \r, etc. If this doesn't solve the problem, your best bet is to reprepare rec.pdb and xtal-lig.pdb from scratch — it is likely that there are some blanks or hidden characters that are causing the problems.
  • check the files stdout and stderr after the run for potential mistakes and error messages. Furthermore, verify that rec.crg still contains all atoms.
  • if you do not want to do anything special with the protein, like tarting some residues or modifying the spheres, go directly to chapter 3.

Removing and modifying files

  • go to ./grids and remove the surplus files from this directory (some would cause error messages from the subsequent programs):
    rm -f PDBPARM chem.* distmap distmap.log rec+sph.phi solvmap tart.txt OUT*
  • modify rec.crg:
  • AH: CYK: put the three missing atoms, delete the surplus hydrogens specific for LYS and rename the carboxylated lysine residue CYK.
  • remove all TER statements that might have been added.
  • AH: set the atom names of the metal ions to ZA and ZB and the residue name to ZN.
  • take care of disulfide bonds. Remove the thiol hydrogens (if they have been added) and change the residue name from CYS to CYX.
  • look at the box and maybe move it, so that the ligands won't stick out. Modify the 'center' and 'coordinates' statement in the preamble.
  • all residues and atoms have to be listed in prot.table.ambcrg.ambH and vdw.parms.amb.mindock, respectively ⇒ do not tart any residues in this file!

Running chemgrid

  • run chemgrid and check OUTPARM for the correct charges of all residues.
  • grep for 0.000 in PDBPARM: if any atom has this value in the 3rd and 4th column, it has not been recognized by chemgrid (because it is not listed in prot.table.ambcrg.ambH) and is thus ignored in the van der Waals-maps. There will be no other errors, the docking will finish showing some "bumping" ligands which have extremely favorable energies (≤ -200).
  • Another sign of a problem with atomic radii are any 'WARNING' issued in OUTPARM
  • if one has to run chemgrid again, remove PDBPARM OUTPARM OUTCHEM and chem.*.

Running distmap

  • the default is to run distmap on rec.crg. If you modified this file, rerun by simply typing distmap.
  • AH: cp rec.crg to rec-dist.crg and remove the Zn atoms in the latter file (otherwise there will be lots of bumping ligands). Edit INDIST to update the filename.
  • run distmap

Tarting the protein

  • cp rec.crg to rec+sph.crg and continue with the latter file.
  • tarted residues can be found in $DOCK_BASE/scripts/grids, they are the files with the extension prot2.
  • take care that the format of the .prot2 file is consistent with the format in the amb.crg.oxt file, e.g., that there is no leading space before an atom name etc.
  • AH: select the appropriate version of amb.crg.oxt depending on the subtype. Files are called amb.crg.oxt.N , where N can be I, III or VI .
  • AH: edit the residues in the binding site (i.e., all residues complexing the metal ions in the binding site), so that their names conform to the names of the modified residues in amb.crg.oxt.N
  • tart the residues that are in contact with a crystallographic ligand, if any.
  • AH: check that ZA and ZB, respectively (left-aligned in the atom column), have corresponding entries in amb.crg.oxt.N and vdw.siz.

Modifying the spheres

  • load match1.sph.pdb (i.e., the DelPhi spheres).
  • delete the spheres that are too close to the solvent.
  • (AH:) add spheres so that there is one sphere between the metals, several spheres around the metals and some spheres close to polar residues.
  • a good number for DelPhi spheres is 120.
  • append the spheres to the end of rec+sph.crg and put a TER statement after each sphere. Don't use tabs for whitespace, can cause problems with DelPhi!
  • do the same for match2.sph.pdb (i.e., the matching spheres); put at least one sphere between the metals and increase the sampling in the region around the metal ions by putting some spheres there. If you selected .useligsph . be careful not to move any spheres based on the ligand atoms.
  • a good number for matching spheres is 40.
  • run matchN.sph.pdb matchN.sph to generate the files that will be read by DelPhi/DOCK.
  • if desired, run sph/match2.sph in the parent directory of the docking run (i.e., .. to grids ) to put some color on your spheres.
  • in any case, put the preamble ("DOCK 5.2 ligand_atoms...") into match2.sph .

Running DelPhi

  • if necessary, modify so that all the paths and file names are appropriate.
  • run > delphi.log and check the output.
  • any 'WARNING' in the log is an indication that some atomic charges might not be correct.

Running solvmap

  • check that all atoms are present in rec.crg and run solvmap .
  • after the run, make sure that the file solvmap contains no blank lines.