Guided Ligand Replacement


Authors: Nigel W. Moriarty and Herbert E. Klei


The Guided Ligand Replacement (GLR) module is intended to facilitate ligand placement when the X-ray crystal structure of at least one similar protein:ligand complex is known. GLR is often employed in structure-assisted drug design where the structures of many related protein:inhibitor complexes are determined. It is especially useful for complicated ligands and when there are multiple independent copies in the asymmetric unit.

The Guided Ligand Replacement (GLR) tool is designed to use prior knowledge from a previously fit ligand to expedite the fitting of a "similar" ligand into the same or "similar" protein. The similarity required between the pair of ligands can be as little as a few atoms.

The proteins are overlayed using phenix.superpose_pdbs. The ligands are matched using the electronic Ligand Builder and Optimization Workbench (eLBOW) via another tool called phenix.superpose_ligands. Once the ligand has been fit, a Real-Space Refinement (RSR) is performed and the final model written to disk.



phenix.guided_ligand_replacement [options] \
  target_pdb_file_name=structure-in-which-target-ligand-is-to-be-placed.pdb \
  ligand_cif_file_name=parameterization-of-target-ligand.cif \
  map_coeffs_file_name=map-coefficients-for-target-electron-density.mtz \
  guide_pdb_file_name=structure-of-reference-protein-ligand-complex.pdb \
  ligand_selection_in_guide_model=selection-to-identify-reference-ligand \

Input files

GLR requires a number of inputs and also has some options, all of which are provided using the phil formalism. The list of inputs:

  • guide_pdb_file_name: Name of the PDB file for the structure of the reference protein:ligand complex. While the protein used in the target and reference structures is usually the same, the protein sequences need only be sufficiently similar to reliably superimpose.
  • protein_pdb_file_name: Name of PDB file in which the target (i.e. new) ligand is to be placed. This structure is typically the apo structure from molecular replacement after one round of refinement.
  • map_coeffs_file_name: Name of the map coefficient file for the electron density in which the target ligand is to be placed. This file is typically the MTZ file from latest run of phenix.refine. If this file contains more than one set of possible map coefficients, the first recognized set is used. If structure factors are provided (e.g. MTZ file with only FOBS and SIGFOBS), the 2Fo-Fc map is calculated and used. Map files (e.g. .map) are used directly.

The input for the ligand can be any input file that is readable by eLBOW. Either or both of the following phil parameters can be used.

  • ligand_input_file_name: this is a file containing chemical information about the ligand.
  • ligand_cif_file_name: this is the of a ligand restraints file for the RSR step. If not supplied, the restraints are generated from ligand_input_file_name.
  • ligand_code: the ligand code from the PDB Chemical Components so that restraints and geometry can be generated.

Optional inputs include:

  • ligand_selection_in_guide_model: Standard PHENIX selection expression to identify the reference ligand in the reference structure. A typical selection might look like "resname LG1 and resid 1". While GLR tries to make a sensible selection based on ligand code or uniqueness, it may be necessary to make an explicit selection.

  • replace_ligand_in_guide_pdb_file_name=[true|false]: By default, this parameter is set to false and can be omitted. If this parameter is set to true, the reference structure is also used as the target structure (protein_pdb_file_name parameter is ignored). This functionality is useful in two circumstances:

    1) the first copy of the ligand is placed manually and it needs
    to be propagated throughout the asymmetric unit
    2) the current ligand is incorrect (e.g. GTP instead of ATP) or
    needs to be updated (e.g. remove non-standard atom names).
  • phenix.superpose_pdbs: this allows selection of the same parameters as in the standalone program.

Output files

By default, GLR will generate a PDB file with the fit ligand.


Example 1

The reference FXa structure, 3KQE, and its reference ligand, LGM, is used to place the target ligand, LGK, in the target FXa structure 3KQC

phenix.guided_ligand_replacement \
  target_pdb_file_name=3KQC-apo.pdb \
  ligand_cif_file_name=LGK.cif \
  map_coeffs_file_name=refine-apo.mtz \
  guide_pdb_file_name=3KQE.pdb \
  ligand_selection_in_guide_model="resname LGM"

Example 2

In all of the above examples, the input can be provided through a PHIL file.

phenix.guided_ligand_replacement < glr.phil > glr.log


Ligand placement based on prior structures: the guided ligand-replacement method. H.E. Klei, N.W. Moriarty, N. Echols, T.C. Terwilliger, E.T. Baldwin, M. Pokross, S. Posy, and P.D. Adams. Acta Crystallogr D Biol Crystallogr 70, 134-43 (2014).

electronic Ligand Builder and Optimization Workbench (eLBOW): a tool for ligand coordinate and restraint generation. N.W. Moriarty, R.W. Grosse-Kunstleve, and P.D. Adams. Acta Crystallogr D Biol Crystallogr 65, 1074-80 (2009).

List of all available keywords