Quantum Mechanical Restraints (QMR) usage

Contents

Authors
Theory and Background
Running an example
- mmtbx.quantum_interface
- phenix.refine
Parameters
Literature

Authors

Nigel W. Moriarty

Theory and Background

In macromolecular crystallographic structure refinement, ligands present challenges for the generation of geometric restraints due to their large chemical variability, their possible novel nature and their specific interaction with the binding pocket of the protein. Quantum-mechanical approaches are useful for providing accurate ligand geometries, but can be plagued by the number of minima in flexible molecules. In an effort to avoid these issues, the Quantum Mechanical Restraints (QMR) procedure optimizes the ligand geometry in situ, thus accounting for the influence of the macromolecule on the local energy minima of the ligand. The optimized ligand geometry is used to generate target values for geometric restraints during the crystallographic refinement. As demonstrated using a sample of >2330 ligand instances in >1700 protein–ligand models, QMR restraints generally result in lower deviations from the target stereochemistry compared with conventionally generated restraints. In particular, the QMR approach provides accurate torsion restraints for ligands and other entities.

QMR is part of the Quantum Interface (QI) module that covers the interaction of Phenix with QM packages.

Running an example

There are two major ways to interact with the QMR modules in Phenix.

Run directly in phenix.refine by supplying a PHIL scope to specify the ligand details. This will modify the restraints in memory.
Run a command line program (mmtbx.quantum_interface) that creates the QMR PHIL scope (which also works for the phenix.refine option) and then using the same program run the QM step to write the restraints file for refinement.

Note: Using the python2 installer requires the setting of the PHENIX_MOPAC environmental variable to point to the MOPAC executable. The python3 installer is shipped with the MOPAC executables.

mmtbx.quantum_interface

For a number of reasons, we will cover the latter first. Using the example of PDB:227L, download from the PDB website using:

phenix.fetch_pdb 227L --mtz

Command to produce PHIL scope for QM minimisation using mmtbx.quantum_interface is:

mmtbx.quantum_interface 227L.pdb format=qi write_qmr_phil=True

which will fail because of the bulk test for hydrogen atoms.

Sorry: Model must have Hydrogen atoms

Also note that the write_qmr_phil=True option writes a PHIL file. In this example, the format of the PHIL file works for QI interface but the default works for phenix.refine.Add Hydrogen atoms to the model using:

phenix.ready_set 227L.pdb

which produces a file with hydrogens atoms – 227L.updated.pdb. It also may produce a restraints file that needs to be used next but in this case because the restraints for BNZ (benzene) are shipped with Phenix it is not necessary to supply to the program unless you want a different list of restraints (not ideal values). This file can be created cheaply but is needed as the skeleton for the QMR values to be inserted in the bond, angle and torsion restraints. So, we can repeat the quantum interface command with the new files.:

mmtbx.quantum_interface 227L.updated.pdb format=qi write_qmr_phil=True

This will prompt for the selection of the ligand which in this case is “5”.

 1 : "chain A and resid  173  and resname  CL"
 2 : "chain A and resid  178  and resname  CL"
 3 : "chain A and resid  901  and resname BME"
 4 : "chain A and resid  902  and resname BME"
 5 : "chain A and resid  169  and resname BNZ"

Enter selection by choosing number or typing a new one ~>

The result is a PHIL scope for BNZ QMR restraints:

227L.updated_A_401_BNZ.phil.

Alternatively, a selection can be supplied directly.:

mmtbx.quantum_interface 227L.updated.pdb format=qi write_qmr_phil=True qi.selection="chain A and resid 169 and resname BNZ"

A quick tour of the PHIL scope is helpful. The complete scope for the QMR module is shown at the end of this page. A reduced scope is shown here.

qi.qm_restraints {
 selection = "chain A and resid 169 and resname BNZ"
 buffer = 3.5
 write_restraints = True
 restraints_filename = Auto
 run_in_macro_cycles = \*first_only all test
 package {
    program = \*mopac orca
    charge = Auto
    multiplicity = Auto
    method = Auto
    basis_set = Auto
    solvent_model = None
    nproc = 1
    read_output_to_skip_opt_if_available = True
    view_output = None
 }
}

The key items are the selection, buffer and package items. The selection is the active part of the QM minimisation. The buffer is the radius from the selection to include in the QM environment. The package scope defines the 3rd party QM package details.

The program will also suggest the next command line for running the QM part of the restraints calculation. The skeleton restraints file may need to be included.

mmtbx.quantum_interface 227L.updated.pdb run_qmr=True 227L.updated_A_169_BNZ.phil

The result is a set of QMR restraints:

227L.updated_A_169_BNZ_3.5_C_PM6-D3H4.cif

Provide this to the desired refinements. The final QM calculation upon which the restraints are based are in:

227L.updated_cluster_final_A_169_BNZ_3.5_C_PM6-D3H4.pdb

The input and output files for the MOPAC calculations are in the directory for inspection.

mopac_227L.updated_A_169_BNZ_3.5_C_PM6-D3H4.out
mopac_227L.updated_A_169_BNZ_3.5_C_PM6-D3H4.arc
mopac_227L.updated_A_169_BNZ_3.5_C_PM6-D3H4.mop

phenix.refine

The restraints file 227L.updated_A_169_BNZ_3.5_C_PM6-D3H4.cif can be used in a refinement or the restraints can be calculated on the fly in the refinement run.

To get a PHIL scope for phenix.refine, remove the format=qi parameter from the write_qmr command.:

mmtbx.quantum_interface 227L.updated.pdb write_qmr_phil=True qi.selection="chain A and resid 169 and resname BNZ"

The refinement command can then be run.:

phenix.refine 227L.updated.pdb 227L.mtz 227L.updated_A_169_BNZ.phil xray_data.r_free_flags.generate=True

The major output is below including the name of the restraints file.

========================== Update restraints from QM ==========================

QM restraints calculations for macro cycle 1

Time for calculation of "chain A and resid 169 and resname BNZ" using PM6-D3H4 None: -

Updating QM restraints: "chain A and resid 169 and resname BNZ"

Core atoms

" C1 BNZ A 169 " ( 2624) " C2 BNZ A 169 " ( 2625) " C3 BNZ A 169 " ( 2626) " C4 BNZ A 169 " ( 2627) " C5 BNZ A 169 " ( 2628) " C6 BNZ A 169 " ( 2629) " H1 BNZ A 169 " ( 2630) " H2 BNZ A 169 " ( 2631) " H3 BNZ A 169 " ( 2632) " H4 BNZ A 169 " ( 2633) " H5 BNZ A 169 " ( 2634) " H6 BNZ A 169 " ( 2635)

Buffer residues

LEU A 7 ILE A 29 GLY A 30 PHE A 67 ASP A 70 VAL A 71 ALA A 74 ALA A 104 BNZ A 169 HOH A 266 HOH A 274 HOH A 404

Starting stats: bond: 0.024 angle: 0.80 dihedral: 0.00

RMS difference in entire QM model : 0.518

Interim stats : bond: 0.009 angle: 0.07 dihedral: 0.00

Writing PDB : qmr_cluster_final_01_A_169_BNZ_3.5_C_PM6-D3H4.pdb

Checking

Transfer

30 " C1 BNZ A 169 " - " C2 BNZ A 169 " 1.389 ~> 1.399 34 " C5 BNZ A 169 " - " C6 BNZ A 169 " 1.389 ~> 1.399 37 " C1 BNZ A 169 " - " C6 BNZ A 169 " 1.390 ~> 1.399 38 " C2 BNZ A 169 " - " C3 BNZ A 169 " 1.390 ~> 1.399 39 " C3 BNZ A 169 " - " C4 BNZ A 169 " 1.390 ~> 1.399 40 " C4 BNZ A 169 " - " C5 BNZ A 169 " 1.390 ~> 1.398 1 " C2 BNZ A 169 " - " C1 BNZ A 169 " - " H1 BNZ A 169 " 120.0 ~> 120.7 2 " C6 BNZ A 169 " - " C1 BNZ A 169 " - " H1 BNZ A 169 " 120.0 ~> 119.4 3 " C4 BNZ A 169 " - " C3 BNZ A 169 " - " H3 BNZ A 169 " 120.0 ~> 119.6 4 " C2 BNZ A 169 " - " C3 BNZ A 169 " - " H3 BNZ A 169 " 120.0 ~> 120.4 5 " C3 BNZ A 169 " - " C2 BNZ A 169 " - " H2 BNZ A 169 " 120.0 ~> 119.7 6 " C1 BNZ A 169 " - " C2 BNZ A 169 " - " H2 BNZ A 169 " 120.0 ~> 120.2 7 " C4 BNZ A 169 " - " C5 BNZ A 169 " - " H5 BNZ A 169 " 120.0 ~> 120.2 8 " C5 BNZ A 169 " - " C6 BNZ A 169 " - " H6 BNZ A 169 " 120.0 ~> 119.8 9 " C6 BNZ A 169 " - " C5 BNZ A 169 " - " H5 BNZ A 169 " 120.0 ~> 119.9 10 " C2 BNZ A 169 " - " C1 BNZ A 169 " - " C6 BNZ A 169 " 120.0 ~> 119.9 11 " C1 BNZ A 169 " - " C6 BNZ A 169 " - " C5 BNZ A 169 " 120.0 ~> 120.1 12 " C4 BNZ A 169 " - " C5 BNZ A 169 " - " C6 BNZ A 169 " 120.0 ~> 120.0 13 " C1 BNZ A 169 " - " C6 BNZ A 169 " - " H6 BNZ A 169 " 120.0 ~> 120.0 14 " C3 BNZ A 169 " - " C4 BNZ A 169 " - " H4 BNZ A 169 " 120.0 ~> 120.0 15 " C1 BNZ A 169 " - " C2 BNZ A 169 " - " C3 BNZ A 169 " 120.0 ~> 120.1 16 " C3 BNZ A 169 " - " C4 BNZ A 169 " - " C5 BNZ A 169 " 120.0 ~> 120.0 17 " C5 BNZ A 169 " - " C4 BNZ A 169 " - " H4 BNZ A 169 " 120.0 ~> 120.0 18 " C2 BNZ A 169 " - " C3 BNZ A 169 " - " C4 BNZ A 169 " 120.0 ~> 120.0

Finished stats : bond: 0.000 angle: 0.00 dihedral: 0.00

planarity (n=1) mean: 0.006 min: 0.006 max: 0.006

Writing restraints : qmr_01_A_169_BNZ_3.5_C_PM6-D3H4.cif

Parameters

qm_restraints
- selection = None selection for core of atoms to calculate new restraints via a QM geometry minimisation
- run_in_macro_cycles = *first_only first_and_last all last_only test the steps of the refinement that the restraints generation is run
- buffer = 3.5 distance to include entire residues into the enviroment of the core
- calculate = *in_situ_opt starting_energy final_energy starting_strain final_strain starting_bound final_bound starting_higher_single_point final_higher_single_point Choose QM calculations to run
- write_files = *restraints pdb_core pdb_buffer pdb_final_core pdb_final_buffer which ligand or cluster files to write
- protein_optimisation_freeze = *all None main_chain main_chain_to_beta main_chain_to_delta torsions the parts of protein residues that are frozen when an amino acid is the main selection
- restraints_filename = Auto restraints filename is based on model name if not specified
- cleanup = all *most None
- ignore_x_h_distance_protein = False
- capping_groups = True
- include_nearest_neighbours_in_optimisation = False include the side chains of protein in the QM optimisation
- include_inter_residue_restraints = False
- do_not_update_restraints = False For testing and maybe getting strain energy of standard restraints
- buffer_selection = None use this instead of distance from selection
- specific_atom_charges
  - atom_selection = None
  - charge = None
- freeze_specific_atoms
  - atom_selection = None
- package
  - program = test
  - charge = Auto
  - multiplicity = Auto
  - method = Auto
  - basis_set = Auto
  - solvent_model = None
  - nproc = 1
  - read_output_to_skip_opt_if_available = False
  - ignore_input_differences = False
  - view_output = None

Literature

In situ ligand restraints from quantum-mechanical methods. D. Liebschner, N.W. Moriarty, B.K. Poon, and P.D. Adams. Acta Cryst. D79, 100-110 (2023).