CHARMM-GUI_GROMOS_conversion_tutorial

CHARMM-GUI creates systems ready for simulation but with charmm parameters, this can be converted to the amber format using charmmlipid2amber.py. Conversion to a united atom forcefield like GROMOS-54A7 requires changing the number of atoms to a united atom represenation which is not supported by charmmlipid2amber.py.

The convert_molecules.py script developed here enables all-atom (AA) to united-atom (UA) conversion using templates, i.e. the included charmm_to_gromos.yaml file. The code does have checkes in place, but double check your conversions and pay attention to any warning about atom names emitted from grompp!

This tool also enables conversion of AA ligand pdb files to UA ATB files. I.e. if you have done ligand docking in all atom represenation and want to change the atom/residue names of the docked ligand based on ATB files.

Tutorial

Files for this tutorial may be found in tutorial-files

• Here I have generated a small CHARMM-GUI system step5_input.pdb (nAChR with nicotine bound and embedded in POPC membrane)
• The step5_assembly.str contains information about the system size
• The NCT_prot_het.pdb file was created by processing the protein through pdb2gmx and adding the ligand (nicotine) to the file.
  • Used the GROMOS-54A7 forcefield for the ligand
  • Generated parameters for nicotine using ATB, they can be found here

Inserting the UA-protein into the CHARMM-GUI system

1. Load both step5_input.pdb and NCT_prot_het.pdb into VMD (assuming they have molid 0 1 respectively)
2. Align the protein from NCT_prot_het on step5_input, commands to be entered in TkConsole

set ref [atomselect 0 "protein and name CA"]
set target [atomselect 1 "protein and name CA"]
set all [atomselect 1 all]
$all move [measure fit $target $ref]

3. Optional step, if your protein is not centered in the box, i.e. poking out the top, use the shift_solvent.tcl vmd scritpt to fix it
4. Save the adjusted NCT_prot_het structure -> NCT_prot_aln.pdb
5. Save the step5_input structure without the protein or ligand atoms -> step5_mem_ion.pdb:
   • Example selection: not (segid "PRO[A-Z]" "HET[A-Z]")
6. Combine the NCT_prot_aln.pdb and step5_mem_ion.pdb files:

cat NCT_prot_aln.pdb step5_mem_ion.pdb > NCT_comb.pdb

IMPORTANT: remove the END and CRYST1 lines between the atoms in the joined files!

Running convert_molecules.py:

convert_molecules.py NCT_comb.pdb NCT_comb_conv.pdb charmm_to_gromos.yaml

Note:

POPC conversion corresponds with (Poger et. al 2009)

Running GMX editconf

GMX editconf is required to set the box dimentions for running PBC simulations

• Extract info about system size from step5_assembly.str: A length: 111.618775, B length: 111.618775, C length: 150.889
  • If you have used the shift_solvent.tcl in step 4,
    add ~2 angstrom padding to the vertical C dimention, i.e. 150.889 -> 152.889
• For rectangular systems:

gmx editconf -f NCT_comb_conv.pdb -o NCT_comb_box.pdb -box 11.1522525 11.1522525 15.2889

• For hexagonal (triclinic) boxes

gmx editconf -f NCT_comb_conv.pdb -o NCT_comb_box.pdb -box 11.1522525 11.1522525 15.2889 -angles 90 90 60 -bt triclinic

IMPORTANT: if the charge is not zero, run genion or remove the appropriate number of ions for neutrality

Notes on creating templates:

Each line in the conversions is formatted as [source_numbering, source, template]

1. Open the united-atom and AA pdb files in a visualisation software like PyMOL and label the atoms by name
   • Helps to use grid mode
2. Copy the atom and residue names from the united-atom pdb file to the template column
   • Helps to use code editors like VS code with column selection mode
3. Copy the corresponding atom/residue names from the AA pdb file to the source column
4. Copy the corresponding atomic index from the AA pdb file to the source_numbering column
   • IMPORTANT: AA file must have atoms numbered from 1