how to handle non standard amino acids e.g. PTR

[michelsanner]

I have a kinase with the phosphorylated tyrosine PTR which I would like to keep

my first attempt was to
fixer1 = PDBFixer(filename="mypdb.pdb")
fixer1.findMissingResidues()
fixer1.findMissingAtoms()
fixer1.addMissingAtoms()
fixer1.addMissingHydrogens(7.4)
fixer1.removeHeterogens(keepWater=False)

but that last call removed the PTR residue (even though the ATOM records are ATOM not HETATM)

so I tried without calling fixer1.removeHeterogens() but then the following code

from simtk.openmm.app import PDBFile, ForceField, GBn2
forcefield = ForceField('amber14-all.xml', 'amber14/tip3pfb.xml')
system = forcefield.createSystem(fixer1.topology, implicitSolvent=GBn2,
soluteDielectric=1.0, solventDielectric=80.0)

fails with:
raise ValueError('No template found for residue %d (%s). %s' % (res.index+1, res.name, _findMatchErrors(self, res)))
ValueError: No template found for residue 443 (GLN). The set of atoms matches GLN, but the bonds are different.

and I can see that GLN443 is missing the peptide bond to PTR 444
and PTR 44 has only 3 bonds:

for b in ptr443.bonds(): print(b) ## missing C - N (443)
Bond(<Atom 7068 (C) of chain 0 residue 443 (PTR)>, <Atom 7082 (N) of chain 0 residue 444 (GLN)>)
Bond(<Atom 7068 (C) of chain 0 residue 443 (PTR)>, <Atom 7082 (N) of chain 0 residue 444 (GLN)>)
Bond(<Atom 7068 (C) of chain 0 residue 443 (PTR)>, <Atom 7082 (N) of chain 0 residue 444 (GLN)>)

so I tried to use replaceNonstandardResidues() to change PTR to TYR
fixer1 = PDBFixer(filename=recFilename)
fixer1.findMissingResidues()
fixer1.findMissingAtoms()
fixer1.findNonstandardResidues()
fixer1.addMissingAtoms()
fixer1.addMissingHydrogens(7.4)
fixer1.replaceNonstandardResidues()
with open('PTRtoTYR.pdb', 'w') as outfile:
PDBFile.writeFile(fixer1.topology, fixer1.positions, file=outfile, keepIds=True)

inspection of PTRtoTYR.pdb still show residue PTR

What am I missing here? what is the proper way to prepare this while for amber while keeping the PTR residue ?

thanks

-Michel

[peastman]

Do you have CONECT records for the nonstandard residue? A PDB file needs to have them for anything except standard amino acids, nucleotides, and water.

[michelsanner]

Hi Peter

Thanks for the rapid reply. Aafter adding the CONECT records for PTR it get

system = forcefield.createSystem(fixer1.topology, implicitSolvent=GBn2,
... soluteDielectric=1.0, solventDielectric=80.0)
Traceback (most recent call last):
File "", line 2, in
File "/home/sanner/local/miniconda3/envs/py37/lib/python3.7/site-packages/simtk/openmm/app/forcefield.py", line 1177, in createSystem
templateForResidue = self._matchAllResiduesToTemplates(data, topology, residueTemplates, ignoreExternalBonds)
File "/home/sanner/local/miniconda3/envs/py37/lib/python3.7/site-packages/simtk/openmm/app/forcefield.py", line 1391, in _matchAllResiduesToTemplates
raise ValueError('No template found for residue %d (%s). %s' % (res.index+1, res.name, _findMatchErrors(self, res)))
ValueError: No template found for residue 444 (PTR). The set of atoms is similar to DC, but it is missing 14 atoms.

using the following code on the attached pdb file

from pdbfixer import PDBFixer
from simtk.openmm.app import PDBFile

fixer1 = PDBFixer(filename='../2src_20pad.pdb')
fixer1.findMissingResidues() # required to be able to call findMissingAtoms
fixer1.findMissingAtoms()
fixer1.addMissingAtoms()
fixer1.addMissingHydrogens(7.4)

from simtk.openmm.app import PDBFile, ForceField, GBn2
forcefield = ForceField('amber14-all.xml', 'amber14/tip3pfb.xml')
system = forcefield.createSystem(fixer1.topology, implicitSolvent=GBn2,
soluteDielectric=1.0, solventDielectric=80.0)

[-Michel
2src_20pad.pdb.zip

[peastman]

The Amber14 force field only contains parameters for standard amino acids. If you want to simulate a nonstandard one, you'll need to provide your own force field parameters for it. Are you sure you really want to do that?

[michelsanner]

Well, as you know that In kinases the phosphorylation of tyrosine residues dictates activation and deactivation of which entails domain rearrangements. If I mutate the PTR to TYR I'd be running MD on active form using a crystal structure of the inactive form. I am not sure how much value there will be in that. The goal of the MD run is to study the stability of a peptide bound to the inactive form.

I recall reading a post from John Chodera specifically about this /export/export/people/sanner/python/collab/parang/tmp
where he mentioned "For example, suppose Im working with an AMBER variant that has the phosphorylated kinase residues PTR, and want to keep these?"

Amber Parameters have been reported for phosphorylated amaino acids
https://pubmed.ncbi.nlm.nih.gov/16240095/
http://amber.manchester.ac.uk/ # has the parameters .off files

How hard would it be in extend OpenMM with these templates ?

thanks

[jchodera]

The Amber14 force field only contains parameters for standard amino acids. If you want to simulate a nonstandard one, you'll need to provide your own force field parameters for it. Are you sure you really want to do that?

We just added the amber/phosaa14SB.xml file to the new openmmforcefields release! This should be out on conda-forge in a day or two.

You can also try amber/phosaa10.xml in the meantime if you're using older force fields.