VMD-L Mailing List

From: Manas Kohli (manaskohli1_at_gmail.com)
Date: Fri Nov 20 2020 - 14:25:18 CST

Dear All,

I'm wondering if anyone has performed a simulation on Glycoproteins with
either NAMD or Gromacs. Specifically I'm trying to find a way to go about
doing this, what force-field and parameter files do people use since I'm a
bit confused/lost. I'm currently trying to simulate a membrane protein
covalently bonded to a sugar tree docked to another extracellular protein.
To summarize the steps I've undertaken:

1. I used the GLYCAM server to build the appropriate glycoprotein. From
what I've seen, the server uses its own format to build the sugar tree. So
for example, a NAG residue could be named 4YB. Atom Names are also
correspondingly different between the two - for example, there is a C8
atomname in typical NAG residues but C2N in 4YB.

2. I then changed the name of the residues back to apparently a more
standard PDB glycan residue name that could be recognized by the
protein-protein docking software HADDOCK. I changed 4YB for instance back
to NAG.

3. HADDOCK recognized the sugar residues but changed the atom naming from
the GLYCAM convention to their own. The sugar tree though was maintained
during the docking run and the protein seems to have docked in a good
conformation/one that I expect.

4. I then tried to build a corresponding topology for the glycoprotein. For
the glycan residues, I tried to use tleap from Amber-tools with GLYCAM
force-fields to build a topology in AMBER for the sugar residues. Then
there is another program to convert an AMBER topology to a GROMACS topology
which I planned to use in Gromacs. However, with the change in residue
naming and atom naming, this is where I'm stuck.

So I had two major questions from this

1) How exactly do I convert between a GLYCAM and PDB format to have an
appropriate naming for residues and atoms for both HADDOCK and then
correspondingly Amber-Tools

2) When I create a topology using the glycam force-field for the sugar
residues, how do I ensure I'm able to capture the covalent bond between the
Asn residue of my glycoprotein and the rest of the sugar tree.

I may be doing something completely wrong with this but I'm very confused.
If anyone has had experience in simulating glycoproteins, I'd appreciate
it! I don't personally mind mapping the atom names back from the PDB format
HADDOCK recognizes to the GLYCAM format but I had the major question as to
how to maintain a covalent bond between the protein and sugar trees in my
topology file if I were to isolate the glycans to use this with the GLYCAM
force-field.

Thanks and Best Regards,
Manas