VMD-L Mailing List

From: Rune Thomas Kidmose (rtk_at_biomed.au.dk)
Date: Mon Oct 22 2018 - 15:45:04 CDT

Hi joão,

I have also been having some troubles with glycosylations in charmm36 for a small MDFF pipeline I am making, so I was happy to read your answer to Steinar.

As you mentioned, I have already altered the "top_all36_carb.rtf" file by changing two RESI names" BGLCNA and BGLCN0" into BGLN.

I then added the modded topology file (together with all the other ones) using the -top flag with the autopsf command in my script

While this seems to work, as autopsf is now able to process my test PDB file containing a bunch of N-linked glycosylations (NAG and BMA only), I now got a problem with the bond between the ASN and the NAG/BGLN

The error from MDFF/NAMD is:

Reason: FATAL ERROR: UNABLE TO FIND BOND PARAMETERS FOR CC321 CT1 (ATOMS 19331 19329)

FATAL ERROR: UNABLE TO FIND BOND PARAMETERS FOR CC321 CT1 (ATOMS 19331 19329)
Charm++ fatal error:
FATAL ERROR: UNABLE TO FIND BOND PARAMETERS FOR CC321 CT1 (ATOMS 19331 19329)

So I am guessing I need to define this specific type of bond in a paramter file in order to tell NAMD how to process the ASN-NAG link?

How can I proceed with this? I am guessing ppl have already made such files maybe?

Currently I am including the following paramter files in my .namd file:

par_all36_lipid.prm
par_all36_prot.prm
par_all36_carb.prm
toppar_water_ions_namd.str
par_all36_cgenff.prm
par_all36_na.prm"

I would assume that the "par_all36_carb.prm" would be the one to alter? Just dont know exactly how.

Ruki

________________________________
Fra: owner-vmd-l_at_ks.uiuc.edu <owner-vmd-l_at_ks.uiuc.edu> på vegne af João Ribeiro <jribeiro_at_ks.uiuc.edu>
Sendt: 22. oktober 2018 18:46
Til: steinar.halldorsson_at_crick.ac.uk
Cc: Vmd l
Emne: Re: vmd-l: AutoPSF error with glycana (end of segment error)

Thank you, Steinar.

Autopsf renames NAG to BGLN automatically. In CHARMM36, there is no "BGLN" residue, as all NAG residues have more than 4 character length residue names (not allowed in the pdb format). To properly use autopsf with NAG, you would need to change the target residue name in the top_all36_carb.rtf file to BGLN, and feed it to autopsf. If your protein is heavily glycosylated (which seems it is the case), I would also consider building your system outside of autopsf, as the inclusion of glycans can be a little bit tricky, and we might probably need to increase the glycan support in autopsf.

I hope this helps

Best

João

On Mon, Oct 22, 2018 at 10:16 AM Steinar Halldorsson <steinar.halldorsson_at_crick.ac.uk<mailto:steinar.halldorsson_at_crick.ac.uk>> wrote:

Hi Joao

I probably should have checked there before, here is the error message:

-----
psfgen) building segment AG1
psfgen) reading residues from pdb file fitted_1flc_autopsf-temp.pdb_AG1.pdb
psfgen) unknown residue type BGLN
psfgen) extracted 2 residues from pdb file
psfgen) setting patch for first residue to none
psfgen) setting patch for last residue to none
psfgen) Info: generating structure...psfgen) unknown residue type BGLN
failed!
-----

Segment AG1 is the first glycan encountered and the three types of glycans there are NDG, NAG and a BMA. I'm not sure if BGLN corresponds to NDG?

Thanks,

Steinar

________________________________
From: João Ribeiro <jribeiro_at_ks.uiuc.edu<mailto:jribeiro_at_ks.uiuc.edu>>
Sent: 22 October 2018 16:05:01
To: Steinar Halldorsson
Cc: Vmd l
Subject: Re: vmd-l: AutoPSF error with glycana (end of segment error)

Hi Steinar,

Would it be possible to send a few lines from the VMD terminal window before these autopsf lines? Sometimes the cause for this error is printed before these lines.

Thank you

Best

João

On Mon, Oct 22, 2018 at 8:58 AM Steinar Halldorsson <steinar.halldorsson_at_crick.ac.uk<mailto:steinar.halldorsson_at_crick.ac.uk>> wrote:
Hi all

I'm fairly new to VMD so this may turn out to be a trivial problem..!
I'm using it for MDFF and I have a crystal structure of a glycoprotein which I would like to fit into a cryo-EM density. The trimeric protein has six modelled N-linked glycosylation sites per monomer and each site has a chain of three sugars, so there are lots of sugars around.
Following the MDFF tutorial, when it comes to the AutoPSF generation I get this error:

------

ERROR: failed on end of segment
MOLECULE DESTROYED BY FATAL ERROR! Use resetpsf to start over.
ERROR: failed on end of segment
MOLECULE DESTROYED BY FATAL ERROR! Use resetpsf to start over.
    while executing
"segment $segid {
      pdb $segfile

      # We alias the C-terminal OXT atoms to OT2 so that psfgen has to guess one atom less.
      # Otherwise psf..."
    (procedure "psfsegments" line 37)
    invoked from within
"psfsegments $logfileout"
    (procedure "::autopsf::afterchains_gui" line 66)
    invoked from within
"::autopsf::afterchains_gui"
    invoked from within
".autopsf.chains.finish invoke"
    ("uplevel" body line 1)
    invoked from within
"uplevel #0 [list $w invoke]"
    (procedure "tk::ButtonUp" line 22)
    invoked from within
"tk::ButtonUp .autopsf.chains.finish"
    (command bound to event)

-------

When I run the AutoPSF without the glycans it's fine.
I guess this has something to do with how the AutoPSF tries to define different chains? There is information in the PDB file about links between the sugars and with the ASN, and VMD does display these links correctly.
The types of sugars are NAG, NDG, BMA and MAN

Has anyone run into a similar problem and found a solution? Or maybe someone can point me in the right direction of how to solve this problem?

I really appreciate any help!
Cheers,
Steinar

The Francis Crick Institute Limited is a registered charity in England and Wales no. 1140062 and a company registered in England and Wales no. 06885462, with its registered office at 1 Midland Road London NW1 1AT 

--
..............................................................
João Vieira Ribeiro
Theoretical and Computational Biophysics Group
Beckman Institute, University of Illinois
http://www.ks.uiuc.edu/~jribeiro/
jribeiro_at_ks.uiuc.edu<mailto:jribeiro_at_ks.uiuc.edu>
+1 (217) 3005851
The Francis Crick Institute Limited is a registered charity in England and Wales no. 1140062 and a company registered in England and Wales no. 06885462, with its registered office at 1 Midland Road London NW1 1AT
--
..............................................................
João Vieira Ribeiro
Theoretical and Computational Biophysics Group
Beckman Institute, University of Illinois
http://www.ks.uiuc.edu/~jribeiro/
jribeiro_at_ks.uiuc.edu<mailto:jribeiro_at_ks.uiuc.edu>
+1 (217) 3005851