GLYCAM - Carbohydrate Parameterisations

GLYCAM - Carbohydrate Parameterisation - FAQs

1. Glycam

From: Jonathan Rhoad
Subject: Glycam
Date: 2003-11-19 6:20 pm

Rob,
Hello, my name is Jonathan Rhoad. I am not sure if you remember, but we
met at the Carbohydrate Gordon Conference this summer. I work with Todd Lowary.

I am attempting to do some MD calculations using Sander in Amber 7. How
do your your students input their molecules into LEaP? I have pdb files. I
examined the output of the MD runs, it seemed to Todd that the exo-anomeric effect
was not properly parameterized.

I looked at your GLYCAM, and thought, great, I will use that. I had noticed in my leap.log
that AC, EC, OG and OL were included in the atom types listed in leaprc.ff99. So I went to a
prep file and changed to atom types to fit the ff99 parameters, using AC for C1 and OG for my
glycosidic O. However, the MD runs give no output and never terminate. I thought
maybe it was because something was not right between GLYCAM and ff99, so I attempted a
run with only ff99 atom types, with the same result: no output, no terminations.

Is there something that I am missing when I change atom types? Or
should I take an entirely different approach? I am learning more in leap
and am considering just building the molecules from scratch.

Thanks

Jonathan Rhoad
E5-56 Gunning-Lemieux Chemistry Centre

"To be is to do"--Socrates
"To do is to be"--Jean-Paul Sartre
"Do be do be do"--Frank Sinatra

REPLY:
This is an I/O problem. From what you tell us, you could make the
input topology and coordinate files. If you had some problem in the
prep files or in the parameter file, you would not have been able
to prepare these files. Please check your local system settings.

2. Dermatan Sulfate params

From: Traian Sulea
Subject: Dermatan Sulfate params
Date: 2003-11-11 9:25 am

Dear Dr. Woods,
I'm trying to model a tetrasaccharide portion of the dermatan sulfate in
the binding site of an enzyme using the AMBER force feld. The sequence of
the oligosaccharide is:
[NAc-galactosamine-4-sulfate]-[iduronic acid]-[NAc-galactosamine-4-sulfate]-[iduronic acid]

Do you have any suggestion for AMBER-compartible parameters for these two
residues, or how those could be approximated from what is currently available in GLYCAM?

I really appreciate your help,

-Traian
==================================================================================================

Traian Sulea Ph.D.
Biotechnology Research Institute,
National Research Council of Canada,
6100 Royalmount Avenue Montreal,
Quebec H4P 2R2, Canada

Research Officer,
Computational Chemistry & Biology,
Tel: (514) 496-1924,
Fax: (514) 496-5143,
E-mail:Traian.Sulea@nrc-cnrc.gc.ca

==================================================================================================

REPLY:
Please refer to the following articles for the parameters involving sulfates:
ALTONA,J.COMP.CHEM,16(1),56-79(1995)
RAGAZZI,THEOCHEM,395/396,107- '97

Oxygens in the S-O terminal bonds are somewhat like the carbonyl oxygens whereas
the oxygen of the penultimate O-S bond is like an ester oxygen.

3. Glycam and GAGs

From: Duhsan Uhrin
Subject: Glycam and GAGs
Date: 2003-11-11 9:25 am

Hi Rob,

I am starting to do some modeling work on glycosaminoglycans (GAGs,
e.g.heparin) oligosaccharides.
Eventually NMR, but at the moment we want to interpret some time-of-fly
MS data, so gas phase at present.
I know very little about this, so I read some papers and stuff,
basically decided it would be a good idea to use your glycam
force field. I have a one year MSc student and we installed AMBER 7.0
and are learning how to use it.
I should have asked, but I have down loaded some stuff from your web
page, hope it is ok.

This is what I did so far:
- took your glycam_200a.dat file from your web page put it into the
right place, deleted the end bit, including the END statement
- looked for "glycam" statement various amber force fields and found
that the last one was ff99, so I copied it as leaprc file into
a new glycam directory and modified it to load your dat file:
# Load the main parameter set.
#
parm99 = loadamberparams parm99.dat
parm99 = loadamberparams glycam_2000a.dat
#

Is this ok? A related question is that the OH and OS atom types defined
in your dat file are also defined in the leaprc.f99.
However as I am reading your dat file afterwards does it mean your
parameters will apply here? Any conflicts?

Oh, another question. You say in your dat file that that these are for
Amber 5.0 parameters. Is it ok to use them with AMBER 7 (8?)
And another one. I am not sure if you are using explicit terms for H
bonds. I can see the in your dat file. As far as I understand, this was
dropped in ff95. How is it?

Anyway, when we loaded one of your prep files, it read ok, checked ok,
so I presume we can save it as topology file and staff.
So the next step would be try to run some energy minimization and MD
using your sugars to see that it is working in our hands.
Could you suggest a disaccharide which you have bench marked? Also, this
MS colleague of mine has protocols ho she runs
her peptide MD in gas phase, I am not sure how will these be behaving on
sugars.

Would it be possible to send me your library of prep files? Or maybe the
closes ones with the heparin such as aGlcNAc,
do you have acids? (glucuronic, iduronic). I will have to deal with the
end sugar, which has C4C5 double bond, I guess you will
not have that one.
As the next step we want add sulfate and sulfametes to the molecules
together with amber parameters
according to C. Altona, J. Comp. Chem. 16, 56-70 (1995). This is what I
am not sure about, what the mixing of charges will do.
I suppose if we get a reasonably looking conformers we could try to
calculate charges using Gaussian.

I will appreciate any comments you might have on this.

All the best,
Dusan

--

Dr. Dusan Uhrin,
Lecturer in NMR,
School of Chemistry
Joseph Black Building,
King's Buildings
West Mains Road
Edinburgh EH9 3JJ, U.K.

http://www.bionmr.chem.ed.ac.uk/~dusan
Phone: 44-131-650-7061
Tel: (514) 496-1924,
Phone: 44-131-650-7061

REPLY:
Partial charges calculated using the 6-31g(d) basis set with the option POP=CHELPG
are compatible with the GLYCAM parameter set. RESP charges are derived from
these partial charges. It is the RESP charges which should be used in the PREP
files.

4. Glycam Parameters

From: Ganesh Suryanarayanan
Subject: Glycam Parameters
Date: 2002-10-08 3:46 pm

Dear Dr. Woods,
We are working to determine the structure of a 21-residue glycopeptide
AHGVTSAPDT*RPAPGSTAPPA using SYBYL. In this glycopeptide Thr (T*) is
glycosylated by the trisaccharide GlcNAc-a-(1-4)-GlcNAc-b-(1-4)-GalNAc. Since
the GLYCAM parameters that you have provided are for 3-a-D-GalNAc
non-terminal residue, can you please tell us what should be the charge of
O4 (or how to determine its charge) in a 4-a-D-GalNAc non-terminal residue.
Thanking you,

Ganesh Suryanarayanan,
Graduate Student,
Eppley Cancer Center

REPLY:
Calculate the 6-31g(d) partial charges for the alpha-methoxy GalNAc
( ie with a methoxy group at the C-1 position). Calculate the net
charge on the methoxy group. Remove the methoxy group from the unit
and add the calculated net charge( of the methoxy group) to the C-1
carbon ie the anomeric carbon. Similarly, add the partial charge on
the 4-HO to 4-O. The net charge on the unbranched non-terminalresidue
is 0.0. For each branching, subtract 0.18 which will be compensated by
the reducing end terminal residue( of that branch) which has a net charge of 0.18.

5. RESP010 charges

From: Achintya Das
Subject: resp010 charges
Date: 2002-11-14 5:57 am

Dear Dr. Pathiaseril,
I am trying to model some carbohydrates using the resp010 charges. The charges
for glucose (ga.prep, etc in resp010) are missing on the website.
I would be very grateful if you could kindly provide these.
Thanking you

Achintya Das
Department of Chemistry
Indian Institute of Technology
New Delhi - 110016 India

6. Glycam Parameters II

From: Ganesh Suryanarayanan
Subject: Glycam parameters II
Date: 2002-12-13 5:57 am

Dear Dr. Woods and Ahamed,
I received the latest set of GLYCAM parameters that you provided.
Can you please brief me a little on the format description?
Previously, I had requested for the GLYCAM parameters for
4-b-D-glycopyranose non-terminal residue. Can you please provide me those?
Thanking you,

Ganesh Suryanarayanan,
Graduate Student,
Eppley Cancer Center

7. Some Glycam References

Woods, R. J.; Dwek, R. A.; Edge, C. J.; Fraser-Reid, B., Molecular
Mechanical and Molecular Dynamical Simulations of Glycoproteins and
Oligosaccharides. 1. GLYCAM_93 Parameter Development. J. Phys. Chem. 1995, 99, 3832-3846.

Use of ensemble averaging to derive charges:

Basma, M.; Sundara, S.; Calgan, D.; Vernali, T.; Woods, R. J., Solvated
Ensemble Averaging in the Calculation of Partial Atomic Charges. J. Comput. Chem.
2001, 22, 1125-1137.

Use of crystal simulations to derive RESP weighting:

Woods, R. J.; Chappelle, R., Restrained Electrostatic Potential Atomic
Charges for Condensed Phase Simulations of Carbohydrates. J. Mol. Struct. (Theochem) 2000, 527, 149-156.

Selection of 1.0 as scale factor for 1-4 interactions, based on solution rotamer populations:

Kirschner, K. N.; Woods, R. J., Solvent Interactions Determine Carbohydrate Conformation.
Proc. Natl. Acad. Sci. USA 2001, 98, (19), 10541-10545.

8. What is the philosophy used to compute partial charges in GLYCAM-04?

We use ensemble-averaged partial charges, selecting 100-200 conformations from the solvated MD run with an initial charge set then optimize the geometries (holding the exocyclic torsions in the MD confromation) at the HF/6-31G* level. The charges are then computed for each structure followed by averaging.

We reported earlier that the fluctuations that are often seen in the charges associated with aliphatic C and H atoms are artifactual, and result from the fact that generally there are more nuclei (especially in the case of apolar groups) than are required to get a best fit to the QM potentials. The solution we chose was to fit the charges to the molecule excluding aliphatic hydrogen atoms. This is not the same as simply adding the partial charges of the hydrogens from an all atom fitting into the carbon atomic charges!

Lastly, when performing the fit, we use the RESP program in AMBER and set the restraint weight to be 0.01 not the default 0.001.

9. How can I generate a prep file for a monosaccharide that is not currently available in GLYCAM?

This can be tedious, but if you start from a closely related prep file (say start with the prepfile for a-D-Gal to make a-D-Ara) then once the connectivities are OK, the only computational effort would be in deriving the charges. Read the related FAQ. Be sure to check that xleap connects your structure correctly to other residues.

At least the force field parameters should be there, even if a particular prep file has not yet been generated.

10. To get the charges on the monosaccharides

To get the charges on the monosaccharides we compute ensemble-averaged ESP-charges. It is quite a lot of work. But for the type of derivatives you are looking at, we would simply compute the charges for a simple analog of your derivative, and merge the charges.
For example, to add a methyl group to an OH, here is how to do it:
1) Choose two or three analogs of methyl ether (since that is what you are forming). For example: Me-O-Me, cyclohex-O-Me (ax and eq), i-Pr--O-Me
2) optimize all of the geometries at theB3LYP/cc-pVTZ quantum level
3) compute the ESP charges at the HF/6-31G* level using RESP with a RESP weighting of 0.01 and constraining all aliphatic H's to zero charge.
4) If all of the Me groups now appear to have approximately equal charges, then we would average them to get a "transferable" Me charge.
5) Remove the hydroxyl H from the monosaccharide and add the Me group and adjust the partial charge on the linking oxygen atom so tha the net charge on the residue is the same as on the original monosaccharide residue.

You would do the same for any other simple derivatives.

If this seems like a lot of work, then we could do it for you, but it would probably take 1-2 months to do the set of derivatives.

Lastly, yes GLYCAM06 is compatible with PARM99.