AllosMod Help Page - Modeling Glycosylation

Protein glycosylation can be modeled as explained in this help page and in our paper (Guttman et al. 2013 Structure). For help to model non-glycosylated proteins, click here. This tool is meant to sample the range of conformations accessible to sugars. There are a few ways to run the server depending on the desired conformational flexibility of the protein and/or sugars:

1) Input: a protein without sugars. Output: a semi-rigid protein (motions will be limited to loops and surface side chains) with flexible sugars.

2) Input: a protein with sugars. Output: a flexible protein with rigid sugars.

3) Input: a protein without sugars. Output: a flexible protein with flexible sugars.

For option 1, input the protein sequence only and a file describing sugar connectivity (glyc.dat). See below for more information and for an example click here. For option 2, run option 1 and then use a single structure from the output as input for another run. In the input sequence, specify the protein sequence and a "." for each sugar. The "." specifies a block residue in MODELLER and will remain rigid. Such rigid bodies can cause problems during dynamics so care should be given here. You will also need to upload a file allosmod.py that was generated in option 1. For option 3, first run simulations of the protein only using AllosMod. Then take the resulting structures and do option 1 (one for each structure) using AllosMod. This can be best achieved using the batch mode (see below). Take all the glycosylated protein structures as input into the FoXS server, if desired.

HINTS:

1) Input your structure from option 1 into the first page and hit submit. Then copy the sequence for that file and hit the back button and paste the new sequence into the input field. Then you just need to make sure to add any HIS tags or other residues missing from the PDB structure. This procedure allows you to submit a job with minimal manual work.

2) For help making the glyc.dat file from a PDB file that includes sugars, see the GlycoSciences webpage and set the "Assign connections by atom distances" to "HETATOM only." This tool will also assess the structural quality of the sugars.

The following help page is organized to run batch jobs. Each input field in single landscape mode has a corresponding file/section below.

• Directories zip file

  Upload a zip file containing a set of directories for each glycosylated protein that you want to model. For an example directory, click here. Each directory should contain the following files:


PDB file(s): the template structures used to model the protein.

Alignment file: (align.ali) contains one entry for each PDB file. Another entry (named pm.pdb) pertains to the sequence to be modeled. This alignment file should be generated after an alignment procedure, as this alignment will be used to generate restraints for modeling. Multiple chains can be specified by using a "/" as a separator, the same specifications used in MODELLER. There are many ways to create an alignment file including: 1) MODELLER and 2) ClustalW.
***WARNING*** Small errors in the alignment can cause big errors during a simulation due to energy conservation problems. Make sure there are no misalignments in which adjacent residues are aligned far apart in sequence (alignment programs often do this at the beginning or end of chains).

Structure list: (list) contains all PDB files used as template structures.

Input file: (input.dat) contains one line per parameter as follows:
NRUNS = X

Number of models to generate.


ATTACH_GAPS = True/False (optional)
Gaps will be inserted into the alignment file at the sugar attachment sites. This allows more flexibility of the protein to accommodate the sugars (default is True) .


SAMPLING = X (optional)
moderate_cm -
(Default) Sampling is performed using simulated annealing at moderate temperatures. The energy function for the protein is composed of restraints from MODELLER while the energy function for sugar molecules is a combination of dihedral terms from CHARMM and harmonic restraints as described in the published paper (Guttman et al. 2013 Structure).



REPEAT_OPTIMIZATION = X (optional)
is the number of optimization steps to be performed. Default is one.


Glycosylation file: (glyc.dat, for an example click here ) contains one line per sugar monomer grouped by chains of monomers. The line with the protein-bonded monomer specifies the beginning of a chain and is followed by all monomers within the chain. Three columns define the sugar types and connectivity: 1) monomer name, 2) O1 bond type, and 3) O1 attachment residue index:


monomer name:

NAG - b-N-Acetyl-D-Glucosamine
NGA - b-N-Acetyl-D-Galactosamine
GLB - b-Galactose
FUC - a-Fucose
MAN - a-Mannose
BMA - b-Mannose
NAN - a-Neuraminic acid


O1 bond type:

NGLA or NGLB - axial or equatorial bond to ASN*
SGPA or SGPB - bond between alpha or beta position and SER*
TGPA or TGPB - bond between alpha or beta position and THR*
16ab - (i) 1->6 (i-1) axial at C1 and equatorial at C6
16fu - (i) 1->6 (i-1) axial at C1 and equatorial at C6
14bb - (i) 1->4 (i-1) equatorial at C1 and equatorial at C4
13ab - (i) 1->3 (i-1) axial at C1 and equatorial at C3
13bb - (i) 1->4 (i-1) equatorial at C1 and equatorial at C3
12aa - (i) 1->2 (i-1) axial at C1 and axial at C2
12ba - (i) 1->2 (i-1) equatorial at C1 and axial at C2
sa23 - og sialic acid alpha 2->3 equatorial
sa26 - og sialic acid alpha 2->6 equatorial


*entry specifies the beginning of a new chain

O1 attachment residue index:

For every new sugar chain, specify the amino acid index of the residue bound to the first O1 atom in the sugar. The amino acid index must correspond to the input sequence in the alignment, i.e. the first residue in the input sequence is 1 and the Nth is N. For sugars within the chain, specify the index of the sugar monomer bound to the O1 atom defined as such: the first listed monomer in a chain has index 1, the second listed monomer in a chain has index 2,...

The output is a zip file containing the same directories that were uploaded with subdirectories containing glycosylated structures named pm.pdb.B99990001.pdb.