PROCHECK-NMR Operating Manual

8. Viewing restraint violations in 3D

A simple utility program called viol2pdb is provided with PROCHECK-NMR to allow you to visualize in 3D a given structure's distance restraints and their violations.

The program generates two output files, in PDB format, for each of the following sets of distance restraints:-

NOE distance restraints
H-bond distance restraints
S-S disulphide distance restraints

One file of each pair contains the satisfied distance restraints while the other contains those that are violated. Viewing these PDB files in a molecular graphics program (such as QUANTA or RasMol), perhaps superimposed on your molecule, can identify which regions are poorly defined and which contain significant numbers of restraint violations.

Various options allow you to specify which types of restraints and violations you want extracted.

Examples

Violated restraints

Satisfied restraints

Running viol2pdb

Before running viol2pdb you must first run AQUA to compute all the restraint violations in your structure and generate all the files that viol2pdb requires.

Then, to run viol2pdb, enter:

viol2pdb coordinate_file [ranges_file or model_number] [restraint_type(s)]

where:

coordinate_file is the name of the coordinates file containing the NMR ensemble.
[ranges_file or model_number] allows you define which model(s) and residue(s) are to be processed. You enter either the name of a ranges file or a single model number.
In the former case, the ranges_file must be present and its format the same as for the ranges files in AQUA and PROCHECK-NMR (see Selecting model- and residue-ranges). The only difference here is that, if no model ranges are defined, the program will assume you require Model 1, rather than all models.
The alternative to the ranges file is entry of a single model number (see examples below). All residues in that model will be included. You can also extract the restraints and violations for all the models in the ensemble by entering an asterisk (*) for the model number. However, the resultant output will probably be far too cluttered and confusing to be of any use.
If no ranges file or model number are entered the program will use the first model encountered in the ensemble.
[restraint_type(s)] allows you to limit the types of restraints extracted. There are three types:-
1. S = Short-range restraints (ie those within same residue)
2. M = Medium-range restraints (those between residues whose sequence separation is 4 residues or less)
3. L = Long-range restraints (those between residues whose the sequence separation is greater than 4 residues)
To define which you want extracted, enter the required combination of the letters S, M and L. For example, to extract only the long- and medium-range restraints you would enter: LM. By default the program includes all restraints (ie equivalent to SML).

Examples

For example,

viol2pdb 1nmr.pdb 2 LM

would run viol2pdb on Model 2 of the file 1nmr.pdb, extracting all the long- and medium-range restraints.

The following command:

viol2pdb 1nmr.pdb 1nmr.ranges L

would run viol2pdb on 1nmr.pdb using the models and residue ranges defined in the ranges file 1nmr.ranges. Only long-range restraints would be included.

Output PDB files

Each run of viol2pdb generates 6 output PDB files, as follows:-

                       Satisfied              Violated         
       NOE:        1nmr_noesatis.pdb     1nmr_noeviol.pdb 
       H-bond:     1nmr_hbsatis.pdb      1nmr_hbviol.pdb  
       S-S:        1nmr_sssatis.pdb      1nmr_ssviol.pdb

where the name of the original PDB file is 1nmr.pdb.

When viewed in QUANTA or RasMol the resultant sets of "bond-lines" should show the following:-

Satisfied restraints - are shown as white lines joining the two restrained atoms and having two blue lines extending from them at either end. The lengths of the blue lines indicate the difference between the upper-bound restraint and the actual distance between the two atoms in question (ie the distance between the furthermost tips of the blue lines corresponds to the length of the upper-bound restraint - which, of course, for satisfied restraints is longer than the actual distance between the atoms). See example.
Thus, if there are many long blue lines it may be that the upper-bounds have been set too loosely and there is room for tightening them.
Violated restraints - The violated restraints can be of two types:-
- Upper-bound violations which are shown as white lines tipped with red lines at either end. The length of the white line corresponds to the length of the upper-bound distance restraint while the red lines show by how much the actual distance extends beyond this to reach the atoms themselves. Thus the red lines end on the atoms. See example.
- Lower-bound violations are indicative of a bad contact and are shown by a white line joining the two atoms and having its central section in red. The length of the red segment in the centre indicates the size of the lower-bound violation.

Note that, other molecular graphics packages may show the restraints in colours that are different from those described above. Also, there may be problems with the way that the restraint lines themselves are linked by the program's "bonding algorithm" (see below).

Viewing the restraints

The restraints can be viewed by loading any, or all, of the PDB files output by viol2pdb into a standard molecular graphics package such as QUANTA or RasMol.

The line-segments corresponding to the red, white and blue lines described above are obtained by "bonds" between oxygen, hydrogen and nitrogen "atoms" in the PDB files created by viol2pdb. All the atoms are defined in HETATM records and the linkages between them are defined in the CONECT records at the end of the file.

For this reason, it is important that the molecular graphics package uses only the CONECT records in determining which "atoms" are "bonded". If it uses its own algorithm (eg on the basis of distances or residue types) the resultant set of bonds will not give the desired result.

Therefore, you may need to switch off the package's standard bonding algorithm when loading in the PDB files generated by viol2pdb.

In RasMol, there is nothing to be done. Just loading the PDB file straight will give the desired results. The main disadvantage of RasMol is that you cannot load in your protein structure at the same time. (Note that splicing the restraints file together with a protein structure PDB file doesn't work!). So you can only view the distance restraints on their own.
In QUANTA, you will probably need to alter the bonding algorithm as follows:-
1. Select the "Edit" option from the menu bar.
2. Select "Bond options".
3. Identify the structure(s) you wish the bonding algorithm to apply to.
4. Alter the "Minimum distance criterion" to 3.0.
5. Click on OK.
Sometimes QUANTA can behave erratically when the bonding criteria are changed, and may even apply the new criteria to all structures on screen, so a little patience and trial-and-error may be required to get everything right.