Fragmented Structures

In many structure determination problems one produces a model where the polypeptide chain is broken into fragments even though it is known that the ``real'' molecule is a single piece. This usually happens because the loops on the surface of a molecule cannot be seen in early electron density maps. Because of the flexibility of TNT there are many ways one can model this situation. This section will discuss two possibilities.

The simplest way to handle this problem is to define the sequence of your protein ignoring the issue of what you can and cannot see. Write RESIDUE statements which describe the chemical structure of your molecule. The parts you can't see, and haven't built, will be indicated as missing but refinement will continue unaffected. As you build more residues into the map you do not have to edit the sequence file.

The other method is to write a sequence file that describes not the molecule you have, but the molecule you have built. This sequence file will be easier to construct from a PDB file but that sequence file will require some editing to become functional.

As an example, suppose that you have two fragments; one numbered from 100 to 125 and the other numbered from 140 to 160. Your sequence file will look like:

RESIDUE 100 ALA 101 PEPTIDE
RESIDUE 101 GLY 102 PEPTIDE
             .
             .
             .

RESIDUE 125 GLY 140 BREAK
RESIDUE 140 GLU 141 PEPTIDE
             .
             .
             .

RESIDUE 160 TYR NULL BREAK
RESIDUE NULL NULL

In the standard TNT geometry library you will see that the link BREAK defines the geometry for a blunt end at the C terminus. It should be used whenever you have a C-terminus without the extra oxygen atom of a ``real'' carboxy terminus. Because BREAK is a link it must link to something. It doesn't matter what you link it to because it defines no restraints involving the ``linkee".

If you don't want to connect your fragments directly, as I have done, you can connect them all to a NULL or dummy residue, a residue with no atoms and no restraints. All residues must be defined on a RESIDUE statement. Therefore, if you create a dummy residue of some sort you must include a RESIDUE statement which defines its type. I would suggest that you link fragments together if you have some confidence of the connectivity, and link them to a dummy residue if you do not.

Even if you cannot see all of the atoms in the residues next to the break, you can still define the break in this fashion. You may receive warnings that some atoms cannot be found, but in this case the warnings can be safely ignored.

Wed Jul 5 13:21:03 PDT 2000