PDB FILE, homology modelling

Question

[Joe]

Hi,

I am abit confused with PDB data, too many links there to find out what is going on with the file I'm interested with.

I downloaded the PDB file for a homologus protein in cow (3Qz1), it seems it has 4 chains in the structure and they all look identical . The enzyme is suppose to be monomeric AFAIK. Perhaps the crystal has 4 of them aggregated together. There is no paper on the structure released though.

So, if I would like to do homology modelling using Discovery studio, I have to do alignment first, but it seems the alignment command (human vs cow ) using the software showed that it aligned to only one of the chains.

Perhaps there is a command feature in DS that allow me to exclude the other 3 chains.

TIA for comments

[Joe]

Hi Joao,

Thanks for the reply. There is one more thing at the moment Id like to ask you/ppl on board here..

If there are crystal structures of an an enzyme which can bind to a few substrates and effector molecules

Which one would be better taken for homology modelling? Is it the enzyme without anything in it.,.. or with the substrate..? Or enzyme with +effector and substrate?

I guess perhaps all should be taken into account to have a better functional insight?

TIA

[João Rodrigues]

Hey Joe, I would have opened a new question or commented my reply.

Depends on your goal. If you want to do docking and the enzyme you are studying has a similar compound in the same binding pocket, then take the bound structure.

Answer 1

[João Rodrigues]

Hi Joe,

The PDB files do not contain the biological assembly (i.e. your case, the monomer) only. If you want this, which might make things easier, go to the RCSB PDB website, find your structure, and on the right side, next to the PDB code, you have a link to 'Download Files'. You can then opt to download the PDB file or, the last entry, the biological assembly. You might have more than one, some attributed by the authors, some automatically by some software, so in case of doubt, download them all and compare.

As for the homology modelling. To keep things simple, you can just remove the other 3 chains from the PDB file. If your sequence is aligning only to one chain, just remove all others. Do however, make sure that whatever chain you keep, is the most complete of them all. Sometimes there are small differences between chains, such as a couple of aminoacids being resolved in one and not another.

I don't know about DS, but in PyMOL you can just type 'remove chain X' where X is the identifier of the chain. You can also just manually edit the PDB file (with any text editor, wordpad is good in windows) and delete the lines that do not match the chain of interest. The chain identifier is the 21st character of each 'ATOM' line. You will only want the 'ATOM' lines (unless you have some cofactors, then HETATM lines are also of interest) for the modelling.

Have fun modelling!

EDIT: You can use this website to edit your PDB files.