Chains Significance In Proteins

Question

[jasdeep002]

In molecular description of proteins in PDB, they can represented by different Polymers consisting variable number of chains as Insulin (PDB ID: 1XDA) forming dimers of dimer. While analyzing 3D environment of an amino acid in a particular chain to mimic physiological behavior, what is significance of other chains? The chains form dimers/multimers, Is it then necessary to eliminate other chain entries form downloaded PDB files to study single amino acid environments?

Insulin Polymer 1 consists Chains A, C, E, G. To study immediate environment of single amino acid in Chain A, my PERL program also displays amino acids form other chains as well. As the chains are dimers of dimers, would it be feasible to take into account other chains? Do these chains exist at normal physiological conditions too?

Answer 1

[Woa]

Sometimes multimers appear as crystallographic artifacts. The biologically relevant multimer can be different than the crystallographic multimer. There are a lot of papers on this and computational resources as well. The best bet is to get the biological multimer and do all your computation on it.

[jasdeep002]

Thanks alot..!!!

Answer 2

[Jose Manuel Duarte]

Most of the structures in the PDB are solved with crystallography. That means that you get a crystal out of the experiment: i.e. an infinite lattice that contains many protein-protein contacts. There's a lot of symmetry in those crystals and one only needs to deposit the Asymmetric Unit (AU) in the PDB - the smallest unit you need to reconstruct the whole crystal. From the AU you can reconstruct the crystal by applying the symmetry operators of the space group.

The AU is what you find in the PDB file and it is not necessarily coinciding with the Biological Unit (BU). The BU can be:

1. coinciding with the AU
2. a superset of the AU: the biologically relevant contacts happen through crystallographic symmetry
3. a subset of the AU: there is more than 1 copy of the BU (this is known as Non-Crystallograpic Symmetry or NCS)

That means that if one wants to analyse crystal protein structures, in the general case the AU is meaningless. Instead one needs the BU which corresponds to how the protein assembles in solution. In order to get the BU from the PDB you need to download the "Biological Assembly" PDB file at the www.pdb.org web site.

One word of caution though. Experimentally, finding out what's the correct BU from the crystal is a difficult task, thus there are often errors in how the BU is annotated in the PDB. There are some cross-validated datasets that one can use.

See also this other question.