What is the method to predict connected protein structure from two separate known protein structures

Question

[moing.gude]

Hello,

My question is how to predict one connected protein structures from two separate known structures?

It should be different from the protein-protein docking method, which is non-covalent interactions. My question is that the two proteins should "contently" link together.

Is that possible first manually link them (two 3D structure) together using some software, and then do the energy minimization?

Any suggestions are appreciated!

Answer 1

[Mensur Dlakic]

What you want to do is possible, but can't be done reliably without external restraints. Protein modeling typically means satisfying a number of angle and distance restraints. If you add a linker connecting two domains of known structures, there will be no restraints specifying their relative orientation unless you add them manually. Energy minimization will help you only if you are already close to the optimal domain interaction, but it isn't meant for making proteins move around a lot.

There are many recent methods predicting intra-protein contacts between protein residues based on co-evolution. Those approaches can be extended to inter-protein contacts as well, and may provide some restraints for your modeling procedure.

• https://academic.oup.com/nar/article/46/W1/W432/5001161
• https://link.springer.com/protocol/10.1007%2F978-1-4939-9873-9_6

[moing.gude]

Thanks, @Mensur. Your suggestions are very helpful and I will look at intra-protein contacts prediction methods.

Answer 2

[jgreener]

MODELLER can do this for you and is well-documented, see for example this section of the docs.

This is a common step in protein structure prediction pipelines and generally works well for native structures or high-quality models, particularly in the context of joining two protein domains. In this case there will be few inter-domain contacts and the relative positions of the domains are already constrained by the length of the linker region.

Extensive energy minimisation - i.e. beyond bond/angle/steric clash fixing and into conformational searching - will be unlikely to do much good since the timescales and energy barriers for relative domain motions are large.

[moing.gude]

@jgreener thank you.