In HADDOCK, you can define Residual Dipolar Coupling restraints that will restrict the relative orientation of the molecules. You could generate synthetic ones for your system and use these for docking. Other advantage: full flexibility of both ligand and receptor.

For the use of RDCs in HADDOCK see:

https://www.researchgate.net/publication/7804238_Various_strategies_of_using_residual_dipolar_couplings_in_NMR-driven_protein_docking_application_to_Lys48-linked_di-ubiquitin_and_validation_against_15N-relaxation_data?ev=srch_pub

Article Various strategies of using residual dipolar couplings in NM...

Jeffrey,

have you considered HADDOCK. We are planning to use that for RNA-proteins but it also works for protein ligand docking.

(Haddock , Bonvin)

Cheers

We are trying to do something similar to the paper below. HADDOCK is the first thing we considered using intervector projection angle restraints between the fiber axis and each angle on the ligand.

In AutoDock one can fix the initial orientation of the ligand for each generation. I believe the translation along z can also be restricted by modifying the source files.

I am not sure as per your docking protocol ! But xplor-nih docking protocol can allow you with AIR restraints and dipolar couplings

( http://nmr.cit.nih.gov/xplor-nih/doc/current/python/examples.html )

All the best !

HADDOCK would be better option as it has many options for the both receptor as well as ligand molecules for precise constraints. extended docking steps might produce the poses of in required orientation.

As I'm sure you've considered, all of the better docking programs will enable you to 'suggest' and 'encourage' specific orientations, though generally not constrain. With Surflex, you can favor a given binding mode by prepositioning your reference ligand in the desired mode and using it to define your protomol (which you can further prune by hand if desired). In AutoDock and Vina, you can focus on a specific location and orientation by carefully specifying your grid. I believe that if you make a grid of the perfect size and orientation to accommodate your desired mode, it should give you good odds of finding it.

By the sounds of it, though, you are only looking at *one* ligand / receptor pair though? If so, perhaps it would be enlightening instead to request a large number of docking poses and just prune off all of those that do not meet your criteria.

It may also be a valuable insight if you could determine what aspect of the current receptor conformation is standing in the way of getting your desired binding mode in unconstrained simulations? Most receptor structures that we use for docking are not perfectly suited to the ligand binding event we're seeking to probe. One way to improve the ligand - receptor compatibility would be to run a constrained molecular dynamics simulation, starting with the ligand locked into the right pose (using a few judicious constraints) and permitting everything else to relax around it. If you do this at a low temperature, you may be able to mimic the crystallization conditions and thus derive a receptor model specially tailored for your ligand.

Dear Dr. Gerald, After doing the docking you are suggesting Dr. Jeffrey to use constrained molecular dynamics. Can you suggest me which software is best suited for the constrained MD simulations. As in case of GROMACS i have faced some troubles while running protein-ligand complexes. so, which tool would be best suited for CMD runs of Protein-ligand intercations?

Use Autodock....u can target the ligand binding residues very precisely

You can use Autodock in X, Y and Z axis by 8 equally boxes, 8 times. Then cluster your result and find the best conformer.

I guess the protocol would go something like this:

1. Dock the molecule without constraints in AutoDock or another program.

3. Rotate the fiber axis of the ligand bound structure to the z-axis

4. Calculate the dipolar coupling for each atom in which I have an RDC

Dear Jeffrey

1. Refine the structures with RDC before docking , and give orientation information ( SANI restraints and VEAN (predict R&D values by using PALES software SVDS matrix for best fit ) ) while docking if you are using HADDOCK .

2.After docking ,Results let you know the the structural quality factor Q and predicted RDC values and restrains violation list for protein complex for given set of constraints . ( if you refine the structures by RDC before docking , it gives the advantage for final results ... i mean it will help best fit for protein complex )

3. Final protein complex . you can do the the unrestrained molecular dynamics to validate the results .

Hex server can orientate the docking, angle axis, residues try it

Hex seems promising. Thanks

In HADDOCK, you can define Residual Dipolar Coupling restraints that will restrict the relative orientation of the molecules. You could generate synthetic ones for your system and use these for docking. Other advantage: full flexibility of both ligand and receptor.

For the use of RDCs in HADDOCK see:

https://www.researchgate.net/publication/7804238_Various_strategies_of_using_residual_dipolar_couplings_in_NMR-driven_protein_docking_application_to_Lys48-linked_di-ubiquitin_and_validation_against_15N-relaxation_data?ev=srch_pub

Article Various strategies of using residual dipolar couplings in NM...

Looking at the problem more carefully, Prof. Lushington's approach makes the most sense for this particular problem. RDCs are usually formed by transient contacts with the aligning medium. In this case the aligning media is the target, so RDCs will only be observed if the koff rate is large enough to get rapid exchange with the molecule in solution. If koff is too slow we will get no signal at all (for NMR experts this is a similar to the STD experiment).

The practical effect is that very tight binders will only be observed through a decrease in the signal and will not give RDCs. So the best docked conformations may not be observed at all in the experiment.

Well, if you generate synthetic RDCs matching the orientation you want, then koff issues are irrelevant...

They wouldn't be irrelevant because the tightest binders may not be observed experimentally if the koff is too low. In this case, the experimentally measured RDCs would reflect only a subset of the true orientations. Unconstrained docking as Prof. Lushington suggested followed by the generation of synthetic RDCs may be the best approach.

We are not speaking about experimental RDCs, but simply an artificial way of restraining orientations provided one knows what the relative orientations should be.

So no experiments required, purely simulated RDC restraints (which was the topic of the question if I got it correctly). So koff is irrelevant in this context.

Cheers

Alexandre

We originally developed TURNIP as a docking program that would do rational searches with constraints which became Daughter of TURNIP or DOT - http://www.ks.uiuc.edu/Development/biosoftdb/biosoft.cgi?&category=7

I think DOT may still do this for you - see http://www.sdsc.edu/CCMS/DOT/