RDC's for intermolecular complexes

Dear all,
We are attempting to calculate the structure of an inter-molecular complex of two proteins (~20 kDa [X-ray] and 7 kDa [NMR] in MW) . The structure of the smaller protein has been determined when in complex with the larger molecule. I have defined the AIRs from chemical shift and mutagenesis data. We have also measured RDC restraints for the two proteins in the complex, but not in the isolated molecules. Thus I am unable to determine the tensor components of the individual proteins. What is the best way to use this data in a HADDOCK structure calculation protocol? May I input the data in the initial stages of calculation directly as RDC’s (SANI)? Any advice on this matter will be of help.


In order to use the RDCs to define the complex, you do need those in the context of the complex. So you are in the perfect situation. There should be a unique tensor for the two data sets (if the complex is rigid). You can try to determine your tensor parameters (anisotropy / rhombicity) from each set separately and compare them.

Also recommend here is to measure how well your proteins in isolation fit the RDC data. If not, you might consider refining them prior to docking. In case you are using a local version of HADDOCK, there is an example for a protein-protein complex based on RDCs and

For the use of RDCs in HADDOCK refer to the following paper:

A.D.J. van Dijk, D. Fushman and A.M.J.J. Bonvin Various strategies of using residual dipolar couplings in NMR-driven protein docking: Application to Lys48-linked di-ubiquitin and validation against 15N-relaxation data. Proteins: Struc. Funct. & Bioinformatics, 60 367-381(2005).

We also have published a protocol in Methods in Molecular Biology demonstrating the use of RDCs for docking in HADDOCK:

G.C.P. van Zundert and A.M.J.J. Bonvin. Modeling protein-protein complexes using the HADDOCK webserver. Methods in Molecular Biology: Protein Structure Prediction. Ed. Daisuke Kihara. Humana Press Inc., 163-179 (2014).

“In this chapter, we explain the use of the HADDOCK webserver based on the real-life Lys48-linked di-ubiquitin case, which led to the 2BGF PDB model. We demonstrate the use of chemical shift perturbation data in combination with residual dipolar couplings”

Unfortunately Springer seems to have lost the extra material required to run that protocol… (I contacted them about it). If you want it, email me directly for now (it should be hopefully be corrected on the Springer site at some point)