An efficient high-throughput resonance assignment procedure for structural genomics and protein folding research by NMR

Sequence specific resonance assignment is the primary requirement for all investigations of proteins by NMR methods. In the present postgenomic era where structural genomics and protein folding have occupied the center stage of NMR research, there is a high demand on the speed of resonance assignment, whereas the presently available methods based either on NOESY or on some triple-resonance experiments are rather slow. They also have limited success with unfolded proteins because of the lack of NOEs, and poor dispersion of amide and carbon chemical shifts. This paper describes an efficient approach to rapid resonance assignment that is suitable for both folded and unfolded proteins, making use of the triple-resonance experiments described recently [HNN and HN(C)N]. It has three underlying principles. First, the experiments exploit the 15N chemical shift dispersions which are generally very good for both folded and unfolded proteins, along two of the three dimensions; second, they directly display sequential amide and 15N correlations along the polypeptide chain, and third, the sign patterns of the diagonal and the sequential peaks originating from any residue are dependent on the nature of the adjacent residues, especially the glycines and the prolines. These lead to so-called "triplet fixed points" which serve as starting points and/or check points during the course of sequential walks, and explicit side chains assignment becomes less crucial for unambiguous backbone assignment. These features significantly enhance the speed of data analysis, reduce the amount of experimentation required, and thus result in a substantially faster and unambiguous assignment. Following the amide and 15N assignments, the other proton and carbon assignments can be obtained in a straightforward manner, from the well-established three-dimensional triple-resonance experiments. We have successfully tested the new approach with different proteins in the molecular mass range of 10-22 kDa, and for illustration, we present here the backbone results on the HIV-1 protease-tethered dimer (molecular mass 22 kDa), both in the folded and in the unfolded forms, the two ends of the folding funnel. We believe that the new assignment approach will be of great value for both structural genomics and protein folding research by NMR