| Random pool|| MD|| SAXS|| NMR|| Validated|
Tanja Mittag; Joseph A. Marsh; Alexander Grishaev; Stephen Orlicky; Hong Lin; Frank Sicheri; Mike Tyers; Julie D. Forman-KayPublication
: Structure/Function Implications in a Dynamic Complex of the Intrinsically Disordered Sic1 with the Cdc4 Subunit of an SCF Ubiquitin LigaseRelease date:
Intrinsically disordered proteins can form highly dynamic complexes with partner proteins. One such dynamic complex involves the intrinsically disordered Sic1 with its partner Cdc4 in regulation of yeast cell cycle progression. Phosphorylation of six N-terminal Sic1 sites leads to equilibrium engagement of each phosphorylation site with the primary binding pocket in Cdc4, the substrate recognition subunit of a ubiquitin ligase. ENSEMBLE calculations using experimental nuclear magnetic resonance and small-angle X-ray scattering data reveal significant transient structure in both phosphorylation states of the isolated ensembles (Sic1 and pSic1) that modulates their electrostatic potential, suggesting a structural basis for the proposed strong contribution of electrostatics to binding. A structural model of the dynamic pSic1-Cdc4 complex demonstrates the spatial arrangements in the ubiquitin ligase complex. These results provide a physical picture of a protein that is predominantly disordered in both its free and bound states, enabling aspects of its structure/function relationship to be elucidated.Calculation procedure:
Ensemble models of intrinsically disordered Sic1 and pSic1 were calculated using essentially the same approach as was described (Marsh and Forman-Kay, 2009). Distance restraints were calculated from PRE measurements. SAXS profiles of the experimentally restrained ensembles were calculated by predicting scattering curves for each individual member using the program CRYSOL (Svergun et al., 1995) and averaged over the members of the ensemble. Chemical shifts were calculated from individual conformers using SHIFTX (Neal et al., 2003). RDCs were calculated using a local alignment approach, in which local alignment tensors are calculated for 15 residue fragments of the sequence in a sliding window fashion (Marsh et al., 2008). 15N R2 relaxation rates were compared to the number of heavy atoms in an 8 A ° radius of each measured nucleus, as previously described (Marsh and Forman-Kay, 2009).
The Sic1 and pSic1 ensemble models comprised residues 1–90 of the full-length Sic1 amino acid sequence plus an N-terminal Gly-Ser sequence remaining after tag cleavage. Glutamate residues were used to represent the phosphorylated residues in pSic1 to facilitate use of TraDES (Feldman and Hogue, 2000). These glutamate residues were converted to the proper phosphorylated threonine or serine residues for electrostatic calculations. Three independent ensembles were calculated for each of free Sic1 and pSic1 and the pSic1 complex. Calculations were performed on a cluster of CPUs, with one main node performing the core conformational selection calculations and 8–12 nodes performing the iterative conformational sampling with CNS (Bru¨ nger et al., 1998), Unfoldtraj, and TraDES (Feldman and Hogue, 2000). The initial temperature for the ENSEMBLE calculations was set to 10,000 and decreased to 0.01 in 200,000 steps. The starting ensembles contained 200 structures and the number of conformers comprising the ensembles was decreased by one after each successful ENSEMBLE calculation in which full agreement with experimental restraints was achieved. Calculations were stopped when a smaller ensemble could not be successfully calculated within 72 hr.