Technical Highlight - May 2014
Short description: A method for prediction of protein dynamics from primary sequence can map functional regions and help predict the effects of mutations.
Protein dynamics is closely related to function for structured domains and intrinsically disordered regions that fold upon binding cognate partners. NMR can provide residue-level information on dynamics, but the experiments require significant effort with data not usually deposited in the BioMagResBank (BMRB) database. In contrast, NMR chemical shift data are required for sequence-specific assignments during early stages of structure determination and interaction analysis. These data are readily accessible in the BMRB, providing a rich source of dynamical and structural information.
Vranken and colleagues have developed DynaMine, a method to estimate backbone flexibility for any protein, using chemical shift data for ∼2,000 proteins. Analysis of over 200,000 residue entries showed that amino acids can be quantitatively divided into the traditional ordered, neutral and disordered categories, based on distinct and statistically significant dynamical preferences. This analysis reveals that amino acids are dynamic across a continuous range rather than in a binary ordered/disordered distribution. Next, the data were used to train regression models based on sequence context, yielding the best results with a 51-residue window centered on a given amino acid. DynaMine compares well to existing disorder prediction software, especially for identification of short disordered fragments in the absence of prior structural data.
The DynaMine algorithm was applied to several paradigmatic test cases including p53 and adenovirus E1A. The p53 tumor suppressor contains a central DNA binding domain (DBD) and a C-terminal tetramerization domain flanked by disordered terminal regions that function as binding hubs for diverse proteins. DynaMine successfully predicted domain boundaries for the entire p53 sequence, in addition to the secondary structure of the DBD. DynaMine further predicted intermediate dynamics for the p53 transactivation region, reflecting the known propensity for sequences within this region to fold upon binding. The adenoviral E1A protein is a hub for host protein interactions and is mostly disordered in the free state. DynaMine identified a zinc finger and transcription-related region, in addition to several potentially preformed sequence elements that fold into stable structures in complexes.
In conclusion, DynaMine promises insights into the backbone dynamics of proteins without structural data and will aid in the assessment of mutational effects. It is available to try online at http://dynamine.ibsquare.be.
E. Cilia et al. From protein sequence to dynamics and disorder with DynaMine.
Nat. Commun. 4, 2741 (2014). doi:10.1038/ncomms3741