BPS2026 – Are short peptides or coil libraries of proteins suitable model systems for determining the coil state of intrinsically disordered proteins (IDPs)?

Contrary to the classical belief amino acid residues in unfolded and disordered peptides and proteins differ in terms of the structural preferences. They do not sample the entire sterically allowed Ramachandran space with similar probabilities. Hence, a description of the extended statistical coil states of unfolded and IDPs requires the knowledge of the Ramachandran distribution of their individual residues. Short peptides in water probed by spectroscopies (NMR and vibrational spectroscopy, DOIs: https://doi.org/10.1002/chem.201705353, https://doi.org/10.3390/biom12050684) and coil libraries constructed from not regularly structures segments of folded proteins were used to determine the respective structural propensities of amino acids. These investigations are complicated by the fact that contrary to Flory’s isolated pair hypothesis (IPH) for random coil polymers population distributions of residues depend on their nearest and even second nearest neighbors. Coil libraries seem to be an ideal tool to obtain the necessary information about neighbor dependent conformational propensities, if as assumed the used ensembles of not regularly structured proteins are representative for the coil state of proteins. The to be presented work compares experimentally determined Ramachandran plots of 24 cationic GXYG tetrapeptides (X and Y represent guest amino acids) with corresponding distributions in the coil library of Ting et al. (doi: https://doi.org/10.1371/journal.pcbi.1000763). Based on the use of several metrics (configurational entropy, Hellinger distances, and mesostate populations), the obtained results reveal that nearest neighbor interactions lead to different and in some cases opposite changes of propensities in short peptides and coil libraries, respectively. This observation suggests that the overall context (i.e., tertiär structures) of proteins used for the construction of coil libraries has a non-negligible influence on conformational distributions that do not average out even for a large number of investigated proteins.