Search for: All records

Modeling unfolded states of proteins has implications for protein folding and stability. Since in unfolded state proteins adopt multiple conformations, any experimentally measured quantity is ensemble averaged, therefore the computed quantity should be ensemble averaged as well. Here, we investigate the possibility that one can model an unfolded state ensemble with the coil model approach, algorithm such as “flexible-meccano” [Ozenne V et al., Flexible-meccano: A tool for the generation of explicit ensemle descriptions of intrinsically disordered proteins and their associated experimental observables, Bioinformatics 28:1463–1470, 2012], developed to generate structures for intrinsically disordered proteins. We probe such a possibility by using generated structures to calculate pKas of titratable groups and compare with experimental data. It is demonstrated that even with a small number of representative structures of unfolded state, the average calculated pKas are in very good agreement with experimentally measured pKas. Also, predictions are made for titratable groups for which there is no experimental data available. This suggests that the coil model approach is suitable for generating 3D structures of unfolded state of proteins. To make the approach suitable for large-scale modeling, which requires limited number of structures, we ranked the structures according to their solvent accessible surface area (SASA). It is shown that in the majority of cases, the top structures with smallest SASA are enough to represent unfolded state. 

This study compares the role of electrostatics in the binding process between microtubules and two dynein microtubule-binding domains (MTBDs): cytoplasmic and axonemal. These two dyneins are distinctively different in terms of their functionalities: cytoplasmic dynein is processive, while axonemal dynein is involved in beating. In both cases, the binding requires frequent association/disassociation between the microtubule and MTBD, and involves highly negatively charged microtubules, including non-structured C-terminal domains (E-hooks), and an MTBD interface that is positively charged. This indicates that electrostatics play an important role in the association process. Here, we show that the cytoplasmic MTBD binds electrostatically tighter to microtubules than to the axonemal MTBD, but the axonemal MTBD experiences interactions with microtubule E-hooks at longer distances compared with the cytoplasmic MTBD. This allows the axonemal MTBD to be weakly bound to the microtubule, while at the same time acting onto the microtubule via the flexible E-hooks, even at MTBD–microtubule distances of 45 Å. In part, this is due to the charge distribution of MTBDs: in the cytoplasmic MTBD, the positive charges are concentrated at the binding interface with the microtubule, while in the axonemal MTBD, they are more distributed over the entire structure, allowing E-hooks to interact at longer distances. The dissimilarities of electrostatics in the cases of axonemal and cytoplasmic MTBDs were found not to result in a difference in conformational dynamics on MTBDs, while causing differences in the conformational states of E-hooks. The E-hooks’ conformations in the presence of the axonemal MTBD were less restricted than in the presence of the cytoplasmic MTBD. In parallel with the differences, the common effect was found that the structural fluctuations of MTBDs decrease as either the number of contacts with E-hooks increases or the distance to the microtubule decreases.