Date Published: June 6, 2019
Publisher: Public Library of Science
Author(s): Ryo Hayama, Mirco Sorci, John J. Keating IV, Lee M. Hecht, Joel L. Plawsky, Georges Belfort, Brian T. Chait, Michael P. Rout, Bostjan Kobe.
Protein-protein interactions are central to biological processes. In vitro methods to examine protein-protein interactions are generally categorized into two classes: in-solution and surface-based methods. Here, using the multivalent interactions between nucleocytoplasmic transport factors and intrinsically disordered FG repeat containing nuclear pore complex proteins as a model system, we examined the utility of three surface-based methods: atomic force microscopy, quartz crystal microbalance with dissipation, and surface plasmon resonance. Although results were comparable to those of previous reports, the apparent effect of mass transport limitations was demonstrated. Additional experiments with a loss-of-interaction FG repeat mutant variant demonstrated that the binding events that take place on surfaces can be unexpectedly complex, suggesting particular care must be exercised in interpretation of such data.
Protein-protein interactions are at the core of any biological system and regulate essential cellular functions; measuring their characteristics, such as stoichiometry, affinity and kinetics, is crucial for understanding their biological roles. There are multiple in vitro methods to characterize protein-protein interactions, among the most popular of which are surface-based such as enzyme-linked immunosorbent assay (ELISA) and surface plasmon resonance (SPR). These surface-based methods have been applied to a wide range of protein-protein interactions, from well-defined antigen-antibody interactions to those involving intrinsically disordered proteins (IDPs), a major class of proteins involved in various functions, many of whose detailed behaviors are still being characterized [1, 2]. Here, we examined the applicability of select surface-based techniques to the characterization of a complex system involving IDPs, specifically the ones involved in the nucleocytoplasmic transport mediated by nuclear pore complexes (NPCs) [3–6].
FG Nup constructs and their interactions with Kap95 were systematically characterized by AFM, QCM-D, and SPR. Nsp1FG grafting conditions were optimized and cross-validated by AFM and QCM-D, which yielded a consistent picture on the viscoelastic behavior of Nsp1FG. Both QCM-D and SPR demonstrated the apparent specificity of the FG-grafted surfaces to the TF, Kap95. Therefore, robust experimental systems were set up to reproducibly collect data on TF binding to FG-grafted surfaces. The results from AFM, QCM-D, and SPR were in agreement with each other and with previous observations in the literature. However, our further investigations demonstrated the existence of significant mass transport limitation and extra binding mode(s). The existence of mass transport limitations has been acknowledged in the nuclear transport field [13, 26, 27]. Indeed, the present work demonstrates the importance of taking mass transport limitations properly into account, but also the importance of considering unexplained physical phenomena that are implicit in the binding data. One could attempt to infer microscopic mechanisms that produced the binding curves, however, such deconvolution processes require assumptions about those mechanisms (e.g. what processes exist and how they depend on each other, along with all their parameters). While there have been attempts to model the generative process of the surface binding data, we and others found that the underlying dynamics of the “fuzzy” FG-TF interactions to be quite complex [17–19]. Since it is difficult to validate the assumptions we would have to make to model the surface binding data, we opted out from making such inferences in this work.