Dynamics of Protein Assemblies by Analytical Ultracentrifugation

National Institute Of Biomedical Imaging And Bioengineering, Bethesda

Investigators

Linked publications, trials & patents

Abstract

In this project we seek to develop an experimental window for the observation of polymorphic dynamic multi-protein complexes that elude classical structural techniques, with the goal to elucidate coarse-grained architectural principles and energetic driving forces in solution. Recently, one major thrust of our work has been the continued improvement of nonideal sedimentation velocity (SV) as a technique for measuring macromolecular size-distributions in highly concentrated solutions. The goal is to study proteins at concentrations closer to the intracellular environment, where weak interactions can govern a wide spectrum of behavior, including dynamic multi-protein complex formation and liquid-liquid phase transition. Our recent breakthrough in the analysis of polydisperse concentrated protein solutions came from the introduction of a mean-field approach to account for hydrodynamic interactions in the sedimenting particle mixtures. This has allowed us to increase the concentration limits in SV for high-resolution size-distribution analyses by more than tenfold, approaching the concentration of total protein in serum. Since then we have improved this approach further by extending the hydrodynamic interaction models to higher-order accuracy using recent theoretical results in statistical fluid mechanics. However, a remaining experimental difficulty when working at high concentration is presented by optical aberrations in the strong refractive index gradients of the sedimentation boundary. While this lensing effect is historically well documented (Wiener skewing) and has been theoretically studied for some model geometries, it is unclear quantitatively to what extent it impacts measurements using the current optical detection systems. Nonetheless, lensing effects can be minimized by short optical pathlengths, and therefore we have previously introduced novel 3D printed sample holders. To experimentally assess the remaining distortions we have now incorporated a pattern of fiduciary markers into the windows of the sample holders, which superimpose a set of characteristic signals onto the low-spatial-frequency sedimentation profiles. In the reporting year we have implemented new mathematical data analysis procedures to localize these fiduciary signals in experimental data, remove them from sedimentation signals, and to evaluate their displacement from lensing artifacts. In future work this should allow us to quantitatively model the actual optical aberrations. We expect that this can raise the experimental concentration limits further. A second major effort in our work on SV is directed at making sedimentation analysis more information-rich for the study of multi-protein interactions. We have previously pursued this goal by multi-signal approaches and creating temporal probes by photoswitching. In addition, an opportunity resides in the stratification of solution during the sedimentation process, since sedimenting systems of dynamically interacting proteins exhibit richly patterned sedimentation profiles, with sedimentation boundaries that depend in complex ways on protein concentrations, equilibrium constants, and sedimentation coefficients associated with the different assembly states. These patterns have remained largely unexploited for data analysis of interacting systems. We have previously developed effective particle theory that provides a physical explanation for sedimentation boundary patterns, and integrated this theory into a new approach for interpreting SV data from sedimenting interacting systems. It allows taking advantage of the entire sedimentation pattern, rather than only the average sedimentation velocity, and thereby enhances the information content of SV experiments. In the reporting period, we have further improved our data analysis software in expanding the user interface that allows creation of customized binding models with embedded boundary pattern analysis, to facilitate the study of multi-protein interactions. To disseminate knowledge of analytical ultracentrifugation we have developed step-by-step protocols both for the calibration of the instrument, as well as the study of protein self-association. Furthermore, we have engaged in collaborative model applications of SV to challenging protein assemblies, including extended multi-step processes. Finally, we have reviewed how these techniques for protein assemblies can be applied to protein/nanoparticle assemblies.