Significant progress during the reporting period was made in the use of the newly purchased fluorescence detector for the analytical ultracentrifuge. Fluorescence-detected analytical ultracentrifugation (FD-AUC) is a relatively new technique that provides sub-nanomolar detection sensitivity. We have carried out comprehensive series of experiments aimed at exploring the properties of this detector. We found that modifications of the standard sedimentation analysis models are necessary. Motivated by the optical design of the detector, we developed models that can account for spatial gradients of signal magnification, temporal drifts of laser intensity, photo-bleaching of the fluorophore, and the obscurement of the detection cone close to the end of the solution column. After accounting for these effects, the resulting data quality was found to be comparable or exceeding that of the best conventional detection, allowing for the first time a consistently reliable and detailed interpretation of the FD-AUC data. These tools were implemented and disseminated in our analysis software SEDFIT. Next, we explored the effect of different labelling and experimental design strategies for FD-AUC in the study of high-affinity protein self-association and hetero-association. Using GluA2 ATD as a model system, we found the fluorescent label, FAM could create multiple classes of molecules with different binding properties, whereas the more polar Dylight label appears to leave the binding properties of GluA2 unaffected, indistinguishable from unlabelled molecules as well as GFP fusion proteins. The broad isotherm caused by the polydispersity of FAM-GluA2 explains previously observed discrepancies of FAM-labeled GluA2 ATD apparent hydrodynamic behaviour with the hydrodynamic expectation. This removes previously noted concerns about the reliability of sedimentation coefficients measured in FD-AUC. Finally, we developed an approach to improve the detection limits of FD-AUC by 1 to 2 orders of magnitude, which now brings interacting systems with picomolar binding constants into the dynamic range of AUC. As an example, we revisited an GFP antibody previously reported in the literature to be of too high affinity for FD-AUC, and to show (unexpectedly) only mono-valent binding. With our new tools, we were able to determine a 10 pM binding constant with bi-valent binding, consistent with the antibody structure. Thus, taken together, these results set FD-AUC on a solid foundation for the quantitative study of high-affinity assembly reactions, extending the corresponding capabilities of conventional AUC by three orders of magnitude. In a different line of investigation we tested existing and developed new calibration procedures for analytical ultracentrifugation (conventional and FD-AUC). The need for this arose from the observation of inconsistent sedimentation coefficients measured in different instruments. We discovered that the current manufacturer data acquisition software misreported experimental times by 10%. We wrote and disseminated software to restore data integrity. Further, we discovered similarly serious deficiencies in the temperature and radial calibration. We developed a new device for temperature measurement in the spinning rotor, and for the accurate determination of the radial magnification of the imaging optics. By combination of these three external calibrations, we were able to restore accuracy of the sedimentation parameters of a test protein across eleven different instruments. Our intention is to expand this study to many laboratories, in order to assure a uniform quality standard for hydrodynamic measurements.
Showing the most recent 10 out of 40 publications
