On the analysis of sedimentation velocity in the study of protein complexes

Abstract

Sedimentation velocity analytical ultracentrifugation has experienced a significant transformation, precipitated by the possibility of efficiently fitting Lamm equation solutions to the experimental data. The precision of this approach depends on the ability to account for the imperfections of the experiment, both regarding the sample and the instrument. In the present work, we explore in more detail the relationship between the sedimentation process, its detection, and the model used in the mathematical data analysis. We focus on configurations that produce steep and fast-moving sedimentation boundaries, such as frequently encountered when studying large multi-protein complexes. First, as a computational tool facilitating the analysis of heterogeneous samples, we introduce the strategy of partial boundary modeling. It can simplify the modeling by restricting the direct boundary analysis to species with sedimentation coefficients in a predefined range. Next, we examine factors related to the experimental detection, including the magnitude of optical aberrations generated by out-of-focus solution columns at high protein concentrations, the relationship between the experimentally recorded signature of the meniscus and the meniscus parameter in the data analysis, and the consequences of the limited radial and temporal resolution of the absorbance optical scanning system. Surprisingly, we find that large errors can be caused by the finite scanning speed of the commercial absorbance optics, exceeding the statistical errors in the measured sedimentation coefficients by more than an order of magnitude. We describe how these effects can be computationally accounted for in SEDFIT and SEDPHAT.

Human Herpesvirus 7 U21 Tetramerizes To Associate with Class I Major Histocompatibility Complex Molecules

“…SV data of U21 samples were modeled with diffusion-deconvoluted sedimentation coefficient distributions [c(s)] (19) in SEDFIT, using algebraic noise decomposition (20), and with the signal-average frictional ratio and the meniscus position refined by nonlinear regression. Accurate scan time intervals were derived from the file timestamps (21), and finite acceleration of the rotor was accounted for in the evaluation of Lamm equation solutions (22,23). Maximum entropy regularization was applied at a confidence level of P ϭ 0.68.…”

“…The global analysis of SE and SV experiments was conducted in the same model in SEDPHAT, joining the SE data with a single family of fringe profiles from an SV experiment conducted side by side with a sample from the same protein preparation. In order to exclude contaminating trace species at very low and very high sedimentation coefficients to bias the SV data set, a partial-boundary analysis was used (22) for the SV set in the global analysis, restricting the consideration to a radial range of boundary profiles corresponding to species between 3 S and 11 S. All SE and SV plots were created in GUSSI (http://biophysics.swmed.edu/MBR/software.html), kindly provided by C. Brautigam.…”

Protein Assembly in Serum and the Differences from Assembly in Buffer

“…The use of AUC approaches to characterize protein binding and assembly phenomena has had a long and successful history (Brown, Balbo, & Schuck, 2009;Cole, Lary, Moody, & Laue, 2008;Laue & Stafford, 1999). The most hydrodynamically and thermodynamically detailed analyses are obtained from both sedimentation equilibrium (SE) and sedimentation velocity (SV) approaches with dilute solutions and simple buffers.…”

Stoichiometry of Signalling Complexes in Immune Cells: Regulation by the Numbers

“…In sedimentation velocity experiments, the analytical programs are commonly used to interpret the data by computationally fitting it to www.intechopen.com the Lamm equation, an approach called discrete Lamm equation modelling (DLEM) (Brown et al, 2009;Schuck, 2009Schuck, , 2010a.…”