Peter J. Sherwood

Simultaneous analysis of hydrodynamic and optical properties using analytical ultracentrifugation equipped with multiwavelength detection.

Analytical ultracentrifugation (AUC) has proven to be a powerful tool for the study of particle size distributions, particle shapes, and interactions with high accuracy and unrevealed resolution. In this work we show how the analysis of sedimentation velocity data from the AUC equipped with a multiwavelength detector (MWL) can be used to gain an even deeper understanding of colloidal and macromolecular mixtures. New data evaluation routines have been integrated in the software SEDANAL to allow for the handling of MWL data. This opens up a variety of new possibilities because spectroscopic information becomes available for individual components in mixtures at the same time using MWL-AUC. For systems of known optical properties information on the hydrodynamic properties of the individual components in a mixture becomes accessible. For the first time, the determination of individual extinction spectra of components in mixtures is demonstrated via MWL evaluation of sedimentation velocity data. In our paper we first provide the informational background for the data analysis and expose the accessible parameters of our methodology. We further demonstrate the data evaluation by means of simulated data. Finally, we give two examples which are highly relevant in the field of nanotechnology using colored silica and gold nanoparticles of different size and extinction properties.

Effect of Kinetics on Sedimentation Velocity Profiles and the Role of Intermediates

We have previously presented a tutorial on direct boundary fitting of sedimentation velocity data for kinetically mediated monomer-dimer systems [Correia and Stafford, 2009]. We emphasized the ability of Sedanal to fit for the k(off) values and measure their uncertainty at the 95% confidence interval. We concluded for a monomer-dimer system the range of well-determined k(off) values is limited to 0.005-10(-5) s(-1) corresponding to relaxation times of approximately 70 to approximately 33,000 s. More complicated reaction schemes introduce the potential complexity of low concentrations of an intermediate that may also influence the kinetic behavior during sedimentation. This can be seen in a cooperative ABCD system (A+B --> C; B+C --> D) where C, the 1:1 complex, is sparsely populated (K(1)=10(4) M(-1), K(2)=10(8) M(-1)). Under these conditions a k(1,off)<0.01 s(-1) produces slow kinetic features. The low concentration of species C contributes to this effect while still allowing the accurate estimation of k(1,off) (although k(2,off) can readily compensate and contribute to the kinetics). More complex reactions involving concerted assembly or cooperative ring formation with low concentrations of intermediate species also display kinetic effects due to a slow flux of material through the sparsely populated intermediate states. This produces a kinetically limited reaction boundary that produces partial resolution of individual species during sedimentation. Cooperativity of ring formation drives the reaction and thus separation of these two effects, kinetics and energetics, can be challenging. This situation is experimentally exhibited by systems that form large oligomers or rings and may especially contribute to formation of micelles and various protein aggregation diseases including formation of beta-amyloid and tau aggregates. Simulations, quantitative parameter estimation by direct boundary fitting and diagnostic features for these systems are presented with an emphasis on the features available in Sedanal to simulate and analyze kinetically mediated systems.

Characterization of therapeutic antibodies in the presence of human serum proteins by AU-FDS analytical ultracentrifugation

The preclinical characterization of biopharmaceuticals seeks to determine the stability, state of aggregation, and interaction of the antibody/drug with other macromolecules in serum. Analytical ultracentrifugation is the best experimental method to understand these factors. Sedimentation velocity experiments using the AU-FDS system were performed in order to quantitatively characterize the nonideality of fluorescently labeled therapeutic antibodies in high concentrations of human serum proteins. The two most ubiquitous serum proteins are human serum albumin, HSA, and γ-globulins, predominantly IgG. Tracer experiments were done pairwise as a function of HSA, IgG, and therapeutic antibody concentration. The sedimentation coefficient for each fluorescently labeled component as a function of the concentration of the unlabeled component yields the hydrodynamic nonideality (ks). This generates a 3x3 matrix of ks values that describe the nonideality of each pairwise interaction. The ks matrix is validated by fitting both 2:1 mixtures of HSA (1-40 mg/ml) and IgG (0.5-20 mg/ml) as serum mimics, and human serum dilutions (10-100%). The data are well described by SEDANAL global fitting with the ks nonideality matrix. The ks values for antibodies are smaller than expected and appear to be masked by weak association. Global fitting to a ks and K2 model significantly improves the fits.

SEDANAL: Model-Dependent and Model-Independent Analysis of Sedimentation Data

SEDANAL (Stafford and Sherwood, Analysis of heterologous interacting systems by sedimentation velocity: curve fitting algorithms for estimation of sedimentation coefficients, equilibrium and kinetic constants. Biophys Chem 108:231–243, 2004) is a suite of routines that are used to analyze data from the analytical ultracentrifuge and other types of centrifugal fluid analyzers. It can handle data from both sedimentation velocity and sedimentation equilibrium experiments. Two general approaches are used: (1) model independent and (2) model dependent. The model-independent modules are based on the time-derivative method for sedimentation velocity and BioSpin for sedimentation equilibrium data. The model-dependent modules use several curve fitting techniques to fit user-specified models both to the sedimentation velocity and sedimentation equilibrium data. Sedanal allows the global analysis of data from multiple runs and multiple optical systems and of absorbance data from multiwavelength instruments. Models are specified in the Model Editor module of Sedanal. This chapter describes the various modules and routines of Sedanal. The Model Editor especially is described in detail.

Techniques for Dissecting the Johnston-Ogston Effect

The development of the fluorescence detection system (Aviv-FDS) for the AUC allows a single fluorescently labeled species to be quantitatively characterized against a highly concentrated and heterogeneous background. During our use of the FDS to characterize ELP, a novel drug delivery vector (see Lyons et al., Biophys J 104:2009–2021, 2013), in serum, we encountered the Johnston-Ogston (J-O) effect. The J-O effect is a classical anomaly in sedimentation velocity theory and practice describing the nonideal sedimentation properties of a component as a function of high concentrations of other components. We examined the J-O effect using recent advances in AUC hardware, the AU-FDS (AVIV Biomedical), and data analysis methods, DCDT+ and SEDANAL global direct boundary fitting. We empirically quantified the self- and cross-sedimentation nonideality properties of ELP and the two most ubiquitous serum proteins, albumin (∼35–40 mg/ml) and γ-globulins (∼10–15 mg/ml). We have verified and measured the presence of cross-term hydrodynamic nonideality by running SV studies on a fluorescently labeled component (∼100 nM) in a titration experiment with high concentrations of unlabeled components. This has been accounted for through the introduction of a 3 × 3 nonideality matrix of Ks values into SEDANAL. ELP experiments with mixtures of albumin and γ-globulins were also performed in an attempt to recapitulate the J-O behavior of a serum solution. Clearly, other components or effects contribute to the serum J-O effect. Additional experiments with lipids, lipidated serum albumin, and PEG solutions are planned. These studies lay the groundwork for bringing quantitative hydrodynamic analyses into crowded environments and will allow measurement of hydrodynamic and equilibrium macromolecular properties in a physiological state.

SEDANAL: Global Analysis of General Hetero- and Self-Associating Systems by Sedimentation Equilibrium

Algorithms have been developed for the analysis of sedimentation equilibrium data by fitting to arbitrary reaction schemes. These have been implemented within the framework of a larger program called Sedanal, which until inclusion of equations for fitting equilibrium data was capable of treating only sedimentation velocity data (Stafford and Sherwood, Biophys Chem 108:231–243, 2004) A predecessor to this program, called NONSIM (first used by Margossian and Stafford, Biochemistry, 1982), forms the basis for the algorithms used in Sedanal. Fitting to the equilibrium equations is carried out by minimization with respect to either the L1 norm (average absolute value of the residuals) using the simplex method of Nelder and Mead (1965) or the L2 norm using the Levenberg-Marquardt method. In cases involving more than one macromolecular component (i.e., hetero-associations or self-associating systems involving nonparticipating species), conservation of mass is invoked, and weight fractions of the components become fitting parameters. Thus, it is possible to fit directly for the weight fraction of incompetent species in a self-associating system without resorting to further mathematical treatment of apparent equilibrium constants. Global fitting of data spanning multiple speeds and loading concentrations (and multiple optical systems) allows the determination of both equilibrium constants for the interacting species and the weight fractions of the several components. Because the weight fraction of components must remain constant upon dilution while the distribution of individual species will vary characteristically with local concentration according to the law of mass action, these types of mixed systems can be resolved as long as data from a sufficiently wide range of loading concentrations and speeds can be combined in a global fit. A large array of arbitrary reaction schemes can be represented using the Model Editor, which is part of Sedanal.