The preclinical characterization of therapeutic proteins seeks to determine the stability, state of aggregation, and interaction of the therapeutic with other macromolecules in serum. Analytical ultracentrifugation (AUC) is one of the experimental methods best suited to determine and understand these factors. Sedimentation velocity, sedimentation equilibrium, synthetic boundary, and band sedimentation experiments are performed in order to determine the behavior of therapeutic antibodies in highly concentrated environments. The Aviv-FDS utilizes a single fluorescently labeled sedimenting species against a concentrated and heterogeneous background of serum proteins which results in nonideality. Nonideality results from excluded volume and charge effects. This describes the hydrodynamic and thermodynamic nonideality of a single component as a function of high concentrations of other components in solution. Thus, the aim of our current work is to study the behavior of a fluorescently labeled therapeutic antibody in the presence of high concentrations of human serum proteins (HSA and human IgG) and in human serum. To study and understand the behavior of therapeutic antibodies in the presence of HSA, human IgG, and other serum components, tracer experiments are done pairwise as a function of HSA, IgG and therapeutic protein concentration. This generates a 3x3 matrix of data that describes self- and cross-term hydrodynamic and thermodynamic nonideality (Ks and BM1). The goal is to develop a preclinical biophysical method for quantitative hydrodynamic and thermodynamic analysis of therapeutic proteins in crowded environments like serum. (Supported by Boehringer-Ingelheim.)

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 gamma-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 (k(s))This generates a 3x3 matrix of k(s) values that describe the nonideality of each pairwise interaction. The k(s), 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 k(s), nonideality matrix. The k(s) values for antibodies are smaller than expected and appear to'be masked by weak association. Global fitting to a k(s) and K-2 model significantly improves the fits.

The goal of this work is to develop a preclinical method for quantitative hydrodynamic and thermodynamic analysis of therapeutic proteins in crowded environments like human serum. The method utilizes tracer amounts of fluorescently labeled monoclonal antibodies and the Aviv AU-FDS optical system. We have performed sedimentation velocity experiments as a function of mAb, human serum albumin and human IgG concentration to extract self- and cross-term hydrodynamic nonideality effects. SV measurements are consistently complicated by weak mAb–mAb and mAb–IgG interactions (Wright et al. in Anal Biochem 550:72–83, 2018). In an attempt to explore different approaches we have investigated measurements of diffusion coefficients by traditional synthetic boundary experiments. Here we present a new technique incorporated into SEDANAL that can globally analyze the full time course of synthetic boundary experiments. This approach also utilizes F-mAb against a high concentration of unlabeled carrier protein (HSA or IgG). In principle both diffusion and sedimentation coefficient information can be extracted including hydrodynamic and thermodynamic nonideality. The method can be performed at a traditional low speed (5–7K rpm) or at high speeds. The high speed method can also be used to measure D and s for small molecules like fluorescein (often contaminants of F-HSA and F-mAb). The advantage of synthetic boundary over the standard sedimentation velocity method is that it allows for higher precision determination of diffusion coefficients. The concentration dependence of D can be corrected for hydrodynamic nonideality effects by plotting D * (1 + kijcj) vs total carrier concentration. The slope of the fitted data allows an alternate approach to determine self- and cross-term thermodynamic nonideality. This method can also explore cross-term diffusion coefficient effects. These results are compared to dynamic light scattering approaches which are limited to kD determinations for solutions of pure protein.

The preclinical characterization of therapeutic proteins seeks to determine the stability, state of aggregation, and interaction with other macromolecules in serum. Sedimentation velocity is one of the experimental methods required to determine and understand these factors. Our goal is to develop analysis methods for sedimentation velocity data acquired using absorbance, multi-wavelength and the Aviv-FDS AUC. The Aviv-FDS utilizes a single fluorescently labeled sedimenting species against a concentrated and heterogeneous background of serum proteins which results in the phenomenon known as the Johnston-Ogston (J-O) effect. The J-O effect describes the hydrodynamic and thermodynamic nonideal sedimentation of a single component as a function of high concentrations of other components in solution. Thus, the aim of our current work is to study the sedimentation of a fluorescently labeled therapeutic antibody in the presence of high concentrations of human serum proteins. The two most ubiquitous serum proteins are human serum albumin, HSA, (∼35-40 mg/ml) and γ-globulins, IgG, (∼10-15 mg/ml). To study and understand the behavior of therapeutic antibodies in the presence of HSA, human IgG, and other serum components, tracer experiments are done pairwise as a function of HSA, IgG and therapeutic protein concentration. A plot of 1/s vs concentration reveals thermodynamic nonideality or a slowing of the sedimentation rate due to itself or another component (s = s0/(1 + Ksc)). This generates a 3x3 matrix of data that describes self- and cross-term hydrodynamic nonideality Ks. The goal is to develop a preclinical method for quantitative hydrodynamic analysis of therapeutic proteins in crowded environments like serum. (Supported by Boehringer-Ingelheim.)

There is a long tradition in analytical ultracentrifugation (AUC) to perform simulation studies to investigate the utility of new data analysis methods or to verify the reliability of established approaches to complex systems. Here we compare the use of sedimentation velocity and sedimentation equilibrium methods for a heterogeneous associating system involving a trimeric cytokine complex (TNFα) and the stepwise binding of three soluble receptors (TNFR1). We will outline the methods for simulating velocity and equilibrium experimental data using the AUC analysis program SEDANAL corresponding to a titration of the ligand to a fixed concentration of trimer. We then outline the methods for analyzing the velocity and equilibrium data and compare the ability to extract the equilibrium constants for the system. The results suggest that complex hetero-associating systems can be successfully analyzed by sedimentation velocity approaches and that velocity analysis may be preferable to sedimentation equilibrium approaches for some complex hetero-reaction mechanisms. A detailed discussion of error analysis by Bootstrap with replacement, Monte Carlo and Fstat is presented in order to establish criteria for comparing and selecting velocity or equilibrium approaches. Experimental data are also compared to simulated data to present the additional challenge of experimental complexity, including sample purity, sample aggregation, and the presence of additional reaction modes. (Funded by UMC AUC Facility.)

The four-stranded i-motif (iM) conformation of cytosine-rich DNA has importance to a wide variety of biochemical systems that range from their use in nanomaterials to potential roles in oncogene regulation. The iM structure is formed at slightly acidic pH, where hemi-protonation of cytosine results in a stable C-C+ base pair. Fundamental studies to understand iM formation from C-rich strands of DNA are described. We present a systematic characterization of the consequences of epigenetic modifications, molecular crowding, degree of hydration, and DNA sequence on the stabilities of iM-forming sequences. We used a number of biophysical techniques to characterize both the folded iM and the folding kinetics of an iM. We established a mechanism for the folding. We observed that the C-C+ hydrogen bonding of certain bases initiates the folding of the iM structure. We also observed that substitutions in the loop regions of iMs give a distinctly different kinetic signature during folding as compared to those bases that are intercalated. Our data reveal that the iM passes through a distinct intermediate form between the unfolded and folded form. In the course of determining this folding pathway, we established that the fluorescent dC analogs tC° and PdC can be used to monitor individual residues of an iM structure and can be used to determine the pKa of an iM. Our results indicate that 5-hydroxymethylation of cytosine destabilized the iMs against thermal and pH-dependent melting, while 5-methylcytosine modification stabilized the iMs. Under molecular crowding conditions, the thermal stability of iMs increased and the pKa was raised to near 7.0. Taken together, our work has laid the foundation for examining folding and structural changes in more complex iMs.

Cytosine-rich nucleic acid sequences found in human DNA can adopt multiple intramolecular structures identified as i-motifs that are dependent on physiochemical solution conditions. The focus of this study is a four-stranded structure from the promoter of the human c-MYC gene. This compact, stable, and monomeric structure forms upon a decrease in pH (<5.0) causing hemi-protonation of a cytosine that results in a stable C-C+ hydrogen bond. Sedimentation velocity experiments were performed with the analytical ultracentrifuge in order to determine the hydrodynamic properties of the folded i-motif. The sedimentation velocity experiments were carried out in buffer conditions that differed in pH (4.5-8.0), salt type (NaCl and KCl), and salt concentration (up to 400 mM). Experiments were run at different pH values in order to observe the linkage between the pKa of the Hoogsteen base pair and i-motif folding. High salt concentration was used to avoid non-ideality (primary charge effect) observed when nucleic acids sediment at low salt concentration. The data indicates that the S20,w value increases when the primary charge effect is overcome at a higher salt concentration. The experimental S20,w values are compared with values obtained by bead model simulations using SOMO as implemented in Ultrascan 3. (Supported by UMC AUC Facility.)

The four-stranded i-motif (iM) conformation of cytosine-rich DNA has importance to a wide variety of biochemical systems that range from their use in nanomaterials to potential roles in oncogene regulation. The iM structure is formed at slightly acidic pH, where hemiprotonation of cytosine results in a stable C-C+ basepair. Here, we performed fundamental studies to examine iM formation from a C-rich strand from the promoter of the human c-MYC gene. We used a number of biophysical techniques to characterize both the hydrodynamic properties and folding kinetics of a folded iM. Our hydrodynamic studies using fluorescence anisotropy decay and analytical ultracentrifugation show that the iM structure has a compact size in solution and displays the rigidity of a double strand. By studying the rates of circular dichroism spectral changes and quenching of fluorescent cytidine analogs, we also established a mechanism for the folding of a random coil oligo into the iM. In the course of determining this folding pathway, we established that the fluorescent dC analogs tC° and PdC can be used to monitor individual residues of an iM structure and to determine the pKa of an iM. We established that the C-C+ hydrogen bonding of certain bases initiates the folding of the iM structure. We also showed that substitutions in the loop regions of iMs give a distinctly different kinetic signature during folding compared with bases that are intercalated. Our data reveal that the iM passes through a distinct intermediate form between the unfolded and folded forms. Taken together, our results lay the foundation for using fluorescent dC analogs to follow structural changes during iM formation. Our technique may also be useful for examining folding and structural changes in more complex iMs.