Maximizing and Assessing the Accuracy of Equilibrium Dissociation Constants for Affinity Complexes

Abstract

Affinity interactions are fundamental to both biophysical technologies and biological systems, with the equilibrium dissociation constant (Kd) serving as a critical parameter that quantifies binding strength. Accurately determining Kd is essential for applications ranging from drug discovery and diagnostics to mechanistic studies in molecular biology. However, despite its importance, Kd values determined by different biophysical methods often vary significantly, highlighting unresolved measurement inaccuracies. Before this work, no comprehensive framework has been established to systematically identify the root causes of these inaccuracies or to provide accessible tools for reliably evaluating Kd accuracy.

This dissertation addresses these challenges through both experimental and theoretical advancements. On the experimental side, I confirmed the robustness and ruggedness of the Accurate Constant via Transient Incomplete Separation (ACTIS) method. Initially applicable only to protein–small molecule systems due to its reliance on diffusivity differences, ACTIS was extended to study protein–DNA complexes by optimizing instrumentation and experimental protocols. On the theoretical side, I performed a systematic analysis to identify the key determinants of Kd accuracy that are independent of measurement method. The study revealed that minimizing the concentration of the limiting component and reducing systematic errors in reagent concentrations and signal measurements are essential for achieving accurate Kd values. Recognizing practical constraints such as the limit of quantitation (LOQ), I investigated common sources of systematic error and proposed mitigation strategies.

To provide researchers with a practical tool for assessing Kd accuracy, I developed a computationally efficient algorithm to estimate the Accuracy Confidence Interval (ACI) for Kd from a single binding isotherm, which serves as the foundation for the web-based tool ACI-Kd. Given the mathematical similarity between Kd and the Michaelis constant (Km), the framework was further extended to enzyme kinetics, resulting in ACI-Km.

Additionally, I developed ACI-ITC, a Monte Carlo-based tool that evaluates the accuracy of Kd, binding enthalpy (ΔH°), and stoichiometry (n) from isothermal titration calorimetry data. All tools are accessible through a unified web platform: https://aci.sci.yorku.ca. Collectively, this research provides rigorous, accessible methodologies for enhancing the reliability of molecular interaction measurements across a wide range of experimental systems.