Accuracy Assessment for Equilibrium Dissociation Constant Using a Single Binding Isotherm

Abstract

The equilibrium dissociation constant (Kd) characterizes stability of non-covalent molecular complexes. Determining Kd for highly stable complexes may be extremely inaccurate if the ratio between the concentration of the limiting component (L0) and the a priori unknown value of Kd exceeds an unknown threshold value (aka threshold ratio). The only known approach to reveal this kind of inaccuracy in Kd requires building multiple experimental binding isotherms; it is resource intensive and, therefore, used very rarely. Here we introduce a single-isotherm approach for assessing Kd accuracy via determining the value of L0/Kd, estimating the threshold ratio, and comparing L0/Kd to the threshold ratio. In this proof-of-concept work, we present the theoretical basis and develop a step-by-step algorithm for our single-isotherm approach. We also demonstrate the experimental use of the developed algorithm.

FRET Assays for the Identification of C. albicans HSP90-Sba1 and Human HSP90α-p23 Binding Inhibitors

“…Large systematic errors in K d values can occur when the ratio between the concentration of the constant component ( L 0 ) and K d is high [ 27 ]. Such errors can be minimized by using L 0 (here either C. albicans HSP90-mNeonGreen or human HSP90α-mNeonGreen) at the level of the limit of quantitation (LOQ) [ 27 , 28 , 29 ]. Furthermore, the attainment of the equilibrium in the investigated binding reaction must be assessed [ 28 ].…”

“…Since K d is computed by nonlinear regression from a binding isotherm and depends on concentration inputs of both a limiting component ( L 0 ) and a varied component ( T 0 , here either C. albicans Sba1-mScarlet-I or human p23-mScarlet-I concentration), errors in both L 0 and T 0 can have a severe impact on the calculated K d values (see also Section 4.3.1 Equation (4)) [ 27 , 29 ]. These ramifications of systematic errors (arising from normally distributed random errors) in L 0 and T 0 should be reported to attain an informative range of K d .…”

“…These ramifications of systematic errors (arising from normally distributed random errors) in L 0 and T 0 should be reported to attain an informative range of K d . A detailed description of how the accuracy confidence interval (ACI) of the K d value can be determined can be found elsewhere [ 29 ]. Reports for ACI determination are included in the Supplementary Materials .…”

“…From these ten concentrations, the relative standard error was calculated and used as input for ACI generation. Detailed descriptions of ACI calculation are described elsewhere [ 29 ].…”

“…From these ten concentrations, the relative standard error was calculated and used as input for ACI generation. Detailed descriptions of ACI calculation are described elsewhere [29]. For C. albicans Sba1 competition experiments for HSP90-mNeonGreen-Sba1-mScarlet-I binding, 1 µM HSP90-mNeonGreen and 1 µM Sba1-mScarlet-I were incubated with varying concentrations of Sba1 in a range of 0-25,000 nM (lowest non-zero concentration 12 nM) in reaction buffer containing 5 mM ATP.…”