Track: Discovery and Basic Research - Bioanalytical - New Approaches
Category: Poster Abstract
Evaluation of Linked Protonation Effects on the Enthalpy of Phenobarbital Binding to Activated Carbon
Purpose: The thermodynamic study of molecular interactions, such as ligand binding to enzymes and proteins, surface sorption, and surface catalysis is of central importance. Occasionally, these interactions show some degree of pH dependence, reflecting that many binding or kinetic reactions proceed with the exchange of protons. In calorimetry, release and/or uptake of protons leads to interferences that mask the heat effect due to the reaction itself. This is problematic since calorimetry is a major tool in drug discovery studies as it may be used as a secondary screening technique to eliminate false positive hits after primary screening.
Activated carbons have widespread use in the treatment of the oral overdose of antiepileptic drugs such as phenobarbital. However, the interaction of phenobarbital with activated carbon depends heavily on the pH and the use of pH buffers. Thus, we aim to study phenobarbital-activated carbon binding as an example of proton exchange for weakly acidic/basic ligands. The theory of linked protonation and its contribution to the measured calorimetric enthalpy is described herein.
Theory: Ligands such as phenobarbital (pKa 7.6) bind predominantly in the neutral form (figure 1). The neutral ligand [LH] is in equilibrium with the deprotonated form [L-]. Upon binding, the concentration of the free [LH] decreases, perturbing the balance between [LH] and [L-]. This prompts a shift in the equilibrium favoring [LH] formation, and consumes [H+] ions in the process. In biological and chemical assays, buffers are used to keep the pH value constant by replenishing (or consuming) [H+]. Therefore, the calorimetrically measured heat (ΔHobserved) is the sum of the true enthalpy of binding (ΔHtrue), the enthalpy of ligand protonation (ΔHligand), and the enthalpy of buffer deprotonation ( ΔHbuffer), with the latter two terms being multiplied by the number of protons exchanged (n) between the buffer and the ligand (Equation 1).
The expression for the number of protons exchanged (n) between the buffer and the ligand is derived by applying the Henderson-Hasselbalch equation to the amount of free ligand before and after binding (equation 7). To the best of our knowledge, this is the first time that has been derived solely from constants such as pH and pKa, though other methods to measure n are used. Methods: An isoperibol calorimeter was used to measure the enthalpy of binding for phenobarbital on the activated carbon surface. One reactant (activated carbon) is loaded into an ampoule and another reactant (phenobarbital) is in solution in the reaction vessel. The reaction is initiated by breaking the ampule and releasing activated carbon to be mixed with the phenobarbital solution, and a heat signal is recorded.
Eight solutions of incremental amounts of phenobarbital are investigated, providing calorimetric data with low, intermediate, and high percent binding site occupancy. The enthalpy of binding is plotted against the amount adsorbed. Typically, a linear plot is observed if all binding sites are energetically equivalent. All phenobarbital solutions were made using 50mM buffer solutions. Blank reference data points were collected by mixing activated carbon and blank buffer solutions. Results: The enthalpy of phenobarbital binding (ΔHligand=24.6 kJ/mol) was evaluated in 5 buffers commonly used in molecular biology. Buffers with a wide range of ΔHbuffer were chosen: Phosphate (5.1 kJ/mol), HEPES (21.0 kJ/mol), MOPS (21.8 kJ/mol), imidazole (36.6 kJ/mol), and Tris (48.7 kJ/mol).
The enthalpy of binding (ΔHobserved ) changed dramatically depending on the buffer media chosen. In buffers with low protonation enthalpy, ΔHobserved was large and exothermic. However, as the ΔHbuffer becomes more endothermic, ΔHobserved turns less exothermic especially at high pH values where the number of protons exchanged (n) is at a maximum (figure 2). At peak difference between tris and phosphate buffer, ΔHobserved was 67% higher.
It's noteworthy that 'n' is independent of the buffer’s pKa. For example, MOPS and HEPES buffers have pKa values of 7.1 and 7.6, respectively. However, ΔHobserved was similar in both buffers due to their similar ΔHbuffer. On the other hand, imidazole and phosphate buffers have similar pKa values, yet the difference in ΔHobserved between them is significant due to the variance in their ΔHbuffer. Strong dependence of ΔHobserved on pH was also observed, reinforcing the proton exchange hypothesis. ΔHobserved was obtained at pH 6.5, 7.0, 7.5, 8.0, and 8.5 in Tris and phosphate buffers. If ΔHtrue is known, ΔHobserved can be predicted (and vice versa) at any pH and buffer using equations 1 and 7. Comparisons between predicted and observed enthalpies are shown in figure 3. Conclusion: The observed enthalpy of binding is susceptible to buffer interferences and exhibits a strong dependence on the pKa of the ligand, experimental pH, and protonation energy of the buffer. The number of protons exchanged per molecules bound ( n) can be calculated from the Henderson-Hasselbalch equation accurately given that buffer is fairly more concentrated than the ligand. Finally, regardless of the pH or the buffer used ΔHtrue can be predicted accurately from ΔHobserved.