Stoichiometry and physical chemistry of promiscuous aggregate-based inhibitors

Abstract

Many false positives in early drug discovery owe to nonspecific inhibition by colloid-like aggregates of organic molecules. Despite their prevalence, little is known about aggregate concentration, structure, or dynamic equilibrium; the binding mechanism, stoichiometry with, and affinity for enzymes remain uncertain. To investigate the elementary question of concentration, we counted aggregate particles using flow cytometry. For seven aggregate-forming molecules, aggregates were not observed until the concentration of monomer crossed a threshold, indicating a "critical aggregation concentration" (CAC). Above the CAC, aggregate count increased linearly with added organic material, while the particles dispersed when diluted below the CAC. The concentration of monomeric organic molecule is constant above the CAC, as is the size of the aggregate particles. For two compounds that form large aggregates, nicardipine and miconazole, we measured particle numbers directly by flow cytometry, determining that the aggregate concentration just above the CAC ranged from 5 to 30 fM. By correlating inhibition of an enzyme with aggregate count for these two drugs, we determined that the stoichiometry of binding is about 10,000 enzyme molecules per aggregate particle. Using measured volumes for nicardipine and miconazole aggregate particles (2.1 x 10(11) and 4.7 x 10(10) A(3), respectively), computed monomer volumes, and the observation that past the CAC all additional monomer forms aggregate particles, we find that aggregates are densely packed particles. Finally, given their size and enzyme stoichiometry, all sequestered enzyme can be comfortably accommodated on the surface of the aggregate.

Figure 1

Critical aggregation points of (A) nicardipine at 0.1% DMSO, (B) miconazole at 0.1% DMSO, and (C) L-755,507 at 1% DMSO in 1.5 μL of 50 mM potassium phosphate buffer, measured using flow cytometry. For (A) and (B), solutions were made by serial dilution, while compound was added directly for (C). Concentrations are represented as the mean and standard deviation of at least three replicates.

Figure 2

Correlation of β-lactamase inhibition to aggregate count for (A) nicardipine and (B) miconazole at 0.1% DMSO. The number of molecules of free enzyme is calculated as the product of percent activity and the total amount of enzyme in 1 mL. Aggregate count is measured by flow cytometry in a volume of 1.5 μL and extrapolated for 1 mL. Empty boxes represent aggregate counts that were extrapolated on the basis of a linear regression analysis of the measured aggregate count data points because the count was above the detection limit of the flow cytometer at higher concentrations of aggregating molecule. Free enzyme values are represented as the mean and standard deviation of three replicates. Aggregate counts are the mean of at least five measurements.

Figure 3

Concentration of monomer and aggregate fractions in (A) nicardipine at 0.1% DMSO, (B) K252c at 0.5% DMSO, and (C) TIPT at 5% DMSO in 50 mM KPi buffer. Aggregates were pulled down by centrifugation at 16000g for 1 h. The supernatant was removed, and the concentration of soluble monomer was determined by UV-visible spectrophotometry. The concentration of compound in the aggregate form was determined by resuspending the pellet in DMSO and performing spectrophotometric analysis. Bars represent the mean and standard deviation of three replicate measurements.

Figure 4

Size distribution histograms of (A) 34 μM nicardipine, (B) 44 μM nicardipine, (C) 4 μM miconazole, and (D) 9 μM miconazole determined by flow cytometry. Dynamic light scattering confirmed that there were no particles smaller than 100 nm in diameter, which is below the detection limit of the flow cytometer (inset of B and D). All samples are in filtered 50 mM KPi, 0.1% DMSO.

Figure 5

Comparison of calculated aggregate volumes to measured volumes for nicardipine and miconazole. Calculated volumes are the product of the molecular volume from Mitools and the predicted number of monomers that form each aggregate. The measured volume is the mean of the size distribution obtained using flow cytometry.

Figure 6

Model of aggregate structure and enzyme binding. Some organic molecules can form densely packed particles (10⁸ small molecules per aggregate for larger particles) in aqueous media. Once formed, these larger particles sequester and then inhibit enzyme with a stoichiometry of approximately 10⁴ enzyme molecules per aggregate. The surface of the aggregate is sufficient to accommodate all bound enzyme.