A Catechol-Containing Screening Hit That Inhibits Everything
A high-throughput screen of 180,000 compounds against a serine protease returns compound 14 with IC50 = 1.2 uM in a fluorogenic peptide assay (10 nM enzyme, 20 uM substrate, Km = 25 uM). The Hill slope is 2.8. In the same assay containing 0.01% Triton X-100 the IC50 shifts to 45 uM. Pre-incubating enzyme with 20 uM compound for 30 min before substrate addition reduces activity by 60%, whereas adding substrate first reduces inhibition to 15%. Compound 14 also inhibits an unrelated caspase and a phosphatase with IC50 values of 3-8 uM. It is a catechol-containing flavonoid with aqueous solubility of 8 uM in assay buffer, and dynamic light scattering at 10 uM shows particles of 150-400 nm, while at 1 uM the scattering is indistinguishable from buffer. An SPR experiment with the compound immobilised on the chip reports binding to three unrelated proteins. The programme director wants to know whether this is a genuine series.
1. Interpret the data set. Which mechanisms explain the detergent shift, the steep Hill slope, and the pre-incubation dependence?
Read the worked answer
Detergent shift is the strongest single signal. A 37-fold IC50 shift on adding 0.01% Triton X-100 is the signature of colloidal aggregation: the detergent coats the particle surface and disperses it, so the effective concentration of the inhibitory species collapses. A genuine active-site inhibitor normally shifts by less than two- to threefold.
Hill slope of 2.8. Steep, non-Langmuirian behaviour is typical of aggregation, because inhibition tracks the number and size of particles rather than a one-to-one binding equilibrium. It can also arise from positive cooperativity, but that is not the simplest explanation here.
Pre-incubation dependence with substrate protection. This pattern is shown by aggregation (the particle forms and adsorbs to the enzyme surface over time) and also by covalent or slowly reversible inhibitors. The distinguishing experiment is a jump-dilution: dilute a pre-incubated enzyme-inhibitor mixture tenfold and measure recovery of activity. Aggregation recovers; covalent inactivation does not.
Solubility and DLS. A solubility of 8 uM with particles only at 10 uM and none at 1 uM places the critical aggregation concentration around a few micromolar, i.e. squarely in the assay range.
Catechol chemistry. A catechol can redox-cycle to generate hydrogen peroxide and can oxidise to a quinone that alkylates active-site cysteines, both of which give time-dependent, broad-spectrum inhibition. This mechanism can coexist with aggregation and must be ruled out separately.
2. Design the counterscreen cascade that would settle the mechanism. Explain what each experiment distinguishes and the order in which you would run them.
Read the worked answer
Step 1, cheapest and most decisive: repeat the IC50 with 0.01% Triton X-100 and with 0.1% CHAPS. If activity vanishes or the IC50 moves above 30 uM, treat the hit as an aggregator and stop investing in it.
Step 2, enzyme-concentration dependence. Measure IC50 at 10 nM and at 200 nM enzyme. For a reversible active-site inhibitor with Ki well above the enzyme concentration the IC50 is unchanged; for aggregation the IC50 rises roughly in proportion to enzyme concentration. This single experiment separates the two mechanisms cleanly.
Step 3, particle and solubility checks. DLS and nephelometry across 1-50 uM, plus centrifugation at 20,000 g for 10 min and retesting the supernatant. Loss of inhibition after centrifugation confirms the inhibitory species is particulate.
Step 4, redox counterscreens. Add catalase (to destroy peroxide) and add DTT or free cysteine in a separate assay. Reversal by catalase indicates peroxide generation; reversal by thiol indicates a quinone or Michael acceptor. Test a catechol-free analogue to see whether activity tracks the catechol.
Step 5, reversibility and orthogonality. Jump-dilution for reversibility, and confirm any residual activity with a different readout (absorbance or mass spectrometry) rather than the same fluorophore, because catechols quench fluorescence and can artefactually appear inhibitory.
Run steps 1 and 2 first: they cost one afternoon and eliminate most of the ambiguity.
3. What is your recommendation on compound 14 and on the series, and what would a compound have to show before you would call it a real inhibitor?
Read the worked answer
Recommendation: deprioritise compound 14 as a chemical probe and do not initiate medicinal chemistry on it. It fails the basic criteria for a genuine, tractable inhibitor: a nanomolar-to-low-micromolar effect that is abolished by detergent, a steep Hill slope, activity against three unrelated enzymes at similar concentrations, particles in the assay buffer at active concentrations, and artefactual SPR binding to unrelated immobilised proteins. The residual possibility of a real weak component is not worth the optimisation effort when the assay signal is contaminated by at least one and possibly two artefacts.
Criteria for a real inhibitor in this programme. IC50 stable within threefold on adding 0.01% Triton X-100; no dependence on enzyme concentration over a 20-fold range; Hill slope between 0.8 and 1.3; no inhibition of two unrelated enzymes above 30 uM; no particles by DLS at 10x IC50; reversible by jump-dilution; activity confirmed in a non-fluorescent orthogonal assay; and, ultimately, a measured Ki consistent with the Cheng-Prusoff relationship for the substrate concentration used.
Series triage. Screen the analogue set for these properties rather than for potency alone. If a catechol-free analogue retains sub-micromolar activity with a Hill slope near 1 and no detergent shift, that analogue, not compound 14, is the starting point.