A curved double-reciprocal plot in a recombinant esterase: substrate inhibition, cooperativity, or cosolvent artefact?
You are characterising a recombinant bacterial esterase for a directed-evolution campaign. Assays run at 25 °C in 25 mM Tris-HCl pH 8.0, 150 mM NaCl, 1 mM DTT, using p-nitrophenyl acetate and following A405 for 60 s. Rates are linear with time up to 2 mM substrate but bend downward above 5 mM. The substrate is dissolved as a 200 mM stock in DMSO, so final cosolvent rises from 0.025% at 0.05 mM substrate to 10% at 20 mM. Specific activities (µmol min⁻¹ mg⁻¹): 0.05 mM, 4.1; 0.2 mM, 11.8; 0.5 mM, 18.9; 1.0 mM, 24.2; 2.0 mM, 27.6; 5.0 mM, 25.0; 10 mM, 16.4; 20 mM, 8.7. A global double-reciprocal fit gives Km = 0.35 mM and Vmax = 33, but the plot is visibly curved and r² = 0.87. One reviewer calls the curvature substrate inhibition, a colleague calls it negative cooperativity, and a third suggests enzyme aggregation at high protein concentration.
1. Which explanation do these data best support, and what is the minimum set of experiments needed to discriminate between the three hypotheses?
Read the worked answer
Start by noting that the descending limb (>5 mM) coincides almost exactly with the cosolvent ramp: 5 mM is 2.5% DMSO, 20 mM is 10%. Esterases are commonly inhibited above 2–5% DMSO, so an artefact is at least as plausible as any kinetic model.
A useful internal check is whether a single inhibition constant fits. Using v = Vmax[S]/(Km + [S] + [S]²/Kis) with Vmax = 33 and Km = 0.35:
- at 10 mM: 33×10/16.4 = 20.1, so 100/Kis = 9.75 → Kis ≈ 10.3 mM
- at 5 mM: 33×5/25.0 = 6.6, so 25/Kis = 1.25 → Kis ≈ 20 mM
A two-fold drift in Kis means the simple substrate-inhibition model does not describe the data.
Discriminating experiments, in order of cost:
- Re-run the entire curve at a constant 2% DMSO (adjust the stock concentration). If the descending limb vanishes, it was cosolvent.
- Assay at three enzyme concentrations (0.05, 0.2, 0.5 µg/mL). Constant specific activity argues against aggregation.
- Repeat with p-nitrophenyl butyrate. A substrate-specific effect favours a genuine active-site phenomenon.
- Fit a Hill equation to the 0.05–2 mM region only; nH ≈ 1 argues against negative cooperativity, which would also be unusual for a monomeric bacterial esterase.
2. Assuming the artefact is fixed and the enzyme behaves Michaelis-Menten with Km = 0.35 mM, at what substrate concentration would you run a 10,000-compound inhibitor screen, and why?
Read the worked answer
Choose [S] ≈ Km, i.e. 0.35 mM, giving v/Vmax = 0.35/(0.35 + 0.35) = 0.50.
Reasoning:
- At [S] = Km the enzyme is half-saturated, so the assay still has signal, but there is a large pool of free enzyme. Competitive inhibitors bind free enzyme, so their effect is maximal when [S] is low. At [S] = 10 × Km a competitive inhibitor with Ki = 2 µM would need to be present at ~22 µM to give 50% inhibition, which many screening libraries cannot reach.
- Going too low is also wrong: at [S] = 0.1 × Km the signal is only ~9% of Vmax and hit confirmation becomes noisy.
- Stay well below 2 mM: that is where the raw rate begins to fall and where the cosolvent problem lives.
Practical constraints to specify: hold DMSO at a constant 1% in all wells including controls; keep total A405 change below ~0.8 AU to remain in the linear range of the plate reader; note that p-nitrophenolate has pKa ≈ 7.15, so at pH 8.0 it is essentially fully ionised and the extinction coefficient is stable, but any buffer or compound that shifts pH will produce a false signal. Include a compound-only blank at A405 to catch intrinsically coloured or precipitating compounds.
3. A confirmed hit gives IC50 = 4 µM at [S] = 0.35 mM and IC50 = 22 µM at [S] = 2.0 mM. What is the inhibition mechanism and the inhibition constant?
Read the worked answer
The IC50 moves with substrate concentration, so the inhibitor competes with substrate for the free enzyme.
For competitive inhibition, IC50 = Ki(1 + [S]/Km). Solving at each point:
- [S] = 0.35 mM: 1 + 0.35/0.35 = 2.0 → Ki = 4/2.0 = 2.0 µM
- [S] = 2.0 mM: 1 + 2.0/0.35 = 6.71 → Ki = 22/6.71 = 3.3 µM
The two estimates agree within about 1.6-fold, which is well inside the error of IC50 determination and confirms a competitive mechanism with Ki ≈ 2–3 µM. If the inhibitor were non-competitive, the IC50 would be unchanged at both substrate concentrations; uncompetitive inhibition would show the IC50 falling as [S] rises, which is the opposite of what is seen.
Two caveats to state before committing chemistry resources:
- Two points cannot exclude a mixed mechanism. Confirm with a full matrix of 5 substrate concentrations × 6 inhibitor concentrations and fit globally to the Cheng-Prusoff-corrected models.
- The 2.0 mM point sits in the region where the raw rate was depressed in the original curve. Repeat it under constant 2% DMSO, otherwise the Ki at high [S] is confounded by cosolvent.