Epistasis arises from shifting the rate-limiting step during enzyme evolution of a β-lactamase

Abstract

Epistasis, the non-additive effect of mutations, can provide combinatorial improvements to enzyme activity that substantially exceed the gains from individual mutations. Yet the molecular mechanisms of epistasis remain elusive, undermining our ability to predict pathogen evolution and engineer biocatalysts. Here we reveal how directed evolution of a β-lactamase yielded highly epistatic activity enhancements. Evolution selected four mutations that increase antibiotic resistance 40-fold, despite their marginal individual effects (≤2-fold). Synergistic improvements coincided with the introduction of super-stochiometric burst kinetics, indicating that epistasis is rooted in the enzyme's conformational dynamics. Our analysis reveals that epistasis stemmed from distinct effects of each mutation on the catalytic cycle. The initial mutation increased protein flexibility and accelerated substrate binding, which is rate-limiting in the wild-type enzyme. Subsequent mutations predominantly boosted the chemical steps by fine-tuning substrate interactions. Our work identifies an overlooked cause for epistasis: changing the rate-limiting step can result in substantial synergy that boosts enzyme activity.

Keywords: Biocatalysis; Enzyme mechanisms; Hydrolases; Molecular evolution.

Fig. 1. Positive epistasis drives the evolution of OXA-48.

a, During directed evolution of OXA-48, selection for resistance against the oxyimino-cephalosporin CAZ was performed at increasing CAZ concentrations from 0.5 to 14 µM. b, CAZ resistance conferred by OXA-48 was improved 43-fold over five rounds of evolution (see Supplementary Table 1 for all IC₅₀ values). c, Mutations acquired during evolution, such as F72L, S212 and T213A, cluster around the active site serine (S70) and the Ω and β5–β6 loops that affect substrate specificity (purple). d, The adaptive landscape of the mutations found during evolution shows high epistasis. Each node represents a unique variant indicated by single-letter amino acid codes. Values within each node reflect the CAZ IC₅₀ fold change relative to wtOXA-48. Purple arrows indicate the trajectory followed during evolution (see Supplementary Table 2 for all IC₅₀ values). e, Comparison of the effects of single mutations (grey, F72L; dark grey, S212A; black, A33V; no expected effect for T213A) on the IC₅₀ fold changes along the evolutionary trajectory reveals a high degree of epistasis (purple). f, Comparison of the fold-change improvements relative to the previous variants reveals diminishing returns (purple area) in CAZ resistance. Source data

Fig. 2. Kinetic changes drive the evolution of OXA-48.

Kinetics are shown for wtOXA-48 (red) and the corresponding variants F72L (purple) and Q4 (blue, A33V/F72L/S212A/T213A). a, Enzymatic hydrolysis of β-lactams proceeds via an enzyme–substrate complex (ES), formation of an acyl–enzyme intermediate (EI), and hydrolytic deacylation (E + P). b, Evolution amplified the super-stoichiometric burst behaviour of OXA-48 (CAZ concentration: 50–400 µM, light to dark colours). c, In vitro burst-phase activities correlate well with the in vivo IC₅₀ fold changes. The line represents the Pearson correlation, and the error bands display the 95% confidence interval. d, Michaelis–Menten kinetics of the burst phase determined at 4 °C. e, Substrate binding was measured by W-fluorescence and was substantially accelerated during evolution (CAZ concentration: 100–1,200 μM, light to dark colours). f, Comparison of k₁ and k_cat/K_M between wtOXA-48 (red), F72L (purple) and Q4 (blue) reveals that binding is no longer rate-limiting in the burst phase of Q4 (determined at 25 °C; point above the diagonal line at k₁ = k_cat/K_M indicates that binding is not rate-limiting). Source data

Fig. 3. Evolution of a catalytically superior ensemble.

a, Ensemble refinement of wtOXA-48, F72L and Q5 reveals increased mobility of the Ω loop. b, ΔRMSF values relative to wtOXA-48, F72L and Q4 from MD simulations reproduce the increased flexibility of the Ω-loop region (see Supplementary Fig. 11 for other variants). Error bands indicate the standard error of the mean. c, PC and cluster analysis show that F72L modulates the conformational landscape in ways likely to accelerate binding displayed as the population shaded in blue (arrow added to highlight change in populations; see Supplementary Figs. 12 and 13 for other variants). d, Cluster representatives indicate that evolution (Q4 variant in blue versus wtOXA-48 in grey) displaced the Ω loop and adjacent α-helix as indicated by the arrow. e, Dynamical correlation analysis reveals that the movement of the acylated S70 becomes tightly coupled with the protein scaffold, particularly the oxyanion hole, by means of the alanine mutations. In contrast, F72L predominantly decreases the interaction of S70 with the Ω loop (increased and decreased correlations relative to wtOXA-48 are shown in blue and red lines, respectively). Line width corresponds to the strength of the correlation. Only statistically significant changes compared with wtOXA-48 are shown (t-test, α = 0.05; see Supplementary Fig. 15 for other variants). Source data