The science
What the news is about. Recent (2026) coverage described a "Pac-Man"-like plastic-degrading enzyme that lets bacteria digest some plastics. The family behind such stories is the PET hydrolases: Ideonella sakaiensis PETase and MHETase (Yoshida et al., Science 2016), engineered variants such as FAST-PETase (Lu et al., Nature 2022) and the leaf-compost cutinase LCC-ICCG (Tournier et al., Nature 2020). This lab is a generic teaching model of that class, not a reproduction of any single paper.
Cascade. PET → MHET (mono-(2-hydroxyethyl) terephthalate) by PETase, then MHET → terephthalic acid (TPA) + ethylene glycol (EG) by MHETase. A small share (10%) of PETase cuts release TPA directly. In the ribbon model, helices are coils, arrows are beta strands, colour runs blue (N-terminus) to red (C-terminus), and the three glowing spheres mark the catalytic Ser-His-Asp triad. The translucent sphere with a wedge removed is a schematic outline, not a computed molecular surface.
Heterogeneous Michaelis-Menten. The enzyme works on a solid, so the "substrate" is accessible surface. With A the surface area per volume, Γ binding sites per area and Kd the dissociation constant, the bound complex follows the exact tight-binding equilibrium
ES = ½[(Kd+E+ΓA) − √((Kd+E+ΓA)² − 4EΓA)].
When sites far outnumber enzyme this reduces to ES = E ΓA/(Kd+ΓA), so v = Vmax A/(Km + A) with Vmax = kcat,eff E and Km = Kd/Γ (an area, cm²/mL). With E comparable to Kd the apparent Km shifts up by about E/(2Γ); Experiment 1 shows both numbers.
Crystallinity. Ordered crystalline regions are far less accessible than loose amorphous chains. The model gives crystalline surface only 3% of the amorphous turnover (an order-of-magnitude choice), so kcat,eff = kcat [(1−X) + 0.03 X]. Amorphous mass is eaten first, the remaining solid is increasingly crystalline, and the mass-loss curve flattens (Ronkvist et al., Macromolecules 2009, report the strong drop in rate with crystallinity).
Glass transition. Dry PET has Tg about 70 °C. Below it, chains are frozen; near and above it they become mobile and the enzyme can thread them into its cleft. Mobility is a sigmoid m(T) = 0.25 + 0.75/(1 + e−(T−(Tg−8))/3.5) (water plasticises the surface, hence the 8 °C offset). In the 3D view the amorphous chains visibly start to wiggle as T approaches Tg. Heating well beyond Tg lets amorphous PET crystallise ("annealing", cold crystallisation), modelled as kx = 0.01 h−1 e(T−75)/6, which converts amorphous to crystalline and shuts the film down.
Arrhenius versus denaturation. kcat(T) = k30 exp[−Ea/R (1/T − 1/303.15)] rises with temperature, but the enzyme unfolds irreversibly, kd(T) = (ln2/τ) exp[Ed/R (1/Tm − 1/T)], with half-life τ = 1 h at its melting temperature Tm and a very steep Ed. Averaged over an assay of length ta, the useful rate is multiplied by (1 − e−kd ta)/(kd ta). Rising Arrhenius times collapsing survival gives the bell curve of Experiment 2. Engineered variants shift Tm (wild-type about 48 °C, FAST-PETase about 66 °C, LCC-ICCG about 94 °C in this model), which is why they can work at or near Tg.
pH and inhibition. Activity follows a plateau f(pH) = 1/(1 + 10pKa1−pH + 10pH−pKa2). Acid release (1 H+ per MHET, 2 per TPA) lowers pH by a linearised buffer capacity (200 mM/pH strong, 8 mM/pH weak). MHET and TPA also inhibit PETase, 1/(1 + [MHET]/Ki + [TPA]/Ki'), which is why a working MHETase partner speeds up the whole job.
Simplifications. Constants are illustrative orders of magnitude, not fits to any dataset. Surface area shrinks as A ∝ M2/3 (shrinking flakes). Enzyme sites are identical and instantly equilibrated; no diffusion limits, no endo/exo distinction, no adsorption to glass, MHETase is assumed thermostabilised. Concentrations are per litre of reaction. The 3D enzymes, chain conformations and the number of molecules shown are schematic; the cutting events you see follow the computed mass loss. This is a conceptual model of enzyme kinetics, not an experimental method.