Advanced Materials Based on Plastic-Eating Enzymes (II): The enzymatic arsenal and its current limits

1. Fundamentals: why a protein can break down a plastic

I assume we all know that an enzyme is a protein that catalyses a specific chemical reaction. If we didn't, note it down — it's an exam question. Its function depends on the three-dimensional structure into which the amino acid chain folds, and in particular on a cavity called the active site, where the substrate fits and the reaction takes place.

In the hydrolases involved in polyester degradation, the active site contains three key amino acids (the catalytic triad), usually Ser-His-Asp or Ser-His-Glu, arranged in a conserved geometry. The serine performs the nucleophilic attack on the carbonyl carbon of the ester bond; the histidine acts as a base and the third position stabilises the histidine. Additionally, near the triad lies what we call the oxyanion hole, formed by several additional amino acids, whose function is to stabilise the negatively charged intermediate that forms during catalysis. This provides an energetic boost to the triad — very important.

Most of these enzymes share the α/β-hydrolase fold: a central parallel β-sheet flanked by α-helices, with the triad positioned on conserved loops. It is a structural scaffold that evolution has reused for highly diverse catalytic functions, and it works beautifully. The practical consequence of this mechanism defines the entire field.

Synthetic polymers can be divided into two groups:

Hydrolysable polymers, which contain ester or urethane bonds accessible to hydrolytic enzymes, such as PET, PLA, PCL and polyester-type polyurethanes.

Non-hydrolysable polymers, with a carbon-carbon backbone, such as PE, PS, PP and PVC. These lack susceptible bonds and require oxidative pretreatment.

This division explains why the enzymatic route, despite its potential, is not a universal solution today: C-C backbone polymers constitute the majority of European plastic waste due to their widespread use. Even so, there is always hope of finding a new enzyme or method to break that solid backbone. And why do I say this? Let's see.

2. Ideonella sakaiensis and the change of approach

The enzymatic degradation of PET had been documented since 2012 through the metagenomic isolation of LCC, a cutinase active at 70 °C. But the discovery of Ideonella sakaiensis provided a bacterium capable of using PET as a carbon and energy source, sustained by two enzymes acting sequentially. The PETase (IsPETase) hydrolyses the polymer generating mainly mono(2-hydroxyethyl) terephthalate (MHET), and MHETase completes the hydrolysis releasing the constituent monomers, terephthalic acid (TPA) and ethylene glycol (EG). Moreover, it does so at 30–37 °C, a relevant fact, since the equivalent chemical depolymerisation of PET requires temperatures between 400 and 800 °C in processes already discussed such as pyrolysis or gasification, or hydrolysis and methanolysis conditions with aggressive reagents. Quite a leap in terms of eco-sustainability and energy savings.

The discovery reoriented the field in a specific direction: the interest lies not in the organism itself, but in its tools.

3. Overview of the advantages of the enzymatic route over alternatives

The documented advantages over conventional methods are as follows.

• Monomer integrity. Enzymatic depolymerisation recovers the original monomers, suitable for repolymerisation at quality equivalent to virgin material.

• Energy efficiency. Enzymatic processes operate under mild temperature and pressure conditions.

• Selectivity. Substrate specificity allows targeting a specific polymer within a mixed waste input. This facilitates processing and reduces dependence on exhaustive prior separation.

• Absence of toxic by-products. The reaction does not generate dioxins nor require gas treatment systems comparable to those used in incineration.

• Contamination tolerance. Enzymatic systems process contaminated plastics that are unsuitable for other methods, expanding the range of recoverable material.

• Compatibility with existing infrastructure. Integration into waste management plants does not require a complete overhaul of facilities.
  

4. The available arsenal: four hydrolase families

Regarding polyester degradation, we have separated enzymes into four families, defined by their evolutionary origin and active site architecture.

Lipases and esterases. Originally associated with lipid metabolism, their preference for simple ester bonds makes them effective against aliphatic polyesters such as PLA and PCL. Many lipases possess a lid domain, a surface loop that covers the active site and undergoes a conformational change upon contact with a lipid interface, exposing the catalytic site. This mechanism, termed interfacial activation, conditions substrate accessibility and can represent a bottleneck for certain substrates. The absence of a lid in certain lipases produces a more exposed active site, functionally analogous to that of cutinases and PETases.

Relevant examples: a thermostable lipase from Brevibacillus thermoruber with a half-life exceeding 5 h at 60 °C on PCL, and the AlcB esterase from Alcanivorax borkumensis, active at 30 °C on PCL, PHB, PHBV and PBS in a marine environment.

Cutinases. Evolved to hydrolyse cutin, the waxy and partially aromatic biopolymer that coats leaf surfaces, they feature an active site in the form of a shallow, solvent-accessible groove suitable for accommodating bulky polymeric substrates. Flexible loops near the catalytic site allow modulation of the binding pocket size and shape.

LCC (leaf-branch compost cutinase), isolated in 2012 from leaf-branch compost by metagenomics and originating from an unidentified prokaryotic organism, is currently one of the reference biocatalysts for PET, with optimal activity at 70 °C. It incorporates two Ca²⁺-binding sites that contribute to its thermal stability. Thermobifida fusca cutinase (TfCUT) shows activity on various polyesters and features a disulphide bond critical for its thermostability.

PETases. Structurally close to cutinases, optimised for aromatic polyesters. Their narrow specificity is advantageous for PET recycling and limiting against other substrates. In addition to IsPETase, promiscuous esterases with PET activity have been identified in Bacteroidetes phylum bacteria, such as that from Kaistella jeonii, which features a Phe-Met-Tyr substrate-binding motif distinct from the Tyr-Met-Trp conserved in IsPETase, LCC and PET2.

C-N bond hydrolases. These act on amide and urethane bonds. The 6-aminohexanoate-dimer hydrolase (EII) from Flavobacterium sp. degrades nylon-6 oligomers and constitutes a documented case of enzymatic adaptation to an anthropogenic substrate: structural analyses indicate it derives from a pre-existing carboxylesterase with a β-lactamase-type fold that acquired new specificity through amino acid substitutions in the catalytic cleft.

Urethanases constitute the most recently developing subgroup. The structural characterisation of metagenomic urethanases has enabled their engineering to improve hydrolysis profiles on low-molecular-weight dicarbamates and thermoplastic polyester-type polyurethanes.

5. Protein engineering: the current state

Most natural enzymes do not initially reach the performance demanded industrially, which has shifted the field towards modified variants. Let's look at some examples:

The LCC variants obtained in 2020 (LCC-ICCG and LCC-WCCG) incorporate four mutations — F243W, D238C, S283C, Y127G — that optimise performance without altering the fold. The D238C and S283C substitutions form an additional disulphide bond that increases thermostability by approximately 10 °C; Y127G and F243W reduce steric hindrance near the active site. The variant degrades 90% of crystalline PET operating at 72 °C. Subsequent variants achieve 98% conversion in 24 h.

FAST-PETase, obtained in 2022 using a convolutional neural network (MutCompute) that predicted favourable mutations on IsPETase, incorporates five substitutions (S121E, D186H, R224Q, N233K, R280A) and operates optimally at around 50 °C, with the ability to depolymerise coloured and transparent post-consumer PET.

HotPETase, obtained by directed evolution using an automated high-throughput platform, was generated after six iterative rounds of random mutagenesis and screening at elevated temperature, evaluating over 13,000 variants. The final variant accumulates 21 mutations, reaches a melting temperature of 82.5 °C and depolymerises semicrystalline PET in the range of the glass transition temperature, between 60 and 70 °C.

Computational approaches have transformed the cost of these processes. AlphaFold has facilitated the acquisition of three-dimensional structural information to guide rational design. Binary classification models based on protein language models achieve average accuracies of around 89% in identifying enzymes with activity against specific plastic substrates, and applied to over 6,000 sequences they have expanded the repertoire of available candidates.

6. Success case: industrial application

Carbios has operated an industrial demonstration plant in Clermont-Ferrand since 2021, at a Michelin Group site. The facility validates the technical, environmental and economic performance of the process using a 20 m³ reactor capable of processing 2 tonnes of plastic (approximately 100,000 crushed bottles), fully depolymerising them to TPA and EG within 10 to 16 hours. The technology employs patented enzymes capable of specifically depolymerising PET contained in diverse plastics and textiles.

Samsara Eco (Australia) has developed enzymatic systems that operate on PET at room temperature in aqueous media with minimal preprocessing requirements. Birch Biosciences applies generative artificial intelligence and enzyme engineering to PET and PEF (polyethylene furanoate) recycling.

At the European level, the ENZYCLE project (Horizon 2020) specifically targets non-recycled plastic fractions (PET trays, clamshell containers, multilayer packaging) exploiting LCC and the metagenome-derived polyester hydrolase PHL7, which degrades 90% of PET within 16 hours. The FuturEnzyme project (Horizon 2020) integrates data mining, machine learning algorithms and protein engineering applied to textiles, detergents and cosmetics; its computational and functional screening of 1,000 preselected enzymes has identified 18 variants with efficacy one hundred times greater than market alternatives, enabling washing temperatures of 20–40 °C.

7. The challenges we face

Despite more than twenty years of research, very few plastic-degrading enzymes achieve the efficiency required for industrial use. The main limiting factors are:

• Substrate accessibility. The reaction occurs on an insoluble substrate, so kinetics are better described as a function of surface area. Moreover, the semicrystalline nature of synthetic polymers generates physical barriers, as densely packed regions are inaccessible. Additional factors such as particle size and surface structure significantly influence degradation rate.

• Additive complexity and contamination. Commercial plastic materials contain mixtures of additives, colourants and multilayer structures. Analysis of Belgian packaging reveals contamination from caps, labels and inseparable polymeric layers reaching 24% in multilayer films; metal contents of 1,133 ppm have been documented in polyethylene bottles. These additives can inhibit enzymatic activity and alter substrate surface properties.

• Limited thermostability. Most lipases lose activity above 60 °C. This creates an operational paradox: thermal treatment improves substrate accessibility by increasing chain mobility and reducing crystallinity, but those same conditions denature the biocatalyst. The PLA case exemplifies this: proteinase K shows optimal activity at 37 °C on PLA films, while degradation rates improve between 50 and 60 °C, where chain mobility increases.

• Economic viability. Enzyme production by microbial fermentation is costly and energy-intensive at the required scales. Combined with deactivation during processing and the difficulty of recovering the catalyst, it raises operating costs to levels that compromise competitiveness against virgin plastic production, especially with low crude oil prices.

• Process integration. Critical parameters for scale-up include polymer crystallinity, surface exchange, reaction temperature, enzymatic efficiency, substrate concentration, depolymerisation yield, product composition and expression yield. The absence of standardised infrastructure for collection, sorting and preprocessing further complicates implementation.

8. Implications

The three main bottlenecks (thermal denaturation, inability to recover the catalyst and substrate accessibility) are not limitations of the catalytic mechanism. They are limitations of stability, catalyst recycling and mass transfer — that is, materials engineering problems applied to a biological catalyst.

And that is a good sign, as it defines the available strategies to improve enzyme performance. Currently, the proposed strategies are grouped into three categories:

Modification of the gene encoding the protein (rational design and directed evolution).

Post-translational chemical modification in vitro of the already synthesised protein (PEGylation, glycosylation, cross-linking, polymer conjugation).

Modification of the reaction medium (ionic liquids and deep eutectic solvents).

The next chapter addresses all three.