Metal–Organic Framework‐Gated Biocatalysis Enables Triggered Depolymerization of Melt‐Processed Polyesters

ABSTRACT

Controlling biocatalytic activity in melt‐processed polymers is a central challenge for triggered depolymerization, because enzymes deactivate at melt‐extrusion temperatures. Here, metal–organic framework‐gated biocatalysis, achieved by encapsulating enzymes within zeolitic imidazolate framework‐8 (ZIF‐8), preserves > 85% activity after 2 min at 180°C while regulating substrate access. Enzyme@ZIF‐8 biocomposite production scales to ∼50 kg day^{− 1} and is compounded by twin‐screw extrusion into poly(ε‐caprolactone) (PCL), poly(butylene adipate‐co‐terephthalate) (PBAT), and polylactide (PLA) at a tonne‐per‐day scale; pellets are compatible with standard thermoforming. The enzyme@ZIF/plastic composites retain mechanical performance comparable to the neat polymers during processing and use. At the end‐of‐life, chemical triggers dissolve the ZIF‐8 gate, releasing the enzyme, Zn^{2 +} and imidazolate to cooperatively accelerate depolymerization. Degradation increases 13.3–62.8‐fold for PCL and PLA in water and 1.7‐fold under industrial composting for PBAT and enables anaerobic PBAT digestion, whereas pristine polyesters show negligible conversion. This melt‐processable platform establishes gated, on‐demand depolymerization compatible with industrial polymer manufacturing.

Metal–organic framework gating functions as a molecular lock, enabling enzyme–MOF composites to be melt‐extruded into polyesters at 180°C under industrial thermomechanical conditions. ZIF‐8 encapsulation keeps enzymes dormant during service. Upon chemical triggering, the framework unlocks, releasing cooperative catalytic species that enable on‐demand depolymerization across diverse polyester systems.

1. Introduction

Controlling enzyme activity in non‐native, melt‐processed polymer matrices underpins applications ranging from on‐demand therapeutic delivery and sensing to programmable disassembly and recycling [1, 2, 3, 4, 5, 6]. These applications require that biological function be preserved in non‐native polymer environments and activated only in response to defined external triggers. To date, strategies that maintain enzymatic activity have largely avoided harsh processing conditions, instead relying on soft fabrication methods such as solution casting at or near room temperature. Such approaches have proven effective in polymeric thermosets, including hydrogels and elastomers, where mild crosslinking enables the entrapment of biomacromolecules while maintaining responsiveness to chemical or physical stimuli such as pH, specific molecular signals, redox processes, or light [7, 8, 9, 10].

Within this context, enzyme‐polymer composites are particularly attractive: as highly selective and efficient biocatalysts, enzymes are desirable for a range of chemical reactions, including polymer‐selective depolymerization [11, 12, 13, 14, 15, 16]. Embedding enzymes within polymers, therefore, offers a route to design programmable materials with self‐degradation functions [17], with direct relevance to the development of sustainable plastics technologies and strategies for mitigating anthropogenic polymer waste. However, extending such concepts to melt‐processed polymers under industrially relevant conditions remains a fundamental challenge, as the elevated temperatures and shear conditions required for industrial thermoplastic processing are generally incompatible with biological activity.

Proof‐of‐concept studies using soft processing have established feasibility; for example, a room‐temperature solvent approach enabled the encapsulation of lipase in a commercial polyester (i.e., PCL) [18, 19], and random‐heteropolymer compatibilization afforded uniform dispersion and quantitative self‐degradation of PCL upon triggering [10, 17, 20, 21, 22]. However, even when cast films show attractive properties [4], solution routes rely on organic solvents that can alter enzyme conformation and activity, complicate scale‐up preparation, and impose size/mechanical limitations on the resulting plastic commodities [23, 24, 25, 26]. Collectively, these limitations hinder mass production of enzyme–plastic composites, necessitating direct enzyme integration into thermoplastics via thermomechanical processing (extrusion, injection/blow molding, and calendering) used in megaton‐scale global plastics manufacturing [27]. Yet the very features that enable throughput, such as elevated temperatures (typically 100°C to 300°C) and intense shear in softened polymers, promote enzyme unfolding and inactivation. Furthermore, hydrophilic–hydrophobic mismatch promotes phase separation and heterogeneity during standard compounding [28]. Prior soft routes (solution casting/solvent systems, heteropolymer compatibilization) demonstrate the potential of bio‐intelligent plastics [29, 30, 31, 32]. Related strategies using hybrid inorganic matrices have been limited to gram‐scale processing of single polymers and simple aqueous degradation tests [33]. Overall, currently available strategies are rarely implemented in fully solvent‐free melt‐processing lines, and the few examples that approach such conditions still lack demonstrated scalability compatible with high‐throughput industrial manufacturing. Together, these limitations have constrained enzyme–plastic composites from reaching industrial relevance, motivating research to target strategies that address preservation, compatibility, and scalability simultaneously.

Here we show that an archetypal metal–organic framework (MOF) named zeolitic imidazolate framework‐8 (ZIF‐8) can preserve enzyme activity up to 180°C when used as an encapsulating shell under industrial melt‐processing conditions. The aqueous spontaneous in situ self‐assembly route of enzyme@ZIF‐8 scales to kilogram‐scale‐per‐day production and enables compounding with commercial polyester plastics via twin‐screw melt extrusion to form enzyme@ZIF‐8/polyester with ton‐scale‐per‐day capability. The enzyme, protected by the ZIF‐8 shell, does not significantly alter the polyester processing window or bulk mechanical performance, preserving processability. The resulting pelleted composite can be further thermoformed with standard methods used for the manufacturing of thermoplastic commodities (e.g., extrusion, hot pressing, and 3D printing). The enzyme remains MOF‐gated (dormant) in polyesters during use. This dormant OFF state is functionally analogous to microbial spores, which remain viable under harsh conditions and are rapidly reactivated when specific environmental conditions are provided. Then the enzyme can be switched ON by defined chemical triggers. In contrast to other framework materials employed solely as inert thermal shields, ZIF‐8 is deliberately selected so that both the embedded enzyme and the framework building blocks (Zn^{2 +} and 2‐methylimidazole, 2‐MeIM) actively co‐participate in the depolymerization chemistry. While water exposure has negligible influence, phosphate buffer, Tris‐EDTA, or acid conditions dissolve ZIF‐8, triggering the co‐release of enzyme and Zn^{2 +}/2‐MeIM and enabling the cooperative depolymerization of polyesters (Figure 1). This ON/OFF behavior combines industrial processability, in‐use stability, and triggered end‐of‐life degradation in one system, providing a practical route to enzyme‐containing, trigger‐responsive thermoplastics compatible with industrial melt compounding.

FIGURE 1.

Construction of enzyme@ZIF‐8/polyester composites with a triggered depolymerization function by enzyme‐metal cooperative catalysis.

2. Results and Discussion

2.1. Synthesis and Scaled Production of CALB@ZIF‐8 With Framework‐Gated Dormancy and High Stability

Candida antarctica lipase B (CALB) was encapsulated within ZIF‐8 by aqueous spontaneous in situ self‐assembly (Methods), affording a 77.5% encapsulation efficiency and ∼8 wt% enzyme loading. Electron microscopy images show rhombic dodecahedral ZIF‐8 crystals with an average size of ∼500 nm, whereas CALB@ZIF‐8 forms smaller particles of ∼200 nm (Figure 2a, Figures S1–S4, and Methods). Elemental maps show homogeneous distributions of Zn (Figure 2a) and N (Figure S5) in both pure ZIF‐8 and CALB@ZIF‐8 particles. After labeling CALB with Rhodamine‐B (RhB), fluorescence microscopy revealed a uniform distribution of CALB within ZIF‐8 crystals (Figure 2b and Methods). Calcination at 325°C for 2 h demonstrates intracrystalline distribution of the enzyme only in CALB@ZIF‐8 (Figure S6). Powder XRD matches the sodalite topology (Figure S7); FTIR shows the typical protein amide‐I band at ∼1654 cm^{− 1} only in CALB@ZIF‐8(Figure S8).

FIGURE 2.

Synthesis and characterization of the CALB@ZIF‐8 nanobiocomposite. (a) Transmission electron microscopy (TEM) image of CALB@ZIF‐8. Inset: energy dispersive spectroscopy (EDS) map showing the distribution of Zn elements in CALB@ZIF‐8. (b) Structured illumination microscopy (SIM) image of CALB@ZIF‐8, where CALB is labeled with Rhodamine B (RhB). Inset: high‐resolution image showing the distribution of RhB‐CALB in a single nanoparticle (scale bar = 500 nm). (c) Enzymatic activity of free CALB, CALB@ZIF‐8, and released CALB from ZIF‐8, where the activity of free CALB is regarded as 100%. (d) Thermal stability of CALB@ZIF‐8 compared with free CALB after heating at 80°C and 180°C. The solid powder samples were heated on a hot plate. After heating, the free CALB and CALB@ZIF‐8 samples were transferred into aqueous solutions, and the residual activity of free CALB and released CALB from ZIF‐8 was measured. The activity of corresponding samples before heating was regarded as 100%. (e) Structures of free CALB and CALB embedded in ZIF‐8 before (purple) and after heat treatment (500K, red) by MD simulation. (f) Thermal signal image of CALB@ZIF‐8 loaded on a silicon chip tested by a scanning thermal microscope. (g) Thermal conductivity of CALB@ZIF‐8 probed from the thermal signal of the same position. (h) Topographic (blue) and thermal conductivity (red) information of two slice observations at the corresponding locations (top for line 1 and bottom for line 2) for a better comparison. Data are presented as mean ± SEM (n = 3). **p < 0.01, ***p < 0.001.

Given the proven enzyme encapsulation, the MOF‐gated dormancy was confirmed by a p‐nitrophenyl butyrate assay (Methods); CALB@ZIF‐8 exhibits < 2% of the apparent activity of free CALB (Figure 2c, Tables S1 and S2), consistent with mass‐transport limitations through sodalite pore windows (∼3.4 Å) [34]. In phosphate buffer (PB), ZIF‐8 dissolution releases CALB with ∼90% native activity (Table S2 and Methods), indicating that ZIF encapsulation has minimal effect on CALB and establishes a trigger‐dependent OFF/ON activation pathway.

Melt extrusion is a standard thermoplastic processing method conducted above the melting temperature of semicrystalline polymers or the glass transition temperature of amorphous polymers [27]. Gross and coworkers [35] demonstrated that immobilized CALB (Novozym 435, N435) catalyzes enzymatic ring‐opening polymerization by reactive extrusion at 130°C. However, the polyacrylic carrier beads (0.3–1 mm) can hinder uniform dispersion, while the active adsorbed enzyme can cause uncontrolled degradation during processing and use [36], leading to mechanical deterioration. These limitations render N435 unsuitable as a functional filler when in‐use stability and predictable end‐of‐life behavior are required, motivating our investigation of CALB@ZIF‐8 as a nanofiller for PCL via thermomechanical compounding. As industrial melt extrusion typically lasts 2–5 min [37], we evaluated the thermal stability of CALB@ZIF‐8 over this processing‐relevant timeframe. CALB@ZIF‐8 retained ∼90% and 80% activity after heating for 5 min at 80°C and 180°C, respectively, whereas free CALB exhibited activity losses of ∼40% and 50% under the same conditions (Methods, Figure 2d). These results demonstrate the thermal robustness of CALB@ZIF‐8 required for thermomechanical processing.

SEM images of CALB@ZIF‐8 before and after heating at 180°C confirm the structural stability (Figure S9a–d). To examine conformational stability, molecular dynamics (MD) simulations were performed. Structural analysis (Figure 2e, Figure S9e) indicates significant disruption of CALB at elevated temperatures (Methods), whereas confinement within ZIF‐8 preserves enzyme structure with only minor changes, consistent with enhanced thermal stability. Previous studies have shown that MOF encapsulation improves stability in aqueous systems [38, 39]; our results demonstrate that these effects extend to thermal stabilization in dry CALB@ZIF‐8.

A scanning thermal microscope was used to measure the thermal conductivity of CALB@ZIF‐8 and probe heat transfer. Local thermal conductivity was calculated from the thermal signal and mapped. Topography (Figure S9f) and thermal information (Figure 2f–g) were collected along two scan lines, and profiles are reported in Figure 2h and Figure S9g. Reduced thermal conductivity was observed within CALB@ZIF‐8 relative to the substrate, indicating suppressed heat transfer. The average thermal conductivity of CALB@ZIF‐8 calculated from both profiles was 0.23 Wm⁻¹ K⁻¹, significantly lower than values reported for nanomaterials such as SiO₂ [40], Fe₃O₄ [41], and graphene [42]. These results demonstrate the reduced heat‐transfer properties of CALB@ZIF‐8. In summary, the enhanced temperature stability of CALB encapsulated in ZIF‐8 arises from restricted conformational mobility and the intrinsically low thermal conductivity of the ZIF lattice.

Next, CALB@ZIF‐8 synthesis was scaled up to a 50 L industrial batch reactor, achieving a production rate of ∼2 kg h⁻¹ (Figure S10a). The resulting material retained the morphology, structure, and activity observed at the laboratory scale (Figure S10b–d). This scale‐up enables the preparation of ∼0.5–5 tonnes day⁻¹ of CALB@ZIF‐8/PCL composite at 10–1 wt% loading.

2.2. Fabrication of CALB@ZIF‐8/PCL Composites by Melt Extrusion Technology

The thermal robustness and scalability of CALB@ZIF‐8 enabled processing with PCL under standard thermomechanical conditions to produce a self‐degradable biohybrid thermoplastic. CALB@ZIF‐8/PCL composites were fabricated via melt extrusion, an established industrial technology for thermoplastics [43]. As shown in Figure 3a, PCL pellets were dry‐mixed with CALB@ZIF‐8 and fed into a twin‐screw extruder operated at 80°C for melt blending (Movie S1). The extrudate was air‐cooled and subsequently processed using a commercial 3D printer (Figure S11). The extrudate was pelletized, demonstrating compatibility of ZIF‐protected enzymes with standard thermoplastic processing steps. CALB@ZIF‐8/PCL pellets were subsequently thermoformed by single‐screw extrusion to produce filaments and by hot pressing to form sheets. Composites containing 1, 5, and 10 wt% CALB@ZIF‐8 (denoted as 1%‐, 5%‐, and 10%‐PCL) were prepared. Electron microscopy confirmed the presence of CALB@ZIF‐8 within the polymer matrix and at the surface (Figure 3b and Figure S12). RhB‐labeled CALB was used to visualize CALB@ZIF‐8 distribution within PCL by fluorescence microscopy (Figure 3b and Figure S13–S15). Although minor nanoparticle agglomeration was observed, CALB@ZIF‐8 was generally uniformly dispersed at the micrometer scale. We note that incorporation of nanoparticles into a polymer matrix is known to introduce interfacial thermal resistance due to phonon mismatch during heat transfer [42, 44]. Accordingly, interfaces between CALB@ZIF‐8 and PCL may further contribute to enzyme preservation during high‐temperature melt extrusion. TGA analysis of CALB@ZIF‐8/PCL showed a characteristic mass loss at 200°C–300°C consistent with protein decomposition, confirming incorporation into PCL (Figure S16a,b).

FIGURE 3.

Synthesis and characterization of CALB@ZIF‐8/PCL composites. (a) Schematic illustration showing the preparation of different forms of CALB@ZIF‐8/PCL composites, using PCL pellets and CALB@ZIF‐8 as starting materials. (b) Scanning electron microscopy (SEM) images showing the cross‐sectional structure of CALB@ZIF‐8/PCL, where the weight percentage of CALB@ZIF‐8 in PCL is 10% (noted as 10%‐PCL). Inset in upper right: a local zoom‐in image illustrating the embedded CALB@ZIF‐8 nanoparticles (scale bar = 500 nm). Inset in upper left: surface of 10%‐PCL (where CALB was labeled with Rhodamine B) under the confocal laser scanning microscopy (CLSM), which demonstrated a uniform distribution of CALB in PCL. (c) Surface scanning of the distribution of Zn elements in 10%‐PCL using X‐ray photoelectron spectroscopy (XPS). (d) Normalized signal intensity of Zn in two directions indicated in (c). (e) High‐resolution Zn 2p XPS spectra at different depths of 10%‐PCL. (f) X‐ray diffraction (XRD) patterns of pure PCL, CALB@ZIF‐8, and CALB@ZIF‐8/PCL (10%‐PCL). (g) Fourier transform infrared spectroscopy (FTIR) spectra of pure PCL, CALB/PCL (containing the same amount of CALB as 10%‐PCL), ZIF‐8@PCL (containing the same amount of ZIF‐8 as 10%‐PCL), and CALB@ZIF‐8/PCL (10%‐PCL). (h) Mechanical testing of pure PCL, 1%‐PCL, 5%‐PCL, and 10%‐PCL. Inset: photos of 10%‐PCL during testing.

Surface photoelectron spectroscopy revealed a uniform Zn signal distribution in 10%‐PCL (Figure 3c–e), with depth profiling and elemental mapping confirming homogeneous surface dispersion of CALB@ZIF‐8 despite minor agglomeration (Figure S16c–j). High‐resolution spectra further showed that Zn remained in the Zn^{2 +} state of the ZIF‐8 lattice irrespective of the composite fabrication procedure (Figure S17). X‐ray diffraction confirmed the coexistence of CALB@ZIF‐8 and PCL phases (Figure 3f), while FTIR spectroscopy showed the characteristic ν(C═N) vibration of imidazole at ∼1579 cm⁻¹ both ZIF‐8@PCL and CALB@ZIF‐8/PCL (Figure 3g). Collectively, these analyses demonstrate successful incorporation of CALB@ZIF‐8 into PCL as a functional filler without alteration of its chemical structure.

Differential scanning calorimetry (DSC) showed minimal change in the crystallization and melting temperatures of PCL with increasing CALB@ZIF‐8 content (Figure S18a,b, Table S3), although a slight reduction in crystallinity was observed for 10%‐PCL (from 46.5% of pure PCL to 44.7%), likely due to nanoparticle‐induced hindered crystallization [45]. Rotational rheology measurements (Figure S18c–e) showed that loss modulus (G^″ ) exceeded the storage modulus (G^’ ) of the composites, indicating viscous‐dominated behavior favorable for polymer processing. The complex viscosity (η^* ) increased with the CALB@ZIF‐8 content due to additional entanglement points introduced by the filler, although changes in modulus and viscosity remained negligible relative to pure PCL at low loading (1%). Mechanical testing revealed a < 10% decrease in yield strength (12.3 MPa for PCL to 11.1 MPa for 10%‐PCL) and a ∼34% increase in elastic modulus (from 152 to 204 MPa, respectively) (Figure 3h, Table S4), indicating preserved polyester performance with increased stiffness from the rigid filler. In contrast, tensile strength and elongation decreased (Figure S19, Table S4), suggesting partial nanoparticle agglomeration. To mitigate this effect, 1 wt% polyvinylpyrrolidone (PVP) was introduced during CALB@ZIF‐8 synthesis as a surfactant/capping agent to improve nanoparticle dispersion (Figure S20) [46, 47, 48, 49]. Compared with 10%‐PCL, 10%‐PCL/PVP exhibited a significant enhancement in yield strength, tensile strength, and elongation (Figure S18f and Table S5). DSC analysis further showed a slight increase in crystallinity for 10%‐PCL/PVP (47.1%) relative to pure PCL (46.5%) (Figure S21, Table S6), suggesting that homogeneously dispersed CALB@ZIF‐8 nanocrystals may act as a nucleating agent during PCL crystallization.

2.3. On‐Demand Degradation of CALB@ZIF‐8/PCL Composites via Enzyme‐Metal Cooperative Catalysis

The stimuli‐responsive dissolution of ZIF‐8, which enables switching from protection to triggered release [50, 51, 52], was here used to design a PCL biocomposite with on‐demand degradation. CALB release from CALB@ZIF‐8 was evaluated using phosphate buffer (PB) and Tris‐HCl buffer containing EDTA (Tris‐HCl‐EDTA) as chemical triggers. Protein release, quantified by UV–vis spectroscopy (Methods, Figure S22, and Figure 4a) was faster in Tris‐HCl‐EDTA, reaching ∼66 wt% after 12 h and ∼100% after 24 h (Figure S23), consistent with visual clarification of the dispersion (Figure S24a). In contrast, negligible release was observed in deionized water or Tris–HCl. XRD and SEM confirm the structural and morphological degradation of ZIF‐8 after 12 h in PB and Tris‐HCl‐EDTA, but not in control media (DI or Tris‐HCl, Figure 4b), consistent with previous reports of ZIF‐8 dissolution under acidic conditions, in the presence of phosphate ions, and chelating agents [53, 54].

FIGURE 4.

On‐demand release of CALB and Zn²⁺/2‐MelM as cooperative catalysts for the hydrolysis of PCL. (a) Concentration of CALB released in the supernatant as a function of time. Condition: CALB@ZIF‐8 (5 mg) was dispersed in 2 mL aqueous solution, including DI water, phosphate buffer (PB, 0.1 M, pH = 7), Tris‐HCl buffer (0.1 M, pH = 7), and Tris‐HCl‐EDTA buffer (0.1 M, pH = 7). All solutions were stirred at 50°C. The supernatant was collected every 2 h by centrifugation at 12 000 rpm for 5 min. The protein was then stained with Coomassie brilliant blue to detect protein released into the supernatant. Values and error bars represent the means and SD from at least three independent measurements. (b) X‐ray diffraction (XRD) patterns and scanning electron microscopy (SEM) images (inset, bar = 100 nm) of CALB@ZIF‐8 after 12‐h incubation in corresponding buffer solutions, which were obtained by centrifugation at 12 000 rpm for 5 min, washed with DI water, and then lyophilized. ZIF‐8 characteristic reflections from the (011), (002), (112), (022), (013), and (222) planes (Figure S7) were retained in DI water and Tris‐HCl. In PB and Tris‐HCl‐EDTA, peak attenuation, broadening, and loss of high‐angle reflections indicate reduced crystallinity and framework degradation. (c) Gel permeation chromatography (GPC) spectra of PCL before degradation and PCL residues after 12‐h hydrolysis by enzymatic degradation and cooperative degradation, respectively. The hydrolysis reaction was performed in Tris‐HCl buffer (0.1 M, pH = 8). (d) Liquid chromatography‐mass spectrometry (LC‐MS) spectra of the PCL hydrolysis products obtained by 12‐h enzymatic degradation and cooperative degradation, respectively. (e) Schematic illustration of the on‐demand release and cooperative catalysis. Data are presented as mean ± SEM (n = 3). **p < 0.01.

Upon ZIF‐8 dissolution, both CALB and the framework components (Zn^{2 +} and 2‐MeIM) were released. CALB alone is known to degrade PCL by random chain scission [17, 18], which we confirmed by gel permeation chromatography (GPC), showing broad molecular‐weight distributions after 12 h (Figure 4c). In contrast, enzymatic degradation of PCL in the presence of CALB, Zn^{2 +} and 2‐MeIM (4, 8, and 12 h) yielded shorter chains with lower molecular weights (Figure 4c, Figure S24b, and Figures S25, S26), indicating a distinct cooperative pathway. Notably, Zn²⁺ acts as a Lewis acid to promote ester bond hydrolysis [55] (Figure S24c). Simultaneously, 2‐MeIM functions as a nucleophile that also facilitates ester cleavage [56] (Figure S24d). Although Zn²⁺ or 2‐MeIM alone had negligible effects on PCL (Figures S27, S28), their combination with CALB significantly enhanced hydrolytic activity (Figure S24e,f). These findings demonstrate that short PCL fragments generated by enzymatic scission were further hydrolyzed to monomers by the ZIF building blocks.

To further examine this synergistic effect, the degradation products after 12 h were analyzed by LC–MS. The peaks at ∼16, 23, 28, and 33 min in the total ion chromatography (TIC) spectra corresponded to caprolactone monomers (n = 1), dimers (n = 2), trimers (n = 3), and tetramers (n = 4) of caprolactones, respectively (Figure S29–S31). By comparing the peak areas in the TIC spectra, the distribution of degradation products was determined (Figure 4d). CALB alone generated 62.6% monomers, whereas the monomer fraction increased to 71% and 91.7% in the presence of Zn²⁺ and 2‐MeIM, respectively.

Together, these results show that released CALB and ZIF‐8 precursors act cooperatively in PCL degradation (Figure 4e). The altered byproduct distribution confirms additional oligomer hydrolysis by framework‐derived species, establishing a cooperative pathway in which enzymatic scission and framework‐assisted catalysis synergistically accelerate depolymerization to monomers.

This stimuli‐responsive, enzyme‐metal cooperative catalysis enabled the on‐demand degradation of CALB@ZIF‐8/PCL composites. At 50°C in DI water, CALB@ZIF‐8/PCL shows negligible mass loss, whereas CALB/PCL exhibited rapid degradation, losing 85.41% of its mass within 1 h (Figure 5a). In long‐term tests, CALB@ZIF‐8/PCL displayed less than 10% mass loss after 4 days in deionized water or Tris–HCl buffer at 50°C (Figure 5b and Figure S32). These results demonstrate that unprotected CALB induces uncontrolled degradation, compromising shelf‐life and in‐use stability, whereas ZIF‐8–gated CALB preserves polymer integrity under aqueous conditions.

FIGURE 5.

On‐demand degradation of enzyme@ZIF‐8/polyester composites. (a) Remaining mass of CALB@ZIF‐8/PCL (10%‐PCL) and CALB/PCL (containing the same amount of CALB as 10%‐PCL) as a function of degradation time in DI water at 50°C. Values and error bars represent the means and SD from at least three independent measurements. (b) Remaining mass of 10%‐PCL in different aqueous solutions as a function of degradation time at 50°C, including DI water, phosphate buffer (PB, 0.1 M, pH = 7), Tris‐HCl buffer (0.1 M, pH = 7), and Tris‐HCl‐EDTA buffer (0.1 M, pH = 7). Values and error bars represented the means and SD from three independent samples. (c) Scale‐up preparation of cutinase/protease@ZIF‐8 in 50 L reactor, enzyme@ZIF‐8/polyester containing 1 wt% enzyme@ZIF‐8 (noted as 1%‐PBAT, 1%‐PLA) by a melt extruder at 130°C–180°C. Shown on the right are cutinase@ZIF‐8/PBAT filament and film. (d) Thermal stability of cutinase@ZIF‐8 and protease@ZIF‐8 compared with free cutinase and protease after heating at 180°C. The activity of corresponding samples before heating was regarded as 100%. (e) Biodegradation of 1%‐PBAT and PBAT in industrial composting. (f) Biodegradation of 1%‐PLA and PLA at 45°C in PB (0.1 M, pH = 8) solution. Data are presented as mean ± SEM (n = 3). ***p < 0.001.

In contrast, when degradation was deliberately triggered in PB or Tris–HCl–EDTA buffer, CALB@ZIF‐8/PCL showed rapid mass loss (> 80% after 4 days) (Figure 5b and Figure S32). XPS depth profiling revealed phosphorus within the PCL matrix after 24 h in PB (Figure S33a), confirming phosphate ion diffusion and ZIF‐8 decomposition. Degradation was thickness‐dependent: reducing the average thickness from 180 to 60 µm decreased the remaining mass from 50% to 20% (Figure S34), consistent with the higher surface‐to‐volume ratio and increased water uptake in thinner films. As controls, pure PCL and ZIF‐8/PCL showed less than 5% mass loss after a 4‐day immersion in Tris‐HCl‐EDTA buffer (Figure S33b), confirming that ZIF‐8 precursors alone do not significantly degrade PCL chains.

To demonstrate generality across polyesters with higher melt‐processing temperatures (up to 180°C), cutinase@ZIF‐8 and protease@ZIF‐8 were synthesized in a 50 L batch reactor to produce 50 kg of nanocomposites, respectively (Figure 5c). After exposure to 180°C for 20 min, cutinase@ZIF‐8 and protease@ZIF‐8 retained 46% and 52% activity, respectively, whereas the corresponding free enzyme retained only 8% and 10% (Figure 5d), confirming the broad thermal protection conferred by ZIF‐8 encapsulation. Cutinase@ZIF‐8 and protease@ZIF‐8 were subsequently melt extruded at 130°C–150°C and 160°C–180°C to produce PBAT and PLA containing 1 wt% cutinase@ZIF‐8 (1%‐PBAT) and 1 wt% protease@ZIF‐8 (1%‐PLA), respectively, at 0.5‐ton scale (Figure 5c, Figure S33c). The resulting enzyme‐containing PBAT and PLA were further processed by film blowing (Figure 5c, Figure S33c). Under industrial composting conditions, acid‐triggered dissociation of cutinase@ZIF‐8 accelerated PBAT degradation, with 1%‐PBAT degrading 1.7‐fold faster than pure PBAT and exceeding 90% mass loss after 60 days (Figure 5e). Interestingly, degradation was also observed under anaerobic digestion, whereas pure PBAT remained stable (Figure S33d). This rapid breakdown under industrial composting and anaerobic digestion addresses a major limitation in post‐use treatment of PBAT plastics. In addition, 1%‐PLA exhibited rapid degradation in PB, with > 80% mass loss within 120 min (Figure 5f), in contrast to the slow hydrolysis of pure PLA [57].

3. Conclusion

In summary, this study demonstrates the thermal robustness and scalable production of enzyme@ZIF‐8, establishing the industrial processability of enzyme–MOF composites. Enzyme@ZIF‐8 was incorporated into polyesters as a “dormant” biocatalyst via melt extrusion, with the ZIF‐8 shell providing protection during processing, gating the enzyme during use, and supplying cooperative catalytic species (Zn^{2 +} and 2‐MeIM) upon triggered dissolution. This dual functionality enabled rapid, on‐demand polyester degradation. Together, these findings define a practical strategy for enzyme‐containing, stimuli‐responsive plastics compatible with industrial manufacturing.

Author Contributions

Shitong Cui: conceptualization, methodology, formal analysis, validation, investigation, visualization, writing – original draft, data curation. Jing Tian: methodology, validation, investigation, data curation. Mengyu Zhu: writing – review and editing, visualization, software. Zihan Liu: validation, data curation. Yufei Cao: software, validation. Yu Liu: investigation. Yuxiao Feng: investigation. Wanghui Xu: resources. Weijian Lai: resources. Hongyi Yang: resources. Jun Xu: supervision, resources. Baohua Guo: supervision. Lin Qiu: supervision, resources. Zhenzhong Yang: supervision, resources. Paolo Falcaro: supervision, writing – review and editing, visualization. Jun Ge: conceptualization, investigation, writing – review and editing, resources, supervision, project administration, funding acquisition.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: anie73397‐sup‐0001‐SuppMat.docx.

Supporting File 2: anie73397‐sup‐0002‐MovieS1.mp4.

Data Availability Statement

The data that support the findings of this study are available in the main text or the supplementary materials.