a Key Laboratory for Waste Plastics Biocatalytic Degradation and Recycling, College of Biotechnology and Pharmaceutical Engineering, Nanjing Tech University, Nanjing 211800, China
b State Key Laboratory of Materials-Oriented Chemical Engineering, Nanjing Tech University, Nanjing 211800, China
c Environment Research Institute, Shandong University, Qingdao 266237, China
d Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences, Tianjin 300308, China
e Department of Biotechnology and Enzyme Catalysis, Institute of Biochemistry, University of Greifswald, Greifswald 17489, Germany
Polyurethane (PU) is highly resistant to biodegradation, primarily due to the intrinsic stability of its urethane bond. Aside from a small number of amidases reported to hydrolyze (poly)urethane bonds, several promiscuous esterases have also been found to catalyze PU degradation. In this study, we clarified the ligand-free crystal structure of Aes72, an esterase enzyme that exhibits promiscuous hydrolytic activity toward carbamate and amide bonds, at a high resolution of 1.80 Å. We investigated the catalytic mechanism underlying urethane bond cleavage by Aes72 using multiscale quantum mechanics/molecular mechanics (QM/MM) simulations. Our findings indicate that the reaction mechanism consists of four concerted elementary steps, with the nucleophilic attack (step i) identified as the rate-determining step. The subsequent structure-guided engineering of Aes72 yielded several enhanced single mutants, ultimately resulting in a superior double mutant, F276A/L141I. This variant exhibited approximately a two-fold increase in catalytic efficacy toward bis(4-hydroxybutyl) (methylenebis(4,1-phenylene)) dicarbamate (BMC) hydrolysis and significantly enhanced degradation performance on two distinct polyether-based PU materials compared to the wild-type enzyme. Our findings provide essential mechanistic insights into the structure-function relationship of the promiscuous esterase Aes72 in PU degradation and demonstrate its potential applicability in bio-based plastic recycling.
Each year, a staggering 413.8 million tonnes of plastic are manufactured across the world, underscoring its role as a fundamental material in modern society [1-3, 4]. Polyurethanes (PUs) are the fifth most-produced synthetic polymer, with a global production of 21.9 million tonnes in 2023 [1]. The widespread use of PU products has led to significant challenges in waste management and consequent environmental pollution. Recycling is a common end-of-life strategy for PU waste that typically involves mechanical processing through shredding and thermo-compression, followed by downcycling applications [5]. Alternatively, PU waste can be chemically recycled to depolymerize the polymer under elevated temperatures and pressures, enabling the recovery of specific building blocks for polymer resynthesis [5]. However, both approaches face limitations, including high energy consumption, the generation of numerous by-products, and stringent requirements for raw material quality [6]. Therefore, the development of broadly applicable green disposal technologies for PU waste is of great interest to both the scientific community and industrial stakeholders.
Enzyme- catalyzed recycling technology can depolymerize various polyester-type plastics into their monomers under mild reaction conditions and without the use of organic solvents [7,8]. This approach builds on over 20 years of extensive research into the discovery and design of polyethylene terephthalate (PET)-depolymerizing enzymes [7,9,10], which are among the most prominent engineered catalysts for the recycling and upcycling of post-consumer PET waste in industrial applications [7,11-13]. Similarly, the biocatalytic recycling and upcycling of other hydrolyzable plastics, including PU, has been proposed as a viable approach for achieving the sustainable and low-carbon disposal of PU waste [6,7,13]. Nonetheless, the vast majority of reported PU hydrolases are polyester hydrolases, which primarily hydrolyze the ester bonds in the polyols and produce the minor fraction of polyester-based PUs [14-18]. Consequently, the identification of carbamate hydrolases that can effectively hydrolyze urethane bonds, the predominant linkages in both polyester- and polyether-based PUs, remains a significant challenge for the depolymerization and recycling of diverse PU wastes.
Recently, several members of the amidase signature superfamily, designated UMG-SP-1, UMG-SP-2, and UMG-SP-3, have been identified from environmental metagenomes, characterized, and shown to efficiently hydrolyze low-molecular-weight carbamates [19], as well as carbamate bonds in PU and amide bonds in nylon 6 [20]. Recently resolved crystal structures, together with mechanistic studies employing protein engineering and computational simulation, have provided insights into the mechanisms underlying PU depolymerization [20-23]. These insights complement recent advances in elucidating the mechanistic details of PET enzymatic hydrolysis, deepening our understanding of the biocatalytic depolymerization process.
In this study, we conducted biochemical and structural analyses of the promiscuous esterase Aes72, an enzyme exhibiting urethanase activity that was first reported in a patent application submitted by Coverstro AG [24]. In addition to confirming its catalytic promiscuity toward ester, amide, and carbamate bonds, we determined the ligand-free crystal structure of Aes72 at 1.8 Å. This crystal structure enabled the computational study of its catalytic mechanism in urethane bond hydrolysis through a quantum mechanics/molecular mechanics (QM/MM) methodology. Additionally, through semi-rational design, we generated Aes72 variants and evaluated their hydrolytic activities toward both short-chain carbamates and various polyether-based PU substrates.
2. Materials and methods
2.1. Strains, plasmids, and chemicals
The Aes72 gene was synthesized and cloned into the pET-29a(+) plasmid by GenScript Biotech Co., Ltd. (China) for recombinant protein expression. All molecular biology reagents were obtained from Takara Bio Inc. (China). All chemical reagents used in this study were sourced from Aladdin Co., Ltd. (China) and were of analytical grade or higher purity, unless specified otherwise.
2.2. Amino acid sequence analysis of Aes72
The amino acid sequence of Aes72 was analyzed for homology using the Protein Data Bank (PDB). Subsequently, multiple sequence alignments of the related proteins were performed using ESPript 3.0 (Fig. S1 in Appendix A).
2.3. Protein expression and purification for crystallography
The pET-29-Aes72 plasmid was introduced into Escherichia coli BL21(DE3) cells, and the resulting protein was purified as described in Ref. [14]. The proteins were then further purified using a diethylaminoethyl (DEAE) 6FF chromatography column (Fig. S2 in Appendix A) [21]. The purity of each protein (> 95%, 10 mg∙mL-1) was used for crystallization screening.
2.4. Crystallization and structure determination
Crystallization experiments were conducted as described in Ref. [21]. In each trial, 1 μL of Aes72 solution (10 mg∙mL-1 in 25 mmol∙L-1 (mM) Tris-HCl, pH 7.5, 150 mM NaCl) was mixed with an equal volume of the reservoir solution in 48-well Cryschem plates. This mixture was allowed to equilibrate with 100 μL of the reservoir solution. The optimal crystallization conditions for Aes72 were 0.2 mol∙L-1 (M) ammonium acetate, 0.1 M trisodium citrate dihydrate (pH 5.6), and 30% w/v polyethylene glycol (PEG) 4000. Crystals suitable for X-ray diffraction were obtained within 3-4 d.
X- ray diffraction data were collected using an in-house Rigaku FR-X (Rigaku, Japan) and the BL17B, BL17U, BL18U1, BL19U1, and BL10U2 beamlines at the National Facility for Protein Science in Shanghai, China. The diffraction images were processed and analyzed using HKL-2000 software [25].
The crystal structure of Aes72 was solved by molecular replacement using the Phaser program [26] from the Collaborative Computational Project No. 4 (CCP4) software suite [27]. The structure of an esterase from an uncultured bacterium (PDB: 7AT3) was used as the search model. Subsequent structure refinement was carried out using the REFMAC5 [28] and Coot [29] programs. Five percent of the reflections were randomly reserved for the calculation of Free R [30].
2.5. Molecular docking
Molecular docking experiments were performed using the Python 3 bindings of AutoDock Vina (1.2.5) [31]. The ligands (bis(4-hydroxybutyl) (methylenebis(4,1-phenylene)) dicarbamate, BMC) and receptor (Aes72 structure) were prepared using the scripts from the MGLTools distribution (1.5.7) [32]. The docking box was manually specified using University of California, San Francisco (UCSF) Chimera (1.16). Docking was performed using the default settings, except that the "exhaustiveness" parameter was increased to 25. The results were visualized using PyMOL.
2.6. System preparation and molecular dynamics simulations
The protonation states of titratable residues were assigned based on the calculated pKa values [33]. The model substrate BMC was constructed and refined at the M06-2X/6-31G(d) theoretical level with Gaussian 16 software [34, 35]. The binding orientation of BMC within the active site was determined by molecular docking using AutoDock [36]. The oxygen atom of the catalytic residue Ser147 was chosen as the center of the grid box, with dimensions of 30 grid points along each of the X, Y, and Z axes. Chemistry at Harvard macromolecular mechanics (CHARMM) force field parameters for BMC were generated using SwissParam [37]. The entire complex was fully solvated in a TIP3 water model with a 30 Å radius and neutralized with 14 sodium ions, resulting in a system of approximately 14 000 atoms. The system was then energy-minimized to eliminate clashes, followed by gradual heating from 0 to 323.15 K. Finally, molecular dynamics simulations were conducted for conformational sampling.
2.7. QM/MM calculations
The calculations utilizing QM/MM were conducted using ChemShell 3.6.0 [38] and DL_POLY software. A single conformation of the enzyme was randomly selected from the molecular dynamics simulation. During the acylation phase, the quantum mechanics (QM) region included the catalytic triad (Ser147, Glu241, and His271), the oxyanion hole (Gly79, Ala80, and Ala148), and the substrate, totaling 112 atoms (Fig. S3 in Appendix A). During the deacylation step, an additional water molecule near the active site was included in the QM region, resulting in 113 atoms for ester bond cleavage and 84 atoms for amide bond cleavage (Figs. S4 and S5 in Appendix A). The active molecular mechanics (MM) region contained approximately 6500 atoms during acylation and approximately 6000 atoms during deacylation. Atoms within a 20 Å radius of the substrate were allowed to move freely, whereas the remaining atoms in the complex were kept fixed during the calculations. The geometries of the reactants, intermediates, and products were optimized using a hybrid delocalized internal coordinates optimizer. The structure with the highest potential energy along the reaction coordinates was further optimized with a micro-iterative TS optimizer to locate the transition state. The reaction potential energy surface and geometry were optimized at the M06-2X/6-31G(d)//CHARMM level, while the single-point energies were calculated at the M06-2X/6-311++G(3df,3pd)//CHARMM level.
2.8. Biochemical characterization of Aes72
2.8.1. Kinetic analysis of the hydrolysis of small-molecule substrates with different chemical bonds
PU model substrates were used to assess the hydrolytic activity of Aes72. The increase in absorbance at 410 nm (para-nitrophenyl-butyrate, pNPB) and 405 nm (butyl(4-nitrophenyl)carbamate, BNC; para-nitroacetanilide, pNAA) was measured to track the generation of p-nitrophenol and p-nitroaniline, respectively [39]. One unit of enzyme activity was defined as the amount of enzyme required to catalyze the formation of 1 μmol of product per minute. The activity of Aes72 was measured at various concentrations (20-1200 μmol∙L-1 (μM)) of pNPB, BNC, and pNAA. The kinetic parameters (limiting rate (Vmax), Michaelis constant (Km), and catalytic rate constant (kcat)) were calculated by fitting the activity data to the Michaelis-Menten equation using the least-squares method. This analysis was performed using Origin version 8.0 [40, 41].
2.8.2. Effects of temperature and pH on the activity and stability of Aes72
BNC, which contains a urethane bond, was used as the substrate for the enzyme activity assay. The optimal reaction temperature was determined in the range of 25-80 °C in 50 mM phosphate buffer (PB, pH 7.0). The thermal stability of Aes72 was evaluated by incubating the enzyme in 50 mM PB (pH 7.0) at 50, 55, and 65 °C. The optimal reaction pH was determined using various buffers at 65 °C. Different buffers were used to cover a broad pH range: 50 mM citrate buffer (pH 3.0-5.0); 50 mM PB (pH 5.0-8.0); and 50 mM glycine-NaOH buffer (pH 8.0-11.0). The pH stability of Aes72 was assessed by incubating the enzyme in varying buffers at 4 °C for 24 h. The residual activities were measured under the enzyme assay conditions as described above, and the results are shown in Fig. S6 in Appendix A.
2.9. Chemical syntheses of BMC and polyether-PU
The low-molecular-weight PU model substrate (BMC) was also synthesized. A solution of 50 g of 1,4-butanediol (BDO) in 100 mL of dimethylformamide (DMF) was prepared, to which 12 g of methylene diphenyl diisocyanate (MDI) dissolved in DMF was gradually added. The mixture was then heated to 100 °C for 2 h under reflux. Subsequently, the solution was transferred into water and extracted with dichloromethane, yielding BMC as a white solid. This solid was further purified by column chromatography, and its chemical structure was confirmed by proton nuclear magnetic resonance (1H NMR) (Fig. S7 in Appendix A).
PEG- PU was synthesized in a jacketed reactor at 80 °C under a nitrogen atmosphere with mechanical stirring. Stoichiometric conditions were applied with a polyol-to-MDI molar ratio of 1:1, corresponding to an average of about 9 repeating units per molecule (personal communication). PEG was first dissolved in dimethyl sulfoxide (DMSO) within the reactor, followed by the slow addition of MDI, also dissolved in DMSO. The reaction mixture was stirred continuously and slowly heated to 85 °C, maintained at this temperature for 12 h, and then freeze-dried. The chemical structure of PEG-PU was confirmed by 1H NMR, and its molecular weight was determined by gel permeation chromatography (GPC) (Fig. S8 in Appendix A).
2.10. Site-directed mutagenesis of Aes72
The amino acid sequence of Aes72 was used to search for closely related sequences in the PDB database. These 12 sequences were aligned using ClustalW in MEGA 10, and the resulting alignment was input into WebLogo for visualization. Finally, WebLogo was used to display the frequency of each amino acid at every position in the sequence as well as the degree of conservation at each position [42].
Plasmids encoding the Aes72 variants were obtained through site-directed mutagenesis, as described in Ref. [21]. The mutated plasmids were sent for sequencing at Tsingke Biotech Co. in Nanjing, China. Following this, the variants were induced and cultured as previously described.
2.11. Enzymatic hydrolysis of BMC
For the BMC degradation assay, a 1 mL reaction mixture containing 0.6 mg of purified wild-type (WT)-Aes72 or its variants, 50 mM PB (pH 7.0), and 100 mg∙L-1 of BMC was incubated at 60 °C for 10 min. After incubation, the reaction was terminated by rapidly cooling the mixture in an ice bath. Subsequently, 4-aminoantipyrine and potassium hexacyanoferrate(III) were added to final concentrations of 1 and 2 mM, respectively. In the presence of potassium hexacyanoferrate(III) as a catalyst, the amines reacted with 4-aminoantipyrine to produce a red compound with a maximum absorbance at 515 nm [43]. The absorbance at this wavelength was measured after a 5 min incubation. One activity unit was defined as the amount of enzyme required to produce 1 μmol of amines per minute. Due to the poor stability of BMC intermediates, their limited commercial availability, and the challenges associated with their synthesis, 4,4′-methylenedianiline (MDA) was selected as the representative amine for correlation measurements in colorimetric assays. Various concentrations of MDA were added to the reaction to achieve final concentrations of 0.025, 0.05, 0.1, 0.2, and 0.25 mM. Subsequently, 4-aminoantipyrine and potassium hexacyanoferrate(III) were added to achieve final concentrations of 1 and 2 mM, respectively. Following a 5 min incubation, the absorbance at 515 nm was recorded to generate a standard curve (R2 = 0.997) (Fig. S9 in Appendix A).
The effect of Aes72 on BMC was evaluated over a BMC concentration range of 0.1-5000 μM. The kinetic parameters Vmax, Km, and kcat were calculated using Origin 8.0 software. We applied a formulation based on the theory of ideal catalysts proposed by Albery and Knowles [44], specifically designed for irreversible conditions, to calculate the efficiency function (Ef) [45]:
where kdif represents the rate of a diffusion-limited process, approximately 109 M-1∙s-1. Ef is a function that describes the catalytic characteristics of an enzyme in relation to the actual substrate concentration [S]. The parameter serves as a practical measure of catalytic efficiency, which is useful for comparing different enzyme variants acting on the same substrate.
2.12. Enzymatic hydrolysis of polyether-PU
PEG- PU was first ground with a mortar and then sieved through a 40-mesh sieve. The sieved particles were used for subsequent degradation experiments. For the PEG-PU degradation assay, a 5 mL reaction mixture containing 2 mg of purified WT-Aes72 or mutant F276A/L141I (12 µM biocatalyst), 10 mg of PEG-PU (2 g∙L-1, enzyme-to-substrate ratio (w/w) 1:5), and 50 mM PB (pH 7.0) was incubated for 36 h at 50 °C with shaking at 800 r∙min-1. Aliquots of 200 μL were taken from the reaction at 2, 4, 6, 9, 12, 24, 36, and 48 h, and the MDA concentration was measured to determine the yield, assuming a theoretical maximum release of 1.1 mM MDA from the complete hydrolysis of 2 mg∙mL-1 substrate. The degradation efficiency of Aes72, UMG-SP-1, and UMG-SP-2 on PEG-PU was evaluated over a 12 h period using the same assay conditions described above. Additionally, the degradation efficiency was assessed for 12 h at the optimal temperature for each urethanase: 50 °C for Aes72 and 37 °C for UMG-SP-1 and UMG-SP-2. Commercial PU foam was first shredded and then used for degradation assays. A 5 mL reaction mixture containing 2 mg of either purified WT-Aes72 or mutant F276A/L141I, 10 mg of PU foam, and 50 mM PB (pH 7.0) was incubated at 50 °C and 800 r∙min-1 for 12 h. Upon reaction completion, the supernatant was collected by centrifugation at 8000 r∙min-1.
MDA released into the supernatant was analyzed by high-performance liquid chromatography (HPLC) [12]. The supernatant obtained from PEG-PU degradation was freeze-dried, re-dissolved in methanol, and filtered through a 0.22 µm nylon membrane. The intermediates formed during PEG-PU degradation were analyzed by quadrupole time-of-flight mass spectrometry (Agilent 6530, Agilent, USA) using electrospray ionization in the positive ion mode with the following parameters: capillary voltage, 4 kV; gas temperature, 350 °C; drying gas, 5 L∙min-1; Oct 1 radiofrequency (RF) peak-to-peak voltage (Vpp), 750 V; mass-to-charge ratio (m/z) range, 50-2000. Recycled PEG-PU or foams were washed with PB containing sodium dodecyl sulfate (SDS) and then dried in an oven at 60 °C to a constant weight. Subsequently, the weight loss was assessed, the molecular weight changes of the treated PEG-PU were analyzed by GPC [11], and the surface morphology of the recovered PEG-PU and foams was examined by scanning electron microscopy (SEM) [11].
3. Results and discussion
3.1. Structural analysis of esterase Aes72
Phylogenetic analysis indicated that Aes72 belongs to lipase family IV, specifically the hormone-sensitive lipase (HSL) group (Figs. S10(a) and (b) in Appendix A). We determined the crystal structure of the recombinant Aes72 in its ligand-free state at a resolution of 1.80 Å (PDB: 9IZT) (Fig. 1(a), Table S1 in Appendix A). The crystal structure exhibits a typical α/β hydrolase fold, characterized by eight β-strands flanked by nine α-helices on either side (Fig. 1(a)). Additionally, it contains a conserved catalytic triad consisting of a nucleophile (S147), a histidine (H271), and a catalytic acid (E241) (Fig. 1). Three loops (L1-L3) form part of the active site in the Aes72 structure, with the HSL family motif (G145-G149: GDSAG) positioned at the center. Loop 1 contains several highly conserved residues, including catalytic S147 (Fig. 1(a)). The catalytic triad residues E241 and H271 are located in loops 2 (G238-L244) and 3 (W266-F273), respectively (Fig. 1(b)). Aes72 possesses a unique lid structure (M1-V47) known as the "cap" domain, commonly found in lipases, which consists of two helices connected by a loop. This lid structure has been reported to play an important role in maintaining enzyme stability, activity, and catalytic specificity [46].
The PDB database includes the HSL esterase E40 (PDB: 4XVC), which shares 47.3% sequence identity with Aes72 and was isolated from a metagenomic library of marine sediment. E40 exhibits the same overall fold as Aes72 (root mean square deviation (RMSD): 0.632 Å for 245 of 291 residues) and a similar active site structure (catalytic center RMSD: 0.131 Å), including the same catalytic triad and oxyanion hole (Fig. S1) [47]. The conserved HGG motif, characteristic of the microbial HSL family, is vital for the formation of the oxyanion hole and is closely linked to the catalytic triad. This oxyanion hole is critical for catalysis, as it provides direct stabilization to the tetrahedral intermediate generated during the reaction [48].
3.2. Kinetic and mechanistic analyses of Aes72-catalyzed hydrolysis of diverse substrates
Given that the ester hydrolase Aes72 has been reported to exhibit promiscuous activity on carbamate bonds, we further investigated the Michaelis-Menten kinetics of its hydrolysis of various model compounds (pNPB, BNC, and pNAA) to better understand its catalytic promiscuity (Fig. 2). As shown in Table 1, Aes72 hydrolyzes pNPB with a kcat/Km value of 1.3 × 10-2 μM-1∙s-1, approximately 295.5 times higher than for pNAA (4.4 × 10-5 μM-1∙s-1) and 34.2 times higher than for BNC (3.8 × 10-4 μM-1∙s-1). Aes72 showed the highest catalytic efficiency for ester bonds compared to urethane and amide bonds, consistent with its role as an intrinsic esterase.
To elucidate the catalytic mechanism of (poly)urethane bond hydrolysis by Aes72, we synthesized a PU model substrate (BMC) composed of a central MDA molecule connected to two BDO molecules through carbamate bonds (Fig. 2, Fig. S11 in Appendix A). In addition to serving as a model compound for the experimental studies of Aes72 and its variants, BMC was used as a model substrate in computational simulations. Molecular docking was carried out to determine the binding orientation of BMC within the active site of Aes72 (Fig. 3(a)). As shown in Fig. 3(b), the backbone NH groups of G79, A80, and A148 within the oxyanion hole form hydrogen bonds with the carbonyl oxygen of BMC. The structure of the enzyme-substrate complex reveals optimal contacts for initiating the chemical reaction, with the distances between S147@O1 and BMC@C1 and between H271@N1 and S147@H1 measuring 3.95 and 1.99 Å, respectively. Notably, S147@O1 refers to the atom O1 of Ser147, and all abbreviations used in this study are defined consistently. Changes in the key active-site distances are provided in Figs. S11 and S12 in Appendix A. This BMC-bound structural complex served as a starting point for molecular dynamics and QM/MM calculations. Aes72-catalyzed carbamate hydrolysis proceeds through acylation (steps i and ii) and deacylation (steps iii and iv) stages (Fig. 3(b)). In step i, the catalytic residue H271 removes a proton from S147, activating it to attack the carbonyl carbon of the urethane bond and forming a tetrahedral intermediate (TI1). In step ii, a stable acetyl-enzyme intermediate (AI) is generated upon ester or amide bond cleavage. In step iii, H271-mediated deprotonation of a water molecule in the active site allows its oxygen to subsequently attack the carbonyl carbon of the acylated intermediate. This interaction gives rise to a tetrahedral intermediate (TI2). In step iv, the product is generated, and the enzyme reverts to its inactive state. The simulation results demonstrate that Aes72-catalyzed urethane hydrolysis can proceed via two distinct pathways, with either the ester or amide moiety being cleaved first. The energy barriers for the four key steps in ester bond cleavage are 31.4, 24.4, 23.8, and 19.9 kcal∙mol-1 (1 kcal = 4184 J), whereas those for amide bond cleavage are 31.4, 25.8, 26.1, and 24.2 kcal∙mol-1, respectively (Figs. 3(c) and (d)). Step i is the rate-limiting step for both pathways, with the cleavage of the ester group being favored over the amide group. These results align with the experimental data shown in Table 1, where the kcat/Km value for pNPB hydrolysis catalyzed by Aes72 (1.3 × 10-2 μM-1∙s-1) is substantially higher than that for pNAA (4.4 × 10-5 μM-1∙s-1).
3.3. Structure-guided engineering of Aes72 for enhanced urethane hydrolytic activity
Next, we aim to enhance the hydrolytic activity of urethane bonds by Aes72 toward carbamate substrates by structure-based semi-rational enzyme design. Based on the active site architecture of Aes72, bulky residues that may interact with the docked BMC substrate were identified and replaced with alanine to produce the following mutants: F83A, D146A, W196A, F276A, E29A, and H89A (Fig. 4). Because BMC resembles a short PU segment, these substitutions were expected to enlarge the Aes72 binding pocket, increasing its catalytic activity toward BMC and the bulkier PU substrate. Other residue substitutions, such as S169A, H224Y, and L141I, were generated using a consensus design strategy (Fig. S13 in Appendix A).
3.4. Characterization of Aes72 variants using the short-chain carbamate substrate BMC
BMC was used as the initial substrate to assess the carbamate hydrolytic activity of Aes72 and its variants. The hydrolysis test was conducted at 50 °C and pH 7.0. BMC was efficiently hydrolyzed by Aes72, yielding MDA as the reaction product (Fig. 2). Compared to WT-Aes72, the variants F276A ((17.39 ± 0.80) mU∙mg-1), H224Y ((13.59 ± 0.65) mU∙mg-1), E29A ((12.52 ± 0.33) mU∙mg-1), L141I ((19.43 ± 1.71) mU∙mg-1), and W196A ((13.44 ± 0.49) mU∙mg-1) showed 1.74-, 1.36-, 1.26-, 1.95-, and 1.35-fold improvements in enzymatic activity, respectively (Fig. 5(a)). Next, double mutants were generated by combining the beneficial mutations (H224Y, E29A, L141I, and W196A) with the superior single mutant F267A. Subsequent BMC hydrolysis tests showed that three of these double mutants, F276A/E29A, F276A/L141I, and F276A/H224Y, exhibited lower activity than F276A, although all remained more active than WT-Aes72. F276A/L141I was the only double variant with improved specific activity ((20.53 ± 1.40) mU∙mg-1), representing a 1.18-fold and 2.06-fold increase over F276A and WT-Aes72, respectively (Fig. 5(a)). We then compared the catalytic efficiencies of WT-Aes72 and the F276A/L141I variant for BMC hydrolysis using Michaelis-Menten kinetics. F276A/L141I exhibited a lower kcat/Km value (2.1 × 10-4 μM-1∙s-1) than WT-Aes72 (5.2 × 10-4 μM-1∙s-1) (Fig. 5(b)), despite the known limitations of using kcat/Km values to predict enzyme variant efficacy [49]. Therefore, we adopted an alternative method proposed by Ceccarelli et al. [45] to estimate the true efficiencies of enzyme-catalyzed reactions, circumventing the limitations of kcat/Km values. The Ef of an enzyme, defined as the ratio of the catalyzed reaction rate to the maximum theoretical rate, is a key parameter for comparing different enzyme variants on the same substrate across a broad range of substrate concentrations. Although F276A/L141I has a lower kcat/Km than WT-Aes72, the Ef plot as a function of BMC concentration suggests that its catalytic efficiency is lower than WT-Aes72 only at low substrate concentrations (< 50.7 μM). At higher BMC concentrations, F276A/L141I catalyzes the hydrolysis of BMC more efficiently than WT-Aes72 (Fig. 5(c)). This emphasizes the benefit of using this enzyme variant, as high substrate loadings are commonly used in PU degradation experiments.
3.5. Degradation of various PEG-PU materials by Aes72 and the F276A/L141I variant
Subsequently, we investigated the hydrolytic activity of Aes72 and its variants toward PU polymers. We synthesized a polyether-based thermoplastic PU (PEG-PU) with a weight-average molecular weight (Mw) of approximately 20 000 using PEG (Mw = 2000) and MDI as the building blocks (Figs. 2(c) and S8). The self-synthesized PEG-PU was mechanically pulverized into powders without chemical pretreatment and subjected to enzymatic depolymerization using either WT-Aes72 or mutant F276A/L141I at a substrate concentration of 2 g∙L-1 and 50 °C in a 5 mL reaction vial. WT-Aes72 achieved an MDA yield of 19.4% ± 1.5% from the PEG-PU powders over 48 h, corresponding to an MDA production rate of 1.89 μmol∙gPU-1∙h-1 (gPU: gram of PU). In contrast, F276A/L141I achieved an MDA yield of 18.9% ± 1.4% after 24 h, equivalent to an MDA production rate of 3.71 μmol∙gPU-1∙h-1 (Fig. 6(a)). This rate was 1.96 times higher than that of WT-Aes72, confirming the enhanced hydrolytic activity of the variant toward BMC. In addition, GPC was used to analyze the molecular weights of PEG-PU samples treated with WT-Aes72 and F276A/L141I after 12 h (Fig. 6(b)). Compared with the negative control (without enzymatic treatment), the Mw and number-average molecular weight (Mn) values of PEG-PU treated with F276A/L141I decreased from 23 596 and 12 510 to 16 159 and 9 611, corresponding to reductions of 31.5% and 23.2%, respectively. Accordingly, the Mw and Mn values of PEG-PU treated with WT-Aes72 decreased by 18.1% and 6.4%, respectively. These results indicate that treatment with F276A/L141I led to more pronounced chain scission and generated a greater proportion of low-molecular-weight fragments compared to WT-Aes72. The catalytic efficiency of Aes72 was also compared with those of other reported PU urethanases (UMG-SP-1 and UMG-SP-2). After 12 h of incubation, UMG-SP-1 and UMG-SP-2 achieved MDA yields of 69.4% and 72.3%, which were 3.9- and 4.1-fold higher than that of mutant F276A/L141I (17.7%), respectively (Fig. S14 in Appendix A). Although Aes72 exhibits lower catalytic efficiency in PEG-PU hydrolysis than amidases UMG-SP-1 and UMG-SP-2, it represents the first reported promiscuous esterase capable of catalyzing this reaction. Remarkably, Aes72 demonstrated markedly superior thermal stability than the UMG-SP enzymes, demonstrating its potential for synergistic applications with other thermally stable plastic-hydrolyzing enzymes in mixed plastic degradation. In contrast, UMG-SP-2 is restricted to tandem applications due to its thermal lability [19]. Next, we compared the depolymerization capabilities of the two enzymes toward commercially available polyether-based PU foams. After a 12 h reaction at 50 °C, 39.8 μM MDA was released through F276A/L141I-catalyzed hydrolysis, 1.6 times higher than that produced by WT-Aes72 (25.5 μM) (Fig. 6(c)). We also collected the two types of PU after enzymatic degradation and measured their weight loss. As expected, the weights of the PU foams remained relatively constant following enzymatic treatment. In contrast, the weight loss of PEG-PU treated with WT-Aes72 and mutant F276A/L141I reached 49.7% and 60.0%, respectively (Fig. 6(d)). Nevertheless, the corresponding MDA yields determined from the biocatalytic degradation of PEG-PU were less than 20%. This discrepancy may arise from the generation of numerous oligomeric degradation intermediates during PEG-PU hydrolysis. In addition to the HPLC analysis of the monomer MDA, mass spectroscopy (MS) was employed to identify other degradation intermediates. Three intermediates were identified as MDA derivatives, each terminated with either 2, 13, or 29 ethylene glycol units, based on molecular ions at m/z 353.16 (M + Na)+, 837.45 (M + Na)+, and 1519.67 (M + Na)+, respectively (Fig. S15 in Appendix A). Although these intermediates are no longer covalently linked to the residual polymer substrates, their size remains below the detection limit of HPLC, potentially due to solubility issues or other unidentified factors. Additionally, the degradation of both PU materials was confirmed through morphological observation using SEM. For SEM observation, PEG-PU films were prepared by dissolving the polymer in dichloromethane and casting the solution into a container, allowing the solvent to evaporate completely to form a film. After 12 h of enzymatic hydrolysis with mutant F276A/L141I, the PU film surface exhibited significant irregularities and decreased thickness (Fig. 6(e)). SEM images of PEG-PU treated with mutant F276A/L141I revealed numerous cracks and increased surface roughness (Fig. 6(e)). Similarly, PU foam treated with mutant F276A/L141I exhibited cracks, thereby confirming enzymatic degradation (Fig. 6(f)).
Degradation tests on thermoplastic and thermoset PU substrates revealed that Aes72 and its variant induced significant erosion effects on thermoplastic PU, whereas their degradation performance on thermoset PU was comparatively inadequate. The highly cross-linked polymer microstructure of thermoset PU effectively restricts the access of both water molecules and enzyme catalysts to hydrolyzable carbamate bonds, making it highly resistant to biodegradation [50,51]. Moreover, the chemical composition of commercially available PU foam is complex, containing additives such as blowing agents, flame retardants, and ultraviolet stabilizers, which may interfere with the biocatalysts and hinder the biodegradation of PU foam [52]. PU foam accounts for the largest share of PU products, representing 68% of global PU production [53]. Consequently, it is essential to develop efficient, environmentally sustainable, and economically feasible PU waste pretreatment techniques, while further enhancing the catalytic efficacy of Aes72 and other recently identified potent PU hydrolases through protein engineering, to facilitate the effective biocatalytic degradation of thermoset PU foam for practical industrial plastic recycling applications [8,19,20,22,23].
4. Conclusions
In this study, we characterized and engineered the promiscuous esterase Aes72, which exhibited amidase and carbamate hydrolase activities on diverse substrates, including oligomeric carbamates such as BMC, which mimics PU segmented, and realistic polyether-based PU. The crystal structure of Aes72, resolved at 1.8 Å, in conjunction with computational QM/MM simulations, enabled a deeper understanding of its catalytic mechanisms and guided the design of variants with potentially enhanced activities. Four of the nine engineered single variants demonstrated increased catalytic activity on BMC, providing the basis for developing the superior double mutant F276A/L141I, which markedly outperformed WT-Aes72 in hydrolyzing BMC and various PU substrates. Despite these encouraging results and the successful enhancement of Aes72 hydrolytic activity via semi-rational redesign, the overall degradation performance of commercially available cross-linked polyether-based PU foam remains limited, resulting in only minimal surface modification. These findings highlight the need to further improve PU waste pretreatment strategies for the efficient biocatalytic recycling of PU wastes on an industrial scale. Simultaneously, insights into the catalytic mechanism of carbamate bonds in PU, gained from structural and simulation studies of the promiscuous esterase Aes72, can guide future efforts in the design of more potent PU-depolymerizing enzymes.
Plastics—the fast facts 2024 [Internet]. Brussels: Plastics Europe; 2024 [cited 2025 Jan 20]. Available from: https://plasticseurope.org/knowledge-hub/plastics-the-fast-facts-2024/.
[2]
LiM, ZhangS. Coupling waste plastic upgrading and CO2 photoreduction to high-value chemicals by a binuclear Re-Ru heterogeneous catalyst. ACS Catal2024; 14(9):6717-27.
[3]
ShiY, DiaoX, JiN, DingH, YaZ, XuD, et al. Advances and challenges for catalytic recycling and upgrading of real-world mixed plastic waste. ACS Catal2025; 15(2):841-68.
[4]
YaZ, TangL, XuD, WangH, ZhangS. Photoreforming of waste plastic by B-doped carbon nitride nanotube: atomic-level modulation and mechanism insights. AIChE J2025; 71(5):e18740.
[5]
ZiaKM, BhattiHN, AhmadBI. Methods for polyurethane and polyurethane composites, recycling and recovery: a review. React Funct Polym2007; 67 (8):675-92.
[6]
LiuJ, HeJ, XueR, XuB, QianX, XinF, et al. Biodegradation and up-cycling of polyurethanes: progress, challenges, and prospects. Biotechnol Adv2021; 48:107730.
[7]
TournierV, TophamCM, GillesA, DavidB, FolgoasC, Moya-LeclairE, et al. An engineered PET depolymerase to break down and recycle plastic bottles. Nature2020; 580(7802):216-9.
[8]
WeiR, WeberG, BlankLM, BornscheuerUT. Process insights for harnessing biotechnology for plastic depolymerization. Nat Chem Eng2025; 2(2):110-7.
[9]
MicanJ, JaradatDMM, LiuW, WeberG, MazurenkoS, BornscheuerUT, et al. Exploring new galaxies: perspectives on the discovery of novel PET-degrading enzymes. Appl Catal B2024; 342:123404.
[10]
SeoH, HongH, ParkJ, LeeSH, KiD, RyuA, et al. Landscape profiling of PET depolymerases using a natural sequence cluster framework. Science2025; 387 (6729):eadp5637.
[11]
XueR, QiuC, ZhouX, ChengY, ZhangZ, ZhangY, et al. Enzymatic upcycling of PET waste to calcium terephthalate for battery anodes. Angew Chem Int Ed Engl2024; 63(1):e202313633.
[12]
ArnalG, AngladeJ, GavaldaS, TournierV, ChabotN, BornscheuerUT, et al. Assessment of four engineered PET degrading enzymes considering large-scale industrial applications. ACS Catal2023; 13(20):13156-66.
[13]
WeiR, von HaugwitzG, PfaffL, MicanJ, BadenhorstCPS, LiuW, et al. Mechanism-based design of efficient PET hydrolases. ACS Catal2022; 12 (6):3382-96.
[14]
LiuJ, XinK, ZhangT, WenY, LiD, WeiR, et al. Identification and characterization of a fungal cutinase-like enzyme CpCut1 from Cladosporium sp. P7 for polyurethane degradation. Appl Environ Microbiol2024; 90(4). e01477-23.
[15]
LiuJ, ZengQ, LeiH, XinK, XuA, WeiR, et al. Biodegradation of polyester polyurethane by Cladosporium sp. P7: evaluating its degradation capacity and metabolic pathways. J Hazard Mater2023; 448:130776.
[16]
Raczyń skaA, GóraA, AndréI. An overview on polyurethane-degrading enzymes. Biotechnol Adv2024; 77:108439.
[17]
JiangZ, ChenX, XueH, LiZ, LeiJ, YuM, et al. Novel polyurethane-degrading cutinase BaCut 1 from Blastobotrys sp. G-9 with potential role in plastic bio-recycling. J Hazard Mater2024; 472:134493.
[18]
ZhangL, CaoK, LiuH, WangY, ZhangB, HanH, et al. Discovery of a polyester polyurethane-degrading bacterium from a coastal mudflat and identification of its degrading enzyme. J Hazard Mater2025; 483:136659.
[19]
BransonY, LiuJ, SchmidtL, GriebelJ, PragerA, StielerL, et al. One-pot depolymerization of mixed plastics using a dual enzyme system. ChemSusChem2025; 18(9):e202402416.
[20]
BayerT, PalmGJ, BerndtL, MeinertH, BransonY, SchmidtL, et al. Structural elucidation of a metagenomic urethanase and its engineering towards enhanced hydrolysis profiles. Angew Chem Int Ed Engl2024; 63(38): e202404492.
[21]
PaivaP, TeixeiraLMC, WeiR, LiuW, WeberG, MorthJP, et al. Unveiling the enzymatic pathway of UMG-SP2 urethanase: insights into polyurethane degradation at the atomic level. Chem Sci2025; 16(5):2437-52.
[22]
RotilioL, BayerT, MeinertH, TeixeiraLMC, JohansenMB, SommerfeldtA, et al. Structural and functional characterization of an amidase targeting a polyurethane for sustainable recycling. Angew Chem Int Ed Engl2025; 64(7): e202419535.
[23]
LiZ, HanX, CongL, SinghP, PaivaP, BransonY, et al. Structure-guided engineering of a versatile urethanase improves its polyurethane depolymerization activity. Adv Sci2025; 12(13):2416019.
[24]
BittnerN, NefzgerH, BehnkenG, JägerG, BehnkenS, ReiskyL; Covestro Deutschland AG, assignee. Novel urethanases for the enzymatic degradation of polyurethanes. United States patent US20210269834A1. 2021 Sep 2.
[25]
OtwinowskiZ, MinorW. Processing of X-ray diffraction data collected in oscillation mode. In: Carter CW, editor. Methods in enzymology. New York City: Academic Press; 1997. p. 307-26.
WinnMD, BallardCC, CowtanKD, DodsonEJ, EmsleyP, EvansPR, et al. Overview of the CCP 4 suite and current developments. Biol Crystallogr2011; 67(Pt 4):235-42.
[28]
MurshudovGN, SkubákP, LebedevAA, PannuNS, SteinerRA, NichollsRA, et al. REFMAC 5 for the refinement of macromolecular crystal structures. Biol Crystallogr2011; 67(Pt 4):355-67.
BrüngerAT, AdamsPD, CloreGM, DeLanoWL, GrosP, Grosse-KunstleveRW, et al. Crystallography & NMR system: a new software suite for macromolecular structure determination. Acta Cryst1998; 54(D5):905-21.
[31]
EberhardtJ, Santos-MartinsD, TillackAF, ForliS. AutoDock Vina 1.2.0: new docking methods, expanded force field, and python bindings. J Chem Inf Model2021; 61(8):3891-8.
[32]
MorrisGM, HueyR, LindstromW, SannerMF, BelewRK, GoodsellDS, et al. AutoDock4 and AutoDockTools4: automated docking with selective receptor flexibility. J Comput Chem2009; 30(16):2785-91.
[33]
CvitkovicJP, PauplisCD, KaminskiGA. PKA17—a coarse-grain grid-based methodology and web-based software for predicting protein pKa shifts. J Comput Chem2019; 40(18):1718-26.
ZhaoY, TruhlarDG. The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics, noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four M06-class functionals and 12 other functionals. Theor Chem Acc2008; 120(1-3):215-41.
[36]
TrottO, OlsonAJ. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem2010; 31(2):455-61.
[37]
ZoeteV, CuendetMA, GrosdidierA, MichielinO. SwissParam: a fast force field generation tool for small organic molecules. J Comput Chem2011; 32 (11):2359-68.
Di BisceglieF, QuartinelloF, VielnascherR, GuebitzGM, PellisA. Cutinase-catalyzed polyester-polyurethane degradation: elucidation of the hydrolysis mechanism. Polymers2022; 14(3):411.
[40]
HaoW, LiY, ShanQ, HanT, LiW, HeS, et al. Characterization of human S-adenosyl-homocysteine hydrolase in vitro and identification of its potential inhibitors. J Enzyme Inhib Med Chem2017; 32(1):1209-15.
[41]
TerholsenH, MyrtollariK, LarvaM, MöllerC, TadenA, KouristR, et al. Spectrophotometric and fluorimetric high-throughput assays for phenolic acid decarboxylase. ChemBioChem2023; 24(16):e202300207.
[42]
ChenC, GuanBH, GengQ, ZhengYC, ChenQ, PanJ, et al. Enhanced thermostability of Candida ketoreductase by computation-based cross-regional combinatorial mutagenesis. ACS Catal2023; 13(11):7407-16.
[43]
WangF, HouY, ZhouJ, LiZ, HuangY, CuiZ. Purification of an amide hydrolase DamH from Delftia sp. T3-6 and its gene cloning, expression, and biochemical characterization. Appl Microbiol Biotechnol2014; 98(17):7491-9.
[44]
AlberyWJ, KnowlesJR. Evolution of enzyme function and the development of catalytic efficiency. Biochemistry1976; 15(25):5631-40.
[45]
CeccarelliEA, CarrilloN, RoveriOA. Efficiency function for comparing catalytic competence. Trends Biotechnol2008; 26(3):117-8.
[46]
MandrichL, MeroneL, PezzulloM, CipollaL, NicotraF, RossiM, et al. Role of the N terminus in enzyme activity, stability and specificity in thermophilic esterases belonging to the HSL family. J Mol Biol2005; 345(3):501-12.
[47]
LiPY, ChenXL, JiP, LiCY, WangP, ZhangY, et al. Interdomain hydrophobic interactions modulate the thermostability of microbial esterases from the hormone-sensitive lipase family. J Biol Chem2015; 290(17):11188-98.
[48]
MandrichL, MenchiseV, AlterioV, De SimoneG, PedoneC, RossiM, et al. Functional and structural features of the oxyanion hole in a thermophilic esterase from Alicyclobacillus acidocaldarius. Proteins2008; 71(4):1721-31.
[49]
EisenthalR, DansonMJ, HoughDW. Catalytic efficiency and kcat/KM: a useful comparator? Trends Biotechnol 2007; 25(6):247-9.
[50]
PetrovićZS, XuY, MilićJ, GlennG, KlamczynskiA. Biodegradation of thermoplastic polyurethanes from vegetable oils. J Polym Environ2010; 18 (2):94-7.
[51]
HamidSH, AminMB, MaadhahAG, editors. Handbook of polymer degradation. New York City: Marcel Dekker Inc.; 1992.
[52]
SkleničkováK, AbbrentS, HaleckýM, BenešH. Biodegradability and ecotoxicity of polyurethane foams: a review. Crit Rev Environ Sci Technol2022; 52(2):157-202.
[53]
MagninA, PolletE, PhalipV, AvérousL. Evaluation of biological degradation of polyurethanes. Biotechnol Adv2020; 39:107457.
PDF
(2155KB)
