Oda (2021) provided a great contribution to the development of Cut190. In 2014, a putative cutinase gene (cut190, from the thermophile Saccharomonospora viridis AHK190) was cloned and successfully expressed (Kawai et al., 2014). The mutant Cut190 (S226P/R228S) was shown to display optimal activity at 65–75°C, pH 6.5–8.0, and could remain stable at temperatures up to 65°C for 24 h; however, 60% of activity was lost after incubation for 1 h at 70°C. In particular, Ca²⁺ could clearly change the tertiary structure of Cut190, and the thermostability and activity could be greatly enhanced by high concentrations of calcium ions. To illustrate the structural basis for the Ca²⁺-enhanced thermostability, Takuya et al. reported the crystal structures of Cut190^S226P in the Ca²⁺-bound and free states (Figure 1(1,2)) (Miyakawa et al., 2015). It mainly adopts a classical α/β hydrolase fold and contains 9 β-strands, 6 α-helices and a disulfide bond linking C287 and C302. The active-site Ser-His-Asp catalytic triad (S176, D222, and H254) was located on the surface without the lid structure. Ca²⁺ was located at the β1-β2 loop (Figure 1(3,4)) formed by residues from S76 to F81 (Figure 1(4)). Furthermore, the bound Ca²⁺ was coordinated by S76, A78, and F81 via their main-chain carbonyl groups. Once Ca²⁺ bound to Cut190^S226P, it induced significant conformational changes, which mainly occurred in the following loops: the β1-β2 loop, the β3-α2 loop (F97-A121), and the β4-α3 loop (I126-E155). Upon Ca²⁺ binding, it will induce N133 to move toward the Ca²⁺ binding site (the β1-β2 loop, ~11.5Å), consequently enabling the movement of Q110 toward D131 and forming an additional hydrogen bond. The β3-α2 loop harbors the active-site sealing residue F106. Although Ca²⁺ binding induced the β3-α2 loop to the flexible open state, it might be helpful for substrate binding.

Based on these findings, attention was focused on insights into the structure–activity relationship of substrate-Cut190 and rational protein engineering. Molecular modeling (ghecom: pocket finding; fkcombu: docking; the myPresto program package: energy minimization) was applied to investigate the reaction mechanism and improve the enzyme activity by Kawabata et al. (2017). The results showed that W201, F106, H175, M177 and F255 are important for substrate binding and enzymatic activities, especially for the aromatic polyester substrate. Numoto et al. (2018) further analyzed the substrate recognition mechanism of Cut190 with the crystal structure of the inactive Cut190 mutant (Cut190^S176A). In this study, it was found that four Ca²⁺ ions were bound to Cut190^S176A (Figure 2), which is different from that observed in a previous study (Miyakawa et al., 2015). From Figure 2, three of them are bound to the surface of the enzyme at site 1, site 2 and site 3. One more bound Ca²⁺ is located at the interface in site 4, which might be caused by the crystal packing that stabilizes the crystalline state of the protein. Structure alignment shows that site 1 is the same Ca²⁺ binding site, and Ca²⁺ binding displays a similar effect on the protein as that observed in the structure of Cut190^S226P (Miyakawa et al., 2015) (Figure 2(1,2)). At site 2, G80 moved slightly away from the Ca²⁺-binding site upon Ca²⁺ binding, and Ca²⁺-binding showed little effect on site 3 (Figure 2(3,4)). Researchers reported that functions for the different Ca²⁺ are believed to be different, in which the Ca²⁺ at sites 1 and 3 might be responsible for the regulation of enzymatic activity, and Ca²⁺ at sites 2 and 3 can accommodate thermostability. It was further verified that Mn²⁺ and Mg²⁺ probably bind to site 2 with high affinity (responsible for stability improvement) and to sites 1 and 3 with low affinity (responsible for activity adjustment) (Senga et al., 2019). These results indicate that it might be possible to regulate the catalytic activity and thermostability by rational engineering of critical residues around site 1, site 2 and site 3.

Structure comparison of Cut190^S176A bound with different substrates (Figure 3) clearly shows that two distinct states occur in the catalytic process, the prereaction state (engaged form, PDB ID: 5ZRR) and postreaction state (open form, PDB ID: 5ZRS). Interestingly, the structure of the postreaction state is almost identical to the structure bound with only one Ca²⁺. This might indicate that Ca²⁺ could greatly accelerate from the closed and the engaged form to the open form, which could also be illustrated by the weak Ca²⁺ binding interaction in site 1.

Recently, it was reported that mutating one Ca²⁺-binding sites of Cut190^D250C/E296C would increase its thermostability while causing little effect on its polyesterase activity (Emori et al., 2021). First, a disulfide bond would be introduced at the interface through the mutation of D250C/E296C (in site 2) without changing the overall structures (Figure 4(1)), which contributes greatly to the increase in thermostability. Furthermore, due to the weak binding of Ca²⁺ in site 1, the catalytic activity was unaffected when the conformational change was altered from the engaged state to the open state. The Ca²⁺ not bound to sites 2 and 3 is mainly caused by mutations in site 2 (D250C and E296C) and the residue nearby site 3 (N202H), which also reveals that Ca²⁺ binding has little effect on its overall structure (Figure 4(2)). Rapid release of the degrading products is a critical process for efficient PET-degrading, and this process has been already determined in the open form of Cut190 through molecular dynamics simulations (Numoto et al., 2018). One possibility is that the final products might be transported back to the engaged form and be tightly regulated. To deal with this issue, recently a novel state, the ejecting form at site 1 (Figure 4(3)) in the Cut190*-mediated catalytic process was determined with three C-terminal residues deleted in complex with metal ions (Senga et al., 2021). The closed-like and ligand-incompatible form displays a special conformation of F81 compared with the previous study (Numoto et al., 2018), in which either F77 and F81 all exhibit the inward conformations despite in the absence of bound Ca²⁺. This form is determined to allow the products to irreversibly dissociate, which should be properly controlled in the presence of an appropriate concentration of Ca²⁺. This phenomenon could also be illustrated by the finding that weak binding affinity of Ca²⁺ plays a critical role in the function of Cut190 (Senga et al., 2019). Once the products are irreversibly released, and the next cycle of reaction would be efficiently performed. In addition, deletion of three C-terminal residues contributes to the increased stability (T_m, 85°C), which is largely the consequence of the decreased fluctuation.

Recently, Cut190 (S226P/R228S) was further engineered (Kawai et al., 2022). Using the best performing Cut190 variant (L136F/Q138A/S226P/R228S/D250CE296C/Q123H/N202H/K305del/L306del/N307del) and amorphous PET powders, more than 90 mM degradation products were obtained in 3 days and approximately 80 mM in 1 day.

Two open reading frames encoding cutinases named Tfu_0882 (TfCut1) and Tfu_0883 (TfCut2) were identified in the thermophilic bacterium Thermobifida fusca (Chen et al., 2008). Although they share a high sequence identity of 93% in the amino acid sequence (261 aa), TfCut2 displayed much higher (~ two times higher) degradation activity than that of TfCut1 (Chen et al., 2010). TfCut2 displays 57.4% similarity in amino acid sequence with LCC. The homology model of Tfu_0883 displays a typical α/β-hydrolase fold with a canonical triad S170-H248-D216, a preformed oxyanion hole and a different nucleophilic serine exposed to the solvent. Although they show similar catalytic properties to fungal cutinases, TfCut1 and TfCut2 are a different subfamily of cutinases due to their sequence and structural differences. For example, they can hydrolyse insoluble triglycerides and soluble esters (e.g., pNPB) in addition to cutin. They also display superior thermostability compared with that of the F. solani pisi cutinase. A study showed that the higher linear degradation rates observed in the first 72 h of incubation are due to the high hydrolysis susceptibility of the mobile amorphous fraction (MAF, with a chain mobility similar to the purely amorphous polymer) of PET; however, a transition of MAF to the rigid amorphous fraction (RAF, with a lower chain mobility probably caused by its presence within the intraspherulitic or interlamellar regions in close vicinity to the crystalline domains) during further incubation resulted in drastically decreased degradation rates (Wei et al., 2019). The RAF could be cleaved only in an endo-type manner at a much lower rate.

To investigate the catalytic mechanisms of the bacteria-derived cutinases, several crystal structures of TfCut2 have been reported [e.g., PDB ID: 4CG2 (Roth et al., 2014), Figure 8, 5ZOA, 7QJR, 7XTT]. The structure of TfCut2 forms a classical α/β-hydrolase fold, contains conserved GXSXG motif, and contains a conserved catalytic triad of S-H-D residues is in a crevice on the surface. In the enzyme-inhibitor complex structure (Figure 8(2)), it was found that covalent modification of S130 would lead to the tighter interactions formed with the main chain of M131 and Y60 mainly through H-bonds interactions (Figure 8(3)). The phenyl ring is further stabilized mainly by hydrophobic interactions with I178 and Y60 (Figure 8(2)).

Efforts have been made to further improve the catalytic performance of TfCut2 through directed evolution and rational mutagenesis. The critical residues associated with substrate recognition were engineered (Wei et al., 2016). First, critical residues around the active site close to (without direct interactions with) the docked 2PET were identified and genetically engineered by performing structure alignment with LCC and TfCut2. Four residues (G62, T63, I178, and I213) were selected for further protein engineering. Among the single mutants, G62A was determined to show markedly increased hydrolytic activity. The improved catalytic performance of G62A likely resulted from a reduction in product inhibition due to MHET rather than the altered surface hydrophobicity or the enlarged substrate binding pocket. Although the double variant G62A/I178V and the triple variant G62A/I178V/I213S showed much higher hydrolytic activity than the wild-type TfCut2, they all displayed slightly lower degrading activity than the single mutant G62A. The single mutant I213S showed the highest thermostability, followed by G62A and G62A/I213S. After 24 h at 65°C, G62A and G62A/I213S caused a 20.5% weight loss of PET films.

In a similar study, the functions of critical residues around the substrate binding site were investigated, which were thought to facilitate the accessibility of the enzyme’s active site (Dong et al., 2020). Eleven bulky residues were selected to increase the substrate binding efficiency by site-directed mutation to Ala, which are located around the catalytic triad, the oxygen hole and the substrate binding region. The results showed that Y60A, T207A, F209A, L90A and I213A exhibited increased pNPB hydrolysing activity, especially the cutin hydrolysing activities of L90A, I213A, and L90A/I213A, which were increased by 5-, 2.4-, and 3.2-fold, respectively. MD simulations revealed that the higher flexibility of the substrate binding site of mutants L90A and I213A might lead to increased hydrolysing activities toward cutin. However, a subsequent study showed that the L90A mutant showed ~11.2-fold reduced degradation activity on commercially sourced PET film, but the L90F mutation displayed considerably increased degradation (~1.5-fold increased) activity (Mrigwani et al., 2023). This might be caused by the more efficient binding of PET resulting from the replacement of the hydrophobic phenylalanine residue.

Based on the abovementioned findings, molecular rational engineering of TfCut2 was performed by docking and MD simulations to modify the space at the active site and streamline the hydrophobicity at the active site or other interacting regions (Mrigwani et al., 2023). The results showed that the selected residues H129W, W155F, H184A, and H184S for engineering the PET-binding cleft exhibited little positive effect on the catalytic activities. Residues (A173C) that were identified for the insertion of thiols/disulfides (A173C/A210C, and A173C/ A206C) might distort the PET-binding site, causing a complete loss of activity against solid PET. Interestingly, residues (L90) distal to the active site could adjust its catalytic activity, especially the L90F mutation, which could considerably increase TfCut2’s PET-degrading activity (~1.5-fold). F209I, F249A and F249R were constructed to modify the surface hydrophobicity proximal/distal to the active site. The results showed that the single mutant (F209I) caused an ~4.4-fold decrease in activity; however, F249R led to an ~1.27-fold increase in activity against solid PET. The double mutants displayed a remarkable increase in degradation activity (G62A/F209I with an ~6.4-fold improvement and G62A/F249R with an ~7.15-fold increase); however, their thermostability was reduced. These results also revealed that TfCut2 possesses high thermal stability, high kinetic-thermal stability, and high thermodynamic stability.

Based on the findings of the double mutation (DM) strategy (Chen et al., 2021), engineered TfCut was investigated for efficient biodegradation of poly(butylene adipate-co-terephthalate) (PBAT) (Yang et al., 2023). The results showed that the mutant (TfCut-DM) is a highly potent catalyst that can completely decompose PBAT films in 36 h and release 3-fold more TPA (the major product) than TfCut within 12 h. This strategy was also determined to be appropriate for BurPL and TcCut. The overall complex structure of the MHET-bound TfCut complex structure is almost identical to that of TfCut-DM (Figure 9(1)), and the DM residues (S224 and I228) are located below the catalytic triad. The ethylene glycol motif of MHET bound points outside the enzyme, and the aliphatic part stretches to fill a cleft across the protein surface (Figure 9(3)), which is lined by several critical residues, such as M171, W195, Y100, I218, and T223. Several water molecules are also shown to be associated with the binding of MHET through H-bond interactions. Interestingly, the side chain of W195 in TfCut-DM exhibits a B-type conformation [Figure 9(2) in purple]; however, wild-type TfCut harbors a TPA-binding Trp in the C-type conformation [Figure 9(2) in brown]. The DM mutation and the B-type W195 lead to a more open and flatter TPA-binding pocket (increased flexibility), which would probably cause less hindrance for the entrance of bulkier substrates. This may be the dominating factor behind the improved hydrolytic activity toward PBAT and PET.

Then et al. (2016) directly introduced a disulfide bridge (D204C-E253C) by substituting the calcium binding site of TfCut2 to eliminate dependence on calcium. The results showed that the thermal stability of the mutant was further improved, the melting point was increased to 94.7°C (TfCut2: 69.8°C), and the half-inactivation temperature was increased to 84.6°C (TfCut2: 67.3°C). Most curiously, the hydrolytic activity of the calcium-independent thermostable and efficient hydrolase was not affected when the calcium binding site was replaced. A similar study was carried out with the anchor peptide-cutinase mutant Tfuc2^D204C/E253C (Liu et al., 2022). The PET degradation rate of the obtained fused protein was determined to be up to 57.9% at 70°C for 96 h, and the catalytic performance was improved by 22.7-fold compared with that of Tfuc2. Based on machine learning, Li et al. successfully designed a highly thermostable Tfuc2 mutant, S121P/D174S/D204P, by mining molecular dynamics simulation trajectories and rational introduction of hydrogen and ionic bonds (Li et al., 2022). The results showed that the PET degradation ratio was further increased by 46.42-fold at 70°C compared to that of wild-type TfCut2.

To further increase the thermostability of the polyester hydrolases from the thermophilic actinomycete Thermobifida fusca, rational engineering of the binding sites of specific metal ions (e.g., Mg²⁺ and Ca²⁺) was also investigated (Then et al., 2015). Based on the MD simulations, residues Asp174, Glu253, and Asp204 of TfCut2 might play roles in facilitating the binding site of Ca²⁺, which were then mutated to screen for stability-enhanced mutants. The results showed that mutants D204R and E253R increased by more than 14°C compared to that of wild-type TfCut2; however, the PET-degrading activities were determined to be slightly reduced. In addition, it was found that the addition of a cationic surfactant (C₁₂-N(CH₃)₃⁺) greatly improved the degradation activity of TfCut2, which might be caused by the aggregation of enzymes near the PET film surface via electrostatic interactions (Arnling Bååth et al., 2022). The results showed that in the presence of cationic surfactant (30 ppm C₁₂-N(CH₃)₃⁺), the activity of the TfCut2^G62A/F209A mutant increased to 31 ± 0.1 nmol min⁻¹ cm⁻², which is approximately 12.7 times higher than that of wild-type TfCut2 in the absence of surfactant.

From the bacterium HR29, a close homolog of the LCC with 94% sequence identity was identified (BhrPETase, NCBI accession No. GBD22443) (Kato et al., 2018). Recently, high-level secretory expression of BhrPETase in Bacillus subtilis was successfully achieved with an expression titer of 0.66 g/L (Xi et al., 2021). More strikingly, a strong preference of BhrPETase for high temperature was determined for BhrPETase when the reaction temperature was increased from 30 to 90°C. Even at 90°C, the catalytic performance remained superior to that of wild LCC, which rapidly dropped to 55% of the maximum activity. The T_m value of BhrPETase was determined to be 101°C without the addition of Ca²⁺ ions, ~11°C higher than that of LCC. Against a crystallinity of 11.2% PET powder for 20 h at 70°C, approximately 6.3 mM degradation products were detected. Molecular modeling suggested that the increased thermostability might be caused by Ser175, Gln202, Pro199 and Pro248 in BhrPETase, which could either result in additional formed hydrogen bonds and/or increase the rigidity of the flexible loop.

In a recent study, molecular engineering of the hydrolase from the bacterium HR29 was reported by means of a protein language model and force-field-based algorithms (under review³). The obtained TurboPETase (BhrPETase^{H218S/F222I/A209R/D238K/A251C/A281C/W104L/F243T}, T_m of 84°C) exhibited a 4.4-fold increase in the degrading activities toward GF-PET films compared to that of wild-type BhrPETase. Moreover, the depolymerization rate of TurboPETase (21%) outperformed all the tested PET hydrolases, which was 4-, 4-, 8- and 33-folds higher than those of wild BhrPETase, LCC^ICCG, HotPETae and FastPETase at 65°C within 3 h, respectively. At 65°C after 24, approximately 100% depolymerization could be obtained. Especially, the full degradation of pretreated PET (up to 300 g L⁻¹) could be accomplished within 10 h, with a maximum production rate of 77.3 gTPAeq L⁻¹ h⁻¹. It shows beyond doubt that this strategy and the obtained TurboPETase might serve as a promising option for the engineering of interfacial enzymes and efficient PET cycling.

Thermobifida alba strain AHK119 (AB298783) was shown to be capable of degrading aliphatic-aromatic copolyester film and polymer particles and releasing TPA (Hu et al., 2010). From AHK119, an esterase gene (est119) was cloned and identified, and the Est119 protein contains the highly conserved motif (G-X-S-X-G) of the serine hydrolase family, a catalytic triad (Ser129, His207, and Asp175) and the oxyanion hole formed by Tyr59 and Met130. The optimal reaction temperature and pH were found to be ~50°C and 6.0, respectively. Kitadokoro et al. determined the crystal structure of Est119 (Figure 10(1)), which shows a typical α/β-hydrolase fold that also consists of a central twisted β-sheet of 9 β-strands flanked by 9 α-helices on both sides (Kitadokoro et al., 2012).

In another study (Thumarat et al., 2012), it was further confirmed that Est119 is a typical cutinase-like enzyme, and Ca²⁺ was necessary to display full catalytic activity (300 mM) and maintain thermostability (200 mM, stable at 50°C for 16 h). Random mutagenesis analysis indicated that mutations with more hydrophobicity at the N-terminus would more likely result in enhanced activity, which led to mutants A68V and A68V/S219P showing the highest activity against pNPB. In particular, the degradation activity of A68V/S219P was further improved by ~50-fold and 1.7-fold compared with that of WT and A68V, respectively. In addition to est119, its homolog est1 was located upstream of est119 (Thumarat et al., 2015). Although showing 95% sequence identity and 98% sequence similarity with Est119, Est1 displays a higher (~2-fold higher) activity toward p-nitrophenyl butyrate. Moreover, Est1 displayed a much higher thermostability, with 300 mM Ca²⁺, and it could maintain nearly 100% of the residual activity at 50°C for 20 h. Est1 and Est119 could also recognize the aromatic polyester and produce EG and TPA. As mentioned above, Est1^A68V/T253P was constructed, and the results showed that the T_m value was increased to 79°C and the specific activity (833 ± 22.5 U/mg) was also improved by ~57.4-fold against pNPB. The increased thermostability was estimated to be caused by the introduction of proline, which could decrease the flexibility and increase the structural rigidity, especially at sites 219 and 253. However, Est1^{A68V/T253P/M259K} was shown to display a higher catalytic activity than Est1^A68V/T253P, with a specific activity of 910 ± 42.0 U/mg, and the thermostability was lower than that of Est1^A68V/T253P over 50°C.

In a recent study, the crystal structure of Est119 complexed with ethyl lactate (EL) and lactic acid (LAC) was obtained (Figure 10(2)) (Kitadokoro et al., 2019). As expected, three Ca²⁺ ions were found in Est119; moreover, LAC and EL are located in the active site cleft with different positions. Based on the crystal structure, it was found that the substrate-binding cleft was a long narrow groove lined with F248, I217, H247, S169, W194, M170, Y99, T100, T102, S105, and N251. In the binding site, EL was stabilized by the H-bonds formed with H247 and water and the hydrophobic interaction with H168. In addition, LAC was regulated by the hydrophobic interaction with W194 and the H-bonds formed with water. The catalytic S169 is located at the bottom of the cleft.

Recently, a novel directional-path modification (DPM) strategy was developed to improve the performance of PET-degrading enzymes, which mainly contains the introduction of positively charged amino acids and binding groove remodeling (Chen et al., 2022). First, the PET binding path was identified. Then, to decrease steric hindrance in the binding site at the exit of the PET binding path, neutral short side-chain amino acids were introduced. At the same time, positively charged amino acids with shorter flexible side chains (e.g., Lys) were selected to optimize the surface positively charged area. These modifications led to the identification of 13 potential residue sites and construction of 23 target mutants, and the results showed that the four-point mutant (4Mz, H184S/Q92G/F209I/I213K) displayed the highest degradation rate of ~90%. The degradation efficiency was increased by approximately 30-fold compared with that of the wild type. Molecular simulations revealed that the PET binding modes of 4Mz were extended to the left side, which was estimated to be more conducive to improving PET degradation. This finding was also confirmed by positive vacuum electrostatics on the left-side subbinding site. When applied to LCC, Est119, and BhrPETase, the constructed mutants LCC^{H218S/Y127G/F243I/S247K}, Est119^{H248S/Q131G/F248I/I252K} and BhrPETase^{H184S/F93G/F209I/S213K} all showed dramatically increased PET degradation efficiency, indicating the high efficiency and universality of the DPM strategy.

In a recent study, a novel thermophilic polyesterase from Deinococcus maricopensis (DmPETase) was characterized (Makryniotis et al., 2023). DmPETase, a phylogenetically distinct and thermophilic polyesterase, was found to display a similar overall structure with that of other bacterial cutinases, such as LCC^ICCG and IsPETase, and the catalytic triad is determined to be Ser185, His263 and Asp231. The optimal temperature for DmPETase was 50°C lower than that of LCC^ICCG, and display the highest catalytic efficiency toward pNPB. It is also capable of degrading various synthetic polymers, including PET, polyurethane, as well as four semi-crystalline aliphatic polyesters. On the whole, DmPETase shows a much lower catalytic activities compared with those of LCC^ICCG, for example against the PCL powder DmPETase displayed a degrading approximately 70% of the tested materials compared with 98.7% obtained by LCC^ICCG. However, at 50°C, it displays a comparable degrading activity compared with that of LCC^ICCG against the semi-crystalline sections of post-consumer PET bottles (crystallinity, 41%).

The cutinase from the fungus Humicola insolens (HIC, Figure 11) is a single-domain protein that shows high hydrolytic activity and stability at 70°C (Carniel et al., 2021), and sequence analysis showed that HiC exhibits high sequence identities to FsC (56%) (Kold et al., 2014).

When used in moist-solid reaction mixtures instead of the typical dilute aqueous solutions or slurries, cutinase from Humicola insolens can directly depolymerize amorphous and crystalline regions of PET equally without any pretreatmentl a 13-fold higher space–time yield and a 15-fold higher enzyme efficiency were achieved than those reported in prior studies with high crystallinity material. Furthermore, this process shows 26-fold selectivity for terephthalic acid over other hydrolysis products (Kaabel et al., 2021).

In an earlier study, it was found that cutinase from Humicola insolens (HIC) could hydrolyse PET, especially the low-crystallinity (7%) film (~10-fold higher than biaxially oriented PET, 35% crystallinity) (Ronkvist et al., 2009). Moreover, HiC shows good thermostability and functions best from 70 to 80°C, and the HiC-catalyzed degradation of the lcPET film resulted in a 97 ± 3% weight loss only releasing the water-soluble TPA after 96 h at 70°C. Based on this finding, a novel synergistic chemoenzymatic hydrolysis of PET was developed to produce terephthalic acid (TA) from textile waste (Quartinello et al., 2017). First, conversion of PET into TA (purity, 85%) and small oligomers was achieved by chemical treatment (T = 250°C, p = 40 bar). Afterwards, the obtained oligomers were further hydrolysed by HiC, yielding ~6.5 mM TA after 6 h of incubation at 50°C (purity, 97%). This is a practical strategy for the efficient recycling of PET in an environmentally friendly way under neutral conditions, especially for promising catalysts displaying lower catalytic activities toward high-crystallinity PET.

Carniel et al. (2021) investigated the HiC-assisted hydrolysis of postconsumed PET (PC-PET) through controlling the pH and performing fractionation of enzyme feeding in stirred reactors. The results showed that pH variation plays a critical role in the HiC-catalyzed reaction, and a 0.86 pH variation (optimal pH 8.95 ± 0.20) finally led to a decrease of 32% in the total of final products. Compared with the tris-buffered reaction, the soluble hydrolysis products from the unbuffered HiC-catalyzed reaction with pH control were increased by 81% at 50°C (26.3 mM). Although enzyme feeding fractionation contributed little activity for the improved PET hydrolysis, an increase of 2.41-fold in the final degrading products released (97 mM) per protein amount was obtained when only half of the enzyme was loaded. These findings are important for the industrial PET biorecycling process.

Additionally, in a recent study, HiC-catalyzed PC-PET degradation was optimized (PET particle size, pH, ionic strength, and enzyme concentrations) for improved performance with experimental and mathematical modeling approaches (Eugenio et al., 2021). The results revealed that the PET particle size might play the most critical role in HIC-catalytic PET hydrolysis, and with a greater surface area (semicrystalline PC-PET), an 8.5-fold increased hydrolysing rate (1.7 mmol·L⁻¹·h⁻¹) was achieved. After 4 days at 70°C, 10.9 mmol·L⁻¹ TPA was detected against the semicrystalline PET (0.075–0.25 mm). The thermostability of wild HiC is much higher than that of the other reported cutinases, the half-life of which was determined to be 110 h at 70°C and pH 7.0. These results indicated that HiC displays great potential as a promising biocatalyst for efficient PET biodepolymerization.

SDS might be the most widely applied anionic surfactant and can regulate the bioactivities of cutinases or cutinase-like PET-degrading enzymes. The interactions of cutinase from Humicola insolens (HIC) with anionic surfactants (SDS) were investigated in several studies (Nielsen et al., 2005; Kold et al., 2014). The protein could specifically bind SDS at a very low molar ratio of SDS/HiC, and amino acids 160–163 and 168–176 (near the active site) were determined to be able to directly interact with SDS by MD simulations. Kold et al. (2014) investigated the detailed molecular interactions between SDS and HiC based on thermodynamic and structural studies. Crystal structures showed that the binding mode of SDS is quite similar to the previously reported substrate analogs to cutinases (Figure 12(2)), and the binding affinity (K) was determined to be ~10⁵ M⁻¹. In addition, mono-ethylphosphate (MEP) was found to covalently bind to S105 and could interact with F132, A108, A109, Q106, S28 and N69 by forming H-bonds (Figure 12(3)). Moreover, ethylated H173 was detected, probably resulting from the reaction with DNPP (diethyl p-nitrophenyl phosphate).

In a recent study, high-crystallinity PET in mixed PET/cotton textiles was reported to be directly and selectively depolymerized to TPA by using a commercial HIC (Novozym 51,032) and the Cellic CTec2 cellulases under moist-solid reaction conditions (Kaabel et al., 2023). Results revealed that TPA is the main degrading products with more than 40-fold selectivity over MHET after 7 days of static incubation at 55°C in the presence of HiC (0.65% w/w) and/or CTec2 (0.7% w/w), and 14.1% of TPA could be obtained when co-applied with CTec2. Interestingly, it was also found that the CTec2 enzymes were most efficient in the absence of HiC with up to 83.4% yield of glucose without any negative influence on the TPA yield.

FsC has long been explored for the hydrolysis of several industrially important esters, insoluble PET polymer films and poly(ε-caprolactone) (PCL) (Ronkvist et al., 2009). FsC is also a single-domain protein consisting of 197 amino acids (~20.8 kDa). The overall structure of FsC is similar to that of Hic, which is different from the bacterial-derived cutinase as a single-domain protein (Figure 13(1,2)). FsC also contains 5 slightly twisted parallel-stranded β-sheets that are covered by 4 α-helices on each side (Figure 13(3)). Among them, 2 flexible loop domains (residues 80–90 and 182–189) and 2 disulfide bridges (DS1:Cys31 and Cys109, DS2:Cys171 and Cys178) are highly conserved and may be associated with substrate binding and the stabilization of the overall structures (especially DS1) (Matak and Moghaddam, 2009).

In a recent study, NMR spectroscopy was used to investigate FsC with PET-hydrolysing activity and the reaction, including the reaction conditions, interactions between FsC and BHET, and product release (Hellesnes et al., 2023). It was found that BHET binding is predominantly regulated by hydrophobic interactions (with adjacent polar residues and polar residues); in particular, a more extended binding cleft around G49 and G192 is needed. Given that FsC^L81A (improved by 4-fold) and FsC^L182A (improved by 5-fold) showed much higher hydrolytic activity on PET and polyamide 6,6 fibers, the enlarged binding site might more likely result in improved cutinase activity, as discovered by Araújo et al. (2007). The main product (80% of the products) from PET hydrolysis is MHET, and after ~400 min, the release of TPA was detected (Hellesnes et al., 2023). More importantly, it was found that the pH and pD conditions might regulate the relative amounts of degradation products, and the low degrading activity of FsC on PET is caused by poor substrate binding and the subsequent slow MHET hydrolysing activity. Previously, it was concluded that increasing accumulation of the final product EG around the active site reduced the overall flexibility of FsC (Groß et al., 2017), which certainly caused decreased thermostability and catalytic performance.

To further exploit the potential of FsC in the biodegradation of parabens, a novel biocatalyst (named SDFsC) was constructed by expressing FsC on the cell surface of Sacchromycese cerevisiae with Aga1p and Aga2p protein subunits and a glycosylphosphatidylinositol (GPI) attachment sequence (Zhu and Wei, 2019). Catalytic results indicated that SDFsC could be used for the efficient removal of parabens with K_m values of 3.1 ± 0.5 mM (for pNPA) and 0.67 ± 0.15 mM (pNPB); thus, the catalytic performance of SDFsC was slightly affected. In addition, the side chain structures of parabens play a critical role in the degradation rates, and parabens with relatively long alkyl or aromatic side chains (propylparaben, butylparaben and benzylparaben) were more efficiently degraded, with efficiencies of 89, 97 and 93%, respectively, after 24 h at 30°C. Moreover, SDFsC showed excellent reusability with 93% residual enzyme activity after being used for paraben degradation 6 times. Altogether, it is believed that biocatalysis based on the enzyme cell surface display would be a green and efficient alternative for efficient PET biodegradation.

In 1997, two distinct serine esterases with cutinolytic activity (52 kDa and 26 kDa) were discovered (Trail and Köller, 1993), and a cutinase-encoding gene, CUTAB1, was then identified from an A. brassicicola cDNA library (Yao and Kller, 1994).

The cutinase CUTAB1 from Alternaria brassicicola (corresponding to the cutinase from the A. brassicicola culture) was successfully heterologously expressed in Pichia pastoris (Koschorreck et al., 2010). It displayed the highest enzymatic activity at 40°C and pH 7–9, and similarly, the chain lengths of different substrates showed significant effects on the hydrolysing activity; the highest catalytic activity was toward tributyrin and the lowest activity was toward p-nitrophenyl palmitate (p-NPP). In contrast to the findings for FsC (Hellesnes et al., 2023), the mutants AbC^L80A, AbC^L181A and AbC^I183A showed a significant decrease in specific activity toward p-NPB and p-NPP. In contrast, the specific activity of AbC^A84F against p-NPP was significantly improved by ~10-fold. This strange phenomenon was thought to be associated with increased interactions with longer chain and hydrophobic substrates caused by the bulkier and more hydrophobic phenylalanine.

Two cutinase genes from Fusarium solani (FsC) and Aspergillus fumigatus (AfC) were also successfully expressed in Pichia pastoris X33 (Ping et al., 2017). The results showed that FsC and AfC display significantly different optimal catalytic temperatures (40°C and 60°C, respectively) but similar pH values (7.5 and 8.0, respectively). This illustrated the superior thermostability of AfC. Both FsC and AfC can hydrolyse p-nitrophenyl substrates with different carbon chain lengths, and pNPD and pNPH were the preferred substrates for FsC and AfC, respectively. Moreover, AfC could completely hydrolyse the PCL film.

Aspergillus oryzae is a well-known filamentous fungus that has long been used in traditional microbial fermentation and the food industry. Recently, the fungus has attracted much attention for the biodegradation of poly(butylene succinate) (PBS) and poly(butylenes succinate-co-adipate) (PBSA) with the successful identification of a PBS-degrading enzyme (AoC) (Maeda et al., 2005). AoC (~21.6 kDa) displayed the highest catalytic activity at 55°C and could remain stable under a wide pH range (pH 6.0–11.0).

In a recent study, the biological properties of AoC were investigated, and the structure–activity relationship was analyzed based on the crystal structure (Liu et al., 2009). AoC showed better thermostability with a T_m value of 59°C. AoC showed a preference for pNPV with a K_m value of 0.04 ± 0.01 μM and displayed the highest catalytic efficiency toward pNPB and pNPV with k_cat/K_m values of 3.49 ± 0.51 μM⁻¹ min⁻¹ and 3.32 ± 0.74 μM⁻¹ min⁻¹, respectively. At 40°C and pH 8.0, poly ε-(caprolactone) (PCL) was almost completely degraded (87% PCL) within 6 h; however, only 30% degradation was achieved catalyzed by FsC. The structure (Figure 14(1)), which is similar to that of FsC and HiC, contains a typical α/β fold characterized by a central β-sheet of 5 parallel strands; this sheet is surrounded by 10 α-helices with a highly conserved catalytic triad S126, D181, and H194. In addition, 3 disulfide bonds were identified, including DS1 between Cys37 and Cys115, DS2 between Cys177-Cys184, and DS3 between Cys63-Cys76. Among them, DS3 is unique to the cutinases from the Aspergillus family connecting helix 2 to strand 2 of the central β-sheet, which might lead to improved thermoactivity of AoC. Different from FsC and HiC, an extended groove was discovered around the substrate binding site that regulates the orientation of polymeric chains (Figure 14(3)), which resulted in a much higher hydrolysing activity of AoC toward the longer chain substrates. Furthermore, residues L87 and V190, similar to the “gate keepers,” may be associated with substrate recognition, especially the different alkyl chains.

To further improve thermostability, Shirke et al. investigated the thermal inactivation mechanisms of AoC (Shirke et al., 2017). First, it was revealed that glycosylation plays a critical role in the improved thermostability of Thiellavia terrestris cutinase (TtC) expressed in Pichia pastoris by inhibiting its thermal aggregation. In view of this finding, kinetic models of cutinase thermal inactivation were created to account for thermal aggregation. However, the models also indicated that the glycosylation of TtC is helpful for the improved kinetic stability (5-15-fold) by inhibiting thermal aggregation at 61°C. Then, one glycosylation site (L185N) was designed and incorporated into AoC, which completely inhibited the thermal aggregation of AoC with a 4-fold improvement in thermostability. However, it should be noted that the undesirable catalytic efficiency of AoC^L185N was lower (~70%) than that of wild-type AoC, and this may be avoided by rationally designed glycosylation sites.

Recently, Lewatit VP OC 1600 was used as the macroporous support for the immobilization of AoC, HIC and TfC for industrial applications of cutinases (Su et al., 2018). The results demonstrated that HiC displays the highest solvent tolerance with increased polarity, and TtC shows the best thermostability at 80°C. Furthermore, in nonane, the three obtained catalysts could maintain ~64% residual enzyme activity even at 90°C, which reveals that this strategy is efficient in improving the organic solvent tolerance.

In a study, the abovementioned cutinases were investigated in detail (Baker et al., 2012). HiC was shown to exhibit the highest stability with significantly improved poly(ε-caprolactone) hydrolysing activities at high temperatures under the tested conditions. Additionally, at low pH conditions, HiC could maintain the overall structure with little affected activity, but AbC exhibited the largest loss in structure. Similarly, FsC and AbC were found to be the least stable at high temperatures.

Recently, two highly similar thermophilic PET hydrolases (PES-H1 and PES-H2) were identified from a compost metagenome library, and PES-H1 and PES-H2 differ in only four residues (A/E1, L/F209, D/N232, and S/A254) (Sonnendecker et al., 2022). Although they contain quite similar amino acid sequences, PES-H1 displayed a much higher hydrolytic activity on amorphous PET films, even higher than that of the wild-type LCC under identical conditions. To investigate the catalytic mechanism, high-resolution crystal structures of PES-H1 complexed with various PET substrates were obtained, including citrate [PDB ID: 7E30, with 4-(2-hydroxyethylcarbamoyl) benzoic acid (MHETA), PDB ID:7W6C, 7W6O, and 7W6Q] and with bis(2-hydroxyethyl) terephthalate (BHET, PDB ID: 7W66) (Pfaff et al., 2022). Regarding the structures, PES-H1 adopts the canonical α/β-hydrolase fold consisting of 9 β-sheets and 10 α-helices and exhibits highly conserved features of the α/β-hydrolase superfamily similar to LCC. Overall, the binding modes of the different ligands in the active sites are quite different but similar to those observed for HEMT bound to IsPETase ^{R103G/S131A28} and MHET bound to the LCC ^ICCG. Given that the multiple noncatalytic intermediate binding modes (e.g., MHETA) far from S130 are less likely to undergo nucleophilic attack, enzyme-mediated PET hydrolysis involves the dynamic reorientation of polymer chains in the substrate-binding site. The effect of W155 was also identified (Figure 15(2–7)), which is critical for PET binding and stabilization. Especially, from Figure 15(2–4), the different conformations of MHETA were obtained, which clearly revel the MHET degrading process. MHETA located away from the catalytic triad is stabilized by hydrophobic residues, including F62, I178, L209, M131 and F62. In addition, the TPA moiety is in the vicinity of F62 rather than close to the catalytic triads. Moreover, the special binding mode of MHETA located at the surface of PES-H1 (Figure 15(4)) indicates that for PET hydrolysis, it might first bind at the surface and be transferred into the deeper active site cavity.

Given the crystal structures and the MD simulations, mutational hotspots were identified for rational engineering of PES-H1, such as the R204C/S250C variant, R6C/S49C variant, L92F/Q94Y, and L92F/Q94Y/R204C/S250C variant. The results showed that mutant PES-H1^L92F/Q94Y displayed a 1.8°C higher T_m and could almost completely degrade the Gf-PET film after 24 h at 72°C, with the highest weight loss of 37.1 mg (62.4%). It outperforms the promising LCC ^ICCG mutant in the biohydrolysis of low-crystallinity (13%) PET powder with 2.2-fold more degradation products. It was believed that the enhanced activity was caused by the “aromatic tunnel” (substitutions of L90F) effect streaming the efficient binding of PET chains to the active site (Cui et al., 2021).