Yifei Zhang

Polyethylene terephthalate (PET) nanoplastics are pervasive aquatic contaminants due to their high dispersibility and persistence, yet their removal at extremely low concentrations remains challenging. Here, we report a “capture‐and‐degrade” strategy for PET nanoparticle degradation using enzyme‐functionalized bowl‐shaped Janus particles. Amphiphilic Janus particles featuring a hydrophobic concave surface were synthesized via interfacial emulsion polymerization. A thermostable PET hydrolase, HotPETase, fused with the hydrophobic anchor peptide LCI, was specifically immobilized on the concave region. This architecture efficiently captures PET nanoparticles from dilute systems, promoting enzyme–substrate colocalization and thereby enhancing their degradation. The strong interaction between the Janus particles and the HotPETase‐LCI fusion protein provides the immobilized enzyme with high operational activity and stability, enabling reuse for more than 17 cycles with over 90% of its initial activity retained. Furthermore, integration into a packed‐bed continuous‐flow reactor achieved 96% PET removal during 20 h of operation, corresponding to a space‐time conversion of 1.5 × 10 ¹⁴ particles L ⁻¹ ·h ⁻¹ . This work establishes a scalable platform based on topology‐engineered Janus materials for PET nanoplastic remediation and may inspire new approaches for the capture and degradation of other nanoplastics from dilute conditions.

Heterologous expression of enzymes from higher organisms limits the construction of microbial cell factories for natural product biosynthesis. Here we develop a ProteinMPNN-based sequence redesign strategy, guided by essential structural features and evolutionary conservation, to improve the bacterial expression of heterologous enzymes. Applied to two plant glycosyltransferases, TOGT and UGT84A56, for esculetin glycosylation, this strategy generates extensively redesigned variants with markedly enhanced soluble expression and catalytic activity in Escherichia coli. In vitro enzyme assays and in vivo whole-cell conversions show consistent improvements, outperforming conventional solubility-enhancing strategies. Molecular simulations suggest that the improved performance arises from global optimization of hydrophobic and hydrophilic residue exposure while preserving productive substrate-binding interactions. Fed-batch fermentation achieved titers of 2.11 g/L cichoriin and 4.05 g/L aesculin. This work establishes a generalizable route for converting difficult-to-express eukaryotic enzymes into efficiently expressible and functional variants, thereby expanding the enzymatic toolbox for microbial cell factory construction.

Heterologous expression of enzymes from higher organisms remains a major bottleneck in the construction of microbial cell factory. To address the problem, we developed a sequence redesign strategy guided by essential structural features and evolutionary conservation. This strategy integrates ProteinMPNN with several scorers to comprehensively redesign enzymes that inherently exhibit poor heterologous expression. We applied this approach to two glycosyltransferases, TOGT and UGT84A56, for esculetin glycosylation. The redesigned enzymes featured alterations in more than 20% of their sequences, resulting in significant improvements in soluble expression in Escherichia coli. Furthermore, both in vitro and in vivo assays demonstrated notable enhancements in the catalytic activity of the redesigned enzymes. Specifically, the best-performing TOGT variant exhibited a 6.4-fold increase in soluble expression and an 11.8-fold improvement in catalytic efficiency, enabling the production of cichoriin at a titer of 299 mg/L in whole-cell catalysis. Similarly, the top-performing UGT84A56 variant achieved 402 mg/L aesculin, representing a 13.9-fold improvement over the native enzyme. Fed-batch fermentation using these redesigned enzymes achieved record-high product titers of 2.11 g/L for cichoriin and 4.05 g/L for aesculin, respectively. This study demonstrates a powerful strategy for enzyme redesign, paving the way for efficient microbial cell factories to synthesize high-value natural products.

A four-enzyme cascade was developed for the enantioselective synthesis of epinephrine in a modular continuous-flow system. This process yielded pharmaceutical-grade epinephrine with a space–time yield of 2.17 g L ⁻¹ h ⁻¹ and 97.1% enantiomeric excess.

Accurate prediction of protein and peptide functions from amino acid sequences is essential for understanding biological processes and advancing biomolecular engineering. Due to the limitations of experimental methods, computational approaches, particularly machine learning, have gained significant attention. However, many existing tools are task-specific and lack adaptability. Here, we propose a BERT–BiLSTM–Attention–TCN Protein Function Prediction Framework (BBATProt), a versatile framework for predicting protein and peptide functions. BBATProt leverages transfer learning with a pretrained bidirectional encoder representations from transformer model to capture high-dimensional features. The custom network integrates bidirectional long short-term memory and temporal convolutional network to align with proteins’ spatial characteristics, combining local and global feature extraction via attention mechanisms to achieve more precise predictions. Evaluations demonstrate that BBATProt consistently outperforms state-of-the-art models in tasks such as hydrolytic catalysis, peptide bioactivity, and post-translational modification (PTM) site prediction. Specifically, BBATProt improves accuracy by 2.96%–41.96% in antimicrobial peptide (AMP) prediction and by 0.64%–23.54% in PTM prediction tasks. In terms of area under the receiver operating characteristic curve, improvements range from 0.71% to 40.51% for AMP prediction and 0.62%–27.82% for PTM prediction. Visualizations of feature evolution and refinement via attention mechanisms validate the framework’s interpretability, providing transparency into the feature-extraction process and offering deeper insights into the basis of property prediction.

Evolutionary constraints significantly limit the diversity of naturally occurring enzymes, thereby reducing the sequence repertoire available for enzyme discovery and engineering. Recent breakthroughs in protein structure prediction and de novo design, powered by artificial intelligence, now enable to create enzymes with desired functions without solely relying on traditional genome mining. Here, a computational strategy is demonstrated for creating new‐to‐nature polyethylene terephthalate hydrolases (PET hydrolases) by leveraging the known catalytic mechanisms and implementing multiple deep learning algorithms and molecular computations. This strategy includes the extraction of functional motifs from a template enzyme (here leaf‐branch compost cutinase, LCC, is used), regeneration of new protein sequences, computational screening, experimental validation, and sequence refinement. PET hydrolytic activity is successfully replicated with designer enzymes that are at least 30% shorter in sequence length than LCC. Among them, RsPETase1 stands out due to its robust expressibility. It exhibits comparable catalytic efficiency (k_cat/K_m) to LCC and considerable thermostability with a melting temperature of 56 °C, despite sharing only 34% sequence similarity with LCC. This work suggests that enzyme diversity can be expanded by recapitulating functional motifs with computationally built protein scaffolds, thus generating opportunities to acquire highly active and robust enzymes that do not exist in nature.

Enzymatic recycling of polyethylene terephthalate (PET) has attracted significant attention in recent years. While the fusion of anchor peptides to PET hydrolases is believed to enhance PET hydrolytic activity, a quantitative analysis is yet lacking. Here, we construct four fusion enzymes by fusing anchor peptides (including hydrophobic LCI, LCIM1 and TA2, and hydrophilic EK4) to the C terminus of HotPETase, one of the most active PET hydrolases for high‐crystallinity PET (HC‐PET). Single‐molecule force spectroscopy (SMFS) demonstrates that hydrophobic anchor peptides promote adhesive interactions between the fusion enzymes and the PET surface. This is also validated by the adsorption kinetics and isotherms, and the saturated adsorption capacity remains unaltered compared to HotPETase. At low substrate loadings, the apparent hydrolytic activity of these fusion enzymes is positively related to the hydrophobicity of the anchor peptides. Among them, HotPETase‐LCI stands out as the most effective enzyme for HC‐PET degradation, demonstrating a 1.5‐fold increase in hydrolytic activity. At high substrate loadings, the advantages of fusion with anchor peptides diminish. We conclude that fusion enzymes only facilitate the hydrolytic rates of reactions for HC‐PET but have little effect on the final conversion extent.