HYHui-Lei Yu

Hui-Lei Yu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Hui-Lei Yu has a total of 184 co-authors. Here are the top co-authors by number of shared publications:

  • Jian-He Xu

    East China University of Science and Technology

    36 shared publications
  • Zhi-Jun Zhang

    State Key Laboratory of Bioreactor Engineering Shanghai Collaborative Innovation Centre for Biomanufacturing East China University of Science and Technology Shanghai 200237 P.R. China

    15 shared publications
  • Jian-He Xu

    State Key Laboratory of Bioreactor Engineering, Shanghai Collaborative Innovation Centre for Biomanufacturing, East China University of Science and Technology, 130 Meilong Road, Shanghai 200237, People's Republic of China

    12 shared publications
  • Jian‐He Xu

    State Key Laboratory of Bioreactor Engineering East China University of Science and Technology Shanghai 200237 China

    11 shared publications
  • Qi Chen

    State Key Laboratory of Bioreactor Engineering Shanghai Collaborative Innovation Centre for Biomanufacturing East China University of Science and Technology Shanghai 200237 P.R. China

    10 shared publications
  • Qiang Geng

    State Key Laboratory of Bioreactor Engineering, Shanghai Collaborative Innovation Center for Biomanufacturing, East China University of Science and Technology, 130 Meilong Road, Shanghai 200237, China

    10 shared publications

Latest Publications

  • Open Access

    CausalMut: a causal deep learning framework for enzyme mutational effect prediction

    by Wenli Gu, Fanxu Zeng, Xiaoyu You, Huilei Yu, Qingchao Jiang

    Accurate prediction of enzyme mutational effects on catalytic kinetics is fundamental to data-driven enzyme engineering, yet existing deep learning models are fundamentally constrained by biases inherent in the data generation process. Experimental datasets are conditioned on prevailing experimental choices rather than uniform mutation interventions, which introduces spurious correlations that compromise model generalization. To address this challenge, CausalMut is proposed as a causality-guided deep learning framework that disentangles genuine mutational causal effects from confounding biases. The data generation mechanism is first formalized via a structural causal graph, revealing the experimental environment as a confounder that induces distributional imbalance across mutation types. Building on this analysis, CausalMut integrates two components: a Multi-scale Interaction Enhanced Network (MIENet), which captures complex interactions among the wild-type enzyme, substrate, and mutant without 3D structural information; and a Joint Focal-IPS optimization strategy, which corrects distributional imbalances at both the amino-acid and mutant levels. Comprehensive experiments on three real-world enzyme kinetic datasets ($k_{\text{cat}}$, $K_{m}$, and $k_{\text{cat}}/K_{m}$) against five state-of-the-art baselines demonstrate average improvements of 16.24% in root mean square error and 22.19% in $R^{2}$. CausalMut maintains consistent superiority under three biologically motivated distribution-shift settings, and case studies further demonstrate its ability to identify high-activity variants within a fixed screening budget, validating its robustness and practical utility.

  • Loop‐Centric Engineering of Terminal Deoxynucleotidyl Transferase for Enhanced Catalysis of 3'‐Modified Nucleotides

    by Kun Shi, An‐Na Li, Jian‐He Xu, Hui‐Lei Yu

    Efficient enzymatic DNA synthesis requires terminal deoxynucleotidyl transferases (TdTs) that can accept reversibly blocked nucleotides, but improving wild‐type enzymes has traditionally depended on labor‐intensive screening. Here, we combine deep learning‐based structure prediction with molecular dynamics simulations to guide loop‐centric engineering of a TdT from Crocodylus porosus . The resulting variant M3 (V253E/L256M/N338R) showed a 26‐fold increase in specific activity toward 3'‐ONH 2 ‐dCTP while maintaining expression and thermostability. Simulations indicated that the mutations reorganized Loop1 dynamics and reproducibly biased the incoming nucleotide toward a more favorable in‐line attack orientation. The key N338R change also improved a second TdT scaffold, supporting N338 as a potentially transferable engineering hotspot. These results provide a practical computational strategy for improving terminal deoxynucleotidyl transferases for incorporation of 3′‐modified nucleotides, reducing reliance on large‐scale screening.

  • Deep Learning-Guided Interface Engineering Stabilizes Oligomeric Enzymes

    by Wei-Jie Zhan, Yang Zhuo, Zhi-Hao He, Xin-Yi Lu, Zhi-Jun Zhang, Xiao-Yu You, Qing-Chao Jiang, Kun Shi, Hui-Lei Yu

    The thermostability of oligomeric enzymes is often limited by the intrinsic flexibility of subunit interfaces. Herein, we report a deep learning-driven interface engineering strategy (DeepIE) to systematically stabilize oligomeric enzymes. Applied to a dimeric formate dehydrogenase, we de novo redesigned six flexible regions at the dimer interface. The lead variant retains wild-type catalytic activity while exhibiting an ∼500-fold increase in half-life at 50 °C. Molecular dynamics analyses revealed that reduced local flexibility and strengthened interfacial hydrophobic packing underpin the enhanced thermostability. This work establishes an artificial intelligence-driven, generalizable framework for rational thermostabilization of oligomeric biocatalysts, effectively overcoming the activity–stability trade-off.

  • Probing the Catalytic Promiscuity and Enhancing the Efficiency of CYP725A4 with a Taxadiene Substrate Mimic

    by Longhao Huang, Mingyuan Lai, Zihan Zhang, Lili Zhu, Ran Hong, Jian-He Xu, Hui-Lei Yu

    The catalytic promiscuity of cytochrome P450 CYP725A4 severely impedes paclitaxel biosynthesis, as it generates taxadien-5α-ol (T-5α-ol) alongside numerous by-products. Despite diverse proposed mechanisms and the reported role of facilitator of taxane oxidation (FoTO1) in promoting T-5α-ol formation, conclusive evidence elucidating the catalytic promiscuity of CYP725A4 remains elusive. Here, we report a substrate mimic strategy using taxadien-2-one (taxadienone), which is selectively epoxidized by CYP725A4 at the C4=C5 double bond, suggesting epoxidation as its major catalytic pathway. This notion was then confirmed by the successful identification of the long-observed yet structurally uncharacterized major taxadiene oxidation product as 4(5)α-epoxy-taxadiene, which can spontaneously rearrange in solution. Furthermore, the minor T-5α-ol produced by CYP725A4 arises from a concurrent hydroxylation pathway rather than from epoxidation, and efficient isomerization of 4(5)α-epoxy-taxadiene to T-5α-ol by FoTO1 is required to achieve improved product selectivity. Employing the promiscuity-free taxadienone as a probe substrate eliminated analytical complications and led to variant M3, which shows markedly enhanced catalytic activity through optimized heme-binding and enhanced FG loop flexibility. Finally, co-expressing the plant-derived FoTO1 with the CYP725A4 variant M3 in engineered Escherichia coli increased T-5α-ol production to 109.1 mg/L in a 5 L bioreactor. Collectively, this strategy provides crucial insights into CYP725A4 catalytic promiscuity and enables more efficient directed evolution, thereby facilitating downstream paclitaxel pathway optimization.


  • Engineering ω-transaminase for enhanced aliphatic semi-aldehyde activity enables efficient nylon monomer synthesis

    by Ju-Mou Li, Yi-Ke Zhang, Zhi-Jun Zhang, Jian-He Xu, Hui-Lei Yu

    Semi-rational design enhanced transaminase activity for aliphatic semi-aldehydes, giving nylon monomer STYs of 1.59–9.29 g L −1 h −1 . MD showed that the "dual-end anchoring" mechanism of M169Q/R417N restricts substrate motion at the catalytic center.

  • Efficient Synthesis of Ergothioneine in Escherichia coli Through Identification of Bacterial‐Type Egt1 and Self‐Assemble of Pathway Enzymes

    by Xin‐Yi Chen, Shuai Yan, Hang‐Cheng Li, Hai‐Ping Liu, Hui‐Lei Yu, Jian‐He Xu, Zhi‐Jun Zhang

    Ergothioneine (EGT) offers diverse physiological benefits and potential applications. Production of EGT by engineered bacterial strain has emerged as an eco‐friendly and cost‐effective alternative to traditional extraction and chemical synthesis methods. The fungal pathway requires only two steps and is simpler than the bacterial pathway, which typically requires three or five steps depending on the catalytic characteristics of EgtB. However, heterologous overexpression of fungal enzymes in Escherichia coli still represents a great challenge, which would limit the practical application of fungal pathway for EGT biosynthesis. In this study, a bacterial‐type Egt1 with catalytic function similar to that of the fungal Egt1 was identified by genome data mining, enabling the construction of the simplest bacterial EGT biosynthetic pathway. The pathway was introduced into E. coli resulting in an EGT titer of 426 mg/L. Cip‐based self‐assemble of enzymes requires only one scaffold protein and occurs spontaneously, which has been employed to aggregate the rate‐limiting enzymes of biosynthetic pathway to speed up the desired reaction via substrate channeling effects. The pathway enzymes were self‐assembled using the CipB scaffold protein. The optimal engineered strain E48 produced 925 mg/L of EGT in shake flask after 72 h cultivation, with a productivity of 13 mg/(L·h). Finally, the production of EGT was scaled up in a 5 L fermenter, with an EGT titer reaching 1934 mg/L and a productivity of 36.5 mg/(L·h) after 55 h of fermentation. This study provides new approach for the efficient synthesis of EGT and other value‐added bioactive molecules.


  • Open Access

    Designing an oxidase toolbox for site-directed oxidation of taxanes

    by Mingyuan Lai, Yu Weng, Jian Wei, Xin-Yi Lu, Zhengyu Huang, Hangcheng Li, Longhao Huang, Zhijun Zhang, Lili Zhu, Jian-He Xu, et al.

    Paclitaxel is one of the most prominent drugs for cancer therapy, and its synthesis has long been pursued by scientific community. Although the biosynthetic pathway has been gradually unraveled, its synthetic efficiency remains severely constrained by the low activity and low product selectivity of the hydrolases involved. In this study, an artificial toolbox of oxidases derived from mutants of TteUPO (unspecific peroxygenase from Thielavia terrestris) is developed for site-directed oxidation of taxanes at the C-4, C-6, C-10, C-11, C-12, and C-13 positions. Furthermore, de novo biosynthesis of oxidized taxadiene products is achieved in an engineered E. coli chassis. These findings offer an unconventional strategy for developing an efficient route toward paclitaxel and constructing paclitaxel analogs with different structures for medicinal evaluation.


  • Flipping of a Non‐productive Substrate Binding Conformation Facilitates Hydroxynitrile Lyase Catalyzed Hydrocyanation

    by Yu‐Cong Zheng, Yong Mao, Qiang Geng, Fu‐Long Li, Xu‐Dong Kong, Yi‐Ke Qi, Lin Zhang, Qi Chen, Zhi‐Jun Zhang, Ran Hong, et al.

    Understanding enzyme–substrate conformational transformations is crucial to the design and engineering of biocatalysts. However, the mechanisms by which substrates undergo dynamic transformations that regulate the function of an enzyme remain poorly understood. Hydroxynitrile lyase from Prunus communis ( Pc HNL5) catalyzes the cleavage of cyanohydrins. Its reverse reaction holds significant synthetic potential for the preparation of pharmaceutical precursors. Using a combination of crystallography and computational experiments, a novel flipped substrate binding state is identified within the substrate tunnel of the Pc HNL5 L331A mutant. This binding state is non‐productive and undergoes a conformational change before the catalytic cycle can proceed. Site‐saturation mutagenesis led to the discovery of a triple mutant, Pc HNL5 L331A/S333V/P340L , that destabilizes the non‐productive substrate binding state thereby facilitating its transition to the catalytically productive conformation and significantly enhancing catalytic efficiency. Crystallographic studies provide a structural description of the factors that stabilize versus destabilize the different binding conformers in the different enzyme variants and thus the differing catalytic efficiencies. These findings demonstrate that destabilizing unfavorable substrate binding conformations within an enzyme active site can improve functionality and provide a promising strategy for designing efficient biocatalysts.