Zhiguang Yuchi

Enzyme catalysis has become a cornerstone of modern synthesis, offering sustainable routes to structurally complex molecules. However, the rugged fitness landscape and limited model capacity to learn from sparse data constrain the development of biocatalysts. Here, we present Rank-guided Exploration for Automated enzyme reProgramming (REAP), a closed-loop platform that combines artificial intelligence with robotic experimentation to accelerate enzyme engineering. At its core, REAP is powered by RankReg, a hybrid loss function that jointly optimizes ranking fidelity and quantitative accuracy, enabling adaptive in-loop learning directly from experimental feedback. This ranking-regression framework allows REAP to identify functional hotspots across both catalytic centers and previously underexplored distal regions, and to progressively extend model-guided exploration from single mutations to productive combinatorial variants. Applied to cytochrome P450 BM3, REAP achieved a 57-fold improvement in desired activity within five cycles, and further produced up to 104-fold enhancement in Staphylococcus aureus Sortase A (SaSrtA). By coupling algorithmic innovation with automation, REAP establishes a scalable and generalizable framework for accelerating biocatalyst discovery and expanding the scope of enzyme-enabled synthesis.

Fipronil (FIP) is a phenylpyrazole insecticide that acts as a noncompetitive pore blocker of insect type A γ-aminobutyric acid chloride channels (GABAARs). Although genetic and mutagenesis studies have long implicated the pore-lining TM2 helix, particularly the 2′ and 6′ positions, in FIP action, the structural basis by which this asymmetric phenylpyrazole is accommodated within the pseudo-fivefold symmetric chloride-channel pore has remained unknown. Here, we determined a 3.7 Å cryo-EM structure of FIP bound to an ELIC-based chimera containing the transmembrane domain of human GABAAR α1 subunit, providing the first structural view of phenylpyrazole recognition within the GABAA receptor pore. The structure reveals a single FIP molecule occupying the lower TM2-lined ion-conduction pathway in an elongated asymmetric pose. The phenyl group is positioned near the hydrophobic 2′ layer, whereas the trifluoromethylsulfinyl moiety extends toward the polar 6′ threonine ring. Molecular dynamics simulations support the stability of this configuration and reveal repeated formation of a hydrogen bond between the sulfoxide group and the 6′T side chain. Consistent with the structural model, substitutions at the equivalent 2′A and 6′T positions in the Drosophila homomeric RDL receptor markedly reduced FIP sensitivity, providing functional validation of the proposed binding mechanism. Structural comparison with picrotoxin- and pregnenolone sulfate-bound GABAARs further demonstrates that FIP exploits a conserved lower-TM2 inhibitory site shared by chemically diverse pore blockers. Together, these findings define the structural basis of FIP recognition and noncompetitive inhibition through direct pore occlusion, provide a mechanistic explanation for resistance-associated pore mutations, and establish a framework for the rational design of next-generation phenylpyrazole insecticides.

Insect Cys-loop receptors including γ-aminobutyric acid receptors (GABAARs) and glutamate-gated chloride channels (GluCls), are major targets of neuroactive insecticides. However, how individual receptor subtypes contribute to whole-animal insecticide susceptibility remains poorly understood. Here, we combined adult feeding bioassays using Drosophila melanogaster knockout (KO) lines lacking either the glutamate-gated chloride channel (GluCl) or the GABAAR subunits RDL, LCCH3, or GRD with RT-qPCR, homology modeling, molecular dynamics (MD) simulations, and MM-GBSA calculations to define the receptor-subtype dependence of abamectin and fluralaner. Deletion of GluCl produced the largest reduction in abamectin susceptibility, increasing the LC₅₀ from 11.9 to 101 mg/L (resistance ratio, RR = 8.49), whereas deletion of RDL, LCCH3, or GRD also caused substantial decreases in susceptibility (RR = 6.13-6.76). These results indicate that abamectin acts through multiple Cys-loop receptors, with GluCl making the predominant contribution. In contrast, fluralaner susceptibility depended primarily on GABAARs. Deletion of LCCH3 caused the greatest reduction in sensitivity, increasing the LC₅₀ from 0.0656 to 0.325 mg/L (RR = 4.95), followed by deletion of RDL and GRD, whereas loss of GluCl had little effect (RR = 1.04), indicating that GluCl makes little, if any, contribution to fluralaner toxicity in vivo. RT-qPCR revealed selective compensatory transcriptional changes among the remaining receptor subunits in different KO backgrounds. MD simulations further identified distinct ligand recognition mechanisms. Abamectin binding was stabilized predominantly by extensive hydrophobic packing together with conserved backbone interactions across multiple receptor interfaces, whereas fluralaner preferentially recognized RDL-containing GABAAR interfaces through a more specific hydrogen-bonding network. Together, our results establish that receptor subtype composition is a major determinant of insecticide susceptibility in vivo. These findings provide a mechanistic framework for understanding target selectivity and resistance and highlight receptor subtype specificity as a guiding principle for the development of next-generation insecticides.

Ependymal multicilia position at one-side on the cell surface and beat synchronously across tissue to propel the flow of cerebrospinal fluid. Loss of ependymal cilia often causes hydrocephalus. However, molecules contributing to their maintenance remain yet fully revealed. Cytosolic carboxypeptidase (CCP) family are erasers of polyglutamylation, a conserved posttranslational modification of ciliary-axoneme microtubules. CCPs possess a unique domain (N-domain) N-terminal to their carboxypeptidase (CP) domain with unclear function. Here, we show that a novel mutant mouse of Agbl5, the gene encoding CCP5, with deletion of its N-terminus and partial CP domain (designated Agbl5M1/M1), developed lethal hydrocephalus due to degeneration of ependymal multicilia. Interestingly, multiciliogenesis was not impaired in Agbl5M1/M1 ependyma, but the initially formed multicilia beat in intercellularly diverse directions, indicative of aberrant tissue-level coordination. Moreover, actin networks are severely disrupted and basal body patches are improperly displaced in mutant cells, suggesting impaired cell level polarity. In contrast, Agbl5 mutants with disruption solely in the CP domain of CCP5 (Agbl5M2/M2) do not develop hydrocephalus despite increased glutamylation levels in ependymal cilia as similarly seen in Agbl5M1/M1. This study revealed an unappreciated role of CCP5, particularly its N-domain, in ependymal multicilia stability associated with their polarization and coordination.

Objective: Xuanfei Baidu decoction (XFBD), one of China’s recommended prescriptions for COVID-19, has demonstrated clinical efficacy in mitigating gastrointestinal inflammation associated with SARS-CoV-2 infection, yet its exact therapeutic mechanism remains incompletely understood, necessitating further in vivo investigation. In this study, we utilized a fruit fly ( Drosophila melanogaster ) model of intestinal inflammation to assess the therapeutic potential and underlying mechanisms of XFBD in ameliorating intestinal inflammation. Methods: We induced intestinal inflammation in D. melanogaster by dextran sulfate sodium (DSS) administration. We examined XFBD’s protective effects on D. melanogaster ’s intestine and explored changes in the IMD and JNK-JAK/STAT signaling pathways. Furthermore, we analyzed the intestinal microbiota through 16S rRNA gene sequencing. Results: XFBD effectively mitigated DSS-induced intestinal inflammation, leading to a significant extension of the flies’ lifespan. It enhanced the integrity of intestinal barrier while reducing reactive oxygen species (ROS) levels in intestine. Additionally, XFBD restored the disrupted intestinal microbiota by regulating immune homeostasis, notably inhibiting the IMD, Toll, and JNKJAK/STAT pathways. Furthermore, the reintroduction of symbiotic microbiota partially alleviated the DSS-induced dysregulation. Conclusion: In conclusion, our study underscores the therapeutic potential of XFBD for intestinal inflammation, providing valuable insights into its key mechanisms and highlighting the pivotal role of symbiotic microbiota in anti-inflammatory responses and immune pathway regulation.

Protein Tyrosine Phosphatase 1B (PTP1B) is a key immune regulator in cancer and an attractive immunotherapy target, yet progress is limited by the lack of selective inhibitors. Here, we identify Fumosorinone (FU), a natural product from Isaria fumosorosea , as a potent and selective allosteric inhibitor of PTP1B. In a murine colon tumor model, FU enhances anti-tumor immunity by reshaping the microenvironment, strengthening CD8⁺ T-cell responses, and promoting M1-like macrophage polarization. Enzymatic and biophysical analyses confirm its potency and direct engagement with PTP1B. A co-crystal structure defines a previously uncharacterized allosteric pocket that stabilizes the inactive state of the enzyme. This pocket is poorly conserved across the PTP family, consistent with minimal activity toward related phosphatases except TCPTP. Guided by this insight, virtual screening identifies additional inhibitors. These findings provide a structural basis for selective PTP1B targeting and support future immunotherapy development and rational drug discovery efforts.

Deoxyribonucleic acid provides unmatched information density and longevity for data storage, yet its easy amplification by polymerase chain reaction enables unauthorized replication at negligible cost. We introduce ZAT-DNA, which encodes information in patterns of canonical adenine and noncanonical 2-aminoadenine. As DNA polymerases cannot distinguish adenine from 2-aminoadenine, polymerase-based amplification erases these patterns, enforcing molecular-layer non-replicability intrinsic to the base-pairing ambiguity. We validate ZAT-DNA for secure key storage, demonstrating error-free encoding, storage, and high-fidelity nanopore retrieval of 32-bit and 64-bit cryptographic keys. ZAT-DNA blocks polymerase-based copying and protects non-fungible tokens by preventing functional duplication. For larger datasets, we present a hybrid “Babel-DNA” architecture: multiple encrypted images are co-encoded in a single regular DNA pool, with each selectively decryptable only via its cognate, non-replicable ZAT-DNA key. This provides a practical framework for molecular access control, secure DNA-encoded databases, and scarce molecular tokens.

Many voltage-gated sodium channel-targeting animal peptide toxins are renowned for their potency and selectivity against insects. Understanding why these toxins selectively target insect sodium channels over their mammalian counterparts is crucial for developing safer and more effective pest control agents. Here, we present the cryoelectron microscopy (cryo-EM) structures of the insect sodium channel Na _v PaS bound to two naturally occurring insect-selective toxins, Av3 from the sea anemone and LqhαIT from the scorpion. Both toxins bind to the voltage-sensing domain 4 (VSD4) of Na _v PaS and disrupt fast inactivation by stabilizing the S4 segment in a deactivated conformation. While Av3 engages a membrane-embedded site between VSD4 and pore domain 1 (PD1), LqhαIT binds to the classical neurotoxin site 3, illustrating distinct binding modes that converge on a shared mechanism of action. These structures reveal the molecular determinants of insect selectivity and highlight the molecular coevolution of toxin-channel interactions, as corroborated by electrophysiology and toxicity assays. Leveraging these insights, we apply AI-driven protein design tools to increase the insecticidal potency of LqhαIT, resulting in a variant with a remarkable doubling in efficacy, as we confirm by insecticidal bioassays. This study illuminates the diverse mechanisms of sodium channel modulation and provides a framework for the structure-guided, AI-driven design of toxin-based biopesticides.