State Key Laboratory of Synthetic Chemistry

Due to its high activity, the synthesis or generation of the N-heterocyclic carbene (NHC) boryl anion remains challenging, with only limited methods available. Traditional stabilization through bulky ligands or aromatic systems restricts its reactivity as a boron synthon. In this work, we present a straightforward strategy for the in situ generation of an NHC boryl anion at room temperature through the reaction of neutral sp2-sp3 diboron reagent (NHC)BH2Bpin with KOtBu. This less sterically hindered and strongly nucleophilic boryl anion demonstrates broad reactivity with diverse electrophiles, including those that are unreactive with other types of boryl anions or traditional diboron reagents. These nucleophilic borylation reactions enable the synthesis of a diverse array of four-coordinate organoboron compounds. Detailed mechanistic studies and DFT calculations confirm the intermediacy of the NHC boryl anion and provide clear insights into the reaction pathways for the newly discovered reactivities.

This review provides a comprehensive examination of smart semiconductor nano‐bio‐interfaces. We start with fundamental aspects of semiconductor nanomaterials—including their structures, properties, and synthesis strategies. We then explore a hierarchical perspective on nano‐bio‐interfaces—from engineered surface modifications and interfacial forces to direct interfaces with biomolecules (DNA, proteins)—and extend to the complex interactions between semiconductor nanomaterials and living cells, addressing both the uptake of exogenous nanomaterials and the endogenous biosynthesis of semiconductors via synthetic biology. Finally, we analyze semiconductor‐mediated electron‐transfer processes in biological media and biodevices, which enable semiconductors to function as sensors and actuators in complex biological systems. By elucidating these principles, we highlight emerging strategies in designing and (re)constructing intelligent, bioresponsive, and energy‐active semiconductor‐bio platforms for next‐generation healthcare and biotechnology.

Anti-icing coatings have gained significant attention to prevent ice accretion on infrastructure surfaces, particularly in aviation, power, and transportation sectors, owing to their energy-saving and satisfactory anti-icing performance. However, dust accumulation on anti-icing coating surfaces significantly weakens their protective performance. Dust particulates adsorbed on the coating surfaces via electrostatic interaction is challenging to remove by wind power or rain washing alone. Herein, the dust-induced icing process is clarified, and the process is effectively suppressed by a self-constraining lubricant (SCL) coating based on ionic liquids (ILs) and zwitterionic copolymers. ILs (EMIES), embedded into the PDMS matrix, enhance the coating's conductivity (≈2.04 S/m) to dissipate surface static electricity and prevent the electrostatic adsorption of charged dust particles. The zwitterionic copolymer in SCL coating is designed to constrain ILs via electrostatic interaction and provide hydrophilic segments to enhance anti-icing as well as deicing properties (heterogeneous ice nucleation temperature of -27.9°C, icing delay time of 1458 s, and ice adhesion strength of 8.1kPa). This SCL coating presents dust-repellent performance and retains excellent anti-icing properties even when suffering from dust deposition. Meanwhile, it still maintained low ice adhesion strength after 30 icing-deicing cycles. This work establishes a new dust-repellent and anti-icing strategy for outdoor infrastructure.

Massively-multitask regression models (MMRMs) have revolutionized activity prediction for drug discovery. MMRMs trained on millions of compounds and many thousands of assays can predict bioactivity with accuracy comparable to 4-concentration IC50 experiments. This report compares six MMRMs: pQSAR, Alchemite, MT-DNN, MetaNN, Macau and IMC. Models were trained by experts in each method, on identical sets of 159 kinase and 4276 diverse ChEMBL assays, employing realistically novel training/test set splits. Results were compared both qualitatively and with statistical rigor. Our use-case is imputing full bioactivity profiles for the very sparse compound collections on which the models were trained. MMRMs performed much better than the single-task random forest regression (ST-RFR) model. Five MMRMs train all models simultaneously, so must leave out test-set measurements from all assays to avoid leakage (here 25% of data), whereas one method trains models one-at-a-time, so only holds out test data for that assay (< 1% of data). Thus, all algorithms were compared both using 75/25 splits, and when possible, 99 + / < 1 splits. Many MMRM evaluations achieved similar accuracy when tested on the same split. However, when evaluated on 75/25 splits, all MMRMs performed much worse than when evaluated on 99 + / < 1% splits. Thus, while many MMRMs produce comparable final production models (trained on all the data), models that require 75/25 splits greatly underestimate the accuracy of the final models. While outstanding for imputations, MMRMs proved little better than ST-RFR for compounds very unlike the training collection. Thus, MMRMs are best for hit-finding, off-target, promiscuity, MoA, polypharmacology or drug-repurposing within the training collection. Since accuracy is not a deciding factor, other pros and cons of each method are also described.

ABSTRACT A general palladium‐catalyzed Sonogashira cross‐coupling reaction of aryl and heteroaryl chlorides with aliphatic, aryl and heteroaryl alkynes is described. The Pd(dba)₂/ PCy 2 ‐Napdole‐phos catalyst system efficiently promotes the reaction, exhibiting broad substrate scope and excellent functional group tolerance. Notably, the catalyst loading can be down to 0.2 mol% Pd. The PCy 2 ‐Napdole‐phos , featuring an extended π‐system provides substantial steric bulk to facilitate the construction of C(sp 2 )—C(sp) bonds.

ABSTRACT Enzyme therapeutics require both catalytic activity and efficient cytosolic delivery—yet protective encapsulation typically compromises enzymatic function, while achieving cellular uptake without lysosomal degradation remains challenging. We address this with a rationally designed supramolecular adhesive photoinitiator ( Gu CD⊃BP‐SH) that unifies surface adhesion, radical initiation, and membrane translocation within a single host‐guest architecture. Guanidinium (Gu + ) motifs on a cyclodextrin scaffold ( Gu CD) enable non‐covalent adhesion to protein surfaces at carboxylate‐rich regions; the cyclodextrin cavity hosts a thiol‐benzophenone guest (BP‐SH) whose photoactivation (365 nm, 60 mW cm −2 for 30 min) initiates localized grafting‐from polymerization, constructing a semi‐permeable polymer jacket. Applied to β ‐galactosidase, this yields sub‐100 nm multi‐enzyme nanoassemblies (containing ∼10 enzymes per particle) retaining ∼30% catalytic activity with exceptional proteolytic resistance: 86% activity retained versus 25% for unprotected enzyme after Proteinase K challenge. The incorporated Gu + motifs enable efficient, energy‐independent cytosolic delivery via membrane translocation, with 91% of cells showing catalytic activity compared to 5% with non‐jacketed enzyme. This modular strategy confers protection and cell‐penetrating capability onto native biomacromolecules while maintaining catalytic function, eliminating the need for enzyme release—a persistent bottleneck in therapeutic delivery.

O-GlcNAc modification is a key cellular signal, but its role in regulating senescence-associated transcription remains poorly understood. Here, we apply a time-resolved chemical genomics strategy to map dynamic O-GlcNAc chromatin-associated proteins (OCPs) during oncogene-induced senescence (OIS) in primary human fibroblasts. Chromatin O-GlcNAc modification continues to accumulate, while 1,987 senescence-associated OCPs undergo dynamic shifts in genomic occupancy across diverse epigenetic chromatin states and display bimodal regulatory activities within the 3,466-gene senescence transcriptome. O-GlcNAc facilitates the formation of dual-function complexes: TF-SWI/SNF activates senescence-associated secretory phenotype (SASP) genes at promoters, whereas NuRD enforces the repression of cell-cycle regulators at enhancers. Furthermore, we identify O-GlcNAc modified JUN and GATAD2A as key regulators of OIS phenotypes in both in vitro and in vivo models of senescence-driven tumorigenesis. These findings reveal dynamic regulation and chromatin organization principles of O-GlcNAc-related epigenetic factors, providing insights into cellular senescence and potential therapeutic strategies.

Transcription factors drive gene expression dysregulation in cancer. However, non-canonical oncogenic mechanisms of these factors are unclear. Utilizing function-centric proteomics to discover RNA-dependent protein-protein interactions, we uncovered an unexpected interaction between transcription factor oncogene SOX2 and methionine cycle enzyme AHCY. Immunofluorescence and CUT&amp;RUN revealed that SOX2 expression sequesters AHCY to the chromatin. A candidate RNA-immunoprecipitation screen identified non-coding RNA RMST as a mediator of the SOX2 and AHCY interaction. The SOX2-AHCY interaction is reduced upon RMST knockdown. SOX2 expression sequesters RMST and AHCY in the nucleus, an activity dependent on the RNA-binding Arginine Rich Motif (ARM) domain of SOX2. Metabolite profiling revealed that SOX2 expression alters methionine cycle intermediates, particularly at the AHCY catalyzed step. Methylation precursor S-adenosylmethionine (SAM) production is also inhibited by SOX2. These metabolic changes are rescued with SOX2 ARM mutation. Whole genome bisulfite sequencing revealed that SOX2 expression induces DNA hypomethylation in cancer cells. DNA hypomethylation and its downstream DNA damage effect are rescued with SAM supplementation or SOX2 ARM mutation. These data suggest that SOX2 mis-expression in cancer sequesters AHCY, through an RMST adaptor, in the nucleus. This reduces the availability of cytoplasmic AHCY to participate in the methionine cycle, reprograming this metabolic process. As a result, SAM levels are reduced, causing DNA hypomethylation and downstream DNA damage. Our findings were validated in cancer patient biopsies. Strikingly, knockdown and pharmacological inhibition of AHCY targets SOX2-expressing cancer cells in culture and in vivo. This suggests that low SAM levels, induced by decreased cytoplasmic AHCY, sensitize SOX2-expressing cancer cells to AHCY inhibition. Overall, our results suggest that a transcription factor can coopt a non-coding RNA to perform non-canonical metabolic reprograming, creating a druggable metabolic dependency in transcription factor-driven cancer.