Communications Chemistry最新文献

Understanding the state-to-state atomic-level dynamics of a chemical reaction is a central topic in modern chemistry. Moving beyond the traditional mode-specific reaction dynamics studies, here we investigate the concept of site and conformer specificity by studying the reaction of the glycine molecule (H₂NCH₂COOH) with the hydroxyl (OH) radical using first-principles theory. Conformer-specific quasi-classical trajectory computations on a 30-dimensional potential energy surface reveal three distinct H-abstraction pathways targeting the different functional groups. CH₂- and NH₂-H-abstraction proceed through direct, single-step mechanisms, whereas a two-step mechanism emerges for COOH-H-abstraction, where initial dehydrogenation frequently leads to fragmentation into CO₂ and CH₂NH₂. COOH-H-abstraction is favored at low energies, while NH₂- and CH₂-H-abstraction are promoted at higher energies. The formation of the unstable H₂NCH₂COO• intermediate becomes increasingly restricted at higher collision energies due to limited interaction time. In specific reactant conformers, the simulations reveal an indirect biradical mechanism and an alternative stabilization pathway via intramolecular H transfer. Product-conformer distributions exhibit a three-step pattern of carboxyl group rearrangement-H-orientation switch, 180° rotation around the C-C axis, and their combination-during NH₂- and CH₂-H-abstraction. Structure-specific product formation arises clearly only in CH₂-H-abstraction, driven by the closed COOH conformation, whereas NH₂-H-abstraction leads to conformational diversity in the products.

Multimodal scientific reasoning remains a significant challenge for large language models (LLMs), particularly in chemistry, where problem-solving relies on symbolic diagrams, molecular structures, and structured visual data. Here, we systematically evaluate 40 proprietary and open-source multimodal LLMs, including GPT-5, o3, Gemini-2.5-Pro, and Qwen2.5-VL, on a curated benchmark of Olympiad-style chemistry questions drawn from over two decades of U.S. National Chemistry Olympiad (USNCO) exams. These questions require integrated visual and textual reasoning across diverse modalities. We find that many models struggle with modality fusion, where, in some cases, removing the image even improves accuracy, indicating misalignment in vision-language integration. Chain-of-Thought prompting consistently enhances both accuracy and visual grounding, as demonstrated through ablation studies and occlusion-based interpretability. Our results reveal critical limitations in the scientific reasoning abilities of current MLLMs, providing actionable strategies for developing more robust and interpretable multimodal systems in chemistry. This work provides a timely benchmark for measuring progress in domain-specific multimodal AI and underscores the need for further advances at the intersection of artificial intelligence and scientific reasoning.

The conversion of soluble amyloid-β peptides into fibrils is central in Alzheimer's disease. Lipids modulate amyloid-β aggregation, but whilst the mechanistic effect of individual lipid species is increasingly addressed, principles explaining their combinatorial contributions in biologically heterogenous membranes remain to be established. We used kinetic analyses to establish an inhibitory mechanism of GM1 gangliosides on the aggregation of amyloid-β variant Aβ(1-42) by which membrane-associated GM1 sequesters soluble Aβ(1-42) and retards primary nucleation. The kinetic inhibition increased in presence of the raft-enabling lipids cholesterol and sphingomyelin, although these lipids, intrinsically, catalysed primary and secondary nucleation respectively. These results decipher important trade-offs between the specific chemical properties of lipids and their general contributions to the physical state of membranes, show principles of competition, and identify low fluidity domains as key regulators of membrane-mediated Aβ(1-42) aggregation. The study thereby highlights a versatile, regulatory role of membranes in the molecular pathology of Alzheimer's disease.

Oxygen vacancies are regarded as crucial defects greatly affecting the electronic and optical properties of oxide films and devices, yet systematic studies on κ-Ga₂O₃ are still lacking. Herein, we investigate the thermodynamic, electronic, and optical properties of oxygen vacancies in κ-Ga₂O₃ using density functional theory calculations with the hybrid functional. The electronic structure reveals that oxygen vacancies create a deep donor defect in the bandgap, with defect levels and transition energies influenced by Ga atom displacement and localized electron dynamics. This interplay explains the stability of vacancies at specific sites and their connection to experimentally observed defect levels. Additionally, oxygen vacancies generate distinct absorption and electron energy loss peaks in the ultraviolet range. Our results elucidate the nature of oxygen vacancies, and offering a foundation for tuning and optimizing the electrical and optical properties of κ-Ga₂O₃ films and improving device performance through defect engineering.

Drug discovery requires understanding disease mechanisms, making the integration of multi-modal data essential. These data types, including omics, disease-associated, and pathway information, must be combined to uncover therapeutic insights. We developed iPANDDA, a computational pipeline that integrates these data through a network-based approach to predict candidate drug targets for specific diseases. We applied iPANDDA to lung squamous cell carcinoma (LUSC), a subtype of non-small cell lung cancer representing ~25% of global cases. Despite advances in cancer therapeutics, targeted treatments for LUSC remain limited, partly due to a lack of robust models to study carcinogenesis and therapeutic response. The SOX2 gene, amplified in ~50% of patients, plays a critical role in sustaining the cancer phenotype. Using iPANDDA, we identified and validated SOX2-dependent therapeutic targets. In vitro inhibition studies confirmed AKT and mTOR complexes as key monotherapy and combination therapy targets and revealed pathways for SOX2-targeted combination therapies.

Standardizing structural isomeric relationships and evaluating their distribution in chemical space remain major challenges in cheminformatics. Conventional molecular fingerprints and dimensionality reduction techniques are often sensitive to dataset size and structural complexity. Here, we introduce a molecular fingerprint, Structural Isomer Cumulative molecular fingerprint (SIC), that quantitatively captures relative structural differences among isomers with high precision. SIC consists of two variables: SIC_em, representing exact mass, and SIC_L, a cumulative descriptor derived from substructural differences. SIC_L enables calculation of relative structural distances within isomeric groups regardless of dataset size or molecular complexity. Using SIC, we successfully quantified structural differences across positional, skeletal, and functional group isomers, which were not adequately captured by existing descriptors. Furthermore, a scatter plot of SIC_em and SIC_L visualized metabolite distributions among cellular compartments, and nine endogenous metabolites were identified whose structural characteristics suggest potential toxicity.

In recent years, industrial biocatalysis has significantly advanced, largely due to innovations in DNA sequencing, bioinformatics, and protein engineering. However, the challenge of implementing biocatalysis at an industrial scale while ensuring sustainability and cost-effectiveness remains a critical barrier. This study presents the development of a flash thermal racemization protocol for chemoenzymatic dynamic kinetic resolution (FTR-CE-DKR) of chiral amines, encompassing an investigation of substrate scope, catalyst screening, and optimization studies. The outcomes of this research facilitated the successful scale-up of an industrially relevant amide within a recycle-batch platform, achieving unprecedented scales of up to 100 grams and space-time yield (STY) values of up to 73.2 g L⁻¹ h⁻¹. Furthermore, the process exhibited very favorable sustainability metrics when benchmarked against previous reports, including atom economy, reaction mass efficiency, and process mass intensity. These findings represent a significant milestone in the biocatalytic production of optically active amines.

Alkynes play a crucial role in chemical synthesis, bio-imaging, and drug design. Despite their significance, the intermolecular interactions involving alkynes have been largely unexplored. In this work, we unveil the previously overlooked alkyne-π interaction by comparing two zirconocene metal-organic cage compounds. The distinct stacking geometry in the single-crystal structures, coupling with the changes in C ≡ C vibrational signals, confirms the alkyne-π interaction as a genuine intermolecular interaction. Combining with computational studies, we reveal that alkyne-π interactions exert a substantial influence on the spectroscopic properties, despite being energetically less potent than π-π interactions. Our findings extend beyond theoretical implications. A comprehensive survey of the Cambridge Crystallographic Data Centre (CCDC) database corroborates the occurrence of alkyne-π interactions across hundreds of crystal structures, which provides a missing piece for fundamentally rationalizing their properties. Meanwhile, the changing C ≡ C vibrational signals, under alkyne-π interactions, may provide strategies for improving bio-imaging resolutions. It could also serve as a signature for desired alkyne-containing supramolecular structures. These results highlight the potential of alkyne-π interactions in designing functional materials for advanced applications in chemistry and biology.

Acid-sensing ion channels (ASICs) are proton-gated cation channels that detect and signal increases in proton concentration. ASIC1a and ASIC3 play a role in pain sensation associated with extracellular acidification. There are few selective modulators of ASIC3, including the tetrapeptide RFamide RPRFa, which slows the acute desensitization of ASIC3. Here we describe the peptide WRPRFa as the most potent ASIC3 activator to date and a more effective pharmacological tool. WRPRFa enhances the pH sensitivity of ASIC3 and effectively removes acute desensitization. Additionally, we demonstrate that ASIC3 can undergo tachyphylaxis at very acidic pH, which is accelerated by WRPRFa. Our work characterizes a selective and effective in vitro tool to study the interaction of RFamides and ASICs, and by extension gating mechanisms of ASIC3.

Quantum dots (QDs) function as photon sensitizers in photoelectrocatalysis (PEC), enhancing the ability of bulk materials to harness a broad spectrum of photon energy. Through precise engineering, QDs facilitate the development of advanced strategies to synthesize high-performance photoelectrodes that improve the efficiency of light-driven technologies. This review highlights valuable insights in integrating QDs into PEC systems, focusing on heterojunction-mediated charge transfer. We explore their unique optoelectronic properties, the enhancement of conventional photoanodes and photocathodes, and strategies to optimize interfacial charge transfer dynamics for efficient photon-to-energy conversion. Finally, we discuss the advantages, limitations, and future prospects of QD-based PEC technology.