Atomically precise metal clusters, characterized by their well-defined structures, have emerged as a versatile platform for energy, catalysis, and biomedicine. Building upon this foundation, the biocatalytic clusterzymes, a class of artificial enzymes with atomic-level programmable activity and renal-excreted properties, have successfully overcome the stability limitations of natural enzymes and biosafety concerns of conventional nanomaterials. This review systematically examines the synthesis, engineering principles, and applications of this programmable platform. First an in-depth analysis of the strategies is provided for programming biocatalytic or enzyme-like activity of metal clusters via atomic and ligand engineering. Meanwhile, infrared emissive metal clusters with tunable electronic structure and optical properties at the atomic level allow to achieve the pathological progression and clinical 3D visualization in deep tissue. Furthermore, semiconductor gold clusters with rich electron carriers can enhance the interface charge transfer between the metal electrode and surface molecular clusters, achieving highly sensitive neuron recording for an efficient brain computer interface. The clusters demonstrate great potential in neuroscience, including neuroinflammation, bioimaging, and neuromodulation. Finally, future challenges are outlined for the rational design and translational development of this programmable platform, poised to address complex challenges in biomedicine.

Abstract Natural enzymes serve as highly efficient biocatalysts yet exhibit limited stability under physiological conditions, including sensitivity to body temperature, narrow pH tolerance, and vulnerability to protease degradation, alongside high preparation and purification costs. Artificial biocatalysts, as highly efficient biocalysts, offer a promising alternative. However, their catalytic activity and selectivity remain inferior to those of natural enzymes, hindering clinical translation. To overcome these limitations, intelligent design strategies have been developed to achieve precise control over the structure and function of artificial biocatalysts. This review summarizes recent progress in the intelligent design of artificial enzymes for biomedical applications. Three major strategies are discussed including biomimetic structural engineering based on atomic and ligand modulation, environment‐responsive design that utilizes endogenous pathological cues or exogenous physical stimuli, and artificial intelligence‐assisted rational design incorporating machine learning and high‐throughput computation. The application of these strategies in treating chronic diseases, including metabolic disorders, neurological diseases, tumors, and infections, is also reviewed. Finally, current challenges and future directions are discussed to inform the rational development and clinical translation of high‐performance artificial enzymes.

Type 2 diabetes mellitus is the most prevalent disease in the world, with one-tenth of the population suffering from the disease, and the most critical challenges are its complications that induce high disability and mortality rates. The state-of-the-art therapeutic agents can manage glucose but fail to prevent renal failure as well as neurodegeneration with immunosuppression. Herein, we developed a deep learning design strategy that exploits the 'size-fitting effect' to engineer an atomic-precision metal cluster for preventing diabetic complications by targeting metabolic abnormality and immunosuppression. The designed AuZn cluster achieves almost 100% alpha-amylase inhibition and 88% alpha-glucosidase inhibition, resulting in the normalized glycated hemoglobin and sustained glucose control. The intrinsic redox properties reduce oxidative stress damage, promoting beta-cell regeneration and metabolic stress alleviation. Consequently, the renal function, the most prevalent complications, shows that glomerular filtration can be restored to normal levels without urinary protein, while the clinical dulaglutide is not show any improvement. The key marker during early neurocognitive disorders, the amyloid precursor protein (APP) induced by complications, can be effectively suppressed, and diabetes induced organelle degeneration in neurons can be restored.

The prevalence of persistent organic pollutants in water bodies demands advanced oxidation processes that are both highly efficient and environmentally sustainable. Single-atom manganese catalysts can enable the green activation of peracetic acid (PAA), a promising alternative oxidant, but the catalytic performance is often limited by the inherent chemical inertness of the Mn sites. In this study, we modulated the electronic structure of the coordination microenvironment of ecofriendly and low-toxicity manganese single-atom catalysts through O doping. This strategy enhanced the electron delocalization ability of tricoordinated MnN2O, established a built-in electric field to strengthen the Mn-O covalency, and significantly boosted its intrinsic mechanism for activating PAA to selectively generate singlet oxygen (1O2). The remarkable Fenton-like performance of MnN2O was reflected by an 8.8-9.9 fold improvement in the bisphenol A (BPA) degradation kinetic rate. Moreover, the practical application potential of this proposed Fenton-like process is enhanced by the average BPA removal rate of 97.96% ± 2.65% for tap water and 95.46% ± 4.96% for the secondary effluent of a sewage treatment plant over 120 h in a continuous-flow device. Density functional theory calculations elucidated that the tricoordinated structure and electron delocalization of MnN2O can effectively optimize the d-band electronic structure of adjacent Mn centers and promote the formation of 1O2 through a Mn-O covalence-dependent mechanism. This study breaks the symmetric coordination constraint to optimize the electron distribution, unlocking broad avenues for designing PAA-based Fenton-like process catalysts.

Artificial enzymes offer enhanced stability and availability over natural enzymes, showing great potential in biomedicine. However, non-oxidoreductases such as artificial hydrolases display inherently low catalytic activity and selectivity, which substantially constrains their effectiveness in regulating metabolic diseases. Here, we developed a machine learning approach using persistent homology to rapidly predict hydrolases, including lipases and cholinesterases. By training on catalytic reaction energy barriers for 133,885 molecules, we identified a series of atomic-precision molecular clusters with high catalytic activity as artificial lipases and cholinesterases. Experimental validation of 153 metal clusters confirmed that iron-gold clusters exhibited 11-fold and 2-fold higher catalytic activities than natural lipase and butyrylcholinesterase, respectively, reducing lipid droplets by 75%. Applied to a model of nonalcoholic steatohepatitis, a common lipid metabolism disorder, the iron-gold clusters significantly alleviated hepatic inflammation and restored liver function by reducing type 1 conventional dendritic cells (cDC1) and CD8+ T cells, while synergistically regulating lipid metabolism by targeting the EHHADH gene, a key factor in fatty acid beta-oxidation.

Building high-performance, multifunctional, and low-cost wearable sensors is of great significance. In this work, utilizing a simple cyclic immersing coating method, we dramatically improved the conductivity of melamine foam with pen ink. By further spraying coating using polydimethylsiloxane solution, the modified foam becomes both conductive and hydrophobic, making it suitable for using as a pressure sensor with water proof property. Significantly, the sensor is capable of effortlessly discerning various pressure attributes, such as intensity and frequency, and also boasts remarkable performance metrics, including a rapid response time of approximately 28 ms and sustained stability across 5000 cycles. Moreover, the rate of resistance change measured by the sensor is linearly correlated with the compression ratio of the foam with a gauge factor of 1.12, allowing us to obtain data on the compression ratio by testing the rate of resistance change. Importantly, with the assistance of deeplearning models, the sensor can recognize different hand gestures with nearly 95 % precision. This research offers novel perspectives on the development of versatile and economically viable wearable sensors for detecting pressure and recognizing human movements, demonstrating significant potential for using in assisting individuals with disabilities, facilitating non-verbal communication, and enhancing human-machine interactions among other areas.

Abstract Inflammatory bowel disease (IBD), with rising global incidence and disease burden, is a multifactorial disorder characterized by chronic and relapsing gastrointestinal inflammation. Its pathogenesis involves immune dysregulation, reactive oxygen species bursts, and microbiota imbalance. Intestinal immune dysregulation drives IBD progression, making precise immune regulation essential for controlling inflammation and improving long‐term outcomes. Despite advances in clinical therapeutics such as corticosteroids and biologics, durable remission remains difficult due to limited targeting and side effects. Nanozymes have emerged as a promising treatment strategy for IBD owing to their structural stability, tunable catalytic performance, and capacity to regulate oxidative stress and inflammation. Deep learning (DL)‐assisted nanozyme engineering facilitates precise matching of catalytic activity and environmental responsiveness to the complex and dynamic inflammatory microenvironment of IBD, thereby enhancing immune regulation and therapeutic efficacy. In this review, we summarize recent advances in nanozyme‐based strategies for IBD, with an emphasis on DL‐assisted design and functional optimization. We discuss current challenges and future perspectives for nanozyme therapeutics, including biosafety evaluation and clinical translation toward precision management of IBD.

Succinate is a critical intermediate for the tricarboxylic acid cycle, whose abnormal accumulation can disrupt energy homeostasis and trigger systemic inflammatory injury across dominant metabolic organs. Targeting succinate thus represents a pivotal strategy to reestablish metabolic equilibrium. Here, through biomimetic electronic structure engineering, we report an atomically precise Pt3Cu2 cluster engineered with intrinsic succinate dehydrogenase (SDH)-mimicking activity that catalyzes succinate oxidation to restore energy metabolism. The Pt3Cu2 cluster exhibits a succinate-binding affinity 4.52-fold superior to that of native SDH, wherein the Pt─Cu dual-metal active center structurally and functionally recapitulates the Fe─S catalytic motifs of SDH, enabling precise substrate recognition and efficient electron transfer. In disease models of energy metabolic dysfunction, Pt3Cu2 reduces succinate accumulation by 43.54% and restores ATP production by 5.31-fold, effectively rescuing hepatic energy metabolic dysfunction. Concurrently, it decreases lipid accumulation by 79.82% and resolves hepatic inflammation through normalization of the PI3K/Akt signaling axis. Beyond the liver, systemic normalization of succinate and inflammatory cytokines attenuates neuroinflammation, restores cerebral energy supply, and improves cognitive function. This work establishes atomically precise metal nanoclusters as a compelling enzyme-mimetic strategy for targeting metabolic-inflammatory diseases.

Sepsis, a life-threatening condition caused by a dysregulated immune response to infection, leads to systemic inflammation, immune dysfunction, and multiorgan damage. Various oxidoreductases play a very important role in balancing oxidative stress and modulating the immune response, but they are stored inconveniently, environmentally unstable, and expensive. Herein, we develop multifunctional artificial enzymes, CeO2 and Au/CeO2 nanozymes, exhibiting five distinct enzyme-like activities, namely, superoxide dismutase, catalase, glutathione peroxidase, peroxidase, and oxidase. These artificial enzymes have been used for the biocatalytic treatment of sepsis via inhibiting inflammation and modulating immune responses. These nanozymes significantly reduce reactive oxygen species and proinflammatory cytokines, achieving multiorgan protection. Notably, CeO2 and Au/CeO2 nanozymes with enzyme-mimicking activities can be particularly effective in restoring immunosuppression and maintaining homeostasis. The redox nanozyme offers a promising dual-protective strategy against sepsis-induced inflammation and organ dysfunction, paving the way for biocatalytic-based immunotherapies for sepsis and related inflammatory diseases.

Two-dimensional MXene materials exhibit significant promise in hydrogen separation membranes, especially for the effective separation of H2/CO2 mixtures, thanks to their structured interlayer channels and profusion of surface functional groups. However, in humid or watery conditions, MXene membranes are susceptible to interlayer swelling, which significantly impairs their ability to separate materials. In order to improve the hydrothermal stability of MXene membranes for effective H2/CO2 separation, we used polyvinyl alcohol (PVA) to intercalate between the MXene nanosheets and cross-link them via heat treatment. The method of heat crosslinking makes it easier for the functional groups of PVA and MXene nanosheets to connect, giving the membrane superior molecular sieving capabilities. Additionally, PVA improves MXene's interlayer spacing, resulting in quicker gas transport paths. MXene's size-selective characteristics and surface adsorption capability efficiently inhibit CO2 diffusion while facilitating H2 permeation. The resultant 20 % PMM composite membrane demonstrates a H2 permeance of 1453 GPU and a H2/CO2 selectivity of 45, sustaining steady separation performance during 72 h of operation, including 20 h in a humid environment.

Heterogeneous catalyst systems hold significant potential for advanced water treatment, yet achieving sustainable catalytic processes capable of continuously generating reactive species remains a substantial challenge. In this work, we develop an integrated oxidation-reduction system that synergistically couples peracetic acid (PAA) with H2O2 under the guidance of an interfacial built-in electric field (BEF). Through a programmable self-assembly approach, a porous nitrogen-doped carbon (NC) layer encapsulating Co/CoO heterojunction was constructed. Experimental and theoretical results confirm that strong electronic coupling between metallic Co and semiconductor CoO spontaneously generates a robust BEF. This field not only optimizes the electronic configuration to enhance PAA adsorption and activation, but also enables the selective adsorption of H2O2 from the mixed oxidant solution. The adsorbed H2O2 acts as an electron donor to sustain the Co(II)/Co(III) redox cycle, facilitating continuous reactive oxygen species (ROS) generation for approximately 120 min. The system demonstrates exceptional catalytic performance, achieving high contaminant removal rate constants (0.3 to 0.6 min-1) with an ultralow catalyst dosage of 15 mg L-1 and significantly improved PAA utilization efficiency. This BEF-mediated "dual-enhancement" strategy offers a sustainable and efficient route for water purification by enabling high oxidant utilization while minimizing catalyst consumption.

Owing to their nanoscale dimensions, well-defined atomic structure, and elevated specific surface area, clusters have emerged as a novel therapeutic platform for neurological disorders. However, efficiently and rationally designing functionalized clusters capable of specifically recognizing and modulating key disease targets remains a major difficulty. The rapid advancement of artificial intelligence (AI) technology offers a revolutionary solution to this bottleneck. By integrating deep learning, generative models, and multi-omics big data, AI can mine vast amounts of biomedical and chemical information with unprecedented speed and precision. It will drive transformative innovations in rational design of clusters, precise modulation of enzymatic activity, and high-throughput screening of therapeutic targets for neurological disorders.

DNA origami technology has revolutionized nanofabrication by enabling the evolution from static nanostructures to programmable dynamic systems. Herein, we present a bioinspired cloverleaf DNA origami system enabling structural reconfiguration, reversible molecular detection, and programmable higher-order assembly. Two basic cloverleaf related DNA origami structures were designed: a flexible four-leaf structure (Leaf4) and a stable four-leaf clover (Lucky4) stabilized by central linker staples, as confirmed by atomic force microscopy (AFM) and simulations (tacoxDNA, CanDo). Leaf4 transitions to a closed doughnut topology via apex strand replacement, mimicking water lily closure dynamics. Lucky4 serves as a sensing platform: its blades are functionalized with split capture strands, enabling target sequence binding visualized by AFM and subsequent toehold-mediated release. Extending this strategy, two Lucky4 units assemble into sandwich structures upon target addition, exhibiting double the height of single units. This reconfigurable system integrates biomimetic nanomechanics with visual molecular detection and modular nanostructure assembly.

Multifunctional composite fabrics combining exceptional water repellency with synergistic performance benefits are emerging as promising candidates for next-generation applications. This study reports a rationally designed superhydrophobic cotton fabric with robust stability and high conductivity, fabricated through sequential layer-by-layer deposition of poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS), pen ink, polydimethylsiloxane (PDMS), and silica nanoparticles (SiO2). The developed fabric achieves superhydrophobicity (water contact angle 155.3(o) +/- 3.0(o)), enhanced electrical conductivity, photothermal properties, UV shielding capability (UV blocking rate > 99.993 %), and temperature sensing functionality. These multifunctional properties demonstrate quantifiable performance in self-cleaning, Joule heating, solar-assisted water evaporation, UV blocking, and temperature detection. Notably, the fabric maintains superhydrophobicity under harsh conditions including acidic/alkaline solutions, UV irradiation, and water impact. Furthermore, the modified fabric exhibits excellent stability after repeated electrical heating cycles. The multifunctional synergy on cotton fabric establishes a viable approach for engineering multifunctional superhydrophobic textiles with tailored performance, addressing the growing demands for advanced materials in intelligent fabric innovation.

DNA nanotechnology has created a wide variety of nanostructures that provide a reliable platform for nanofabrication and DNA computing. However, constructing programmable finite arrays that allow for easy pre-functionalization remains challenge. We aim to create more standardized and controllable DNA origami components, which could be assembled into finite-scale and more diverse superstructures driven by instruction sets. In this work, we designed and implemented DNA origami building block pieces (DOBPs) with eight mutually independent programmable edges and formulated DNA instructions that tailored such components. This system enables DOBPs to be assembled into one or more specific 2D arrays according to the instruction sets. Theoretically, a two-unit system can generate up to 48 distinct DNA arrays. Importantly, experiments results demonstrated that DOBPs are capable of both deterministic and nondeterministic assemblies. Moreover, after examining the effects of different connection strategies and instruction implementations on the yield of the target structures, we assembled more complex 2D arrays, including limited self-assembly arrays such as 'square frames', 'windmills' and 'multiples of 3' long strips. We also demonstrated examples of Boolean logic gates 'AND' and 'XOR' computations based on these assembly arrays. The assembly system provides a model nano-structure for the research on controllable finite self-assembly and offers a more integrated approach for the storage and processing of molecular information.

Constructing high-performance wearable sensors with multifunctional properties and low cost is of great importance. In this work, a multifunctional wearable sensor with highly conductive and superhydrophobic properties was fabricated using commercially available melamine foam as cheap substrate. The sensor was prepared by simply coating the melamine foam with poly(3,4-ethylenedioxythiophene)-poly (styrene sulfonate) (PEDOT: PSS), polydimethylsiloxane (PDMS), and hydrophobic silica nanoparticles. It exhibited high conductivity and superhydrophobicity, making it suitable for use as a pressure sensor with water proof properties. SEM and XPS analysis demonstrated the relationship between the surface of the sensor as well as its interior properties and morphology. Importantly, the sensor not only can easily distinguish between different pressure features (i.e. strength and frequency), but also displays superior properties such as short response time (similar to 93 ms), high sensitivity (14.66 kPa(-1)), and long-term stability (500 cycles). Moreover, by incorporating deep-learning artificial intelligence (AI) algorithms, the sensor can recognize different human motions with high accuracy (similar to 95 %). In conclusion, this study provides a novel insight into the manufacture of multifunctional and cost-effective wearable sensors for pressure detection and human motion recognition, which shows great application potential in the fields of disabled person assistance, non-verbal communication, and human-machine interaction.

The development of programmable DNA origami architectures with combinatorial complexity remains a critical challenge in molecular nanotechnology. This study develops a programmable nucleic acid detection platform by integrating DNA origami nanostructures with molecular logic gates, advancing the field of dynamic molecular computation. Triangular DNA origami modules, designed with edge-specific hybridization sites, successfully emulate Boolean logic operations (YES, AND, and OR gates) to achieve target-driven hierarchical self-assembly. As a proof of concept, significant biomarkers for early lung cancer diagnosis, were detected as targets, demonstrating the platform’s multiplexed analytical capabilities. By coupling the programmability of DNA nanostructures with molecular recognition logic, the platform constructs autonomous systems capable of interpreting biological signals via predefined algorithms. The modular architecture supports the scalability of multi-layered logic circuits, while atomic force microscopy (AFM) provides nanoscale-resolution validation of assemblies. Toehold-mediated strand displacement enables dynamic disassembly of structures, endowing the system with resettable and adaptive feedback functionalities. This technology lays the groundwork for transformative applications in precision diagnostics, synthetic biology, and adaptive nanomedicine.

Two-dimensional WSe2 has garnered significant interest owing to its special electronic properties and high catalytic activity. However, the substrate catalytic selectivity of WSe2 has still been poorly studied. In this study, ultrasmall M-WSe2 clusters (M = Cu, Ru, Zn) were designed by a single-atom doping strategy and catalytic selectivity. The Cu-WSe2 and Zn-WSe2 clusters show high antioxidant activity, both about a 7.2-fold enhancement higher than the WSe2 clusters. Meanwhile, Ru-WSe2 and Cu-WSe2 prefer to exhibit catalase-like (CAT-like) activities, about 5.6- and 2.5-fold enhancement after single-atom doping. Ru-WSe2 displays significant peroxidase-like (POD-like) activity with a better affinity than natural HRP, showing a Km value of 0.43 mM. In contrast, pure WSe2 clusters exhibit NADH oxidase-like (NOX-like) activity that catalyzes the regeneration of NAD+. Density functional theory calculations reveal that catalytic selectivity originates from differential electron transfer around different single atoms as well as d-band center modulation. In vitro cell experiments demonstrated that the M-WSe2 clusters can modulate the oxidation-reduction balance through single-atom engineering.

Ultrasmall Au25(MPA)18 clusters show great potential in biocatalysts and bioimaging due to their well-defined, tunable structure and properties. Hence, in vivo pharmacokinetics and toxicity of Au nanoclusters (Au NCs) are very important for clinical translation, especially at high dosages. Herein, the in vivo hematological, tissue, and neurological effects following exposure to Au NCs (300 and 500 mg kg-1) were investigated, in which the concentration is 10 times higher than in therapeutic use. The biochemical and hematological parameters of the injected Au NCs were within normal limits, even at the ultrahigh level of 500 mg kg-1. Meanwhile, no histopathological changes were observed in the Au NC group, and immunofluorescence staining showed no obvious lesions in the major organs. Furthermore, real-time near-infrared-II (NIR-II) imaging showed that most of the Au25(MPA)18 and Au24Zn1(MPA)18 can be metabolized via the kidney. The results demonstrated that Au NCs exhibit good biosafety by evaluating the manifestation of toxic effects on major organs at ultrahigh doses, providing reliable data for their application in biomedicine.