Accurate characterization of multi-state protein conformations is crucial for understanding their functional mechanisms and advancing targeted therapies. Extracting coevolutionary constraints from homologous sequences helps reveal protein structure and function, which can be automatically captured by MSA Transformer leveraging attention mechanisms. Making use of the multi-conformational coevolutionary signals captured by MSA Transformer, we introduce in this study EvoSplit to disentangle coevolutionary signals associated with distinct conformations to guide protein structure predictions. EvoSplit outperforms AF-Cluster on 85 fold-switching proteins and successfully models the conformations of proteins beyond AlphaFold2’s training set. We then identify 54 candidates with potential conformational diversity for cancer-related human proteins. Notably, for five GTPases, EvoSplit consistently predicts two conformations, one of which has not been previously reported. As an important example, the protein–protein interaction analysis provides new insights into novel HRAS function-associated conformations. Furthermore, the validity of these newly identified conformations is examined by evolutionary analysis and extensive molecular dynamics simulations. Understanding multi-state protein conformations is essential for elucidating their functions and developing targeted therapies. Here, the authors introduce EvoSplit, leveraging MSA Transformer to disentangle coevolutionary signals associated with distinct conformations, outperforming AF-Cluster in modeling fold-switching proteins and identifying new conformations of GTPases and HRAS.

Liquid-liquid phase-separation (LLPS) controls protein activity and dynamically organizes (macro)molecules in living systems without the need for membrane-bound compartments. Biomolecular condensates of water-soluble proteins have extensively been studied, but little is known about LLPS of membrane proteins. In this work we induce in vivo condensation of lactose permease (LacY), a widely-studied model monomeric inner membrane protein in E. coli, and evaluate how it affects LacY function. We fused LacY with engineered, condensate-forming protein PopTag. We observe major changes in the localization and mobility of LacYPop. Molecular dynamics simulations show how the PopTag domain drives the condensate-like association dynamics of LacYPop through hydrophobic sticker interactions. LacYPop preserves native-level transport activity and outperforms the non-condensed LacY under mild hyperosmotic stress (osmotic upshift). In osmotically stressed cells, membrane-bound biomolecular condensates also reduce deformation of the cytoplasmic membrane. Perturbation experiments suggest that membrane curvature drives the accumulation of LacYPop at the poles of E. coli. Co-condensation of LacY and β-galactosidase LacZ slightly reduces their activity and results in remarkable cellular reorganization of the proteins. Our research shows the localization, dynamics, and function of phase-separated membrane proteins in bacteria and highlights the potential of LLPS for engineering complex metabolic networks in vivo. Phase-separation of the transport protein LacY alters its membrane localization and mobility, preserves its transport activity, reduces cell membrane stress under osmotic upshift, and LacY forms mixed condensates with the enzyme β-galactosidase LacZ.

RNA pseudouridylation is one of the most prevalent post-transcriptional modifications, occurring universally across all organisms. Although pseudouridines have been extensively studied in bacterial tRNAs and rRNAs, their presence and role in bacterial mRNA remain poorly characterized. Here, we used a bisulfite-based deep sequencing approach to provide a comprehensive and quantitative measurement of bacterial pseudouridines using E. coli, to provide proof of concept. We identified 1,954 high-confidence sites in 1,331 transcripts, which is 29 times above previous estimates and representing almost 30% of the transcriptome. Furthermore, pseudouridines were significantly associated with mRNA stability and enriched in transcripts associated with secondary metabolite production and adaptation to diverse environments. Finally, we mapped pseudouridines in oral microbiome samples of human subjects, demonstrating the broad applicability of our approach in complex microbiomes. This way, we observe that, although uridines are required for modification, mRNAs from GC-rich bacteria harbored more pseudouridine sites than AT-rich genomes in our dataset. Altogether, our work highlights the advantages of mapping bacterial pseudouridines and provides a tool to study posttranscription regulation in microbial communities. Pseudouridines are RNA modifications occurring in all organisms. Here, Sharma et al. identify and quantify pseudouridines in E. coli bacteria, where this modification is associated with mRNA stability, and map pseudouridines in human oral microbiome samples.

Hyperspectral imaging (HSI) is an advanced sensing modality that captures spatial and spectral information simultaneously, enabling non-invasive and label-free characterization of material, chemical and biological properties. This Primer overviews HSI, with an emphasis on the imaging workflow and data processing pipeline. We introduce the essential physical principles and sensor architectures, using Earth observation HSI systems as a representative example. Key steps in data acquisition, calibration and correction that determine data structure and quality are discussed. We summarize common hyperspectral data forms and highlight core analytical techniques, including dimensionality reduction, classification and spectral unmixing, together with emerging artificial intelligence-based methods that increasingly influence current HSI research. Six representative application fields are briefly surveyed to illustrate how HSI enables sub-visual feature extraction for quantitative interpretation and decision-making. We also discuss persistent challenges, including hardware trade-offs, acquisition variability and high-dimensional complexity alongside promising advances such as computational imaging, physics-guided modelling, cross-modal fusion and scalable learning frameworks. Best practices for dataset sharing, reproducibility and metadata documentation to support transparent research are included. Scalable, real-time and embedded HSI, enabled by sensor miniaturization and artificial intelligence, will push HSI into a cross-disciplinary platform for transformative scientific and societal impact. This Primer introduces hyperspectral imaging (HSI) through a concise, imaging-centric perspective, linking sensor platforms, data types and representative datasets across application domains. It highlights how platform characteristics shape data properties and downstream analysis, providing a unified reference for understanding and comparing HSI systems and data.

Nano-zymes have progressed from simple enzyme mimics to a versatile class of artificial catalysts that expand functionality beyond the reach of natural enzymes. In this Perspective, we examine how the materials design of nanozymes enables catalytic behaviors that are inaccessible to biological systems, including activity in non-biological environments, promotion of non-natural reactions and the integration of multiple catalytic functions within a single nanostructure. Through nanoscale control over physicochemical parameters and emulation of key enzyme architectures, nanozymes can achieve finely tunable activity and selectivity. Facile surface functionalization and inherent electrical connectivity further enhance their performance in complex systems. Collectively, these attributes extend catalysis beyond the constraints of natural enzymes, driving innovations in an array of different fields, including biomedicine, agriculture, environmental remediation and energy conversion. Nanocatalyst ‘nanozymes’ provide a versatile alternative to natural enzymes. Nanozymes can operate in conditions inimical to enzymes and can catalyse reactions that their natural analogues cannot. This Perspective discusses design principles, strengths, challenges and applications of nanozymes.