The biosynthesis of ribosomally synthesized and posttranslationally modified peptides (RiPPs) leverages iterative catalysis to enhance structural and biological diversity. Traditionally, iterative enzymes install posttranslational modifications on linear peptides, rather than mature RiPPs with intricate three-dimensional structures, which require complex changes in substrate binding. Here we present a prolific class of GCN5-related N-acetyltransferases (GNATs) that iteratively and consecutively acylate two Lys residues within the loop and ring motifs of lasso peptides, diverging from the typical iterative modification of linear peptides. Utilizing high-resolution cryogenic-electron microscopy and enzymatic reconstitution, we define the lasso peptide-binding pocket of IatT and pinpoint key residues that distinguish its two distinct acetylation steps. Structure-based engineering of IatT’s acetyl-recognition site expands the cavity to accommodate longer-chain acyl groups, enabling the creation of lipolasso peptides, a class of ribosomal lipopeptide. This engineering strategy can be applied to any RiPP biosynthetic gene cluster encoding GNAT, facilitating the efficient diversification of ribosomal lipopeptides. Traditional iterative catalysis in RiPP biosynthesis typically modifies linear peptides. Here the authors discover GNAT enzymes that iteratively modify fully folded lasso peptides and engineer them to create lipolasso peptides with expanded chemical diversity.

Plasmids are mobile genetic elements that can rapidly spread across bacterial populations, promoting the dissemination of antimicrobial resistance (AMR) genes in bacteria. They are enriched in insertion sequences (IS), which are small transposable elements able to translocate between genetic locations. Here we combined experimental, bioinformatic and computational approaches to investigate the role of plasmids promoting AMR through IS-mediated gene inactivation. We find that plasmid pOXA-48, which encodes two IS1 elements, increases the rate of resistance acquisition to multiple antibiotics in clinical strains of Klebsiella pneumoniae through IS1-mediated gene disruption. Screening of genome databases confirmed that the inactivation of genes through plasmid-encoded IS elements is an extended mechanism of AMR evolution. Both our experiments and computational model revealed that conjugative plasmids can promote this route of AMR acquisition while invading complex bacterial communities. Overall, we show that conjugative plasmids contribute to AMR not only through the dissemination of resistance genes, but also through IS-mediated gene inactivation. Inactivation of chromosomal genes through plasmid-encoded IS elements is an extended mechanism of antimicrobial resistance evolution in bacteria.

Modern agriculture requires rapid, affordable tools to monitor crops and soils directly in the field. Cellulose, the structural polymer of plants, is emerging as a versatile foundation for lightweight, biodegradable sensors that measure nutrients, moisture, stress, and disease without laboratory infrastructure. Spanning simple paper assays to flexible wearables, these platforms enable distributed, real-time insight into plant and soil health. As materials engineering converges with digital connectivity, cellulose-based sensors could accelerate the transition toward more data-driven, adaptive, and sustainable agricultural systems. Agriculture is turning to precision application of agrochemicals and growth conditions, but to achieve this affordable tool to monitor crops and soils in field are needed. Here, the authors review the use of cellulose-based materials to make flexible wearable sensors that allow for in situ monitoring of agricultural conditions.

Large Language Models (LLMs) are increasingly applied to genomic tasks, yet core challenges remain concerning tokenization, evaluation, and data scarcity. This study focuses on promoter classification and systematically evaluates four tokenization methods: non-overlapping 6-mer, overlapping 6-mer, Byte Pair Encoding (BPE), and WordPiece (WPC). We show that the commonly used k-mer approach, specifically the non-overlapping variant, outperforms BPE and WPC across eight organisms, challenging assumptions derived from natural language processing. To ensure robustness, we evaluated performance under two distinct negative data strategies: positive-promoter-shuffled and random-non-promoter-fragments. Using a positional SHAP framework, we demonstrate that the model learns biologically plausible positional patterns rather than exploiting artifacts from these negative data generation processes. Furthermore, evolutionary-informed transfer learning experiments and external validation on an unseen organism reveal that training on phylogenetically related species significantly improves performance, particularly in low-data regimes. These findings underscore the significant impact of tokenization and negative data design, providing practical guidance for refining genomic classifiers.

Microbial communities are structured through complex interactions that are difficult to observe directly. Co-occurrence networks offer a way to infer community structure, revealing (not exclusively) potential biotic interactions. Such networks have been inferred for diverse biomes and repeatedly found to be modular, yet the ecological significance of this modularity remains underexplored. We tested whether clusters within co-occurrence networks (“cohorts”), are universal and ecologically meaningful units by assessing their ubiquity, stability, and environmental specificity across diverse ecosystems. Our meta-analysis spans 25 previously published 16S rRNA gene amplicon sequencing datasets (14 160 samples) and covers high environmental variability ranging from aquatic, terrestrial to anthropogenic environments. Microbial co-occurrence networks consistently exhibited high modularity across biomes. Inferred cohorts were ubiquitous and represented up to 90% of the community composition. Our findings demonstrate that modularity is a fundamental and generalizable feature of microbial community organization, indicating the existence of stable subcommunities. Highly similar cohorts were inferred even across different, unconnected environments and datasets, and showed consistent responses to environmental gradients, indicating that their composition is to a large degree deterministic and predictable. The overall cohort structure and environmental preferences were independent of the sample size and the inference algorithm, underlining the robustness and applicability of the results. Recognizing these microbial cohorts as a meaningful level of microbial organization will refine microbial community ecology, cultivation strategies, and predictive modelling of microbial dynamics.

Expanding the genetic code has revolutionized our ability to study and manipulate biological systems through site-specific incorporation of noncanonical amino acids (ncAAs). However, current methods are primarily limited to single-type ncAA incorporation in mammalian cells owing to translation inefficiency. Here we introduce a multi-type rare codon recoding strategy that addresses this limitation. By systematically evaluating and repurposing rare codons, alongside engineering mutually orthogonal aminoacyl-tRNA synthetase/tRNA pairs, we achieve the expression of proteins containing two or three distinct ncAAs at site-specific positions with recoding rates of up to 90% at wild-type protein expression levels in mammalian cells. This approach facilitates a broad range of applications, including dual bioorthogonal labelling and sequential protein activation. We further demonstrate the utility of this strategy by incorporating up to five distinct ncAAs into a single protein, revealing a redefinable nature of the genetic code and opening unprecedented avenues for future applications in biomedicine and synthetic biology. The site-specific incorporation of noncanonical amino acids (ncAAs) has so far been limited to single-type ncAA incorporation in mammalian cells. Now, the repurposing of rare codons and engineering of mutually orthogonal aminoacyl-tRNA synthetase/tRNA pairs enable up to five distinct ncAAs in a single protein, which can be applied to study mammalian pathways of interest.

Experimental evolution has been a useful tool for investigating long-term temporal evolutionary dynamics and molecular mechanisms underlying adaptation. However, extracting fundamental principles and predictive features of evolutionary outcomes from these datasets remains challenging. Here we sought to circumvent these challenges by comparing distant yeast species that share several evolutionary features but differ in evolutionary history and genome architecture, that is Saccharomyces cerevisiae and Schizosaccharomyces pombe. We evolved ten populations of the fission yeast for 10,000 generations in the same conditions as a pre-existing budding yeast dataset, allowing us to observe repeatable evolutionary outcomes within species but diverse molecular targets of adaptation across species. The most frequent route of adaptation was through changes in carbon flux metabolism, which was previously unseen in S. cerevisiae evolved populations, but similar evolutionary paths have been observed in wild populations. This suggests that parallelism is pervasive and that mechanisms of adaptation can be shared among closely related or distant species. Despite similar gene content and identical environments, recurrent adaptation across S. pombe populations involved different genes than in S. cerevisiae and was detectable mostly at the transcriptomic level. This indicates that trans-regulatory effects and contingency may contribute to differences in evolutionary outcomes between these species. Experimental evolution of fission yeast (Schizosaccharomyces pombe) in the same environment as a previous experiment with budding yeast (Saccharomyces cerevisiae) reveals parallel evolution but distinct molecular mechanisms and targets of adaptation in the two species.

Protein structure serves as a critical bridge between sequence and functional annotation, particularly in establishing functional links among distantly homologous proteins with low sequence similarities. However, systematic protein structure-based functional annotations have been lacking in plants, where functions for a significant portion of the proteomes are still elusive. In this study, we leveraged protein structural data from 17 angiosperms to uncover previously unannotated protein functions in plants. After structural clustering, we used the plant clusters to query the UniProtKB/Swiss-Prot database (the expertly curated component of UniProtKB), a repository of expertly curated and reliably annotated proteins, and identified structural matches for thousands of plant clusters that were undetectable by sequence-based BLAST searches. We further selected 120 clusters, which are highly reliable in structural quality and alignment and are well-conserved across plant species, and uncovered various protein functions that are potentially widely important in plants. Finally, we experimentally analyzed one plant cluster structurally resembling the yeast peroxisomal peroxin 8 (PEX8) protein and verified that plant PEX8-like proteins can functionally complement yeast pex8 mutants. Our findings highlight the power of structural comparison in uncovering protein functions in plants.

A device capable of sampling natural gas under aseptic conditions and in complete safety has been deployed along the transmission grid for the first time. Microbial endospores, resilient enough to survive the extreme conditions of gas transmission and storage, have been detected and isolated throughout high-pressure pipelines and underground reservoirs. In four underground gas storage (UGS) facilities, three in deep aquifers and one in a depleted reservoir, endospores of the same hydrogenotrophic bacterial species from the family Peptococcaceae have been identified, sometimes separated by hundreds of kilometers, and at two different points in the pipeline network. Cultural and genomic analyses show these bacteria can perform acetogenesis, biofilm formation, and produce formate. Hidden within pipelines, these microbes survive long journeys and actively participate in biogeochemical cycles in UGS. Several recent studies on dihydrogen injection into deep aquifers have shown the ubiquity of bacteria similar to these, responsible for formate formation through modified acetogenesis. This formate can serve as a carbon source or inhibit sulfate reduction at high concentrations. Understanding their role offers critical insights into microbial life in the deep biosphere and the potential impacts of future dihydrogen injection into natural gas systems. Their ability to thrive in extreme environments makes these microbes key players in the evolving landscape of underground energy storage and transport.