Domain insertion is an established method to engineer ligand-mediated control of activity in protein scaffolds. Whether this strategy can be systematically applied to large, structured RNAs remains unclear. In this study, we investigated the feasibility of engineering ligand-activated splicing ribozymes (LASRs) from group I catalytic introns. Using domain-insertion profiling coupled with high-throughput screening, we mapped the nucleotide-resolution landscape of aptamer insertion across the ribozyme and identified sites that support robust ligand-dependent control. We showed LASRs function across multiple kingdoms of life, including diverse species of bacteria and even fungi, and can be used to regulate various genetic outputs. Finally, we integrated LASRs with a genetic recorder that writes information into ribosomal RNA, enabling sequencing-based recovery of intracellular chemical signals from microbial consortia. This work establishes LASRs as an RNA-based inducible control platform for sensing diverse chemical inputs, regulating the expression of diverse genes of interest, and recording intracellular information.

Genes for CO2 fixation occur in soil microorganisms, but little is known about the pathways that are most common across ecosystem types, the organisms with these genes, where different CO2 fixation pathways are most prevalent, and the energy sources that support autotrophy across ecosystems. Here, we investigated microbial capacity for autotrophy in soils using 853 metagenomes and 201 metatranscriptomes from a wide range of terrestrial ecosystems (agricultural soils, wetlands, weathering rock). Autotrophy-associated RuBisCO (Form I and II) is widely encoded across all soils and occurs in bacteria from numerous lineages (38 phyla). RuBisCO Form IE is consistently more phylogenetically diverse in soils than in marine ecosystems, suggesting that it may have evolved to function in soil-like environments. A newly discovered deeply branching Form I RuBisCO, Form I triple prime, supports the hypothesis that Form I RuBisCO originated in anaerobic environments. Further, saturated soils harbor more, and more distinct, autotrophic microbes, many of which may use the Calvin-Benson-Bassham cycle or Wood-Ljungdahl pathway for CO2 fixation. Overall, the results indicate that autotrophy is a particularly important metabolism in deep, saturated soils and weathering rock.

Auxotrophy, the absence of biosynthetic capacity for essential metabolites, is widespread in microbes and is thought to shape interactions within communities. Auxotrophies are often treated as fixed properties of organisms; however, recent evidence indicates that auxotrophic phenotypes can depend on environmental context, thereby affecting community assembly or cross-feeding. Here, we systematically quantify how nutrient environments shape both auxotrophy and cross-feeding. Using matched sets of six amino acid auxotrophs in Escherichia coli and Bacillus subtilis, we measured monoculture and pairwise coculture growth across 40 carbon and nitrogen environments. We find that auxotrophy itself is highly environment dependent, with strains growing in a substantial fraction of amino acid-free conditions despite lacking key biosynthetic enzymes. Cross-feeding likewise varies widely across species, environments, and auxotroph pairs. Despite this variability, cross-feeding outcomes exhibit consistent patterns across species. In particular, cross-feeding growth is better predicted by auxotroph type (i.e., which amino acids the strain requires) than by environmental context. A machine-learning model recapitulates this pattern, identifying auxotroph type as the strongest predictor of cross-feeding growth, exceeding the contributions of nutrient environment, prototroph growth, and species identity. Together, these results show that environmental context reshapes both metabolic need and exchange, yet cross-feeding follows emergent patterns linked to auxotrophy. More broadly, our findings suggest that metabolic interdependence is shaped by both gene essentiality in an environmental context and intrinsic constraints of metabolic pathways, with implications for community assembly and the evolution of gene loss.

Bacteriocins are ribosomally synthesized antimicrobial peptides with promising applications in biotechnology, particularly in food preservation and animal and human health. Circular bacteriocins are especially attractive due to their head-to-tail cyclized structure, which confers enhanced stability and antimicrobial potency relative to linear peptides. Here, we report an in vitro cell-free protein synthesis system coupled with an enhanced Split Intein-Mediated Ligation platform (IV-CFPS/SIML) for the efficient synthesis of circular bacteriocins through systematic evaluation of cyclization sites and alternative split inteins. Using enterocin AS-48 as a model, we systematically evaluated multiple serine-based cyclization sites in combination with three split inteins, NpuDnaE, Gp41-1, and SspGyrB, to identify configurations supporting efficient splicing and high antimicrobial activity. Gp41-1 emerged as the most effective intein and was subsequently applied to the production of garvicin ML, amylocyclicin, and 27 naturally occurring sequence variants, demonstrating that cyclization site selection, intein identity, and minor sequence variations strongly influence antimicrobial potency and target range. Finally, SIML expression cassettes encoded in pUC-derived vectors enabled in vivo production and functional expression of selected circular bacteriocins in recombinant E. coli. Collectively, these results establish SIML as a versatile platform for in vitro and in vivo production, screening, and functional characterization of known and putative circular bacteriocins.

Strain-resolved metagenomics characterizes microbial communities at nucleotide-level resolution, enabling researchers to differentiate identical from closely related organisms and characterize population structure and gene content variation. Here we introduce ZipStrain, a program that performs highly accurate strain-resolved metagenomics over 500 times faster than available methods while offering superior RAM management. Applied to a dataset of 2,754 samples spanning human populations, we identify a strain-sharing gradient across social relationships, reveal striking variation in clonal structure across bacteria and bacteriophage, and pinpoint genes whose nucleotide identity deviates from genome-wide expectations. ZipStrain is distributed as an open-source Python package and accompanying Nextflow pipeline at https://github.com/OlmLab/ZipStrain.

Flagella are large transenvelope nanomachines but how they transit the peptidoglycan in Gram positive bacteria is poorly understood. A recent model suggested that flagellar basal bodies diffuse in the membrane and become captured at locations in the peptidoglycan with a pore diameter that could accommodate the axle-like flagellar rod. Mutation of penicillin binding protein 1 (PBP1/PonA), a cell wall repair protein thought to decrease peptidoglycan pore frequency and/or size, resulted in a severe growth defect and cell lysis in the ancestral strain of Bacillus subtilis that was dependent on flagellar synthesis. Genetic analysis indicated that toxicity was due to completion of the flagellar hook, which activated the flagellar sigma factor SigD. SigD, in turn, activated a suite of peptidoglycan hydrolases that caused cellular lysis when PBP1 was absent. In addition, mutations that resulted in high levels of the stress response factor Spx could lessen the toxicity, while PBPX, a putative teichoic acid D-alanylase, was required for autolysis. In sum our results indicate that flagellar synthesis, not normally associated with cell viability, causes cell wall stress and under some conditions, cell death. Moreover, our work indicates that cost of envelope integrity by flagellar synthesis may be underappreciated due to strain domestication, and suggests that specialized systems may compensate for the cost of assembly of transenvelope machines in general.

β-galactosidases (BGs) are essential enzymes widely used in the food industry, particularly in the production of lactose-free products. Among them, the BG from Aspergillus oryzae is of industrial relevance due to its activity at acidic pH and moderate thermal tolerance. However, enhancing its catalytic performance remains a key challenge. Traditional enzyme engineering methods are time-consuming and resource-intensive, limiting their scalability. Recent advances in Artificial Intelligence (AI), particularly those based on Natural Language Processing, offer a promising alternative by enabling efficient exploration of protein sequence space and prediction of beneficial mutations. In this study, we introduce an ensemble-based, zero-shot Protein Language Model pipeline that reconciles predictions from six independent models (ESM2 and the five ESM1v variants) combined with a diversity-aware candidate selection strategy. Applied to the BG from A. oryzae, this approach identified beneficial mutations leading to novel enzyme variants with up to a four-fold increase in catalytic efficiency on oNPGal, a two-fold increase on lactose, and, independently, a T338I variant with markedly enhanced thermostability (≈80% residual activity after 24 h at 60 °C), all without requiring supervised fine-tuning on experimental fitness data. Our results demonstrate that consensus across an ensemble of PLMs can efficiently enrich beneficial substitutions in industrially relevant enzymes and substantially reduce the number of wet-lab candidates that need to be screened.