Modern genome editing methods permit the flexible modification of organisms at the genome level. However, bacteriophages, despite their small genomes, pose unique challenges due to the need to edit during their infection cycle, then select/screen for the modified genomes against a background of the wild type phage. Direct genome synthesis enabled by High-Complexity Golden Gate Assembly (HC-GGA) offers an alternative approach that permits rapid, accurate, and flexible genome modification. Here, we demonstrate HC-GGA’s bacteriophage engineering potential, particularly in addressing the public health challenge of detecting hazardous pathogens and nonpathogenic bacteria as indicators of fecal contamination (indicator organisms) in water supplies. A bacteriophage-based biosensor was developed by recoding the genome to enable in vivo incorporation of the alkyne-modified noncanonical amino acid L-homopropargylglycine into the capsid. The modification enabled a bio-orthogonal cycloaddition reaction with azide-conjugated magnetic nanoparticles resulting in magnetized phages which were able to bind, capture, and concentrate their host E. coli. In parallel, the engineered phage expressed luciferase during infection, allowing detection of E. coli at concentrations below 10 CFU per 100 mL in drinking water samples. The approach significantly reduces assay time and cost associated with such assays, particularly in field-based applications, thereby illustrating the practical benefits of synthetic biology in environmental monitoring and public health initiatives.

Transposon mutagenesis enables genome-wide interrogation of gene function with a single self-contained genetic construct. However, its application to non-model bacteria remains limited because transposition efficiency depends on multiple host factors that are difficult to predict a priori including transposase activity, antibiotic resistance marker performance, and regulatory element compatibility. Here, we present a scalable system to identify functional transposon configurations in non-model bacteria through pooled library screening. We selected 18 promoters across multiple bacterial phyla to independently drive expression of the transposase and antibiotic resistance marker, generating 324 promoter combinatorial variants for each of six antibiotic resistance markers. We developed a high-throughput, automated workflow to deliver all 1,944 mariner-based transposon variants in a single experiment and applied this to 92 non-model bacteria spanning multiple phyla. From this, we identified functional transposons for 43 strains, with high-level mutagenesis (102-104 unique insertions) in 13 species, including seven with no previously described transposon mutagenesis. We then expanded to a dual-transposase system, mariner or Tn5, and devised a single transposon insertion sequencing method for high-throughput screening of 3,888 configurations. To demonstrate the practical utility of our screening approach, we used a top-performing variant to generate a genome-wide transposon mutant library for Comamonas testosteroni KF-1, a bacterium that metabolizes plastic- and lignin-derived polymers. We assayed this C. testosteroni mutant library to identify enzymatic pathways, transporter genes, and regulators essential for the metabolism of plastics-associated monomer terephthalate and lignin-associated monomer 4-hydroxybenzoate. Together, this work establishes a scalable approach to construct and identify genetic perturbation systems in non-model bacteria, expanding our ability to systematically probe gene function across the bacterial tree of life.

Disruption of protein–protein interactions (PPIs) is a major mechanism of a variant's deleterious effect. Computational tools are needed to assess such variants at scale, yet existing predictors rarely consider loss of specific interactions, particularly when variants perturb binding interfaces without significantly affecting protein stability. To address this problem, we present MutPred-PPI, a graph attention network that predicts interaction-specific (edgetic) effects of missense variants by operating on AlphaFold 3-based protein complex contact graphs with protein language model embeddings imposed upon nodes. We systematically evaluated our model with stringent group cross-validation as well as benchmark data recently collected within the IGVF Consortium. MutPred-PPI outperformed all baseline methods across all evaluation criteria, achieving an AUC of 0.85 on seen proteins and 0.72 on previously unseen proteins in cross-validation, demonstrating strong generalizability despite scarce training data. To demonstrate biomedical relevance, we applied MutPred-PPI to variants from ClinVar, HGMD, COSMIC, gnomAD, and two de novo neurodevelopmental disorder-linked datasets. Disease-associated variants from ClinVar and HGMD showed strong enrichment for both quasi-null and edgetic effects, whereas population variants from gnomAD increasingly preserved interactions with higher allele frequencies. Notably, we observed a strong edgetic disruption signature in highly recurrent cancer variants from both the full COSMIC dataset and a subset of variants from oncogenes. Recurrent tumor suppressor gene variants and autism spectrum disorder-associated variants exhibited moderate quasi-null enrichment, whilst neurodevelopmental disorder-linked variants showed a weak edgetic disruption signature. These results indicate distinct PPI perturbation mechanisms across disease types and show that MutPred-PPI captures functionally relevant molecular effects of pathogenic variants.

Reengineering translation initiation provides a powerful route to develop new translation systems that enable precise control of protein synthesis. While many engineered translation systems show promise, their orthogonality and impact on host physiology is largely uncharacterized, limiting broader application. Here, we develop an initiator-tRNA with an AAC anticodon mutation (i-tRNA-AAC), enabling translation initiation at a GUU start codon. Using fluorescence assays, proteomics, and tRNA sequencing, we assess the i-tRNA-AAC initiation orthogonality and effects on the host, guiding the optimization of its translational efficiency. We find the i-tRNA-AAC mutant initiates translation exclusively from its GUU start codon and is improved by overexpression of valyl-tRNA synthetase and methionyl-tRNA formyltransferase. However, this intervention perturbs aminoacylation and base modifications of endogenous tRNAs along with proteome-wide changes likely due to increased valine demand. Our findings demonstrate how an orthogonal translation initiation system reshapes host physiology and reveals the adaptive responses that accompany translational reprogramming.

mRNA design plays a central role in synthetic biology, nucleic acid therapeutics, and vaccine development. Although large language models are applied in many biological fields, generative language models for de novo mRNA design remains largely unexplored. Here, we introduce mRNA-GPT, a series of generative mRNA language models which for the first time covers the three domains of life as pretraining datasets. Based on a GPT-2 transformer architecture with 302 million parameters, we pre-trained three separate models on 19,676 bacterial, 4,688 eukaryotic, and 702 archaeal species, leveraging 80 million, 83 million, and 2 million mRNA coding sequences, respectively. Distinct clustering of mRNA coding sequence embeddings from animals, plants, and fungi in the pretrained mRNA-GPT-eukaryote indicates that the model captures organism-specific sequence features. Following unsupervised pre-training, we fine-tuned mRNA-GPT on a translation-efficiency dataset to generate high-performance mRNA sequence. Compared to the pretrained model, the fine-tuned mRNA-GPT produced mRNA sequences with significantly higher translation efficiency scores, demonstrating the ability of mRNA-GPT to capture sequence features underlying high translation efficiency. We further fine-tuned mRNA-GPT on datasets for mRNA stability and mRNA expression, where it likewise produced high-performance mRNA sequences. Our pretrained models are publicly available, enabling other researchers to adapt mRNA-GPT to specialized tasks such as tissue-specific mRNA expression or stability by fine-tuning on their own data. Together, our study demonstrates that generative mRNA language modeling as a promising approach for accelerating mRNA design across diverse biological fields.

Intracellular biopolymers serve versatile functions, allowing microbes to adapt to fluctuating environmental conditions. The metabolic interdependence of dual biopolymers polyphosphate (polyP) and polyhydroxyalkanoates (PHA) is a defining feature of polyphosphate-accumulating organisms (PAOs), the key agents enabling enhanced biological phosphorus removal (EBPR). Beyond EBPR, PAOs accumulating both polyP and PHA (PHA–PAOs) in natural environments remain unexplored due to limited detection tools, despite hypothesized roles in nutrient cycling. This study presents a novel phenotype-targeted approach integrating triple-stained fluorescence-activated cell sorting with 16S rRNA gene sequencing (TriFlow-Seq) to detect, quantify, and phylogenetically characterize PHA–PAOs, validated by fluorescence imaging and single-cell Raman spectroscopy. TriFlow-Seq of EBPR samples confirmed known PAOs, including Candidatus Phosphoribacter, Tetrasphaera, and Ca. Accumulibacter, and revealed new candidate PHA–PAOs within the Rhodobacteraceae family. In maize rhizosphere, TriFlow-Seq led to the discovery of diverse PHA–PAOs, Pseudomonas, Halomonas, and Nannocystis. These genera participate in plant-growth-promoting functions including phosphate solubilization and phytohormone production, yet their simultaneous polyP and PHA accumulation has never been reported. Our findings revealed high PHA–PAO prevalence and distinct phylogenetic patterns associated with different maize genotypes, suggesting an overlooked role of biopolymer costorage in rhizosphere dynamics. This study establishes a pioneering approach to investigating PHA–PAO identities and roles across ecosystems.

Geogenic arsenic (As) contamination in groundwater, widely used as drinking water, poses a global health risk, yet the microbial pathways linking electron donor oxidation to the reduction of As-bearing Fe(III) oxyhydroxides remain poorly understood. Here, we deployed Fe(III) (oxyhydr)oxide-coated pumice stones in high-As, high-methane groundwater in Cambodia for 270 days to capture planktonic metal-reducing microbial communities in-situ. These were used to inoculate anaerobic microcosms with methane or volatile fatty acids (VFAs) as electron donors over 200 days. Genome-resolved metagenomics revealed that methane oxidation via reverse methanogenesis led to acetate production, which in turn provided the primary electrons for Fe(III) and As(V) reduction in the microcosms, resulting in As(III) release. Our findings highlight an indirect coupling between methane oxidation and arsenic mobilization, with acetate as the key intermediate. This study offers new insights into the role of methane in subsurface biogeochemical cycling and its implications for arsenic contamination in groundwater systems.

Plasmids are essential tools in molecular biology and biotechnology. In research laboratories, it is common to use antibiotic selection markers to ensure that plasmids are stably maintained in a cellular population. However, the use of antibiotics poses a significant challenge in the industrial scale-up process due to the high cost and the risk of spreading resistance. Therefore, methods for antibiotic-free plasmid maintenance are in high demand. Here, we present an essential gene-based plasmid selection strategy utilizing the Escherichia coli tryptophan tRNA (trpT) gene. We developed a workflow using a base strain with a trpT deletion and a temperature-sensitive trpT-expressing plasmid to circumvent the need for remaking chromosomal trpT deletions for every transformation. We evaluated the stability of a range of antibiotic gene-free trpT plasmids with different copy numbers and determined that the system is as efficient as, or better than, systems using antibiotics. Furthermore, the system is stable when producing a biochemical at industrially relevant fermentation conditions, and due to the small size of trpT, it allows for plasmid minimization. The approach constitutes a significant contribution toward developing simpler and more effective antibiotic-free bioprocesses and combating the spread of multiresistant infections.

In response to environmental changes, bacteria have evolved sophisticated regulatory networks that incorporate small RNAs (sRNAs) and transcription factors (TFs) to fine-tune cellular physiology. Both sRNAs and TFs modulate gene expression, but the former function via post-transcriptional mechanisms, while the latter act at the transcriptional level. However, it remains unclear why both regulatory layers are conserved through evolution, rather than one being sufficient. Here, we experimentally identified that CRP, a global regulator, regulates 25 small RNAs (sRNAs) in Escherichia coli. Interestingly, CRP also controls 80% of the target genes of these sRNAs. This architecture led us to identify 34 novel sRNA-mediated feed-forward loops (sFFLs) circuits where CRP regulates both an sRNA and its target within the CRP regulon. Quantitative PCR analysis of 16 such sFFLs revealed that each type possesses a distinct cAMP dose-response profile, suggesting that different sFFL structures embody unique regulatory logic. Specifically, coherent and incoherent type 3 and 4 sFFLs exhibit broader dynamic ranges in their dose-response compared to open-loop controls. Coherent and incoherent type 1 and 2 sFFLs appear more energy-efficient. Altogether, we propose that sRNAs cooperate with TFs through sFFLs to optimize both the energy efficiency and the diversity of signal response profiles. Therefore, sRNAs serve as critical components integrating transcriptional and post-transcriptional networks across diverse cellular pathways.

Antibiotics revolutionized medicine in the 20th century by drastically reducing mortality from bacterial infections. However, their effectiveness is threatened by the global rise of antimicrobial resistance (AMR), driven by misuse, overuse, and environmental dissemination. This review explores the historical trajectory of antibiotics, the mechanisms of bacterial resistance, and the urgent need for innovation amid a declining antibiotic development pipeline. Herein, we highlight the scientific and economic barriers that have discouraged investment by major pharmaceutical companies and examine emerging strategies to address this crisis. Key advances in microbial bioprospecting, including cultivation improvement techniques and genome mining, are discussed alongside the role of high-throughput sequencing and bioinformatics in unlocking the metabolic potential of uncultivated microorganisms. Particular emphasis is placed on the integration of artificial intelligence and machine learning to accelerate drug discovery, predict antimicrobial activity, and identify resistance genes. Additionally, we present alternative therapeutic strategies beyond traditional antibiotics, such as phage therapy, antimicrobial peptides, quorum sensing inhibitors, synthetic conjugates, and vaccine development. Together, these interdisciplinary approaches offer promising pathways to revitalize the antimicrobial pipeline and address the growing threat of antibiotic resistance.