Engineering DNA polymerases to efficiently synthesize artificial or noncognate nucleic acids remains an essential challenge in synthetic biology. Here we describe an evolutionary campaign designed to convert a family of highly selective DNA polymerases into an unnatural homolog with strong RNA synthesis activity. Starting from a homologous recombination library, a short evolutionary path was achieved using a single-cell droplet-based microfluidic selection strategy to produce C28, a newly engineered polymerase that can synthesize RNA with a rate of ~3 nt s−1 and of >99% fidelity. C28 is capable of long-range RNA synthesis, reverse transcription and chimeric DNA–RNA amplification using the PCR. Despite strong discrimination against other genetic systems, C28 readily accepts several 2′F and base-modified RNA analogs. Together, these findings highlight the power of directed evolution as an approach for reprogramming DNA polymerases with activities that could help drive future applications in biotechnology and medicine. Engineering polymerases to synthesize alternative genetic polymers remains a challenging problem in synthetic biology. Using DNA shuffling and droplet microfluidics, the current study provides a short evolutionary path from a DNA polymerase to one with robust RNA-synthesizing activity.

Site-specific DNA recombinases are powerful tools for genome engineering. The recent convergence of DNA recombinase-mediated technology with prime editing has established a promising new frontier for large-scale genomic integration. However, this strategy has limitations, including low intrinsic catalytic efficiency of DNA recombination enzymes and inefficiency of their long DNA landing pad insertions. To overcome these challenges, we developed a novel directed evolution strategy using progressively shortening landing pads as a selective pressure. The resulting evolved variants, VK and AVK, exhibited substantially enhanced intrinsic activity that is maintained on short-length landing pads. This enables a powerful synergistic effect with increased insertion efficiency of prime editing, leading to a dramatic increase in overall integration efficiency while maintaining high genomic specificity. This work thus provides a new class of highly efficient recombinases and a robust engineering strategy with broad applicability for both the development of gene therapies and fundamental research.

Noncoding single-nucleotide variants (SNVs) that alter transcription factor (TF) binding can affect gene expression and contribute to disease. Sequence-based methods can excel at predicting TF binding, but rely on training data and can exhibit TF-specific biases. Here, we propose a structure-guided approach for noncoding SNVs, using AlphaFold 3 (AF3) to model TF–DNA complexes and FoldX for downstream physics-based assessment. Benchmarked against single nucleotide polymorphism-systematic evolution of ligands by exponential enrichment (SNP-SELEX) data for six TFs (SPIB, ELK3, ETV4, SF-1, PAX5, and MEIS2), the FoldX-based strategy showed good agreement with experimental allele preferences. Interestingly, differences in AF3’s interface-predicted template modelling (ipTM) score aligned even more closely with SNP-SELEX results, generally surpassing energy-based metrics. Application to known disease-associated variants recapitulated most reported effects for TFs including NKX2-5, GATA3, and USF2A-USF1. In these examples, considering both ΔipTM and FoldX energies proved more reliable than either metric alone. While less accurate than state-of-the-art sequence-based methods, this work demonstrates that structural modelling can yield interpretable insights into how noncoding variants influence TF binding. By highlighting both the promise and limitations of AF3 in this context, our study provides a framework for complementary structural evaluation of regulatory variants.

Some prokaryotes carry CRISPR-associated transposons (CASTs), Tn7-like elements that incorporate genes encoding CRISPR-Cas effectors. CAST insertion is directed by CRISPR-Cas effectors through RNA-guided DNA binding and interactions with transposition-associated proteins. Although efficient sequence-specific DNA integration requires both precise target DNA recognition and coordinated interactions between effectors and transposition-associated proteins, the underlying mechanism remains elusive. Here, we determined three cryo-EM structures of target DNA-bound type I-F3 TniQ-Cascade from Vibrio parahaemolyticus, revealing how Cas8/5 recognizes the protospacer adjacent motif (PAM) and identifying a key residue responsible for the cytidine preference at position -2 of the PAM. We revealed mismatch tolerance at the PAM-proximal site. Structural analyses showed that correct base pairing at the PAM-distal site correlates with conformational changes in the Cas8/5 helical bundle and TniQ, bending the DNA to guide its downstream region toward the transposition machinery. Together, these dynamic rearrangements at the PAM-distal region provide insights into the licensing mechanism of type I-F3 CAST transposition and highlight its potential for genome engineering applications.

In all domains of life, tRNAs mediate the transfer of genetic information from mRNAs to proteins. As their depletion suppresses translation and, consequently, viral replication, tRNAs represent long-standing and increasingly recognized targets of innate immunity. Here we report Cas12a3 effector nucleases from type V CRISPR–Cas adaptive immune systems in bacteria that preferentially cleave tRNAs after recognition of target RNA. Cas12a3 orthologues belong to one of two previously unreported nuclease clades that exhibit RNA-mediated cleavage of non-target RNA, and are distinct from all other known type V systems. Through cell-based and biochemical assays and direct RNA sequencing, we demonstrate that recognition of a complementary target RNA by the CRISPR RNA triggers Cas12a3 to cleave the conserved 5′-CCA-3′ tail of diverse tRNAs to drive growth arrest and antiphage defence. Cryogenic electron microscopy structures further revealed a distinct tRNA-loading domain that positions the tRNA tail in the RuvC active site of the nuclease. By designing synthetic reporters that mimic the tRNA acceptor stem and tail, we expanded the capacity of current CRISPR-based diagnostics for multiplexed RNA detection. Overall, these findings reveal widespread tRNA inactivation as a previously unrecognized CRISPR-based immune strategy that broadens the application space of the existing CRISPR toolbox. Cas12a3 nucleases constitute a distinct clade of type V CRISPR–Cas bacterial immune systems that preferentially cleave the 3′ tails of tRNAs after recognition of target RNA to induce growth arrest and block phage dissemination.

Understanding the extrinsic and intrinsic environmental factors that influence lag phase duration is critical for developing strategies to control the growth of foodborne pathogens. Although the exponential growth rate can be predicted as a straightforward response to the growth environment, lag phase duration is difficult to predict because it depends on not only the current growth conditions but also the previous growth environment and physiological status of the bacterial cells. Therefore, this article aims to provide a comprehensive understanding of the dynamic, adaptable, and evolvable nature of the lag phase. It is based on relevant literature (experimental studies, literature summaries, observational data) on the effect of pre- and postgrowth environments. We discuss the modeling strategies employed to incorporate physiological heterogeneity and dynamic food environments into predictive modeling frameworks. Overall, we summarize the empirical and mechanistic modeling strategies for quantifying lag phase duration.

Although genome sequencing technologies have advanced rapidly, microbial genomes still contain numerous genes with unknown functions, posing ongoing challenges for comprehensive genome annotation. Traditional annotation methods are constrained by a lack of scalable experimental techniques and the limitations of conventional homology-based computational approaches. Recent computational innovations, particularly deep learning, have substantially improved gene function prediction, facilitating more efficient annotation of transcription factors, enzymes and other protein classes. Integrating computational and experimental approaches has enabled the development of workflows that systematize gene function discovery, paving the way for faster, more accurate and comprehensive genome annotation. Continued refinement of these integrated methods holds great promise for deepening our understanding of microorganisms. Here we review recent advances in artificial intelligence for gene function discovery and discuss future directions for achieving interpretable and high-throughput artificial intelligence-guided annotation. In this Review, the authors discuss how artificial intelligence can aid microbial gene function discovery.

Bacteriophage (phage) genome annotation is essential for understanding their functional potential and suitability for use as therapeutic agents. Here, we introduce Phold, an annotation framework utilizing protein structural information that combines the ProstT5 protein language model and structural alignment tool Foldseek. Phold assigns annotations using a database of over 1.36 million predicted phage protein structures with high-quality functional labels. Benchmarking reveals that Phold outperforms existing sequence-based homology approaches in functional annotation sensitivity whilst maintaining speed, consistency, and scalability. Applying Phold to diverse cultured and metagenomic phage genomes shows it consistently annotates over 50% of genes on an average phage and 40% on an average archaeal virus. Comparisons of phage protein structures to other protein structures across the tree of life reveal that phage proteins commonly have structural homology to proteins shared across the tree of life, particularly those that have nucleic acid metabolism and enzymatic functions. Phold is available as free and open-source software at https://github.com/gbouras13/phold.

Gene co-expression networks are commonly used to identify functionally related genes, but they often suffer from spurious associations and fail to capture genes with restricted, context-specific responses. To address these limitations, we constructed YeastCoDEGNet, a global co-differential expression network in Saccharomyces cerevisiae, by reanalyzing microarray data from 143 carefully curated experiments comprising 425 comparisons. In this network, gene connections are defined by shared context-specific responses, enabling the identification of distinct topological groups enriched in either essential or nonessential genes, often associated with metabolic or nonmetabolic biological processes, respectively. We further characterized each gene by its responsiveness, essentiality, and number of co-differentially expressed partners, uncovering positional clustering of these features across chromosomal locations. To explore higher-order functional organization, we built a cross-pathway coordination network based on context-specific responses between pathway pairs, revealing modules of functionally related pathways. This network not only recovered well-known associations but also identified novel links, with particularly strong coordination observed among pathways involved in central carbon metabolism, amino acid and antioxidant processes, protein synthesis, trafficking, degradation, and gene regulation. By capturing context-specific gene expression dynamics, YeastCoDEGNet provides a powerful framework for studying how genes and pathways adapt to genetic and environmental perturbations.

Magnetotactic bacteria produce membrane-bound organelles known as magnetosomes, which consist of chains of magnetite crystals and function as sensors for orientation within the Earth’s magnetic field. Magnetosome biosynthesis is a complex, multistep process that depends on iron availability and suboxic conditions. However, the expression of the ca. 30 core magnetosome biosynthetic genes has previously been described as constitutive and largely unaffected by environmental conditions. A recent study by Pang et al., published in this journal, reported the identification of a transcriptional regulator, NsrRMg, which was proposed to activate magnetosome biosynthesis in Magnetospirillum gryphiswaldense inresponse to the endogenous signaling molecule nitric oxide (NO). Furthermore, the study suggested that NO also regulates a putative, previously unrecognized nitrification pathway that presumably supports endogenous NO production via denitrification, even in the absence of extracellular nitrate.

Here, we present results of genetic and transcriptional analyses demonstrating that, contrary to the findings of Pang et al., NsrRMg is not required for magnetosome biosynthesis. We also refute the existence of the proposed nitrification pathway and conclude that there is no compelling evidence supporting a role of NsrRMg as a key regulator of magnetosome formation in M. gryphiswaldense.

RNA editing enzymatically modifies RNA molecules post-transcriptionally, enabling precise sequence alterations. Advantages include reversibility and temporal control without genomic DNA changes, allowing dynamic regulation of gene expression while preserving original genetic information. In this study, we characterized McAgo derived from Monosporascus cannonballus, which functions as a programmable nuclease guided by 14–30 nt gRNAs, demonstrating robust RNA cleavage activity at physiological temperature. Furthermore, we delivered McAgo RNP (ribonucleoprotein) complexes into mammalian cells, achieving >90% RNA knockdown efficiency with minimal innate immune responses. A catalytically inactive mutant (dMcAgo) using a gRNA as short as 20 nt, conjugated to the hADAR2 deaminase domain (hADAR2dd E488Q), achieved up to 90% RNA editing efficiency in vitro. This study establishes, for the first time, the effective targeting of endogenous RNA by a heterologous Argonaute in mammalian cells, alongside its demonstrated utility for RNA editing—thereby expanding the functional repertoire of Argonaute proteins.

Mammalian genomes contain millions of regulatory elements that control the complex patterns of gene expression. Previously, the ENCODE consortium mapped biochemical signals across hundreds of cell types and tissues and integrated these data to develop a registry containing 0.9 million human and 300,000 mouse candidate cis-regulatory elements (cCREs) annotated with potential functions. Here we have expanded the registry to include 2.37 million human and 967,000 mouse cCREs, leveraging new ENCODE datasets and enhanced computational methods. This expanded registry covers hundreds of unique cell and tissue types, providing a comprehensive understanding of gene regulation. Functional characterization data from assays such as STARR-seq, massively parallel reporter assay, CRISPR perturbation and transgenic mouse assays have profiled more than 90% of human cCREs, revealing complex regulatory functions. We identified thousands of novel silencer cCREs and demonstrated their dual enhancer and silencer roles in different cellular contexts. Integrating the registry with other ENCODE annotations facilitates genetic variation interpretation and trait-associated gene identification, exemplified by the identification of KLF1 as a novel causal gene for red blood cell traits. This expanded registry is a valuable resource for studying the regulatory genome and its impact on health and disease. The existing ENCODE registry of candidate human and mouse cis-regulatory elements is expanded with the addition of new ENCODE data, integrating new functional data as well as new cell and tissue types.