Genome-editing technologies that use recombinases to insert kilobase-scale DNA sequences into mammalian genomes canonically require large double-stranded DNA (dsDNA) donors. However, dsDNA molecules evoke problematic and toxic innate immune responses, limiting integration efficiencies and generally constraining applicability to ex vivo or immune-deficient contexts. By harnessing mechanisms of integrative prokaryotic viruses and mobile genetic elements, here we demonstrate that recombinases are compatible with immune evasive circular single-stranded DNA molecules optimally bearing a partial-duplex region that reconstitutes the recombinase recognition sequence. This approach, which we term integration through nucleus-synthesized template addition of large lengths (INSTALL), is compatible with diverse protein and RNA-guided recombinases for high-fidelity kilobase-scale human genome writing. INSTALL minimizes innate immune responses in primary human cells and in mice, improving recombinase-mediated integration efficiencies and supporting systemic in vivo non-viral DNA delivery by substantially increasing tolerability and broadening the dosing range compared with lipid nanoparticle-delivered dsDNA molecules. Together, INSTALL overcomes fundamental challenges for DNA delivery and integration methods by synergizing immune-stealth nucleic acids with recombinases to enable kilobase-scale integration strategies without viral vectors. INSTALL overcomes fundamental challenges for DNA delivery and integration methods by synergizing immune-stealth nucleic acids with recombinases to enable kilobase-scale integration strategies without viral vectors.

Engineered cell therapies are transforming precision medicine by enabling real-time, context-responsive interventions that act upon disease-specific cues. Inspired by the success of CAR-T cells in oncology, next-generation platforms are being developed using diverse immune cells and stem cells to address a broader spectrum of diseases. These living therapeutics harness synthetic gene circuits to induce targeted cytotoxicity, to modulate the secretion of effector proteins or to coordinate both functions in response to endogenous signals or externally delivered molecular and physical triggers. Ex vivo engineering of autologous cells remains the norm, but challenges in scalability, cost and accessibility are fuelling efforts towards allogeneic products and in vivo reprogramming. Advances in targeted delivery — using viral vectors, mRNA-loaded nanoparticles and virus-like particles — are expanding the toolkit for direct programming of cells within the body. This Review discusses emerging strategies for engineering human cells with therapeutic functions, highlighting modular control systems, delivery innovations and the translational hurdles that lie ahead. In this Review, Teixeira et al. discuss emerging strategies for developing and improving engineered-cell therapies. They outline progress from ex vivo engineered autologous cells to in vivo reprogramming, advances in delivery systems and the remaining translational barriers.

Plant diseases pose a great risk to global food security, and recent research indicates that pathogen pressures on plant productivity will substantially increase under ongoing climate change (that is, increasing CO2 levels and global warming). However, our mechanistic and predictive knowledge of the impacts of climate extremes, such as heatwaves and prolonged droughts, and their interaction with other climatic factors, on plant pathogens, hosts and microbiomes, remains largely unknown. This is an important knowledge gap that limits our ability to develop effective strategies to mitigate the socioeconomic impacts of climate change-induced plant disease outbreaks. This Review examines the impacts of key climate extremes on soil-borne pathogens, plant microbiomes and host physiology that ultimately determine disease outcomes. We explore evidence that suggests that the responses of pathogen–host–microbiome interactions to climate extremes may differ in many ways from those to long-term climate change. Climate extremes may increase the virulence and distribution of many pathogens, suppress certain plant immune responses, and weaken the core functions of host microbiomes within the disease triangle, thereby facilitating disease outbreaks. We propose an integrated pathway for harnessing microbiomes to address the critical challenges posed by climate extremes. These insights offer new approaches to mitigate disease risks by harnessing microbiomes and metabolites under climate extremes, with the potential to support climate-resilient and sustainable agricultural and natural ecosystems. In this Review, Singh BK and colleagues discuss the impacts of key climate extremes on soil-borne pathogens, plant microbiomes and host physiology that ultimately determine disease outcomes, as well as the eco-evolutionary mechanisms by which pathogens, hosts and microbiomes may adapt to climate extremes. Finally, they propose an integrated pathway for harnessing microbiomes to address the critical challenges posed by climate extremes.

Artificial intelligence has rapidly entered the field of life sciences, including vaccine development. Here, we introduce ‘reverse vaccinology 3.0’, a transformative approach that could revolutionize the field of vaccinology and contribute to addressing global health challenges. Andreano, McLellan and Rappuoli introduce ‘reverse vaccinology 3.0’, a transformative approach that could revolutionize the field of vaccinology and contribute to addressing global health challenges.

Functional nucleic acids (FNAs) are essential elements for designing advanced molecular tools, yet their de novo design faces challenges due to the vast sequence space and inefficiency of experimental screening methods. Nucleic acid large language models (NA-LLMs) offer new opportunities for FNA design, but their generative capability remains underexplored. Here we introduce InstructNA, a framework leveraging NA-LLMs and high-throughput systematic evolution of ligands by exponential enrichment (HT-SELEX) to guide de novo design of FNAs without relying on structural information. InstructNA encodes semantically rich FNA representations and robustly decodes FNA sequences, enabling the generation of various types of FNA such as transcription factor-binding DNA and protein-binding aptamers with enhanced functionality and high sequence diversity. Compared with the traditional HT-SELEX, InstructNA generates 100% and 200% more strong aptamer binders for two protein targets, with a sequence similarity to the original HT-SELEX aptamers as low as 38%. These results underscore the efficacy and robustness of InstructNA, demonstrating its potential for FNA design. InstructNA leverages nucleic acid large language models with HT-SELEX for de novo generation of functional nucleic acids, exhibiting high efficiency and general applicability in designing aptamers for various targets.

The transcriptional interference model suggests that RNA polymerases elongating through overlapping transcription units mutually inhibit transcription and disrupt associated cis-regulatory elements. As a longstanding fundamental concept of gene regulation, the idea of reciprocal inhibition between sense and antisense transcription has been supported by a significant body of research. However, despite the model’s biophysical plausibility and historical significance, evidence from large-scale transcriptome studies raises questions about its universal applicability. In particular, the new data indicate that a measurable influence of transcriptional interference is absent from the majority of loci with overlapping transcription. Here we highlight key aspects of overlapping transcription and propose potential solutions to this emerging puzzle. Gaining a better understanding of the molecular mechanisms that render loci sensitive or resistant to interference could lead to groundbreaking insights into the biology of gene regulation. The transcriptional interference model has been instrumental in shaping our understanding of gene regulation. However, given that its effects are largely absent from genome-wide data, this Perspective reexamines the underlying mechanisms and suggests how they may be resolved at most loci.

Biohybrid microrobots integrate biological components with synthetic structures to navigate complex biological environments, for example, for the delivery of drugs, microsurgery and in vivo diagnostics. In this Review, we propose a biophysics-informed design framework for biohybrid microrobots by connecting biophysical principles with biohybrid solutions. We first identify the biophysical constraints imposed by the human body that limit microrobot integrity, locomotion, navigation and functionality. We then examine the biophysical mechanisms through which biological cells, microorganisms and their derivatives adapt to these challenges, and explore how these can be utilized to improve the performance of microrobots. Building on these insights, we describe how biohybrid microrobots translate biophysical strategies into engineering solutions across four design domains: deformation, actuation, navigation and programming. Finally, we discuss persisting in vivo challenges, key considerations for clinical translation and future developments. By articulating design logics that span biological and synthetic domains, this framework provides a functional definition of biohybrid microrobots and offers a shared language for researchers across disciplines. Biohybrid microrobots combine biological and synthetic components to navigate complex in vivo environments for applications such as drug delivery, microsurgery and diagnostics. This Review introduces a biophysics-informed design framework for biohybrid microrobots that translates natural adaptive strategies into engineering solutions across deformation, actuation, navigation and programming.

Proteins exhibit complex phase behavior as they convert between the native state, the liquid condensate (or droplet) state and the solid condensate (or amyloid) state. To facilitate the study of these processes, we describe the FuzDrop method of predicting the condensation propensity of proteins to undergo liquid–liquid phase separation and to subsequently form amyloid aggregates. The method is based on the principle that liquid condensations reflect a balance between enthalpic and entropic contributions; FuzPred is an algorithm that provides sequence-based estimates for these contributions in stoichiometric complexes ( https://fuzpred.bio.unipd.it/predictor ). FuzDrop extends this algorithm to protein condensates, and enables prediction of the propensity for amyloid formation within liquid condensates, known as the condensation pathway to protein aggregation ( https://fuzdrop.bio.unipd.it/predictor ). This prediction is based on the principle that the sequence regions that promote aggregation within liquid condensates have a multiplicity of binding modes, because they have a strong propensity for both entropic-driven interactions to stabilize the droplet state and enthalpic-driven interactions to stabilize the amyloid state. The time required for FuzDrop predictions on the web server scales linearly with protein length and is typically ~30 s for a protein of 500 residues. By enabling predictions of protein phase behavior, FuzDrop may facilitate experimental studies directed at the development of therapies for protein condensation diseases. FuzDrop predicts the condensation propensity of proteins on the basis of their amino acid sequences. This protocol describes the underlying theory and how to use the results to understand liquid-liquid phase separation and amyloid aggregate formation.

Bacteria harness diverse defence systems that protect against phage predation, many of which are encoded on horizontally transmitted mobile genetic elements. In turn, phages evolve counter-defences, driving a dynamic arms race that remains underexplored in human disease contexts. For the diarrhoeal pathogen Vibrio cholerae, a higher burden of its lytic phage ICP1 in patient stool correlates with reduced disease severity. However, direct molecular evidence of lytic phages driving selection of epidemic V. cholerae has not been demonstrated. Here, through clinical surveillance in cholera-endemic Bangladesh, we capture the acquisition of a parasitic antiphage mobile genetic element, PLE11, that initiated a selective sweep coinciding with the largest cholera outbreak in recent records. PLE11 showed potent anti-phage activity against cocirculating ICP1, explaining its rapid and dominating emergence. We identify PLE11-encoded Rta as the defence responsible and provide evidence that Rta restricts phage tail assembly. Using experimental evolution, we predict phage counteradaptations against PLE11 and document the eventual emergence and selection of clinical ICP1 that achieve a convergent evolutionary outcome. Finally, we discover how PLEs balance their dependence on ICP1 tail proteins for horizontal transmission with the restriction of phage tail assembly by Rta: PLEs construct chimeric tails composed of both mobile genetic element-encoded and phage-encoded proteins to ensure their transmission. Collectively, our findings reveal the molecular basis of the natural selection of a globally important pathogen and its virus in a clinically relevant context. The acquisition of a parasitic anti-phage mobile genetic element, PLE11, showing potent anti-phage activity against cocirculating ICP1, and the subsequent evolution of ICP1 to escape this defense, are captured, revealing the molecular basis of the natural selection of a globally notable pathogen and its virus.

Microbial serine proteases are valuable for industrial applications due to broad substrate specificity and stability. However, heterologous overexpression in microbial hosts is often limited by cytotoxicity and poor secretion. This study developed an integrated strategy combining protein engineering and signal peptide optimization to enhance extracellular production of PrtA—a key acid-stable alkaline serine protease—in Komagataella phaffii.  Directed evolution generated the Q245K variant, showing 1.35-fold higher extracellular expression than wild-type PrtA. A machine learning model, MPEPE (Mutation Predictor for Enhanced Protein Expression), was used to identify critical residues involved in protein secretion; saturation mutagenesis at the top-predicted site generated the I342D mutant with 1.48-fold improved productivity. The double mutant PrtA-Q245K/I342D achieved synergistic enhancement (1.84-fold higher secretion) without altering enzymatic properties. Evaluation of nine signal peptides revealed that serum albumin, α-factor (without pro-region), and PrtA’s native signal peptides each doubled the combinatorial mutant’s secretion, yielding 4.98-fold higher expression than the wild-type. In contrast, α-factor pro-region inclusion drastically reduced yields. In a 15-L fed-batch bioreactor, the optimized strain produced PrtA-Q245K/I342D at 4807.5 U/mL, equivalent to 1.5 g/L protein.  The combined approach of directed evolution, machine learning-guided mutagenesis, and signal peptide engineering significantly boosted PrtA secretion while maintaining functional integrity. This strategy demonstrates strong potential for scalable industrial production of challenging heterologous proteases.

Cross-kingdom RNA interference is an emerging concept in plant–pathogen interactions. Here we provide evidence that cross-kingdom RNA interference also occurs in a beneficial plant symbiosis called arbuscular mycorrhiza AMF. The arbuscular mycorrhizal fungus Rhizophagus irregularis transfers small RNAs into plant cells, promoting the colonization of host roots. This finding establishes inter-organismal RNA communication as a new regulatory mechanism of this ancient and widespread symbiosis. The authors provide evidence that symbiotic fungi forming widespread arbuscular mycorrhiza symbioses use cross-kingdom RNA interference to silence plant genes and promote their colonization of host roots.

TnpB is a diverse family of RNA-guided endonucleases associated with prokaryotic transposons. Because of their small size and putative evolutionary relationship to CRISPR–Cas12, TnpB enzymes hold great potential for genome editing. However, most TnpBs lack robust gene-editing activity. Here, we mapped comprehensive sequence–function landscapes of a TnpB ribonucleoprotein using deep mutational scanning and we discovered activating mutations in both the RNA and the protein. Leveraging the protein’s mutational landscape, we constructed a combinatorial library of activating mutations, from which we identified two enhanced TnpB variants. These variants increased editing in human cells, Nicotania benthamiana, pepper and rice. While editing efficiencies varied by target site, engineered variants achieved up to 55% insertion and deletion frequencies (a 50-fold increase over wild type) in N. benthamiana, surpassing ISYmu1 (<7%), AsCas12f-HKRA (<9%) and other compact editors. These findings highlight elements critical for regulating TnpB endonuclease activity and demonstrate latent activity accessible through mutation. TnpB endonucleases are engineered for improved genome editing.

Synthetic microbial consortium inoculants are emerging nature-based solutions for promoting sustainable agriculture and mitigating environmental challenges. However, despite promising results in simpler lab-scale trials, many inoculants fail to establish or perform satisfactorily in field conditions. One most critical yet least understood factor influencing inoculant effectiveness is the complex microbial interactions, both within consortium inoculants (“within-community” interactions) and between consortium inoculants and native soil communities (“cross-community” interactions). Here, we first discuss major negative and positive “within-community” interactions and highlight the importance to design consortium inoculants with positive interactions for improved stability and functionality. We then examine the bidirectional “cross-community” interactions once introducing consortium inoculants to soils. Soil native communities often create strong resistance to the invasion of inoculants. We discuss major drivers controlling the invasibility of native communities and various strategies increasing the invasiveness of consortium inoculants. We then discuss how consortium inoculants can reshape native communities, with implications for long-term ecosystem resilience and functioning. We propose future research efforts including advancing strategies for harnessing natural species from relatively untapped soil reservoirs and using high-throughput interaction profiling with computational tools to build compatible synthetic consortia with desirable functions; leveraging positive interactions and prebiotics to facilitate inoculant establishment; and assessing fully soil functional resilience over longer terms, including recognizing the importance of rare keystone taxa. By integrating with ecological theory, this review provides a comprehensive insight into microbial interactions to advance the design, application, and monitoring of synthetic consortium inoculants for enhancing soil health and ecosystem sustainability.