Precise control of protein function in living cells is essential for engineering programmable biological systems and therapeutic applications. However, most current strategies act indirectly by altering protein stability, localization, or proximity rather than directly modulating binding activity. Here we present VIPbodies, nanobodies engineered with self-cleaving viral proteases that convert protease inhibition into drug-dependent antigen recognition. Using anti-mCherry nanobodies as prototypes, we create variants responsive to orthogonal viral protease–inhibitor pairs and extend the design to diverse nanobody scaffolds. Each VIPbody functions independently within the same cell, enabling multiplexed regulation of distinct targets. When incorporated into transcriptional circuits, VIPbodies mediate drug-tunable gene expression and execute all six canonical Boolean logic operations, providing a compact framework for programmable cellular computation. Beyond gene regulation, VIPbody circuits mediate bidirectional control of pyroptosis and selectively activate apoptotic or pyroptotic programs via caspase coupling, thereby enabling chemogenetic control of protein function and cell fate. Precise control of protein function in living cells is essential for engineering programmable biological systems. Here the authors present VIPbodies, nanobodies engineered with self-cleaving viral proteases that convert protease inhibition into drug-dependent antigen recognition.

Nitritation, the conversion of ammonia to nitrite without further oxidation, offers an energy-efficient route for nitrogen removal, but its application is limited by the difficulty of selectively suppressing nitrite-oxidizing bacteria (NOB). The underlying biological mechanisms that enable such suppression remain poorly understood. Here we show that quorum sensing (QS), a cell–cell communication system, enables nitritation by regulating NOB behaviour. Using multi-omics and single-cell Raman spectroscopy, we demonstrate that QS signalling induces the overexpression of nirB in the dominant NOB genus Nitrospira, triggering an altruistic nitrite reduction causing self-inactivation. In contrast, ammonia-oxidizing bacteria refrain from this altruistic metabolism, gaining a decisive competitive advantage and directing nitrification flux towards nitritation. QS manipulation confirms that active QS is required to maintain nitritation, and single-cell analysis reveals that QS drives a stress-tolerant Nitrospira cell into a susceptible state, markedly reducing survival. These findings uncover an unknown social behaviour in the nitrifier community and offer new insights for nitritation stabilization. Nitritation is limited by the difficulty of suppressing nitrite-oxidizing bacteria, the mechanisms of which are clear. This study shows that quorum sensing drives altruistic self-inactivation in Nitrospira, favouring ammonia oxidizers and stabilizing nitritation.

The design of molecules with desired properties is a key challenge in drug discovery and materials science. Traditional methods rely on trial-and-error, while recent deep-learning approaches accelerate molecular generation. However, existing models struggle with generating molecules based on specific textual descriptions. We introduce Mol-CADiff, a diffusion-based framework that uses causal attention mechanisms for text-conditional molecular generation. Our approach explicitly models the causal relationship between textual prompts and molecular structures, overcoming limitations in existing methods. We enhance dependency modeling both within and across modalities, enabling precise control over the generation process. While primarily designed for text-guided tasks, this architecture inherently supports unconditional generation, providing the added capability to autonomously sample the broader chemical space without explicit constraints. Here we show that Mol-CADiff outperforms alternative methods in generating diverse, chemically valid molecules, with better alignment to specified properties, enabling more intuitive language-driven molecular design. By bridging these modalities, our framework provides a versatile method for drug discovery. Computational approaches to molecular design often explore only limited regions of the vast chemical space. This study presents a causality-aware diffusion model that generates valid and diverse molecules with or without text prompts, improving controllability in molecular design.

Accurate decoding of peptide sequences is crucial in proteomics. However, achieving this goal is a technical challenge, owing to the compositional and structural complexity of peptides. Studies inspired by the success of nanopore nucleic acid sequencing have shown that nanopore-based techniques can also be applied to peptide sequencing. The key is to generate narrowly distributed, consistent and sequence-dependent events during sequential nanopore readout. Here we introduce a nanopore-based strategy termed transient pore analyte looping (tPAL). We develop an engineered Mycobacterium smegmatis porin A (MspA) nanopore that is dual modified with a nickel-ion-bound nitrilotriacetic acid (NTA-Ni) adapter and the target peptide. This distinctive sensing configuration enables precise recognition of the N terminus of the immobilized peptide by multiple re-readings. With the aid of cholesterolized aminopeptidase, the immobilized peptide can be shortened sequentially in single-amino-acid increments, yielding sequence-dependent, stepwise and narrowly distributed signal alterations that provide clues to allow peptide sequence decoding. Our strategy achieves single-amino-acid resolution and effectively identifies single-amino-acid mutations, post-translational modifications and unnatural-amino-acid insertions—indicative of its versatility in nanopore proteomics and chiral peptide analyses. A nanopore-based ‘chop and measure’ method sequences peptides at single-amino-acid resolution by using enzymatic digestion to progressively shorten the N terminus one residue at a time, together with repetitive N-terminus re-reading.

No principled framework exists for conditioning sequence generative models for protein engineering on auxiliary information, such as experimental data, without additional training of a generative model. Here we present ProteinGuide, a method for such ‘on-the-fly’ conditioning. ProteinGuide is amenable to a broad class of protein generative models including masked language models such as ESM3, any-order autoregressive models such as ProteinMPNN and diffusion and flow-matching models on discrete state-spaces such as MultiFlow. ProteinGuide stems from a unifying statistical framework for these model classes. As proof of principle, pretrained generative models are used to design proteins with user-specified properties, such as higher stability or activity. Proteins are additionally designed to optimize for two desired properties that are in tension with each other. Lastly, we apply ProteinGuide jointly with wet-lab data generation to increase the editing activity of an adenine base editor in vivo, resulting in a base editor with higher editing efficiency than was previously achieved using seven rounds of directed evolution. On-the-fly conditioning of pretrained protein generative models guides protein generation toward specific properties.

Despite lacking a stable three-dimensional structure, intrinsically disordered protein regions (IDRs) are ubiquitous across all kingdoms of life and have essential cellular roles. While rational design of folded proteins has seen substantial recent progress, our ability to design IDRs remains more limited. Here we present GOOSE (Generate disOrdered prOteins Specifying propErties), a comprehensive computational framework for the rational design of IDRs. GOOSE’s versatility and throughput enable us to design and test thousands of IDR sequences to reveal distinct sequence-to-function relationships. Using GOOSE to explore these relationships, we examine how sequence properties influence IDR structural ensembles in cells, design IDRs that respond to structural changes associated with cell volume decrease, create scaffold IDRs that self-assemble and recruit specific clients, and design novel IDRs that protect cells from desiccation. Our work uses rational sequence design as a powerful method for exploring function in IDRs and provides a versatile tool for designing functional disordered proteins. GOOSE enables the design and testing of thousands of disordered protein region sequences to reveal distinct sequence-to-function relationships.

Pseudomonas aeruginosa is a ubiquitous, Gram-negative bacterium that forms biofilms and is responsible for antibiotic-resistant hospital-acquired infections in humans. The P. aeruginosa BqsRS two-component system regulates biofilm formation and dispersal by sensing extracytoplasmic Fe2+, but the mechanistic details of this process are poorly understood. In this work, we report the crystal and solution structures of the PaBqsR response regulator receiver domain, comprising a (βα)5 response regulator assembly, and the DNA-binding domain, comprising a helix-turn-helix motif. Consistent with its cognate stimulus being Fe2+, we show that BqsR binds directly to the promoter region of the feo operon that encodes the bacterial Fe2+ transport system FeoABC. Corroborating these in vitro results, transcriptional studies show that BqsR is a global regulator controlling many important genes in PAO1, including the feo operon. Intriguingly, promoter-based assays reveal that BqsR is a dynamic regulator that responds to bioavailable Fe2+, likely through the ability of BqsR to bind Fe2+ directly via a His-rich motif, independent of the BqsS membrane His kinase. To our knowledge, this mode of regulation has not been reported previously among OmpR-like response regulators but represents an important level of control over Fe2+ acquisition in P. aeruginosa that could be an attractive therapeutic target to treat hospital-acquired infections. The BqsRS two-component system regulates biofilms in the infectious pathogen Pseudomonas aeruginosa. Here, the authors show that the global response regulator BqsR controls Fe2+ acquisition in PAO1 through a previously unrecognized regulatory mechanism.

Artificial intelligence (AI) is transforming scientific research, including proteomics. In this Perspective, we highlight key mass spectrometry (MS)-based proteomics areas where AI is driving innovation, ranging from protein identification to building AI virtual cells. These include improving peptide and protein identification and quantification; characterizing protein-protein interactions and protein complexes; advancing spatial and perturbation proteomics; integrating multi-omics data; and, ultimately, enabling AI virtual cells. Finally, we call for global collaboration among data producers, data consumers and other stakeholders to establish an AI-friendly ecosystem for MS-based proteomics, laying the foundation for transformative advancements in proteomics driven by AI. This Perspective highlights key research areas within mass spectrometry-based proteomics where AI is poised to drive significant advances.