Poly(ADP-ribose) polymerase inhibitors (PARPi) exploit homologous recombination (HR) deficiency in BRCA1/2-mutant cancers to induce synthetic lethality. PARPi also kill ataxia telangiectasia-mutated (ATM)-deficient cells, but the underlying mechanism has remained unclear. Here, we define the mechanism of PARPi cytotoxicity in mammalian ATM-deficient cells, uncovering a critical role for ATM in DNA replication. In the absence of ATM, unrestrained PRIMPOL-dependent repriming at oxidative base lesions generates post-replicative DNA gaps that activate PARP. This defect is driven by aberrant BRCA1-A recruitment to stalled replication forks in ATM-deficient cells, suppressing HLTF/ZRANB3-mediated replication fork slowing. These DNA gaps require HR for repair and underlie PARPi synthetic lethality. Suppressing repriming or base excision repair, or reducing oxidative stress, alleviates these defects. Our findings reveal how spontaneous oxidative DNA damage synergizes with dysregulated replication to drive PARPi sensitivity, establishing a paradigm of post-replicative repair addiction in ATM-deficient cells and linking oxidative DNA damage to genome instability in ataxia telangiectasia.

Despite the significant progress made in recent years, current AI-driven protein engineering methods often suffer from iterative experimental validation, limited success rates, modest improvements, and/or high computational costs. Moreover, their utility for evolving plant proteins remains unexplored. Here we present unZipro, an efficient, scalable, and generalizable framework for zero-shot, in silico protein evolution. unZipro integrates a compact, pre-trained inverse folding model with meta-learning to derive family-specific fitness landscapes. We demonstrate unZipro's ability to directly identify high-activity variants from minimal libraries (∼10 candidates) with an average success rate of 61% across nine diverse proteins, achieving up to a 28-fold increase in the gene-editing activity of T5E-CasΦ2 fusion variants. Leveraging unZipro, we developed state-of-the-art plant gene-editing nucleases, superior firefly luciferase variants, enhanced rice transcription factors, and barley-derived antiviral proteins with elevated potency. Overall, unZipro represents a new paradigm for cost-effective, widely accessible, and transformative protein engineering in biology, biotechnology, and agriculture.

Cognitive functions require synaptic plasticity, specifically long-term potentiation (LTP). LTP is thought to require Ca[2+]/CaM-dependent protein kinase II (CaMKII) binding to the NMDA-type glutamate receptor subunit GluN2B, but this poses a conundrum: truncated CaMKII monomers (without the hub domain that forms 12meric holoenzymes) fail to bind GluN2B within cells but still potentiate synapses when made constitutively active. We hypothesized that CaMKII monomer binding to GluN2B has just eluded detection. Instead, even though full-length CaMKII monomers (with hub-domain mutations) were found to indeed bind and even form condensates with GluN2B, truncated monomers were not. Nonetheless, truncated monomers still potentiated synapses, even in neurons with GluN2B mutations that ablate CaMKII binding. However, potentiation occurred only with monomers that were made Ca[2+]-independent by artificial phosphatase-resistant thio-autophosphorylation, not by regular autophosphorylation of T286. These findings support that CaMKII binding to GluN2B is required during physiological LTP induction because it generates the phosphatase-resistant autonomous activity that mediates LTP expression.

Linker histone H1 is an essential chromatin architectural protein that compacts chromatin into transcriptionally silent regions by interacting with nucleosomal and linker DNA while rapidly exchanging in vivo. How H1 targets nucleosomes while being dynamic remains unanswered. Using a single-molecule strategy, we investigated human H1.0 interactions with DNA and nucleosomes. H1.0 directly binds nucleosomes and naked DNA with a preference toward nucleosomes. DNA-bound H1.0 exhibited a range of bound lifetimes with both mobile and immobile states, where H1.0 was largely mobile, diffused over kilobase distances in seconds, but surprisingly did not load onto nucleosomes. Linker histone chaperones significantly facilitated H1.0-nucleosome loading by enabling H1.0 loading through DNA sliding, reducing DNA resident times without impacting nucleosome resident times, increasing mobility along DNA, and targeting H1.0 loading onto the nucleosome dyad. These findings reveal the multiple distinct mechanisms linker histones and their chaperones use to interact with nucleosomes to regulate chromatin accessibility.

Programmable DNA integration using CRISPR-associated transposase elements (CASTs) offers powerful capabilities for genome engineering. The large single effector Cas12k CAST examples evolved from a minimal TnpB nuclease protein. Here, we engineer a de novo RNA-guided transposition systems in bacteria, where the single guide RNA effector components are repurposed nuclease-dead TnpB-family proteins. These compact systems mediate high-efficiency guide-RNA-directed DNA insertion with preserved orientation control, target immunity, and release of a host factor requirement and can be paired with an exonuclease domain to mediate cut-and-paste transposition. In this engineered context, the TnpB derivatives show features not predicted from the original enzymes, suggesting untapped avenues for improvement. In parallel, we show that mutations at the TniQ-TnsC interface in the Cas12k CAST system selectively attenuate off-site insertions while enhancing on-site activity. These results establish how Cas12 proteins and antecedent TnpB proteins can be engineered for high performance and specificity with guide-RNA-directed systems.

Traditional deep mutational scanning (DMS) encodes every single amino-acid substitution from a wild-type sequence. We hypothesize that combinatorial DMS (CDMS) libraries, incorporating all mutations in all combinations, can enable the discovery of high-affinity protein (super)binders by capturing epistatic, non-linear amino-acid interactions. Here, we introduce origin-independent and context-exhaustive high-throughput integration of combinatorial DMS libraries (ORCHID), which systematically maps regions of wild-type-independent epistasis across all mutational contexts and trajectories. For benchmarking, we build a high-throughput peptide display assay measuring PIN1[WW-domain] affinity for a CDMS peptide library containing phosphoserine via amber codon suppression. ORCHID raises prediction accuracy by 45% over non-epistatic models. We identify and structurally characterize two epistatic superbinders, SPY-tide and LYR-tide, binding 3- to 5-fold tighter than current optimal PIN1[WW-domain] binders through "fold-and-turn" and register-shifted conformational changes. We also identify two molecular determinants of epistasis, PIN1[F25] and PIN1[R14], that natively encode non-linear binding and, when mutated, abolish it. Hence, natural proteins recognize peptides non-linearly, offering opportunities for improved binder design.

Rapid protein turnover is essential for cellular stress adaptation. HUWE1 (HECT, UBA, and WWE domain containing 1), a large HECT-type E3 ligase, regulates many short-lived stress-responsive proteins, yet the mechanisms underlying its substrate selectivity remain unclear. Here, we reveal that HUWE1 functions as a ubiquitin chain amplifier that captures pre-ubiquitinated substrates and amplifies the degradation signal by assembling long ubiquitin chains containing K11-K48 branch points, a process regulated by its partners HUWE1-associated protein stress response 1 (HAPSTR1) and USP7 (ubiquitin-specific-processing protease 7). Structural and biochemical analyses show that HAPSTR1 engages HUWE1's ubiquitin-binding motifs to drive nuclear import and modulate substrate recruitment. A cryo-EM structure of the HUWE1-USP7 complex reveals a bidirectional regulatory mechanism: HUWE1 activates USP7's catalytic activity, while USP7 modulates HUWE1 conformational states. Global proteomic analyses demonstrate that this axis drives extensive remodeling of the short-lived nuclear proteome. These findings establish the HUWE1-HAPSTR1-USP7 complex as a key ubiquitin code modifier, providing a molecular rationale for HUWE1 dysregulation in neurodevelopmental disorders and cancer.

In a recent issue of Science, Gao et al.[1] discover a link between pattern-triggered immunity (PTI) initiated by plant cell surface receptors and effector-triggered immunity (ETI) initiated upon detection of pathogen effectors by intracellular nucleotide-binding leucine-rich repeat (NLR) immune receptors. PTI activates alternative splicing that results in mRNAs for the potato NLR Rpi-vnt1.1 that lack a repressive N-terminal prodomain.

In this issue of Molecular Cell, Koo et al.[1] describe how the yeast Rad55-Rad57-SHU complex dynamically nucleates Rad51 filaments on DNA. Functional coupling between Rad55-Rad57 and Rad51 on ssDNA seeds filament growth, whilst ATP hydrolysis promotes complex turnover.

Cohesin organizes 3D genome architecture yet its acute depletion in vitro minimally affects transcription, leaving its regulatory role puzzling. We hypothesized that cohesin function is relevant when first encounters chromatin after mitosis. Using in vivo labeling and cell-cycle synchronization in human cells, we demonstrate that cohesin complexes persist through mitosis. This inherited pool dominates over newly translated cohesin during telophase-G1, highlighting the relevance of such transmission. Acute depletion of cohesin Rad21 at mitotic exit impairs the expression of hundreds of genes, mostly downregulated, enriched on functions related to tissue growth and development. Rad21 binds the promoter of this group of genes. Remarkably, we uncover a direct interaction during the M-G1 transition between Rad21 and Cdk9, the kinase component of transcription elongation factor b (pTEFb) complex. Rad21 loss impairs Cdk9 chromatin association, reducing active-elongating RNAPol2 at down-regulated genes. Overall, preserving cohesin across mitosis enables post-mitotic transcriptional reactivation via Cdk9, establishing tissue-specific expression programs.

Kaposin is the most abundantly expressed viral RNA in tumors caused by the oncogenic virus Kaposi's sarcoma-associated herpesvirus (KSHV); however, its role in viral replication is not understood. Here, we show that kaposin, previously viewed as a protein-coding transcript, exists primarily as a nuclear viral long non-coding RNA (lncRNA) that rebuilds cellular nuclear speckles (NSs) adjacent to the viral genome to enhance viral gene expression. Kaposin is both necessary and sufficient to drive substantial NS remodeling, and this effect depends on repetitive elements within the RNA. Absence of kaposin-mediated NS remodeling, depletion of the essential NS protein, serine/arginine repetitive matrix 2 (SRRM2), or steric blocking of the kaposin repetitive elements impair viral gene expression. This work defines kaposin as a viral architectural RNA that drives nuclear speckle seeding beside the viral genome and reframes our understanding of lncRNA function and the spatial organization of transcription in the infected cell nucleus.

Faithful gene expression requires coordination between RNA polymerase II elongation and co-transcriptional RNA processing, yet the chromatin-associated assemblies supporting this coupling remain incompletely defined. Using native chromatin proteomics in human cells, we identify ECAS (elongation complex associated with splicing), a bipartite assembly linking an elongation module (AFF4-ELL2-EAF1) to a U2 snRNP-associated intron binding complex. ECAS occupies promoter-proximal chromatin and is molecularly distinct from canonical P-TEFb-containing super elongation complex (SEC). Acute perturbation of Aquarius (AQR)/XAB2-dependent ECAS functions impairs productive gene-body elongation without altering pause release, increases intron retention, and preferentially affects long genes. AQR loss also reduces capping-factor recruitment and nascent 5' capping and increases promoter-proximal accumulation of cleavage/polyadenylation and RNA exosome factors. In mouse hematopoietic cells, Aqr is highly expressed, and its perturbation profoundly impairs leukemia initiation, maintenance, and leukemia stem cell function in MLL-rearranged leukemia. Together, these findings establish ECAS as a chromatin-associated assembly that couples productive elongation with RNA processing.

5-Azacytidine (5-azaC) is a DNA hypomethylating agent clinically used in myeloid malignancies. As a ribonucleoside analogue, 5-azaC incorporates more readily into nascent RNA than DNA. Here, we demonstrate that RNA 5-methylcytosine (m[5]C) depletion by 5-azaC treatment, particularly at early time points, is sufficient to induce leukemia cell death. In contrast to its DNA demethylation function, the RNA-dependent effects of 5-azaC lead to transcriptional repression, disrupting genes involved in cell-cycle regulation and DNA repair. Mechanistically, depletion of m[5]C by 5-azaC in chromatin-associated RNA (caRNA) disrupts the MBD6-mediated H2AK119ub deubiquitination. This also impairs SRSF2 recruitment and the downstream H3K27ac deposition by p300. Consistently, loss of the caRNA methyltransferase NSUN2 caused prolonged cell cycle, defective DNA repair, and shifted hematopoietic lineage commitment toward erythropoiesis, mirroring the effects of 5-azaC treatment. Our findings highlight the transcriptional repression by 5-azaC through depleting caRNA m[5]C, providing additional insights into the mechanism of action for 5-azaC.

Mouse pericentric heterochromatin forms compacted, transcriptionally silent domains, termed chromocenters, that are enriched in heterochromatin protein 1 (HP1). Whether HP1α represses chromocenters by binding locally or by maintaining a phase-separated compartment is unresolved. We investigated this question by recruiting transcriptional activators to mouse fibroblast chromocenters and quantifying repression at a transcription reporter. HP1α established a promoter-proximal barrier that suppressed weaker activators (VP16) but was overcome by stronger ones (VP64-p65-Rta [VPR] and p65). Activator-induced decondensation and transcription occurred without displacing HP1α or H3K9 trimethylation, and HP1α retained its granular distribution and stoichiometric binding dynamics. Multi-color super-resolution imaging revealed spatial segregation of transcribed and HP1α-bound regions at the nanodomain scale. A nanodomain model captured how repeat clusters independently transition between silenced and activated states. These data establish that HP1α and H3K9me3 act through independently switchable nanodomains, accounting for chromocenter regulation without invoking phase separation.

Bacterial gene expression is strongly influenced by local mRNA secondary structure, yet the impact of long-range folding remains poorly understood. Here, we show that sequences hundreds of nucleotides from the mRNA 5' end can act as potent repressors of gene expression through long-range base pairing to the ribosome binding site (RBS), subjecting anti-RBS sequences to negative selection. Using massively parallel reporter assays in Bacillus subtilis, we identify anti-RBS sequences as among the strongest determinants of reduced mRNA abundance across the transcript body. We demonstrate that distal anti-RBS elements engage in long-range folding with the Shine-Dalgarno sequence, blocking ribosome entry and promoting mRNA decay. Consistent with these repressive effects, anti-RBS-like sequences are depleted throughout diverse bacterial coding sequences but not from leaderless transcripts, and introducing distal anti-RBS to native genes reduces expression. Our findings establish that long-range mRNA folding is a conserved force shaping gene expression and constrains coding sequence evolution.