CD8+ T cells differentiate into diverse states that shape immune outcomes in cancer and chronic infection1-4. To define systematically the transcription factors (TFs) driving these states, we built a comprehensive atlas integrating transcriptional and epigenetic data across nine CD8+ T cell states and inferred TF activity profiles. Our analysis catalogued TF activity fingerprints, uncovering regulatory mechanisms governing selective cell state differentiation. Leveraging this platform, we focused on two transcriptionally similar but functionally opposing states that are critical in tumour and viral contexts: terminally exhausted T (TEXterm) cells, which are dysfunctional5-8, and tissue-resident memory T (TRM) cells, which are protective9-13. Global TF community analysis revealed distinct biological pathways and TF-driven networks underlying protective versus dysfunctional states. Through in vivo CRISPR screening integrated with single-cell RNA sequencing (in vivo Perturb-seq) we delineated several TFs that selectively govern TEXterm cell differentiation. We also identified HIC1 and GFI1 as shared regulators of TEXterm and TRM cell differentiation and KLF6 as a unique regulator of TRM cells. We discovered new TEXterm-selective TFs, including ZSCAN20 and JDP2, with no previous known function in T cells. Targeted deletion of these TFs enhanced tumour control and synergized with immune checkpoint blockade but did not interfere with TRM cell formation. Consistently, their depletion in human T cells reduces the expression of inhibitory receptors and improves effector function. By decoupling exhaustion TEX-selective from protective TRM cell programmes, our platform enables more precise engineering of T cell states, accelerating the rational design of more effective cellular immunotherapies.

Magic-angle twisted multilayer graphene stands out as a highly tunable class of moiré materials that exhibit strong electronic correlations and robust superconductivity1-4. However, understanding the relationships between the low-temperature superconducting phase and the preceding correlated parent states remains a challenge. Here we use scanning tunnelling microscopy (STM) and spectroscopy to track the formation sequence of correlated phases established by the interplay of dynamic correlations, intervalley coherence and superconductivity in magic-angle twisted trilayer graphene (MATTG). We discover the existence of two well-resolved gaps pinned at the Fermi level within the superconducting doping range. Although the outer gap, previously associated with the pseudogap phase5,6, persists at high temperatures and magnetic fields, the newly revealed inner gap is more fragile, in line with previous transport experiments1,2,4. Andreev reflection spectroscopy taken at the same location confirms a clear trend that closely follows the doping behaviour of the inner gap and not the outer one. Moreover, spectroscopy taken at nanoscale domain boundaries further corroborates the contrasting behaviour of the two gaps, with the inner gap remaining resilient to structural variations. By comparing our results with recent topological heavy fermion (THF) models that include dynamical correlations7,8, we find that the outer gap probably arises from a splitting of the Abrikosov-Suhl-Kondo resonance9,10 owing to the breaking of the valley symmetry. Our results indicate an intricate yet tractable hierarchy of correlated phases in twisted multilayer graphene.

Multiple mosquito species serve as competent vectors to carry and transmit numerous flaviviruses1,2. Several long-standing scientific questions remain to be answered, including identification of the fundamental factors that facilitate flavivirus infectivity in mosquitoes and the genetic basis that contributes to the naturally occurring interspecies specificity of mosquitoes to flaviviruses3-8, such as Aedes aegypti mosquitoes to dengue virus (DENV). Here we report that circulating mature virions are inactivated by the acidity of mosquito haemolymph; thus, extracellular vesicles carrying replication-competent viral nucleocapsids serve as the predominant means of intercellular viral dissemination. Mechanistically, mosquito valosin-containing protein (VCP) binds to the viral capsid, thereby allowing the incorporation of nucleocapsids into extracellular vesicles. The capsid of a flavivirus (such as DENV) selectively binds to the VCP of its natural vector (Ae. aegypti), but not to that of an incompetent vector (for example, Culex quinquefasciatus). Replacing the DENV capsid with that of Japanese encephalitis virus (JEV) renders DENV infectious in the haemolymph of the natural JEV vector, Cx. quinquefasciatus. Furthermore, two amino residues in Aedes (D723/N728) and Culex (E723/E728) VCP determine its binding specificity for viral capsid, thus contributing to interspecies specificity of mosquitoes to flaviviruses. In vivo ectopic expression of the Cx. quinquefasciatus VCP mutant E723D/E728N renders Cx. quinquefasciatus susceptible to DENV2 via intrathoracic microinjection. Our study provides a major molecular mechanism contributing to the selectivity and compatibility between mosquito vectors and flavivirus species, enabling systemic virus dissemination after the virus reaches the haemocoel. Upstream mechanisms that determine specificity at the midgut level remain to be determined.

Copy number variants (CNVs) are key drivers of human diversity and disease risk1. Here we evaluate the role of CNVs across a broad range of human phenotypes and diseases by analysing CNVs from 470,727 UK Biobank whole-genome sequences and conducting a variant- and gene-level phenome-wide association study (PheWAS) with 2,941 plasma protein abundance measurements, 13,336 binary clinical phenotypes and 1,911 quantitative traits. Proteomic analyses validated functional associations of CNVs with nearby genes (cis-protein quantitative trait loci; cis-pQTLs)-with deletions and duplications typically associated with reduced and increased protein levels, respectively-and uncovered previously unknown protein-protein interactions (trans-pQTLs). Our PheWAS recapitulated known associations and uncovered associations in both coding and non-coding regions. Notably, we identified a rare deletion in ZNF451 associated with increased leukocyte telomere length and a non-coding deletion of a SLC2A9 enhancer associated with reduced gout risk. In addition, by combining CNVs with protein-coding single nucleotide variants and indels, we enhanced the power of our study to detect gene-disease associations. Finally, we leveraged this multiomics dataset to identify several pQTLs that constitute candidate biomarkers, including TMPRSS5 for Charcot-Marie-Tooth disease type 1A. This multiancestry whole-genome-sequence CNV PheWAS offers insights into the roles of CNVs in human health outcomes and could serve as a valuable resource for therapeutic development.

Animals use 18-33-nucleotide PIWI-interacting RNAs (piRNAs) to silence transposons in germ cells1-3. In addition to transposon-silencing piRNAs, placental mammals make pachytene piRNAs4,5, an abundant class of testis-specific small RNAs derived from long noncoding RNA precursors. Although the sites of pachytene piRNA precursor transcription are often conserved among placental mammals, the sequences of the piRNAs themselves are rapidly diverging, even in the human population6. Consequently, the biological function and mechanism of action of pachytene piRNAs remain debated. Here we report that most mouse pachytene piRNAs have no biological function but instead 'selfishly' promote their own production. Our data suggest that pachytene piRNAs direct endonucleolytic cleavage of partially complementary targets and neither activate nor repress mRNA translation. Although many pachytene piRNAs guide cleavage of specific mRNAs, few alter the steady-state abundance of their targets. The minority of pachytene piRNAs that reduce target mRNA abundance enhance sperm fitness, thereby ensuring production of the entire pachytene piRNA repertoire. Together, our findings explain the lack of conservation of most pachytene piRNA sequences and suggest that these 'selfish' small RNAs persist in mammalian evolution because target cleavage by a tiny minority of piRNAs supports male fertility.

ZFTA-RELA+ ependymomas are malignant brain tumours defined by fusions formed between the putative chromatin remodeller ZFTA and the NF-κB mediator RELA1. Here we show that ZFTA-RELA+ cells produce itaconate, a key macrophage-associated immunomodulatory metabolite2. Itaconate is generated by cis-aconitate decarboxylase 1 (ACOD1; also known as IRG1). However, the production of itaconate by tumour cells and its tumour-intrinsic role are not well established. ACOD1 is upregulated in a ZFTA-RELA-dependent manner. Functionally, itaconate enables a feed-forward system that is crucial for the maintenance of pathogenic ZFTA-RELA levels. Itaconate epigenetically activates ZFTA-RELA transcription by enriching for activating H3K4me3 via inhibition of the H3K4 demethylase KDM5. ZFTA-RELA+ tumours enhance glutamine metabolism to supply carbons for itaconate synthesis. Antagonism of ACOD1 or glutamine metabolism reduces pathogenic ZFTA-RELA levels and is potently therapeutic in multiple in vivo models. Mechanistically, ZFTA-RELA epigenetically suppresses PTEN expression to upregulate PI3K-mTOR signalling, a known driver of glutaminolysis. Finally, suppression of ACOD1 or a combination of glutamine antagonism with PI3K-mTOR inhibition abrogates spinal metastasis. Our data demonstrate that ZFTA-RELA+ ependymomas subvert a macrophage-like itaconate metabolic pathway to maintain expression of the ZFTA-RELA driver, which implicates itaconate as a candidate oncometabolite. Taken together, our results position itaconate upregulation as a previously unappreciated driver of ZFTA-RELA+ ependymomas. Our work has implications for future drug development to reduce pathogenic ZFTA-RELA expression for this brain tumour, and will advance our understanding of oncometabolites as a new class of therapeutic dependencies in cancers.

Telomere-to-telomere (T2T) assembly is the ultimate goal for de novo genome assembly. Existing algorithms1,2 capable of near-T2T assembly all require Oxford Nanopore Technologies (ONT) ultra-long reads, which are costly and experimentally challenging to obtain and are thus often unavailable for samples without established cell lines3. Here we introduce hifiasm (ONT), an algorithm that can produce near-T2T assemblies from standard ONT simplex reads, eliminating the need for ultra-long sequencing. Compared with existing methods, hifiasm (ONT) reduces computational demands by an order of magnitude and reconstructs more chromosomes from telomere to telomere on the same datasets. This advance substantially broadens the feasibility of T2T assembly for applications previously limited by the high cost and experimental requirement of ultra-long reads.

The loss of fur during human evolution has long mystified scientists and the public1-5. Reduced hair density coincides with acquisition of epidermal rete ridges, the developmental timing and molecular mechanisms of which are poorly understood despite their prominence in humans1,6-9. Examination of human and pig skin development has shown that rete ridges form through a mechanism independent from those of hair follicles10,11 and sweat glands3,4,12-15 by establishing interconnected epidermal invaginations. Here we document the occurrence of rete ridges across Mammalia, including in grizzly bears and dolphins, and show that neonatal pig wounds can regenerate them de novo. Multispecies spatiotemporal transcriptomics identifies significant signalling interactions between epidermal and dermal cells during rete ridge morphogenesis, particularly through bone morphogenetic proteins (BMP). We also demonstrate that mouse fingerpad skin forms rete ridges and functionally requires epidermal BMP signalling. We propose that evolution of rete ridges in mammalian skin involved replacement of the molecular program for formation of discrete microscopic appendages, including hair follicles and sweat glands, with a distinct program for the interconnected appendage network. Broad epidermal activation of BMP is required for the development of rete ridge networks organized around underlying dermal pockets. Understanding rete ridge mechanisms may enable development of therapeutic approaches to regenerate epidermal appendages lost during wounding or disease in humans.

The essential chaperonin T-complex protein ring complex (TRiC) (also known as chaperonin containing TCP-1 (CCT)) mediates protein folding in cooperation with the co-chaperone prefoldin (PFD)1-5. In vitro experiments have shown that the cylindrical TRiC complex facilitates folding through ATP-regulated client protein encapsulation6-9. However, the functional dynamics of the chaperonin system in vivo remain unexplored. Here we developed single-particle tracking in human cells to monitor the interactions of TRiC-PFD with newly synthesized proteins. Both chaperones engaged nascent polypeptides repeatedly in brief probing events typically lasting around one second, with PFD recruiting TRiC. As shown with the chaperonin client actin8, the co-translational interactions of PFD and TRiC increased in frequency and lifetime during chain elongation. Close to translation termination, PFD bound for several seconds, facilitating TRiC recruitment for post-translational folding involving multiple reaction cycles of around 2.5 s. Notably, the lifetimes of TRiC interactions with a folding-defective actin mutant were markedly prolonged, indicating that client conformational properties modulate TRiC function. Mutant actin continued cycling on TRiC until it was targeted for degradation. TRiC often remained confined near its client protein between successive binding cycles, suggesting that the chaperonin machinery operates within a localized 'protective zone' in which free diffusion is restricted. Together, these findings offer detailed insight into the single-molecule dynamics and supramolecular organization of the chaperonin system in the cellular environment.

Bacteria use diverse mechanisms to protect themselves against phages1-6. Many antiphage systems form large oligomeric complexes, but how oligomerization is regulated during phage infection remains mostly unknown7-12. Here we demonstrate that the bacterial immunity protein ring-activated zinc-finger RNase (RAZR) assembles into an active, 24-meric ring around the circumference of large ring structures formed by two unrelated phage proteins: a putative recombinase and a portal protein. Each multi-layered, megadalton-scale complex enables RAZR to cleave RNA nonspecifically to inhibit translation and restrict phage propagation. The recognition of unrelated phage proteins that form rings with similar diameters indicates that these proteins not only bind to RAZR but also enforce a geometry crucial to activation. The lack of large ring structures in the host probably prevents auto-immunity and RAZR activation before infection. The infection-triggered oligomerization of RAZR mirrors pathogen-induced oligomerization in eukaryotic innate immune complexes13, underscoring a common principle of immunity across biology.

In pursuit of a practical quantum advantage1, analogue quantum systems provide an invaluable way to simulate the physics of quantum materials2-4, quantum systems out of equilibrium5,6 or interaction-induced localization7. Notable recent progress to realize such systems has been achieved in ultracold atoms8-12, superconducting circuits13-15 and twisted van der Waals materials16-19. However, so far, these platforms have struggled to simulate large-scale strongly interacting fermionic systems at low temperatures, at which electronic correlations dominate materials properties and numerical simulations remain restricted in accuracy and scope20,21. Here we demonstrate the realization of a new platform consisting of large-scale 2D arrays of sub-nanometre precision-engineered atom-based quantum dots (15,000 sites) to simulate strongly interacting, low-temperature physics. By observing a metal-insulator (MI) transition on a 2D square lattice of atom-based quantum dots, we demonstrate independent and precise control of the on-site interaction U and tunnelling t. Magneto-transport measurements further indicate the formation of an insulating state driven by Mott-Hubbard/Anderson physics and promising signatures of correlated electron physics. These precision-engineered analogue quantum simulators provide a unique platform to simulate quantum materials on arbitrary 2D lattices and to explore many unanswered questions in the formation of quantum magnetism, interacting topological quantum matter and unconventional superconductivity.

Exposure to cytosolic DNA triggers innate immune responses through cyclic GMP-AMP (cGAMP) synthase (cGAS)1,2,3. After binding to DNA, cGAS produces cGAMP as a second messenger that binds to stimulator of interferon genes (STING), a signalling adaptor protein anchored to the endoplasmic reticulum (ER)3-5. STING then traffics from the ER through the Golgi to perinuclear vesicle clusters, which leads to activation of the kinases TBK1 and IKK and subsequent induction of interferons and other cytokines6-9. Here we show that phosphatidylinositol 3,5-bisphosphate (PtdIns(3,5)P2; also known as PI(3,5)P2) is an endogenous ligand of STING that functions together with cGAMP to induce STING activation. Proteomic analyses identified a constitutive interaction between STING and PIKFYVE, an enzyme that produces PtdIns(3,5)P2 in mammalian cells. Deletion of PIKFYVE blocked STING trafficking from the ER and TBK1 activation. In vitro reconstitution uncovered a strong and selective effect of PtdIns(3,5)P2 on STING activation by cGAMP. PtdIns(3,5)P2 bound directly to STING in fluorescence resonance energy transfer assays. Consistently, cryo-electron microscopy revealed that PtdIns(3,5)P2 promotes cGAMP-induced STING oligomerization10, functioning as a molecular glue. Similar to PIKFYVE depletion, mutation of the PtdIns(3,5)P2-binding residues in STING largely blocked its trafficking and downstream signalling. These findings reveal that PtdIns(3,5)P2 is a lipid ligand of STING with essential roles in innate immunity.

Photoreceptor proteins regulate fundamental biological processes such as vision, photosynthesis and circadian rhythms1. A large photoreceptor subfamily uses vitamin B12 derivatives for light sensing2, contrasting with the well-established mode of action of these organometallic derivatives in thermally activated enzymatic reactions3. The exact molecular mechanism of B12 photoreception and how this differs from the thermal pathways remains unknown. Here we provide a detailed description of photoactivation in the prototypical B12 photoreceptor CarH4,5 from nanoseconds to seconds, combining time-resolved and temperature-resolved structural and spectroscopic methods with quantum chemical calculations. Building on the crystal structures of the initial tetrameric dark and final monomeric light-activated states5, our structural snapshots of key intermediates in the truncated B12-binding domain illustrate how photocleavage of a cobalt-carbon (Co-C) bond within the B12 chromophore adenosylcobalamin triggers a series of structural changes that propagate throughout CarH. Breakage of the photolabile Co-C5' bond leads to the formation of a previously unknown adduct that links the C4' position of the adenosyl moiety to the Co ion and can subsequently be cleaved thermally over longer timescales to allow release of the adenosyl group, ultimately causing tetramer dissociation4,5. This adduct, which differentiates CarH from thermally activated B12 enzymes, steers the photoactivation pathway and acts as the molecular bridge between photochemical and photobiological timescales. The biological relevance of our study is corroborated by kinetic data on full-length CarH in the presence of DNA. Our results offer a spatiotemporal understanding of CarH photoactivation and pave the way for designing B12-dependent photoreceptors for optogenetic applications.

Stimulator of interferon genes (STING) is an essential adaptor in the cytosolic DNA-sensing innate immune pathway1. STING is activated by cyclic GMP-AMP (cGAMP) produced by the DNA sensor cGAMP synthase (cGAS)2-5. cGAMP-induced high-order oligomerization and translocation of STING from the endoplasmic reticulum to the Golgi and post-Golgi vesicles are critical for STING activation6-11. Other studies have shown that phosphatidylinositol phosphates (PtdInsPs) and cholesterol also have important roles in STING activation, but the underlying mechanisms remain unclear12-17. Here we demonstrate that cGAMP-induced high-order oligomerization of STING is enhanced strongly by phosphatidylinositol 3,5-bisphosphate (PtdIns(3,5)P2 and PtdIns(4,5)P2, and by PtdIns4P to a lesser extent. Our cryo-electron microscopy structures reveal that PtdInsPs together with cholesterol bind at the interface between STING dimers, directly promoting the high-order oligomerization. The structures also provide an explanation for the preference of the STING oligomer to different PtdInsPs. Mutational and biochemical analyses confirm the binding modes of PtdInsPs and cholesterol and their roles in STING activation. Our findings shed light on the regulatory mechanisms of STING mediated by specific lipids, which may underlie the role of intracellular trafficking in dictating STING signalling.