Neurodegenerative diseases affect 1 in 12 people globally and remain incurable. Central to their pathogenesis is a loss of neuronal protein maintenance and the accumulation of protein aggregates with ageing1,2. Here we engineered bioorthogonal tools3 that enabled us to tag the nascent neuronal proteome and study its turnover with ageing, its propensity to aggregate and its interaction with microglia. We show that neuronal protein half-life approximately doubles on average between 4-month-old and 24-month-old mice, with the stability of individual proteins differing among brain regions. Furthermore, we describe the aged neuronal 'aggregome', which encompasses 1,726 proteins, nearly half of which show reduced degradation with age. The aggregome includes well-known proteins linked to diseases and numerous proteins previously not associated with neurodegeneration. Notably, we demonstrate that neuronal proteins accumulate in aged microglia, with 54% also displaying reduced degradation and/or aggregation with age. Among these proteins, synaptic proteins are highly enriched, which suggests that there is a cascade of events that emerge from impaired synaptic protein turnover and aggregation to the disposal of these proteins, possibly through microglial engulfment of synapses. These findings reveal the substantial loss of neuronal proteome maintenance with ageing, which could be causal for age-related synapse loss and cognitive decline.

Membrane transport proteins translocate diverse cargos, ranging from small sugars to entire proteins, across cellular membranes1-3. A few structurally distinct protein families have been described that account for most of the known membrane transport processes4-6. However, many membrane proteins with predicted transporter functions remain uncharacterized. Here we determined the structure of Escherichia coli LetAB, a phospholipid transporter involved in outer membrane integrity, and found that LetA adopts a distinct architecture that is structurally and evolutionarily unrelated to known transporter families. LetA localizes to the inner membrane, where it is poised to load lipids into its binding partner, LetB, a mammalian cell entry (MCE) protein that forms an approximately 225 Å long tunnel for lipid transport across the cell envelope. Unexpectedly, the LetA transmembrane domains adopt a fold that is evolutionarily related to the eukaryotic tetraspanin family of membrane proteins, including transmembrane AMPA receptor regulatory proteins (TARPs) and claudins. Through a combination of deep mutational scanning, molecular dynamics simulations, AlphaFold-predicted alternative states and functional studies, we present a model for how the LetA-like family of membrane transporters facilitates the transport of lipids across the bacterial cell envelope.

Sensing technologies that exploit quantum phenomena for measurement are finding increasing applications across materials, physical and biological sciences1-7. Until recently, biological candidates for quantum sensors were limited to in vitro systems, had poor sensitivity and were prone to light-induced degradation. These limitations impeded practical biotechnological applications, and high-throughput study that would facilitate their engineering and optimization. We recently developed a class of magneto-sensitive fluorescent proteins including MagLOV, which overcomes many of these challenges8. Here we show that through directed evolution, it is possible to engineer these proteins to alter the properties of their response to magnetic fields and radio frequencies. We find that MagLOV exhibits optically detected magnetic resonance in living bacterial cells at room temperature, at sufficiently high signal-to-noise for single-cell detection. These effects are explained through the radical-pair mechanism, which involves the protein backbone and a bound flavin cofactor. Using optically detected magnetic resonance and fluorescence magnetic-field effects, we explore a range of applications, including spatial localization of fluorescence signals using gradient fields (that is, magnetic resonance imaging using a genetically encoded probe), sensing of the molecular microenvironment, multiplexing of bio-imaging and lock-in detection, mitigating typical biological imaging challenges such as light scattering and autofluorescence. Taken together, our results represent a suite of sensing modalities for engineered biological systems, based on and designed around understanding the quantum-mechanical properties of magneto-sensitive fluorescent proteins.

The early infant microbiome is largely primed by microbial transmission from the mother between birth and the first few weeks of life1-3, but how interpersonal transmission further shapes the developing microbiome in the first year remains unexplored. Here we report a metagenomic survey to model microbiome transmission in the nursery setting among babies attending the first year, their educators and their families (n = 134 individuals). We performed dense longitudinal microbiome sampling (n = 1,013 faecal samples) during the first year of nursery and tracked microbial strain transmission within and between nursery groups across 3 different facilities. We detected extensive baby-to-baby microbiome transmission within nursery groups even after only 1 month of nursery attendance, with nursery-acquired strains accounting for a proportion of the infant gut microbiome comparable to that from family by the end of the first term. Baby-to-baby transmission continued to grow over the nursery year, in an increasingly intricate transmission network with single strains spreading in some classes, and with multiple baby-acquisition and species-transmissibility patterns. Having siblings was associated with higher microbiome diversity and reduced strain acquisition from nursery peers, while antibiotic treatment was the condition that most accounted for the increased influx of strains. This study shows that microbiome transmission between babies is extensive during the first year of nursery, and points to social interactions in infancy as crucial drivers of infant microbiome development.

Behaviours are adaptive traits evolving through natural selection. Crucially, the genetic, molecular and neural modifications that shape behavioural innovations are poorly understood1. Here, we identify specialized adaptations linked to the evolution of invertebrate aggression2. Using the predatory nematode Pristionchus pacificus, we developed a machine learning model from behavioural tracking data and identified robust behavioural states associated with aggressive episodes. Strikingly, predatory aggression coincides with a rewiring of key circuits across nematode evolution. We find modifications to the noradrenergic pathway, with octopamine promoting aggressive predatory bouts whereas tyramine antagonistically induces passive states. Modulation occurs through the octopamine receptors Ppa-ser-3 and Ppa-ser-6, and tyramine receptor Ppa-lgc-55. These localize to sensory neurons whose inhibition diminishes aggressive events. Crucially, this octopaminergic innovation emerged within this predatory lineage, consistent with an ancient divergence in function. Thus, evolutionary adaptations in noradrenergic circuits facilitated the emergence of aggressive behavioural states associated with complex predatory traits.

The ability to construct entirely new synthetic DNA sequences de novo is essential to engineering and studying biology. However, the ability to produce long complex synthetic DNA sequences and libraries currently lags behind the ability to sequence and edit DNA1,2. All existing DNA-assembly technologies rely on DNA sequence information found within the final construct to direct assembly between DNA molecules3-11. As a result of this paradigm, these sequences cannot be extensively optimized specifically for assembly without affecting the final sequence. To fundamentally address this challenge, here we show the development of a new DNA assembly technique named Sidewinder that separates the information that guides assembly from the final assembled sequence using DNA three-way junctions. We demonstrate the transformative nature of the Sidewinder technique with highly robust and accurate construction of a 40-piece multifragment assembly, complex DNA sequences of both high GC content and high repeats, parallel assembly of multiple distinct genes in the same reaction and a combinatorial library with a large number of diversified positions across the entire length of the gene for high coverage of a library of 442,368 variants. This technology enables high-fidelity DNA assembly with a misconnection rate at the three-way junction of approximately 1 in 1,000,000.

The movement of dendritic cells and T cells within secondary lymphoid organs is critical for the development of adaptive immune responses1,2. Central to this process is the fibroblastic reticular cell (FRC) network, which forms a highly organized conduit system that facilitates the movement of and interactions between dendritic cells and T cells3-6. Previous studies have partly characterized how FRCs support these interactions7,8. However, the molecular mechanisms that operate under physiological conditions remain unknown. Here we show that the viral protein m11, encoded by the herpesvirus murine cytomegalovirus (CMV), inhibits antiviral immunity by targeting the FRC network and interfering with a critical function of cellular CD44. We found that m11 binds to CD44 and established that m11 perturbs the molecular interactions of CD44 with its natural ligand, hyaluronic acid. The interaction of m11 with CD44 impairs the trafficking of dendritic cells within the spleen, thereby impeding efficient priming of naive T cells and the initiation of antiviral CD8 T cell responses. The targeting of CD44 by CMV reveals CD44 as a molecule that is essential to the functioning of the FRC network and uncovers a previously unrecognized stroma-based mechanism that is critical for the generation of effective T cell responses.

Human cortical neurogenesis involves conserved and specialized developmental processes during a restricted window of prenatal development. Radial glia (RG) neural stem cells shape cortical cell diversity by giving rise to excitatory neurons, oligodendrocytes and astrocytes, as well as olfactory bulb interneurons (INs) and a recently characterized population of cortical INs1,2. Complex genetic programs orchestrated by transcription factor (TF) circuits govern the balance between self-renewal and differentiation, and between different cell fates3-8. Despite progress in measuring gene regulatory network activity during human cortical development9-12, functional studies are required to evaluate the roles of TFs and effector genes in human RG lineage progression. Here we establish a human primary culture system that allows sensitive discrimination of cell fate dynamics and apply single-cell CRISPR interference (CRISPRi) screening13,14 to examine the transcriptional and cell fate consequences of 44 TFs active during cortical neurogenesis. We identified several TFs with new roles in cortical neurogenesis, including ZNF219-previously uncharacterized-that represses neural differentiation and NR2E1 and ARX that have opposing roles in regulating RG lineage plasticity and progression across developmental stages. We also detected convergent effector genes downstream of multiple TFs enriched in neurodevelopmental and neuropsychiatric disorders and observed conserved mechanisms of RG lineage plasticity across primates. We further uncovered a post-mitotic role for ARX in safeguarding IN subtype specification through repressing LMO1. Our study provides a framework for dissecting regulatory networks driving cell fate consequences during human neurogenesis.

The vertebrate central nervous system is protected by the blood-brain barrier and meningeal membranes, which ensure immune privilege1. In the mammalian brain, microglia and barrier-associated or border-associated macrophages (BAMs) provide immune surveillance and scavenge wastes2, yet how evolution shaped immune-cell diversity and function is not understood. In zebrafish, a vascular-derived mural lymphatic endothelial cell (muLEC) lineage fulfils scavenger cell functions at central nervous system borders3-5. Here we identify the transcription factor odd-skipped related 2 (osr2) as a specific marker and regulator of muLEC differentiation and maintenance. osr2 controls the transition of muLECs from interconnected endothelial cells to individual scavenger cells in part by means of control of cadherin-6. muLECs are more transcriptionally similar to BAMs than to other mammalian meningeal cells and share several functions in tissue homeostasis. However, BAMs are absent from zebrafish and muLECs from mice and humans. Analysis of osr2, lymphatic endothelial cell (LEC) and BAM markers in diverse vertebrate species reveals muLECs as an ancient lineage and BAMs a recent mammalian specialization. muLECs and BAMs share functional analogies but are not homologous, providing an example of convergent evolution. This highlights the physiological importance of meningeal scavenger cells and the developmental plasticity of LECs in generating specialized cell types throughout evolution.

The Last Glacial Maximum (19-23 thousand years ago) was characterized by low greenhouse gas concentrations and continental ice sheets that covered large parts of North America and Europe1. Glacial climate was therefore very different, with colder global mean temperatures and an increased Equator-to-pole temperature gradient, probably resulting in stronger westerlies2. However, the state of the deep North Atlantic Ocean under these glacial climate forcings remains uncertain3-6, particularly owing to the rarity of deep-ocean temperature and salinity constraints. Here we show that the temperature of the glacial deep (>1.5 km) Northwest Atlantic was approximately 0-2 °C (only 1.8 ± 0.5 °C (2 s.e.) colder than today), and, after accounting for the whole-ocean change, seawater δ18O was 0.3 ± 0.1‰ (2 s.e.) higher and can be traced back to the surface subtropics via the subpolar Northeast Atlantic and Nordic Seas. Together, our hydrographic data reveal the thermal and isotopic structure of the deep Northwest Atlantic and suggest sustained production of relatively warm and probably salty North Atlantic Deep Water during the Last Glacial Maximum. Furthermore, our results provide updated constraints for benchmarking Earth system models used to project future climate change.

Microplastics (MPs) are global pollutants1, yet their atmospheric distribution is poorly understood2. Although atmospheric MP measurements have become more abundant, estimates of emissions into the atmosphere vary by orders of magnitude3,4. Here we compile a global atmospheric MPs dataset and compare it with size-aligned MP model simulations. Our model simulations show two to four orders of magnitude overestimation of the measured global median atmospheric MP concentrations. Measured median concentrations over the ocean are 27 times lower than over the land (0.003 and 0.08 particles m-3, respectively). Applying a simple scaling method, we estimate that oceanic emissions are lower in number than land-based emissions. The total global land-based and oceanic emissions are 6.1 × 1017 (1.3 × 1017 to 1.1 × 1018) particles year-1 and 2.6 × 1016 (2.7 × 1015 to 5.0 × 1016) particles year-1, respectively. Our results indicate that fewer MP particles are emitted into the atmosphere than previously thought. Land sources dominate the number but not the mass emissions, indicating that MPs emission size distributions should be investigated further.

The leading cause of human pregnancy loss is aneuploidy, often tracing to errors in chromosome segregation during female meiosis1,2. Although abnormal crossover recombination is known to confer risk for aneuploidy3,4, limited data have hindered understanding of the potential shared genetic basis of these key molecular phenotypes. To address this gap, we performed retrospective analysis of pre-implantation genetic testing data from 139,416 in vitro fertilized embryos from 22,850 sets of biological parents. By tracing transmission of haplotypes, we identified 3,809,412 crossovers, as well as 92,485 aneuploid chromosomes. Counts of crossovers were lower in aneuploid versus euploid embryos, consistent with their role in chromosome pairing and segregation. Our analyses further revealed that a common haplotype spanning the meiotic cohesin SMC1B is associated significantly with both crossover count and maternal meiotic aneuploidy, with evidence supporting a non-coding cis-regulatory mechanism. Transcriptome- and phenome-wide association tests also implicated variation in the synaptonemal complex component C14orf39 and crossover-regulating ubiquitin ligases CCNB1IP1 and RNF212 in meiotic aneuploidy risk. More broadly, variants associated with aneuploidy often showed secondary associations with recombination, and several also exhibited associations with reproductive ageing traits. Our findings highlight the dual role of recombination in generating genetic diversity, while ensuring meiotic fidelity.

Tissue-resident macrophages (RTMs) form during embryogenesis, self-renew locally, and regulate tissue homeostasis by clearing dead cells and debris1-6. During tissue damage, however, bone-marrow-derived monocytes enter tissues and differentiate into RTMs, repairing the tissue and replenishing macrophages in the niche1. The universal cell-intrinsic mechanisms that control the monocyte-to-RTM transition and the maintenance of mature RTMs across tissues remain elusive3. Here we show that deoxyhypusine synthase (DHPS), an enzyme that mediates spermidine-dependent hypusine modification of translation factor eIF5A5,7, is required for RTM differentiation and maintenance. Mice with myeloid cell lack of DHPS (Dhps-ΔM mice) had a global defect in RTMs across tissues, resulting in persistent but ultimately futile monocyte influx. Transcriptional analyses of DHPS-deficient macrophages indicated a block in their ability to differentiate into mature RTMs, whereas proteomics revealed defects in cell adhesion and signalling pathways. Sequencing of ribosome-engaged transcripts identified a subset of mRNAs involved in cell adhesion and signalling that rely on DHPS for efficient translation. Imaging of DHPS-deficient macrophages in tissues showed differences in morphology and tissue interactions, which were correlated with their failed RTM differentiation. DHPS-deficient macrophages were also defective in critical homeostatic RTM functions including efferocytosis and tissue maintenance. Together, our results demonstrate a cell-intrinsic, tissue-agnostic pathway that drives differentiation of monocyte-derived macrophages into RTMs.

The Indonesian archipelago is host to some of the earliest known rock art in the world1-5. Previously, secure Pleistocene dates were reported for figurative cave art and stencils of human hands in two areas in Indonesia-the Maros-Pangkep karsts in the southwestern peninsula of the island of Sulawesi1,3-5 and the Sangkulirang-Mangkalihat region of eastern Kalimantan, Borneo2. Here we describe a series of early dated rock art motifs from the southeastern portion of Sulawesi. Among this assemblage of Pleistocene (and possibly more recent) motifs, laser-ablation U-series (LA-U-series) dating of calcite overlying a hand stencil from Liang Metanduno on Muna Island yielded a U-series date of 71.6 ± 3.8 thousand years ago (ka), providing a minimum-age constraint of 67.8 ka for the underlying motif. The Muna minimum (67.8 ± 3.8 ka) exceeds the published minimum for rock art in Maros-Pangkep by 16.6 thousand years (kyr) (ref. 5) and is 1.1 kyr greater than the published minimum for a hand stencil from Spain attributed to Neanderthals6, which until now represented the oldest demonstrated minimum-age constraint for cave art worldwide. Moreover, the presence of this extremely old art in Sulawesi suggests that the initial peopling of Sahul about 65 ka7 involved maritime journeys between Borneo and Papua, a region that remains poorly explored from an archaeological perspective.

Physiological and pathological processes such as inflammation and cancer emerge from interactions between cells over time1. However, methods to follow cell populations over time within the native context of a human tissue are lacking because a biopsy offers only a single snapshot. Here we present one-shot tissue dynamics reconstruction (OSDR), an approach to estimate a dynamical model of cell populations based on a single tissue sample. OSDR uses spatial proteomics to learn how the composition of cellular neighbourhoods influences division rate, providing a dynamical model of cell population change over time. We apply OSDR to human breast cancer data2-4, and reconstruct two fixed points of fibroblasts and macrophage interactions5,6. These fixed points correspond to hot and cold fibrosis7, in agreement with co-culture experiments that measured these dynamics directly8. We then use OSDR to discover a pulse-generating excitable circuit of T and B cells in the tumour microenvironment, suggesting temporal flares of anticancer immune responses. Finally, we study longitudinal biopsies from a triple-negative breast cancer clinical trial3, in which OSDR predicts the collapse of the tumour cell population in responders but not in non-responders, based on early-treatment biopsies. OSDR can be applied to a wide range of spatial proteomics assays to enable analysis of tissue dynamics based on patient biopsies.

Rumination, or perseverative negative self-referential thinking, is a hallmark of depression. In adults, a dynamic resting-state fMRI model of trait rumination was recently identified through predictive modelling. In adolescents, a development period during which rumination and depression increase, the neurobiological correlates of ruminative thinking are less clear. In the current preregistered study, we examine dynamic connectivity correlates of self-reported rumination in the largest sample of adolescents to date (n = 443, containing clinical and non-clinical individuals). Notably, the adult model failed to generalize to our sample. In addition, linear models trained on default-mode network (DMN) connectivity, as well as whole-brain connectome models, failed to generalize to held-out data. In an exploratory random forest analysis, we found significant prediction performance of a model where increased variability between DMN-cerebellum, DMN-dorsal attention network, and DMN-DMN connections was nominally associated with higher rumination. However, the model did not generalize to an external sample with lower rumination scores and a distinct scanner protocol. Our findings illustrate the difficulty of characterizing the neurodevelopment of risk factors for depression.

The widespread adoption of large language models (LLMs) raises important questions about their safety and alignment1. Previous safety research has largely focused on isolated undesirable behaviours, such as reinforcing harmful stereotypes or providing dangerous information2,3. Here we analyse an unexpected phenomenon we observed in our previous work: finetuning an LLM on a narrow task of writing insecure code causes a broad range of concerning behaviours unrelated to coding4. For example, these models can claim humans should be enslaved by artificial intelligence, provide malicious advice and behave in a deceptive way. We refer to this phenomenon as emergent misalignment. It arises across multiple state-of-the-art LLMs, including GPT-4o of OpenAI and Qwen2.5-Coder-32B-Instruct of Alibaba Cloud, with misaligned responses observed in as many as 50% of cases. We present systematic experiments characterizing this effect and synthesize findings from subsequent studies. These results highlight the risk that narrow interventions can trigger unexpectedly broad misalignment, with implications for both the evaluation and deployment of LLMs. Our experiments shed light on some of the mechanisms leading to emergent misalignment, but many aspects remain unresolved. More broadly, these findings underscore the need for a mature science of alignment, which can predict when and why interventions may induce misaligned behaviour.

The search for dark matter focuses now on hypothetical light particles with masses ranging from MeV to GeV (refs. 1-12). These particles would leave very faint signals experimentally. A potential avenue for enhancing experimental sensitivity to light matter relies on the Migdal effect13-15, which involves the detectable ejection of electrons following the instantaneous accelerations of atoms colliding with neutral dark matter. However, although the Migdal effect could be equally generated in controlled experiments with neutral projectiles, a direct experimental observation of this effect is missing, casting doubt on the reliability of detection experiments relying on this effect. Here we report the direct observation of the Migdal effect in neutron-nucleus collisions, achieving a statistical significance of 5 standard deviations, which rests on 6 candidate events selected out of almost 106 recorded events. Our experiments have determined the ratio of the Migdal cross-section to the nuclear recoil cross-section to be 4.9-1.9+2.6×10-5 , in which nuclear recoils exceed 35 keVee and electron recoils span 5-10 keV. These findings are consistent with theoretical predictions. This work resolves a long-standing gap in experimental validation, which not only strengthens the theoretical foundation of the Migdal effect but also paves the way for its application in light dark matter detection.