Biology subjects

Du, S.

Publications and source records attributed to Du, S..

At least 19 recordsLinked to original sources

bioRxiv2026

A self-organized geometry sensing oscillator spatially regulates cell division in archaea

Cell division must be tightly regulate in time and space. Many bacteria and eukaryotes employ geometry sensing systems to regulate cell division spatiotemporally. However, how archaea ensure they divide in the right place at the right time is little understood. Here, we report the discovery of a geometry sensing oscillator that determines the division plane in the model haloarcheon Haloferax volcanii, which relies on two tubulin-like proteins, FtsZ1 and FtsZ2, for division. This archaeal oscillator (named as Arco system) is composed of three components, ArcoA, ArcoB and ArcoC, which are a membrane associated Ras superfamily GTPase, a SepF-like protein and an antagonist of FtsZ1, respectively. ArcoA interacts with both ArcoB and ArcoC. ArcoAB sense the geometry of the cell to oscillate between the two cell poles, generating a time-averaged ArcoC gradient which is high at the cell poles but low at the midcell, where FtsZs can assemble into the Z ring to initiate cytokinesis. As a result, deletion of the Arco system caused aberrant FtsZ assembly throughout the cell, leading to irregular division and formation of minicells. Interestingly, the Arco system is globally distributed among archaea and co-occurs with FtsZ1, suggesting that it emerged in the last archaeal common ancestor (LACA). The oscillatory behavior and phenotypes of the archaeal Arco system is remarkably analogous to the bacterial Min system and yet share no protein homology, representing a remarkable example of convergent evolution. Overall, these results suggest that archaea, similar to bacteria and eukaryotes, independently evolved self-organized Turing reaction diffusion systems to sense geometry and spatially regulate cell division.

microbiology↗

bioRxiv2026

The brain-meningeal interface functions as a reservoir and entry site for brain parenchymal macrophages

Microglia are regarded as self-maintaining brain parenchymal macrophages without contribution from adult hematopoiesis. Nevertheless, peripheral macrophage engraftment into the brain has been reported, but the biological variables governing central nervous system (CNS) macrophage niche access remain unclear. We show that CNS macrophage engraftment is determined by the balance between tissue-resident macrophage (TRM) self-renewal and the temporal alignment of niche opening with the availability of engraftment-competent cells and proximity to the vacated niche, rather than prolonged niche vacancy. Fate mapping revealed that monocytes entering the subdural space undergo border-associated macrophage (BAM)-like differentiation, clonal expansion, and transpial migration into the parenchyma, whereas mature BAMs directly repopulate selectively vacated parenchymal niches. We further identify a parenchymal macrophage population with a peripheral BAM-like transcriptional program in aged and neurodegenerative human brains, challenging the dogma of MG-exclusivity. Together, these findings establish a predictive framework for interpreting macrophage maintenance and replacement and for developing macrophage-based therapies.

neuroscience↗

bioRxiv2026

Systematic identification of cell division proteins in haloarchaea and the discovery of a membrane anchor for the Z ring

Most archaea rely on the tubulin-like protein FtsZ for division. In the last decade, several novel cell division proteins have been discovered in the model haloarchaeon Haloferax volcanii which contains two FtsZ proteins, FtsZ1 and FtsZ2. However, the composition of this FtsZ-based archaeal divisome is largely uncharacterized. Here, using in vivo crosslinking coupled with immunoprecipitation and mass spectrometry (CLIP-MS), we identified ten proteins that localize to the division site, including wide-spread proteins predicted to be involved in DNA binding and energy metabolism. Deletion analysis indicate that most of these proteins have a modest or minor impact on cell division, but HVO_0399, renamed as Cell Division Protein C (CdpC), is important for cell division and functions as a membrane anchor for FtsZ1. CdpC consists of three domains, an N-terminal domain (NTD) containing an amphipathic helix critical for membrane binding, a long intrinsically disordered linker, and a C-terminal domain (CTD) that binds FtsZ1 with high affinity and likely promotes its polymerization. Notably, both the NTD and CTD of CdpC harbor a zinc finger that is indispensable for its function. Phylogenetic analysis indicates that CdpC sequences evolved fast across haloarchaea, but exhibited high conservation in domain structure and critical residues found in this study. Overall, our study expands the repertoire of candidate division proteins and establishes CdpC as a membrane anchor for FtsZ1 in haloarchaea. These findings pave the way for in-depth studies of arcaheal cell division and broaden the function of zinc finger proteins in archaea.

microbiology↗

bioRxiv2026

Ependymomas are cancers of the pre-neural crest/roof plate lineage

Distinct molecular variants of the brain cancer ependymoma are distributed along the rostral-caudal extent of the central nervous system (CNS). Historically proposed to arise from ventricular ependyma, recent studies have suggested conflicting cellular origins, including the neural radial glia and the roof plate lineages. Using single-cell transcriptomics, immunohistochemistry, and lineage tracing, we demonstrate that ependymomas across all CNS compartments transcriptionally mirror MSX1+ve pre-neural crest/roof plate (Pre-NC/RP) lineage derivatives. Ependymoma subgroups recapitulate the spatial and molecular diversity of regional Pre-NC/RP populations, while retaining conserved MSX1 expression. Expression of the oncogenic fusion ZFTA-RELA within the murine Pre-NC/RP lineage generated tumors that faithfully resembled human ependymoma. These findings identify a common embryonic cellular origin for ependymomas and reconcile previously conflicting models of tumorigenesis.

developmental biology↗

bioRxiv2026

Microbiota-derived TMAO links PM2.5 exposure to cognitive impairment: A gut liver brain axis

Although fine particulate matter (PM2.5) is linked to cognitive impairment, the mechanistic links between pulmonary exposure and neurodegeneration remain poorly understood. This study investigates the role of the gut-liver-brain axis in mediating PM2.5-induced neurotoxicity. We demonstrate that three weeks of intratracheal PM2.5 instillation in mice causes significant memory deficits, impaired hippocampal adult neurogenesis, and reduced synaptic plasticity in both sexes. Metagenomic analysis revealed that PM2.5 alters gut microbiota composition, specifically by upregulating pathways involved in trimethylamine (TMA) synthesis. This microbial shift led to a systemic increase in the metabolite trimethylamine N-oxide (TMAO) and its accumulation in the hippocampus. Mechanistic experiments revealed that TMAO drives neurotoxicity by activating the hippocampal PERK-mediated endoplasmic reticulum (ER) stress pathway. Critically, these deficits were reversed by suppressing hepatic TMAO production via either pharmacological inhibitors or genetic knockdown. Additionally, hippocampal-specific PERK silencing or dietary resveratrol intervention attenuated hippocampal ER stress, thereby protecting against PM2.5-induced cognitive and hippocampal impairments. These results identify the microbial metabolite TMAO as a key mediator of air pollution-related neurotoxicity and highlight the gut-liver-brain axis as a promising therapeutic strategy to counteract the neurological impacts of air pollution.

animal behavior and cognition↗

bioRxiv2026

An AI-agent-orchestrated grey-box Transformer framework for sparse pharmacokinetic curve reconstruction and pharmacometric model initialization

Clinical pharmacokinetic (PK) modelling is constrained by sparse sampling, limited general-isability of single-drug models, and labour-intensive workflows, making it difficult to infer complete drug exposure from limited concentration observations. We present the Pharmacokinetic Foundation Model (PKFM), a grey-box Transformer framework pre-trained across 32 drugs that reconstructs concentration-time profiles from sparse concentration observations, dosing events, molecular descriptors, and physiological covariates while preserving output interpretability. In representative oral PK curves, three sparse input points recovered the principal absorption-elimination trajectory, achieving coefficient of determination (R2) = 0.992 for Midazolam oral and R2 = 0.990 for Verapamil oral. Using reconstructed curves in NONMEM (nonlinear mixed-effects modelling) improved covariance stability and individual prediction accuracy. Contrastive-learning embeddings supported Top-10 physiologically based pharmacokinetic (PBPK) candidate retrieval, with 75.6% of observations within the 2-fold range. A pharmacometrics-informed AI Agent (PM Agent) outperformed general-purpose programming tools in stability and pairwise win rate on a standardised modelling benchmark, with each run requiring human pharmaco-metrician confirmation before downstream use. These results support cross-drug pre-trained PK models as an information-completion layer for sparse PK evidence and a structured scaffold for the modelling workflow; clinical or regulatory use requires prospective validation, broader external benchmarking, and independent expert assessment.

pharmacology and toxicology↗

bioRxiv2026

A weak-interaction model defines the cell-autonomous function of DNA methylation in gastrulation

DNA methylation is essential for mammalian embryonic development, yet its functional implications remain debated. Here, we introduce a quantitative model that reconciles conflicting perspectives on the roles of methylation in genome regulation and cell specification. Using single-cell multiome analysis in chimeric and whole embryo mutants during mouse gastrulation and organogenesis, we separate cell-autonomous from indirect effects of methylation and demethylation machinery. We show that while the transcriptional and chromatin programs defining basic lineages can be established independently of methylation, the fidelity of differentiated states is severely impaired in its absence. Specifically, we identify hundreds of genes and thousands of cis-regulatory elements (CREs) dependent on methylation for precise regulation. CRE accessibility alterations in methylation mutants are linked with CpG dinucleotide content, and correlate with multiple transcription factor binding motifs. Our data support a weak-interaction model in which DNA methylation moderates, but does not block or instruct the potency of trans-acting regulatory machineries genome-wide.

genomics↗

bioRxiv2026

Systematic detection of abnormal samples reveals widespread mislabeling in metagenomic studies

The human microbiome plays a critical role in health and disease, and its dynamic nature has made longitudinal sampling a key strategy for elucidating microbiome-disease relationships. Although the gut microbiome generally stabilizes over time, a subset of samples frequently shows marked deviations from an individuals baseline profile. We refer to these as abnormal samples. To analyze these abnormal samples, we developed a three-stage workflow to identify and classify these abnormal samples to figure out the underlying causes of these abnormal samples. Moreover, we systematically investigated abnormal samples across 16 publicly available metagenomic datasets, comprising a total of 5,171 metagenomes. Our analysis revealed that abnormal samples are often the result of mislabeling during sample collection, processing, or sequencing. Of which, fecal samples from family are more likely mislabeled. We found evidence of mislabeling in 75% of longitudinal datasets, involving up to dozens of samples per study, and in 25% of randomly selected cross-sectional datasets. Additional factors such as disease status (e.g., inflammatory bowel disease), sampling intervals, and sampling density may also contribute to sample abnormalities owing to true biological variations. These findings highlight that mislabeling is a common yet underrecognized issue in microbiome research. Our work underscores the importance of identifying and correcting abnormal samples to ensure data integrity in microbiome studies and provides a practical solution for quality control in large-scale metagenomic datasets.

microbiology↗

bioRxiv2026

A long-read human pangenome initiative for comprehensive interpretation of nuclear-embedded mitochondrial DNA

Nuclear-embedded mitochondrial DNA segments (NUMT) preserve a record of ongoing mitochondrial-to-nuclear DNA transfer during evolution, with important implications for disease mechanisms and genome organization. Here, we developed a Pangenome Graph-based NUMT Detection (PG-NUMT) approach to improve NUMT detection sensitivity by 2.52-fold compared to short-read based approaches and comprehensively resolved seven concatenated mega-NUMTs with a maximum length of 127.7 kbp. We generated a high-resolution human NUMT map comprising 774 fixed, 280 polymorphic, and 123 pericentromeric NUMTs, alongside 74 superpopulation-stratified NUMTs. Both fixed and polymorphic NUMTs are hypermethylated and enriched within segmental duplications, whereas fixed NUMTs preferentially localize to intergenic regions and avoid transposable elements. Interestingly, NUMTs derived from the 3-end of the mtDNA D-loop are less frequently fixed in human genomes and exhibit potential cis-regulatory activity in the nuclear genome. These observations suggest that selective pressures shape the genomic features of fixed NUMTs. Seven NUMTs associated with gene expression or alternative splicing were further identified, suggesting potential modulatory functions of common NUMTs. Using 20 complete non-human primate genomes, we identified variable NUMT insertion rates across primate lineages (1.5-18.7 insertions per million years), with particularly high rates in the Pan lineage. Notably, we uncovered two NUMT-derived variable number tandem repeats (VNTRs), establishing NUMTs as a novel source of VNTRs. In summary, the integrated analysis enhances our understanding of NUMT genomic architecture, population dynamics, and evolutionary implications in human and non-human primates, establishing NUMTs as dynamic genomic components of biomedical relevance.

genomics↗

bioRxiv2026

CodonRL: Multi-Objective Codon Sequence Optimization Using Demonstration-Guided Reinforcement Learning

Optimizing synonymous codon sequences to improve translation efficiency, RNA stability, and compositional properties is challenging because the search space grows exponentially with protein length and objectives interact through long range RNA structure. Dynamic programming-based methods can provide strong solutions for fixed objective combinations but are difficult to extend to additional constraints. Deep generative models require large-scale, high-quality mRNA sequence datasets for training, limiting applicability when such data are scarce. Reinforcement learning naturally handles sequential decision-making but faces challenges in codon optimization due to delayed rewards, large action spaces, and expensive structural evaluation. We present CodonRL, a reinforcement learning framework that learns a structural prior for mRNA design from efficient folding feedback and demonstration-guided replay, and then enables user-controlled multi-objective trade-offs during inference. CodonRL uses LinearFold for fast intermediate reward computation during training and ViennaRNA for final evaluation, warms up learning with expert sequences to accelerate convergence for global structure objectives, and introduces milestone-based intermediate rewards to address delayed feedback in long range optimization. On a benchmark of 55 human proteins, CodonRL outperforms GEMORNA, a state-of-the-art codon optimization method, across multiple metrics, achieving 9.5% higher codon adaptation index (CAI), 25.4 kcal/mol more favorable minimum free energy (MFE), and 3.4% lower uridine content on average, while improving codon stabilization coefficient (CSC) in over 90% of benchmark proteins under matched constraints. These gains translate into designs that are predicted to be more efficiently translated, more structurally stable, and less immunogenic, while supporting continuous objective reweighting at inference time.

bioinformatics↗

bioRxiv2026

Molecular basis of enfumafungin inhibition and resistance in Candida glabrata β-1,3-Glucan synthase FKS1

{beta}-1,3-Glucan synthase FKS1 is essential for fungal cell wall biogenesis and serves as a validated target of widely prescribed antifungal drugs. However, the molecular mechanism of pathogenic FKS1, its modes of inhibition, and the associated resistance mechanisms remain elusive, hindering the antifungal development. Here, we focus on Candida glabrata FKS1, a clinically relevant target frequently linked to antifungal resistance. We present cryo-EM structures of C. glabrata FKS1 in multiple detergent environments and in complex with the triterpenoid antifungal enfumafungin. Integrated functional studies revealed a key catalytic residue within the conserved ED motif and uncovered structural adaptations to different membrane-mimetic environments, rationalizing FKS1s sensitivity to its surrounding membrane context. Enfumafungin binds at the extracellular membrane leaflet of cgFKS1 transmembrane domain, engaging a convex surface separated from the cytosolic active site, consistent with its role as a non-competitive inhibitor. Further analyses identified key enfumafungin-binding residues from TM5-TM6 and elucidated the molecular basis of drug resistance in C. glabrata. Collectively, these findings establish a mechanistic framework for pathogenic FKS1 function and inhibition and provide a molecular basis for the rational design of next-generation antifungal therapeutics.

microbiology↗

bioRxiv2026

Borrelia Burgdorferi binds Serum Amyloid A and Modulates Subcutaneous Adipose Tissue Immune Signaling

Borrelia burgdorferi causes over 470,000 infections annually [1]. Following a tick bite, Borrelia spirochetes replicate in human skin and disseminate through tissues, including subcutaneous adipose tissue. Adipose tissue is a complex organ containing non-immune and immune cells that play a significant role in the immune response [2], yet little is known about adipose tissue signaling after Borrelia infection. We investigated the landscape of immune signaling within adipose tissue-resident cells during Borrelia infection in human tissue (skin and adipose tissue) ex vivo. Immune pathways overall were downregulated in adipose tissue during Borrelia infection. Despite this, adipose stem/progenitor cells exhibited increased pro-inflammatory and extracellular matrix (ECM)-related signaling (IL-6, MIF, collagen, laminin, fibronectin), positioning them as key hubs of intercellular communication during infection. Myeloid lineage cells showed cluster-specific upregulation of immune genes such as IL1B and TNF. Network analyses highlighted laminin, and strong outgoing MIF signals in all adipose-resident cell clusters. Among the identified genes modulated after Borrelia infection, we observed reduced expression levels of serum amyloid A in adipose tissue-resident cells, a marker typically elevated in Lyme disease patient serum during acute infection. Here, we show that human serum amyloid A1 and serum amyloid A2 directly bind B. burgdorferi spirochetes, using flow cytometry. Functionally, serum amyloid A2 reduces spirochete viability in vitro in a dose-dependent manner and enhances opsonization and phagocytosis by macrophages. Together, these findings show that adipose tissue is a site of active immune signaling in the setting of Borrelia infection and identifies SAA as a host defense factor against Borrelia.

immunology↗

bioRxiv2026

Global mRNA 3'UTR lengthening in small-cell neuroendocrine carcinoma

Small-cell neuroendocrine carcinoma (SCNC) is a rare but highly malignant tumor subtype that primarily arises in the lung, also rarely in other organs, and as a consequence of treatment induced lineage transdifferentiation of prostate adenocarcinomas. The molecular convergence of SCNC across diverse tissues enables its identification through conserved SCNC-specific molecular markers, facilitating tumor subtype classification. As a critical post-transcriptional regulatory mechanism, alternative polyadenylation (APA) modulates 3'UTR length and significantly impacts tumor progression. However, its role in SCNC remains largely unclear. Here, we report a global 3'UTR lengthening pattern driven by APA in SCNC. We identified a set of conserved 3'UTR lengthening events across SCNCs of different tissue origins, which are strongly associated with neural development and related signaling pathways. Furthermore, we developed a neural network-based prediction model to classify SCNC by leveraging these specific APA signatures. Our study provides new insights into the post-transcriptional landscape of SCNCs and highlights APA signatures as promising biomarkers for SCNC identification.

cancer biology↗

bioRxiv2026

RNA Structure Coordinates Translation Across the Meiotic Program

mRNA structure clearly modulates translation efficiency for individual transcripts, yet its role in coordinating translation across many transcripts remains poorly defined. Meiosis offers a compelling test case, involving large-scale gene expression changes coordinated across hundreds of mRNAs under conditions where transcription is constrained by chromosome condensation. We profiled mRNA structures across yeast meiosis, generating a high-resolution structurome encompassing ~70% of annotated mRNAs, including multi-time-point measurements for 2,084 transcripts. Transcripts upregulated during meiosis generally adopt flexible structures that enhance translation, whereas mRNAs with complex 5' UTR and coding-region structures show suppressed translation, indicating that RNA structure globally shapes meiotic translation. We further observed a high-low-high oscillation in cytoplasmic RNA helicase levels across meiosis. Together, these findings support a model in which RNA structure and helicase expression act jointly to program stage-specific translation of hundreds of mRNAs, with highly structured transcripts selectively translated at late stages. Consistent with this model, disrupting the structure of the abundant transcript CCW22, or altering Ded1p helicase levels, impaired meiotic progression in an RNA structure-dependent manner. Our study reveals that the concerted action of RNA structure and helicases coordinates cell-wide translation dynamics to meet stage-specific demands, especially critical when transcription is limited during meiotic divisions.

genomics↗

bioRxiv2026

Oxytocin Modulation of Spinal Circuits Drives Therapeutic Benefits of Massage

Across social species, social touch enhances well-being and reduces pain -- two seemingly distinct benefits that enhance survival. Yet where and how the nervous system integrates these functions, and whether a single mechanism could serve both, remains unknown. Here we show that massage triggers oxytocin release, which shapes both pain and touch reward at the earliest stage of central processing -- the spinal cord -- through a single, state-dependent circuit mechanism. We report that in humans, massage enhances well-being, effects that correlate with endogenous oxytocin release. In mice, gentle touch activates hypothalamic oxytocin neurons that project directly to the spinal dorsal horn. Genetic manipulation of spinal oxytocin circuits alters behavioral responses to both gentle touch and noxious stimuli. Spinal calcium imaging and slice electrophysiology reveal that oxytocin acts on both excitatory and inhibitory spinal neurons to sculpt the relative activity of spinal ascending systems that convey both social touch and pain to the brain. Extending these findings to humans, we show that oxytocin receptors are also expressed on spinal excitatory and inhibitory neurons, and that endogenous oxytocin during massage correlates with altered spinal touch processing. Thus, spinal oxytocin signaling provides an evolutionarily conserved mechanism for the therapeutic benefits of massage.

neuroscience↗

bioRxiv2026

RodZ acts through MreBCD to activate the elongasome in Escherichia coli

Most bacteria are surrounded by a peptidoglycan (PG) matrix that maintains cell shape and provides protection against turgor pressure. In many rod-shaped bacteria, synthesis of PG along the cell cylinder is organized by the elongasome, also called the Rod complex, which consists of six highly conserved proteins, including the actin-like MreB, the PG synthase complex RodA-PBP2 and three regulatory membrane proteins, MreC, MreD and RodZ. However, how these proteins interact with each other to form the elongasome and synthesize lateral PG remains elusive. In this study, by characterizing MreC mutations affecting elongasome activity in Escherichia coli, we provide evidence that MreC alternates between an active and an inactive state and these mutations affect MreC interaction with PBP2. We also find that RodZ interacts with MreC via a periplasmic region and disruption of this interaction compromises elongasome function. Additionally, we show that the cytoplasmic region of RodZ, which interacts with MreB, work synergistically with activating MreC mutations to promote rod shape formation. These results indicate that RodZ activates the elongasome by initiating two signaling cascades, one in the periplasm via MreCD, and another in the cytoplasm through MreB. This role of RodZ in regulating elongasome activity is analogous to that of FtsN in activating the divisome, suggesting that the mechanisms governing the activation of PG synthesis in cell elongation and division are similar in E. coli. ImportanceThe elongasome, or Rod complex, mediates lateral peptidoglycan synthesis during cell elongation in many rod-shaped bacteria. It consists of the cytoskeletal protein MreB, the peptidoglycan synthase RodA-PBP2, and three regulatory proteins MreCD and RodZ. Although it has been extensively studied, how its activity is controlled remains incompletely understood. Here, we reveal the roles of MreCD and RodZ in regulating elongasome activity in Escherichia coli. Our results indicate that RodZ triggers lateral peptidoglycan synthesis by interacting with MreB in the cytoplasm and MreCD in the periplasm. These interactions likely switch MreBCD to the active state such that they can stimulate the activity of the RodA-PBP2 complex. This regulatory mechanism resembles the activation mechanism of the divisome that mediates septal peptidoglycan synthesis.

microbiology↗

bioRxiv2025

P2RY2 is a purinergic immune checkpoint linking extracellular ATP to immune evasion and adaptive resistance to immunotherapy

Extracellular ATP (eATP) accumulates substantially in the tumor microenvironment (TME) and rises further during immunotherapy. While canonically an immune-activating "danger" signal, eATP also promotes immunosuppression in tumors, thus far largely attributed to its metabolite, adenosine. Here, we identify direct eATP signaling through P2RY2 as a dominant, adenosine-independent mechanism of immune resistance. Specifically, eATP-P2RY2 signaling serves as the primary upstream driver of COX-1/2 upregulation and consequent accumulation of immunosuppressive PGE2 in the TME, uncovering the long-sought TME-specific trigger of pathological COX-PGE2 hyperactivation in solid tumors. Genetic deletion or pharmacologic inhibition of P2RY2 eliminates both baseline and therapy-induced intratumoral PGE2, restores antitumor T cell responses, and reverses resistance to CAR-T, TCR-T, checkpoint blockade, and TIL therapies. Given that persistently elevated eATP is a hallmark of solid tumors, our work reveals a fundamental mechanism by which tumors hijack innate "danger" signaling to establish immune suppression and develop adaptive resistance to immunotherapy. These findings establish P2RY2 as a purinergic immune checkpoint with translational potential for combinatorial cancer immunotherapies. HIGHLIGHTSO_LIeATP in the TME drives baseline and therapy-induced immune evasion via P2RY2 C_LIO_LIeATP-P2RY2 signaling is the primary cause of PGE2 accumulation in the TME C_LIO_LIP2RY2 propels a feedback loop that enforces adaptive resistance to immunotherapies C_LIO_LIP2RY2 blockade reprograms the TME and restores immunotherapy responsiveness C_LI

immunology↗

bioRxiv2025

Evolutionary innovation within conserved gene regulatory networks underlying biomineralized skeletons in Bilateria

Biomineralized skeletons have evolved convergently across animals, yet exhibit remarkable diversity in structure and development. However, the evolutionary origins of gene regulatory networks underlying biomineralized skeletons remains elusive. Here, we report comprehensive developmental profiling of transcriptomic and chromatin dynamics in a bivalve mollusc, Crassostrea nippona. We provide evidence for a biphasic regulatory program orchestrating larval and adult shell formation, involving the coordinated activity of ancient transcription factors and dynamic chromatin remodeling. Comparative analyses suggest a conserved developmental toolkit was co-opted for larval exoskeleton formation in the common lophotrochozoan ancestor. In contrast, limited regulatory conservation was observed between lophotrochozoans and echinoderms with regards to the formation of biomineralized skeletons, despite both relying on a heterochronic activation of ancestral regulators. Together, our findings support a hierarchical model where dynamic chromatin decouples rapidly evolving effectors from deeply conserved regulators, allowing modular innovations within conserved gene regulatory networks. This study highlights how epigenetic dynamics bridge evolutionary conservation and novelty, offering a framework for understanding the independent evolution of biomineralization across Bilateria through combinatorial regulatory evolution.

evolutionary biology↗