Biology subjects

Liu, S.

Publications and source records attributed to Liu, S..

At least 181 records · Page 10Linked to original sources

bioRxiv2025

Information Network Flux interprets the importance of regulatory circuits in the protein and gene spaces

Cells receive information from environment, make decisions, and execute biological functions. This sensor-actuator controlling mode is common, and has benefits for system stability especially when there exist feedback-loops for action correction. In cells, transcription factors (TFs) play the role of "brain" since they control the expression levels of about 20,000 protein-coding genes. However, the detailed topological structure of the regulating relationships among them is far from clear. Circuitlike pathways, which usually contain protein-protein interactions (PPIs), TF-gene regulations and gene expressions, are pervasive in cells, however, their importance is yet to be systematically elucidated by unifying information at both protein and gene levels. Here we developed Information Network Flux (INF), an algorithm that could simulate information transmit on multi-layer networks, and integrate protein and gene level information for cell type specific pathways that respond to given perturbation. We used topological analysis to identify regulatory circuits and found that the TFs participating in these circuits, especially those acted as "sensor" or "actuator" are highly relevant to cellular response to external signals. At the gene level, circuit genes serve as efficient cell type indicators, which showed better performance than commonly used gene expression levels. At the circuit level, shorter circuits enrich housekeeping functions and longer circuits exhibit increased tissue-specificity. These findings suggest that these circuit pathways may contribute to the identity and stability of cell states.

biophysics↗

bioRxiv2025

A foundation model for microbial growth dynamics

Microbial growth dynamics contain rich information about microbial populations, which support applications from antibiotic testing to microbiome engineering. However, the high dimensionality of growth data and the scarcity of large, task-specific datasets have limited generalizable modeling analysis across systems. Here, we develop a foundation model for microbial growth dynamics. It is a large-scale, self-supervised representation model trained on [~]370,000 experimental and simulated growth curves spanning diverse microbial species, environmental conditions, and community contexts. The model learns lower-dimensional latent embeddings that capture essential dynamical features of raw growth data and enable accurate reconstruction of these data. The concise representations enhance predictive performance in diverse downstream applications. Using these embedding, we achieve few-shot learning for antibiotic classification and concentration prediction, accurate forecasting of simulated and experimental communities, and inference of total abundance from relative-abundance data. By extracting transferable representations from heterogeneous datasets, our model provides a general framework for analyzing and predicting microbial community dynamics from limited measurements.

synthetic biology↗

bioRxiv2025

Gut metabolism links precision nutrition, exercise, and healthspan in Drosophila melanogaster

Longevity-promoting interventions commonly entail functional trade-offs, raising the unresolved question of whether lifespan extension necessarily compromises physiological performance. Here, utilizing a chemically defined diet (CDD) in Drosophila melanogaster, we systematically evaluated a multimodal intervention combining methionine restriction (MR), taurine supplementation (Tau), and moderate exercise. This combinatorial approach synergistically extended lifespan, preserved reproductive capacity, and improved locomotor function. Integrative targeted metabolomics and stable isotope tracing revealed increased mitochondrial tricarboxylic acid (TCA) cycle flux and enhanced redox homeostasis in the gut as central metabolic features. Notably, Lactobacillus plantarum was identified as a key microbial mediator responsive to dietary and behavioral stimuli, potentially coordinating host energy metabolism and maintaining physiological integrity. Together, these observations outline a "nutrition-behavior-microbiota" framework that uncouples the traditional trade-off between lifespan and functional health, offering new perspectives for promoting healthy aging.

physiology↗

bioRxiv2025

Phylogenomics of Rhinogobius gobies reveal northern-southern divergence and trait evolution in China

The goby genus Rhinogobius represents a prominent radiation of freshwater fishes in East Asia, with China serving as its evolutionary hotspot harboring over 39 endemic species. Despite their ecological dominance in riverine ecosystems, phylogenetic relationships among mainland Chinese Rhinogobius species remain poorly resolved due to limited sampling and morphological conservatism, particularly in meristic traits traditionally used for classification. To address these knowledge gaps, we employed a nuclear gene enrichment approach to target 4,434 single-copy nuclear loci across 26 species spanning all major river basins in eastern and southern China. Our phylogenomic analyses based on sequence of 2,137 cleaned-up loci (477,822 bp) recovered a robust topology dividing Rhinogobius into three primary lineages: an early-diverging of R. similis sister to two major clades, the northern vs. southern clades exhibiting strong geographic structuring. Fossil-calibrated molecular dating placed the crown age of Rhinogobius at [~]9.33 million years ago (Ma) during the late Miocene, with a pivotal north-south divergence at [~]5.96 Ma coinciding with major paleo-drainage reorganizations in East Asia. Ancestral state reconstruction revealed that the southern clade uniquely evolved unstable first dorsal fin formula and near-universal presence of predorsal scales. By integrating the most extensive geographic and taxonomic sampling to date with phylogenomic data, this study establishes: (1) a revised evolutionary timeline for Rhinogobius diversification linked to East Asian geomorphological history, (2) diagnostic morphological characters for major clades, and (3) a robust phylogenetic framework to guide future taxonomic revisions and conservation efforts for this ecologically significant group. HighlightsO_LIPhylogenomics resolves north/south split in Chinese Rhinogobius gobies. C_LIO_LIFossil-dated 9.3 MYA origin of Rhinogobius and rapid divergence at 5.96 MYA. C_LIO_LISouthern lineages exhibit more predorsal scales and unstable first dorsal fin formula. C_LIO_LIFirst genomic backbone for species revision and adaptation study. C_LI

zoology↗

bioRxiv2025

Interpretable Distillation Reveals that Deep-learning-based Splicing Models Suffer from Pervasive Confounders and Blind Spots

Despite their growing popularity, genomic deep-learning-based models function largely as black boxes, raising concerns about their trustworthiness. Here we develop a framework to explain model prediction logic using interpretable distillation. Applying our framework, we find that RNA splicing prediction models suffer from pervasive confounders and blind spots, leading to poor performance on non-reference sequences. Our findings illuminate fundamental limitations of training models on genomic sequences and suggest ways to overcome them.

genomics↗

bioRxiv2025

Molecular interplay between the DNA damage checkpoint kinase Mec1-Ddc2 and its activator Dpb11 on gapped DNA

The eukaryotic DNA damage and replication stress checkpoint is initiated by activation of the apical kinase complex ATR-ATRIP on RPA-coated ssDNA. In Saccharomyces cerevisiae, the Mec1-Ddc2 (hATR-ATRIP) activator and checkpoint mediator Dpb11 (hTopBP1) is recruited to the 9-1-1 checkpoint clamp (another Mec1-Ddc2 activator) at 5 ss-dsDNA junctions. It remains unclear how Mec1-Ddc2 encounters its activators on damaged DNA due to their differential DNA binding preferences. Using real-time single-molecule imaging, we show that Dpb11 binds to ssDNA directly and localizes to ss-dsDNA junctions in an RPA-dependent manner. Furthermore, Dpb11 recruits Mec1-Ddc2 to ss-dsDNA junctions. Single-molecule force spectroscopy was used to demonstrate that Dpb11 forms bridges on ssDNA, both alone and in the presence of RPA, reducing the end-to-end distance of gapped DNA. These data support a model in which Dpb11 facilitates Mec1-Ddc2 colocalization with its activators directly by recruiting Mec1-Ddc2 to gap junctions and indirectly by decreasing the effective gap length.

biophysics↗

bioRxiv2025

Fungal Mst3-proteins are involved in fungal innate immunity needed for the recognition of bacteria surrounding the hyphae, as well as for plant pathogenicity

Fungal innate immunity overall responses are as fast as and resemble mammalian innate immunity. It does not, however, employ membrane toll-like receptors (TLRs), but should employ endocytosis of non-fungal molecular patterns recognized by nuclear-localizing receptors (NLR). Downstream, both types of receptors in mammals are Mammalian Ste20 kinases (MSTs). We identified single MST3 orthologs in the plant pathogens Fusarium graminearum (FgMST3) and Magnaporthe oryzae (MoMST3). We knocked out both genes and investigated mutants using a standard panel of tests for growth, development, and pathogenicity for each respective fungus. Both{Delta} FgMST3 and{Delta} MoMST3 strains showed reduced pathogenicity. The deletions negatively affected conidia production and conidia germination but had little effect on growth rate. However, the two mutants reacted differently to some stress treatments, such as Zn2+ and gentamicin. Furthermore, we constructed an innate immunity reporter system for F. graminearum based on a gene with fast responses to bacterial MAMPs to detect less than 4-hour responses to non-self-molecular patterns (NSMP) like bacterial outer membrane vesicles (OMVs) as well as trace levels of sucrose, as a plant NSMP. The reporter gene responses to OMVs of MST3 mutant strains were severely reduced. Our results suggest that both MoMst3 and FgMst3 are involved in fungal innate immunity downstream of unknown NLR proteins, motivating studies to try to identify genes for the NLR-like receptors. Finding such and investigating how they work and vary between fungal species and strains should be useful for understanding fungal biotic interactions with viruses, bacteria, plants, and animals.

immunology↗

bioRxiv2025

NanoporeDB: A Structural Resource Of Multimeric Protein Nanopores In Single-Molecule Sensing

Protein nanopores are essential molecular gateways in biology and have inspired transformative technologies in biosensing and single-molecule sequencing. While this technology has transformed genomics and biosensing, the discovery of novel nanopore scaffolds remains limited due to the scarcity of experimentally resolved pore structures. Here, we present NanoporeDB, an open-access structural resource comprising over 6,600 high-confidence multimeric models across four representative pore types. Candidate proteins were systematically mined from large-scale datasets, including the AlphaFold Protein Structure Database, UniRef90, and MGnify90, and assembled using AlphaFold-Multimer and AlphaFold3. We performed membrane embedding, pore axis annotation, and constriction profiling to enable functional interpretation. NanoporeDB features an interactive web interface with 3D visualization and quantitative metrics such as insertion depth, tilt angle, and pore geometry. This resource provides a structural foundation for advancing nanopore-based molecular sensing, precision diagnostics, and synthetic biology.

bioinformatics↗

bioRxiv2025

TissueNarrator: Generative Modeling of Spatial Transcriptomics with Large Language Models

The intricate spatial organization and molecular communication among cells are fundamental to multicellular systems. Spatial transcriptomics (ST) enables gene expression profiling while preserving spatial context, providing rich data for studying cellular interactions and tissue dynamics. However, most existing computational approaches focus on embedding-based tasks and provide limited generative capacity for simulating cell behavior in situ. Moreover, accurately interpreting spatial interactions requires extensive biological knowledge, which current models do not incorporate. Here, we introduce TO_SCPLOWISSUEC_SCPLOWNO_SCPLOWARRATORC_SCPLOW, a framework that reformulates spatial omics analysis as a language modeling problem. By representing tissue sections as spatial sentences - rank-based gene lists augmented with spatial coordinates and metadata - TO_SCPLOWISSUEC_SCPLOWNO_SCPLOWARRATORC_SCPLOW leverages pretrained large language models (LLMs) to learn spatially conditioned gene expression patterns. The model generates realistic, context-aware cellular profiles, predicts intercellular interactions, and performs in silico perturbation analyses. Across multiple ST technologies (MERFISH, Perturb-FISH, and CosMx SMI), TO_SCPLOWISSUEC_SCPLOWNO_SCPLOWARRATORC_SCPLOW achieves superior quantitative performance and recovers biologically meaning-ful ligand-receptor and signaling pathways. Furthermore, a conversational inference mode enables natural-language querying of tissue organization. By integrating pretrained biological knowledge with spatial context, TO_SCPLOWISSUEC_SCPLOWNO_SCPLOWARRATORC_SCPLOW establishes a new, scalable generative paradigm for modeling, simulating, and reasoning about tissue systems.

bioinformatics↗

bioRxiv2025

Sustained Humoral Activation through self-amplifying mRNA Vaccination Enhances Longitudinal Antibody Function in a Phase III Trial

Sustained and functional antibody responses to respiratory pathogens through vaccination is critical for global public health. The development and deployment of mRNA vaccines during the coronavirus disease 2019 (COVID19) pandemic was a landmark achievement in modern medicine and ushered in a new age of vaccine innovation. The mRNA-based vaccines elicited strong antibody responses, both neutralizing and extra-neutralizing, against the viral Spike protein. The antibody levels waned with time since vaccination, and that coupled with the antigenic drift of the virus prompted updates to the mRNA vaccine composition and evaluation of other mRNA modalities. Self-amplifying mRNA (sa-mRNA) vaccines such as ARCT-154 can prolong antigen production and durability of humoral immune response post-immunization, and can thus be administered at a lower dose. How this translates into the overall humoral architecture compared to that shaped by conventional mRNA vaccinations, however, is unclear. Here, we analyze serum-based antibody responses from a recent Phase III trial comparing humoral responses elicited by ARCT-154 and mRNA BNT162B2. All participants had received three doses of mRNA COVID-19 vaccines and were randomized to receive a booster dose of ARCT-154 or BNT162B2. Primary outcomes were to quantify waning responses against ancestral/wild type SARS-CoV-2 Spike (WT Spike) and a panel of antigenically SARS-CoV-2 variants Spikes. Through a systems serology approach, we identified that the sa-mRNA vaccine ARCT-154 elicited a unique antibody response compared to BNT162B2 defined by a sustained, activating profile to the vaccine-encoded Spike protein and a broad spectrum of drifted Spikes. Notably, potently activating Fc{gamma}RIIIA-binding antibodies showed a sustained stimulation in the ARCT-154-treatment arm, and this translated to an enhanced natural killer (NK) cell activation. The NK-activation through ARCT-154 was present for both target WT Spike and the antigenically drifted BA.5 Spike, which was the predominant form of SARS-CoV-2 during the observation period. Our results support a model whereby prolonged antigen expression and presentation moves immune profiles towards activating phenotypes with broad antigenic coverage.

immunology↗

bioRxiv2025

Dissociation between physical reasoning and tool use in individuals with left hemisphere brain damage

Many everyday tasks, from chopping vegetables to catching a ball, require understanding both how objects respond to physical forces and how to use them effectively. These capacities, tool use and physical reasoning, are often assumed to rely on shared cognitive and neural mechanisms. At some level, this correspondence is expected: using an object typically requires understanding its physical properties. However, both capacities are complex and multicomponential, so the relationship between them may vary across levels of representation and task demands. Here, we asked whether third-person physical reasoning about object dynamics can dissociate from tool use (i.e., performing a tools typical action, such as using a hammer to drive a nail) in individuals with left-hemisphere stroke. We tested 11 patients, five of whom showed impairments in tool use. Physical reasoning was assessed using a novel collection of tasks probing judgments about mass, velocity, and timing across static and dynamic scenes. Tool use was evaluated using a classic gesture-to-sight task, a pantomime-based measure in which participants are shown pictures of familiar tools and asked to demonstrate how each would be used. We identified an individual-level dissociation: patient I.A.* showed impairment in gesturing the use of objects despite preserved physical reasoning, often outperforming neurotypical controls. This pattern was complemented by patient N.P., who showed the reverse profile, with intact tool use gestures but difficulties in some physical reasoning tasks. These findings suggest that the ability to reason about the physical world and tool use can dissociate behaviorally and be independently disrupted by brain damage. This challenges the view that physical reasoning and tool use draw on the same underlying cognitive and neural mechanisms and suggests that at least some of their components are distinct.

neuroscience↗

bioRxiv2025

Cadmium toxicity to the human gut microbiome varies depending on composition

Cadmium (Cd(II), hereafter Cd) is a toxic heavy metal with detrimental impacts on the gut microbiota. We investigated the effects of acute Cd exposure on human fecal microbiotas using 24-hour in vitro cultures from 21 healthy adult donors. Regression analysis of butyrate production in the absence (-Cd) versus presence (+Cd) of Cd identified three categories of microbial responses: sensitive, intermediate, and resilient. Under Cd stress, sensitive microbiomes exhibited significant decreases in butyrate coupled with elevated acetate and lactate production, while resilient microbiomes did not show significant changes in butyrate and exhibited attenuated increases in lactate compared with sensitive microbiomes. Several genera were significantly different between sensitive and resilient communities after exposure to Cd, but the most striking was Anaerostipes. Network analysis revealed a significantly greater disruption of microbial interactions in sensitive communities compared with resilient. In resilient communities, butyrate production was primarily associated with Faecalibacterium in the absence of Cd and Anaerostipes in the presence of Cd. Furthermore, supplementation of sensitive microbiota with Anaerostipes species restored butyrate production in the presence of Cd. These findings highlight distinct gut microbial responses to acute Cd exposure and provide a foundation to investigate microbiota features underlying Cd sensitivity or resilience. IMPORTANCECadmium is a widespread environmental contaminant that reaches the human intestine, where it can disrupt the gut microbial community and negatively impact digestive and systemic health. However, this study demonstrates that human gut microbiomes vary in their responses to cadmium exposure: sensitive communities exhibit losses of beneficial organisms, particularly butyrate-producing taxa that contribute to intestinal integrity and metabolic balance, whereas resilient communities retain microorganisms with this key functional capacity. Anaerostipes appeared to be involved, at least in part, with Cd resilience. This work advances our understanding of how gut microbial functions may mitigate the adverse effects of cadmium exposure by identifying the compositional features that distinguish sensitive from resilient microbiomes. These findings highlight the importance of elucidating microbiome-mediated mechanisms that help sustain host health and lay the groundwork for deeper mechanistic studies aimed at mitigating cadmium toxicity.

microbiology↗

bioRxiv2025

Gut metabolite IPA alleviates white matter injury post-ICH by enhancing myelin debris phagocytosis via Stap1 inhibition

BackgroundIntracerebral hemorrhage (ICH) causes neurological dysfunction and white matter injury (WMI) characterized by myelin loss, axonal injury and myelin debris accumulation. Microglia-mediated debris clearance is critical for WMI repair. The microbiota-gut-brain axis plays an essential role in the central nervous system diseases, one of the ways in which gut microbiota affects brain is via producing metabolites. Indole-3-proprionic acid (IPA), a tryptophan-derived metabolite that mainly produced by Clostridium sporogenes, exhibits anti-inflammatory and neuroprotective properties. However, its effect on ICH remains unclear. This study aims to investigate the IPA level after ICH and the therapeutic effects of IPA on neurological deficits and WMI, as well as the potential mechanisms underlying IPA-mediated neuroprotection. MethodsAn ICH model was established using C57BL/6 mice, which then received intragastric IPA (20 mg/kg/day). Fecal abundance of IPA-related genes and IPA levels in feces and plasma were measured by qPCR and UPLC-MS/MS. Behavioral tests, qPCR, and immunofluorescence staining were used to assess neurological function, myelin integrity, and axonal injury. In vitro, BV2 microglia, with or without Stap1 knockdown, were co-cultured with myelin debris to assess IPAs effects on phagocytosis. Additionally, targeted plasma IPA profiling was performed in 30 ICH patients and 30 matched healthy controls. ResultsThe relative abundance of IPA production-related genes and IPA levels in feces and plasma were significantly decreased after ICH and remained low into the chronical phase. After IPA administration, the neurological deficits, myelin loss and axonal injury of mice with ICH were significantly improved. In vitro, IPA increased BV2 microglia myelin debris phagocytosis by inhibiting Stap1 expression. IPA levels were significantly reduced in ICH patients, consistent with ICH mouse model. ConclusionsOur findings demonstrated that the gut microbiota-derived metabolite IPA could facilitate neurological deficits recovery and alleviate WMI, which could be a promising therapeutic strategy to improve ICH prognosis.

neuroscience↗

bioRxiv2025

A disease mutation reveals a non-swapped pore and rewired gating cycle of the mechanically activated PIEZO1 channel

PIEZO1 is a versatile mechanotranduction channel critical for diverse cellular and physiological processes1-3. Gain-of-function mutations, such as R2482H, which slow channel inactivation, are linked to dehydrated hereditary stomatocytosis (DHS) and iron overload4-7. PIEZO1 forms a gigantic bowl-shaped trimer comprising flattenable peripheral blades and a central pore module with a switchable cap8-12. However, the structural dynamics governing its gating process and pathogenesis remain incompletely understood. Here, we determined distinct conformational states of the DHS-associated PIEZO1-R2482H mutant in both detergents and proteoliposomes. Remarkably, the R2482H mutation in the central pore induces a conformational change that propagates [~]25 nm to the most distal blade, enabling resolution of all 38 transmembrane helices. PIEZO1-R2482H is inherently flatter than PIEZO1, lowering the gating force and energy barrier and thus explaining its heightened mechanosensitivity. While the flattened PIEZO1 has its cap in the up-state adopting an inactivated state, the flattened PIEZO1-R2482H has its cap in the down-state and twisted by [~]75{degrees}, drastically switching the pore module from a swapped to a non-swapped organization. Such twisted conformation of PIEZO1-R2482H might hinder the transition to the inactivated state, instead proceeding through a deactivation-transition state characterized by curved blades and a partially twisted cap, thereby accounting for the slowed inactivation and deactivation kinetics. Taken together, we reveal the full transmembrane topology and a non-swapped pore conformation induced by R2482H, and propose the structural dynamic model that delineates the distinct gating cycle of activation, inactivation, and deactivation for PIEZO1 and PIEZO1-R2482H.

biophysics↗

bioRxiv2025

Multimodal AI Decodes Extreme Environment Functional Dark Matter Beyond Homology

Functional annotation of proteins from extreme environments represents a major bottleneck for bioresource discovery, as a vast reservoir of functional dark matter defies existing homology-based methods. We demonstrate that environmental pressures impart conserved physicochemical energy signatures that co-determine protein function with sequence and structure. Here we developed ACCESS, a multimodal graph neural network employing hierarchical contrastive learning with a tailored label-sample co-embedding to fuse energy, sequence, and structural information and overcome homology scarcity. ACCESS surpasses state-of-the-art methods including BLASTp and CLEAN in annotating low-identity enzymes. Applied to extreme environmental metagenomics, we constructed a function map of extremophile enzymes to expand the biocatalyst library, pinpointed functionally critical residues to guide rational design, and enabled large-scale, function-based macro-evolutionary analyses. This paradigm transcends the limitations of homology, illuminating protein dark matter and accelerating the exploration of the biospheres functional diversity for applications in biotechnology and therapeutic development.

bioinformatics↗

bioRxiv2025

Modular mRNA lipid nanoparticle platform rescues diverse genetic male infertility

Genetic male infertility arises from diverse mutations that disrupt spermatogenesis, yet therapeutic strategies capable of restoring germ-cell function remain limited. Messenger RNA (mRNA) therapeutics offer a programmable and non-integrating approach, but their functional compatibility in structured and immune-privileged tissues remains unclear. Here we identify a temporal mismatch between mRNA delivery kinetics and the progression of spermatogenesis as a key barrier to functional rescue. We show that delivery efficiency does not predict outcome: formulations with high peak expression but short duration fail to support spermiogenesis, whereas sustained expression aligned with developmental timing enables recovery. Guided by this principle, we establish a testis-targeted mRNA delivery platform that restores spermatogenesis across multiple genetic models and generates functional sperm capable of supporting embryo development and multigenerational inheritance. The approach is modular, extends to human seminiferous tubules ex vivo, and is supported by comprehensive safety analyses. Together, these findings define a design constraint for RNA therapeutics in structured tissues: functional rescue requires temporal alignment between delivery kinetics and the therapeutic window defined by tissue-specific biological programs.

bioengineering↗

bioRxiv2025

Antibody feedback establishes an affinity brake in the germinal center

Recent advances in mRNA vaccine technology have opened the door to novel types of antigen display, but little is yet known as to how B cells recognize and respond to these formats. By delivering an mRNA-LNP encoded membrane-bound immunogen displaying three conserved HIV-1 Envelope (Env) epitopes to knock-in mouse models with B cell receptors (BCRs) of defined affinities, we investigated how epitope-specific competition shapes germinal center (GC) responses. Co-activation of B cells targeting different epitopes did not alter GC kinetics observed in individual activations, but a striking inverse correlation was observed between BCR affinity and GC residence time in either scenario: high-affinity B cells exhibited shorter persistence in GCs, while those with lower affinity to the antigen were maintained. Furthermore, B cells were able to engage in GC reactions at equivalent rates in the presence or absence of clonal lineages binding the same epitope with similar affinities, while higher-affinity clones suppressed lower-affinity counterparts targeting the same epitope. Spatial transcriptomics revealed plasma-like cells within and adjacent to the GC which, together with the detection of early IgG in draining lymph nodes, suggests that local antibody production from these cells may contribute to feedback-driven kinetics. These findings indicate that a self-modulated local antibody feedback loop may act as a "brake" on epitope-specific recognition--dampening further affinity enhancement for high-affinity B cells and facilitating epitope spreading by redirecting the response toward alternative epitopes.

immunology↗

bioRxiv2025

N6-methyladenosine regulation of mRNA translation is essential for early human erythropoiesis

N6-methyladenosine (m6A) is an abundant modification of mRNA with important regulatory roles in normal and malignant hematopoiesis. We previously reported that in human erythroid leukemia (HEL) cells, m6A mRNA marking selectively regulates translation of essential erythropoiesis genes required for in vitro differentiation and human erythroid colony formation. Here, we further investigated the timing and nature of requirement for m6A-methyltransferase (MTase) activity during human erythropoiesis, using a standardized in vitro erythroid differentiation assay for hHSPCs. We identified two critical m6A regulated developmental windows in BFU-E and during the transition from CFU-E to proerythroblasts. These windows of m6A-MTase requirement coincide with rising global m6A levels, which peak in proerythroblasts. After proerythroblast formation, however, m6A -MTase activity is dispensable for differentiation, proliferation, and survival. In BFU-E, m6A-MTase promotes proliferation but is dispensable for differentiation, while, in CFU-E, both m6A -MTase and the YTHDF family of m6A readers are essential for differentiation to proerythroblasts. Mechanistically, in CFU-E, m6A MTase activity enhances translation of ribosomal and oxidative phosphorylation (OXPHOS) genes, thereby elevating global protein synthesis rates and enabling efficient erythroblast formation. We propose that this form of translational regulation by m6A emerged as an evolutionary adaptation to meet the high translational demands of human erythropoiesis.

developmental biology↗