Biology subjects

Pan, F.

Publications and source records attributed to Pan, F..

12 recordsLinked to original sources

bioRxiv2026

Riboflavin Deficiency-Induced Angular Stomatitis and Conjunctivitis in Associated with Gut Microbiota Dysbiosis

PurposeRiboflavin deficiency causes ariboflavinosis, and the purpose of this study is to investigate the potential biological factors underlying its occurrence. MethodsMale F344 rats were randomly assigned to R6 and R0 groups. The healthy controls (R6 group) was fed a riboflavin-normal diet (6 mg/kg riboflavin), while the R0 group was fed a riboflavin-deficient diet (0 mg/kg riboflavin) for 16 weeks. Following this period, the R0 group was switched to a riboflavin replenishment diet (6 mg/kg riboflavin) for an additional 4 weeks (weeks 17-20). The bacterial communities were analyzed using high-throughput 16S rRNA gene sequencing. ResultsRiboflavin deficiency induces ariboflavinosis in rats (66.7%), characterized by angular stomatitis and conjunctivitis. With riboflavin replenishment, ariboflavinosis was completely resolved. Riboflavin deficiency altered the gut microbiota co-occurrence network and composition. The potential mechanism is predicted to involve an increase in glycan biosynthesis and metabolism within the gut microbiota, particularly in lipopolysaccharide biosynthesis pathways. ConclusionsThese findings suggest that riboflavin deficiency induces ariboflavinosis by altering the gut microbiota, providing new insights into the mechanisms of riboflavin deficiency and its association with chronic inflammation. Clinical perspectivesRiboflavin is an essential nutrient that must be obtained through the diet, as the human body cannot synthesize it. Since riboflavin is not stored in the body, regular dietary intake is necessary. Riboflavin deficiency is prevalent worldwide and can lead to ariboflavinosis. However, the specific mechanisms underlying riboflavin deficiency-induced ariboflavinosis remain unclear. This study demonstrates that riboflavin deficiency causes ariboflavinosis (characterized by angular stomatitis and conjunctivitis) in rats and reveals a close relationship between riboflavin deficiency and alterations in gut microbiota composition. Gut dysbiosis appears to be a hallmark of ariboflavinosis and may serve as a promising diagnostic biomarker and therapeutic target for the condition.

microbiology↗

bioRxiv2026

Endothelin 2-Mediated Neuroinflammation Drives Hypertension-Associated Retinal Dysfunction

Systemic hypertension is a significant risk factor for glaucoma, a leading cause of irreversible blindness, yet the mechanistic basis connecting the two conditions remains elusive. Here, we demonstrate that hypertensive rats develop accelerated retinal degeneration characterized by retinal thinning, functional impairment, reduced perfusion, and progressive neuroinflammation. Critically, these changes occur despite persistently lower intraocular pressure (IOP), indicating an IOP-independent mechanism of pathology. Transcriptomic and biochemical analyses identified a robust activation of the retinal endothelin system, with endothelin-2 (EDN2) being the earliest and most significantly upregulated component. AAV-delivered retinal Edn2 knockdown completely restored retinal function, ameliorated neuroinflammation, and enhanced retinal blood flow in hypertensive rats. Pharmacological blockade of endothelin receptor type A (EDNRA), but not type B, replicated these protective effects without altering IOP. Our findings establish EDN2 as a central, IOP-independent driver of hypertension-associated retinal neurodegeneration and identify the endothelin-EDNRA axis as a potential therapeutic target to prevent vision loss in hypertensive patients.

neuroscience↗

bioRxiv2025

Bioinformatics classification of the MgtE Mg2+ channel and de novo protein design for the stabilization of its novel subclass

MgtE channels play crucial roles in Mg{superscript 2} homeostasis and are implicated in bacterial survival under antibiotic exposure. Previous structural and biophysical studies have predominantly focused on the Thermus thermophilus MgtE, leaving the structural and mechanistic diversity of MgtE family proteins largely unexplored. In this study, using a genome mining approach, we identified diverse MgtE homologs, including a novel subclass termed the "mini-N type," which lacks the canonical cytoplasmic N and CBS domains but possesses a unique small N-like domain. Despite extensive expression screening, mini-N type homologs could not be stably purified. To address this issue, we designed a series of de novo proteins and determined their crystal structures. A selected de novo protein was fused to a mini-N type MgtE, enabling successful purification and preliminary cryo-EM imaging. Our findings demonstrate that de novo designed protein fusions can serve as powerful tools for stabilizing and purifying otherwise unstable membrane proteins, opening new avenues for the structural and functional studies of otherwise inaccessible membrane proteins.

biophysics↗

bioRxiv2025

Lethal streptococcal infection is associated with widened lymph node bottlenecks

Streptococcus suis serotype 2 (SS2) infection causes streptococcal toxic shock-like syndrome (STSLS), which can be lethal in both pigs and humans. In mouse models, relatively subtle differences in the inoculum size result in drastically different infection outcomes, from clearance to acute death. Here, we leveraged barcoded SS2 in a murine model to define the bacterial population dynamics associated with death. Barcode sequencing analysis demonstrated that a 10-fold increase in inoculum size led to a 100-1,000-fold increase in the number of bacteria that initiated infection in certain lymph nodes, ultimately driving lethality. This lethal dose threshold was in part driven by suilysin (SLY), the major SS2 virulence factor. At higher doses, the clearance and replication kinetics of bacteria lacking SLY mimicked WT SS2 at sub-lethal doses. Our findings suggest that acute death in SS2 infection is caused by a wholesale failure of the immune system, enabling a large number of bacteria to initiate infection. These results demonstrate how subtle changes in inoculum size can have dramatic consequences for infection outcome and how lethal infection is driven by widened infection bottlenecks. Author SummaryLethal infections are often associated with a specific dose threshold, where increasing inoculum size is associated with an increased likelihood of death. However, a quantitative explanation how this numeric threshold is established is lacking. For example, lethal dose thresholds may reflect a dose where at least one microbe can pass through host-imposed barriers to infection. Alternatively, lethal dose thresholds may reflect a failure of the host immune system, leading to numerous microbes initiating infection. Here, using barcoded Streptococcus suis serotype 2 (SS2), we compare the within-host population dynamics associated between lethal and non-lethal infections. We find that lethal doses are associated with dramatic widening of infection bottlenecks, leading to a large number of bacteria that initiate infection. Lethal infection is also driven by suilysin, the major SS2 virulence factor. These observations demonstrate how lethal doses are quantitatively established by the relationship between host bottlenecks and bacterial virulence factors.

microbiology↗

bioRxiv2023

DisDock: A Deep Learning Method for Metal Ion-Protein Redocking

The structures of metalloproteins are essential for comprehending their functions and interactions. The breakthrough of AlphaFold has made it possible to predict protein structures with experimental accuracy. However, the type of metal ion that a metalloprotein binds and the binding structure are still not readily available, even with the predicted protein structure. In this study, we present DisDock, a physics-driven deep learning method for predicting protein-metal docking. DisDock takes distogram of randomly initialized protein-ligand configuration as input and outputs the distogram of the predicted binding complex. It combines the U-net architecture with self-attention modules to enhance model performance. Taking inspiration from the physical principle that atoms in closer proximity display a stronger mutual attraction, this predictor capitalizes on geometric information to uncover latent characteristics indicative of atom interactions. To train our model, we employ a high-quality metalloprotein dataset sourced from the Mother of All Databases (MOAD). Experimental results demonstrate that our approach outperforms other existing methods in prediction accuracy for various types of metal ions.

bioinformatics↗

bioRxiv2023

Three-dimensional chromatin landscapes in MLLr AML

Rearrangements of the mixed lineage leukemia (MLLr) gene are frequently associated with aggressive acute myeloid leukemia (AML). However, the treatment options are limited due to the genomic complexity and dynamics of 3D structure, which regulate oncogene transcription and leukemia development. Here, we carried out an integrative analysis of 3D genome structure, chromatin accessibility, and gene expression in gene-edited MLL-AF9 AML samples. Our data revealed profound MLLr-specific alterations of chromatin accessibility, A/B compartments, topologically associating domains (TAD), and chromatin loops in AML. The local 3D configuration of the AML genome was rewired specifically at loci associated with AML-specific gene expression. Together, our results show that the 3D chromatin undergoes extensive reorganization at multiple architectural levels, which underpins the remodeling of the transcriptome in MLLr AML.

genomics↗

bioRxiv2023

BioKG: a comprehensive, high-quality biomedical knowledge graph for AI-powered, data-driven biomedical research

To address the rapid growth of scientific publications and data in biomedical research, knowledge graphs (KGs) have become a critical tool for integrating large volumes of heterogeneous data to enable efficient information retrieval and automated knowledge discovery (AKD). However, transforming unstructured scientific literature into KGs remains a significant challenge, with previous methods unable to achieve human-level accuracy. In this study, we utilized an information extraction pipeline that won first place in the LitCoin NLP Challenge (2022) to construct a large-scale KG named iKraph using all PubMed abstracts. The extracted information matches human expert annotations and significantly exceeds the content of manually curated public databases. To enhance the KGs comprehensiveness, we integrated relation data from 40 public databases and relation information inferred from high-throughput genomics data. This KG facilitates rigorous performance evaluation of AKD, which was infeasible in previous studies. We designed an interpretable, probabilistic-based inference method to identify indirect causal relations and applied it to real-time COVID-19 drug repurposing from March 2020 to May 2023. Our method identified 600-1400 candidate drugs per month, with one-third of those discovered in the first two months later supported by clinical trials or PubMed publications. These outcomes are very challenging to attain through alternative approaches that lack a thorough understanding of the existing literature. A cloud-based platform (https://biokde.insilicom.com) was developed for academic users to access this rich structured data and associated tools.

bioinformatics↗

bioRxiv2021

Characterization of epigenetic alterations in esophageal cancer by whole-genome bisulfite sequencing

Esophageal carcinoma is a common and aggressive malignancy, and its patients have dismal clinical outcomes. The epigenetic dysregulation in both major subtypes, esophageal squamous cell carcinoma (ESCC) and adenocarcinoma (EAC), awaits further characterization. Here, we perform whole-genome bisulfite sequencing (WGBS) on a total of 43 esophageal cancer and normal samples, generating one of the largest WGBS datasets in this cancer to date. Focusing on hypomethylated regions in cancer, we show that they are associated with increased chromatin activity and enhancer RNA expression. Using this large collection of WGBS dataset, we reveal and validate novel clusters in both ESCC and EAC. We further identify specific molecular features in each cluster, with potential clinical implications. These data together advance our understanding of the epigenetic alterations in esophageal cancer and provide a rich resource for the research community of this disease.

cancer biology↗

bioRxiv2021

METTL3 promotes homologous recombination repair and modulates chemotherapeutic response by regulating the EGF/Rad51 axis

METTL3 and N6-methyladenosine (m6A) are involved in many types of biological and pathological processes, including DNA repair. However, the function and mechanism of METTL3 in DNA repair and chemotherapeutic response remain largely unknown. In present study, we identified that METTL3 participates in the regulation of homologous recombination repair (HR), which further influences chemotherapeutic response in breast cancer (BC) cells. Knockdown of METTL3 sensitized BC cells to Adriamycin (ADR) treatment and increased accumulation of DNA damage. Mechanically, we demonstrated that inhibition of METTL3 impaired HR efficiency and increased ADR-induced DNA damage by regulating m6A modification of EGF/RAD51 axis. METTL3 promoted EGF expression through m6A modification, which further upregulated RAD51 expression, resulting in enhanced HR activity. We further demonstrated that the m6A "reader," YTHDC1, bound to the m6A modified EGF transcript and promoted EGF synthesis, which enhanced HR and cell survival during ADR treatment. Our findings reveal a pivotal mechanism of METTL3-mediated HR and chemotherapeutic drug response, which may contribute to cancer therapy.

biochemistry↗

bioRxiv2021

Estimating the Designability of Protein Structures

The total number of amino acid sequences that can fold to a target protein structure, known as "designability", is a fundamental property of proteins that contributes to their structure and function robustness. The highly designable structures always have higher thermodynamic stability, mutational stability, fast folding, regular secondary structures, and tertiary symmetries. Although it has been studied on lattice models for very short chains by exhaustive enumeration, it remains a challenge to estimate the designable quantitatively for real proteins. In this study, we designed a new deep neural network model that samples protein sequences given a backbone structure using sequential Monte Carlo method. The sampled sequences with proper weights were used to estimate the designability of several real proteins. The designed sequences were also tested using the latest AlphaFold2 and RoseTTAFold to confirm their foldabilities. We report this as the first study to estimate the designability of real proteins.

bioinformatics↗

bioRxiv2021

Protein loop modeling and refinement using deep learning models

Loops in proteins play essential roles in protein functions and interactions. The structural characterization of loops is challenging because of their conformational flexibility and relatively poor conservation in multiple sequence alignments. Many experimental and computational approaches have been carried out during the last few decades for loop modeling. Although the latest AlphaFold2 achieved remarkable performance in protein structure predictions, the accuracy of loop regions for many proteins still needs to be improved for downstream applications such as protein function prediction and structure based drug design. In this paper, we proposed two novel deep learning architectures for loop modeling: one uses a combined convolutional neural network (CNN)-recursive neural network (RNN) structure (DeepMUSICS) and the other is based on refinement of histograms using a 2D CNN architecture (DeepHisto). In each of the methods, two types of models, conformation sampling model and energy scoring model, were trained and applied in the loop folding process. Both methods achieved promising results and worth further investigations. Since multiple sequence alignments (MSA) were not used in our architecture, the energy scoring models have less bias from MSA. We believe the methods may serve as good complements for refining AlphaFold2 predicted structures.

bioinformatics↗

bioRxiv2018

TRAF6 directs Foxp3 localization and facilitates Treg function through K63-type ubiquitination

Regulatory T cells (Treg) are crucial mediators of immune control. The characteristic gene expression and suppressive function of Treg depend considerably on the stable expression and activity of the transcription factor Foxp3. While transcriptional regulation of the Foxp3 gene has been studied in depth, both the expression and function of Foxp3 are also modulated at the protein level. However, the molecular players involved in posttranslational Foxp3 regulation are just beginning to be elucidated. Here we found TRAF6-deficient Tregs were dysfunctional in vivo; mice with Treg-restricted deletion of TRAF6 were resistant to B16 melanomas and displayed enhanced anti-tumor immunity. We further determined that Foxp3 undergoes lysine-63 chain (K63) ubiquitination at lysine 262 mediated by the E3 ligase TRAF6. When deprived of TRAF6 activity or rendered insensitive to K63 ubiquitination, Foxp3 displayed aberrant, perinuclear accumulation, disrupted function. Thus, Foxp3 ubiquitination by TRAF6 ensures proper localization of Foxp3 and facilitates Foxp3s gene-regulating activity in Tregs. These results implicate TRAF6 as a key posttranslational, Treg-stabilizing force that may be targeted in novel tolerance-breaking therapies.

immunology↗