Enhanced DNA repair through droplet formation and p53 oscillations.

Living organisms are constantly exposed to DNA damage, and optimal repair is therefore crucial. A characteristic hallmark of the response is the formation of sub-compartments around the site of damage, known as foci. Following multiple DNA breaks, the transcription factor p53 exhibits oscillations in its nuclear concentration, but how this dynamics can affect the repair remains unknown. Here, we formulate a theory for foci formation through droplet condensation and discover how oscillations in p53, with its specific periodicity and amplitude, optimize the repair process by preventing Ostwald ripening and distributing protein material in space and time. Based on the theory predictions, we reveal experimentally that the oscillatory dynamics of p53 does enhance the repair efficiency. These results connect the dynamical signaling of p53 with the microscopic repair process and create a new paradigm for the interplay of complex dynamics and phase transitions in biology.

Intestinal epithelial dopamine receptor signaling drives sex-specific disease exacerbation in a mouse model of multiple sclerosis.

Although the gut microbiota can influence central nervous system (CNS) autoimmune diseases, the contribution of the intestinal epithelium to CNS autoimmunity is less clear. Here, we showed that intestinal epithelial dopamine D2 receptors (IEC DRD2) promoted sex-specific disease progression in an animal model of multiple sclerosis. Female mice lacking Drd2 selectively in intestinal epithelial cells showed a blunted inflammatory response in the CNS and reduced disease progression. In contrast, overexpression or activation of IEC DRD2 by phenylethylamine administration exacerbated disease severity. This was accompanied by altered lysozyme expression and gut microbiota composition, including reduced abundance of Lactobacillus species. Furthermore, treatment with N2-acetyl-L-lysine, a metabolite derived from Lactobacillus, suppressed microglial activation and neurodegeneration. Taken together, our study indicates that IEC DRD2 hyperactivity impacts gut microbial abundances and increases susceptibility to CNS autoimmune diseases in a female-biased manner, opening up future avenues for sex-specific interventions of CNS autoimmune diseases.

Intestinal-derived HDL: The portal guardian of the liver.

The trafficking and function of intestine-derived high-density lipoprotein (HDL) have not been identified. In a recent issue of Science, Han et al. (2021) find that intestine-derived HDL neutralizes intestinal-leaked LPS in the portal vein, serving as a host disease tolerance strategy to restrain liver damage of enteric origin under physiological conditions.

Emergence of large-scale cell death through ferroptotic trigger waves.

Large-scale cell death is commonly observed during organismal development and in human pathologies1-5. These cell death events extend over great distances to eliminate large populations of cells, raising the question of how cell death can be coordinated in space and time. One mechanism that enables long-range signal transmission is trigger waves6, but how this mechanism might be used for death events in cell populations remains unclear. Here we demonstrate that ferroptosis, an iron- and lipid-peroxidation-dependent form of cell death, can propagate across human cells over long distances (≥5 mm) at constant speeds (around 5.5 µm min-1) through trigger waves of reactive oxygen species (ROS). Chemical and genetic perturbations indicate a primary role of ROS feedback loops (Fenton reaction, NADPH oxidase signalling and glutathione synthesis) in controlling the progression of ferroptotic trigger waves. We show that introducing ferroptotic stress through suppression of cystine uptake activates these ROS feedback loops, converting cellular redox systems from being monostable to being bistable and thereby priming cell populations to become bistable media over which ROS propagate. Furthermore, we demonstrate that ferroptosis and its propagation accompany the massive, yet spatially restricted, cell death events during muscle remodelling of the embryonic avian limb, substantiating its use as a tissue-sculpting strategy during embryogenesis. Our findings highlight the role of ferroptosis in coordinating global cell death events, providing a paradigm for investigating large-scale cell death in embryonic development and human pathologies.

Histone Deacetylase 3 Couples Mitochondria to Drive IL-1ß-Dependent Inflammation by Configuring Fatty Acid Oxidation.

Immune cell function depends on specific metabolic programs dictated by mitochondria, including nutrient oxidation, macromolecule synthesis, and post-translational modifications. Mitochondrial adaptations have been linked to acute and chronic inflammation, but the metabolic cues and precise mechanisms remain unclear. Here we reveal that histone deacetylase 3 (HDAC3) is essential for shaping mitochondrial adaptations for IL-1ß production in macrophages through non-histone deacetylation. In vivo, HDAC3 promoted lipopolysaccharide-induced acute inflammation and high-fat diet-induced chronic inflammation by enhancing NLRP3-dependent caspase-1 activation. HDAC3 configured the lipid profile in stimulated macrophages and restricted fatty acid oxidation (FAO) supported by exogenous fatty acids for mitochondria to acquire their adaptations and depolarization. Rather than affecting nuclear gene expression, HDAC3 translocated to mitochondria to deacetylate and inactivate an FAO enzyme, mitochondrial trifunctional enzyme subunit α. HDAC3 may serve as a controlling node that balances between acquiring mitochondrial adaptations and sustaining their fitness for IL-1ß-dependent inflammation.

Cholesterol Homeostatic Regulator SCAP-SREBP2 Integrates NLRP3 Inflammasome Activation and Cholesterol Biosynthetic Signaling in Macrophages.

Cholesterol metabolism has been linked to immune functions, but the mechanisms by which cholesterol biosynthetic signaling orchestrates inflammasome activation remain unclear. Here, we have shown that NLRP3 inflammasome activation is integrated with the maturation of cholesterol master transcription factor SREBP2. Importantly, SCAP-SREBP2 complex endoplasmic reticulum-to-Golgi translocation was required for optimal activation of the NLRP3 inflammasome both in vitro and in vivo. Enforced cholesterol biosynthetic signaling by sterol depletion or statins promoted NLPR3 inflammasome activation. However, this regulation did not predominantly depend on changes in cholesterol homeostasis controlled by the transcriptional activity of SREBP2, but relied on the escort activity of SCAP. Mechanistically, NLRP3 associated with SCAP-SREBP2 to form a ternary complex which translocated to the Golgi apparatus adjacent to a mitochondrial cluster for optimal inflammasome assembly. Our study reveals that, in addition to controlling cholesterol biosynthesis, SCAP-SREBP2 also serves as a signaling hub integrating cholesterol metabolism with inflammation in macrophages.

One-Carbon Metabolism Supports S-Adenosylmethionine and Histone Methylation to Drive Inflammatory Macrophages.

Activated macrophages adapt their metabolic pathways to drive the pro-inflammatory phenotype, but little is known about the biochemical underpinnings of this process. Here, we find that lipopolysaccharide (LPS) activates the pentose phosphate pathway, the serine synthesis pathway, and one-carbon metabolism, the synergism of which drives epigenetic reprogramming for interleukin-1ß (IL-1ß) expression. Glucose-derived ribose and one-carbon units fed by both glucose and serine metabolism are synergistically integrated into the methionine cycle through de novo ATP synthesis and fuel the generation of S-adenosylmethionine (SAM) during LPS-induced inflammation. Impairment of these metabolic pathways that feed SAM generation lead to anti-inflammatory outcomes, implicating SAM as an essential metabolite for inflammatory macrophages. Mechanistically, SAM generation maintains a relatively high SAM:S-adenosylhomocysteine ratio to support histone H3 lysine 36 trimethylation for IL-1ß production. We therefore identify a synergistic effect of glucose and amino acid metabolism on orchestrating SAM availability that is intimately linked to the chromatin state for inflammation.

Independent insulin signaling modulators govern hot avoidance under different feeding states.

Thermosensation is critical for the survival of animals. However, mechanisms through which nutritional status modulates thermosensation remain unclear. Herein, we showed that hungry Drosophila exhibit a strong hot avoidance behavior (HAB) compared to food-sated flies. We identified that hot stimulus increases the activity of α'ß' mushroom body neurons (MBns), with weak activity in the sated state and strong activity in the hungry state. Furthermore, we showed that α'ß' MBn receives the same level of hot input from the mALT projection neurons via cholinergic transmission in sated and hungry states. Differences in α'ß' MBn activity between food-sated and hungry flies following heat stimuli are regulated by distinct Drosophila insulin-like peptides (Dilps). Dilp2 is secreted by insulin-producing cells (IPCs) and regulates HAB during satiety, whereas Dilp6 is secreted by the fat body and regulates HAB during the hungry state. We observed that Dilp2 induces PI3K/AKT signaling, whereas Dilp6 induces Ras/ERK signaling in α'ß' MBn to regulate HAB in different feeding conditions. Finally, we showed that the 2 α'ß'-related MB output neurons (MBONs), MBON-α'3 and MBON-ß'1, are necessary for the output of integrated hot avoidance information from α'ß' MBn. Our results demonstrate the presence of dual insulin modulation pathways in α'ß' MBn, which are important for suitable behavioral responses in Drosophila during thermoregulation under different feeding states.

Key processes required for the different stages of fungal carnivory by a nematode-trapping fungus.

Nutritional deprivation triggers a switch from a saprotrophic to predatory lifestyle in soil-dwelling nematode-trapping fungi (NTF). In particular, the NTF Arthrobotrys oligospora secretes food and sex cues to lure nematodes to its mycelium and is triggered to develop specialized trapping devices. Captured nematodes are then invaded and digested by the fungus, thus serving as a food source. In this study, we examined the transcriptomic response of A. oligospora across the stages of sensing, trap development, and digestion upon exposure to the model nematode Caenorhabditis elegans. A. oligospora enacts a dynamic transcriptomic response, especially of protein secretion-related genes, in the presence of prey. Two-thirds of the predicted secretome of A. oligospora was up-regulated in the presence of C. elegans at all time points examined, and among these secreted proteins, 38.5% are predicted to be effector proteins. Furthermore, functional studies disrupting the t-SNARE protein Sso2 resulted in impaired ability to capture nematodes. Additionally, genes of the DUF3129 family, which are expanded in the genomes of several NTF, were highly up-regulated upon nematode exposure. We observed the accumulation of highly expressed DUF3129 proteins in trap cells, leading us to name members of this gene family as Trap Enriched Proteins (TEPs). Gene deletion of the most highly expressed TEP gene, TEP1, impairs the function of traps and prevents the fungus from capturing prey efficiently. In late stages of predation, we observed up-regulation of a variety of proteases, including metalloproteases. Following penetration of nematodes, these metalloproteases facilitate hyphal growth required for colonization of prey. These findings provide insights into the biology of the predatory lifestyle switch in a carnivorous fungus and provide frameworks for other fungal-nematode predator-prey systems.

Monocytes Release Pro-Cathepsin D to Drive Blood-to-Brain Transcytosis in Diabetes.

BACKGROUND: Microvascular complications are the major outcome of type 2 diabetes progression, and the underlying mechanism remains to be determined. METHODS: High-throughput RNA sequencing was performed using human monocyte samples from controls and diabetes. The transgenic mice expressing human CTSD (cathepsin D) in the monocytes was constructed using CD68 promoter. In vivo 2-photon imaging, behavioral tests, immunofluorescence, transmission electron microscopy, Western blot analysis, vascular leakage assay, and single-cell RNA sequencing were performed to clarify the phenotype and elucidate the molecular mechanism. RESULTS: Monocytes expressed high-level CTSD in patients with type 2 diabetes. The transgenic mice expressing human CTSD in the monocytes showed increased brain microvascular permeability resembling the diabetic microvascular phenotype, accompanied by cognitive deficit. Mechanistically, the monocytes release nonenzymatic pro-CTSD to upregulate caveolin expression in brain endothelium triggering caveolae-mediated transcytosis, without affecting the paracellular route of brain microvasculature. The circulating pro-CTSD activated the caveolae-mediated transcytosis in brain endothelial cells via its binding with low-density LRP1 (lipoprotein receptor-related protein 1). Importantly, genetic ablation of CTSD in the monocytes exhibited a protective effect against the diabetes-enhanced brain microvascular transcytosis and the diabetes-induced cognitive impairment. CONCLUSIONS: These findings uncover the novel role of circulatory pro-CTSD from monocytes in the pathogenesis of cerebral microvascular lesions in diabetes. The circulatory pro-CTSD is a potential target for the intervention of microvascular complications in diabetes.

Analysis of bacterial transcriptome and epitranscriptome using nanopore direct RNA sequencing.

Bacterial gene expression is a complex process involving extensive regulatory mechanisms. Along with growing interests in this field, Nanopore Direct RNA Sequencing (DRS) provides a promising platform for rapid and comprehensive characterization of bacterial RNA biology. However, the DRS of bacterial RNA is currently deficient in the yield of mRNA-mapping reads and has yet to be exploited for transcriptome-wide RNA modification mapping. Here, we showed that pre-processing of bacterial total RNA (size selection followed by ribosomal RNA depletion and polyadenylation) guaranteed high throughputs of sequencing data and considerably increased the amount of mRNA reads. This way, complex transcriptome architectures were reconstructed for Escherichia coli and Staphylococcus aureus and extended the boundaries of 225 known E. coli operons and 89 defined S. aureus operons. Utilizing unmodified in vitro-transcribed (IVT) RNA libraries as a negative control, several Nanopore-based computational tools globally detected putative modification sites in the E. coli and S. aureus transcriptomes. Combined with Next-Generation Sequencing-based N6-methyladenosine (m6A) detection methods, 75 high-confidence m6A candidates were identified in the E. coli protein-coding transcripts, while none were detected in S. aureus. Altogether, we demonstrated the potential of Nanopore DRS in systematic and convenient transcriptome and epitranscriptome analysis.

Uncoupling transcription and translation through miRNA-dependent poly(A) length control in haploid male germ cells.

As one of the post-transcriptional regulatory mechanisms, uncoupling of transcription and translation plays an essential role in development and adulthood physiology. However, it remains elusive how thousands of mRNAs get translationally silenced while stability is maintained for hours or even days before translation. In addition to oocytes and neurons, developing spermatids display significant uncoupling of transcription and translation for delayed translation. Therefore, spermiogenesis represents an excellent in vivo model for investigating the mechanism underlying uncoupled transcription and translation. Through full-length poly(A) deep sequencing, we discovered dynamic changes in poly(A) length through deadenylation and re-polyadenylation. Deadenylation appeared to be mediated by microRNAs (miRNAs), and transcripts with shorter poly(A) tails tend to be sequestered into ribonucleoprotein (RNP) granules for translational repression and stabilization. In contrast, re-polyadenylation might allow for translocation of the translationally repressed transcripts from RNP granules to polysomes. Overall, our data suggest that miRNA-dependent poly(A) length control represents a previously unreported mechanism underlying uncoupled translation and transcription in haploid male mouse germ cells.

Fast and accurate protein intrinsic disorder prediction by using a pretrained language model.

Determining intrinsically disordered regions of proteins is essential for elucidating protein biological functions and the mechanisms of their associated diseases. As the gap between the number of experimentally determined protein structures and the number of protein sequences continues to grow exponentially, there is a need for developing an accurate and computationally efficient disorder predictor. However, current single-sequence-based methods are of low accuracy, while evolutionary profile-based methods are computationally intensive. Here, we proposed a fast and accurate protein disorder predictor LMDisorder that employed embedding generated by unsupervised pretrained language models as features. We showed that LMDisorder performs best in all single-sequence-based methods and is comparable or better than another language-model-based technique in four independent test sets, respectively. Furthermore, LMDisorder showed equivalent or even better performance than the state-of-the-art profile-based technique SPOT-Disorder2. In addition, the high computation efficiency of LMDisorder enabled proteome-scale analysis of human, showing that proteins with high predicted disorder content were associated with specific biological functions. The datasets, the source codes, and the trained model are available at https://github.com/biomed-AI/LMDisorder.

SUMOylation and coupling of eNOS mediated by PIAS1 contribute to maintenance of vascular homeostasis.

Endothelial dysfunction (ED) is commonly considered a crucial initiating step in the pathogenesis of numerous cardiovascular diseases. The coupling of endothelial nitric oxide synthase (eNOS) is important in maintaining normal endothelial functions. However, it still remains elusive whether and how eNOS SUMOylation affects the eNOS coupling. In the study, we investigate the roles and possible action mechanisms of protein inhibitor of activated STAT 1 (PIAS1) in ED. Human umbilical vein endothelial cells (HUVECs) treated with palmitate acid (PA) in vitro and ApoE-/- mice fed with high-fat diet (HFD) in vivo were constructed as the ED models. Our in vivo data show that PIAS1 alleviates the dysfunction of vascular endothelium by increasing nitric oxide (NO) level, reducing malondialdehyde (MDA) level, and activating the phosphatidylinositol 3-kinase-protein kinase B-endothelial nitric oxide synthase (PI3K-AKT-eNOS) signaling in ApoE-/- mice. Our in vitro data also show that PIAS1 can SUMOylate eNOS under endogenous conditions; moreover, it antagonizes the eNOS uncoupling induced by PA. The findings demonstrate that PIAS1 alleviates the dysfunction of vascular endothelium by promoting the SUMOylation and inhibiting the uncoupling of eNOS, suggesting that PIAS1 would become an early predictor of atherosclerosis and a new potential target of the hyperlipidemia-related cardiovascular diseases.

Identifying MiR-140-3p as a stable internal reference to normalize MicroRNA qRT-PCR levels of plasma small extracellular vesicles in lung cancer patients.

Exploration of a stably expressed gene as a reference is critical for the accurate evaluation of miRNAs isolated from small extracellular vesicles (sEVs). In this study, we analyzed small RNA sequencing on plasma sEV miRNAs in the training dataset (n = 104) and found that miR-140-3p was the most stably expressed candidate reference for sEV miRNAs. We further demonstrated that miR-140-3p expressed most stably in the validation cohort (n = 46) when compared to two other reference miRNAs, miR-451a and miR-1228-3p, and the commonly-used miRNA reference U6. Finally, we compared the capability of miR-140-3p and U6 as the internal reference for sEV miRNA expression by evaluating key miRNAs expression in lung cancer patients and found that miR-140-3p was more suitable as a sEV miRNA reference gene. Taken together, our data indicated miR-140-3p as a stable internal reference miRNA of plasma sEVs to evaluate miRNA expression profiles in lung cancer patients.

Identification of whole-genome mutations and structural variations of bile cell-free DNA in cholangiocarcinoma.

Bile cell-free DNA (cfDNA) has been reported as a promising liquid biopsy tool for cholangiocarcinoma (CCA), however, the whole-genome mutation landscape and structural variants (SVs) of bile cfDNA remains unknown. Here we performed whole-genome sequencing on bile cfDNA and analyzed the correlation between mutation characteristics of bile cfDNA and clinical prognosis. TP53 and KRAS were the most frequently mutated genes, and the RTK/RAS, homologous recombination (HR), and HIPPO were top three pathways containing most gene mutations. Ten overlapping putative driver genes were found in bile cfDNA and tumor tissue. SVs such as chromothripsis and kataegis were identified. Moreover, the hazard ratio of HR pathway mutations were 15.77 (95% CI: 1.571-158.4), patients with HR pathway mutations in bile cfDNA exhibited poorer overall survival (P = 0.0049). Our study suggests that bile cfDNA contains genome mutations and SVs, and HR pathway mutations in bile cfDNA can predict poor outcomes of CCA patients.

Comprehensive Glycomic and Glycoproteomic Analyses of Human Programmed Cell Death Protein 1 Extracellular Domain.

Human programmed cell death protein 1 (hPD-1) is an essential receptor in the immune checkpoint pathway. It has played an important role in cancer therapy. However, not all patients respond positively to the PD-1 antibody treatment, and the underlying mechanism remains unknown. PD-1 is a transmembrane glycoprotein, and its extracellular domain (ECD) is reported to be responsible for interactions and signal transduction. This domain contains 4 N-glycosylation sites and 25 potential O-glycosylation sites, which implicates the importance of glycosylation. The structure of hPD-1 has been intensively studied, but the glycosylation of this protein, especially the glycan on each glycosylation site, has not been comprehensively illustrated. In this study, hPD-1 ECD expressed by human embryonic kidney 293 (HEK 293) and Chinese hamster ovary (CHO) cells was analyzed; not only N- and O-glycosylation sites but also the glycans on these sites were comprehensively analyzed using mass spectrometry. In addition, hPD-1 ECD binding to different anti-hPD-1 antibodies was tested, and N-glycans were found functioned differently. All of this glycan information will be beneficial for future PD-1 studies.

Inhibition of CC chemokine receptor 1 ameliorates osteoarthritis in mouse by activating PPAR-γ.

BACKGROUND: Osteoarthritis (OA) is a degenerative joint disease characterized by cartilage destruction and inflammation. CC chemokine receptor 1 (CCR1), a member of the chemokine family and its receptor family, plays a role in the autoimmune response. The impact of BX471, a specific small molecule inhibitor of CCR1, on CCR1 expression in cartilage and its effects on OA remain underexplored. METHODS: This study used immunohistochemistry (IHC) to assess CCR1 expression in IL-1ß-induced mouse chondrocytes and a medial meniscus mouse model of destabilization of the medial meniscus (DMM). Chondrocytes treated with varying concentrations of BX471 for 24 h were subjected to IL-1ß (10 ng/ml) treatment. The levels of the aging-related genes P16INK4a and P21CIP1 were analyzed via western blotting, and senescence-associated ß-galactosidase (SA-ß-gal) activity was measured. The expression levels of inducible nitric oxide synthase (iNOS), cyclooxygenase-2 (COX-2), aggrecan (AGG), and the transcription factor SOX9 were determined through western blotting and RT‒qPCR. Collagen II, matrix metalloproteinase 13 (MMP13), and peroxisome proliferator-activated receptor (PPAR)-γ expression was analyzed via western blot, RT‒qPCR, and immunofluorescence. The impact of BX471 on inflammatory metabolism-related proteins under PPAR-γ inhibition conditions (using GW-9662) was examined through western blotting. The expression of MAPK signaling pathway-related molecules was assessed through western blotting. In vivo, various concentrations of BX471 or an equivalent medium were injected into DMM model joints. Cartilage destruction was evaluated through Safranin O/Fast green and hematoxylin-eosin (H&E) staining. RESULTS: This study revealed that inhibiting CCR1 mitigates IL-1ß-induced aging, downregulates the expression of iNOS, COX-2, and MMP13, and alleviates the IL-1ß-induced decrease in anabolic indices. Mechanistically, the MAPK signaling pathway and PPAR-γ may be involved in inhibiting the protective effect of CCR1 on chondrocytes. In vivo, BX471 protected cartilage in a DMM model. CONCLUSION: This study demonstrated the expression of CCR1 in chondrocytes. Inhibiting CCR1 reduced the inflammatory response, alleviated cartilage aging, and retarded degeneration through the MAPK signaling pathway and PPAR-γ, suggesting its potential therapeutic value for OA.

Phenylboronic Acid-Functionalized Ratiometric Surface-Enhanced Raman Scattering Nanoprobe for Selective Tracking of Hg2+ and CH3Hg+ in Aqueous Media and Living Cells.

The development of appropriate molecular tools to monitor different mercury speciation, especially CH3Hg+, in living organisms is attractive because its persistent accumulation and toxicity are very harmful to human health. Herein, we develop a novel activity-based ratiometric SERS nanoprobe to selectively monitor Hg2+ and CH3Hg+ in aqueous media and in vivo. In this nanoprobe, a new bifunctional Raman probe bis-s-s'-[(s)-(4-(ethylcarbamoyl)phenyl)boronic acid] (b-(s)-EPBA) was synthesized and immobilized on the surface of gold nanoparticles via a Au-S bond, in which the phenylboronic acid group was employed as the recognition unit for Hg2+ and CH3Hg+ based on the Hg-promoted transmetalation reaction. In the presence of Hg2+ and CH3Hg+, a new surface-enhanced Raman scattering (SERS) peak aroused from of C-Hg appeared at 1080 cm-1, and the SERS intensity at 1002 cm-1 belonged to the B-O symmetric stretching decreased simultaneously. The quantitative tracking of Hg2+ and CH3Hg+ was realized based on the SERS intensity ratio (I1080/I1303) with rapid response (∼4 min) and high sensitivity, with detection limits of 10.05 and 25.13 nM, respectively. Moreover, the SERS sensor was used for the quantitative detection of Hg2+ and CH3Hg+ in four actual water samples with a high accuracy and excellent recovery. More importantly, cell imaging experiments showed that AuNPs@b-(s)-EPBA could quantitatively detect intracellular CH3Hg+ and had a good concentration dependence in ratiometric SERS imaging. Meanwhile, we demonstrated that AuNPs@b-(s)-EPBA could detect and image CH3Hg+ in zebrafish. We anticipate that AuNPs@b-(s)-EPBA could potentially be used to study the physiological functions related to CH3Hg+ in the future.

Predicting the effects of mutations on protein solubility using graph convolution network and protein language model representation.

Solubility is one of the most important properties of protein. Protein solubility can be greatly changed by single amino acid mutations and the reduced protein solubility could lead to diseases. Since experimental methods to determine solubility are time-consuming and expensive, in-silico methods have been developed to predict the protein solubility changes caused by mutations mostly through protein evolution information. However, these methods are slow since it takes long time to obtain evolution information through multiple sequence alignment. In addition, these methods are of low performance because they do not fully utilize protein 3D structures due to a lack of experimental structures for most proteins. Here, we proposed a sequence-based method DeepMutSol to predict solubility change from residual mutations based on the Graph Convolutional Neural Network (GCN), where the protein graph was initiated according to predicted protein structure from Alphafold2, and the nodes (residues) were represented by protein language embeddings. To circumvent the small data of solubility changes, we further pretrained the model over absolute protein solubility. DeepMutSol was shown to outperform state-of-the-art methods in benchmark tests. In addition, we applied the method to clinically relevant genes from the ClinVar database and the predicted solubility changes were shown able to separate pathogenic mutations. All of the data sets and the source code are available at https://github.com/biomed-AI/DeepMutSol.