Biology subjects

Xiong, C.

Publications and source records attributed to Xiong, C..

At least 37 records · Page 2Linked to original sources

bioRxiv2024

Modelling cell type-specific lncRNA regulatory network in autism with Cycle

Autism spectrum disorder (ASD) is a class of complex neurodevelopment disorders with high genetic heterogeneity. Long non-coding RNAs (lncRNAs) are vital regulators that perform specific functions within diverse cell types and play pivotal roles in neurological diseases including ASD. Therefore, studying the specific regulation of lncRNAs in various cell types is crucial for deciphering ASD molecular mechanisms. Existing computational methods utilize bulk transcriptomics data across all of cells or samples, which could reveal the commonalities of lncRNA regulation in the pathogenesis of ASD, but ignore the specificity of lncRNA regulation across various cell types. Here, we present Cycle (Cell type-specific lncRNA regulatory network) to construct the landscape of cell type-specific lncRNA regulation in ASD. We have found that each ASD cell type is unique in lncRNA regulation, and more than one-third and all of cell type-specific lncRNA regulatory networks are characterized as scale-free and small-world, respectively. Across 17 ASD cell types, we have discovered 19 rewired and 11 conserved modules, and eight rewired and three conserved hubs underlying within the discovered cell type-specific lncRNA regulatory networks. Moreover, the discovered rewired and conserved modules and hubs are significantly enriched in ASD-related terms. Furthermore, more similar ASD cell types tend to be connected with higher strength in the constructed cell similarity network. Finally, the comparison results demonstrate that Cycle is a potential method for uncovering cell type-specific lncRNA regulation.

bioinformatics↗

bioRxiv2024

Enhanced Recognition of a Herbal Compound Epiberberine by a DNA Quadruplex-Duplex Structure

The small molecule epiberberine (EPI) is a natural alkaloid with versatile bioactivities against several diseases, including cancer and bacterial infection. EPI can induce the formation of a unique binding pocket at the 5' side of a human telomeric G-quadruplex (HTG) sequence Q4, resulting in a nanomolar binding affinity (KD approximately 26 nM) with significant fluorescence enhancement upon binding. It is important to understand (1) how EPI binding affects HTG structural stability and (2) how enhanced EPI binding may be achieved through the engineering of the DNA binding pocket. In this work, the EPI binding-induced HTG structure stabilization effect was probed by a peptide nucleic acid (PNA) invasion assay in combination with a series of biophysical techniques. We show that the PNA invasion-based method may be useful for the characterization of compounds binding to DNA (and RNA) structures in physiological conditions without the need to vary the solution temperature or buffer components, which are typically needed for structural stability characterization. Importantly, the combination of theoretical modeling and experimental quantification allows us to successfully engineer the Q4 derivative Q4-ds-A by a simple extension of a duplex structure to Q4 at the 5' end. Q4-ds-A is a superb EPI binder with a KD of 8 nM, with the binding enhancement achieved through the preformation of a binding pocket and a reduced dissociation rate. The tight binding of Q4 and Q4-ds-A with EPI allows us to develop a novel magnetic bead-based affinity purification system to effectively extract EPI from Rhizoma coptidis (Huang Lian) extracts.

biochemistry↗

bioRxiv2024

Investigating spatial dynamics in spatial omics data with StarTrail

Spatial omics technologies revolutionize our view of biological processes within tissues. However, existing methods fail to capture localized, sharp changes characteristic of critical events (e.g. tumor development). Here, we present StarTrail, a novel gradient based method that powerfully defines rapidly changing regions and detects "cliff genes", genes exhibiting drastic expression changes at highly localized or disjoint boundaries. StarTrail, the first to leverage spatial gradients for spatial omics data, also quantifies directional dynamics. Across multiple datasets, StarTrail accurately delineates boundaries (e.g., brain layers, tumor-immune boundaries), and detects cliff genes that may regulate molecular crosstalk at these biologically relevant boundaries but are missed by existing methods. For instance, StarTrail precisely pinpointed the cancer-immune interface in a HER2+ breast cancer dataset, unveiled key cliff genes including a potential prognostic biomarker IGSF3, highlighting NK-, B-cell mediated immunity, and B cell receptor signaling pathways missed by all spatial variable gene methods attempted. StarTrail, filling important gaps in current literature, enables deeper insights into tissue spatial architecture.

genetics↗

bioRxiv2024

Effect of Electroacupuncture Pretreatment on Cognitive Impairment and Mitophagy in Aging Model Rats and its mechanism

ObjectiveInvestigate the effects of "Shuanggu Yitong" EA pretreatment on cognitive impairment, mitochondrial function, and mitophagy in aging model rats, and to analyze the related mechanisms. MethodsForty 3-month-old male SD rats were randomly divided into a blank group, a model group, an EA group, and a sham EA group, with 10 rats in each group. And the Morris water maze test was performed after the intervention. HE staining, to observe the morphological changes of the hippocampus of the model rats. Nissl staining was used to observe the changes in the number of hippocampal neurons in rats, Western Blotting (WB) was used to observe the expression of endogenous PTEN-induced hypothetical kinase 1 (PINK1) and human Parkinsons protein 2 (Parkin) in the hippocampus, spectrophotometry was employed to detect the activity of respiratory chain complex I in the hippocampus of the model rats, and flow cytometry was utilized to detect hippocampal mitochondrial membrane potential (MMP) and hippocampal mitochondrial permeability transition pore (MPTP) opening. Results"Shuanggu Yitong" EA pretreatment relieved the cognitive impairment induced by D-galactose in subacute aging model rats. The mechanism of "Shuanggu Yitong" EA pretreatment in the improvement of cognitive impairment of subacute aging model rats may be related to the enhancement of Pink1/Parkin mediated mitophagy and the timely removal of accumulated abnormal mitochondria, thus improving mitochondrial function. Conclusion"Shuanggu Yitong" EA pretreatment can significantly improve the cognitive impairment induced by D-galactose in subacute aging model rats.

cell biology↗

bioRxiv2023

Modeling Human Spine-Spinal Cord Organogenesis by hPSC-Derived Neuromesodermal Progenitors

Human trunk development, including spine and spinal cord organogenesis, is a coordinated, orderly, and interdependent process with spatiotemporal tissue patterning. However, the underlying cellular and molecular mechanisms remain largely unclear due to the lack of an effective model that can simulate the early development of human body axis. Here, we reported the long-term patterning and dynamic morphogenesis of human trunk through the formation of spine-spinal cord organoids (SSCOs) self-organized from three-dimensional culture of human PSC-derived neuromesodermal progenitors (NMPs). The SSCOs resembled the morphogenetic features of spine and spinal cord along the anterior-posterior axis, and showed the chondro-osteogenic and neural trajectories consistent with developmental dynamics of spine and spinal cord in gestational embryo through single-cell RNA sequencing (scRNA-seq). In addition, we identified a new HMMR+ bipotent cell population with self-renewal ability and neural/mesodermal competence but distinct from NMPs, which may be involved in trunk development and represent an invaluable tool for disease modeling of spine- and spinal cord-related disorders. Graphic Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=200 SRC="FIGDIR/small/549829v1_ufig1.gif" ALT="Figure 1"> View larger version (80K): org.highwire.dtl.DTLVardef@14bfd30org.highwire.dtl.DTLVardef@1db7fe8org.highwire.dtl.DTLVardef@1aa66a1org.highwire.dtl.DTLVardef@380932_HPS_FORMAT_FIGEXP M_FIG C_FIG

developmental biology↗

bioRxiv2023

Functional variations in gamma-secretase activity are critical determinants of the clinical, biomarker, and cognitive progression of autosomal dominant Alzheimer's disease

BackgroundThe balance between production, clearance, and toxicity of A{beta} peptides is central to Alzheimers disease (AD) pathobiology. Though highly variable in terms of age at symptom onset (AAO), hundreds of variants in PSEN1 cause autosomal dominant forms of AD (ADAD) with nearly complete penetrance. PSEN1 forms the catalytic core of the {gamma}-secretase complex and thereby directly mediates the production of longer, aggregation-prone A{beta} peptides relative to shorter, non-aggregating peptides. We hypothesized that the broad AAO and biomarker heterogeneity seen across ADAD would be predictable based on mutation-specific differences in the production of A{beta} species. MethodsA{beta}-37, 38, 40, 42, and 43 production was quantified from 161 unique PSEN1 variants expressed in HEK293 cells. Prediction of AAO was carried out in 106 variants with available AAO and then replicated in 55 variants represented across 190 PSEN1 mutation carriers who have detailed cognitive and biomarker data from the Dominantly Inherited Alzheimers Network (DIAN). ResultsVariations in A{beta} production across the 161 mutations examined in cell-based models were highly predictive of AAO. In those with corresponding in vivo data from the DIAN study, our cell-based {gamma}-secretase composite was strongly associated with biomarker and cognitive trajectories. ConclusionsThese findings elucidate the critical link between {gamma}-secretase function, A{beta} production, and AD progression and offer mechanistic support for the amyloid hypothesis. The approach used here represents a powerful tool to account for heterogeneity in disease progression in ADAD clinical trials and to assess the pathogenicity of variants of unknown significance or with limited family history.

neuroscience↗

bioRxiv2023

Inhibition of muscarinic receptor signaling protects human enteric inhibitory neurons against platin chemotherapy toxicity

GI toxicity is a common dose-limiting adverse effect of platin chemotherapy treatment. Up to 50% of cancer survivors continue to experience symptoms of chronic constipation or diarrhea induced by their chemotherapy for many years after their treatment. This drug toxicity is largely attributed to damage to enteric neurons that innervate the GI tract and control GI motility. The mechanisms responsible for platin-induced enteric neurotoxicity and potential preventative strategies have remained unknown. Here, we use human pluripotent stem cell derived enteric neurons to establish a new model system capable of uncovering the mechanism of platin-induced enteric neuropathy. Utilizing this scalable system, we performed a high throughput screen and identified drug candidates and pathways involved in the disease. Our analyses revealed that excitotoxicity through muscarinic cholinergic signaling is a key driver of platin-induced enteric neuropathy. Using single nuclei transcriptomics and functional assays, we discovered that this disease mechanism leads to increased susceptibility of specific neuronal subtypes, including inhibitory nitrergic neurons, to platins. Histological assessment of the enteric nervous system in platin-treated patients confirmed the selective loss of nitrergic neurons. Finally, we demonstrated that pharmacological and genetic inhibition of muscarinic cholinergic signaling is sufficient to rescue enteric neurons from platin excitotoxicity in vitro and can prevent platin-induced constipation and degeneration of nitrergic neurons in mice. These studies define the mechanisms of platin-induced enteric neuropathy and serve as a framework for uncovering cell type-specific manifestations of cellular stress underlying numerous intractable peripheral neuropathies.

pharmacology and toxicology↗

bioRxiv2022

Sequencing-free Tissue-wide Spatial Profiling of Post-transcriptional Regulations

The importance of genetic or epi-genetic heterogeneity has been increasingly recognized, but it has been challenging to profile intracellular post-transcriptional targets with sufficient throughput and resolution at across large-scale tissue samples. This study describes a technique, Spectrum-FISH, for high-throughput, sequencing-free, and tissue-wide spatial profiling of various post-transcriptional targets in acute tissue sections with subcellular resolution. The platform uses a biochip with an array of vertically aligned nanoprobes to effectively extract intracellular molecules for downstream analysis in the coordinates of the large-scale of cells within a tissue slice. As a proof-of-concept, the Spectrum-FISH is used to profile the spatial dynamics of 24 miRNAs and 9 m6A-modified messenger RNAs (m6A-mRNA) in acute olfactory bulb (OB) slices of millimeter scale. The results showed potentially multiomics spatial heterogeneity for the examined post-transcriptional regulations in rodent OB, especially in the outer plexiform layer and granule layer, where highly correlated miRNAs and m6A-mRNAs groups were identified, indicating a potential cooperative involvement of different post-transcriptional regulations at these OB regions.

genomics↗

bioRxiv2022

Loss of estrogen unleashing neuro-inflammation increases the risk of Alzheimer's disease in women

The risk of Alzheimers disease (AD) in women is about 2 times greater than in men. The estrogen hypothesis is being accepted as the essential sex factor causing the sex difference in AD. Also, the recent meta-analysis using large-scale medical records data indicated estrogen replacement therapy. However, the underlying molecular targets and mechanisms explaining this sex difference in AD disease development remain unclear. In this study, we identified that estrogen treatment can strongly inhibition of neuro-inflammation signaling targets, using the systems pharmacology model; and identified ESR1/ESR2 (the receptors of estrogen) are topologically close to the neuroinflammation biomarker genes using signaling network analysis. Moreover, the estrogen level in women decreased to an extremely lower level than in men after age 55. Pooling together the multiple pieces of evidence, it is concluded that the loss of estrogen unleashing neuro-inflammation increases the womens risk of Alzheimers disease. These analysis results provide novel supporting evidence explaining the potential mechanism of the anti-neuroinflammation role of estrogen causing the sex difference of AD. Medications boosting the direct downstream signaling of ESR1/ESR2, or inhibiting upstream signaling targets of neuroinflammation, like JAK2 inhibitors, on the signaling network can be potentially effective or synergistic combined with estrogen for AD prevention and treatment.

neuroscience↗

bioRxiv2022

Brain aerobic glycolysis and resilience in Alzheimer disease

The distribution of brain aerobic glycolysis (AG) in normal young adults correlates spatially with amyloid-beta (A{beta}) deposition in individuals with dementia of the Alzheimer type (DAT) and asymptomatic individuals with brain amyloid deposition. Brain AG decreases with age but the functional significance of this decrease with regard to the development of DAT symptomatology is poorly understood. Using PET measurements of regional blood flow, oxygen consumption and glucose utilization--from which we derive AG--we find that cognitive impairment is strongly associated with loss of the typical youthful pattern of AG. In contrast, amyloid positivity without cognitive impairment was associated with preservation of youthful brain AG, which was even higher than that seen in typical, cognitively unimpaired, amyloid negative adults. Similar findings were not seen for blood flow nor oxygen consumption. Finally, in cognitively unimpaired adults, white matter hyperintensity burden was found to be specifically associated with decreased youthful brain AG. Our results implicate preserved AG as a factor in brain resilience to amyloid pathology and suggest that white matter disease may be a cause and/or consequence of this impaired resilience.

neuroscience↗

bioRxiv2022

Configurable Compartmentation Enables In vitro Reconstitution of Sustained Synthetic Biology Systems

The compartmentalized and communicative nature of biological cells contributes to the complexity and endurance of living organisms. Current in vitro compartmentalization systems such as droplet emulsions reproduce the compartmentalization property of cells yet fail to recapture the configurability of cellular communication with the environment. To mimic biological cells a step further and expand the capabilities of in vitro compartmentalization, we present here a general strategy that inherits the passive transport phenomenon of biology. The strategy incorporates layered, micrometer-sized, hydrogel-based compartments featuring configurability in composition, functionality, and selective permeability of biomolecules. We demonstrated the unique advantage of our strategy in two scenarios of synthetic biology. First, a compartmentalized cell-free protein synthesis system was reconstituted that could support multiple rounds of reactions. Second, we constructed living bacteria-based biosensors in the hydrogel compartments, which could achieve long-lasting functioning with markedly enhanced fitness in complex environments. Looking forward, our strategy should be widely applicable for constructing complex, robust, and sustained in vitro synthetic molecular and cellular systems, paving the way for their practical applications.

synthetic biology↗

bioRxiv2021

ASF1 activation PI3K/AKT pathway regulates sexual and asexual development in filamentous ascomycete

Sexual and asexual reproduction is ubiquitous in eukaryotes. PI3K/AKT signaling pathway can modulate sexual reproduction in mammals. However, this signaling pathway modulating sexual and asexual reproduction in fungi is scarcely understood. SeASF1, a SeH4 chaperone, could manipulate sexual and asexual reproduction of Stemphylium eturmiunum. SeDJ-1, screened from Se{Delta}asf1 transcriptome, was confirmed to regulate sexual and asexual development by RNAi, of which the mechanism was demonstrated by detecting transcriptional levels and protein interactions of SeASF1, SeH4 and SeDJ-1 by qRT-PCR, and Y2H, Co-IP and Pull-down, respectively. SeASF1 coupling SeH4 bound SeDJ-1 to arouse the sexual and asexual activity. In S. eturmiunum genome, SeDJ-1 was upstream while SeGSK3 was downstream in PI3K/AKT signaling pathway. Moreover, SeDJ-1 interacted with SePI3K or SeGSK3 in vivo and in vitro. Significantly, SeDJ-1 or SePI3K could effectively stimulate sexual activity alone, but SePI3K could recover the sexual development of SiSeDJ-1. Meanwhile, SeDJ-1-M6 was a critical segment for interaction of SeDJ-1 with SePI3K. SeDJ-1-M6 played a critical role in irritating sexual reproduction in SiSePI3K, which further uncovered the regulated mechanism of SeDJ-1. Summarily, SeASF1 coupling SeH4 motivates SeDJ-1 to arouse SePI3K involved in sexual reproduction. Thus, SeASF1 can activate PI3K/AKT signaling pathway to regulate sexual and asexual development in filamentous ascomycete.

microbiology↗

bioRxiv2021

Weakly activated core inflammation pathways were identified as a central signaling mechanism contributing to the chronic neurodegeneration in Alzheimer's disease

Neuro-inflammation signaling has been identified as an important hallmark of Alzheimers disease (AD) in addition to amyloid {beta} plaques (A{beta}) and neurofibrillary tangles (NFTs). However, our knowledge of neuro-inflammation is very limited; and the core signaling pathways associated with neuro-inflammation are missing. From a novel perspective, i.e., investigating weakly activated molecular signals (rather than the strongly activated molecular signals), in this study, we uncovered the core neuro-inflammation signaling pathways in AD. Our novel hypothesis is that weakly activated neuro-inflammation signaling pathways can cause neuro-degeneration in a chronic process; whereas, strongly activated neuro-inflammation often cause acute disease progression like in COVID-19. Using the two large-scale genomics datasets, i.e., Mayo Clinic (77 control and 81 AD samples) and RosMap (97 control and 260 AD samples), our analysis identified 7 categories of signaling pathways implicated on AD and related to virus infection: immune response, x-core signaling, apoptosis, lipid dysfunctional, biosynthesis and metabolism, and mineral absorption signaling pathways. More interestingly, most of genes in the virus infection, immune response and x-core signaling pathways, are associated with inflammation molecular functions. Specifically, the x-core signaling pathways were defined as a group of 9 signaling proteins: MAPK, Rap1, NF-kappa B, HIF-1, PI3K-Akt, Wnt, TGF-beta, Hippo and TNF, which indicated the core neuro-inflammation signaling pathways responding to the low-level and weakly activated inflammation and hypoxia, and leading to the chronic neuro-degeneration. The core neuro-inflammation signaling pathways can be used as novel therapeutic targets for effective AD treatment and prevention.

neuroscience↗

bioRxiv2021

Botrytis cinerea BcPTP1 is a late infection phase, cysteine rich protein cytotoxic effector

Botrytis cinerea is a broad-host-range necrotrophic phytopathogen responsible for serious crops diseases. To facilitate infection, B. cinerea secretes a large number of effectors that induce plant cell death. In screening secretome data of B. cinerea during infection stage, we identified a phytotoxic protein (BcPTP1) that can also induce immune resistance in plants. BcPTP1 is a small (90 aa), cysteine rich protein without any known domains. Transiently expression of BcPTP1 in leaves caused chlorosis that intensifies with time and eventually lead to cell death. Point mutations in eight of the 10 cysteine residues of BcPTP1 abolished the toxic effect, however residual toxic activity remained after heating the peptide, suggesting contribution of unknown epitopes to protein phytotoxic effect. The transcript level of the bcptp1 gene was low during the first 36 h after inoculation and increased sharply upon transition to the late infection stage, suggesting a role of BcPTP1 in lesion spreading. While statistically insignificant, deletion of the bcptp1 gene led to slightly smaller lesions on bean leaves. Further analyses indicated that BcPTP1 is internalized into plant cells after secreting into the apoplast and its phytotoxic effect is negatively regulated by the receptor-like kinases BAK1 and SOBIR1. Collectively, our findings show that BcPTP1 is a virulence factor that toxifies the host cells and facilitates lesion spreading during the late infection stage.

pathology↗

bioRxiv2021

Deep neural language modeling enables functional protein generation across families

Bypassing natures evolutionary trajectory, de novo protein generation--defined as creating artificial protein sequences from scratch--could enable breakthrough solutions for biomedical and environmental challenges. Viewing amino acid sequences as a language, we demonstrate that a deep learning-based language model can generate functional artificial protein sequences across families, akin to generating grammatically and semantically correct natural language sentences on diverse topics. Our protein language model is trained by simply learning to predict the next amino acid for over 280 million protein sequences from thousands of protein families, without biophysical or coevolutionary modeling. We experimentally evaluate model-generated artificial proteins on five distinct antibacterial lysozyme families. Artificial proteins show similar activities and catalytic efficiencies as representative natural lysozymes, including hen egg white lysozyme, while reaching as low as 44% identity to any known naturally-evolved protein. The X-ray crystal structure of an enzymatically active artificial protein recapitulates the conserved fold and positioning of active site residues found in natural proteins. We demonstrate our language models ability to be adapted to different protein families by accurately predicting the functionality of artificial chorismate mutase and malate dehydrogenase proteins. These results indicate that neural language models successfully perform de novo protein generation across protein families and may prove to be a tool to shortcut evolution.

synthetic biology↗

bioRxiv2020

Deciphering the mechanism underlying circRNA-mediated immune responses of western honeybees to Nosema ceranae infection

Nosema ceranae is a widespread fungal parasite for adult honeybees, severely damaging bee health and sustainable development of apiculture. Circular RNAs (circRNAs) are a class of newly discovered noncoding RNAs (ncRNAs) that regulate a number of biological processes such as immune defense and development. In this current work, based on previously obtained whole transcriptome data, 8 199 and 8 711 circRNAs were predicted from the midguts of Apis mellifera ligustica workers at 7 days (AmT1) and 10 days (AmT2) post inoculation (dpi) with N. ceranae using bioinformatics. Additionally, in combination with transcriptome data from uninfected midguts (AmCK1 and AmCK2) (Xiong et al., 2018), 4 464 circRNAs were found to be shared by the aforementioned four groups, whereas the numbers of specifically transcribed circRNAs in each group were 1 389, 1 696, 1 019 and 1 871, respectively. Furthermore, 10 226 circRNAs were homologous to Apis cerana cerana circRNAs, while 20 circRNAs had homology with Homo sapiens circRNAs; in addition, 16 circRNAs were highly conserved in these three species. Differential expression analysis showed that 168 (306) differentially expressed circRNAs (DEcircRNAs) were identified in AmCK1 vs AmT1 (AmCK2 vs AmT2) comparison group, including 61 (143) upregulated circRNAs and 107 (163) downregulated circRNAs. Moreover, RT-qPCR results showed that the expression trend of eight DEcircRNAs was consistent with that of the transcriptome dataset. Based on GO database annotation, we observed that source genes of DEcircRNAs in AmCK1 vs AmT1 (AmCK2 vs AmT2) were engaged in 27 (35) functional terms, including two (two) cell renewal-associated terms, seven (seven) cell structure-associated terms, and one (one) immunity-associated terms. Additionally, DEcircRNA source genes in AmCK1 vs AmT1 were involved in two cell renewal-related pathways, Hippo and Wnt signaling pathways, and three carbohydrate metabolism-related pathways, galactose metabolism, starch and sucrose metabolism, fructose and mannose metabolism, only one energy metabolism-related pathway (oxidative phosphorylation pathway), three cellular immune-related pathways, endocytosis, phagosome, and lysosome, and a humoral immune-related pathway (FoxO signaling pathway). In AmCK2 vs AmT2 comparison group, more source genes of DEcircRNAs were associated with the abovementioned pathways relative to cell renewal, carbohydrate metabolism, and cellular and humoral immune pathways. In addition, 122 (234) DEcircRNAs in the host midgut at 7 dpi (10 dpi) with N. ceranae targeted 82 (106) miRNAs. Furthermore, 75 (103) miRNAs targeted by 86 (178) DEcircRNAs in AmCK1 vs AmT1 (AmCK2 vs AmT2) further bound to 215 (305) mRNAs. These targets could be annotated as an array of functional terms and pathways related to cellular renewal, cellular structure, carbohydrate and energy metabolism, and cellular and humoral immunity. In a word, we for the first time explored immune responses mediated by DEcircRNAs in the midguts of A. m. ligustica workers to N. ceranae infection. Our data provide a foundation for clarifying the molecular mechanism underlying immune response of western honeybee to N. ceranae invasion, but also a new insight into further understanding the host-pathogen interaction during bee microsporidiosis.

genomics↗

bioRxiv2020

BERTology Meets Biology: Interpreting Attention in Protein Language Models

Transformer architectures have proven to learn useful representations for protein classification and generation tasks. However, these representations present challenges in interpretability. Through the lens of attention, we analyze the inner workings of the Transformer and explore how the model discerns structural and functional properties of proteins. We show that attention (1) captures the folding structure of proteins, connecting amino acids that are far apart in the underlying sequence, but spatially close in the three-dimensional structure, (2) targets binding sites, a key functional component of proteins, and (3) focuses on progressively more complex biophysical properties with increasing layer depth. We also present a three-dimensional visualization of the interaction between attention and protein structure. Our findings align with known biological processes and provide a tool to aid discovery in protein engineering and synthetic biology. The code for visualization and analysis is available at https://github.com/salesforce/provis.

bioinformatics↗

bioRxiv2020

Human Induced Pluripotent Stem Cell Derived Sensory Neurons are Sensitive to the Neurotoxic Effects of Paclitaxel

Chemotherapy-induced peripheral neuropathy (CIPN) is a dose-limiting adverse event associated with treatment with paclitaxel and other chemotherapeutic agents. The prevention and treatment of CIPN are limited by a lack of understanding of the molecular mechanisms underlying this toxicity. In the current study, a human induced pluripotent stem cell–derived sensory neuron (iPSC-SN) model was developed for the study of chemotherapy-induced neurotoxicity. The iPSC-SNs express proteins characteristic of nociceptor, mechanoreceptor and proprioceptor sensory neurons and show Ca2+ influx in response to capsaicin, α,β-meATP and glutamate. iPSC-SNs are relatively resistant to the cytotoxic effects of paclitaxel, with IC50 values of 38.1 μM (95% CI: 22.9 – 70.9 μM) for 48 hr exposure and 9.3 μM (95% CI: 5.7 – 16.5 μM) for 72 hr treatment. Paclitaxel causes dose- and time-dependent changes in neurite network complexity detected by βIII-tubulin staining and high content imaging. The IC50 for paclitaxel reduction of neurite area was 1.4 μM (95% CI: 0.3 - 16.9 μM) for 48 hr exposure and 0.6 μM (95% CI: 0.09 - 9.9 μM) for 72 hr exposure. Decreased mitochondrial membrane potential, slower movement of mitochondria down the neurites and changes in glutamate-induced neuronal excitability were also observed with paclitaxel exposure. The iPSC-SNs were also sensitive to docetaxel, vincristine and bortezomib. Collectively, these data support the use of iPSC-SNs for detailed mechanistic investigations of genes and pathways implicated in chemotherapy-induced neurotoxicity and the identification of novel therapeutic approaches for its prevention and treatment.Competing Interest StatementThe authors have declared no competing interest.View Full Text

pharmacology and toxicology↗