Biology subjects

Lin, Z.

Publications and source records attributed to Lin, Z..

At least 181 records · Page 10Linked to original sources

bioRxiv2022

Molecular basis of stepwise cyclic tetra-adenylate cleavage by the type III CRISPR ring nuclease Crn1/Sso2081

The cyclic oligoadenylates (cOAs) act as second messengers of type III CRISPR immunity system through activating the auxiliary nucleases for indiscriminate RNA degradation. The cOA-degrading nucleases (ring nucleases) provide an off-switch regulation of the signaling, thereby preventing cell dormancy or cell death. Here, we describe the crystal structures of the CRISPR-associated ring nuclease 1 (Crn1) from Saccharolobus solfataricus (Sso) 2081 in its apo or bound to cA4 in both pre-cleavage and transient intermediate states. Sso2081 harbors a unique helical insert that encloses cA4 in the central cavity. Two free phosphates symmetrically bind the catalytic site of apo Sso2081 and overlap with the two scissile phosphates of cA4, supporting a bilaterally symmetrical cleavage. The structure of transient intermediate state captured by Ser11Ala mutation immediately illustrates a stepwise cleavage of cA4 by Sso2081. Our study establishes atomic mechanisms of cA4 recognition and degradation by the type III CRISPR ring nuclease Crn1/Sso2081.

molecular biology↗

bioRxiv2022

Spatiotemporally resolved transcriptomics reveals subcellular RNA kinetic landscape

Spatiotemporal regulation of the cellular transcriptome is crucial for proper protein expression and cellular function1. However, the intricate subcellular dynamics of RNA synthesis, decay, export, and translocation remain obscured due to the limitations of existing transcriptomics methods2-8. Here, we report a spatiotemporally resolved RNA mapping method (TEMPOmap) to uncover subcellular RNA profiles across time and space at the single-cell level in heterogeneous cell populations. TEMPOmap integrates pulse-chase metabolic labeling of the transcriptome with highly multiplexed three-dimensional (3D) in situ sequencing to simultaneously profile the age and location of individual RNA molecules. Using TEMPOmap, we constructed the subcellular RNA kinetic landscape of 991 genes in human HeLa cells from upstream transcription to downstream subcellular translocation. Clustering analysis of critical RNA kinetic parameters across single cells revealed kinetic gene clusters whose expression patterns were shaped by multi-step kinetic sculpting. Importantly, these kinetic gene clusters are functionally segregated, suggesting that subcellular RNA kinetics are differentially regulated to serve molecular and cellular functions in cell-cycle dependent manner. Together, these single-cell spatiotemporally resolved transcriptomics measurements provide us the gateway to uncover new gene regulation principles and understand how kinetic strategies enable precise RNA expression in time and space.

genomics↗

bioRxiv2022

Transcriptome-wide characterization of piRNAs during the developmental process of European honey bee larval guts

piRNAs play pivotal roles in suppressing transposons, maintaining genome stability, regulating gene expression, and modulating development and immunity. However, there are few piRNA-associated studies on honey bee, and the regulatory role of piRNAs in the development of bee guts is largely unknown. In this current work, the differential expression pattern of piRNAs during the developmental process of the European honey bee (Apis mellifera) was analyzed, and target prediction of differentially expressed piRNAs (DEpiRNAs) was then conducted, followed by investigation of regulatory networks and the potential function of DEpiRNAs in regulating larval gut development. Here, a total of 843 piRNAs were identified in the larval guts of A. mellifera; among these, 764 piRNAs were shared by 4- (Am4 group), 5- (Am5 group), and 6-day-old (Am6 group) larval guts, while 11, 67, and 1, respectively, were unique. The first base of piRNAs in each group had a cytosine (C) bias. Additionally, 61 up-regulated and 17 down-regulated piR-NAs were identified in the Am4 vs. Am5 comparison group, further targeting 9, 983 genes, which were involved in 50 GO terms and 142 pathways, while two up-regulated and five down-regulated piRNAs were detected in the Am5 vs. Am6 comparison group, further targeting 1, 936 genes, which were engaged in 41 functional terms and 101 pathways. piR-ame-742536 and piR-ame-856650 in the Am4 vs. Am5 comparison group as well as piR-ame-592661 and piR-ame-31653 in the Am5 vs. Am6 comparison group were found to link to the highest number of targets. Further analysis indicated that targets of DEpiRNAs in these two comparison groups were seven development-associated signaling pathways such as Hippo and Notch signaling pathways, seven immune-associated pathways including endocytosis and the Jak/STAT signaling pathway, and three energy metabolism pathways, namely sulfur metabolism, nitrogen metabolism, and oxidative phosphorylation. Moreover, the expression trends of five randomly selected DEpiRNAs were verified based on stem-loop RT-PCR and RT-qPCR. These results were suggestive of the overall alteration of piRNAs during the larval developmental process and demonstrated that DEpiRNAs potentially modulate development-, immune-, and energy metabolism-associated pathways by regulating expression of corresponding genes via target binding, further affecting the development of A. mellifera larval guts. Our data offer novel insights into the development of bee guts and lay a basis for clarifying the developmental mechanism underlying the larval guts of European honey bee.

molecular biology↗

bioRxiv2022

Molecular and structural characterization of lenalidomide-mediated sequestration of eIF3i

Lenalidomide is a ligand of the E3 ligase substrate adaptor cereblon (CRBN) that achieves its clinical effects in part by promotion of substrate recruitment and degradation. In contrast to prior substrates, eIF3i is recruited but not degraded upon complex formation with lenalidomide and CRBN, although the structural details and mechanistic outcomes of this interaction are unresolved. Here, we characterize the structural basis and mechanistic outcomes of lenalidomide-induced sequestration of eIF3i from the eIF3 complex. Identification of the binding interface on eIF3i by a covalent lenalidomide probe and chemical proteomics rationalizes the sequestration event. We further connect eIF3i and CRBN to known lenalidomide effects on angiogenic markers, Akt1 phosphorylation, and associated antiangiogenesis phenotypes. Finally, we find that eIF3i sequestration is observed in MM.1S and MOLM13 cells after the degradation of other substrates, such as IKZF1. The defined binding interface elucidated by chemical proteomics, and connection of eIF3i sequestration to phenotypic outcomes of lenalidomide open future directions in designing new chemical adaptors for protein sequestration as a strategy to selectively control translational profiles and downstream cellular function.

biochemistry↗

bioRxiv2022

Language models of protein sequences at the scale of evolution enable accurate structure prediction

Artificial intelligence has the potential to open insight into the structure of proteins at the scale of evolution. It has only recently been possible to extend protein structure prediction to two hundred million cataloged proteins. Characterizing the structures of the exponentially growing billions of protein sequences revealed by large scale gene sequencing experiments would necessitate a break-through in the speed of folding. Here we show that direct inference of structure from primary sequence using a large language model enables an order of magnitude speed-up in high resolution structure prediction. Leveraging the insight that language models learn evolutionary patterns across millions of sequences, we train models up to 15B parameters, the largest language model of proteins to date. As the language models are scaled they learn information that enables prediction of the three-dimensional structure of a protein at the resolution of individual atoms. This results in prediction that is up to 60x faster than state-of-the-art while maintaining resolution and accuracy. Building on this, we present the ESM Metage-nomic Atlas. This is the first large-scale structural characterization of metagenomic proteins, with more than 617 million structures. The atlas reveals more than 225 million high confidence predictions, including millions whose structures are novel in comparison with experimentally determined structures, giving an unprecedented view into the vast breadth and diversity of the structures of some of the least understood proteins on earth.

synthetic biology↗

bioRxiv2022

Graphene Biointerface for Cardiac Arrhythmia Diagnosis and Treatment

Heart rhythm disorders, known as arrhythmias, cause significant morbidity and are one of the leading causes of mortality. Cardiac arrhythmias are primarily treated by implantable devices, such as pacemakers and defibrillators, or by ablation therapy guided by electroanatomical mapping. Pharmacological treatments are mostly ineffective. Both implantable and ablation therapies require sophisticated biointerfaces for electrophysiological measurements of electrograms and delivery of therapeutic stimulation or ablation energy. In this work, we report for the first time on graphene biointerface for in vivo cardiac electrophysiology. Leveraging sub-micrometer thick tissue-conformable graphene arrays, we demonstrate sensing and stimulation of the open mammalian heart both in vitro and in vivo. Furthermore, we demonstrate graphene pacemaker treatment of a pharmacologically-induced arrhythmia, AV block. The arrays show effective electrochemical properties, namely interface impedance down to 40 Ohmxcm2 at 1kHz, charge storage capacity up to 63.7 mC/cm2, and charge injection capacity up to 704 C/cm2. Transparency of the graphene structures allows for simultaneous optical mapping of cardiac action potentials and optogenetic stimulation while performing electrical measurements and stimulation. Our report presents evidence of the significant potential of graphene biointerfaces for the future clinical device- and catheter-based cardiac arrhythmias therapies.

bioengineering↗

bioRxiv2022

Spatial Atlas of the Mouse Central Nervous System at Molecular Resolution

Spatially charting molecular cell types at single-cell resolution across the three-dimensional (3D) volume of the brain is critical for illustrating the molecular basis of the brain anatomy and functions. Single-cell RNA sequencing (scRNA-seq) has profiled molecular cell types in the mouse brain1, 2, but cannot capture their spatial organization. Here, we employed an in situ sequencing technique, STARmap PLUS3, 4, to map more than one million high-quality cells across the whole adult mouse brain and the spinal cord, profiling 1,022 genes at subcellular resolution with a voxel size of 194 X 194 X 345 nm in 3D. We developed computational pipelines to segment, cluster, and annotate 231 molecularly defined cell types and 64 tissue regions with single-cell resolution. To create a transcriptome-wide spatial atlas, we further integrated the STARmap PLUS measurements with a published scRNA-seq atlas1, imputing 11,844 genes at the single-cell level. Finally, we engineered a highly expressed RNA barcoding system to delineate the tropism of a brain-wide transgene delivery tool, AAV-PHP.eB5, 6, revealing its single-cell resolved transduction efficiency across the molecular cell types and tissue regions of the whole mouse brain. Together, our datasets and annotations provide a comprehensive, high-resolution single-cell resource that integrates a spatial molecular atlas, cell taxonomy, brain anatomy, and genetic manipulation accessibility of the mammalian central nervous system (CNS).

neuroscience↗

bioRxiv2022

The domain-separation low-dimensional language network dynamics in the resting-state support the flexible functional segregation and integration during language and speech processing

Modern linguistic theories and network science propose that the language and speech processing is organized into hierarchical, segregated large-scale subnetworks, with a core of dorsal (phonological) stream and ventral (semantic) stream. The two streams are asymmetrically recruited in receptive and expressive language or speech tasks, which showed flexible functional segregation and integration. We hypothesized that the functional segregation of the two streams was supported by the underlying network segregation. A dynamic conditional correlation approach was employed to construct frame-wise time-varying language networks and investigate the temporal reoccurring patterns. We found that the time-varying language networks in the resting-state robustly clustered into four low-dimensional states, which dynamically reconfigured following a domain-separation manner. Spatially, the hub distributions of the first three states highly resembled the neurobiology of primary auditory processing and lexical-phonological processing, motor and speech production processing, and semantic processing, respectively. The fourth state was characterized by the weakest functional connectivity and subserved as a baseline state. Temporally, the first three states appeared exclusively in limited time bins ([~]15%), and most of the time (> 55%), the language network kept inactive in state 4. Machine learning-based dFC-linguistics prediction analyses showed that dFCs of the four states significantly predicted individual linguistic performance. These findings suggest a domain-separation manner of language network dynamics in the resting-state, which forms a dynamic "meta-networking" (network of networks) framework. HighlightsO_LIThe time-varying language network in the resting-state is robustly clustered into four low-dimensional states. C_LIO_LISpatially, the first three dFC states are cognitively meaningful, which highly resemble the neurobiology of primary auditory processing and lexical-phonological representation, speech production processing, and semantic processing, respectively. C_LIO_LITemporally, the first three states appeared exclusively in limited time bins ([~]15%), and most of the time (> 55%), the language network kept inactive in state 4. C_LIO_LIA dynamic "meta-networking" framework of language network in the resting-state is proposed. C_LI

neuroscience↗

bioRxiv2022

A Modular Chemoenzymatic Approach to C14-Functionalized Steroids

C14-functionalized steroids belong to a unique class of steroids with important biological activities. However, the lack of efficient methods to access C14-functionalized steroids impede related steroidal drug discovery. Herein we report a modular chemoenzymatic approach to access diversified C14-functionalized steroids. We first identified a novel C14-hydroxylase (CYP14A) from Cochliobolus lunatus with high catalytic efficiency and substrate promiscuity. Protein engineering of CYP14A generated three variants I111A, M115K and V124A that greatly improved the C14-hydroxy regioselectivity. Based on this efficient biocatalytic method, a range of C14-OH steroids with C17 side chain were prepared in good yields, which was then transformed into {Delta}14 olefins through a facile elimination. The newly formed {Delta}14 olefin served as a versatile handle to install diversified functional groups (e.g. epoxide, {beta}-OH, F, Cl and N3) at C14 position through hydrofunctionalization. Furthermore, the synthetic utility of this powerful chemoenzymatic methodology was demonstrated by performing a 7-step semisynthesis of periplogenin and the diversity-oriented synthesis of cardenolide (+)-digitoxigenin and its three diastereomers in a concise manner.

biochemistry↗

bioRxiv2022

Multi-task learning for single-cell multi-modality biology

Current biotechnologies can simultaneously measure multi-modality high-dimensional information from the same cell and tissue samples. To analyze the multi-modality data, common tasks such as joint data analysis and cross-modal prediction have been developed. However, current analytical methods are generally designed to process multi-modality data for one specific task without considering the underlying connections between tasks. Here, we present UnitedNet, a multi-task deep neural network that integrates the tasks of joint group identification and cross-modal prediction to analyze multi-modality data. We have found that multi-task learning for joint group identification and cross-modal prediction significantly improves the performance of each task. When applied to various single-cell multi-modality datasets, UnitedNet shows superior performance in each task, achieving better unsupervised and supervised joint group identification and cross-modal prediction performances compared with state-of-the-art methods. Furthermore, by considering the spatial information of cells as one modality, UnitedNet substantially improves the accuracy of tissue region identification and enables spatially resolved cross-modal prediction.

genomics↗

bioRxiv2022

Design of a systemic small molecule clinical STING agonist using physics-based simulations and artificial intelligence

The protein STING (stimulator of interferon genes) is a central regulator of the innate immune system and plays an important role in antitumor immunity by inducing the production of cytokines such as type I interferon (IFN). Activation of STING stems from the selective recognition of endogenous cyclic dinucleotides (CDNs) by the large, polar, and flexible binding site, thus posing challenges to the design of small molecule agonists with drug-like physicochemical properties. In this work we present the design of SNX281, a small molecule STING agonist that functions through a unique self-dimerizing mechanism in the STING binding site, where the ligand dimer approximates the size and shape of a cyclic dinucleotide while maintaining drug-like small molecule properties. SNX281 exhibits systemic exposure, STING-mediated cytokine release, strong induction of type I IFN, potent in vivo antitumor activity, durable immune memory, and single-dose tumor elimination in mouse models via a Cmax-driven pharmacologic response. Bespoke computational methods - a combination of quantum mechanics, molecular dynamics, binding free energy simulations, and artificial intelligence - were developed during the course of the project to design SNX281 by explicitly accounting for the unique self-dimerization mechanism and the large-scale conformational change of the STING protein upon activation. Over the course of the project, we explored millions of virtual molecules while synthesizing and testing only 208 molecules in the lab. This work highlights the value of a multifaceted computationally-driven approach anchored by methods tailored to address target-specific problems encountered along the project progression from initial hit to the clinic.

cancer biology↗

bioRxiv2022

Multifunctional flexible electro-optical arrays for simultaneous spatiotemporal cardiac mapping and modulation

Bioelectronic devices that allow simultaneous accurate monitoring and control of the spatiotemporal patterns of cardiac activity provide an effective means to understand the mechanisms and optimize therapeutic strategies for heart disease. Optogenetics is a promising technology for cardiac research due to its advantages such as cell-type selectivity and high space-time resolution, but its efficacy is limited by the insufficient number of modulation channels and lack of simultaneous spatiotemporal mapping capabilities in current cardiac optogenetics tools. Here we present soft implantable electro-optical cardiac devices integrating multilayered highly uniform arrays of transparent microelectrodes and multicolor micro-light-emitting-diodes in thin, flexible platforms for mechanically compliant high-content electrical mapping and single-/multi-site optogenetics and electrical stimulation without light-induced artifacts. Systematic benchtop characterizations, together with ex vivo and in vivo evaluations on healthy and diseased small animal and human hearts demonstrate their functionalities in real-time spatiotemporal mapping and control of cardiac rhythm and function, with broad applications in basic and ultimately clinical cardiology.

bioengineering↗

bioRxiv2022

Machine learning directed organoid morphogenesis uncovers an excitable system driving human axial elongation

The human embryo breaks symmetry to form the anterior-posterior axis of the body. As the embryo elongates along this axis, progenitors in the tailbud give rise to axial tissues that generate the spinal cord, skeleton, and musculature. The mechanisms underlying human axial elongation are unknown. While ethics necessitate in vitro studies, the variability of human organoid systems has hindered mechanistic insights. Here we developed a bioengineering and machine learning framework that optimizes symmetry breaking by tuning the spatial coupling between human pluripotent stem cell-derived organoids. This framework enabled the reproducible generation of hundreds of axially elongating organoids, each possessing a tailbud and an epithelial neural tube with a single lumen. We discovered that an excitable system composed of WNT and FGF signaling drives axial elongation through the induction of a signaling center in the form of neuromesodermal progenitor (NMP)-like cells. The ability of NMP-like cells to function as a signaling center and drive elongation is independent of their potency to generate mesodermal cell types. We further discovered that the instability of the underlying excitable system is suppressed by secreted WNT inhibitors of the secreted frizzled-related protein (SFRP) family. Absence of these inhibitors led to the formation of ectopic tailbuds and branches. Our results identify mechanisms governing stable human axial elongation to achieve robust morphogenesis.

developmental biology↗

bioRxiv2022

Learning inverse folding from millions of predicted structures

We consider the problem of predicting a protein sequence from its backbone atom coordinates. Machine learning approaches to this problem to date have been limited by the number of available experimentally determined protein structures. We augment training data by nearly three orders of magnitude by predicting structures for 12M protein sequences using AlphaFold2. Trained with this additional data, a sequence-to-sequence transformer with invariant geometric input processing layers achieves 51% native sequence recovery on structurally held-out backbones with 72% recovery for buried residues, an overall improvement of almost 10 percentage points over existing methods. The model generalizes to a variety of more complex tasks including design of protein complexes, partially masked structures, binding interfaces, and multiple states.

systems biology↗

bioRxiv2022

ViPal: A Framework for Virulence Prediction of Influenza Viruses with Prior Viral Knowledge Using Genomic Sequences

Influenza viruses pose significant threats to public health and cause enormous economic loss every year. Previous work has revealed the viral factors that influence the virulence of influenza viruses. However, taking prior viral knowledge represented by heterogeneous categorical and discrete information into account is scarce in the existing work. How to make full use of the preceding domain knowledge into virulence study is challenging but beneficial. This paper proposes a general framework named ViPal for virulence prediction that incorporates discrete prior viral mutation and reassortment information based on all eight influenza segments. The posterior regularization technique is leveraged to transform prior viral knowledge to constraint features and integrated into the machine learning models. Experimental results on influenza genomic datasets validate that our proposed framework can improve virulence prediction performance over baselines. The comparison between ViPal and other existing methods shows the computational efficiency of our framework with superior performance. Moreover, the interpretable analysis through SHAP identifies the scores of constraint features contributing to the prediction. We hope this framework could provide assistance for the accurate detection of influenza virulence and facilitate flu surveillance.

bioinformatics↗

bioRxiv2022

Saturated reconstruction of living brain tissue

Complex wiring between neurons underlies the information-processing network enabling all brain functions, including cognition and memory. For understanding how the network is structured, processes information, and changes over time, comprehensive visualization of the architecture of living brain tissue with its cellular and molecular components would open up major opportunities. However, electron microscopy (EM) provides nanometre-scale resolution required for full in-silico reconstruction1-5, yet is limited to fixed specimens and static representations. Light microscopy allows live observation, with super-resolution approaches6-12 facilitating nanoscale visualization, but comprehensive 3D-reconstruction of living brain tissue has been hindered by tissue photo-burden, photobleaching, insufficient 3D-resolution, and inadequate signal-to-noise ratio (SNR). Here we demonstrate saturated reconstruction of living brain tissue. We developed an integrated imaging and analysis technology, adapting stimulated emission depletion (STED) microscopy6,13 in extracellularly labelled tissue14 for high SNR and near-isotropic resolution. Centrally, a two-stage deep-learning approach leveraged previously obtained information on sample structure to drastically reduce photo-burden and enable automated volumetric reconstruction down to single synapse level. Live reconstruction provides unbiased analysis of tissue architecture across time in relation to functional activity and targeted activation, and contextual understanding of molecular labelling. This adoptable technology will facilitate novel insights into the dynamic functional architecture of living brain tissue.

neuroscience↗

bioRxiv2022

Zfp281 inhibits the pluripotent-to-totipotent state transition in mouse embryonic stem cells

The cell-fate transition between pluripotent and totipotent states determines embryonic development and the first cell-lineage segregation. However, limited by the scarcity of totipotent embryos, regulators on this transition remain largely elusive. A novel model to study the transition has been recently established, named the 2-cell-like (2C-like) model. The 2C-like cells are rare totipotent-like cells in the mouse embryonic stem cell (mESC) culture. Pluripotent mESCs can spontaneously transit into and out of the 2C-like state. We previously dissected the transcriptional roadmap of the transition. In this study, we revealed that Zfp281 is a novel regulator for the pluripotent-to-totipotent transition in mESCs. Zfp281 is a transcriptional factor involved in the cell-fate transition. Our study shows that Zfp281 represses transcripts upregulated during the 2C-like transition via Tet1 and consequentially inhibits mESCs from transiting into the 2C-like state. Interestingly, we found that the inhibitory effect of Zfp281 on the 2C-like transition leads to an impaired 2C-like-transition ability in primed-state mESCs. Altogether, our study reveals a novel mediator for the pluripotent-to-totipotent state transition in mESCs and provides insights into the dynamic transcriptional control of the transition.

cell biology↗

bioRxiv2022

Genome-wide CRISPR screens identify novel regulators of wild-type and mutant p53 stability

Tumour suppressor p53 (TP53) is the most frequently mutated gene in cancer. Several hotspot p53 mutants not only lose tumour suppressive capabilities, but also function in a dominant-negative manner, suppressing canonical wild-type p53 function. Furthermore, some hotspot p53 mutants promote oncogenesis by gain-of-function mechanisms. Levels of p53 are regulated predominantly through regulation of protein stability and while wild-type p53 is normally kept at very low levels at steady-state, p53 mutants are often stabilized in tumours, which may be vital for their oncogenic properties. Here, we systematically profiled the factors that regulate protein stability of wild-type and mutant p53 using marker-based genome-wide CRISPR screens. We found that most proteins that regulate wild-type p53 also regulate a subset of p53 mutants with the exception of p53 R337H regulators, which are largely private to this mutant. Mechanistically, we identified FBXO42 as a novel positive regulator of a subset of p53 mutants comprising R273H, R248Q and R248W. We show that FBXO42 acts together with CCDC6 to regulate USP28-mediated p53 stabilization. Our work also identifies C16orf72 as a negative regulator of the stability of wild-type p53 and of all p53 mutants tested. C16orf72 is amplified in breast cancer, and we show that C16orf72 regulates p53 levels in mammary epithelium of mice and its overexpression results in accelerated breast cancer with reduced p53 levels. Together, this work provides a network view of the processes that regulate p53 stability, which might provide clues for reinforcing wild-type p53 or targeting mutant p53 in cancer.

cancer biology↗