Biology subjects

Zhong, F.

Publications and source records attributed to Zhong, F..

At least 19 recordsLinked to original sources

bioRxiv2025

From Black Box to Biological Insight: AttentioFuse Unlocks Multi-Omics Dynamics in Lung Cancer

Lung adenocarcinoma (LUAD) and squamous cell carcinoma (LUSC), the major subtypes of non-small cell lung cancer (NSCLC), exhibit distinct molecular landscapes that demand precision in prognosis and therapy. While deep learning models achieve high predictive accuracy, their "black-box" nature limits clinical translation. To address this, we propose AttentioFuse, an interpretable deep learning framework employing a Reactome-guided mid-fusion strategy for multi-omics integration. AttentioFuse innovates through three pillars: (1) dual-phase learning to preserve omics-specific patterns via independent sub-networks, (2) hierarchical attention mechanisms (cross-omics, feature-level, and fusion-layer) to dynamically quantify layer contributions, and (3) integrated explainability combining DeepSHAP and global attention weights for gene-to-pathway interpretation. Evaluated on TCGA LUAD/LUSC cohorts, AttentioFuse matches state-of-the-art performance in TNM staging, while uncovering actionable biological insights. The framework validates pan-NSCLC mechanisms like AKT/mTOR metabolic control and histology-divergent Notch signaling roles, while revealing novel pathways--developmental reactivation (T-stage), microbiota-driven metastasis (M-stage), and ECM remodeling--providing testable hypotheses for progression and personalized therapy. Crucially, AttentioFuse bridges computational predictions to clinical practice by proposing molecularly-guided combination therapies. This paradigm shifts toward interpretable-aware AI advances oncology by transforming black-box predictions into biologically grounded decision-support tools.

bioinformatics↗

bioRxiv2025

Fast in vivo deep-tissue 3D imaging with selective-illumination NIR-II light-field microscopy and aberration-corrected implicit neural representation

Near-infrared II (NIR-II) microscopy, which enables in vivo deep-tissue visualization of vasculature and cell activities, has been a promising tool for understanding physiological mechanisms. However, the volumetric image speed of the NIR-II microscopy is hindered by scanning strategy, causing limitations for observing instantaneous biological dynamics in 3D space. Here, we developed a NIR-II light-field microscopy (LFM) based on selective-illumination and self-supervised implicit neural representation (INR)-reconstruction, which allows ultra-fast 3D in vivo imaging (100 volumes/s). Through integrating INR with view-wise aberration correction, we could conquer the artifacts induced by the angular subsampling and refractive index variation, achieving single-cell resolution at a volume of 550 m diameter and 200 m thickness. The volumetric selective-illumination overcomes the influence of out-of-focus background, together with the low scattering advantage of NIR-II wavelength, extending the imaging depth to 600 m. The developed aberration-corrected implicit neural representation reconstruction (AIR) NIR-II LFM showcases its capability by monitoring hemodynamics of mouse brain under norepinephrine and flow redistribution of ischemic stroke in 3D vasoganglion, as well as noninvasively tracking immune cell activities inside subcutaneous solid tumor through intact skin. This approach represents a significant advancement in 3D in vivo imaging, holding great potential in biomedical research and preclinical studies.

neuroscience↗

bioRxiv2025

Excavating Podocyte-Protective Targets of phloroglucinol-terpene hybrids Utilizing AI Models and Omics Analysis

Lupus nephritis (LN) is a multifactorial autoimmune inflammatory disorder, with podocyte dysfunction or structural abnormalities playing a pivotal role in its pathogenesis. Eucalyptus leaves, traditionally employed in folk medicine, have demonstrated notable efficacy in managing non-infectious inflammations. Eucarbwenstols D (R14) is a novel phloroglucinol-terpene that has been isolated from Eucalyptus robusta. To further refine potential targets, we utilized an AI scoring system and successfully identified the epidermal growth factor receptor (EGFR) as the novel target of R14. This discovery was built upon the foundation of transcriptome and connectivity mapping (CMap) analysis. Molecular docking and molecular dynamics simulations were employed to analyze the interaction between R14 and EGFR, surface plasmon resonance (SPR) and homogeneous time-resolved fluorescence (HTRF) assays were utilized to accurately quantify the binding affinity between R14 and EGFR. Western blot (WB) and immunofluorescence assays were then conducted to explore the downstream signaling pathways initiated by EGFR, specifically the PI3K/AKT/mTOR pathway. These assays demonstrated that R14 modulates the expression of apoptotic proteins Bax and Bcl-2 through this pathway, thereby exerting a protective effect on podocytes from undergoing apoptosis. The modulation of key podocyte proteins, nephrin and podocin, further substantiates this protective effect. Overall, we have applied a novel AI algorithm model that transcends bioinformatics analysis to precise identification of the novel podocyte target of R14. In this study, R14 was identified for the first time as a novel scaffold compound for EGFR inhibition. Moreover, R14 has demonstrated its capacity to inhibit podocyte apoptosis, thereby protecting these critical cells. This discovery establishes a robust theoretical foundation for the use of phloroglucinol-terpene hybrids (PTHs) in the treatment of LN.

cell biology↗

bioRxiv2024

Uncovering Latent Biological Function Associations through Gene Set Embeddings

The complexity of biological systems has increasingly been unraveled through computational methods, with biological network analysis now focusing on the construction and exploration of well-defined interaction networks. Traditional graph-theoretical approaches have been instrumental in mapping key biological processes using high-confidence interaction data. However, these methods often struggle with incomplete or/and heterogeneous datasets. In this study, we extend beyond conventional bipartite models by integrating attribute-driven knowledge from the Molecular Signatures Database (MSigDB) using the node2vec algorithm. Our approach explores unsupervised biological relationships and uncovers potential associations between genes and biological terms through network connectivity analysis. By embedding both human and mouse data into a shared vector space, we validate our findings cross-species, further strengthening the robustness of our method. This integrative framework reveals both expected and novel biological insights, offering a comprehensive perspective that complements traditional biological network analysis and paves the way for deeper understanding of complex biological processes and diseases.

bioinformatics↗

bioRxiv2024

Proposing a methodology for axon-centric analysis of IOP-induced mechanical insult

PurposeIOP-induced mechanical insult on retinal ganglion cell axons within the optic nerve head (ONH) is believed to be a key factor in axonal damage and glaucoma. However, most studies focus on tissue-level mechanical deformations, overlooking that axons are long and thin, and that their susceptibility to damage likely depends on the insults type (e.g. stretch/compression) and orientation (longitudinal/transverse). We propose an axon-centric approach to quantify IOP-induced mechanical insult from an axon perspective. MethodsWe used optical coherence tomography (OCT) scans from a healthy monkey eye along with histological images of cryosections to reconstruct the axon-occupied volume including detailed lamina cribrosa (LC) pores. Tissue-level strains were determined experimentally using digital volume correlation from OCT scans at baseline and elevated IOPs, then transformed into axonal strains using axon paths estimated by a fluid mechanics simulation. ResultsAxons in the LC and post-LC regions predominantly experienced longitudinal compression and transverse stretch, whereas those in the pre-LC and ONH rim mainly suffered longitudinal stretch and transverse compression. No clear patterns were observed for tissue-level strains. ConclusionsOur approach allowed discerning axonal longitudinal and transverse mechanical insults, which are likely associated with different mechanisms of axonal damage. The technique also enabled quantifying insult along individual axon paths, providing a novel link relating the retinal nerve fiber layer and the optic nerve through the LC via individual axons. This is a promising approach to establish a clearer connection between IOP-induced insult and glaucoma. Further studies should evaluate a larger cohort.

bioengineering↗

bioRxiv2024

Impact of elevated IOP on lamina cribrosa oxygenation; A combined experimental-computational study on monkeys

PurposeOur goal is to evaluate how lamina cribrosa (LC) oxygenation is affected by the tissue distortions resulting from elevated IOP. DesignExperimental study on monkeys SubjectsFour healthy monkey eyes with OCT scans with IOP of 10 to 50 mmHg, and then with histological sections of LC. MethodsSince in-vivo LC oxygenation measurement is not yet possible, we used 3D eye-specific numerical models of the LC vasculature which we subjected to experimentally-derived tissue deformations. We reconstructed 3D models of the LC vessel networks of 4 healthy monkey eyes from histological sections. We also obtained in-vivo IOP-induced tissue deformations from a healthy monkey using OCT images and digital volume correlation analysis techniques. The extent that LC vessels distort under a given OCT-derived tissue strain remains unknown. We therefore evaluated two biomechanics-based mapping techniques: cross-sectional and isotropic. The hemodynamics and oxygenations of the four vessel networks were simulated for deformations at several IOPs up to 60mmHg. The results were used to determine the effects of IOP on LC oxygen supply, assorting the extent of tissue mild and severe hypoxia. Main Outcome MeasuresIOP-induced deformation, vasculature structure, blood supply, and oxygen supply for LC region ResultIOP-induced deformations reduced LC oxygenation significantly. More than 20% of LC tissue suffered from mild hypoxia when IOP reached 30 mmHg. Extreme IOP(>50mmHg) led to large severe hypoxia regions (>30%) in the isotropic mapping cases. ConclusionOur models predicted that moderately elevated IOP can lead to mild hypoxia in a substantial part of the LC, which, if sustained chronically, may contribute to neural tissue damage. For extreme IOP elevations, severe hypoxia was predicted, which would potentially cause more immediate damage. Our findings suggest that despite the remarkable LC vascular robustness, IOP-induced distortions can potentially contribute to glaucomatous neuropathy.

bioengineering↗

bioRxiv2023

Enhancing Recognition and Interpretation of Functional Phenotypic Sequences through Fine-Tuning Pre-Trained Genomic Models

Decoding high-quality human genomic sequences requires comprehensive analysis of DNA sequence functionality. Through computational and experimental approaches, researchers study the genotype-phenotype relationship and generate important datasets that help unravel complicated genetic blueprints. This study explores the use of deep learning, particularly pre-trained models like DNA_bert_6 and human_gpt2-v1, in interpreting and representing human genome sequences. We meticulously construct multiple datasets linking genotypes and phenotypes to fine-tune pre-trained models for precise DNA sequence classification. Furthermore, we specifically focused on the human endogenous retrovirus (HERV) dataset with commendable classification performance (both binary and multi-classification accuracy and F1 values above 0.935 and 0.888, respectively). We evaluate the influence of sequence length on classification results and analyze the impact of feature extraction in the models hidden layers using the HERV dataset. To further understand the phenotype-specific patterns learned by the model, we perform enrichment, pathogenicity and conservation analyzes of specific motifs in the HERV sequence with high average local representation weight (LRAW) scores. Overall, the generated datasets further provide numerous additional genotype-phenotype datasets for evaluating the performance of genomic models. The findings highlight the potential of large models in learning DNA sequence representations, particularly when utilizing the HERV dataset, and provide valuable insights for future research. This work represents an innovative strategy that combines pre-trained model representations with classical omics methods for analyzing the functionality of genome sequences, fostering cross-fertilization between genomics and advanced AI. The source code and data are available at https://github.com/GeorgeBGM/Genome_Fine-Tuning.

bioinformatics↗

bioRxiv2023

EquiScore: A generic protein-ligand interaction scoring method integrating physical prior knowledge with data augmentation modeling

Developing robust methods for evaluating protein-ligand interactions has been a long-standing problem. Here, we propose a novel approach called EquiScore, which utilizes an equivariant heterogeneous graph neural network to integrate physical prior knowledge and characterize protein-ligand interactions in equivariant geometric space. To improve generalization performance, we constructed a dataset called PDBscreen and designed multiple data augmentation strategies suitable for training scoring methods. We also analyzed potential risks of data leakage in commonly used data-driven modeling processes and proposed a more stringent redundancy removal scheme to alleviate this problem. On two large external test sets, EquiScore outperformed 21 methods across a range of screening performance metrics, and this performance was insensitive to binding pose generation methods. EquiScore also showed good performance on the activity ranking task of a series of structural analogs, indicating its potential to guide lead compound optimization. Finally, we investigated different levels of interpretability of EquiScore, which may provide more insights into structure-based drug design.

bioinformatics↗

bioRxiv2023

Direct measurements of collagen fiber recruitment in the posterior pole of the eye

Collagen is the main load-bearing component of the peripapillary sclera (PPS) and lamina cribrosa (LC) in the eye. Whilst it has been shown that uncrimping and recruitment of the PPS and LC collagen fibers underlies the macro-scale nonlinear stiffening of both tissues with increased intraocular pressure (IOP), the uncrimping and recruitment as a function of local stretch have not been directly measured. This knowledge is crucial for the development of constitutive models associating micro and macro scales. In this project we measured local stretch-induced collagen fiber bundle uncrimping and recruitment curves of the PPS and LC. Thin coronal samples of PPS and LC of sheep eyes were mounted and stretched biaxially quasi-statically using a custom system. At each step, we imaged the PPS and LC with instant polarized light microscopy and quantified pixel-level (1.5 m/pixel) collagen fiber orientations. We used digital image correlation to measure the local stretch and quantified collagen crimp by the circular standard deviation of fiber orientations, or waviness. Local stretch-recruitment curves of PPS and LC approximated sigmoid functions. PPS recruited more fibers than the LC at the low levels of stretch. At 10% stretch the curves crossed with 75% bundles recruited. The PPS had higher uncrimping rate and waviness remaining after recruitment than the LC: 0.9{degrees} vs. 0.6{degrees} and 3.1{degrees} vs. 2.7{degrees}. Altogether our findings support describing fiber recruitment of both PPS and LC with sigmoid curves, with the PPS recruiting faster and at lower stretch than the LC, consistent with a stiffer tissue. Statement of SignificancePeripapillary sclera (PPS) and lamina cribrosa (LC) collagen recruitment behaviors are central to the nonlinear mechanical behavior of the posterior pole of the eye. How PPS and LC collagen fibers recruit under stretch is crucial to develop constitutive models of the tissues but remains unclear. We used image-based stretch testing to characterize PPS and LC collagen fiber bundle recruitment under local stretch. We found that fiber-level stretch-recruitment curves of PPS and LC approximated sigmoid functions. PPS recruited more fibers at a low stretch, but at 10% bundle stretch the two curves crossed with 75% bundles recruited. We also found that PPS and LC fibers had different uncrimping rates and non-zero wavinesss when recruited.

bioengineering↗

bioRxiv2021

Ultra-high-speed multi-parametric photoacoustic microscopy

Multi-parametric photoacoustic microscopy (PAM) is uniquely capable of simultaneous, high-resolution mapping of blood hemoglobin concentration, oxygenation, and flow in vivo. However, its speed has been limited by the dense sampling required for blood flow quantification. To overcome this limitation, we have developed an ultra-high-speed multi-parametric PAM system, which enables simultaneous acquisition of ~500 densely sampled B-scans by superposing the rapid laser scanning across the line-shaped focus of a cylindrically focused ultrasonic transducer over the conventional mechanical scan of the optical-acoustic dual foci. A novel optical-acoustic combiner is designed and implemented to accommodate the short working distance of the transducer, enabling convenient confocal alignment of the dual foci in the reflection mode. This new system enables continuous monitoring of microvascular hemoglobin concentration, blood oxygenation, and flow over a 4.5 x 3 mm2 area in the awake mouse brain with high spatial and temporal resolution (6.9 m and 0.3 Hz, respectively).

bioengineering↗

bioRxiv2021

A Parameter-free Deep Embedded Clustering Method for Single-cell RNA-seq Data

Clustering analysis is widely utilized in single-cell RNA-sequencing (scRNA-seq) data to discover cell heterogeneity and cell states. While many clustering methods have been developed for scRNA-seq analysis, most of these methods require to provide the number of clusters. However, it is not easy to know the exact number of cell types in advance, and experienced determination is not always reliable. Here, we have developed ADClust, an automatic deep embedding clustering method for scRNA-seq data, which can accurately cluster cells without requiring a predefined number of clusters. Specifically, ADClust first obtains low-dimensional representation through pre-trained autoencoder, and uses the representations to cluster cells into initial micro-clusters. The clusters are then compared in between by a statistical test, and similar micro-clusters are merged into larger clusters. According to the clustering, cell representations are updated so that each cell will be pulled toward centres of its assigned cluster and similar clusters, while cells are separated to keep distances between clusters. This is accomplished through jointly optimizing the carefully designed clustering and autoencoder loss functions. This merging process continues until convergence. ADClust was tested on eleven real scRNA-seq datasets, and shown to outperform existing methods in terms of both clustering performance and the accuracy on the number of the determined clusters. More importantly, our model provides high speed and scalability for large datasets.

bioinformatics↗

bioRxiv2021

Structural basis of main proteases of coronavirus bound to drug candidate PF-07321332

The high mutation rate of COVID-19 and the prevalence of multiple variants strongly support the need for pharmacological options to complement vaccine strategies. One region that appears highly conserved among different genus of coronaviruses is the substrate binding site of the main protease (Mpro or 3CLpro), making it an attractive target for the development of broad-spectrum drugs for multiple coronaviruses. PF-07321332 developed by Pfizer is the first orally administered inhibitor targeting the main protease of SARS-CoV-2, which also has shown potency against other coronaviruses. Here we report three crystal structures of main protease of SARS-CoV-2, SARS-CoV and MERS-CoV bound to the inhibitor PF-07321332. The structures reveal a ligand-binding site that is conserved among SARS-CoV-2, SARS-CoV and MERS-CoV, providing insights into the mechanism of inhibition of viral replication. The long and narrow cavity in the cleft between domains I and II of main protease harbors multiple inhibitor binding sites, where PF-07321332 occupies subsites S1, S2 and S4 and appears more restricted compared with other inhibitors. A detailed analysis of these structures illuminated key structural determinants essential for inhibition and elucidated the binding mode of action of main proteases from different coronaviruses. Given the importance of main protease for the treatment of SARS-CoV-2 infection, insights derived from this study should accelerate the design of safer and more effective antivirals.

biophysics↗

bioRxiv2021

M-CAMP™: A cloud-based web platform with a novel approach for species-level classification of 16S rRNA microbiome sequences

The M-CAMP (Microbiome Computational Analysis for Multiomic Profiling) Cloud Platform was designed to provide users with an easy-to-use web interface to access best in class microbiome analysis tools. This interface allows bench scientists to conduct bioinformatic analysis on their samples and then download publication-ready graphics and reports. The core pipeline of the platform is the 16S-seq taxonomic classification algorithm which provides species-level classification of Illumina 16s sequencing. This algorithm uses a novel approach combining alignment and kmer based taxonomic classification methodologies to produce a highly accurate and comprehensive profile. Additionally, a comprehensive proprietary database combining reference sequences from multiple sources was curated and contains 18056 unique V3-V4 sequences covering 11527 species. The M-CAMP 16S taxonomic classification algorithm was validated on 52 sequencing samples from both public and in-house standard sample mixtures with known fractions. Compared to current popular public classification algorithms, our classification algorithm provides the most accurate species-level classification of 16S rRNA sequencing data.

bioinformatics↗

bioRxiv2021

A high-accuracy and high-efficiency digital volume correlation method to characterize in-vivo optic nerve head biomechanics from optical coherence tomography

In-vivo optic nerve head (ONH) biomechanics characterization is emerging as a promising way to study eye physiology and pathology. We propose a high-accuracy and high-efficiency digital volume correlation (DVC) method for the purpose of characterizing the in-vivo ONH deformation from volumes acquired by optical coherence tomography (OCT). Using a combination of synthetic tests and analysis of OCTs from monkey ONHs subjected to acute and chronically elevated intraocular pressure, we demonstrate that our proposed methodology overcomes several challenges for conventional DVC methods. First, it accounts for large ONH rigid body motion in the OCT volumes which could otherwise lead to analysis failure; second, sub-voxel-accuracy displacement can be guaranteed despite high noise and low image contrast of some OCT volumes; third, computational efficiency is greatly improved, such that the memory consumption of our method is substantially lower than with conventional methods; fourth, we introduce a parameter measuring displacements confidence. Test of image noise effects showed that the proposed DVC method had displacement errors smaller than 0.028 voxels with speckle noise and smaller than 0.037 voxels with Gaussian noise; The absolute (relative) strain errors in the three directions were lower than 0.0018 (4%) with speckle noise and than 0.0045 (8%) with Gaussian noise. Compared with conventional DVC methods, the proposed DVC method had substantially improved overall displacement and strain errors under large body motions (lower by up to 70%), with 75% lower computation times, while saving about 30% memory. The study thus demonstrates the potential of the proposed technique to investigate ONH biomechanics.

bioengineering↗

bioRxiv2020

Inferring intestinal mucosal immune cell associated microbiome species and microbiota-derived metabolites in inflammatory bowel disease

Inflammatory bowel disease (IBD) is a complex, chronic inflammatory disease of the gastrointestinal tract with subtypes Crohns disease (CD) and ulcerative colitis (UC). While evidence indicates IBD is characterized by alterations in the composition and abundance of the intestinal microbiome, the challenge remains to specify bacterial species and their metabolites associated with IBD pathogenesis. By the integration of microbiome multi-omics data and computational methods, we provide analyses and methods for the first time to identify microbiome species and their metabolites that are associated with the human intestine mucosal immune response in patients with CD and UC at a systems level. First, we identified seven gut bacterial species and seventeen metabolites that are significantly associated with Th17 cellular differentiation and immunity in patients with active CD by comparing with those obtained in inactive CD and non-IBD controls. The seven species are Ruminococcus gnavus, Escherichia coli, Lachnospiraceae bacterium, Clostridium hathewayi, Bacteroides faecis, Bacteroides vulgatus, and Akkermansia muciniphila, and a few associated metabolites include the secondary bile acid lithocholate and three short-chain fatty acids (SCFAs): propionate, butyrate, and caproate. We next systematically characterized potential mechanistic relationships between the Th17-involved metabolites and bacterial species and further performed differential abundance analysis for both microbiome species and their metabolites in CD and UC relative to non-IBD controls with their metagenomic and metabolomic data. Based on the deconvolution of immune cell compositions from host intestinal bulk RNA-seq, we investigated changes in immune cell composition and abundance in CD and UC in comparison to non-IBD controls. Finally, we further extended our species and metabolite associations with immune cells from Th17 and Th2 cells to B cells, plasma B cells, plasmablasts, CD4+ T cells, and CD8+ T cells. While a set of associations of immune cells with bacterial species and metabolites was supported by published evidence, the new findings in this work will help to furthering our understanding of immune responses and pathogenesis in IBD.

bioinformatics↗

bioRxiv2020

Early Postnatal Tobacco Smoke Exposure Aggravates Experimental Autoimmune Encephalomyelitis in Adult Rats

Although substantial evidence supports smoking as a risk factor for the development of multiple sclerosis (MS) in adulthood, it remains controversial as to whether early-life exposure to environmental tobacco smoke (ETS) increases the risk of MS later in life. Here, using experimental autoimmune encephalomyelitis (EAE) as an animal model for MS, we show that exposing neonatal rats during the 1st week (ETS1-EAE), but not the 2nd week (ETS2-EAE) and the 3rd week (ETS3-EAE) after birth, increased the severity of EAE in adulthood in comparison to pups exposed to filtered compressed air (AIR-EAE). The EST1-EAE rats showed a worse neurological deficit score and a significant increase in CD4+ cell infiltration, demyelination, and axonal injury in the spinal cord compared to AIR-EAE, ETS2-EAE, and ETS3-EAE groups. Flow cytometry analysis showed that the ETS1 group had decreased numbers of regulatory T (Treg) cells and increased effector T (Teff) cells in the brain and spinal cord. The expressions of Treg upstream regulator Foxp3 and downstream cytokines such as IL-10 were also altered accordingly. Together, these findings demonstrate that neonatal ETS exposure suppresses Treg functions and aggravates the severity of EAE, confirming early-life exposure to EST as a potential risk factor for multiple sclerosis in adulthood.

neuroscience↗

bioRxiv2020

Crystal structure of SARS-CoV-2 main protease in complex with a Chinese herb inhibitor shikonin

Main protease (Mpro, also known as 3CLpro) has a major role in the replication of coronavirus life cycle and is one of the most important drug targets for anticoronavirus agents. Here we report the crystal structure of main protease of SARS-CoV-2 bound to a previously identified Chinese herb inhibitor shikonin at 2.45 angstrom resolution. Although the structure revealed here shares similar overall structure with other published structures, there are several key differences which highlight potential features that could be exploited. The catalytic dyad His41-Cys145 undergoes dramatic conformational changes, and the structure reveals an unusual arrangement of oxyanion loop stabilized by the substrate. Binding to shikonin and binding of covalent inhibitors show different binding modes, suggesting a diversity in inhibitor binding. As we learn more about different binding modes and their structure-function relationships, it is probable that we can design more effective and specific drugs with high potency that can serve as effect SARS-CoV-2 anti-viral agents.

biophysics↗

bioRxiv2020

Structure of SARS-CoV-2 main protease in the apo state reveals the inactive conformation

Mpro is of considerable interest as a drug target in the treatment of COVID-19 since the proteolytic activity of this viral protease is essential for viral replication. Here we report the first insight of the structure Mpro for SARS-CoV-2 in the inactive conformation under conditions close to the physiological state (pH 7.5) to an overall resolution of 1.9 [A]. The comparisons of Mpro in different states reveal that substrate binding site and the active site are more flexible in the inactive conformation than that in the active conformations. Notably, compared with the active conformation of the apo state structure in pH7.6 of SARS, the SARS-CoV-2 apo state is in the inactive conformation under condition close to physiological state (pH7.5). Two water molecules are present in the oxyanion hole in our apo state structure, whereas in the ligand-bound structure, water molecular is absence in the same region. This structure provides novel and important insights that have broad implications for understanding the structural basis underlying enzyme activity, and can facilitate rational, structure-based, approaches for the design of specific SARS-CoV-2 ligands as new therapeutic agents.

biochemistry↗