Biology subjects

Yang, X.

Publications and source records attributed to Yang, X..

At least 739 records · Page 41Linked to original sources

bioRxiv2021

PDX1 directs a core developmentally and evolutionarily conserved gene program in the pancreatic islet

Pancreatic and duodenal homeobox 1 (PDX1) is crucial for pancreas organogenesis, yet the dynamic changes in PDX1 targets in mouse or human pancreas development have not been examined. We integrated the PDX1 cistrome with cell lineage-specific gene expression in both mouse and human developing pancreas. We identified a core set of developmentally and evolutionarily conserved PDX1 bound genes that reveal the broad multifaceted role of PDX1 in pancreas development. Despite the well-known, dramatic changes in PDX1 function and expression, we showed that PDX1 binding is largely stable from embryonic pancreas to adult islet. This may point towards a dual role of PDX1, activating or repressing the expression of its targets at different ages, dependent on other functionally-congruent or directly-interacting partners. Our work also suggests that PDX1 functions not only in initiating pancreas differentiation, but also as a potential keepsake of the progenitor program in the adult beta cells.

developmental biology↗

bioRxiv2021

Single-cell sequencing characterizes intratumoral profile of immune cells in old mice

It has long been thought that aging is a major risk factor for cancer incidence. However, accumulating evidence indicates increased resistance of old animals to tumor growth. A systematic understanding of how old individuals defend against tumor invasion is currently lacking. Here we investigated the differences of age-associated alterations in tumor-infiltrating immune cells between young and old mice using single-cell RNA analysis. Our results showed that a higher proportion of cytotoxic CD8+ T cells, nTregs, cDC, and M1-type macrophages, while a higher percentage of exhausted CD8+ T cells, iTregs, pDC, and M2-type macrophages were found in young mice. Importantly, TCR diversity analysis showed top 10 TCR clones consisted primarily of exhausted CD8+ T cells in young mice whereas tip clones were predominantly cytotoxic CD8+ T cells in old mice. Consistently, trajectory inference demonstrated that CD8+ T cells preferentially differentiated into cytotoxic cells in old mice in contrast to exhausted cells in young mice. Meanwhile, we confirmed the main distinctions between young and old mice by flow cytometry. Collectively, our data revealed that a significantly higher proportion of effector immune cells in old mice defend against tumor progression, providing a framework for the immunotherapy of elderly patients with tumors.

immunology↗

bioRxiv2021

The proximal proteome of 17 SARS-CoV-2 proteins links to disrupted antiviral signaling and host translation

Viral proteins localize within subcellular compartments to subvert host machinery and promote pathogenesis. To study SARS-CoV-2 biology, we generated an atlas of 2422 human proteins vicinal to 17 SARS-CoV-2 viral proteins using proximity proteomics. This identified viral proteins at specific intracellular locations, such as association of accessary proteins with intracellular membranes, and projected SARS-CoV-2 impacts on innate immune signaling, ER-Golgi transport, and protein translation. It identified viral protein adjacency to specific host proteins whose regulatory variants are linked to COVID-19 severity, including the TRIM4 interferon signaling regulator which was found proximal to the SARS-CoV-2 M protein. Viral NSP1 protein adjacency to the EIF3 complex was associated with inhibited host protein translation whereas ORF6 localization with MAVS was associated with inhibited RIG-I 2CARD-mediated IFNB1 promoter activation. Quantitative proteomics identified candidate host targets for the NSP5 protease, with specific functional cleavage sequences in host proteins CWC22 and FANCD2. This data resource identifies host factors proximal to viral proteins in living human cells and nominates pathogenic mechanisms employed by SARS-CoV-2. Author SummarySARS-CoV-2 is the latest pathogenic coronavirus to emerge as a public health threat. We create a database of proximal host proteins to 17 SARS-CoV-2 viral proteins. We validate that NSP1 is proximal to the EIF3 translation initiation complex and is a potent inhibitor of translation. We also identify ORF6 antagonism of RNA-mediate innate immune signaling. We produce a database of potential host targets of the viral protease NSP5, and create a fluorescence-based assay to screen cleavage of peptide sequences. We believe that this data will be useful for identifying roles for many of the uncharacterized SARS-CoV-2 proteins and provide insights into the pathogenicity of new or emerging coronaviruses.

microbiology↗

bioRxiv2021

BTBD9 is a novel component of IGF signaling and regulates manganese-induced dopaminergic dysfunction

Restless legs syndrome (RLS) is a common neurological disorder associated with iron deficiency and dopaminergic (DAergic) neuronal dysfunction. BTBD9 is a genetic risk factor for RLS. However, its molecular function remains largely unknown. Here, we report the interaction between BTBD9, manganese (Mn) and insulin/insulin-like growth factor (IGF) signaling in Caenorhabditis elegans, mouse Neuro2a cells and humans. We found that elevated Mn downregulated BTBD9 mRNA levels; in turn, BTBD9 expression attenuated Mn-induced cellular stress and dopaminergic neurodegeneration. As Mn is a known co-factor for insulin receptor and IGF-1 receptor, which activates IGF signaling, we posited that BTBD9 negatively regulates IGF signaling. Our results showed that the protective effects of BTBD9 against Mn toxicity were dependent on the forkhead box O (FOXO) protein. Furthermore, BTBD9 overexpression significantly elevated FOXO level and decreased PKB level, while phosphoinositide-dependent kinase-1 (PDK1) level remained unchanged. We conclude that BTBD9 acts as a key component in the IGF signaling pathway. Meanwhile, the roles of Mn in DAergic neurotoxicity and regulating BTBD9 shed new light on the etiology of RLS.

neuroscience↗

bioRxiv2021

Assembly of chromosome-scale and allele-aware autotetraploid genome of the Chinese alfalfa cultivar Zhongmu-4 and identification of SNP loci associated with 27 agronomic traits

Alfalfa (Medicago sativa L.), the most valuable perennial legume crop, referred to as "Queen of the Forages" for its high nutritional value and yield production among forage crops. Comprehensive genomic information of germplasm resources from different ecological regions and modern breeding strategies, such as molecular-marker assisted breeding are of great importance to breed new alfalfa varieties with environmental resilience. Here, we report assembly of the genome sequence of Zhongmu-4 (ZM-4), one of the most planted cultivars in China, and identification of SNPs associated with alfalfa agronomic traits by Genome-wide Association Studies (GWAS). Sequence of 32 allelic chromosomes was assembled successfully by single molecule real time sequencing and Hi-C technique with ALLHiC algorithm. About 2.74 Gbp contigs, accounting for 88.39% of the estimated genome, were assembled with 2.56 Gbp contigs anchored to 32 pseudo-chromosomes. In comparison with M. truncatula A17, distinctive inversion and translocation on chromosome 1, and between chromosome 4 and 8, respectively, were detected. Moreover, we conducted resequencing of 220 alfalfa accessions collected globally and performed GWAS analysis based on our assembled genome. Population structure analysis demonstrated that alfalfa has a complex genetic relationship among germplasm with different geographic origins. GWAS identified 101 SNPs associated with 27 out of 93 agronomic traits. The updated chromosome-scale and allele-aware genome sequence, coupled with the resequencing data of most global alfalfa germplasm, provides valuable information for alfalfa genetic research, and further analysis of major SNP loci will accelerate unravelling the molecular basis of important agronomic traits and facilitate genetic improvement of alfalfa.

plant biology↗

bioRxiv2021

Non-canonical role of Hippo tumor suppressor serine/threonine kinase 3 STK3 in prostate cancer

Serine/threonine kinase 3 (STK3) is an essential member of the highly conserved Hippo Tumor suppressor pathway which regulates Yes 1 Associated protein (YAP1) and TAZ. STK3 and its paralog STK4 initiate a phosphorylation cascade that regulate YAP1/TAZ activation and degradation, which is important for regulated cell growth and organ size. Deregulation of this pathway leads to hyper-activation of YAP1 in various cancers. Counter to the canonical tumor suppression role of STK3, we report that in the context of prostate cancer (PC), STK3 has a pro-tumorigenic role. Our investigation started with the observation that STK3, but not STK4, is frequently amplified in PC. A high STK3 expression is associated with decreased overall survival and positively correlates with androgen receptor (AR) activity in metastatic castrate resistant PC. XMU-MP-1, an STK3/4 inhibitor, slowed cell proliferation, spheroid growth and matrigel invasion in multiple models. Genetic depletion of STK3 decreased proliferation in several PC cell lines. In a syngeneic allograft model, STK3 loss slowed tumor growth kinetics in vivo and biochemical analysis suggest a mitotic growth arrest phenotype. To further probe the role of STK3 in PC, we identified and validated a new set of selective STK3 inhibitors, with enhanced kinase selectivity relative XMU-MP-1, that inhibited tumor spheroid growth and invasion. Consistent with the canonical role, inhibition of STK3 induced cardiomyocyte growth and had chemo-protective effects. Our results contend that STK3 has a non-canonical role in PC progression and inhibition of STK3 may have therapeutic potential for PC that merits further investigation. SignificanceOur findings illuminate a new actionable target for PC therapy that would traditionally be overlooked due to its canonical role as a tumor suppressor in other cancer types.

cancer biology↗

bioRxiv2021

Pharmacologic Therapy for Engraftment Arrhythmia Induced by Transplantation of Human Cardiomyocytes

BackgroundEngraftment arrhythmias (EAs) are observed in large animal studies of intramyocardial transplantation of human pluripotent stem cell-derived cardiomyocytes (hPSC-CMs) for myocardial infarction. Although transient, the risk posed by EA presents a barrier to clinical translation. ObjectivesWe hypothesized that clinically approved antiarrhythmic drugs can prevent EA-related mortality as well as suppress tachycardia and arrhythmia burden. MethodshPSC-CM were transplanted into the infarcted porcine heart by surgical or percutaneous delivery to induce EA. Following a screen of antiarrhythmic agents, a prospective study was conducted to determine the effectiveness of amiodarone plus ivabradine in preventing cardiac death and suppressing EA. ResultsEA was observed in all subjects, and amiodarone-ivabradine treatment was well-tolerated. None of the treated subjects experienced the primary endpoint of cardiac death, unstable EA or heart failure compared to 5/8 (62.5%) in the control cohort (hazard ratio 0.00; 95% confidence interval, 0-0.297; p = 0.002). Overall survival including two deaths in the treated cohort from immunosuppression-related infection showed borderline improvement with treatment (hazard ratio 0.21; 95% confidence interval, 0.03-1.01; p = 0.05). Without treatment, peak heart rate averaged 305 {+/-} 29 beats per min (bpm), whereas in treated subjects peak daily heart rate was significantly restricted to 185{+/-}9 bpm (p = 0.006). Similarly, treatment reduced peak daily EA burden from 96.8 {+/-} 2.9% to 76.5 {+/-} 7.9% (p = 0.003). Antiarrhythmic treatment was safely discontinued after approximately one-month of treatment without recrudescence of arrhythmia. ConclusionsThe risk of engraftment arrhythmia following hPSC-CM transplantation can be reduced significantly by combined amiodarone and ivabradine drug therapy. Condensed AbstractEngraftment arrhythmia (EA) is a transient but serious complication of cardiac remuscularization therapy. Using a porcine model of cardiac remuscularization and EA, ivabradine and amiodarone were independently effective in suppressing tachycardia and arrhythmia, respectively. Baseline amiodarone combined with adjunctive ivabradine successfully prevented cardiac death, unstable EA and heart failure (hazard ratio 0.00; 95% confidence interval, 0-0.297; p = 0.002) and significantly suppressed peak daily heart rate and arrhythmia burden (p=0.006 and 0.003, respectively). Antiarrhythmic treatment was successfully discontinued after one-month without recrudescence of arrhythmia. We conclude that EA can be suppressed by combined amiodarone and ivabradine drug therapy.

pharmacology and toxicology↗

bioRxiv2021

Transfer learning via multi-scale convolutional neural layers for human-virus protein-protein interaction prediction

To predict interactions between human and viral proteins, we combine evolutionary sequence profile features with a Siamese convolutional neural network (CNN) architecture and a multi-layer perceptron (MLP). Our architecture outperforms various feature encodings-based machine learning and state-of-the-art prediction methods. As our main contribution, we introduce two types of transfer learning methods (i.e., frozen type and fine-tuning type) that reliably predict interactions in a target human-virus domain based on training in a source human-virus domain, by retraining CNN layers. Our transfer learning strategies can effectively apply prior knowledge transfer from large source dataset/task to small target dataset/task to improve prediction performance. Finally, we utilize the frozen type of transfer learning to predict human-SARS-CoV-2 PPIs, indicating that our predictions are topologically and functionally similar to experimentally known interactions. Source code and datasets are available at https://github.com/XiaodiYangCAU/TransPPI/.

bioinformatics↗

bioRxiv2021

Rapid spread of a symbiotic virus in a major crop pest following wide-scale adoption of Bt-cotton in China

Bacillus thuringiensis (Bt) crops have been widely planted and the effects of Bt-crops on populations of the target and non-target insect pests were well studied. However, the effects of Bt-crops exposure on microorganisms that interact with crop pests havent previously been quantified. Here, we use laboratory and field data to show that infection of Helicoverpa armigera with a symbiotic densovirus (HaDV2) is associated with its enhanced growth and resistance to Bt-cotton. Moreover, field monitoring showed a much higher incidence of cotton bollworm infection with HaDV2 in regions cultivated with Bt-cotton than in regions without it, with the rate of densovirus infection increasing with increasing use of Bt-cotton. RNA-seq suggested resistance to both baculovirus and Cry1Ac were enhanced via the immune-related pathways. These suggest that the exposure to Bt-crops has selected for beneficial interactions between the target pest and a symbiotic microorganism that enhances its performance on Bt-crops under field conditions.

ecology↗

bioRxiv2021

Mitochondriomics reveals the underlying neuoprotective mechanism of TrkB receptor agonist R13 in the 5xFAD mice

Decreased energy metabolism and mitochondrial biogenesis defects are implicated in the pathogenesis of Alzheimers disease (AD). In present study, mitochondriomics analysis revealed significant effects of R13, a prodrug of 7,8-dihydroxyflavone, on mitochondrial protein expression profile, including the proteins related to the biological processes: fatty acid beta-oxidation, fatty acid metabolic process, mitochondrial electron transport, and mitochondrial respiratory chain. Cluster analysis of mitochondriomics demonstrated that R13 promoted mitochondrial oxidative phosphorylation (OXPHOS). The functional analysis showed that R13 increased ATP levels and enhanced OXPHOS including complex I, complex II, complex III and complex IV. R13 treatment increased mitochondrial biogenesis by regulating the levels of p-AMPK, p-CREB, PGC-1, NRF1 and TFAM as a consequence of activation of TrkB receptor in the 5xFAD mice. Finally, R13 significantly reduced the levels of tau phosphorylation and A{beta} plaque. Our data suggest that R13 may be used for treating AD via enhancing mitochondrial biogenesis and metabolism.

pharmacology and toxicology↗

bioRxiv2021

Neutralization of SARS-CoV-2 VOC 501Y.V2 by human antisera elicited by both inactivated BBIBP-CorV and recombinant dimeric RBD ZF2001 vaccines

Recently, the emerged and rapidly spreading SARS-CoV-2 variant of concern (VOC) 501Y.V2 with 10 amino acids in spike protein were found to escape host immunity induced by infection or vaccination. Global concerns have been raised for its potential to affect vaccine efficacy. Here, we evaluated the neutralization activities of two vaccines developed in China against 501Y.V2. One is licensed inactivated vaccine BBIBP-CorV and the other one is recombinant dimeric receptor-binding domain (RBD) vaccine ZF2001. Encouragingly, both vaccines largely preserved neutralizing titres, with slightly reduction, against 501Y.V2 authentic virus compare to their titres against both original SARS-CoV-2 and the currently circulating D614G virus. These data indicated that 501Y.V2 variant will not escape the immunity induced by vaccines targeting whole virus or RBD.

microbiology↗

bioRxiv2021

Myocardial BDNF regulates cardiac bioenergetics through the transcription factor Yin Yang 1

Circulating Brain-derived Neurotrophic Factor (BDNF) is markedly decreased in heart failure patients. Both BDNF and its receptor, Tropomyosin Related Kinase Receptor (TrkB), are expressed in cardiomyocytes, however the role of myocardial BDNF signaling in cardiac pathophysiology is poorly understood. We found that cardiac-specific TrkB knockout (cTrkB KO) mice displayed a blunted adaptive cardiac response to exercise, with attenuated upregulation of transcription factor networks controlling mitochondrial biogenesis/metabolism, including Peroxisome proliferator-activated receptor gamma coactivator 1 alpha (PGC-1). The cTrkB KO mice developed an exacerbated heart failure progression with transaortic constriction. The downregulation of PGC-1 in cTrkB KO mice exposed to exercise or TAC resulted in decreased cardiac energetics. We further unraveled that BDNF induces PGC-1 upregulation and bioenergetics through a novel signaling pathway, the pleiotropic transcription factor Yin Yang 1 (YY1). Taken together, our findings suggest that myocardial BDNF plays a critical role in regulating cellular energetics in the cardiac stress response.

molecular biology↗

bioRxiv2021

Spatial drivers and pre-cancer populations collaborate with the microenvironment in untreated and chemo-resistant pancreatic cancer

Pancreatic Ductal Adenocarcinoma (PDAC) is a lethal disease with limited treatment options and poor survival. We studied 73 samples from 21 patients (7 treatment-naive and 14 treated with neoadjuvant regimens), analyzing distinct spatial units and performing bulk proteogenomics, single cell sequencing, and cellular imaging. Spatial drivers, including mutant KRAS, SMAD4, and GNAQ, were associated with differential phosphosignaling and metabolic responses compared to wild type. Single cell subtyping discovered 12 of 21 tumors with mixed basal and classical features. Trefoil factor family members were upregulated in classical populations, while the basal populations showed enhanced expression of mesenchymal genes, including VIM and IGTB1. Acinar-ductal metaplasia (ADM) populations, present in 95% of patients, with 46% reduction of driver mutation fractions compared to tumor populations, exhibited suppressive and oncogenic features linked to morphologic states. We identified coordinated expression of TIGIT in exhausted and regulatory T cells and Nectin receptor expression in tumor cells. Higher expression of angiogenic and stress response genes in dendritic cells compared to tumor cells suggests they have a pro-tumorigenic role in remodeling the microenvironment. Treated samples contain a three-fold enrichment of inflammatory CAFs when compared to untreated samples, while other CAF subtypes remain similar. A subset of tumor and/or ADM-specific biomarkers showed differential expression between treatment groups, and several known drug targets displayed potential cross-cell type reactivities. This resolution that spatially defined single cell omics provides reveals the diversity of tumor and microenvironment populations in PDAC. Such understanding may lead to more optimal treatment regimens for patients with this devastating disease. HIGHLIGHTSO_LIAcinar-ductal metaplasia (ADM) cells represent a genetic and morphologic transition state between acinar and tumor cells. C_LIO_LIInflammatory cancer associated fibroblasts (iCAFs) are a major component of the PDAC TME and are significantly higher in treated samples C_LIO_LIReceptor-ligand analysis reveals tumor cell-TME interactions through NECTIN4-TIGIT C_LIO_LITumor and ADM cell proteogenomics differ between treated and untreated samples, with unique and shared potential drug targets C_LI

cancer biology↗

bioRxiv2020

A novel fermented Yi traditional medicine efficiently suppresses the replication of SARS-CoV-2 in vitro

Currently, the coronavirus disease 2019 (COVID-19) caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has a worldwide epidemic, causing more than 80 million infections and more than 1.7 million deaths. The pandemic has led to the closure of enterprises and schools in many countries, resulting in serious disruption of the global economy and social activities. Remdesivir is currently approved by the FDA for the treatment of COVID-19, but the WHO declared that Remdesivir is almost ineffective against COVID-19. The research and development of vaccines has made great progress, but it will take at least several months for safe and effective vaccines to be widely used clinically. Clinical studies revealed that some Traditional Chinese Medicines, such as Lianhua Qingwen Capsule and Huoxiang Zhengqi Water, exhibited excellent therapeutic effect on COVID-19. However, until now, there is still no cure for COVID-19. Therefore, there is an urgent need to find medicines that can effectively fight against the SARS-CoV-2. In this study, JIE BEN No. 10 (JB10), a fermentation broth produced by Yi traditional medicine fermentation technology, was explored for its anti-coronavirus activity. The in vitro data showed that JB10 could significantly suppresses the replication of the SARS-CoV-2 with an EC50 of 769.1 times dilution and a selection index of 42.68. Further studies indicated that JB10 had significant anti-inflammatory and antioxidant activities. The analysis of active components suggested that JB10 contained a large amount of superoxide dismutase (SOD), flavones, polyphenols, crude polysaccharide, etc. which may explain the anti-coronavirus activity, anti-inflammatory and antioxidant effects. Our study provides a new potentially therapeutic strategy for COVID-19.

microbiology↗

bioRxiv2020

Secondary rewards acquire enhanced incentive motivation via increasing anticipatory activity of the lateral orbitofrontal cortex

The motivation to strive for and consume primary rewards such as palatable food is bound by internal satiation and devaluation mechanisms, yet secondary rewards such as money may not be bound by these regulatory mechanisms. The present study therefore aimed at determining diverging devaluation trajectories for primary (chocolate milk) and secondary (money) reinforcers on the behavioral and neural level. Satiation procedures combined with a choice (Experiment 1) and an incentive delay (Experiment 2) paradigm consistently revealed decreased hedonic value for the primary reward as reflected by decreasing hedonic evaluation and choice preference, while hedonic value and preferences for the secondary reward increased. Concomitantly acquired functional near-infrared spectroscopy (fNIRS) data during the incentive delay paradigm revealed that increasing value of the secondary reward was accompanied by increasing anticipatory activation in the lateral orbitofrontal cortex, while during the consummatory phase the secondary reinforcer associated with higher medial orbitofrontal activity irrespective of devaluation stage. Overall, the findings suggest that - in contrast to primary reinforcers - secondary reinforcers can acquire progressively enhanced incentive motivation with repeated receipt, suggesting a mechanism which could promote escalating striving to obtain secondary rewards.

neuroscience↗

bioRxiv2020

An Ultrasensitive Biosensor for Quantifying the Interaction of SARS-CoV-2 and Its Receptor ACE2 in Cells and in vitro

The severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) is currently spreading and mutating with increasing speed worldwide. Therefore, there is an urgent need for a simple, sensitive, and high-throughput (HTP) assay to quantify virus-host interaction in order to quickly evaluate infectious ability of mutant virus and develop or validate virus-inhibiting drugs. Here we have developed an ultrasensitive bioluminescent biosensor to evaluate virus-cell interaction by quantifying the interaction between SARS-CoV-2 receptor binding domain (RBD) and its cellular receptor angiotensin-converting enzyme 2 (ACE2) both in living cells and in vitro. We have successfully used this novel biosensor to analyze SARS-CoV-2 RBD mutants, and evaluated candidate small molecules (SMs), antibodies, and peptides that may block RBD:ACE2 interaction. This simple, rapid and HTP biosensor tool will significantly expedite detection of viral mutants and anti-COVID-19 drug discovery processes.

biochemistry↗

bioRxiv2020

Age-related Impairment of Implant Osseointegration is Associated with Immune Activation and Suppression of Angiogenic, Notch, and Wnt Pathways

The number of total joint replacements (TJRs) in the United States is increasing annually. Cementless implants are intended to improve upon traditional cemented implants by allowing bone growth directly on the surface to improve implant longevity. One major complication of TJR is implant loosening, which is related to deficient osseointegration in cementless TJRs. Although poor osseointegration in aged patients is typically attributed to decreased basal bone mass, little is known about the molecular pathways that compromise the growth of bone onto porous titanium implants. To identify the pathways important for osseointegration that are compromised by aging, we developed an approach for transcriptomic profiling of peri-implant tissue in young and aged mice using our murine model of osseointegration. Based on previous findings of changes of bone quality associated with aging, we hypothesized that aged mice have impaired activation of bone anabolic pathways at the bone-implant interface. We found that pathways most significantly downregulated in aged mice relative to young mice are related to angiogenic, Notch and Wnt signaling. Downregulation of these pathways is associated with markedly increased expression of inflammatory and immune genes at the bone-implant interface in aged mice. These results identify osseointegration pathways affected by aging and suggest that an increased inflammatory response in aged mice may compromise peri-implant bone healing. Targeting the Notch and Wnt pathways, promoting angiogenesis, or modulating the immune response at the peri-implant site may enhance osseointegration and improve the outcome of joint replacement in older patients.

immunology↗

bioRxiv2020

Self-Supervised Representation Learning of Protein Tertiary Structures (PtsRep): Protein Engineering as A Case Study

In recent years, deep learning has been increasingly used to decipher the relationships among protein sequence, structure, and function. Thus far these applications of deep learning have been mostly based on primary sequence information, while the vast amount of tertiary structure information remains untapped. In this study, we devised a self-supervised representation learning framework (PtsRep) to extract the fundamental features of unlabeled protein tertiary structures deposited in the PDB, a total of 35,568 structures. The learned embeddings were challenged with two commonly recognized protein engineering tasks: the prediction of protein stability and prediction of the fluorescence brightness of green fluorescent protein (GFP) variants, with training datasets of 16,431 and 26,198 proteins or variants, respectively. On both tasks, PtsRep outperformed the two benchmark methods UniRep and TAPE-BERT, which were pre-trained on two much larger sets of data of 24 and 32 million protein sequences, respectively. Protein clustering analyses demonstrated that PtsRep can capture the structural signatures of proteins. Further testing of the GFP dataset revealed two important implications for protein engineering: (1) a reduced and experimentally manageable training dataset (20%, or 5,239 variants) yielded a satisfactory prediction performance for PtsRep, achieving a recall rate of 70% for the top 26 brightest variants with 795 variants in the testing dataset retrieved; (2) counter-intuitively, when only the bright variants were used for training, the performances of PtsRep and the benchmarks not only did not worsen but they actually slightly improved. This study provides a new avenue for learning and exploring general protein structural representations for protein engineering.

bioinformatics↗