Research On Characteristics Of Mammalian Virus Receptors And Receptor Prediction Based On Bioinformatics Methods

Virus-receptor interaction is essential for the viral infection of host cells and the host specificity of the virus.However,only a few viral receptors have been identified,and the question of how viruses select receptors remains unanswered.Here,we characterized the mammalian virus receptors and built machine-learning methods for predicting human virus receptors using the bioinformatics methods.The main findings of this dissertation was listed as follows:(1)Systematic characterization of mammalian virus receptors.To understand the mechanism of receptor usage by viruses,we first manually curated a high-quality database named Viral Receptor which included 322 pairs of mammalian virus–host receptor interactions,142 unique viral species or sub-species and 150 virus receptors.Systematic analysis of these mammalian virus receptors showed that the mammalian viral protein receptors are structurally and functionally diverse,yet they had several common features when compared to other cell membrane proteins: more protein domains,higher level of Nglycosylation,higher ratio of self-interaction and more interaction partners in host proteinprotein interaction network(PPIN),and higher expression in most tissues of the host.Then we systematically characterized the mammalian virus protein receptors by the viral group,entry mode,binding mode and envelop,and found that the protein receptors of retrovirus in mammals had less protein domains,lower level of N-glycosylation,smaller node degrees and betweennesses in the host PPIN and lower expressions in common human tissues,than those of DNA and RNA viruses;the protein receptors of mammalian viruses which enter the host cell by endocytosis had more interaction partners in the host PPIN and higher expressions in common human tissues,than those of viruses which enter the host cells by membrane-fusion;the entry receptors of mammalian viruses had more interaction partners in the host PPIN and higher expressions in common human tissues,than the primary receptor;the protein receptors of non-enveloped mammalian viruses had more protein domains,higher level of N-glycosylation,larger node degrees and betweennesses in the host PPIN and higher expressions in common human tissues,than those of enveloped viruses.(2)Systematic comparison of viral receptors of mammalian viruses and phages.To explore whether the characteristics of mammalian viral receptor are universal in phage receptors,we first manually curated a high-quality database named phage Receptor which included 427 pairs of phage–host receptor interactions,341 unique viral species or subspecies and 207 virus receptors.Then we systematically compared viral receptors of bacteriophages and mammalian viruses from the perspectives of the location,structure,function and expression level.We found that the phage receptors not only located on the cell membrane of the host cell,but also on the outer membrane,cell wall and flagella,compared to mammalian viral receptors;although both of their viral receptors have more protein domains,they had no similar functions;mammalian virus and phage protein receptors have more interaction partners in host PPIN and higher expression levels than other proteins.This indicated that more partners and higher expression levels in host PPIN may be common features of viral receptors.(3)Systematic analysis of the relationship between mammalian virus receptors and the tissue tropism and host specificity of the mammalian viruses.The interaction between virus and protein receptors determines the tissue tropism,host range,and cross-species infection of the virus to a large extent.By comparing the expression levels of human virus receptors in major human tissues infected or non-infected by viruses,we found that tissues with higher expression levels of viral receptors are more likely to be infected by viruses;by comparing the sequence similarity between viral receptors and their homologous proteins in infected or non-infected mammals,we found that the sequence similarity between viral receptors and their homologous proteins can be used as an important feature for predicting the crossspecies infection of mammalian viruses.(4)Prediction of the receptorome for the human-infecting viruses based on computational modeling.A random-forest model was built to identify potential receptors for human viruses from human cell membrane proteins based on the feature of mammalian virus receptors and protein sequence features,and a total of 1424 human cell membrane proteins were predicted to be the potential receptors of human-infecting viruses.In addition,the combination of the receptorome of human-infecting virome with protein–protein interactions between human and viruses predicted in previous studies enabled further prediction of the receptors for 693 human-infecting viruses,such as the enterovirus,norovirus and West Nile virus.Finally,the candidate alternative receptors of the SARSCo V-2 were also predicted.In summary,this dissertation systematically characterized the mammalian viral receptors and analyzed the relationship between the viral receptors and tissue tropism and host specificity of virus,which helps understand the mechanism of virus-receptor interaction and virus-host interactions.A computational model was further built to the predict the receptorome of the human-infecting virome based on the features of mammalian virus receptors identified in this dissertation,which provides a basis for the identification and indepth study of the viral receptors.