Screening And Functional Study For Splicing Mutation In Familial Hypercholesterolemia

Background: Familial hypercholesterolemia(FH,OMIM #143890)is a common and serious autosomal dominant monogenic disease,which is characterized by elevated total cholesterol and low density lipoprotein cholesterol levels in plasma,resulting in skin and tendon xanthomata,and premature cardiovascular disease.FH is a disease with highly genetic heterogeneity.Based on the pathogenic gene mapping strategy of pedigree samples,13 genes have been found to be associated with cholesterol metabolism disorders,including LDLR,apoB-100,PCSK9,ABCG5,and ABCG8.Traditional research suggests that the main pathogenic mechanism of FH is that mutations in the coding region of the causative gene cause structural or functional changes in the protein,such as mutations in the LDLR gene,causing absence of LDLR or structural dysfunction on the surface of the cell membrane,resulting in whole body tissues to utilize disorders of the low density lipoprotein in blood circulation and overdeposition in the tissues.Recent studies have suggested that the related mutations in the protein coding regions do not explain all the cause of FH.Therefore,more in-depth research is needed.RNA splicing is an important step in the transcription process of eukaryotes.It cuts off the non-coding portion(intron)of the gene sequence and connects the coding portion(exon)to form the precursor mRNA.Abnormal splicing can also cause changes in the structure and function of proteins,which is a common cause of disease.There are quite a few mutations located in the intron of non protein coding region and may cause disease by affecting the splicing pattern.It is currently speculated that about one-third of the variations in monogenic genetic disease cause abnormal splicing of mRNA.Therefore,aberrant splicing is another important pathogenesis of monogenic genetic disease and requires a systematic approach to screening for abnormal splicing.RNA splicing is widespread in eukaryotes.Facing huge and complex genomic data,only using traditional experimental methods can not meet the needs of theresearch of splicing sites.It is urgent to request a bioinformatics method that can precisely predict splice sites to guide the smooth progress of the experiment.The prediction methods of commonly used splice sites are MaxEntScan,NNSPLICE and NetGene2.However,these methods have not yet reached consensus on the interpretation of the results,and the cut-off value for the consistent interpretation of the splice site score has not been determined.In addition,there is a limit to the length of the sequence of analysis.In this study,we used gkm-SVM to screen the splicing mutation of FH gene and evaluate whether these mutations would affect splicing.Screening of the splicing mutation of FH related pathogenic genes through bioinformatics theory will greatly reduce the scope of the experiment and provide guidance for the exploration of the splicing mechanism.Objective:1.Establish a scheme of splicing mutation prediction based on gkm-SVM.2.Systematic screening of FH related splicing mutations.3.Verify the effect of splicing mutation on FH at cell level.Methods:1.The real splice site and false splice site sequences were obtained from the HS3 D database,and new data set is formed by intercepting the sequences according to a certain length.The gkm-SVM model is trained using positive and negative training sets.Use the test set data to verify the correctness of the model.78 pathogenic splicing mutations were obtained from the HGMD professional database,and 53 benign polymorphisms were obtained from the ClinVar and 1000 genomes to evaluate the predictive effect of the gkm-SVM method on splicing mutation and compare with other methods.2.Use a R package called Shiny in R language,and user visual interface is built based on Shiny framework to facilitate users to predict splicing mutation.3.The splicing mutation near the splice site was screened for mutations of the existing FH pathogenic gene and the gkm-SVM model was used to predict the FH related splicing mutation.4.Minigene technique was used to verify the function of FH related splicing mutation: a mutant IVS7+2 G>A wild type and mutant plasmid of ABCG5 gene wasconstructed,HepG2 cells were transfected,and the total cell RNA was extracted,and RT-PCR analysis and sequencing were carried out.The splicing results were identified by electrophoresis and sequencing of RT-PCR products.Results:1.The gkm-SVM splice mutation prediction model was established.For the donor site,the recognition rate of positive sample and negative number was 95.9%and 89.9%,and the correlation coefficient was 0.8286.For the acceptor site,the recognition rate of the positive sample and the negative number were 86.6% and82.7%,respectively,and the correlation coefficient was 0.6653,indicating that the gkm-SVM model was reached a high sensitivity and specificity.The gkm-SVM model was evaluated and the results of ROC curve analysis showed that both the donor site and the acceptor site showed better predictive performance of the gkm-SVM method.2.Based on the Shiny framework,a visual interface for predicting splice mutations was successfully established to facilitate the user's analysis of the splicing mutations of interest and to assess whether mutations have an impact on splicing.3.Screening for FH related splicing mutations: There were 108 LDLR splicing mutations,and 14 splicing mutations in other FH related genes(apoB-100,PCSK9,ABCG5 and ABCG8).4.The results of agarose gel electrophoresis showed that the RT-PCR products of the wild type plasmids were normal,the band size was 295 bp,and the mutant was not the normal band,which was replaced by the band of 81 bp.After sequencing,it was confirmed that the target band of the mutant was 81 bp and the mature mRNA excised the entire exon 8,resulting in shortened RT-PCR products from mutant plasmids.Conclusion:1.The gkm-SVM model established in this study has achieved a high sensitivity and specificity.Whether it is a donor site or a acceptor site,gkm-SVM shows better prediction performance than other methods.2.According to the gkm-SVM prediction scheme,for the donor site,84.1% of the LDLR splicing mutation may affect the splicing,and for the receptor site,86.7%of the LDLR splicing mutation may affect the splicing.3.The minigene technique was used to verify that the abnormal splicing of the ABCG5 gene mutation IVS7+2 G>A resulted in the deletion of the exon 8,which led to the abnormal structure and function of the protein.