Gene Expression And Alternative Splicing Analysis Of Three Different Hepatic Cells

Liver, the largest organ of the human body, plays important roles in metabolismof a variety of substances, such as lipids, carbohydrates, protein and vitamins.Besides as the center of human metabolism, liver is also responsible fordetoxification, immunity, glycogen storage, synthesis of bile, hematopoietic and soon. So liver is called the busiest organ. The complex functions of liver depend on thecomposition of various types of liver cells. Liver cells are grouped into twocategories: hepatocyte and non-parenchymal cells. The former is the main cell type,accounting for about80%of liver volume and playing the majority of liver functions.Non-parenchymal cells include sinusoidal endothelial cells, Kupffer cells, hepaticstellate cells and other immune cells. Currently, most of studies about liver cells tendto focus on one particular type of cell. Comprehensive analysis of transcriptome andproteome profiles based on the liver cells can contribute to deep understanding ofthe mechanism for liver function.In the post-genomic era, transcriptome and proteome have become the popularobjects of research. The former conveys genetic information and the latter performsa vast array of functions. The protein coverage degree to proteome of one sampleoften refers to the size of coding genes, so we try to assist in deeper coverage ofliver proteome by the next generation sequencing technology. RNA-Seq, one of thenext generation high-throughout sequence technologies, has unparalleled advantagesin terms of deep coverage, high sensitivity and high accuracy. RNA-Seq couldidentify new alternative splice isoforms, RNA-editing sites and some new geneticregions and has been a powerful tool for fast and comprehensive profiling oftranscriptome.Firstly, we constructed liver transcriptome profile in the cellular level for thefirst time. C57BL/6mice were used to isolate HC, LSEC and KC by flow cytometry sorting, ensuring the purity and activity of various types of cells above95%and85%. We performed RNA-Seq on cDNA library built from3types of cells andcompleted qualitative and quantitative analysis after a series of data processingincluding low-quality filter, reads mapping, expression quantification and transcriptreconstruction.8802,10789and10125genes were identified in HC, LSEC and KCat the FDR<5%. Abundance correlation analysis showed the highest correlationcoefficient between HC and liver tissue, reflecting the fact of that HC counts for themajority type in liver cells. This is currently the most scale of mouse livertranscriptome data in cellular levels, providing a comprehensive reference resource.Secondly, functional annotation of liver cells was systematically analyzedbased on transcriptome expression profiles.1) Expressed genes in each cell wereclassed into low, medium and high expression groups. Functional enrichmentanalysis for each cells indicated genes involved in a particular function areexpressed at similar abundances.2) Differently expressed genes of each cell couldbetter characterize the biological features of the respective cells.3) Cell-specificgenes and cell-common genes differed from abundance distribution and functionalenrichment. Cell-specific genes preferred to low abundance distribution, membraneand cell surface location and was enriched in small molecular metabolism.Additionally, pathway enrichment analysis for cell-specific genes suggested somebiological processing involved cooperation between multiple types of cells.4)Unsupervised clustering for gene expression in each cell revealed that different celltypes had significantly different gene expression patterns. In short, systematicalfunction analysis deepens our understanding of distinct liver cells.Thirdly, we focused on the identification of alternative splicing in differenttypes of liver cells. In all reconstructed transcripts from RNA-Seq data, weadditionally identify over20%of new transcripts, corresponding to different splicingtypes. In the3types of cells, exon skipping is the most widespread splicing type.Identification and verification of new transcripts enhanced the corresponding genesannotation. Importantly, we firstly discovered that differences between differentliver cells presented the switching-like manner, indicating different cells showedpreferences for different splicing isoforms of one gene to achieve their ownbiological functions. In addition, we also attempted to explain the splicing differences between cells by splicing regulator proteins.Lastly, we employed MS and RNA-Seq in parallel into mouse liver tissue andcaptured a considerable catalogue of both transcripts and proteins. We thendeveloped a bioinformatics workflow for building a customized protein databasethat for the first time included new splicing-derived peptides andRNA-editing-caused peptide variants, allowing us to more completely identifyprotein isoforms. Using this experimentally determined database, we totallyidentified150peptides not present in standard biological databases at falsediscovery rate (FDR) of <1%, corresponding to72novel splicing isoforms,43newgenetic regions and15RNA-editing sites. Of these,11randomly selected novelevents passed experimental verification by PCR and Sanger sequencing. Newdiscoveries of gene products with high confidence in two omics levels demonstratedthe robustness and effectiveness of our approach and its potential application intodeep coverage of proteome profiles of liver cells.