Characterization Of Plant Imprinted Genes At The Sequence And Gene Functional Network Levels

Imprinted genes are a class of genes showing parent-of-origin allelic expression that are subjected to epigenetic regulation.Gene imprinting is widely found in both the placentae of mammalian and endosperm of higher plants and plays critical roles in the regulation of normal growth and development of organisms.Abnormal allele-specific expression of imprinted genes can lead to some human diseases and has a great influence on the size and starch content of seed in Angiosperm plants.As so far,dozens to hundreds of imprinted genes have been identified in human,mouse,and different higher plants using high-throughput sequencing technology.However,except for a few imprinted genes that have been functionally characterized by experimental methods,the functions of most imprinted genes remain unknown,especially in plants.Besides,the functional relationship among imprinted genes remains unclear.More importantly,there has been no detailed report regarding the differences in sequence features and functional network property between imprinted genes and non-imprinted genes.Also,the evolutionary force driving the gain and loss of imprinted genes in different species is still unresolved.Thus,by analyzing imprinted genes at both sequence and gene functional network levels,we expected to understand the characteristics of imprinted genes and to provide some theoretical basis for the mechanism of genomic imprinting.In this paper,taking imprinted genes identified in eight plant species in genome-wide as the research objects,we analyzed their sequence features,such as gene length,CpG island,tandem repeat elements,and made several comparisons between imprinted and non-imprinted genes regarding sequence characteristics.We also explored the conservation of gene imprinting across plant species and investigated the evolution rate of imprinted genes.By analyzing the gene expression levels of imprinted genes across multiple tissues and conditions,we found a distinct expression pattern or expressional tissue specificity between imprinted genes and non-imprinted genes.By building the Arabidopsis functional gene network,we analyzed the network properties of imprinted genes and revealed closed functional relationships among imprinted genes.The main research contents and results are as follows.(1)By performing sequence analysis on plant imprinted genes,we found that the imprinted genes in Arabidopsis,rice,tomato,maize and caster bean have longer gene length than that of the non-imprinted genes,especially for paternally expressed genes(PEGs).We further observed that the average length of the exon of imprinted genes in Arabidopsis lyrata,Capsella rubella and Solanum lycopersicum is longer than that of the non-imprinted genes.For introns,the sequence length of Arabidopsis and maize imprinted genes is averagely longer than that of the non-imprinted genes.(2)Analyzing the tandem repeat elements(TREs)in plants imprinted genes revealed that the frequency of the TRE in both promoter and gene body of imprinted genes is significantly higher than that in non-imprinted genes in most of our investigated plant species,suggesting these tandem repeat elements potentially play important roles in mediating genomic imprinting.(3)Analyzing the cis-regulatory elements of imprinted genes showed that imprinted genes were preferentially regulated by transcription factors like AIB3,MYC4,MYC3,HBI1,HAT2,AP3,AGL13,TCP2 and TGA3,and several of them are involved in epigenetic regulation and thus may be involved in gene imprinting.(4)Gene family analysis revealed that imprinted genes were conserved across plant species at the sequence level,and less species-specific imprinted genes were observed.However,the gene imprinting was not conserved across plant species.For example,a gene is imprinted in one species,but it is not in other species.We also showed that imprinted genes mostly belong to multiple-copy genes,implying that gene imprinting may result from the neofunctionalization following gene duplications.Besides,we also observed that imprinted genes had a faster evolution rate compared with their paralogous genes.(5)Gene expression profiles showed that the expression specificity of MEGs was greater than PEGs,and the former tend to be expressed in the tissues like endosperm and seed,while the latter is more widely expressed.By analyzing the maize gene expression profile of endosperm,we found that imprinted genes had highly dynamic expression profiles during the early development of the seed.(6)The analysis of the Arabidopsis gene functional network showed that Arabidopsis imprinted genes tend to interact with each other and form a tightly connected sub-network in gene functional network.Specifically,PEGs have a higher frequency of functional interaction than MEGs.These results suggested that imprinted genes in plants are not functionally unrelated.Besides,compared with non-imprinted genes,imprinted genes have higher network degree and betweenness centralities in the functional network,indicating the regulatory importance of imprinted genes in the whole network.The importance of imprinted genes is also reflected by the closed functional relationship with essential genes and immune-related genes.Finally,we found the imprinted genes in Arabidopsis have similar network properties with that in several other species by orthologous gene mapping,although gene imprinting is not conserved across species.(7)Chromatin modification analysis on leaf tissues of Arabidopsis and rice showed that imprinted genes and non-imprinted genes have similar chromatin modification patterns,and it was feasible to predict gene expression level based on the combinatory signal of multiple chromatin modifications.Analysis of the signal profile of individual chromatin modification revealed that,except H3K27me1,there was no significant difference between the imprinted genes and non-imprinted genes regarding the relationship between chromatin modification intensity and gene expression level.In summary,the imprinted genes from eight Angiosperms were analyzed using bioinformatics methods in this study.The analysis included the sequence features,cis-regulatory elements,cross-species comparisons,gene functional network analysis,etc,which provided some theoretical basics for exploring the characteristics of imprinted genes and the mechanism of genomic imprinting in plants.