The Origin And Evolution Of Human Imprinting Clusters

BackgroundIn diploid organisms, the somatic cells have two copies of each gene, one is inherited from the father and the other from the mother. For most autosomal genes, both alleles possess the same expression capability. While a small fraction (less than 1%) of genes are imprinted in mammals, which means mono-allelic expression. Genomic imprinting is an epigenetic phenomenon independent of the classical Mendelian inheritance. In 1991, DeChiara and colleagues found the first imprinted gene—insulin-like growth factor type II (IGF2) in mice by using gene knock-out technology. With the development of molecular technology, more and more imprinted genes have been identified during 1990s and 2000s, and considerable studies reveal that these imprinted genes play key roles in many physiological processes.The majority of imprinted genes form clusters in the genome. Each imprinting cluster contains the regulatory elements called imprinting control regions, which are often located near the long noncoding RNA genes. LncRNAs are transcripts longer than 200 nucleotides and lack protein conding capacity. With our understanding of genome increases, we have changed the point of ncRNAs are "junk DNA" or "transctipt noise" to they are functional moleculars. There are cases that genetically modified mice and human patients accompanied by imprinting disorder demonstrate aberrant expression of lncRNAs, indicating a close relationship between genomic imprinting and lncRNAs.There are two principal models that explain imprinted gene expression. One is the insulator model, and the other is the lncRNA mediated silencing model. In the silencing model, lncRNAs can recruit DNA methyltransferases (such as Dnmt1) and Polycomb Repressive Complexes (such as PRC2), and direct them to specific genome loci. But, the details of how remain unclear. One possible mechanism is that lncRNAs can utilize the 3D structure of chromosomes and quickly locate themselves to target regions. The other is that lncRNAs can bind to DNA to form RNA:DNA triplexes with Hoogsteen or reverse Hoogsteen base pairing. Taking the lncRNA PARTICLE as an example, PARTICLE represses the tumor suppressor MAT2A via triplex formation at the upstream of a MAT2A promoter and interaction with a polycomb repressor complex.LncRNAs’ DNA binding sites have been widely concerned recently. If the lncRNAs’ binding sites can be identified, the target genes of many lncRNAs can be quickly recognized. Many lncRNAs involved in imprinted gene expression have been confirmed. In order to find out the origin and evolution of human imprinting clusters, we identified orthologues of 7 human lncRNAs that control human imprinting clusters in other species, analyzed features of these orthologous lncRNAs, and use LongTarget to predict the incRNAs’ DNA binding motifs (TFO) and binding sites (TTS).MethodsWe examined 7 well-characterised imprinting clusters in both human and mouse controlled by 7 lncRNAs. Since compensatory mutations may make lncRNA to have divergent sequences but conserved secondary structure, we used RNA sequence search program Infernal to identify orthologues of human lncRNAs in other 24 species. Meanwhile, we use GENEWISE to annotate protein coding genes in imprinting clusters whenever necessary. Finally, we used the program LongTarget to predict each lncRNAs’DNA binding motifs and binding sites in species covering multiple mammalian clades.ReslutsOrthologues of human H19 exist in early placental mammals. Interestingly, while H19 is highly conserved in simians, chimp H19 lacks exon2, and human H19 exon2 contains 2 specific simple repeats. The predicted H19 DNA binding motif is located in exon4 (-570-670bp), and its DNA binding sites are mainly distributed at the promoter of IGF2 and CpG islands in the imprinting cluster.Orthologues of human Aim are conserved only in simians, as it is not found in rodents, and only very short orthologous sequences are found in some other species. The experimentally identified Aim in mice has the same function as that in human. Comparing the transposon elements (TE) contained in human Aim and mouse Aim, we found that these TEs are very different. The predicted DNA binding motif in both human Aim and mouse Aim shows the same feature—A-rich, and their DNA binding sites are distributed equally at the promoter of slc22a2 and slc22a3.Orthologues of human Gnas-AS1 is conserved in primates. During evolution, it seems that its exon5 becomes conserved yet its exon8 is primate-specific. The predicted DNA binding motif in simians is located at the 5’end of exonl (～1-100bp).Highly conserved orthologues of human Kcnqlotl exists only in simians; in other species the length of orthologous sequence is shorter than the half of human Kcnqlot1. Human and chimp Kcnqlot1 have the same DNA binding motif, and their DNA binding sites are distributed at the promoter of about ten genes, as well as transposons and CpG islands in the Kcnql cluster.ConclusionLncRNAs that control human imprinting clusters exist only in placental mammals. Some exons of these lncRNAs (such as exon5 of Gnas-AS1) share conserved features, and some exons of these lncRNAs (such as exonl of H19) show lineage-specific features. Only in simians, do the predicted DNA binding motifs of these lncRNAs demonstrate same features (such as A-rich). These lncRNA’s DNA binding sites are mainly distributed at the promoter of imprinted genes, as well as transposons and CpG islands.