Predicted Controlling Element Motifs Using Conserved Sequences in Mammalian Imprinted Gene Clusters Reveal Evolutionary Differences

There are around 150 genes proven to be imprinted in mouse and/or human genomes, but the number of imprinted genes with corroborated status in other mammalian species is low. Most of these genes are found in a few clusters, sharing mechanisms of imprinting expression control. We carried out sequence comparative analysis of five main clusters containing 55 imprinted genes, using the genome sequences of human, mouse, cattle, horse, dog and rat, as well as partial sequences from other mammals, and those of chicken and lizard, as outgroups. We found high degrees of conservation in gene order and size in all the mammalian genomes, as well as in chicken and lizard, expect for a few genes, mostly of retrotransposon origin. We also found a large number of short, highly conserved segments in intergenic and intronic regions, some of them being repetitive elements (especially L1 and L2 LINEs) conserved in most mammals, but with lower conservation in mouse, rat and rabbit. We predicted some region-specific motifs that were found throughout each region. Long, non-coding and micro-RNAs are found in all clusters. Although most miRNAs are only annotated in the human and mouse genomes, they show high degree of conservation in all mammals and are expected to be expressed in these other mammals, as potentially useful for gene prediction in unannotated genomes based on this comparative approach. Phylogenetic analyses of the expressed sequences and the short conserved segments show differences in the pattern of evolution, compared to the most accepted mammalian phylogenetic tree.