The Regulation of Genomic Imprinting and X Inactivation in Mice Marisa S. Bartolomei, Ph.D. Assistant Investigator, University of Pennsylvania School of Medicine Summary: Marisa Bartolomei is interested in the processes that control genomic imprinting and X inactivation in mice. Genomic imprinting results in the differential expression of the parental alleles of a gene. This parent-of-origin-specific effect is observed for a small number of genes in mammals and is likely to involve novel regulatory events. The mechanism by which parental origin affects this subset of genes remains largely unknown and is the focus of our research. To this end, we are using the maternally expressed mouse H19 gene as a model imprinted gene. The mouse H19 gene was one of the first imprinted genes to be identified. Although the H19 transcript is typical of protein-coding RNAs, it does not encode a protein, and the active gene product may be the RNA itself. The function of the H19 gene, which is predominantly expressed in mesodermal and endodermal derivatives in embryonic and neonatal tissues, is not, however, entirely clear. Interestingly, H19 is closely linked to the oppositely imprinted insulin-like growth factor II gene (Igf2). Evidence suggests that these genes are part of an imprinting domain in which a common set of enhancers drives expression of H19 on the maternal chromosome and Igf2 on the paternal chromosome. Our initial studies investigated how the two parental alleles of imprinted genes are distinguished so that the transcriptional machinery knows which allele to express. A 2-kb region located from ?2 kb to ?4 kb relative to the start of transcription (designated the imprinting control region [ICR] or differentially methylated domain [DMD]) is preferentially methylated on the repressed paternal allele throughout development. While hypermethylation of this region represents the best candidate for conferring the allele-specific imprinting mark, this region also exhibits maternal-specific DNase hypersensitivity in somatic cells. We have deleted the differentially methylated sequence at the endogenous locus and found that the imprinting and methylation of both the H19 and the Igf2 genes are disrupted, indicating that this region is required for the reciprocal imprinting of H19 and Igf2. We have also used conditional mutants to demonstrate that the ICR is required in the germline for the establishment of the imprinting of these two genes; in differentiated cells, however, the ICR is required for Igf2 imprinting but is no longer required for H19 imprinting. More recently, it has been shown in vitro that the ICR displays insulator activity that is sensitive to methylation. Insulator elements can isolate a gene and its regulatory elements from position effects and block enhancer-mediated transcription when placed between a gene and its enhancer. At the H19 locus, it is postulated that the ICR insulates H19 and its enhancers from Igf2 on the maternal allele, thereby causing the exclusive expression of H19. On the paternal allele, the ICR is methylated and inactivated, and Igf2 has access to the 3' enhancers. (This work has been supported by the National Institutes of Health.) To determine which part of the 2-kb ICR is essential to imprinting, we cloned and characterized the corresponding sequence from human and rat. Sequence comparison revealed a conserved 21-bp CpG-rich element that is repeated four times in mouse and rat and six times in human. The ubiquitous transcriptional regulator CTCF binds this sequence in a methylation-sensitive manner in vitro. CTCF also binds to the 5' chicken ß-globin insulator via a sequence that is similar to the H19 repeats. While in vitro experiments point to the repeat sequence as mediating the ICR insulator activity, the insulating capacity of these sequences has not been demonstrated in vivo. To understand the nature of the insulator, we have mutated the CpG dinucleotides within the CTCF sites at the endogenous locus. Although the mutant ICR retains insulating capacity, preliminary analysis indicates that the ICR does not become methylated in the male germline, resulting in mice harboring hypomethylated parental alleles. These results suggest that the H19 locus possesses distinct 5' sequences that permit insulator activity on the maternal allele and attract methylation that leads to repression on the paternal allele. While deletion of the ICR disrupts imprinted expression and methylation at the H19 locus, there is still a small amount of parental-specific methylation that remains on the mutant alleles. These results suggest that other cis-acting sequences may be crucial to imprinted expression. Such elements are likely located 3' to the transcription unit, since additional conserved sequences have been described in this region and deletion of a large segment of DNA harboring these sequences perturbs imprinting of H19 transgenes. We are performing gene targeting in embryonic stem cells and transgenic analysis to investigate which sequence is required for H19 and Igf2 imprinting. It is also essential to determine at what point during gametogenesis the parental-specific epigenetic mark is conferred. Since mature oocytes and sperm are differentially methylated, we have analyzed H19 methylation in earlier populations of gametes. Determination of the time and tissue in which the methylation pattern is established will ultimately allow us to dissect the mechanism by which the parental alleles are identified and subsequently marked. We have found that primordial germ cells in 13.5-day male embryos lack methylation on both parental alleles, indicating that during male germ cell development, the paternal-specific methylation inherited from sperm is erased. By embryonic day 15.5, the paternal H19 allele is methylated in prospermatogonia. In contrast, the maternal H19 allele acquires significant methylation after birth. These experiments indicate that the parental alleles of H19 remain distinct in the male germline, despite the fact that methylation differences are transiently erased. Thus an epigenetic modification such as differential chromatin structure may act in the absence of methylation to maintain parental identity. Once the embryo recovers its ability to methylate genomic DNA, allelic methylation differences may either resume a primary role in allelic identity or act with another epigenetic modification to confer parental identity. Experiments are in progress to assess the nature of such an allele-distinguishing epigenetic modification. We are also interested in the processes that govern X inactivation in mice. X inactivation is the dosage compensation mechanism that female mammals use to silence one X chromosome and to achieve X-linked expression equivalent to that of males. Certain aspects of this complex multistep process have been well established, but the molecular and genetic mechanisms controlling this process remain poorly characterized. While all factors known to be involved in X inactivation map to the X chromosome, it is probable that unidentified autosomal factors are essential to the process. To isolate such factors, we used N-ethyl-N-nitrosourea (ENU) mutagenesis in the mouse to select for mutations that affect X inactivation. In collaboration with Huntington Willard (Case Western Reserve University School of Medicine), we have recovered two independent autosomal-dominant mutations that perturb X-inactivation patterns. Affected heterozygous females exhibit alterations in the proportion of cells expressing a given X chromosome. The observation that 6.5-day embryos are affected by the mutations suggests that we have disrupted autosomal factors that act early in the X-inactivation pathway. Such factors may regulate the choice process of X inactivation. These results are the first evidence of an autosomal mutation affecting any component of the X-inactivation pathway. We have mapped one mutation to chromosome 15 and are refining its location, mapping the second mutation, and characterizing the phenotype of the mutant animals.