Identification of an Xist RNA binding protein and a novel genetic element at the X inactivation center

In mammals, X chromosome inactivation (XCI) equalizes expression of X-linked genes between sexes, transcriptionally silencing an X chromosome in females. Inactivation involves both counting of X chromosomes and choice of which X to inactivate. During differentiation, counting ensures that in the diploid cell, all X chromosomes except for one are inactivated. The choice mechanism determines which X chromosome becomes inactivated. Choice, counting, and silencing are mediated by the noncoding RNA genes within the X inactivation center ( Xic ). Xist causes transcriptional silencing in cis, and is negatively regulated by the antisense gene, Tsix . The proper execution of silencing, counting, and choice are fundamentally dependent upon Xist and Tsix function, and their reciprocal relationship. The counting mechanism and the regulation of Xist and Tsix are not well defined. To test the hypothesis that the 3' end of Xist is a regulatory locus, we generated a deletion of the 3' end of Xist . Our findings indicate that the region has an effect on choice, but not on counting. We then demonstrated that the 3' region functions as an insulator. We propose a model where the boundary between Xist and Tsix organizes two exclusive chromatin structures, one that allows Tsix to predominate, and another that favors Xist and results in inactivaton. In our second study, we addressed the mechanism of Xist silencing. We performed a computational search for conserved and repeated motifs within the Xist RNA, which we considered more likely to have a conserved function. Repeat A, the domain required for silencing, was our top candidate. To find proteins that might assist Xist in its silencing function, we performed an affinity purification of proteins that bind to Repeat A. We found that the adenosine to inosine (A to I) editing protein ADAR1 was specifically purified with Xist Repeat A. We did not find evidence of extensive editing of either Xist or Tsix , but we did find two specific sites and one mRNA that appear to be edited at low levels. Our finding opens a new area of investigation regarding a role for editing or an alternative function of ADAR1 in XCI.