To achieve dosage compensation, female mammals silence one of their two sex chromosomes in a process called X chromosome inactivation (XCI). In mice two forms of XCI exist: In 2-4 cell staged embryos, imprinted XCI (iXCI) results in exclusive silencing of the paternal X (Xp). Around implantation, female epiblast cells that give rise to somatic tissues reactivate the silenced Xp and undergo random XCI (rXCI), which inactivates the Xp or the maternal X (Xm) with equal probability. Both forms of XCI require the long non-coding Xist RNA, which coats the inactive X chromosome (Xi) from which it is transcribed1,2.

The X-linked gene Rnf12 encodes the RING finger protein Rnf12/RLIM, a transcriptional coregulator3. Rnf12/RLIM has been identified as a ubiquitin ligase4 that regulates activities of different classes of transcription factors5,6 in part by adjusting cellular levels of various nuclear cofactors via the ubiquitin/proteasome system (UPS)4,7. In female mice, sex-specific parent-of-origin effects have been reported upon Rnf12 deletion. Deletion of the maternal Rnf12 allele in oocytes results in early embryonic lethality specifically in female embryos due to a failure to develop extraembryonic trophoblast tissues in the placenta8. In contrast, in pregnant and lactating adult females Rnf12/RLIM expressed from the paternal allele serves as a survival factor specifically for milk-producing alveolar cells in mammary glands9. Thus, Rnf12 functions as a sex-specific epigenetic regulator in female mouse nurturing tissues.

In an embryonic stem cell (ESC) model, Rnf12/RLIM has been identified as a major regulator for the initiation of rXCI. Indeed, forced overexpression of Rnf12/RLIM triggers ectopic Xist clouds in male and female ESCs10, and homozygous deletion of Rnf12 resulted in a failure to initiate rXCI11. In mice, the early lethality of female embryos upon deletion of the maternal Rnf12 allele was shown to be the consequence of defective initiation of iXCI8. Strong evidence has been provided that Rnf12/RLIM exerts its functions for XCI initiation via promoting Xist expression8,10,11 in a dose-dependent manner10. Intriguingly, cellular Rnf12/RLIM levels increase at the onset of rXCI10,11, likely due to a release from transcriptional repression by pluripotency factors12. Together, these studies indicate crucial functions of Rnf12/RLIM for the initiation of XCI via activation of Xist.

However, these studies left unanswered the question of how Rnf12/RLIM promotes the activation of Xist. A recent study has addressed this point13. Gontan et al. used affinity purification followed by mass spectrometry to identify the transcription factor REX1 as a major interaction partner of Flag-tagged Rnf12/RLIM in ESC nuclei. In these cells, REX1 appears to repress genes that promote differentiation14 and its expression correlates with the pluripotent ESC state15. In an elegant series of experiments, the authors showed that endogenous Rnf12/RLIM interacts with, polyubiquitinates, and targets REX1 for degradation via the UPS, thereby dynamically adjusting levels of REX1 in various cell types including ESCs. In agreement with these results, the upregulation of Rnf12/RLIM upon differentiation of ESCs was accompanied with a downregulation of REX1. While Rnf12/RLIM appears to be a major regulator of REX1 during differentiation, the authors also observed downregulation of REX1 in ESCs lacking Rnf12/RLIM, indicating that other proteins contribute to the reduction of REX1 levels upon ESC differentiation. Interestingly, the authors showed that the RING finger of RLIM, which mediates its ubiquitin ligase activity, plays important roles in the Rnf12/RLIM-induced activation of ectopic XCI in male ESCs.

To identify roles of REX1 in XCI, day-3-differentiated male ESCs expressing lower REX1 levels (+/− for Rex1) or female ESCs overexpressing REX1 were examined for XCI as measured by the occurrence of Xist clouds coating an X chromosome. Indeed, whereas a significant number of the male Rex1+/− ESCs underwent ectopic XCI, REX1-overexpressing female ESCs experienced severe inhibition of XCI. Importantly, knockdown of Rex1 in Rnf12−/− ESCs resulted in a significant rescue of XCI. These results reveal functions of REX1 downstream of Rnf12, directly linking the functions of both proteins during XCI and the formation of Xist clouds.

Using chromatin immunoprecipitation in undifferentiated ESCs, Gontan et al.13 detected REX1 occupancy of promoter/regulatory elements of the Xist gene as well as those of Tsix, a gene that overlaps with the Xist locus on the X chromosome but is transcribed in the antisense direction, thereby inhibiting Xist transcription2. Indeed, consensus recognition sequences for REX1 binding were found on regulatory elements of both genes, suggesting that this transcription factor directly regulates the expression of these genes. To address this possibility, they performed transient cotranfections and detected a downregulation of Xist in cells overexpressing REX1, similar to cells that lack Rnf12/RLIM. These results indicate that REX1 functions as a repressor of Xist, thereby inhibiting XCI. To explain their findings that REX1 binds to regulatory sequences of Tsix and represses XCI in ESCs, the authors suggest that REX1 may function by activating Tsix expression. Based on these results, the scenario for female ESC differentiation and XCI would be as follows: Differentiation-induced downregulation of pluripotency factors trigger the upregulation of Rnf12 transcription and an increase in cellular Rnf12/RLIM levels. Lower cellular levels of REX1 due to its increased proteasomal targeting by Rnf12/RLIM would then lead to downregulation of Tsix and activation of Xist transcription followed by initiation of XCI. Due to the fact that similar to Xist, Rex1 is found only in placental mammals but not in marsupials that also lack Xist-mediated XCI, the authors propose that Xist and Rex1 may have co-evolved with the appearance of an Xist-dependent XCI mechanism.

While this report identifies important mechanisms of Rnf12/RLIM functions during XCI activation, it also raises several questions. Going forward, one of the most pressing question concerns the relevance of the Rnf12-Rex1 regulatory module for XCI in vivo as seemingly conflicting results have been reported for the requirement of Rnf12/RLIM for the initiation of rXCI. In an ES cell model Rnf12/RLIM is essential for initiation of rXCI11 and the results obtained by Gontan et al.13 are largely based on the same model. In contrast, evidence suggests that Rnf12/RLIM may be dispensable for rXCI in mice8 and Rex1−/− mice are viable with no apparent phenotype16. Thus, it will be important to address the functional requirement of Rnf12/RLIM in conjunction with REX1 for rXCI in vivo and how they connect to the choice of which X chromosome to inactivate. Moreover, it will be interesting to examine whether the crucial functions of Rnf12/RLIM during iXCI are also connected to Rex1 and whether targeted degradation of REX1 is the only role of Rnf12/RLIM in XCI. The further deciphering of the roles of Rnf12 and Rex1 in XCI promises to reveal new and exciting mechanisms of epigenetic regulation.