Sex chromosomes present a gene expression dosage problem for the organisms that carry them: if the male is XY and the female XX, the female has a double dose of X chromosome genes. This has been overcome in very different ways in different organisms: whereas Drosophila melanogaster doubles gene expression from the single X chromosome in males, vertebrates take a different approach, controlling gene expression by inactivating one of the two X chromosomes in the female. Noncoding RNAs have been implicated in the mechanism of dosage compensation in both cases, but although there are many potential cis-elements that may be involved in mammalian X chromosome inactivation, the order and mechanism by which they act are still being elucidated. To further complicate the issue, mammals have two modes of X chromosome inactivation. In the placenta, the paternal X chromosome is always inactivated, in so-called 'imprinted X inactivation', whereas in the embryo, one of the two X chromosomes is randomly inactivated. In recent work from Lee and colleagues (Dev. Cell 12, 57–71, 2007), a common layer of upstream regulation has been found to control X chromosome inactivation in mammals. The authors discovered that a repeat element regulates X chromosome inactivation in the context of both imprinted and random inactivation.

A single locus, Xic, is known to regulate both the 'counting' of X chromosomes and the subsequent silencing of one chromosome. Several factors encoded at this locus have been implicated in X inactivation in the past, including the noncoding RNA Xist, which is expressed from and coats the inactive X chromosome. A second important transcript is the antisense version of this RNA, Tsix, which is expressed from the active X and antagonizes Xist. A Tsix promoter has been defined, but the results of two targeted deletions made by Lee and colleagues now indicate that this region is dispensable for Tsix expression and X chromosome inactivation. As a larger deletion more strongly depletes Tsix expression, it seems that a different sequence is driving Tsix expression. The authors thus focused on the nearby DXPas34 repeat element and found that it is related in sequence to retrotransposon elements, suggesting its evolutionary origin in an endogenous retrovirus element (ERVL). Using a reporter construct, Lee and colleagues further show that DXPas34 acts as a bidirectional promoter that is active in embryonic stem cells but not in more differentiated cells (and thus is less active in increasingly differentiated cells, as expected of an element involved in X inactivation).

To test whether DXPas34 is required for Tsix expression (and thus usually antagonizes X inactivation), the authors generated a targeted deletion of this element on one X chromosome and found that Tsix expression is specifically lost from that chromosome. The chromosome carrying the DXPas34 deletion also has higher transcription of genes usually expressed from the inactive X, and is biased towards inactivation, as assessed by Xist co-localization; in contrast, the intact X chromosome (red spot) was not coated with Xist (larger green patch) in 80% of cases, suggesting that it is active. This phenotype supports the idea that the intact DXPas34 element antagonizes inactivation on the active X. Aberrant later expression of Tsix from the inactive X in the DXPas34 deletion strain suggests that the maintenance of silencing is also dependent on this repeat.

In addition to its involvement in random X inactivation, the authors also have evidence suggesting that imprinted silencing is affected by DXPas34. They found that deletion of this locus on the maternal X chromosome is embryonic lethal, whereas embryos lacking this locus on the paternal X are unaffected. Although the mechanism of DXPas34 action in this context is unclear, this result indicates disruption of imprinted X inactivation.

Overall, this work suggests that DXPas34 acts upstream to enhance Tsix expression and downstream to maintain Tsix silencing, and it also begins to reveal which previously implicated elements ultimately contribute to the regulation of Xist expression. Furthermore, the study indicates that imprinted and random X inactivation share a common upstream element in their regulation. This opens the door to further understanding inactivation through identification of the factors that regulate DXPas34 and the molecular mechanisms by which it acts to enhance and silence gene expression.