Is Tsix repression of Xist specific to mouse?

Because X-chromosome inactivation is a uniquely mammalian method of dosage compensation and much of the mammalian X chromosome has been conserved¹, the mechanisms underlying the single active X are probably the same for most mammals. But some features differ, even among tissues of an individual. These differences have to do with whether X inactivation is imprinted (the maternal X is always active) or random (either X can be active) and the stability of the inactive state. Underlying such variations are species differences in the physical map of the X inactivation center, temporal differences in the onset of developmental events and the role of tinkering in evolution of biological processes^2,3. Such changes usually eliminate or add elements that modify, but do not interfere with, the basic blueprint. Because X dosage compensation is an essential developmental program for mammalian cells, clearly these variations do not meddle with the basic mechanisms.

The gene Tsix represents an important species difference, perhaps related to the fact that X inactivation is imprinted in mouse but not human placenta⁴. In the mouse, Tsix transcripts are antisense to mouse Xist and inhibit expression of the maternal Xist allele in placental cells⁵ and of the future active X in embryonic stem cells^6,7. By repressing the accumulation of Xist transcripts, Tsix blocks the cascade of events that lead to transcriptional inactivation. On the sole basis of observations in mice, Tsix has been proposed to have an essential role in protecting the future active X from inactivation—not only when the X is maternal, but also when randomly chosen^8,9.

But the human version of the gene does not share this function^10,11. Human TSIX is antisense to XIST but carries a deletion of the CpG island, which was shown by Lee and colleagues to be essential for function of Tsix^7,9. There is evidence that CpG islands, like the one missing from TSIX, are needed for imprinting^12,13, consistent with lack of imprinted X inactivation in human placenta. Most important, TSIX transcripts are ineffectual; they do not repress XIST in cis and, in fact, are co-expressed with XIST from the inactive X throughout human embryonic development¹¹. Despite being homozygous with respect to this TSIX mutation, human females undergo random X inactivation. Therefore, TSIX can not be essential for this function in our species. In addition, X inactivation in bovine placenta is imprinted¹⁴, but evolutionary changes in the region 3′ to the bovine Xist gene¹⁵ suggest that bovine Tsix may also be defective. Conceivably, Tsix regulation of Xist is an exclusive feature of X inactivation in the mouse. After all, the evolution of laboratory mice has been subject not only to chance events but also to considerable artificial selection.

See Reply to "Is Tsix repression of Xist specific to mouse?" by Jeanie T. Lee