"Much less attention has been paid to the question of X/autosome dosage," says Avner. "X inactivation would be expected to lead to halved quantities of X-linked gene products compared to autosomal gene products. Haploinsufficiency for the entire X would a priori be expected to be catastrophic to the organism and lead to lethality during early embryonic development." Haploinsufficiency was the issue that Gupta and colleagues set out to address. "There was a lack of evidence for a germline dosage compensation machinery," notes Gupta, citing studies showing that in Drosophila the X chromosomes are not coated with MSL complexes or hyperacetylated on H4 K16 in male germ cells [5] (see the 'Behind the scenes' box for more on the rationale for the work). "Also, Parisi et al. [6] showed that a subset of ribosomal protein encoding genes are equally expressed in both testis and ovaries and we had seen that X;AA and XX;AA tumors showed very similar gene expression profiles," adds Gupta. "I've been thinking about this problem since I was a graduate student in the late 80s," recalls Brian Oliver who heads the research group at the National Institute of Diabetes and Digestive and Kidney Diseases in Bethesda, USA. "Until microarrays appeared, we didn't really see a good way to do a convincing experiment."

Not content with their impressive results in flies, Oliver's group went on to reanalyze microarray data from Caenorhabditis elegans and mice. Their results led them to a similar conclusion: that the single X chromosomes of X;AA nematodes and mice were expressed at similar levels to two autosomes. Di Nguyen and Christine Disteche, at the University of Washington, Seattle, USA, recently reported a similar study of gene expression in a range of human and mouse somatic tissues [7]. They also found that doubling of the global expression level of the X chromosome leads to dosage compensation in mammalian somatic tissues. "Interestingly, X-chromosome expression levels appear even more markedly increased in the brain," comments Avner. "Dare one suggest that the X chromosome has become specialized in cognitive function and sex? Also, although some X-linked genes are expressed in post-meiotic cells of the male and female germlines, upregulation is absent, allowing X/A ratios to be maintained in an equivalent fashion to somatic tissues."

Gupta thinks that studying the germline X-chromosome gene expression in other organisms (such as mice and C. elegans) using a similar approach might bring new insights. Oliver is very keen to get a handle on the mechanism. "We'll need to probe chromatin structure in the germline by ChIP-chip [chromatin immunoprecipitation] type methods," he says. He cites recent studies investigating how many of the X-chromosome genes might be regulated by the MSL complex [8,9]. "There may be other players besides the MSLs in the soma," says Oliver. "This supports our observation of MSL-independent dosage compensation in the germline." But he is also keeping an open mind about dosage compensation. He is interested in how the apparent moderate dosage compensation on the autosomes works and what this can tell us about general properties of gene expression networks. "And what is the best reference for measuring dosage compensation?" asks Oliver provocatively. "It is still possible, as long suggested by Jim Birchler, that autosomal expression is down in X;AA individuals. 2X = AA or X = (AA)/2?" Whether it is indeed the expression on the X chromosome that goes up or expression of the autosomes that goes down, the mysteries of the X will continue to fascinate both sexes for many years to come.