Sex complicates many aspects of life but is sufficiently beneficial that virtually every eukaryotic species uses it. An especially sticky consequence is that an unequal complement of sex-determining chromosomes exists between the two sexes. One sex, typically female, carries two X chromosomes, and the other sex, usually male, has a single X chromosome (accompanied by a Y chromosome in some, but not all, species). This inequality puts the sex chromosomes in an evolutionarily vulnerable position.

Two hypotheses propose different outcomes for sex-linked genes that are preferentially expressed in one sex over the other. Rice's hypothesis1 states that, because males carry a single X chromosome, any recessive allele arising on the X that gives males a reproductive advantage is immediately available for positive selection. Conversely, any negative effects of that allele in females are masked by the presence of a second X chromosome. Thus, genes expressed preferentially in males are more likely to be present on the X chromosome. An alternate hypothesis states that the X should be feminized because the X chromosome is present in females 50% more frequently than in males, providing evolution with more opportunity to act on genes benefiting females. Thus, genes with female-biased expression should reside on the X chromosome. So which hypothesis is correct, or are the two hypotheses mutually compatible? In this issue, Pavel Khil and colleagues2 look at sex-biased expression profiles in mice and show that, as for many relationships, the answer involves a compromise.

On again, off again

Khil et al. adopted a computational approach using publicly available mouse expressed-sequence tag and microarray databases to define genes expressed preferentially in various somatic and gonadal tissues. They then mapped these genes onto the 20 mouse chromosomes and looked for biases in location. Genes expressed in various somatic tissues, such as liver and lung, showed no significant bias, but genes expressed preferentially in gonadal tissues told a different story. In particular, genes expressed in female tissues such as ovary and placenta were located on the X chromosome more frequently than expected, whereas genes expressed in testes showed the opposite trend. On the surface, this result supports feminization of the X. Complicating this interpretation, however, a previous study by Wang et al.3 reported the opposite result: genes expressed specifically in male germ cells were X-linked much more frequently than expected, supporting Rice's hypothesis.

What could account for these conflicting observations? The key issue, it seems, is timing. Wang et al. focused on male germ cells in the earliest stage of spermatogenesis, before meiosis, when the cells are still diploid and mitotic. By contrast, Khil et al. examined intact testes, of which mitotic germ cells compose only a small percentage. Most germ cells in the testes are in various stages of meiosis and are subject to a process known as meiotic sex chromosome inactivation (MSCI), in which the sex chromosomes become heterochromatic and transcriptionally inactive4. The reason MSCI occurs is unknown, but it avoids problems that might occur if nonequivalent sex chromosomes tried to recombine using autosomal, sequence-dependent pairing mechanisms5. Because of MSCI, genes expressed during meiosis cannot reside on the X or Y chromosomes because they would be inactivated by this chromosome-wide silencing mechanism.

To examine whether MSCI accounts for the observed paucity of testes-expressed genes on the X chromosome, Khil et al. used Spo11 knockout mice, in which spermatogenesis is blocked early in meiosis I, at the onset of MSCI. Using microarrays, they compared expression profiles in testes of Spo11−/− and wild-type mice. They found that genes expressed before MSCI are over-represented on the X chromosome and genes expressed after MSCI are under-represented. In the testes, then, timing is indeed everything: genes expressed early during spermatogenesis are enriched on the X chromosome, in accordance with Rice's hypothesis, and genes expressed later are prevented from residing on the X chromosome by MSCI (Fig. 1).

Figure 1: Genes expressed during early spermatogenesis (before MSCI) are enriched on the X chromosome (purple), in accordance with Rice's hypothesis.
figure 1

The onset of MSCI during early meiosis prevents genes expressed in male germ cells after this time from residing on the X chromosome (blue).

Full size image

Conflicts and resolutions

The Spo11 data set suggests that the X is masculinized with respect to testes-enriched genes not expressed during spermatogenic meiosis. But this masculinization of the X chromosome for some genes can occur simultaneously with feminization of the X chromosome for others, as increased X-chromosome localization is also seen for genes with preferential expression in two female-specific tissues, ovary and placenta2. Both of these tissues are composed primarily of somatic cells, with germ cells comprising only a small fraction of the total mass of the ovary. In all somatic tissues, dosage compensation occurs to equalize gene expression between the sexes. For unknown reasons, however, dosage compensation is not implemented in germ cells, and so the sex chromosomes have nonequivalent expression levels between males and females. Once MSCI occurs in male germ cells, the imbalance in sex-linked gene expression is further exacerbated. This imbalance probably results in sensitivity of the sex chromosomes to distinct selective pressures in the germ line relative to the soma. Thus, it is possible that one force (feminization) predominates in the soma and another (masculinization) predominates in the germ line.

Global analysis of sex-biased expression has also been done for Caenorhabditis elegans and Drosophila melanogaster6,7, and in both species, demasculinization of the X chromosome has occurred. These observations, together with the findings of Khil et al., suggest that expression from the X chromosome in each organism is subject to a curious mosaic of influences. Some superficial parallels exist between the species, and the more divergent organisms seem to be more similar in sex chromosome composition. In both C. elegans and mice, demasculinization of the X chromosome seems to be primarily driven by MSCI2,8. In flies, however, MSCI does not have a major role; instead, a gene-by-gene selective mechanism acts in both the soma and germ line to demasculinize the X chromosome7. Consistent with this observation, MSCI-dependent imprinting of the X chromosome occurs in paternal germ cells of worms and mice, but apparently not in flies9,10.

Together, the data from global expression studies2,6,7 indicate that the X chromosome is buffeted by prevailing selective forces that vary from one organism to the next. Ultimately, the forces governing gene composition of the sex chromosomes probably reflect divergent mechanisms controlling reproduction, dosage compensation and sex determination in each organism.