Lack of global meiotic sex chromosome inactivation, and paucity of tissue-specific gene expression on the Drosophila X chromosome

Abstract

Background: Paucity of male-biased genes on the Drosophila X chromosome is a well-established phenomenon, thought to be specifically linked to the role of these genes in reproduction and/or their expression in the meiotic male germline. In particular, meiotic sex chromosome inactivation (MSCI) has been widely considered a driving force behind depletion of spermatocyte-biased X-linked genes in Drosophila by analogy with mammals, even though the existence of global MCSI in Drosophila has not been proven.

Results: Microarray-based study and qRT-PCR analyses show that the dynamics of gene expression during testis development are very similar between X-linked and autosomal genes, with both showing transcriptional activation concomitant with meiosis. However, the genes showing at least ten-fold expression bias toward testis are significantly underrepresented on the X chromosome. Intriguingly, the genes with similar expression bias toward tissues other than testis, even those not apparently associated with reproduction, are also strongly underrepresented on the X. Bioinformatics analysis shows that while tissue-specific genes often bind silencing-associated factors in embryonic and cultured cells, this trend is less prominent for the X-linked genes.

Conclusions: Our data show that the global meiotic inactivation of the X chromosome does not occur in Drosophila. Paucity of testis-biased genes on the X appears not to be linked to reproduction or germline-specific events, but rather reflects a general underrepresentation of tissue-biased genes on this chromosome. Our analyses suggest that the activation/repression switch mechanisms that probably orchestrate the highly-biased expression of tissue-specific genes are generally not efficient on the X chromosome. This effect, probably caused by dosage compensation counteracting repression of the X-linked genes, may be the cause of the exodus of highly tissue-biased genes to the autosomes.

Figure 1

X-linked and the autosomal testis-biased genes show similar patterns of activation in testis development. (A) X-linked genes; the profile for the known spermatocyte-specific gene Sdic is outlined as a black dotted line. Other genes are CG15450, CG1314, CG1338, CG15711, CG15452, CG11227 and CG1324 (gray lines). (B) autosomal genes; the profiles for the known spermatocyte-specific genes fzo, dj, and β(2)Tubulin are outlined as black dotted lines. The profile for the gene ocn which is the source of promoter in the transgene based study of Hense et al. [15] is outlined as a black solid line. Other genes are CG6262, CG3483, CG7813, CG3492, CG15874, CG3494, CG16837, CG4439, CG4750, CG15873, CG15925, CG7848, CG15710, Eyc, Mst35Ba, and Mst35Bb (gray lines). (C) Box plot analysis of gene expression data (A, B) shows no significant differences in the expression patterns between the X-linked (orange) and the autosomal (blue) gene sets. For each gene, expression level in pupae served as the reference.

Figure 2

Lack of global X-chromosome inactivation in developing testes. Expression of X-linked (orange) and autosomal (blue) genes was measured as signal intensity of corresponding microarray probes, after normalization. cDNAs hybridized to the microarrays were isolated from testes of either feeding (f) or wandering (w) larvae grown for the indicated number of days at 18°C. The analysis traces the first wave of germline differentiation; pupation indicative of the meiotic divisions occurred at day 11.

Figure 3

Paucity of X-linked genes with high expression bias toward testis. The testis::somatic tissue ratio of microarray signal intensities, observed between adult testis and a panel of somatic tissues [24], serves as a measure of testis bias of gene expression and is shown at bottom. Bars indicate proportion of genes up-regulated in testis development in the four bias categories. Orange bars, X-linked genes; blue bars, autosomal genes.

Figure 4

Highly tissue-biased genes show a skewed representation on the X chromosome. The ratio of microarray signal intensities observed between the tissue sample indicated at bottom including midgut, malpigian tubule, accessory gland, salivary gland, head, and ovary and a panel of other tissue samples [24] was used as a measure of tissue bias. The bars correspond to the frequencies of the genes on the X chromosome normalized against the genome averages, and are shown for the genes with at least two-fold expression bias toward indicated tissues (light gray), for the genes with at least five-fold bias (dark gray), and for highly biased genes with at least ten-fold bias (black).

Figure 5

'Inversed' correlations between gene regulation in testis development and chromatin modifications in somatic cells. For each of the developmental time points indicated on the horizontal axis, correlations were calculated between up-regulation of genes in testis relative to the earliest time point (from microarray data) and binding of the same genes to the shown chromatin proteins in embryos or somatic cells [29,30]. Yellow, orange, and purple colors represent proteins that are generally associated with gene silencing, and blue and green colors represent proteins generally associated with active gene expression.

Figure 6

The autosomal and the X-linked tissue-biased genes show different patterns of chromatin modifications. The frequencies of binding targets for the proteins indicated at the bottom were calculated among the X-linked (orange) and autosomal (blue) tissue-biased genes. The sets of genes showing at least two-fold expression bias toward the indicated tissues (such as testis, midgut, accessory gland, salivary gland, malpigian tubule, and ovary) were generated from the genome-wide expression data [24]. Protein binding genes were defined using chromatin immunoprecipitation and DamID data [29,30] with arbitrarily set thresholds. Bars show relative increase or decrease in binding target frequency within analyzed gene sets as compared to the entire genome.