We identified the gene carrying the juvenile spermatogonial depletion mutation (jsd), a recessive spermatogenic defect mapped to mouse chromosome 1 (refs. 1,2). We localized jsd to a 272-kb region and resequenced this area to identify the underlying mutation: a frameshift that severely truncates the predicted protein product of a 2.3-kb genomic open reading frame. This gene, Utp14b, evidently arose through reverse transcription of an mRNA from an X-linked gene and integration of the resulting cDNA into an intron of an autosomal gene, whose promoter and 5' untranslated exons are shared with Utp14b. To our knowledge, Utp14b is the first protein-coding retrogene to be linked to a recessive mammalian phenotype. The X-linked progenitor of Utp14b is the mammalian ortholog of yeast Utp14, which encodes a protein required for processing of pre-rRNA and hence for ribosome assembly. Our findings substantiate the hypothesis that mammalian spermatogenesis is supported by autosomal retrogenes that evolved from X-linked housekeeping genes to compensate for silencing of the X chromosome during male meiosis. We find that Utp14b-like retrogenes arose independently and were conserved during evolution in at least four mammalian lineages. This recurrence implies a strong selective pressure, perhaps to enable ribosome assembly in male meiotic cells.

One X chromosome, selected at random, is silenced in each female mammalian cell. Xist encodes a noncoding RNA that influences the probability that the cis-linked X chromosome will be silenced. We found that the A-repeat, a highly conserved element within Xist, is required for the accumulation of spliced Xist RNA. In addition, the A-repeat is necessary for X-inactivation to occur randomly. In combination, our data suggest that normal Xist RNA processing is important in the regulation of random X-inactivation. We propose that modulation of Xist RNA processing may be part of the stochastic process that determines which X chromosome will be inactivated.

Early in the development of female mammals, one of the two X chromosomes is silenced in half of cells and the other X chromosome is silenced in the remaining half. The basis of this apparent randomness is not understood. We show that before X-inactivation, the two X chromosomes appear to exist in distinct states that correspond to their fates as the active and inactive X chromosomes. Xist and Tsix, noncoding RNAs that control X chromosome fates upon X-inactivation, also determine the states of the X chromosomes prior to X-inactivation. In wild-type ES cells, X chromosomes switch between states; among the progeny of a single cell, a given X chromosome exhibits each state with equal frequency. We propose a model in which the concerted switching of homologous X chromosomes between mutually exclusive future active and future inactive states provides the basis for the apparently random silencing of one X chromosome in female cells.

Anumber of mammalian genes are expressed from only one of the two homologous chromosomes, selected at random in each cell. These include genes subject to X-inactivation, olfactory receptor (OR) genes, and several classes of immune system genes. The means by which monoallelic expression is established are only beginning to be understood. Using a cytological assay, we show that the two homologous alleles of autosomal random monoallelic loci differ from each other in embryonic stem (ES) cells, before establishment of monoallelic expression. The Polycomb Group gene Eed is required to establish this distinctive behavior. In addition, we found that when Eed mutant ES cells are differentiated, they fail to establish asynchronous replication timing at OR loci. These results suggest a common mechanism for random monoallelic expression on autosomes and the X chromosome, and implicate Eed in establishing differences between homologous OR loci before and after differentiation.

No abstract