Role of the Xist Gene in X Chromosome Choosing

Review Paper

York Marahrens, Jan Loring, and Rudolph Jaenisch. 1998. Cell, Vol 92, 657-664.

In female mammals, one of the two X chromosomes becomes inactive in all cells. The X chromosome to be silenced is chosen at the X controlling element (Xce) (Cattanach and Papworth, 1981; Simmler et al., 1993). Different alleles indicate different probabilities of being chosen. The Xist gene has been mapped to the Xce locus and encodes a 15kb untranslated RNA that coats the inactive X chromosome (Clemson et al., 1996).

In a previous experiment, a line of mice with a deletion in the Xist gene was developed (Marahrens et al., 1997). Female progeny who had inherited the mutated gene paternally died during embryogenesis due to a failure to undergo X inactivation. However, female progeny who inherited the mutation maternally were healthy, with the wild type paternal X chromosome inactive in every cell. This scheme is indicative of nonrandom selection of the X chromosome for inactivation in mice, in that the paternal X chromosome is always the one that is inactivated. The paper I am reviewing further examines this mutation by attempting to determine if the consistent inactivation of the paternal wild type X chromosome in this mice strain occurred by random choice or by primary nonrandom selection.

The Xist gene is transcribed into RNA which upon X inactivation forms a heterochromatin coat around the inactive chromosome. Transcription of RNA from the X chromosome that remains active is silenced. If the Xist gene is mutated, as with a deletion in this case, heterochromatin formation will not occur as readily on that particular chromosome. Therefore, while the mutant X chromosome could be chosen for inactivation, inactivation could not actually occur because heterochromatin does not form. Two X chromosomes will be expressed in the female mice.

If choice were random as to which X chromosome would become inactive, some cells in the strain of mice would have one active X and some two. Cells with one active X would result from the inactivation of the wild type X which can form the heterochromatin coat. Cells with two active X chromosomes would result from inactivation of the X chromosome carrying the deletion. Although this chromosome could be chosen because of a wild type Xce site, it could not actually become inactive due to the lack of ability to make heterochromatin. Those cells with two active X chromosomes would be lost, which previous experiments had shown could happen (McMahon and Monk, 1983). Loss of cells would occur between embryonic day seven and embryonic day eight. Therefore, in order to support their conclusion that primary nonrandom selection was occurring, the authors had to prove that no cells were lost.

According to previous experiments, if the cells were lost, progeny would be smaller at birth (Lyon et al., 1964). However, in this study female pups born heterogenous for the Xist deletion had comparable birthweights to wild type siblings. Figure one demonstrates this fact, providing previously published data in which cells were lost and pups born were smaller than wild type (Figure 1A) and data from this experiment, in which the Xist +/- mice are shown to be the same birthweight as wild type mice.

Despite this fact, one still cannot rule out loss of cells in which both X chromosomes were active. Compensatory growth could have occurred in which the cells lost were replaced. This would have to happen very quickly. The authors attempt to rule out this factor by comparing the embryos' sizes immediately after X chromosome inactivation, at E7.5 to E9.5. This comparison is shown in figure two. If compensatory growth were occuring, we would see smaller embryos in the mutated strain at this stage. It appears that the embryos are of comparable size, indicating that there is no immediate cell loss and subsequently no compensatory growth occurs. Because all cells are retained, some evidence is provided that the wild type X chromosome is inactivated in all these mice and this inactivation is nonrandom.

There are some problems with figure two. As the authors provide no evidence of genotype, we must take them at their word that the genotypes indicated are in fact the genotypes present. To increase the credibility of this figure, the authors should have done an Single Stranded Conformation Polymorphism analysis of the chromosome or a Restriction Fragment Length Polymorphism analysis. Either one of these methods would have provided evidence that a deletion was present where the figure legend indicated and would have given this figure more reliability.

The authors sought to detect unbalanced cells, that is, cells with two activated X chromosomes, which would be indicative of random inactivation selection. RNA Flourescence in situ Hybridiziation (RNA-FISH) was used to detect the Xist transcript shortly after inactivation. This would determine if Xist +/- embryos retain cells with two active X chromosomes shortly after inactivation. Prior to inactivation, two small dots of flourescence would be detected, a low level biallelic signal (Figure 3B). Unbalanced cells would continue this signal. In male cells, one dot would be detected, in a low level monoallelic signal. However, during X inactivation, RNA transcribed from the Xist gene coats the inactivated chromosome, resulting in a differential biallelic signal (Figure 3B) consisting of one large and one small flourescent spot. When the Xist on the active chromosome is silenced, one sees a high level monoallelic signal (Figure 3B), just one large flourescent spot. At E7.5, fifteen percent of wild type females show a differential biallelic pattern and eighty five percent show high level monoallelic patterns. Fifteen percent of wild type males demonstrated the presence of an unstable transcript. This indicates that X inactivation is complete at E7.5 but that the transcript from the silenced Xist may still be expressed in both male and female cells. (Panning et al., 1997). Therefore, to determine X inactivation status, RNA-FISH had to be performed after E7.5.

Figure 3A shows a full probe, which recognizes both wild type Xist transcript and RNA transcribed from the Xist with the deletion. It also shows a deletion probe, which recognizes only wild type Xist RNA. When cells from Xist- males were probed with the full probe, a small amount of Xist RNA was detected, indicated by a tiny pink spot (Figure 4A). According to the text, this is low level monoallelic expression and is indistinguishable from wild type expression (data not shown). When Xist - cells from males were probed with the deletion probe, no Xist wild type RNA was detected, as indicated by the absence of any florescence. The possibility of technical error was ruled out when the cells were probed for Pgk-1 , which appears as a green spot (Figure 4B). These two experiments serve as controls. The first one is a positive control, ensuring that the full probe will bind to small amounts of Xist RNA. The second experiment serves as a negative control. It indicates that the probe for wild type Xist RNA only binds to wild type RNA. The authors then probed an Xist +/+ female with the full probe. Pink spots appeared in the cells (Figure 4C) indicating that Xist was expressed in a differential biallelic pattern (Figure 3B). This also serves as a positive control.

When the authors probed an Xist +/- female with the full probe, all cells demonstrated either differential biallelic expression of transcribed RNA or high level monoallelic expression (Figure 4D). This would only occur if primary nonrandom selection were occurring (Figure 3C). If selection were random, we would expect that half the cells would silence the X chromosome with the mutated Xist, resulting in low level biallelic expression and that half the cells would silence the wild type X chromosome, resulting in high level monoallelic expression or differential biallelic expression. However, if selection were nonrandom, we would expect only high level monoallelic expression or differential biallelic expression. This is, in fact, what occurs, supporting the primary nonrandom selection hypothesis.

Finally, the authors probed Xist +/- females with the deletion probe. Two large pink spots, one in each cell, were seen (Figure 4E). This could only occur if it were the wild type Xist RNA that was being transcribed and formed the heterochromatin coat. If the wild type X chromosome were remaining active in these cells, we would expect to see small or no spots of flourescence when cells were probed with the deletion probe for wild type RNA. Therefore, the wild type X chromosome is being inactivated in all cells and is being coated with a heterochromatin coat. Because this occurs in all cells, it indicates again primary nonrandom inactivation.

The authors conclude that the Xist allele is required for choice in X chromosome inactivation and that the selection is nonrandom in mice heterozygous for this particular deletion. One mechanism that they suggest for this is that an initiation factor that results in X inactivation binds to this segment of the chromosome. The Xist gene with the deletion would be unable to bind this initiation factor, and thus the wild type X chromosome is always the one that is inactivated (Figure 5A). However, this explanation has problems. It does not explain how inactivation could occur in organisms with more than one X chromosome. X chromosome inactivation operates by the N-1 rule. Thus, an organism with three X chromosomes must inactivate two of them, and an organism with four X chromosomes must inactivate three of them. To explain how this could occur, the authors further their model. They suggest that a cell produces a blocking factor that binds to the Xist gene and switches between the two chromosomes. The initiation factor binds to all X chromosomes not already attached to the blocking factor (Figure 5B). A mutation in the Xist gene would cause the initiation factor to bind to the wild type X chromosome .

The authors hypothesize that the Xist locus is where heterochromatin formation begins. Any proteins (i.e. initiation factors) that recognizes Xist may serve to stabilize resulting RNA, leading to heterochromatin formation and subsequent inactivation of that X chromosome. The upstream Xce locus may in fact be an enhancer that affects binding of the initiation factor to Xist.

From here, research could take many directions. Experiments could begin by confirming the presence of initiation factors binding to the Xist gene. To do this, one would perform column chromatography. The Xist gene would be attached to beads and protein from the cell poured over the column. The nonbinding protein would bewashed off. In other words, this test would confirm that initiation factors did in fact bind to the protein. A funtional assay could then be performed to determine if these initiation factors in fact inactivated the X chromosome.

Yet another direction for research would be to determine the functional elements in this mechanism. For example, just what part of the Xist gene is necessary for heterochromatin formation and subsequent X chromosome choosing? To determine this, site directed mutagenesis would be performed on the Xist gene. A wild type gene would be used as a positive control in which we would expect initiation factors to always bind and one X chromosome would always be inactivated. As a negative control, one would use a gene known not to bind to Xist initiation factors. It would always be transcribed. If fluorescence in situ hybridization were performed on these controls and the experimental mutagenesis subjects, one would expect to see a wide variety of results. If the positive control were probed for Xist RNA, we would expect to see a differential biallelic or high level monoallelic pattern. If the negative control were probed for the gene's RNA, we would expect to see it expressed in an equal level in all cells. Finally,. when the experimental cells are probed for Xist RNA, we would expect to see a differential biallelic pattern or a high level monoallelic pattern if the binding of initiation factors were not affected. However, if the binding of initiation factors were affected, we would expect to see only the low level biallelic transcription of RNA.

Finally, the role of the Xce gene as an enhancer for binding of initiation factors could be examined, again by site directed mutagenesis. The Xce gene could be deleted entirely, in which case if it was the enhancer we would expect no initiation factors to bind and both X chromosomes to be expressed. A wild type cell would serve as a positive control. Site directed mutagenesis would be used to alter the Xce locus and the effect on inactivation of the X chromosome could be observed.

Works Cited:

Cattanach, B.M. and Papworth, D. (1981). Controlling elements in the mouse V. linkage tests with X-linked genes. Genet. Res. 38., 57-70.

Clemsonm C.M., McNeil., J.A., Willard, H.F., and Lawrence, J.B. (1996). XIST RNA paints the inactive X chromosome at interphase: evidence for a novel RNA involved in nuclear/chromosome structure. J. Cell Bio. 132, 259 275.

Lyon, M.F., Searle, A.G., Ford, C.E., and Ohno, S. (1964). A mouse translocation suppressing sex-linked variegation. Cytogenics 3, 306-323.

Marahrens, Y., Panning, B., Dausman, J., Strauss, W., and Jaenisch, R. (1997). Xist deficient mice are defective in dosage compensation but not spermatogenesis. Genes Dev. 11. 156-166.

McMahon, A. and Monk, M. (1983). X chromosome activity in female mouse embryos heterozygous for Pgk-1 and SearleÕs translocation, T(X; 16) 16H. Genet. Res. 41, 69-83.

Panning, B., Dausman, J., amd Jaenisch, R. (1997). X chromosome inactivation is mediated by Xist RNA stabilization. Cell 90, 907-916.

Penny, G.D., Kay, G.F., Sheardown, S.A., Rastan, S. and Brockdorff, N. (1996). Requirement for the Xist in X chromosome inactivation. Nature 379, 131 137.

Simmler. M.C., Cattanach, B.M., Rasberry, C., Rougeulle, C. and Avner, P. (1993). Mapping the murine Xce locue with (CA)n repeats. Mamm. Genome 4, 523-530.

To return to my homepage, click here.