Review Paper: Role of the Xist Gene in X Chromosome Choosing

York Marahrens, Jan Loring, and Rudolf Jaenisch

This study tries to explain how X inactivation occurs. Past studies have shown that in female mammals, when deciding which of the two X chromosomes will be inactivated, or silenced, a random choice mechanism occurs. This study contradicts this postulated random X inactivation mechanism, by demonstrating that Xist+/- females go thru nonrandom primary inactivation of the wt X chromosome in every cell. More specifically, the study claims that if an internal deletion occurs in the Xist gene, then there will be primary nonrandom inactivation of the wt X chromosome in a heterozygous female.

A basic background is discussed in the paper so as to better understand the study. Xic is the X chromosome inactivation center. In order for an X chromosome to be inactivated, two Xic regions must be there. The Xic region contains two elements within it: the Xce and the Xist.

The Xce is the X controlling element. X chromosome choosing occurs here. It has been postulated that the way it works is by the Xce blocking one of the X chromosomes, while the other X is inactivated. Their affinity for the blocking factor determines which X chromosome will be blocked. There is an equal chance for X chromosomes to be chosen only if both have the same Xce allele.

The Xist region within the Xic, is separate from the Xce region. The Xist gene functions downstream of the X choice mechanism. It plays an important role in choosing an X chromosome and it is necessary for the inactivation of an X chromosome to occur. It is a 15kb of untranslated RNA that is stabilized during X inactivation and it coats the inactive chromosome. Xist is required for heterochromatin formation.

A past study showed that the mutant epiblast of offspring with maternally inherited Xist mutations, made: balanced cells with inactive wt X chromosome, and unbalanced cells with two active X chromosomes at the time of inactivation. These unbalanced cells were lost. The loss of cells expressing both XÕs occurs in the offspring inheriting SearleÕs autosome translocation. These offspring, when they undergo random choice, result in : wt X inactivation (balanced cells) and inactivation of the translocation product with the Xic (unbalanced). Like mentioned before, these latter unbalanced cells are lost.

If random choice is the mechanism, then half the cells should not inactivate the X chromosome, and will result in cell loss and smaller size of the offspring. This study claims that an internal deletion occurs in the Xist gene, so that there will be primary nonrandom inactivation of the wt X chromosome in a heterozygous female.

Analysis of the Data

Fig 1 These are graphs comparing the relative weights of newborn mice. 1A is previously published data used to compare with this paperÕs experimental data. The wt were assigned a value of 100% and the rest were weighed in relation to them.

In 1a, 14 heterozygous females for SearleÕs translocation were compared to 14 normal siblings. Naturally, those heterozygous for SearleÕs translocation lose most unbalanced cells, however, they still develop into healthy pups. They are different from the wt because they are smaller in size. Fig 1A shows the XX16 to be 79% of the weight of the wild type pups at birth.

In 1b, 10 Xist+/- females (those with a deletion in the Xist gene) were compared to 31 wildtype siblings. The XXxist- pups were the same weight at birth as the wt ones.

There are two explanations for this. One, is that primary nonrandom inactivation occured. The other is that random X inactivation occured. In the former case, the Xist+/- embryos conserved all the epiblast because they inactivated the wt X on every cell. In the latter explanation, the Xist+/- tried to do random X inactivation by losing half of its epiblast. But this explanation requires that the Xist+/- embryos experience alot of compensatory growth so that normal birthweights remained in the end.

SearleÕs translocation would cause heterozygous embryos to lose most of their unbalanced cells and rarely fully compensate for the cell loss. This occurs at about embryonic day 7 and 8. Losing the majority of these unbalanced cells (epiblast) from Xist+/- embryos would result in smaller embryos at embryonic days 8.5 and 9.5 when compared to the wt littermates. Since random choice is still a possibility, they further tested the embryos in Fig 2 to determine which mechanism was taking place.

Fig2 This figure compares the Xist+/+(wt) and Xist+/- littermates phenotypically at different points of gestation.

In 2A, it compares them at embryonic day 7.5, in B at embryonic day 8.5, and in C, at embryonic day 9.5. Comparing the Xist+/+ and Xist+/- embryos at these three different points of gestation showed no difference in size from the wild types. This histological comparison confirmed that indeed the Xist+/- embryos are not smaller than their wt littermates. Seeing that even after the Xist+/- inactivated their X chromosome, they still had the same size as the wts, gave further support to primary nonrandom X inactivation occuring here. There is still, however, a possibility that random inactivation occured. It could have occured so that the loss of the unbalanced cells was so gradual and slow that compensatory growth rate kept up with the loss, thus the embryos maintained their normal size all throughout. Further evidence for nonrandom inactivation is therefore needed.

Fig 3 To see if after X inactivation, the Xist+/- embryos still have cells with two active X chromosomes, the FISH method was used. FISH checks the inactivation status of the individual cells in Xist+/- embryos. Figure 3 is a prediction of the patterns of Xist expression.

Figure 3A is a map of the wt and the mutant of the Xist gene. It shows the probes used: a full probe, and a deletion probe. The full probe recognizes both the wt and the truncated Xist RNAÕs. The deletion probe was used with FISH to confirm that Xist +/- embryos undergo primary nonrandom X inactivation since it only recognizes the wt (not the mutant).

In figure 3B one sees the pattern of Xist expression by FISH in female cells. Before the X chromosome is inactivated, one can detect two dot-like RNA signals (at each Xist locus) which are the unstable Xist transcripts expressed from both X chromosomes. This is seem at the low-level biallelic expression pattern. The differential biallelic pattern seem in the figure shows that during the inactivation of X, the Xist RNA from one allele is stable and it coats the X chromosome. The high level monoallelic pattern shows the Xist gene on an active X chromosome as silenced. Nevertheless, even though X inactivation is complete at embryonic day 7.5, silencing in all cells of the expression of the unstable transcript has not occured.

Figure 3C predicts the pattern of random choosing of either X chromosome and shows that it is very different from the pattern also predicted in this figure for primary nonrandom selection of wt X chromosome when the deletion probe is used. A low level monoallelic pattern should be witnessed in cells that selected the mutant X but had not yet silenced the wt Xist locus, if the chosing is done at random in 1/2 of the cells. Conversely, all cells should show a high level monoallelic pattern if wt X went thru nonrandom inactivation. The experimenters then put it to the test in the following figure 4.

Fig 4 This figure shows the actual Xist expression in female Xist+/- embryos at embryonic day 7.5. FISH was used to do this. In figure A, C, and D the full probe was used. In B and E the deletion probe was used.

4A is a Xist- male. It was used as a control. Probing it with the full probe showed this control had a low-level monoallelic expression pattern. This was just like, or indistinguishable from the expression seen in wt males.

4B is a Xist- male. Probing with the deletion probe showed no detection of the Xist signal in these embryos. They wanted to make sure the deletion had occured. An additional probe that recognizes the Pgk-1 transcript was used as a control. This was done to demonstrate that the absence of the Xist signal was not due to technical failure.

4C is a Xist +/+ female or wt. Probing with the full probe showed that X inactivation was complete in wt female embryos showing a high level monoallelic or differential biallelic expression. Past studies have shown that truncated Xist RNA will not coat or inactivate an X chromosome, therefore it was expected that those cells that had picked the mutant X chromosome should not display the signal expression pattern of high level monoallelic and differential biallelic cells.

4D is a Xist +/- female. Probing with the full probe revealed that high level monoallelic or differential biallelic expression was indeed seen in all cells of Xist+/- embryos. This implied that the Xist RNA was coating the wt X chromosome and the mutant X chromosome was not chosen.

4E is a Xist +/- female. Probing with the deletion probe showed the high level monoallelic signal. There was no low level monoallelic signal. Female wt E7.5 embryos produced both differential biallelic and high level monoallelic expression patterns. There was considerably more high level monoallelic expression (83%) than biallelic (43%). Overall, the results show that Xist+/- embryos go thru primary nonrandom inactivation of the wt X chromosome.

So What Does this All Mean?

Nevertheless, the researchers draw the conclusion that the embryos having a deletion in one copy of the Xist gene, went thru primary nonrandom activation of the chromosome. They further confirmed that the Xist gene is needed in X chromosome choosing because there was no random choice in the Xist +/- cells. The deletion aparently cut out a positive element that was needed for an X chromosome to be chosen. Once again, the data in this study contradicts the past study that claims random choice of X inactivation by action of a blocking factor. The wt X chromosome in this study was always inactivated rather than blocked and remain active. Thus they located a postive element that starts heterochromatin formation.

They discuss two roles for the Xist in X inactivation. A prior study (Penny) in which a deletion in the Xist gene produced cells with inactive wt X and cells with two active XÕs, differs from this experimentÕs results. PennyÕs study showed that the random choice mechanism was NOT disturbed by the deletion in the Xist. How can this be? The experimenters claim that it is due to the location of the deletion. PennyÕs deletion covers the promoter and part of exon 1, while the deletion in this study occured from exon 1 to intron 5. PennyÕs deletion affects transcription, while this experiments deletion does not. This shows that a 6kb segment is absolutely necessary for the random choice mechanism.

The experimenters try to explain the X inactivation patterns seen in the experiment using two models. In the first model, figure 5A, they propose that a female cell makes one initiation factor that binds and inactivates one X chromosome. This initiation factor cannot recognize a mutant Xist allele caused by a deletion. Therefore, it would always bind and inactivate the wt X chromosome. If the deletion doesnÕt affect the initiation factor binding, then cells with two active XÕs and cells with inactive wt X chromosomes result. The other model for choosing the X chromosome to be deleted is shown in figure 5B. The binding factor in this mechanism would bind to the Xist gene but it shifts to and from between the two X chromosomes. The X chromosome not bound to the binding factor, would have the initiation factor bind to it. The binding of the initiation factor and the blocking factor is mutually exclusive. Therefore, each chromosome has a Ôset fateÕ. If a deletion of that initiation binding site occurs, all cells go to the blocking factor on the mutant allele.

The experimenters suggest that the Xist location is the nucleation site where heterochromatin begins to form when X chromosome is inactivated. They suggest that it works by the Xist transcript being stabilized by proteins recognizing it. Xce might act by enhancing the initiator binding to Xist. This is assuming Xist to be different from Xce, of course.

It is my opinion that it would have been helpful to know the percentages of the cells showing the highlevel monoallelic pattern and those that didnÕt, for the random choice mechanism prediction in figure 3C. They failed to mention whether half the cells showed high level monoallelic and half didnÕt. Of course, I assume like they said that absolutely all cells expressed the high level monoallelic pattern since they underwent nonrandom choosing, but I was curious as to the actual percentages and if there were any cells at all that did not show high level monoallelic after the deletion. I also questioned the certainty in which they drew their conclusions from the histological comparisons in figure 2. Although the embryos do appear to be equal in size, I fear that this might be a very superficial comparison and differences could easily be missed. One other thing that the researchers did was introduce information in their discussion with no supporting data. They claim that the location of the deletion on the Xist gene was between exon 1 and intron 5, but we are to blindly accept this. They could have placed the evidence in the paper if they were going to claim a deletion causes nonrandom choosing. The fact that past studies show deletions also lead to random choosing, should have geared them to including this evidence for it specifies the particular deletion. Overall, I think that the use of a functional test, such as testing expression by FISH, was solid and good support to use for their study. It provided solid support for their hypothesis of primary nonrandom inactivation.

Their suggestions for future studies include researching on whether Xce and Xist are synonymous. Also some studies have shown skewed inactivation patterns in some families, perhaps by studying the reason for this will allow us to see more clearly the role of the XIST promoter.

In the future, studies could be geared to learning the specifics of X inactivation. The elements and sites involved, and factors that bind these elements should be studied to clarify the mechanism of X inactivation further. They proposed two possible mechanisms for inactivation, they should in the future put them to the test to see which of the two, if any, is the actual X choosing mechanism. For example, they could test if the blocking factor is specific to these X chromosomes, and thus unique, or is it like any other blocking factor. One could test its binding to different chromosomes using a probe to detect the binding. They could perform functional tests such as FISH to test for active and inactive chromosomes after the blocking factor has bound. The possibilities in testing for the specifics of the nonrandom choosing of the X chromosome are endless.


Cell Vol. 92: 657-664, March 6, 1998. Copyright 1998 by Cell Press

To link back to my homepage click here.

Click here to return to the Molecular Biology Home Page,

To send me email click here,