Role of the Xist Gene in X Chromosome Choosing* York Marahrens Jan Loring Rudolf Jaenisch
A Review by Jeremy Rusin
X inactivation is a process that occurs early in female mammal development in which one X chromosome is transcriptionally silenced. One chromosome is chosen for inactivation over the other by a choice mechanism. X inactivation was thought to be controlled at a specific region, known as the X-controlling element (Xce). As one would expect, two X chromosomes with the same allele are inactivated with the same frequency, but two with different alleles are chosen at different rates.
X choosing is apparently random in somatic tissues, but imprints, found in rodent development, inactivate the paternal X chromosome. It is not known whether random choosing and imprinting are related, but a large deletion in the Xist gene with an intact Xce can be chosen for, but fails to inactivate that chromosome. Therefore Xist is necessary for heterochromatin formation and is downstream of the random choice mechanism. In addition, female pups that inherited the mutated X from their fathers died, while female pups that inherited the mutation from their mothers survived, despite reduced birth weights.
Based on these previous findings the authors predicted that Xist+/-, if undergoing random choice, should inactivate an X chromosome in only half of the cells, causing cell loss and a reduced birth weight. However, in the present study, the researchers are trying to show that Xist+/- actually proceed with nonrandom inactivation that inactivates an X chromosome in every cell.
The relative birth weights of 14 heterozygous Searle's mutation and 14 normal mice from a previous publication showed that the heterozygous mice had birthweights 79% of those of normal littermates (Fig. 1A). When 31 normal and 10 Xist+/- mice birthweights were compared, however, the normal and heterozygous mice showed nearly identical birthweights (Fig. 1B).
The authors offer two possible explanations for these data: 1) Most likely, the cells "chose" to inactivate the wild-type chromosome in every cell 2) Less likely, embryos, while gradually losing cells, experienced compensatory growth of new cells, resulting in the normal birthweight. The latter is less likely because heterozygous cells with Searle's mutation do not experience such regrowth.
The data in Figure 1 are visually convincing for the authors' arguments, but they would be more convincing with a greater number of Xist+/- sample mice, as 10 compared to 31 is not a great sample size. Nevertheless, the error is within reason.
The authors then did phenotypic size comparisons of Xist+/- and Xist+/+ female littermates at 7.5, 8.5, and 9.5 days of gestation (Fig 2A, 2B, and 2C, respectively). The authors argue that there are no substantial differences in each comparison. Some small shape variation is present (Fig. 2A and 2B), but size differences appear to be negligible. These data suggest that Xist+/- do undergo nonrandom inactivation, as the fetuses were the same size. The data do not, however address the question of whether the same size could be attributed to compensatory growth.
A good complement to these data would be a comparison of a Searle's mutant and normal mouse to see if any visible differences exist at 7.5, 8.5, and 9.5 days. That could a better measure of relative difference in size, as Searle's mutants are known to lose a considerable proportion of cells between these periods of gestation.
To determine whether the cells retained two active chromosomes after X inactivation, the researchers used Fluorescence In Situ Hybridization (RNA FISH) to observe X inactivation in Xist+/- embryos. Several expression patterns are visible with the FISH probes during the course on inactivation. Prior to X inactivation (E6) females express an unstable transcript on each chromosome, two RNA dots, that can be observed (low level biallelic expression). Males also show a single dot pattern (low level monoallelic expression) for the unstable transcript on their single X chromosome. During X inactivation, however, one X transcript is stabilized causing a differential biallelic expression pattern (Fig. 3B). At E7.5, the majority of the cells show high-level monoallelic expression, indicating that the X inactivation is complete at E7.5, but the transcript has not been fully silenced.
The researchers predicted models for Xist expression according to random and nonrandom choice inactivation (Fig. 3C). They used two probes: one was a full probe that idnetified expression patterns in wild-type and truncated chromosomes during development, the other was a deletion probe that bound exclusively to the Xist+ (Fig. 3A). Known wild-type female expression patterns were shown in Figure 3B, exhibiting both differential biallelic and high-level monoallelic expression, with full probes, at different times in development. Using the deletion probe, the researchers predicted that random choice inactivation Xist+/- females should show high-level monoallelic expression half of the time and no expression half of the time, while nonrandom inactivation Xist+/- females should show high-level monoallelic expression 100% of the time (Fig. 3C).
FISH Xist expression in Xist+/- females at E7.5 was observed (Fig. 4). Pink indicates Xist RNA and green indicates Pgk-1 RNA, a control. Differential or high-level monoallelic expression was seen in all Xist+/- females, supporting the idea that the full-length RNA was always chosen for inactivation and the mutant was never chosen for inactivation (Fig. 4D).
Using the deletion probe, only high-level monoallelic expression was seen in Xist+/- females, indicating primary nonrandom inactivation (Fig. 4E). High-level monoallelic and biallelic expression was observed using the full probe in Xist+/- females, the smaller dot indicates the imcomplete silencing of the chromosome. All males expressed as expected: a small dot in wild-type males (Xist-) with the full probe and no expression in Xist- males with the deletion probe.
The control probe for the Pgk-1 transcript was recognized in Fig. 4B, but the control would have been more helpful in each of the expression photos. Instead, one must take the word of the authors that Pgk-1 was expressed accurately as a control.
From these and Figure 1-3 data, the authors conclude that Xist+/- female mice experience primary nonrandom inactivation of the Wild-type X chromosome. The inactivation factor is therefore a positive mechanism, as only the wild-type X chromosomes were inactivated. The authors argue that because random choice in somatic cells is affected by the mutation, the random choice and nonrandom choosing must be independent of one another. Contrasting results with previous work appear to be the result of differing deletions in the mutant chromosome. Previous work had deletions that spanned the promoter region affecting transcription; however in this work the deletion only spanned exon 1 to intron 5. Because random choosing can ossur in the absence of Xist, the nonrandom choosing mechanism may lie in one of the deleted introns.
Finally, the authors propose a model for nonrandom inactivation that fits with diploid, triploid, and tetraploid organsims (Fig. 5B). According to their model, several inactivation factors are present within the cell and one inactivation blocker is present. After the inactivation factor binds, the blocker is set to bind on the other chromosome. If a deletion is present, the inactivation factor can only bind to the wild-type chromosome, leaving the mutant to receive the blocking factor and transcriptional activation. This proposed model is in contrast to previous thought that a single inactivation factor is present inactivating an X chromosome (Fig. 5A). This model, however, helps to explain how multiple X chromosomes are inactivated in triploid and tetraploid organsims, for example.
This paper was well-written, concise, and the figures, for the most part were presented well and applicable. Data in these figures support the hypothesis that X chromosome choosing is nonrandom when the two X chromosomes have different alleles in the Xist gene. Future research could bolster findings in this paper and explore the blocking and inactivation factors further, as well as whether or not compensatory growth is experienced in embyos as cells are lost.
Future research on this topic could take several directions:
1) The inactivation model at the conclusion of the paper, though explaining inactivation factor binding, fails to explain how a blocking factor could bind to a deleted site that the inactivation factor could not. Examining the Xist gene in further detail, I.e. with varying deletion size and placement, will probably show more about the binding factors, as at some point deletions will make binding by each impossible.
2) Further study of X inactivation could explore the functional transcription pathway. An insertion of GFP encoding region, for example, would show which X chromosome was actively transcribing in the presence of the blocking factor. Glowing chromosomes or regions of chromosomes would be indicators of transcription and might also lead reserchers to more specific sites of importance to the Xist gene on the X chromosome because it is a functional study.
3) A helpful test would be SSCP, or single strand conformation polymorphism, on ÒactiveÓ versus ÒinactiveÓ chromosomes on analogous regions. Differences in sequences, otherwise undetectible, should lead to a better understanding of the threshold at which inactivation binding and blocking factors still operate.
4) Along the same lines, creating deletions of varying size and in different regions of Xist will be helpful in observing interactions between binding factors and chromosomes, given the ÒchoiceÓ of two X chromosomes with different mutant alleles. What is the necessary region for inactivation? Can a single bp substitution affect inactivation processes?
5) One of the most important questions to address is the identity of the binding factors. Examining the regions that are bound by the blocking and inactivation factors could lead to probes for the factors and the structures of these factors. 6) The question of compensatory growth could be addressed by labeling chromosomes with a fluorescent marker and observing whether new growth accounts for similar birth size Xist+/- offspring with Xist +/+ littermates.
Continued research on X chromosome choosing is critical, as X chromosome inactivation determines allelic representation in somatic expression.
*from Cell 92: 657 1998
or Return to Jeremy's Home Page