"The Role of Chromosomal RNAs in Marking the X for Dosage Compensation".

Richard L. Kelley ¹ and Mitzi I Kuroda ²

¹ Baylor College of Medicine and ² Howard Hughes Medical Institute, Department of Molecular and Cellular Biology, One Baylor Plaza, Mail Stop BCM235, Houston, Texas 77030, USA

¹ E-mail: rkelley@bcm.tmc.edu
² E-mail: mkuroda@bcm.tmc.edu 

Both flies and mammals remodel the architecture of the X chromosome to achieve dosage compensation. A novel class of noncoding RNAs that paint entire chromosomes are centrally involved in this process. The genes encoding these unusual RNAs are themselves located on the X, and are key sites that target the X for dosage compensation.

Introduction:

Many species use sex chromosome number or composition to determine sex. For instance, in both flies and mammals, females are XX:AA and males are XY:AA, where A represents a set of autosomes. The Y chromosome carries few genes, so that males are effectively monosomic for the X, and yet they seem to cope with this genetic imbalance remarkably well. The key is dosage compensation, the process that enables each sex to produce similar amounts of X-linked products from different numbers of sex chromosomes.

There is every reason to believe that the mechanisms of dosage compensation in Drosophila and mammals arose independently long after they shared a common ancestor. The primary pathway of dosage compensation in Drosophila operates in males to hypertranscribe most genes along the single male X to match the output of the female's two X chromosomes (reviewed in [1-4]). This is carried out by the MSL (male specific lethal) complex which appears to remodel chromatin architecture by a site-specific acetylation on histone H4 [5*, 6*]. In mammals, it is the female which dosage compensates, by silencing most of the genes on one of the two chromosomes via the formation of a heterochromatic Barr body (reviwed in [7, 8]).

In this review, we focus on how the cell distinguishes the X from the autosomes as the correct target of dosage compensation. Recent results point to a novel mechanism operating through a new class of stable RNAs which coat the dosage-compensated chromosome. Studies of two such novel RNAs in flies have changed our understanding of the composition of the MSL complex and how it is targeted to the X, and revealed unexpected mechanistic parallels with mammals.

Chromosomal RNAs:

A pivotal discovery regarding dosage compensation in Drosophila came with the identification of two unusual roX (RNA on the X chromosome) RNAs [9, 10]. The roX1 and roX2 RNAs are large (~3.7 and 0.6-1.0 kb, respectively), male-specific, spliced, and polyadenylated transcripts showing almost no primary sequence similarity to each other. However, both lack significant open reading frames and remain nuclear. The most dramatic observation is that roX RNAs paint the Drosophila male X chromosome [9, 11**, 12*]. This indicates that roX RNAs are members of an entirely new class of chromosomally associated RNA whose only other member is the Xist (X inactive specific transcript) RNA which mediates dosage compensation in mammals. Xist is also large (~15-17 kb), spliced, and polyadenylated [13, 14]. Like the mammalian Xist gene, the roX1 and roX2 genes both map to the x chromosome, a genomic location which is central to their functions [10, 15, 16].

roX RNAs are Components of the MSL Complex:

New evidence supports the idea that roX RNAs are physically associated with the MSL proteins in a high molecular weight complex which remodels chromatin. First, both roX1 and roX2 RNAs precisely colocalize with the MSL proteins at hundreds of bands along the male X chromosome [11**, 12*]. Second, roX RNAs are exquisitely unstable in the absence of a complete set of MSL proteins, suggesting that the MSL proteins bind the roX RNAs directly to protect them from destruction [12*]. Third, MSL complex formation is either delayed or blocked when both roX1 and roX2 are removed [11**]. Finally, antibodies directed against the protein subunits of the MSL complex can specifically immunoprecipitate roX RNA [5*, 12*].

The MSL Complex Assembles at roX Genes:

Analysis of the roX genes indicates that they may provide the targeting mechanism which delivers the MSL complex to the x chromosome. When roX genes are moved from their normal locations on the X to autosomes, they provide strong MSL target sites at each new location tested [ 17**].

Removal of an MSL subunit through mutation appears to arrest maturation of the complex at a defined step which can be visualized on polytene chromosomes. The five MSL proteins characterized to date can be divided into two groups. MSL1 (a novel protein) and MSL2 (a RING finger protein) seem to form the core of the complex. Removing either one prevents the remaining subunits from binding anywhere along the X [18-20]. By contrast, the MLE (a helicase), MOF (a histone acetyltransferase), or MSL3 (a chromo domain protein) subunits seem more peripheral. Removing any of the last three results in a nonfunctional partial complex bound to ~35 unusual sites scattered along the X [19-22] These sites, termed 'chromatin entry sites' have been postulated to be locations where the MSL subunits assemble into large complexes [17**, 20].

An important insight came with the realization that two of these 35 chromatin entry sites are the roX1 and roX2 genes [17**]. This led to the idea that some MSL proteins bind to the roX genes and may  subsequently capture nascent roX transcripts to begin complex assembly. The molecular nature of the remaining chromatin entry sites is not known. Will they produce additional roX RNAs?

It is also not known whether there are separate MSL complexes containing either roX1 or roX2, or if all complexes contain both RNAs. Examination of partial complexes, however, shows that the roX RNAs are incorporated at different stages of maturation. When MSL3 is removed, roX2 RNA is still transported to the other chromatin entry sites, but roX1 is found only at its site of synthesis [12*]. This indicates that roX2 probably enters the complex before MSL3, and that roX1 enters together with, or after MSL3. When the MLE helicase is removed, not even roX2 RNA is exported from its site of synthesis [12*]. Perhaps the MLE helicase is needed at an early point to help process or fold roX2 RNA into a structure competent to bind the first MSL proteins. Such incomplete complexes would then be transported to other chromatin entry sites for further maturation.

The MSL Complex Spreads from roX Genes into Flanking Chromatin:

Autosomal roX transgenes not only attract MSL complexes to themselves but can also serve as nucleation sites for epigenetic spreading of the MSL complex into flanking chromatin [17**]. The MSL complex can spread several hundred kilobases to >1 Mb in exceptional cases. The affected chromatin is enriched in histone H4 acetylated at lysine 16 suggesting that the segment of the autosome is suffering inappropriate dosage compensation. Depending upon the insertion site, the extent of spreading can be highly variable from cell to cell. Spreading along autosomes is far less consistent than the X. The mechanism of spreading is not undestood but any model must account for the fact that spreading appears discontinuous [17**] and requires enzymatically active MOF histone acetyltransferase [22] and MLE ATPase [23].

Several conclusions can be drawn from this result which change our view of how dosage compensation is targeted to the X in flies. First, individual X-linked genes may not have discrete MSL binding sites. The new model (Fig. 1) postulates that the MSL complex assembles in a stepwise pathway at ~35 chromatin entry sites along the X chromosome where it captures roX RNAs [12*, 17**]. Mature complexes subsequently spread through an unknown mechanism in cis to surrounding genes. In this model, being closely linked (within ~1Mb) to a chromatin entry site will be dosage compensated with little regard to gene-specific  regulatory sequences.

This model predicts that individual X-derived genes should not be recognized by the MSL complex or be dosage compensated when moved to autosomes because they are no longer linked to chromatin entry sites. Attempts to detect MSL complex bound to X-derived transgenes inserted into autosomes have failed for four genes examined in the salivary gland [24, 25*]. However, thee are well documented reports of X-derived genes being at least partially dosage compensated when moved to some autosomal sites [26 - 29]. The MSL complex can spread from one autosome carrying a roX1 transgene to a paired homolog lacking the roX1 gene [17**]. Perhaps such a mechanism allows relocated genes to pair transiently with the homologous region on the X and receive MSL complex. The larger the transgene, the more likely pairing would take place and partial dosage compensation would be observed.

In a process distinct from spreading, MSL complexes appear to exchange freely between different chromatin entry sites, presumably in a soluble form. The most compelling evidence for this is the presence of roX2 RNA in MSL complexes surrounding an autosomal roX1 transgene [17**]. The roX2 RNA must have come from its site of synthesis on the X chromosome, and yet reached an ectopic autosomal site via a roX1 transgene. Likewise, in males in which the only source of roX1 RNA is autosomal, roX1 RNA coats the X in trans as well as a small region surrounding the roX1 transgene [9, 17**]. This behavior would help explain why both roX1 and roX2 coat the entire length of  the X in a precisely coincident pattern rather than non-overlapping small regions surrounding their respective sites of synthesis.

Parallels to Mammalian Xist:

Understanding the role of roX genes in targeting dosage compensation to the X reveals a surprising similarity to the situation in mammals. In both cases, the chromosome targetted for dosage compensation is marked by large noncoding RNAs spreading in cis from their site of synthesis. In Drosophila, this happens to the single X in males and leads to hypertranscription. In mammals, Xist RNA spreads over one of two female X chromosomes inactivating most of the genes along its length. In each animal. the RNA has the notable ability to spread over previously unrecognized chromatin when a roX or Xist transgene is inserted into an autosome[17**, 30-32]. This shows that both RNAs bind chromosomes through a novel mechanism based on linkage to the site of synthesis rather than a strict dependence upon the DNA sequence of the chromosome. We and others (e.g. [17**, 33]) have speculated that repetitive chromosomal sequences, which are enriched on the X compared to the autosomes, may aid RNA spreading.

One fundamental difference between the way roX and Xist RNAs spread is that Xist propagates excusively in cis from a single site of synthesis (Figure 2). This is essential so that Xist RNA does not travel between the two homologs, catstophically silencing both.  By contrast, flies produce at least two distinct species of roX RNAs which appear to spread locally up to ~1 Mb around their sites of synthesis but also seem to reinitiate spreading from multiple chromatin entry sites along he X chromosome. If the site of roX1 synthesis is moved to an autosome, roX1 RNA locates the X in trans and spreads to give a
wild-type RNA distribution along the X [9, 17**]. As roX RNAs operate in cells containing only one X chromosome, they have never been subjected to the same functional constraint which demands mammalian Xist to act exclusively in cis. The ability to 'jump' between chromatin entry sites, even in trans, may provide redundant coverage to ensure the entire X is reliably coated with MSL complex. Alternatively, travel between the different chromatin entry sites may reflect a requirement to gather several distinct roX RNA molecules into a single MSL complex.

Genetic Analysis of Chromosomal RNAs:

Xist knockout mutants have clearly demonstrated that this gene is necessary to establish X silencing in females [34, 35]. Furthermore, autosomal Xist transgenes show that it is sufficient to silence large segments of a chromosome [30, 32, 36, 37*, 38**].  Once silencing has been established, it is epigenetically inherited even if Xist is removed [38**, 39, 40]. The inactive X is late-replicating, deficient in certain acetylated histones [41*], and the DNA is heavily methylated at CpG islands (reviewed in [7]). After Xist RNA inactivates one X, the Barr body becomes enriched for a variant histone - macro H2A [42, 43*]. The biochemical role of Xist remains a mystery but recent work shows that the primary sequence is not critical. Mouse and human Xist show little sequence similarity and yet transgenic XIST RNA is able to coat mouse autosomes and silence genes in a mosaic pattern [44, 45]. Oddly, the silenced autosome is not late replicating and histones are not dramatically hypoacetylated.

Genetic analysis of roX genes has been hampered by two problems. First, mutations in any essential gene on the X result in male lethality, so distinguishing true male-specific lethals relevant to dosage compensation from mutations in unrelated genes is technically challenging. Second, different roX genes may have overlapping functions such that mutating only one at a time may not affect dosage compensation.

RNA null mutations of roX1 have been recovered and found to have no effect on dosage compensation or male viability [ 9, 17**]. Simple mutations in roX2 have not been reprted but a large deletion of the roX2 region was examined in dying embros. This study indicated that removing roX1 or roX2 separately had either little or no effect on assembly of the MSL complex on the male X chromosome but removing both either blocked or significantly delayed MSL complex formation [11*]. If this result can be confirmed with smaller mutations removing just the roX2 gene, it would indicate that these two unrelated RNAs perform overlapping functions vital to dosage compensation.

Regulation of Chromosomal RNA Expression:

Ideas about regulating roX expression are controversial. It is clear that the MSL complex plays some role in regulating sex-specific roX RNA accumulation. Males mutant for one of the msl genes lose roX RNAs. Conversely, females forced to make a full set of MSL proteins, assemble functional MSL complexes and roX RNAs along both their X chromosomes [9, 10]. Do the MSL proteins directly control the synthesis of roX RNA by acting as male-specific transcription factors on the roX genes, or do the MSL proteins merely stabilize RNA which is constitutively transcribed in both sexes? It is clear that the MSL complex is essential for roX RNA stability [12*]. However, the question of transcriptional regulation remains unresolved. Although MSL proteins associate with roX exons in vivo, it has not been rigorously shown whether the MSL proteins recognize nascent transcripts, genomic DNA, or both [17**]. Even if the target is roX DNA, it is not clear whether this affects transcription initiation or is a step in complex assembly or spreading into flanking chromatin.

The initial characterization of the roX1 expression pattern showed that RNA accumulates to similar levels in both sexes from cellular blastoderm through germ-band extension. The RNA is subsequently lost from females but remains high in males [9]. More recent work reports excusively male-specific expression throughout embryonic development [17**]. It is not clear if this discrepancy is caused by some technical difficulty or a real difference in the stocks examined. An early burst of female expression would indicate that roX1 is not strictly dependent upon the MSL complex for transcription.

Early regulation of Xist expression requires counting the X chromosomes in a cell, choosing one as active X, and silencing the remaining X chromosome(s) with Xist RNA. These activities are contained within an 80 kb transgene [32], but the regulatory mechanisms at work are contentious areas of research. There is general agreement that murine Xist undergoes a transition during differentiation between early, low level expression over both Xist alleles to high Xist RNA accumulation over only the inactive X. Some authors argue that this reflects selective stabilization of Xist RNA [46, 47]. One school of thought postulates that Xist expression is controlled from two distinct promoters [48] (Figure 3). 

Figure 3. Models of early Xist regulation. Only female cells are illustrated. The thick horizontal lines represent Xist alleles, with straight arrows representing transcripts. RNA in undifferentiated cells is found at low levels only over the site of synthesis and may turn over rapidly.

(a). The promoter switch hypothesis states that early, unstable Xist transcripts are derived from P₀, and stable transcripts are made from P₁ following differentiation [48].

(b). The antisense model proposes that Tsix is expressed early and somehow blocks Xist RNA accumulation but later shuts off, allowing Xist RNA to spread over the inactive X [49**, 52**].

(c). The transcription model postulates that Xist RNA may be synthesized below a critical threshold (thin arrow) in undifferentiated cells such that it rapidly turns over. Differentiation may trigger an increase in Xist transcription (thick arrow) so that there is enough RNA to spread along the chromosome (curved arrows in each model). Precocious RNA spreading and gene silencing are seen in undifferentiated cells when large amounts of Xist RNA are synthesized from an inducible promoter [38**]. Xa, active X; Xi, inactive X.

The P₁ promoter is activated on one allele in response to differentiation and makes a stable RNA capable of spreading along the chromosome. The P₀ promoter lies ~6 kb upstream and produces an unstable transcript from both alleles in undifferentiated cells. In this model, choosing which X will become inactive is determined largely by which allele of Xist successfully makes the transition from unstable P₀ transcripts to stable P₁ transcripts during a small window of developmental time.

This view has been challenged by an alternative proposal that Xist is negatively regulated by an antisense Tsix transcript in differentiating cells [49**, 50]. Recent results question the existence of the early P₀ promoter in part because a pseudogene of ribosomal protein S12 was found at that site, and all RNA between P₁ and P₀ mapped to the antisense strand [51*]. An Xist transgene controlled by only P₁ was able to produce unstable RNA prior to but not after differentiation [51*]. This indicates that P₀ is not essential early but important regulatory sequences do map 5' of P₁. When only the promoter driving antisense Tsix expression is deleted, Xist is constitutively expressed from the mutant X in female cells, arguing that Tsix is a pivotal control factor [52**]. The Tsix promoter mutant did not trigger stable Xist expression (and subsequent lethal X inactivation) in male cells, however, indicating that the mutant retained the ability to sense that only one X was present. A 65 kb deletion 3' of Xist was reported to cause consitutive expression of Xist in cells with only one X indicating that additional regulatory sequences may be defined by the larger deletion [53]. This deletion may have removed ~3 kb from the last Xist exon, however, complicating the interpretation [54*]. If the Tsix antisense model withstands further experimental tests, it will reveal a remarkable approach to regulating a new class of RNAs.

Recent studies using an inducible Xist cDNA transgene have prompted a third model of Xist regulation postulating that X inactivation is triggered by a transcriptional upregulation of an Xist promoter ]38**]. Remarkably, robust transcription of an Xist cDNA can lead to reversible silencing of flanking genes in undifferentiated cells. In this rapidly developing field, it is possible that elements of several different models will prove correct. Perhaps Xist is controlled both at the level of transcription initiation and RNA stability, either of which could be strongly affected by by antisense transcription of Tsix.

Conclusions:

Knowledge of the biochemical functions of RNAs has expanded dramatically since the days of mRNA, rRNA, and tRNA. RNAs are now known to play critical roles in splicing and editing mRNAs, and modifying rRNA. They are also components of telomerases and signal recognition particles involved in protein secretion. To this list we must now consider unusual chromosomal RNAs which coat the X chromosome undergoing dosage compensation. The genes producing these RNAs reside on the chromosome upon which they act. These RNAs spread extensively in cis, remodeling the chromatin as they go to alter gene expression on the scale of an entire chromosome. As whole genome sequences are compared to sets of cDNAs, it will be interesting to look for RNAs lacking open reading frames which may be unappreciated components of other chromatin-remodeling machines. 

We would like to thank Barbara Panning, Hubert Amrein, and Carsten Stuckenholz for helpful comments on the manuscript. We apologize to those whose work was not cited due to limitations on the space and scope of the review.

References and Recommended Reading:
Papers of particular interest, published within the annual period of review, have been highlighted as:

* of special interest
** of outstanding interest

1. Cline TW, Meyer BJ, "Vive la difference: males vs. females in flies vs. worms." Annu. Rev. Genet (1996), 30: 637-702.

2. Lucchesi JC, "Dosage Compensation in flies and worms: the ups and downs of X-chromosome regulation", Curr. Opin. Genet. Dev. (1998), 8: 179-184.

3. Stuckenholz C, Kageyama Y, Kuroda MI, "Guilt by association: non-coding RNAs, chromosome-specific proteins, and dosage compensation in Drosophila", Trends Genet. (1999), 15: 454-458.

4. Meller VH, "Dosage Compensation: Making 1X equal 2X", Trends Cell Biol. (2000), 10: 54-59.

5. *  Smith ER, Pannuti A, Gu W, Steurnagel A, Cook RG, Allis CD, Lucchesi JC, "The Drosophila MSL complex acetylates histone H4 at lysine 16, a chromatin modification linked to dosage compensation", Mol. Cell Biol. (2000), 20: 312-318.

The MOF (males absent on first) acetyltransferase is a component of the MSL complex and here is shown to acetylate histone H4 at lysine 16 in vitro, exactly the same modification associated with the male X. This, along with previous genetic analysis, implicates MOF as a central player in remodeling chromatin on the male X. Upon partial purification of the MSL complex, the authors also show that roX2 RNA is a component.

6. * Akhtar A, Becker PB, "Activation of transcription through histone H4 acetylation by MOF, an acetyltransferase essential for dosage compensation in Drosophila".

MOF is shown to acetylate histone H4 at lysine 16 in vitro. When recombinant MOF is added to an in vitro transcription system, the resulting acetylation overcomes the general repressive effect of nucleosomes, allowing robust transcription.

7. Heard E, Clerc P, Avner P, "X-chromosome inactivation in mammals", Annu. Rev. Genet. (1997), 31: 571-610.

8. Lee JT, Jaenisch R, "The (epi)genetic control of mammalian X-chromosome inactivation", Curr. Opin. Genet. Dev. (1997), 7: 274-280.

9. Meller VH, Wu KH, Kuroda MI, Davis RL, "roX1 RNA paints the X chromosome of male Drosophila and is regulated by the dosage compensation system", Cell (1997), 88: 445-457.

10. Amrein H, Axel R, "Genes expressed in neurons of adult male Drosophila", Cell (1997), 88: 459-469.

11. ** Franke A, Baker BS, "The roX1 and roX2 RNAs are essential components of the compensasome, which mediates dosage compensation in Drosophila", Mol. Cell (1999), 4: 117-122.

This paper shows that roX2 RNA coats the male X in the same banded pattern as roX1, both of which are precisely colocalized with the MSL proteins. They also find that male embryos carrying a large deletion of the roX2 region of the X and a mutation in roX1 fail to assemble MSL complex before dying. Either mutant alone forms a complex, suggesting that the two dissimilar RNAs somehow perform overlapping functions.

12. * Meller VH, Gordadze PR, Park Y, Chu X, Stuckenholz C, Kelley RL, Kuroda MI, "Ordered assembly of roX RNAs into MSL complexes on the dosage compensated X chromosome in Drosophila", Curr. Biol. (2000), 10: 136-143.

The MSL proteins block destruction of roX RNAs, suggesting that they are physically bound in vivo. Both roX RNAs precisely colocalize with the MSL proteins at hundreds of bands along the male X, and the roX2 RNA can be immunoprecipitated along with MSL proteins. roX1 and roX2 RNAs were found to be incorporated into the MSL complex at different points of assembly.

13. Brockdorff N, Ashworth A, Kay GF, McCabe VM, Norris DP, Cooper PJ, Swift S, Rastan S, "The product of the mouse Xist gene is a 15 kb inactive X-specific transcript containing no conserved ORF and located in the nucleus", Cell (1992), 71: 515-526.

14. Brown CJ, Hendrich BD, Rupert JL, Lafreniere RG, Xing Y, Lawrence J, Willard HF, "The human Xist gene: analysis of a 17 kb inactive X-specific RNA that contains conserved repeats and is highly localized within the nucleus", Cell (1992), 71: 527-542.

15. Brockdorff N, Ashworth A, Kay GF, Cooper P, Smith S, McCabe VM, Norris DP, Penny GD, Patel D, Rastan S, "Conservation of position and exclusive expression of mouse Xist from the inactive X chromosome", Nature (1991), 351: 329-331.

16. Borsani G, Tonlorenzi R, Simmler MC, Dandolo L, Arnaud D, Capra V, Grompe M, Pizzuti A, Muzny D, Lawrence C, et al, "Characterization of a murine gene expressed from the inactive X chromosome", Nature (1991), 351: 325-329.

17. ** Kelley RL, Meller VH, Gordadze PR, Roman G, Davis RL, Kuroda MI, "Epigenetic spreading of the Drosophila dosage compensation complex from roX genes into flanking chromatin", Cell (1999), 98: 513-522.

In this paper, we propose a new model in which the MSL complex first assembles at chromatin entry sites and then spreads in cis to coat the male X. The model rests on the finding that two of the ~35 chromatin entry sites are roX genes, and that the MSL complex will spread in cis from a roX transgene into flanking autosomal chromatin. In such transgenic animals, the MSL complex is also found to move in trans  between chromatin entry sites on the X and autosomes.

18. Gorman M, Kuroda MI, Baker BS, "Regulation of the sex-specific binding of the maleness dosage compensation protein to the male X chromosome in Drosophila", Cell (1993), 72: 39-49.

19. Palmer MJ, Richman R, Richter L, Kuroda MI, "Sex-specific regulation of the male-specific-lethal-1 dosage compensation gene in Drosophila", Genes Dev. (1994), 8: 698-706.

20. Lyman LM, Copps K, Rastelli L, Kelley RL, Kuroda MI, "Drosophila male-specific-lethal-2 protein: structure/function analysis and dependence on MSL-1 for chromosome association", Genetics (1997), 147: 1743-1753.

21. Gorman M, Franke A, Baker BS, "Molecular characterization of the male-specific-lethal-3 gene and investigation of the regulation of dosage compensation in Drosophila", Development (1995), 121: 463-475.

22. Gu W, Szauter P, Lucchesi JC, "Targeting of MOF, a putative histone acetyl transferase, to the X chromosome of Drosophila", Dev. Genetics (1998), 22: 56-64.

23. Lee C-G, Chang KA, Kuroda MI, Hurwitz J, "The NTPase/helicase activities of Drosophila Maleless, an essential factor in dosage compensation", EMBO J. )1997) 16: 2671-2681.

24. Bone JR, "Dosage Compensation of the Drosophila X chromosome (PhD Thesis)", Houston, TX: Baylor College of Medicine; 1996.

25. * Bhadra U, Pal-Bhadra M, Birchler JA, "Role of the male specific lethal (msl) genes in modifying the effects of sex chromosomal dosage in Drosophila", Genetics (1999), 152: 249-268.

These authors favor the inverse theory of dosage compensation which postulates that the primary function of the MSL complex is to sequester other positive regulatory factors away from the autosomes in males. They also show that very large (>100kb) X--autosome transpositions bind MSL complex but small transgenes carrying only the white gene do not. They propose that te MSL complex binds the X through nucleation sites present on the large fragments but absent on the smaller transgenes.

26. Gutierrez AG, Christensen AC, Manning JE, Lucchesi JC, "Cloning and dosage compensation of the the 6-phosphogluconate dehydrogenase gene (Pgd) of Drosophila melanogaster", Dev. Genet. (1989), 10: 155-161.

27. Cooper MK, Hamblen-Coyle MJ, Liu X, Rutila JE, Hall JC, "Dosage compensation of the period gene of Drosophila melanogaster", Genetics (1994), 138: 721-732.

28. Qian S, Pirrotta V, "Dosage compensation of the Drosophila white gene requires both the X chromosome environment and multiple intragenic elements", Genetics (1995), 139: 733-744.

29. Roseman RR, Swan JM, Geyer PK, "A Drosophila insulator protein facilitates dosage compensation of the X chromosome mini-white gene located at autosomal insertion sites", Development (1995), 121: 3573-3582.

30. Lee JT, Jaenisch R, "Long-range cis effects of ectopic X-inactivation centres on a mouse autosome", Nature (1997), 386: 275-279.

31. Herzing LBK, Romer JT, Horn JM, Ashworth A, "Xist has properties of the X-chromosome inactivation centre", Nature (1997), 386: 272-275.

32. Lee T, Lu N, and Han Y, "Genetic analysis of the mouse X inactivation center defines an 80-kb multifunction domain", Proc. Natl. Acad. Sci. USA (1999), 96: 3836-3841.

33. Lyon MF, "X-chromosome inactivation: a repeat hypothesis", Cytogenet. Cell Genet. (1998), 80: 133-137.

34. Penny GD, Kay GF, Sheardown  SA, Rastan S, Brockdorff N, "Requirement for Xist in X chromosome inactivation", nature (1996), 379: 131-137.

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

36. Lee JT, Strauss WM, Dausman JA, Jaenisch R, "A 450 kb transgene displays properties of the mammalian X-inactivation center", Cell (1996), 86: 83-94.

37. * Heard E, Mongelard F. Amaud D, Avner P, "Xist yeast artificial chromosome transgenes function as X-inactivation centers only in multicopy arrays and not as single copies", Mol. Cell Biol. (1999), 19: 3156-3166.

Analysis of several transgenics shows that strong activity is only found in lines where multicopy arrays of  Xist genes are inserted into a chromosome. Lines containing only a sigle copy of an Xist transgene could not silence genes in cis or be counted as a second X. There are several reasonable interpretations but one possibility is appealing in light of the findings of Wutz and Jaenisch [38 **]. Perhaps multicopy arrays supply enough Xist RNA at one site to consistently spread over the chromosome. Because of unpredictable position effects, single-copy transgenes may often fail to produce the minimum threshold RNA level needed to trigger silencing.

38. ** Wutz A, Jaenisch R, "A shift from reversible to irreversible X inactivation is triggered during ES cell differentiation", Mol. Cell (2000), 5: 695-705.

The authors were able to turn on high levels of Xist expression at any time in male embryonic stem cells using an Xist cDNA driven by an inducible promoter. This RNA should have approximately the same 5' end as endogenous RNA transcribed from the P₁ promoter but displayed several properties not observed before.  The RNA was stable in undifferentiated cells and spread over flanking chromatin. Not only did the Xist RNA silence flanking genes in cis by spreading from the transgene insertion site but when undifferentiated cells were tested, Xist-mediated silencing was not associated with histone hypoacetylation or late DNA replication. Significantly, silencing could be reversed by withdrawing Xist RNA prior to differentiation, but became irreversible following differentiation. Perhaps early Xist RNA can directly repress flanking gene transcription by a chromatin-independent mechanism, which is subsequently preserved by a heritable epigenetic alteration in chromatin architecture. The authors suggest that their inducible promoter mimics a strong transcriptional upregulation which may normally occur at one Xist allele during differentiation.

39. Brown CJ, Willard HF, "The human X-inactivation centre is not required for maintenance of X-chromosome inactivation", Nature (1994), 368: 154-156.

40. Czankovszki G, Panning B, Bates B, Pehrson JR, Janisch R, "Conditional deletion of Xist disrupts histone macroH2A localization but not maintenance of X inactivation", Nature Genet. (1999), 22: 323-324.

41. * Gilbert SL, Sharp PA, "Promoter-specific hypoacetylation of X inactivated genes", Proc. Natl. Acad. Sci. USA, (1999), 96: 13825-13830.

Previous work using immunocytochemistry of metaphase chromosomes indicated that histones on the inactive X are hypoacetylated on a global scale. These authors use antibodies against acetylated histone H4 to immunoprecipitate acetylated interphase chromatin and then amplify different regions of X-linked genes by PCR. The genes from the active and inactive X had similar acetylation profiles except near the 5' ends, where genes from the inactive X were significantly hypoacetylated relative to genes from the active X. The locations of the hypoacetylated promoters correlate well with the location of heavily methylated CpG islands.

42. Costanzi C, Pehrson JR, "Histone macroH2A1 is concentrated in the inactive X chromosome in female mammals", Nature (1998), 393: 599-601.

43. * Mermoud JE, Costanzi C, Pehrson JR, Brockdorff N, "Histone macroH2A1.2 relocates to the inactive X chromosome after initiation and propagation of X-inactivation", J. Cell Biol. (1999), 147: 1399-1408.

An unusual histone variant, histone H2A1.2, had previously been reported to colocalize with Xist RNA on the inactive X in somatic female cells, suggesting that it might play a role in forming heterochromatin. In this work, XX and XY ES cells were stained with antibodies to macro H2A1.2 over the course of differentiation. Undifferentiated cells contained a discrete macroH2A1.2 domain which did not colocalize to an X. By the end of differentiation, this was lost in males but somehow came to reside over the inactive X in female cells. The late arrival of macroH2A1.2 suggests that it plays either little or no role in initiating or propagating X silencing in ES cells.

44. Migeon BR, Kazi E, Haisley-Royster C, Hu J, Reeves R, Call L, Lawter A, Moore CS, Morrison H, Jeppesen P, "Human X inactivation center induces random X chromosome inactivation in male transgenic mice", Genomics (1999), 59: 113-121.

45.  Heard E, Mongelard F, Arnaud D, Chureau C, Voruc'h C, Avner P, "Human XIST yeast artificial chromosome transgenes show partial X inactivation center function in mouse embryonic stem cells", Proc. Natl. Acad. Sci. USA (1999), 96: 6841-6846.

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

47. Sheardown SA, Duthie SM, Johnston CM, Newall AET, Formstone EJ, Arkell RM, Nesterova TB, Alghisi GC, Rastan S, Brockdorff N, "Stabilization of Xist RNA mediates initiation of X chromosome inactivation", Cell (1997), 91: 99-107.

48. Johnston CM, Nesterova TB, Formstone EJ, Newall AE, Duthie SM, Sheardown SA, Brockdorff N, "Developmentally regulated Xist promoter switch mediates initiation of X inactivation", Cell (1998), 94: 809-817.

49. ** Lee JT, Davidow LS, Warshawsky D, "Tsix, a gene antisense to Xist at the x-inactivation centre", Nature Genet. (1999), 21: 400-404.

This paper reports the discovery of the Tsix gene, which is antisense to Xist and extends beyond Xist at both the 5' and 3' ends. The Tsix transcript is produced at both alleles in undifferentiated female ES cells but is lost from one chromosome (the prospective X chromosome) during differentition. Tsix transcripts are later lost from the second (active) X as well.

50. Mise N, Goto Y, Nakajima N, Takagi N, "Molecular cloning of antisense transcripts of the mouse Xist gene", Biochem. Biophys. Res. Commun. (1999), 258: 537-541.

51. * Warshawsky D, Stavropoulos N, Lee TJ, "Further examination of the Xist promoter-switch hypothesis on X inactivation: evidence against the existence and function of a P₀ promoter", Proc. Natl. Acad. Sci. USA, (1999), 96: 14424-14429.

A pseudogene of ribosomal protein S12 (rpS12) is discovered at the reported location of P₀, and RT-PCR P₀ primers amplify rpS12 sequences derived from elsewhere in the genome. In this study, only antisense transcripts are detected between P₀ and P₁. The P₁ promoter is active in undifferentiated ES cells and produces an unstable transcript. Earlier work from the same lab [32] showed that this transgene lacking sequences upstream of P₁ was not able to make stable RNA following differentiation. This indicates that 5' sequences do play some role in normal Xist expression.

52. ** Lee TJ, Lu N, "Targeted mutagenesis of Tsix leads to nonrandom X inactivation", Cell (1999), 99: 47-57.

If Tsix transcripts somehow negatively regulate Xist, then removing Tsix should de-repress Xist expression. Deletion of the Tsix promoter leads to constitutive Xist expression and inactivation of the mutant X in differentiating XX cells, consistent with the antisense model. One curious finding is that the Tsix promoter deletion does not result in constitutive, high levels of Xist RNA accumulation in undifferentiated cells (compare to [38*]). The authors postulate that additional regulatory factors are needed.

53. Clerc P, Avner P, "Role of the region 3' to Xist exon 6 in the counting process of X-chromosome inactivation", Nature Genet. (1998), 19: 249-253.

54. * Hong Y-K, Ontiveros SD, Chen C, Strauss WM, "A new structure for the murine Xist gene and its relationship to chromosome choice/counting during X-chromosome inactivation", Proc. Natl. Acad. Sci. USA (1999) 96: 6829-6834.

The authors re-evaluate the structure of the murine Xist locus with particular attention to the 3' end. They find a small change in the intron/exon structure and a series of potential polyadenylation sites. One important finding is that the 65 kb deletion of Clerc and Avner [53] removes not only 3' flanking sequences but ~3 kb of the last Xist exon.

55. Miller AP, Willard HF, "Chromosomal basis of X chromosome inactivation: identification of a multigene domain in Xp11.21-p11.22 that escapes X inactivation", Proc. Natl. Acad. Sci. USA (1998), 95: 8709-8714.

1. "Nuclear Polyanions as De-Repressors of Synthesis of Ribonucleic Acid".

2. "Mated Models of Gene Regulation in Eukaryotes".

3. "Riboregulation in E. coli: DsrA RNA Acts by RNA:RNA Interactions at Multiple Loci".

4. "Oncogenes as Molecular Targets within Active Chromatin".

5. "A Trans-Acting RNA as a Control Switch in Escherichia coli: DsrA Modulates Function by Forming Alternative Structures".

6. "Switching On and Off with RNA".