- Email: [email protected]
Dosage compensation in flies: Mechanism, models, mystery
FEBS Letters 579 (2005) 3258–3263
FEBS 29483
Minireview
Dosage compensation in flies: Mechanism, models, mystery Tobias Straub1, Ina K. Dahlsveen1, Peter B. Becker* Adolf-Butenandt-Institut, Molekularbiologie, Schillerstrasse 44, 80336 Mu¨nchen, Germany Accepted 26 March 2005 Available online 2 April 2005 Edited by Horst Feldmann
Abstract Dosage compensation involves fine-tuning of gene expression at the level of entire chromosomes. The principles that assure selective targeting of the male X chromosome in Drosophila and the mechanism by which transcription levels are adjusted in a twofold range are still mysterious. We discuss the prevalent models in the context of recent experimental observations. Ó 2005 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. Keywords: Transcription; Dosage compensation; Chromosome; Chromatin; Histone acetylation
1. Dosage compensation, a chromosome-wide challenge Sexual dimorphism is brought about by different sex chromosomes, which are unequally distributed between the sexes. For example in humans and fruit flies females carry two X chromosomes, but males have only a single X chromosome in addition to the gene-poor Y chromosome. Hence, male cells contain a halved dose of all X-linked genes. The level of Xlinked gene products, however, must be similar in cells of either sex. A prime principle of gene dosage compensation is the chromosome-wide adjustment of transcription. The mechanisms by which this may be achieved vary largely between different organisms and involve different molecular factors. Whereas female mammals inactivate one of their two X chromosomes in order to approximate the gene dosage of males, C. elegans hermaphrodites (XX) reduce transcription of each X by about half. In Drosophila, the male flies increase transcription from their single X chromosome about twofold to equal the expression of the two X chromosomes in females (for extensive reviews see [1–5]). This adjustment is essential for male survival. Dosage compensation in flies relies on the integrity of a ribonucleoprotein complex, known as the dosage compensation complex (DCC). This complex comprises the male-specific lethal (MSL) proteins and two non-coding RNA molecules, roX1 and roX2. Male-specific expression of
* Corresponding author. Fax: +49 89 5996 425. E-mail address: [email protected] (P.B. Becker). 1
These authors contributed equally to this work.
Abbreviations: DCC, dosage compensation complex; MSL, male-specific lethal; H4K16, lysine 16 of histone H4
MSL2 leads to enrichment of the DCC on the single X chromosome. The cooperation of structural and enzymatic complex components is necessary for proper function and targeting. So far, the catalytic activities known to be essential for the compensation process are the ATPase/helicase activity of MLE and the histone acetyl transferase activity of MOF. Accumulation of the DCC on the X chromosome results in increased acetylation of lysine 16 on histone H4. This specific chromatin modification is thought to be at least in part responsible for elevated X-linked gene expression. The mechanisms that assure that a fairly global process acting at the level of an entire chromosome leads to rather subtle (twofold) tuning of transcription are still mysterious. However, integrating the limited experimental observations into models has been instructive. Not surprisingly, different kinds of models have been proposed trying to explain the mode of precise transcription tuning and to describe how the X chromosome gets exclusively furnished with the DCC. Experimental progress now allows for the critical evaluation of these models.
2. Opposing models for dosage compensation The prevailing model (Fig. 1, model A) for the compensation of the X chromosomal genes in males states that the DCC is directly involved in activating gene expression on the male X chromosome. It is based on the following observations: (a) in Drosophila males, the MSL proteins are almost exclusively located on the X chromosome [4]; (b) H4K16 acetylation follows the distribution of the MSLs [6]; (c) acetylation of H4K16 can activate transcription in a chromatin context [7]; (d) the male X chromosome has a more diffuse appearance, reminiscent of reduced chromatin condensation [1]; (e) knock-down of any of the MSL proteins in males reduces expression of X chromosomal genes relative to that of autosomal genes and leads to loss of compensation and male lethality [8,9]. Accordingly, in Drosophila males MSL2 stabilizes and tethers the DCC to the X chromosome. Concomitant enrichment of H4K16 acetylation results in local chromatin opening and a twofold increase in transcription, conceivably at the level of elongation [10]. Reflecting its role as a direct modulator of gene expression, the model implies a flexible and direct association of the DCC with active genomic loci [11]. Although the model is straightforward and intuitive, not all phenomena can be smoothly incorporated and key aspects have not yet been experimentally verified. For example, all MSL proteins (with the exception of MSL2) are present in female cells as well, albeit some of them at reduced levels, where
0014-5793/$30.00 Ó 2005 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2005.03.050
T. Straub et al. / FEBS Letters 579 (2005) 3258–3263
XX
3259
XY mle-
XY
model A
model B
X chromosome and gene product
autosome and gene product
dosage compensation complex
Fig. 1. Different models for how compensation of the X-linked genes is achieved in order to equalize gene doses between male (XY) and female (XX) flies. Model A suggests male-specific assembly of the dosage compensation complex (DCC) that directly activates gene expression on the X. Model B proposes that compensation of the X is accomplished by a genome-wide increase in transcription that is due to activator imbalance inferred by X chromosomal monosomy. Tethering of the activators (members of the DCC) onto the male X alleviates the transcription enhancement on the autosomes. While the expression status in wild type males is the same for both models, MLE mutants (XY mle-) uncover major differences characterizing the two hypotheses.
they may affect chromatin structure and function globally. An unexplained issue is how the targeted H4K16 acetylation will bring about a general and precise twofold elevation of transcription [10]. Increased H4K16 acetylation might induce a chromosome-wide chromatin decondensation leading to an elevated transcription in the proper range of magnitude. Unfortunately, there is no experimental evidence for such an effect. In vitro, MOF-mediated histone acetylation has been shown to relieve chromatin-mediated transcriptional repression [7]. However, the activation observed was much more than twofold. A precise doubling of transcription is likely a result of a fine balance between activation and feedback repression and thus it is possible that DCC harbours repressive activities in addition to the activating histone acetylation function, activities that are not yet accounted for by model A. In general, histone acetylation is involved in gene activation in all eukaryotes from yeast to man. Surprisingly, acetylation of H4K16 is an exception to this rule, since in yeast, where its occurrence has been correlated with gene activity on a genome-wide scale, it was shown to mark inactive, rather than active loci [12]. Accordingly, H4K16ac has been postulated to prevent binding of transcription factors in silent chromatin regions in yeast. While we appreciate that the language of histone modifications is modulated by species-specific dialects [13], the case of yeast chromatin suggests complex effects of H4K16 acetylation, which may impact on model building for dosage compensation in Drosophila. The most important objection to model A is, however, the lack of absolute quantification of gene expression confirming the direct involvement of the DCC in compensation of Xlinked genes.
Birchler and colleagues have proposed an alternative and less intuitive model (Fig. 1, model B) that employs a genome-wide mechanism underlying the compensation of the X chromosome [14]. Central to this model is the inverse dosage effect, which can be observed when large chromosomal segments are rendered aneuploid. Such deletions have been observed to lead to a global increase in gene expression. This may be explained by a preferential loss of transcriptional repressors given a higher prevalence of those factors. Thus, deletion of a chromosomal segment will offset the balance between global positive and negative regulators, resulting in a dominance of positive factors. According to Birchler, the X chromosomal monosomy in males is a special case of an aneuploidy that did not arise by abrupt deletion, but by the gradual process of diversification of sex chromosomes. The evolutionary time scales to be considered provided ample opportunity for co-evolution of compensatory principles that re-establish the balance of gene expression. According to the model, loss of one X chromosome in males leads to a global inverse regulation of gene expression, i.e. to hyperactivation of all autosomes and the remaining X chromosome due to a relative preponderance of positive transcription regulators [14]. In Drosophila metafemales harbouring three X chromosomes, the same principle would compensate the increase in X chromosomal dosis by a general reduction of gene expression to 2/3 [14]. Although the inverse dosage effect compensated for the aneuploidy of the X chromosome, it obviously would not allow male survival, since all autosomal gene expression is doubled relative to that in females (Fig. 1B, centre panel). In order to counteract autosomal hyperactivation, Birchler proposes that global activators of transcription, con-
3260
ceivably MOF, are removed from the autosomes by an MSL2/MLE-mediated sequestration to the X chromosome (Fig. 1B, right panel). Consequently, MOF-mediated histone acetylation is enriched on the X and gene expression on the autosomes drops. Now, in order to prevent excessive activation of X chromosomal target genes, repressive mechanisms are required to limit gene activation to the twofold effect, brought about by the inverse dosage effect. Such negative effectors my reside in the MSL complex itself [14] or involve other factors, such as the chromatin remodeling ATPase ISWI [15,16]. Apparently, one of the fundamental differences between the models is stipulated by the role of MSL2: whereas in model A MSL2 and the DCC it organises are directly involved in activation, in model B dosage compensation is brought about by the inverse dosage effect, which may have no direct mechanistic connection to the MSL proteins. Here, MSL2 may function in two distinct ways: first, to tether global autosomal activators (the other MSL proteins) to the X chromosome in order to counteract autosomal hyperactivation, and second, to prevent excessive activation by increased H4K16 acetylation on the X. Accordingly, the presence of DCC on the X chromosome might not necessarily correlate with activated transcription. A prediction of model B is that disruption of MSL2 or MLE would result in a release of activators and an increased autosomal expression. Under these circumstances X chromosomal expression would remain unchanged due to the retained inverse dosage effect. The strongest support for this model stems from experiments employing absolute quantification of gene expression [14,16]. These measurements demonstrated, for instance, the predicted absolute increase in autosomal gene expression upon knockdown of MLE for a number of genes (Fig. 1, compare the two models in the XY mle-background) [17]. However, the global nature of the inverse dosage effect appears to be less robust. Although some genes respond as predicted in the experiments performed by Birchler and coworkers [18], a significant number of them do not. MSL mutants that should re-establish the inverse dosage effect on the autosomes by releasing the activators exhibit a rather heterogeneous expression response of autosomal genes. Although highly suggested by the model, MOF does not appear to contribute to the compensation of the X by the inverse dosage effect: in MOF mutants X chromosomal gene expression remained unchanged [16]. At this point it remains unclear why MOF is recruited to the X in order to relieve the inverse dosage effect on the autosomes and, most importantly, why a mof mutation is lethal for males only. An obvious drawback of the model is, therefore, that the inverse dosage effect it employs to explain dosage compensation remains even more mysterious than the phenomenon it tried to explain. OckhamÕs razor would find ample opportunities for trimming. Absolute quantification of gene expression as applied by Birchler and colleagues is an important step towards defining the actual compensation process and testing models. However, care has to be taken regarding the specimen analyzed. The MSL mutant phenotype of embryos or larvae could be compromised to various extend by maternal contribution of MSL proteins and by secondary effects on gene expression in dying larvae. Crucial for the validation of any model will be to define the functional contribution of H4K16 acetyla-
T. Straub et al. / FEBS Letters 579 (2005) 3258–3263
tion to the compensation process as the simple rule ‘‘histone acetylation equals gene activation’’ turns out to be obsolete [12].
3. Recruitment of DCC to the X chromosome: Models in transition The two contrasting models for how the DCC functions in dosage compensation discussed above have different implications for the importance of the targeting and distribution of the complex. Model A implies that the correct distribution of the DCC is vital for its function in up-regulating transcription, whereas in model B the precise location of the complex may not be so important. A longstanding model for how localisation of the DCC to the X is achieved involves primary recognition of specific sites (Ôentry sitesÕ) followed by distribution along the chromosome by ÔspreadingÕ in cis from these sites to less defined chromatin [19]. Recent observations have challenged this Ôentry-siteÕ model and led to the proposition of an alternative model involving a larger number of specific sites of varying affinities [4,20,21]. Both models are based on observations using salivary gland polytene chromosomes where the distribution of the DCC can be easily visualised by immunostaining (Fig. 2, top pictures). In wild type male nuclei, the complex binds mainly to interbands in a characteristic pattern [23]. The Ôentry siteÕ model was originally based on the characterisation of the two first specific binding sites for the DCC found within the roX1 and roX2 genes. The roX genes are part of a relatively small set of DCC binding sites that is observed in females expressing MSL2 and mutant for msl3, mle or mof [4]. This very visually suggested the presence of a special type of binding site, proposed to be the initial sites of recruitment, from which the complex would subsequently spread to cover the chromosome. Accordingly, these site were termed Ôchromatin entry sitesÕ (CES) [19]. Detailed studies of the roX1 and roX2 sites provided convincing evidence in favour of the model [22–24]. When the roX genes were inserted on autosomes via P-element transformation, the DCC was recruited to the insertion site. Importantly, the DCC was occasionally seen to ÔspreadÕ to neighbouring, autosomal interbands in a pattern reminiscent of that seen on the X, suggesting that this recruitment and spreading mirrored the real mechanism for targeting and distribution of the DCC. The Ôentry site modelÕ (illustrated in Fig. 2a) therefore predicts that the CES are the only specific recruitment sites for the DCC. Binding to an entry site would be a prerequisite for the occupancy of secondary, non-specific sites in the vicinity, a process that has been referred to as Ôepigenetic spreadingÕ [5,23]. It has been proposed that these secondary sites may represent sites of active transcription where the complex is required to implement dosage compensation. Using transposable elements on the X that contain UAS sequences, it has been demonstrated that the DCC can be recruited to these elements upon activation of transcription by Gal4 [11]. Two aspects of the Ôentry siteÕ model have been questioned by recent results. Firstly, on several occasions insertion of X chromosomal sequences lacking a CES still led to recruitment of the DCC to autosomes in wild type males ([21], IKD unpublished). In addition, several large translocations from the X to
T. Straub et al. / FEBS Letters 579 (2005) 3258–3263
3261
Fig. 2. Sequential occupancy of the X chromosome might employ different ways of complex distribution. Depicted are two Drosophila polytene chromosome stainings for MSL1 at low (upper panel) or high (lower panel) concentration of complex. The way the complexes target new ÔbandsÕ (long-range distribution, a–c) and the chromosome binding within a ÔbandÕ (short-range distribution, d–f) are controversially discussed.
autosomes, also lacking a CES, have been shown to recruit the DCC [20]. These observations demonstrate that DCC recruitment is not restricted to a small number of dedicated Ôentry sitesÕ. The second aspect of the Ôentry siteÕ model that has been challenged by recent data is whether spreading actually occurs as part of the mechanism for distribution of the DCC over the X chromosome. Extensive spreading from autosomal roX genes, which depended on the transcription of the roX RNA, the abundance of DCC [27] and was particularly evident if the number of roX genes was reduced [25,26], turns out to be the exception rather than the rule. The third CES to be described, 18D, was found to induce only minimal spreading in rare cases [21]. Importantly, when parts of autosomes get
translocated to the X, spreading of the DCC into autosomal sequences does not occur, even if the translocations lie close to entry sites, including roX genes [20,21]. This suggests that spreading is not the prevailing mechanism for DCC distribution and that the X chromosome harbours recognition elements in addition to the known entry sites. The observations described above have led to a revised model for the targeting and distribution of the DCC, which we refer to as the ÔaffinitiesÕ model (Fig. 2b) [4,20,21]. This model assumes that most, if not all, binding sites for the DCC are specific, but of varying affinities for the DCC. Accordingly, one major determinant of recruitment is the concentration of intact DCC in a nucleus. Studies using females expressing varying levels of MSL2 support this prediction as when complex levels
3262
are limiting the DCC binds only to the sites that were observed in the msl mutant backgrounds [27]. In this scenario, the entry sites are not special recruitment sites, but merely represent the highest affinity sites. Partial and/or complete complexes would bind first to these high affinity sites, and as complex levels rise, moderate and lower affinity sites would get filled until the complex is covering the X chromosome. Lowering the amount of DCC in general or depletion of a DCC component will still allow interaction with high affinity sites, but leave sites of lower affinity unbound. Thus a major difference to the Ôentry siteÕ model is that binding to lower affinity sites does not depend on prior binding to a nearby CES, which explains why many X derived sequences lacking entry sites can recruit the complex. A model based purely on affinities predicts that any part of the X, which is normally bound by the DCC, would be able to recruit the complex to an autosomal position in wild type males. However, early attempts at recruiting the DCC to autosomal insertions of dosage compensated genes such as white and yellow failed. In fact a transposable element carrying a 179 kb segment between white and roughest, which is bound by the complex on the X, failed to recruit the DCC to its autosomal location [17]. It is therefore probable that many of the proposed low affinity sites profit from high local concentrations of DCC on the X chromosome and are not able to recruit the complex if moved out of this context. Under those circumstances, the term ÔspreadingÕ may therefore be used to describe how low affinity sites benefit from the presence of higher affinity sites that increase the local concentration of complex (Fig. 2c). This might explain why the roX genes can induce extensive spreading on autosomes when they are the only source of RNA or when the levels of complex are increased. The very high local concentrations of DCC generated on the autosomal insertion site in this case would allow recruitment to such low affinity ÔsitesÕ on autosomes. The recruitment of the DCC to sites of active transcription [11] may suggest that low affinity sites could be defined by features other than DNA sequence alone, such as the presence of active marks on chromatin or the transcription machinery itself. As previously mentioned, the models described above are entirely based on studies using polytene chromosomes as these provide a way of visualizing the global distribution of chromatin-associated proteins. However, the limited resolution means that we know little about how the DCC proteins are distributed within what we see as a fluorescent ÔbandÕ on polytenes. Fig. 2 (models d–f) illustrates some of the various possibilities for short-range distribution. Each ÔbandÕ of DCC signal could represent a single binding site for the complex. In support of this, single transcription factor binding sites give a similar signal when visualised by immunofluorescence [28]. On the other hand, it is possible that each ÔbandÕ contains many specific binding sites for the complex. Analysis of the 18D entry site indicates that this site may be made up of multiple elements distributed over up to 8.8 kb [21]. Finally, it is also conceivable that each ÔbandÕ results from several binding events: one primary, sequence-specific binding and additional contacts within the domain with no sequence definition. In this context, the term ÔspreadingÕ may be used to indicate an equilibration of chromatin binding within a domain, which may be limited by domain boundaries. Alternatively, the distribution of DCC within a band may be determined by active chromatin itself. Such a scenario gains support from mapping of the distri-
T. Straub et al. / FEBS Letters 579 (2005) 3258–3263
bution of the H4K16 acetylation mark by chromatin immunoprecipitation, which correlated the acetylation mark with active transcription units [10].
4. Outlook In order to resolve the issues of how dosage compensation is achieved on a transcriptional level and what determines the distribution of the DCC, further molecular studies will be essential. Regarding the mode of hyperactivation, applying absolute quantification of gene expression to cell systems in which the DCC can be knocked down by RNA interference will reveal the actual phenotype of mutant flies at the gene level. Identification and characterisation of more DCC binding sites that are not related to the roX genes should finally lead to a molecular explanation for their sequential occupancy. Increasing the resolution of mapping experiments by applying chromatin IP will help to uncover significant correlations between DCC binding, histone acetylation and compensated gene expression. The comprehensive collection of molecular data will hopefully facilitate the integration of the various models discussed here into one model for dosage compensation in flies. Acknowledgements: Work in the Becker lab on dosage compensation is supported by Deutsche Forschungsgemeinschaft through Transregio5 and ‘‘Fonds der Chemischen Industrie’’.
References [1] Bashaw, G.J. and Baker, B.S. (1996) Dosage compensation and chromatin structure in Drosophila. Curr. Opin. Genet. Dev. 6, 496–501. [2] Lucchesi, J.C. (1998) Dosage compensation in flies and worms: the ups and downs of X-chromosome regulation. Curr. Opin. Genet. Dev. 8, 179–184. [3] Meller, V.H. and Kuroda, M.I. (2002) Sex and the single chromosome. Adv. Genet. 46, 1–24. [4] Gilfillan, G.D., Dahlsveen, I.K. and Becker, P.B. (2004) Lifting a chromosome: dosage compensation in Drosophila melanogaster. FEBS Lett. 567, 8–14. [5] Kelley, R.L. (2004) Path to equality strewn with roX. Dev. Biol. 269, 18–25. [6] Bone, J.R., Lavender, J., Richman, R., Palmer, M.J., Turner, B.M. and Kuroda, M.I. (1994) Acetylated histone H4 on the male X chromosome is associated with dosage compensation in Drosophila. Genes. Dev. 8, 96–104. [7] Akhtar, A. and Becker, P.B. (2000) Activation of transcription through histone H4 acetylation by MOF, an acetyltransferase essential for dosage compensation in Drosophila. Mol. Cell 5, 367–375. [8] Belote, J.M. and Lucchesi, J.C. (1980) Control of X chromosome transcription by the maleless gene in Drosophila. Nature 285, 573– 575. [9] Chiang, P.W. and Kurnit, D.M. (2003) Study of dosage compensation in Drosophila. Genetics 165, 1167–1181. [10] Smith, E.R., Allis, C.D. and Lucchesi, J.C. (2001) Linking global histone acetylation to the transcription enhancement of Xchromosomal genes in Drosophila males. J. Biol. Chem. 276, 31483–31486. [11] Sass, G.L., Pannuti, A. and Lucchesi, J.C. (2003) Male-specific lethal complex of Drosophila targets activated regions of the X chromosome for chromatin remodeling. Proc. Natl. Acad. Sci. USA 100, 8287–8291. [12] Kurdistani, S.K., Tavazoie, S. and Grunstein, M. (2004) Mapping global histone acetylation patterns to gene expression. Cell 117, 721–733.
T. Straub et al. / FEBS Letters 579 (2005) 3258–3263 [13] Loidl, P. (2004) A plant dialect of the histone language. Trends. Plant Sci. 9, 84–90. [14] Birchler, J.A., Pal-Bhadra, M. and Bhadra, U. (2003) Dosage dependent gene regulation and the compensation of the X chromosome in Drosophila males. Genetica 117, 179–190. [15] Corona, D.F., Langst, G., Clapier, C.R., Bonte, E.J., Ferrari, S., Tamkun, J.W. and Becker, P.B. (1999) ISWI is an ATPdependent nucleosome remodeling factor. Mol. Cell 3, 239–245. [16] Pal Bhadra, M., Bhadra, U., Kundu, J. and Birchler, J.A. (2005) Gene expression analysis of the function of the MSL complex in Drosophila. Genetics. [17] Bhadra, U., Pal-Bhadra, M. and Birchler, J.A. (1999) Role of the male specific lethal (msl) genes in modifying the effects of sex chromosomal dosage in Drosophila. Genetics 152, 249– 268. [18] Bhadra, U., Pal-Bhadra, M. and Birchler, J.A. (2000) Histone acetylation and gene expression analysis of sex lethal mutants in Drosophila. Genetics 155, 753–763. [19] Park, Y. and Kuroda, M.I. (2001) Epigenetic aspects of Xchromosome dosage compensation. Science 293, 1083–1085. [20] Fagegaltier, D. and Baker, B.S. (2004) X chromosome sites autonomously recruit the dosage compensation complex in Drosophila males. PLoS. Biol. 2, e341. [21] Oh, H., Bone, J.R. and Kuroda, M.I. (2004) Multiple classes of MSL binding sites target dosage compensation to the X chromosome of Drosophila. Curr. Biol. 14, 481–487.
3263 [22] Kageyama, Y., Mengus, G., Gilfillan, G., Kennedy, H.G., Stuckenholz, C., Kelley, R.L., Becker, P.B. and Kuroda, M.I. (2001) Association and spreading of the Drosophila dosage compensation complex from a discrete roX1 chromatin entry site. Embo J. 20, 2236–2245. [23] Kelley, R.L., Meller, V.H., Gordadze, P.R., Roman, G., Davis, R.L. and Kuroda, M.I. (1999) Epigenetic spreading of the Drosophila dosage compensation complex from roX RNA genes into flanking chromatin. Cell 98, 513–522. [24] Meller, V.H., Gordadze, P.R., Park, Y., Chu, X., Stuckenholz, C., Kelley, R.L. and Kuroda, M.I. (2000) Ordered assembly of roX RNAs into MSL complexes on the dosage-compensated X chromosome in Drosophila. Curr. Biol. 10, 136–143. [25] Park, Y., Kelley, R.L., Oh, H., Kuroda, M.I. and Meller, V.H. (2002) Extent of chromatin spreading determined by roX RNA recruitment of MSL proteins. Science 298, 1620–1623. [26] Oh, H., Park, Y. and Kuroda, M.I. (2003) Local spreading of MSL complexes from roX genes on the Drosophila X chromosome. Genes. Dev. 17, 1334–1339. [27] Demakova, O.V., Kotlikova, I.V., Gordadze, P.R., Alekseyenko, A.A., Kuroda, M.I. and Zhimulev, I.F. (2003) The MSL complex levels are critical for its correct targeting to the chromosomes in Drosophila melanogaster. Chromosoma 112, 103–115. [28] Schwartz, B.E., Werner, J.K. and Lis, J.T. (2004) Indirect immunofluorescent labeling of Drosophila polytene chromosomes: visualizing protein interactions with chromatin in vivo. Methods Enzymol. 376, 393–404.
FEBS 29483
Minireview
Dosage compensation in flies: Mechanism, models, mystery Tobias Straub1, Ina K. Dahlsveen1, Peter B. Becker* Adolf-Butenandt-Institut, Molekularbiologie, Schillerstrasse 44, 80336 Mu¨nchen, Germany Accepted 26 March 2005 Available online 2 April 2005 Edited by Horst Feldmann
Abstract Dosage compensation involves fine-tuning of gene expression at the level of entire chromosomes. The principles that assure selective targeting of the male X chromosome in Drosophila and the mechanism by which transcription levels are adjusted in a twofold range are still mysterious. We discuss the prevalent models in the context of recent experimental observations. Ó 2005 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. Keywords: Transcription; Dosage compensation; Chromosome; Chromatin; Histone acetylation
1. Dosage compensation, a chromosome-wide challenge Sexual dimorphism is brought about by different sex chromosomes, which are unequally distributed between the sexes. For example in humans and fruit flies females carry two X chromosomes, but males have only a single X chromosome in addition to the gene-poor Y chromosome. Hence, male cells contain a halved dose of all X-linked genes. The level of Xlinked gene products, however, must be similar in cells of either sex. A prime principle of gene dosage compensation is the chromosome-wide adjustment of transcription. The mechanisms by which this may be achieved vary largely between different organisms and involve different molecular factors. Whereas female mammals inactivate one of their two X chromosomes in order to approximate the gene dosage of males, C. elegans hermaphrodites (XX) reduce transcription of each X by about half. In Drosophila, the male flies increase transcription from their single X chromosome about twofold to equal the expression of the two X chromosomes in females (for extensive reviews see [1–5]). This adjustment is essential for male survival. Dosage compensation in flies relies on the integrity of a ribonucleoprotein complex, known as the dosage compensation complex (DCC). This complex comprises the male-specific lethal (MSL) proteins and two non-coding RNA molecules, roX1 and roX2. Male-specific expression of
* Corresponding author. Fax: +49 89 5996 425. E-mail address: [email protected] (P.B. Becker). 1
These authors contributed equally to this work.
Abbreviations: DCC, dosage compensation complex; MSL, male-specific lethal; H4K16, lysine 16 of histone H4
MSL2 leads to enrichment of the DCC on the single X chromosome. The cooperation of structural and enzymatic complex components is necessary for proper function and targeting. So far, the catalytic activities known to be essential for the compensation process are the ATPase/helicase activity of MLE and the histone acetyl transferase activity of MOF. Accumulation of the DCC on the X chromosome results in increased acetylation of lysine 16 on histone H4. This specific chromatin modification is thought to be at least in part responsible for elevated X-linked gene expression. The mechanisms that assure that a fairly global process acting at the level of an entire chromosome leads to rather subtle (twofold) tuning of transcription are still mysterious. However, integrating the limited experimental observations into models has been instructive. Not surprisingly, different kinds of models have been proposed trying to explain the mode of precise transcription tuning and to describe how the X chromosome gets exclusively furnished with the DCC. Experimental progress now allows for the critical evaluation of these models.
2. Opposing models for dosage compensation The prevailing model (Fig. 1, model A) for the compensation of the X chromosomal genes in males states that the DCC is directly involved in activating gene expression on the male X chromosome. It is based on the following observations: (a) in Drosophila males, the MSL proteins are almost exclusively located on the X chromosome [4]; (b) H4K16 acetylation follows the distribution of the MSLs [6]; (c) acetylation of H4K16 can activate transcription in a chromatin context [7]; (d) the male X chromosome has a more diffuse appearance, reminiscent of reduced chromatin condensation [1]; (e) knock-down of any of the MSL proteins in males reduces expression of X chromosomal genes relative to that of autosomal genes and leads to loss of compensation and male lethality [8,9]. Accordingly, in Drosophila males MSL2 stabilizes and tethers the DCC to the X chromosome. Concomitant enrichment of H4K16 acetylation results in local chromatin opening and a twofold increase in transcription, conceivably at the level of elongation [10]. Reflecting its role as a direct modulator of gene expression, the model implies a flexible and direct association of the DCC with active genomic loci [11]. Although the model is straightforward and intuitive, not all phenomena can be smoothly incorporated and key aspects have not yet been experimentally verified. For example, all MSL proteins (with the exception of MSL2) are present in female cells as well, albeit some of them at reduced levels, where
0014-5793/$30.00 Ó 2005 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2005.03.050
T. Straub et al. / FEBS Letters 579 (2005) 3258–3263
XX
3259
XY mle-
XY
model A
model B
X chromosome and gene product
autosome and gene product
dosage compensation complex
Fig. 1. Different models for how compensation of the X-linked genes is achieved in order to equalize gene doses between male (XY) and female (XX) flies. Model A suggests male-specific assembly of the dosage compensation complex (DCC) that directly activates gene expression on the X. Model B proposes that compensation of the X is accomplished by a genome-wide increase in transcription that is due to activator imbalance inferred by X chromosomal monosomy. Tethering of the activators (members of the DCC) onto the male X alleviates the transcription enhancement on the autosomes. While the expression status in wild type males is the same for both models, MLE mutants (XY mle-) uncover major differences characterizing the two hypotheses.
they may affect chromatin structure and function globally. An unexplained issue is how the targeted H4K16 acetylation will bring about a general and precise twofold elevation of transcription [10]. Increased H4K16 acetylation might induce a chromosome-wide chromatin decondensation leading to an elevated transcription in the proper range of magnitude. Unfortunately, there is no experimental evidence for such an effect. In vitro, MOF-mediated histone acetylation has been shown to relieve chromatin-mediated transcriptional repression [7]. However, the activation observed was much more than twofold. A precise doubling of transcription is likely a result of a fine balance between activation and feedback repression and thus it is possible that DCC harbours repressive activities in addition to the activating histone acetylation function, activities that are not yet accounted for by model A. In general, histone acetylation is involved in gene activation in all eukaryotes from yeast to man. Surprisingly, acetylation of H4K16 is an exception to this rule, since in yeast, where its occurrence has been correlated with gene activity on a genome-wide scale, it was shown to mark inactive, rather than active loci [12]. Accordingly, H4K16ac has been postulated to prevent binding of transcription factors in silent chromatin regions in yeast. While we appreciate that the language of histone modifications is modulated by species-specific dialects [13], the case of yeast chromatin suggests complex effects of H4K16 acetylation, which may impact on model building for dosage compensation in Drosophila. The most important objection to model A is, however, the lack of absolute quantification of gene expression confirming the direct involvement of the DCC in compensation of Xlinked genes.
Birchler and colleagues have proposed an alternative and less intuitive model (Fig. 1, model B) that employs a genome-wide mechanism underlying the compensation of the X chromosome [14]. Central to this model is the inverse dosage effect, which can be observed when large chromosomal segments are rendered aneuploid. Such deletions have been observed to lead to a global increase in gene expression. This may be explained by a preferential loss of transcriptional repressors given a higher prevalence of those factors. Thus, deletion of a chromosomal segment will offset the balance between global positive and negative regulators, resulting in a dominance of positive factors. According to Birchler, the X chromosomal monosomy in males is a special case of an aneuploidy that did not arise by abrupt deletion, but by the gradual process of diversification of sex chromosomes. The evolutionary time scales to be considered provided ample opportunity for co-evolution of compensatory principles that re-establish the balance of gene expression. According to the model, loss of one X chromosome in males leads to a global inverse regulation of gene expression, i.e. to hyperactivation of all autosomes and the remaining X chromosome due to a relative preponderance of positive transcription regulators [14]. In Drosophila metafemales harbouring three X chromosomes, the same principle would compensate the increase in X chromosomal dosis by a general reduction of gene expression to 2/3 [14]. Although the inverse dosage effect compensated for the aneuploidy of the X chromosome, it obviously would not allow male survival, since all autosomal gene expression is doubled relative to that in females (Fig. 1B, centre panel). In order to counteract autosomal hyperactivation, Birchler proposes that global activators of transcription, con-
3260
ceivably MOF, are removed from the autosomes by an MSL2/MLE-mediated sequestration to the X chromosome (Fig. 1B, right panel). Consequently, MOF-mediated histone acetylation is enriched on the X and gene expression on the autosomes drops. Now, in order to prevent excessive activation of X chromosomal target genes, repressive mechanisms are required to limit gene activation to the twofold effect, brought about by the inverse dosage effect. Such negative effectors my reside in the MSL complex itself [14] or involve other factors, such as the chromatin remodeling ATPase ISWI [15,16]. Apparently, one of the fundamental differences between the models is stipulated by the role of MSL2: whereas in model A MSL2 and the DCC it organises are directly involved in activation, in model B dosage compensation is brought about by the inverse dosage effect, which may have no direct mechanistic connection to the MSL proteins. Here, MSL2 may function in two distinct ways: first, to tether global autosomal activators (the other MSL proteins) to the X chromosome in order to counteract autosomal hyperactivation, and second, to prevent excessive activation by increased H4K16 acetylation on the X. Accordingly, the presence of DCC on the X chromosome might not necessarily correlate with activated transcription. A prediction of model B is that disruption of MSL2 or MLE would result in a release of activators and an increased autosomal expression. Under these circumstances X chromosomal expression would remain unchanged due to the retained inverse dosage effect. The strongest support for this model stems from experiments employing absolute quantification of gene expression [14,16]. These measurements demonstrated, for instance, the predicted absolute increase in autosomal gene expression upon knockdown of MLE for a number of genes (Fig. 1, compare the two models in the XY mle-background) [17]. However, the global nature of the inverse dosage effect appears to be less robust. Although some genes respond as predicted in the experiments performed by Birchler and coworkers [18], a significant number of them do not. MSL mutants that should re-establish the inverse dosage effect on the autosomes by releasing the activators exhibit a rather heterogeneous expression response of autosomal genes. Although highly suggested by the model, MOF does not appear to contribute to the compensation of the X by the inverse dosage effect: in MOF mutants X chromosomal gene expression remained unchanged [16]. At this point it remains unclear why MOF is recruited to the X in order to relieve the inverse dosage effect on the autosomes and, most importantly, why a mof mutation is lethal for males only. An obvious drawback of the model is, therefore, that the inverse dosage effect it employs to explain dosage compensation remains even more mysterious than the phenomenon it tried to explain. OckhamÕs razor would find ample opportunities for trimming. Absolute quantification of gene expression as applied by Birchler and colleagues is an important step towards defining the actual compensation process and testing models. However, care has to be taken regarding the specimen analyzed. The MSL mutant phenotype of embryos or larvae could be compromised to various extend by maternal contribution of MSL proteins and by secondary effects on gene expression in dying larvae. Crucial for the validation of any model will be to define the functional contribution of H4K16 acetyla-
T. Straub et al. / FEBS Letters 579 (2005) 3258–3263
tion to the compensation process as the simple rule ‘‘histone acetylation equals gene activation’’ turns out to be obsolete [12].
3. Recruitment of DCC to the X chromosome: Models in transition The two contrasting models for how the DCC functions in dosage compensation discussed above have different implications for the importance of the targeting and distribution of the complex. Model A implies that the correct distribution of the DCC is vital for its function in up-regulating transcription, whereas in model B the precise location of the complex may not be so important. A longstanding model for how localisation of the DCC to the X is achieved involves primary recognition of specific sites (Ôentry sitesÕ) followed by distribution along the chromosome by ÔspreadingÕ in cis from these sites to less defined chromatin [19]. Recent observations have challenged this Ôentry-siteÕ model and led to the proposition of an alternative model involving a larger number of specific sites of varying affinities [4,20,21]. Both models are based on observations using salivary gland polytene chromosomes where the distribution of the DCC can be easily visualised by immunostaining (Fig. 2, top pictures). In wild type male nuclei, the complex binds mainly to interbands in a characteristic pattern [23]. The Ôentry siteÕ model was originally based on the characterisation of the two first specific binding sites for the DCC found within the roX1 and roX2 genes. The roX genes are part of a relatively small set of DCC binding sites that is observed in females expressing MSL2 and mutant for msl3, mle or mof [4]. This very visually suggested the presence of a special type of binding site, proposed to be the initial sites of recruitment, from which the complex would subsequently spread to cover the chromosome. Accordingly, these site were termed Ôchromatin entry sitesÕ (CES) [19]. Detailed studies of the roX1 and roX2 sites provided convincing evidence in favour of the model [22–24]. When the roX genes were inserted on autosomes via P-element transformation, the DCC was recruited to the insertion site. Importantly, the DCC was occasionally seen to ÔspreadÕ to neighbouring, autosomal interbands in a pattern reminiscent of that seen on the X, suggesting that this recruitment and spreading mirrored the real mechanism for targeting and distribution of the DCC. The Ôentry site modelÕ (illustrated in Fig. 2a) therefore predicts that the CES are the only specific recruitment sites for the DCC. Binding to an entry site would be a prerequisite for the occupancy of secondary, non-specific sites in the vicinity, a process that has been referred to as Ôepigenetic spreadingÕ [5,23]. It has been proposed that these secondary sites may represent sites of active transcription where the complex is required to implement dosage compensation. Using transposable elements on the X that contain UAS sequences, it has been demonstrated that the DCC can be recruited to these elements upon activation of transcription by Gal4 [11]. Two aspects of the Ôentry siteÕ model have been questioned by recent results. Firstly, on several occasions insertion of X chromosomal sequences lacking a CES still led to recruitment of the DCC to autosomes in wild type males ([21], IKD unpublished). In addition, several large translocations from the X to
T. Straub et al. / FEBS Letters 579 (2005) 3258–3263
3261
Fig. 2. Sequential occupancy of the X chromosome might employ different ways of complex distribution. Depicted are two Drosophila polytene chromosome stainings for MSL1 at low (upper panel) or high (lower panel) concentration of complex. The way the complexes target new ÔbandsÕ (long-range distribution, a–c) and the chromosome binding within a ÔbandÕ (short-range distribution, d–f) are controversially discussed.
autosomes, also lacking a CES, have been shown to recruit the DCC [20]. These observations demonstrate that DCC recruitment is not restricted to a small number of dedicated Ôentry sitesÕ. The second aspect of the Ôentry siteÕ model that has been challenged by recent data is whether spreading actually occurs as part of the mechanism for distribution of the DCC over the X chromosome. Extensive spreading from autosomal roX genes, which depended on the transcription of the roX RNA, the abundance of DCC [27] and was particularly evident if the number of roX genes was reduced [25,26], turns out to be the exception rather than the rule. The third CES to be described, 18D, was found to induce only minimal spreading in rare cases [21]. Importantly, when parts of autosomes get
translocated to the X, spreading of the DCC into autosomal sequences does not occur, even if the translocations lie close to entry sites, including roX genes [20,21]. This suggests that spreading is not the prevailing mechanism for DCC distribution and that the X chromosome harbours recognition elements in addition to the known entry sites. The observations described above have led to a revised model for the targeting and distribution of the DCC, which we refer to as the ÔaffinitiesÕ model (Fig. 2b) [4,20,21]. This model assumes that most, if not all, binding sites for the DCC are specific, but of varying affinities for the DCC. Accordingly, one major determinant of recruitment is the concentration of intact DCC in a nucleus. Studies using females expressing varying levels of MSL2 support this prediction as when complex levels
3262
are limiting the DCC binds only to the sites that were observed in the msl mutant backgrounds [27]. In this scenario, the entry sites are not special recruitment sites, but merely represent the highest affinity sites. Partial and/or complete complexes would bind first to these high affinity sites, and as complex levels rise, moderate and lower affinity sites would get filled until the complex is covering the X chromosome. Lowering the amount of DCC in general or depletion of a DCC component will still allow interaction with high affinity sites, but leave sites of lower affinity unbound. Thus a major difference to the Ôentry siteÕ model is that binding to lower affinity sites does not depend on prior binding to a nearby CES, which explains why many X derived sequences lacking entry sites can recruit the complex. A model based purely on affinities predicts that any part of the X, which is normally bound by the DCC, would be able to recruit the complex to an autosomal position in wild type males. However, early attempts at recruiting the DCC to autosomal insertions of dosage compensated genes such as white and yellow failed. In fact a transposable element carrying a 179 kb segment between white and roughest, which is bound by the complex on the X, failed to recruit the DCC to its autosomal location [17]. It is therefore probable that many of the proposed low affinity sites profit from high local concentrations of DCC on the X chromosome and are not able to recruit the complex if moved out of this context. Under those circumstances, the term ÔspreadingÕ may therefore be used to describe how low affinity sites benefit from the presence of higher affinity sites that increase the local concentration of complex (Fig. 2c). This might explain why the roX genes can induce extensive spreading on autosomes when they are the only source of RNA or when the levels of complex are increased. The very high local concentrations of DCC generated on the autosomal insertion site in this case would allow recruitment to such low affinity ÔsitesÕ on autosomes. The recruitment of the DCC to sites of active transcription [11] may suggest that low affinity sites could be defined by features other than DNA sequence alone, such as the presence of active marks on chromatin or the transcription machinery itself. As previously mentioned, the models described above are entirely based on studies using polytene chromosomes as these provide a way of visualizing the global distribution of chromatin-associated proteins. However, the limited resolution means that we know little about how the DCC proteins are distributed within what we see as a fluorescent ÔbandÕ on polytenes. Fig. 2 (models d–f) illustrates some of the various possibilities for short-range distribution. Each ÔbandÕ of DCC signal could represent a single binding site for the complex. In support of this, single transcription factor binding sites give a similar signal when visualised by immunofluorescence [28]. On the other hand, it is possible that each ÔbandÕ contains many specific binding sites for the complex. Analysis of the 18D entry site indicates that this site may be made up of multiple elements distributed over up to 8.8 kb [21]. Finally, it is also conceivable that each ÔbandÕ results from several binding events: one primary, sequence-specific binding and additional contacts within the domain with no sequence definition. In this context, the term ÔspreadingÕ may be used to indicate an equilibration of chromatin binding within a domain, which may be limited by domain boundaries. Alternatively, the distribution of DCC within a band may be determined by active chromatin itself. Such a scenario gains support from mapping of the distri-
T. Straub et al. / FEBS Letters 579 (2005) 3258–3263
bution of the H4K16 acetylation mark by chromatin immunoprecipitation, which correlated the acetylation mark with active transcription units [10].
4. Outlook In order to resolve the issues of how dosage compensation is achieved on a transcriptional level and what determines the distribution of the DCC, further molecular studies will be essential. Regarding the mode of hyperactivation, applying absolute quantification of gene expression to cell systems in which the DCC can be knocked down by RNA interference will reveal the actual phenotype of mutant flies at the gene level. Identification and characterisation of more DCC binding sites that are not related to the roX genes should finally lead to a molecular explanation for their sequential occupancy. Increasing the resolution of mapping experiments by applying chromatin IP will help to uncover significant correlations between DCC binding, histone acetylation and compensated gene expression. The comprehensive collection of molecular data will hopefully facilitate the integration of the various models discussed here into one model for dosage compensation in flies. Acknowledgements: Work in the Becker lab on dosage compensation is supported by Deutsche Forschungsgemeinschaft through Transregio5 and ‘‘Fonds der Chemischen Industrie’’.
References [1] Bashaw, G.J. and Baker, B.S. (1996) Dosage compensation and chromatin structure in Drosophila. Curr. Opin. Genet. Dev. 6, 496–501. [2] Lucchesi, J.C. (1998) Dosage compensation in flies and worms: the ups and downs of X-chromosome regulation. Curr. Opin. Genet. Dev. 8, 179–184. [3] Meller, V.H. and Kuroda, M.I. (2002) Sex and the single chromosome. Adv. Genet. 46, 1–24. [4] Gilfillan, G.D., Dahlsveen, I.K. and Becker, P.B. (2004) Lifting a chromosome: dosage compensation in Drosophila melanogaster. FEBS Lett. 567, 8–14. [5] Kelley, R.L. (2004) Path to equality strewn with roX. Dev. Biol. 269, 18–25. [6] Bone, J.R., Lavender, J., Richman, R., Palmer, M.J., Turner, B.M. and Kuroda, M.I. (1994) Acetylated histone H4 on the male X chromosome is associated with dosage compensation in Drosophila. Genes. Dev. 8, 96–104. [7] Akhtar, A. and Becker, P.B. (2000) Activation of transcription through histone H4 acetylation by MOF, an acetyltransferase essential for dosage compensation in Drosophila. Mol. Cell 5, 367–375. [8] Belote, J.M. and Lucchesi, J.C. (1980) Control of X chromosome transcription by the maleless gene in Drosophila. Nature 285, 573– 575. [9] Chiang, P.W. and Kurnit, D.M. (2003) Study of dosage compensation in Drosophila. Genetics 165, 1167–1181. [10] Smith, E.R., Allis, C.D. and Lucchesi, J.C. (2001) Linking global histone acetylation to the transcription enhancement of Xchromosomal genes in Drosophila males. J. Biol. Chem. 276, 31483–31486. [11] Sass, G.L., Pannuti, A. and Lucchesi, J.C. (2003) Male-specific lethal complex of Drosophila targets activated regions of the X chromosome for chromatin remodeling. Proc. Natl. Acad. Sci. USA 100, 8287–8291. [12] Kurdistani, S.K., Tavazoie, S. and Grunstein, M. (2004) Mapping global histone acetylation patterns to gene expression. Cell 117, 721–733.
T. Straub et al. / FEBS Letters 579 (2005) 3258–3263 [13] Loidl, P. (2004) A plant dialect of the histone language. Trends. Plant Sci. 9, 84–90. [14] Birchler, J.A., Pal-Bhadra, M. and Bhadra, U. (2003) Dosage dependent gene regulation and the compensation of the X chromosome in Drosophila males. Genetica 117, 179–190. [15] Corona, D.F., Langst, G., Clapier, C.R., Bonte, E.J., Ferrari, S., Tamkun, J.W. and Becker, P.B. (1999) ISWI is an ATPdependent nucleosome remodeling factor. Mol. Cell 3, 239–245. [16] Pal Bhadra, M., Bhadra, U., Kundu, J. and Birchler, J.A. (2005) Gene expression analysis of the function of the MSL complex in Drosophila. Genetics. [17] Bhadra, U., Pal-Bhadra, M. and Birchler, J.A. (1999) Role of the male specific lethal (msl) genes in modifying the effects of sex chromosomal dosage in Drosophila. Genetics 152, 249– 268. [18] Bhadra, U., Pal-Bhadra, M. and Birchler, J.A. (2000) Histone acetylation and gene expression analysis of sex lethal mutants in Drosophila. Genetics 155, 753–763. [19] Park, Y. and Kuroda, M.I. (2001) Epigenetic aspects of Xchromosome dosage compensation. Science 293, 1083–1085. [20] Fagegaltier, D. and Baker, B.S. (2004) X chromosome sites autonomously recruit the dosage compensation complex in Drosophila males. PLoS. Biol. 2, e341. [21] Oh, H., Bone, J.R. and Kuroda, M.I. (2004) Multiple classes of MSL binding sites target dosage compensation to the X chromosome of Drosophila. Curr. Biol. 14, 481–487.
3263 [22] Kageyama, Y., Mengus, G., Gilfillan, G., Kennedy, H.G., Stuckenholz, C., Kelley, R.L., Becker, P.B. and Kuroda, M.I. (2001) Association and spreading of the Drosophila dosage compensation complex from a discrete roX1 chromatin entry site. Embo J. 20, 2236–2245. [23] Kelley, R.L., Meller, V.H., Gordadze, P.R., Roman, G., Davis, R.L. and Kuroda, M.I. (1999) Epigenetic spreading of the Drosophila dosage compensation complex from roX RNA genes into flanking chromatin. Cell 98, 513–522. [24] Meller, V.H., Gordadze, P.R., Park, Y., Chu, X., Stuckenholz, C., Kelley, R.L. and Kuroda, M.I. (2000) Ordered assembly of roX RNAs into MSL complexes on the dosage-compensated X chromosome in Drosophila. Curr. Biol. 10, 136–143. [25] Park, Y., Kelley, R.L., Oh, H., Kuroda, M.I. and Meller, V.H. (2002) Extent of chromatin spreading determined by roX RNA recruitment of MSL proteins. Science 298, 1620–1623. [26] Oh, H., Park, Y. and Kuroda, M.I. (2003) Local spreading of MSL complexes from roX genes on the Drosophila X chromosome. Genes. Dev. 17, 1334–1339. [27] Demakova, O.V., Kotlikova, I.V., Gordadze, P.R., Alekseyenko, A.A., Kuroda, M.I. and Zhimulev, I.F. (2003) The MSL complex levels are critical for its correct targeting to the chromosomes in Drosophila melanogaster. Chromosoma 112, 103–115. [28] Schwartz, B.E., Werner, J.K. and Lis, J.T. (2004) Indirect immunofluorescent labeling of Drosophila polytene chromosomes: visualizing protein interactions with chromatin in vivo. Methods Enzymol. 376, 393–404.