Identification of the Drosophila X chromosome: The long and short of it

Abstract

The different dose of X chromosomes in males and females produces a potentially fatal imbalance in X-linked gene products. This imbalance is addressed by dosage compensation, a process that modulates expression from an entire X chromosome in one sex. Dosage compensation acts on thousands of genes with disparate expression patterns. Both flies and mammals accomplish this with remarkable specificity by targeting epigenetic chromatin modifications to a single chromosome. Long noncoding RNAs that are expressed from the X chromosome are essential elements of the targeting mechanism in both lineages. We recently discovered that the siRNA pathway, as well as small RNA from satellite repeats that are strikingly enriched on the fly X chromosome, also promote X recognition. In this article we review the current understanding of X recognition in flies and discuss potential mechanisms by which the siRNA pathway, repetitive elements and long noncoding RNAs might cooperate to promote X recognition.

Keywords:

• 1.688^X repeats
• 359 bp repeats
• dosage compensation
• Drosophila melanogaster
• roX RNA
• siRNA
• satellite repeats
• X recognition

Introduction

Sex chromosomes produce genetic imbalance

Although the primary signals that direct sexual development are remarkably varied, the adaptation of a pair of chromosomes to control this process is common. Organisms as diverse as flies and humans have XY males and XX females. The emergence of dimorphic sex chromosomes is driven by degradation of the Y chromosome in response to transmission solely through one sex.¹ In consequence, the Y chromosomes of flies and humans, which are of completely different origin, are both gene poor and largely heterochromatic. In contrast, the X chromosomes are gene rich. Heterogametic sex chromosomes therefore produce an imbalance in X-linked gene dosage between males and females. This imbalance is addressed early in development by a process known as dosage compensation.²

Chromosome-wide modulation of gene expression

Maintenance of a balanced X to autosome gene expression ratio is essential in flies and mammals. Eutherian mammals accomplish this by a global increase of X-linked gene expression that restores the X to autosome ratio in males.³ Females accommodate increased expression by inactivation of a single, randomly selected X chromosome in every cell during early embryogenesis.² Inactivation is initiated by expression of the X inactive specific transcript (Xist) from one of the 2 X chromosomes in female cells.⁴ Xist, a long noncoding RNA, recruits complexes that modify and silence chromatin. Silencing spreads in cis to inactivate a single X chromosome. Additional features of the X chromosome, such as interphase organization and repetitive elements on the X, facilitate recognition and silencing.^4,5

Flies also increase the expression of X-linked genes, but through a mechanism that is limited to males. Almost all genes on the single X chromosome of males are bound by a complex of proteins and RNA that modifies chromatin to increase expression.⁶ This complex, known as the Male Specific Lethal (MSL) complex, is composed of 5 MSL proteins and one of 2 non-coding RNA on the X RNAs (roX1, roX2). Females block expression of a key MSL protein, and this limits X upregulation to males.⁷ The MSL complex is recruited to transcribed genes.⁸ A histone acetyltransferase in the complex, Males absent on the first (MOF), places H4Ac16 marks at sites of MSL complex binding.^9,10 The precise molecular mechanism by which the MSL complex elevates expression remains a topic of debate. The H4K16Ac mark partially decondenses chromatin and increases transcriptional elongation at X-linked genes.^11-14 Compensation may also involve a modest enrichment of RNA polymerase II (RNAP II) at promoters, and the mammalian homolog of one complex member is an E3 ubiquitin ligase that modifies H2B and promotes early transcriptional elongation.^15-17 Regardless of the transcriptional steps that are modulated to achieve elevated expression, association of the MSL complex with the male X chromosome is remarkably specific, limiting this process to X-linked genes.

Identification of X chromatin is a multi-step process

Elegant studies by several groups have established a sequential model for recruitment of the MSL complex to X chromatin. A number of X-linked Chromatin Entry Sites (CES) were initially characterized by their robust recruitment of remaining MSL proteins in males mutated for some members of the complex.¹⁸ A key feature of CES is the MSL Recognition Element (MRE), a 21 bp sequence that is modestly enriched on the X chromosome.^19,20 An RNAi screen designed to find genes required for X-localization identified CLAMP, a zinc finger protein that binds MREs and is essential for MSL recruitment.²¹ Although CLAMP is necessary for compensation, it also binds to autosomal MREs that do not recruit the MSL complex. Furthermore, CLAMP is an essential gene in both sexes, indicating a role outside of dosage compensation. After recruitment by CES the MSL complex spreads into nearby transcribed genes through binding of one complex member, Male Specific Lethal 3 (MSL3) to the cotranscriptional H3K36me3 mark.^22-24 In accord with this the Set2 methyltransferase, responsible for H3K36me3 deposition, is essential for spreading of the complex from CES into nearby genes.^23,25 These studies were used to formulate a well-supported model of local recruitment in which initial binding to CES is followed by spreading into active genes.²⁶ However, both recruitment to the CES and subsequent spreading into nearby genes rely on features that are not specific to the X chromosome. How X chromatin is recognized with sufficient specificity remains an open question. We recently demonstrated that small RNA contributes to X recognition.²⁷ In this article we propose mechanisms by which small RNAs might act to ensure efficient X recognition by the MSL complex.

The roX genes play multiple roles in MSL recruitment to the X chromosome

RNA produced by roX1 and roX2 assembles with the MSL proteins and can be visualized along the length of the male X chromosome.²⁸ Mutation of a single roX gene has no obvious phenotype, but simultaneous loss of roX1 and roX2 results in mislocalization of MSL proteins to ectopic autosomal sites.²⁹ Reduced expression of X-linked genes in roX1 roX2 males leads to lethality.³⁰ The roX genes have unusual properties that may contribute to X recognition. Both roX genes are X-linked and overlap CES. In accordance with this, the roX genes themselves recruit the MSL complex and promote spreading into flanking chromatin.³¹ Remarkably, autosomal insertions of roX transgenes recruit MSL proteins, which then spread into autosomal chromatin flanking the transgene.^32,33 The ability of roX to recruit modifying activities to chromatin in cis is reminiscent of Xist, and undoubtedly contributes to normal X recognition in flies. However, roX RNA from autosomal transgenes also assembles with the MSL proteins and travels to the X chromosome, restoring compensation and rescuing roX1 roX2 males.²⁹ Although roX RNA is essential for X recognition, the location of both roX genes on the X chromosome is thus not essential for identification of X chromatin.

The siRNA pathway contributes to X recognition

A clue to how X chromatin is recognized lies in the discovery that the siRNA pathway contributes to this process.²⁷ Loss of the siRNA binding protein Argonaut 2 (Ago2) has little effect on MSL localization or viability in otherwise normal males, but is almost completely male-lethal in weakly hypomorphic roX1 roX2 flies. Lethality is accompanied by severe disruption of MSL localization, which is not otherwise observed in this particular roX1 roX2 mutant. Several additional genes that are necessary for production of siRNA also interact genetically with roX1 roX2 mutants.²⁷ This raises intriguing questions, such as the nature of small RNAs involved and the mechanism by which this pathway contributes to X chromosome recognition. Small RNA pathways have been shown to modulate chromatin and gene expression in flies, but typically act to repress expression or promote heterochromatin formation. For example, the Piwi pathway, responsible for germ line silencing of mobile elements, and the siRNA pathway have both been implicated in formation of heterochromatin in somatic tissues.^34-37 Ago2 binds chromatin and is involved in transcriptional repression of some genes.³⁸ It is possible that repression, at least in some instances, is mechanistically similar to that in Schizosaccharomyces pombe, where nascent RNAs are bound by siRNA-containing effector complexes that recruit epigenetic modifiers.^39,40

Although the involvement of the siRNA pathway in X recognition is intriguing, a possible mode of action is not immediately clear. First, there is no evidence that the MSL complex interacts directly with proteins of the siRNA system, even in the chromatin-bound context.⁴¹ An additional complication is that the MSL complex deposits activating marks, but the siRNA pathway is typically associated with repression.⁴² These considerations suggest that the siRNA pathway could modulate MSL complex localization through an indirect mechanism. Speculation about how small RNA contributes to X recognition was inspired by the identification of candidate siRNAs.

A family of X-linked repeats

The discovery that the siRNA pathway contributes to X recognition prompted a search for the source of the small RNA involved. Obvious candidates are small RNAs produced from a family of satellite repeats with remarkable enrichment on the X chromosome.⁴³ These comprise a large block of pericentric heterochromatin, and related repeats are distributed in short, tandem clusters throughout X euchromatin. These AT-rich sequences are variously known as the 359 bp (repeating unit), or 1.688 g/cm³ (buoyant density in cesium chloride) satellites. We have adopted the 1.688^X designation for the euchromatic repeats to reflect a physical property and genomic location. The remarkable distribution of the 1.688^X satellites has long prompted speculation that they might participate in dosage compensation.⁴⁴ In Drosophila melanogaster many 1.688^X satellites are transcribed from both strands, and small RNA from these repeats is detectable in some fly tissues.^45,46 Unlike most repetitive DNA, the euchromatic 1.688^X satellites are often near or within genes.⁴⁷ Several properties of the 1.688^X satellites suggest functionality. First, enrichment for satellite repeats on the X chromosome is conserved in related species, even if the sequence of these repeats is not.⁴⁸ Strikingly, a neo-X chromosome produced by an X and autosome fusion has been rapidly invaded by satellite repeats.⁴⁸

siRNA from one 1.688^X repeat promotes X recognition

The large number and dispersed localization of 1.688^X repeats on the X precludes functional testing by deletion. Instead, we decided to test whether ectopic production of short RNA from selected repeat clusters was biologically active in flies. On average, X-linked 1.688^X repeat family members share 73% sequence identity.⁴⁷ We selected 3 clusters, designated by superscripts indicating cytological position. One is immediately distal to roX1 (1.688^3F), one is at the tip of the X chromosome (1.688^1A, 89% identity to 1.688^3F), and a third is situated between these (1.688^3C, 69% identity to 1.688^3F). Ectopic production of double stranded hairpin RNAs from these repeats, which are readily processed into siRNA, had no apparent effect on otherwise wild type males. We then tested the survival of males with partial to complete loss of roX1 roX2 function. Double stranded RNA from 1.688^1A or 1.688^3C had little or no effect, but double stranded RNA from 1.688^3F dramatically rescued male survival.⁴⁵ Amazingly, recovery of adult males with a roX1 roX2 chromosome that supports fewer than 1% adult escapers was increased to over 30%! In parallel with increased survival, localization of the MSL proteins to the X chromosome was partially restored in these males. It is particularly intriguing that only siRNA from 1.688^3F was capable of rescuing roX1 roX2 males, as the regions tested share considerable similarity.⁴⁵ The 1.688^3F repeat cluster is located just a few hundred bases distal to roX1. Interestingly, roX1 is expressed several hours earlier than roX2 and is solely responsible for initial X recognition.⁴⁹ The adjacent location of 2 elements involved in X recognition suggests spatial integration of cooperating pathways.

The discovery of siRNAs that promote compensation is exiting, but interpretation of these findings is far from obvious. Dosage compensation and siRNA-dependent epigenetic processes both alter chromatin structure. Although it is unlikely that the siRNA pathway directly recruits the MSL complex, genetic interactions between siRNA and roX1 roX2 mutants suggests that both participate in X-recognition. One possibility is that the siRNA pathway modifies chromatin at the 1.688^X repeats, and this change promotes X recognition.

Do 1.688^X repeats enhance transcription of nearby genes?

One possibility is that the 1.688^X repeats act as enhancers to facilitate transcription of X-linked genes, which in turn might impact MSL recruitment. The transcriptional status of compensated genes, as well as epigenetic factors associated with activation, are factors in MSL recruitment.^22,23,50 In support of this idea, repetitive elements have been adapted for gene regulation in other organisms. Recent studies in mammals describe the evolution of Alu repeats to acquire the chromatin features of poised and active enhancers (enrichment for H3K4me1 or H3K27ac and P300).⁵¹ Transcripts originating from mammalian enhancer regions have been implicated in the recruitment of RNAP II to gene promoters.^52-54 We observe widespread transcription from the fly 1.688^X repeats, which could be related to enhancer function.⁴⁵ Enhancer activity might be modulated by siRNA-directed modification of 1.688^X repeats. Although usually implicated in repression, Ago2 has also been linked to activation in a few instances in flies and humans.^55-58 It would be interesting to determine whether epigenetic features of Drosophila enhancers are found at 1.688^X repeats.⁵⁹ In spite of the large number of genome-wide studies of chromatin modifications that have been performed in flies, highly repetitive sequences are typically removed from analysis because they are difficult or impossible to map.

Do 1.688^X repeats influence chromatin architecture or topology?

Features of the 1.688^X repeats suggest a potential role in modulation of chromosome architecture. The 1.688^X repeats are AT rich, a common feature of DNA that interacts with proteins of the nuclear matrix, such as satellite binding proteins and Topoisomerase II (Top2).⁶⁰ Top2, a major component of the nuclear matrix, has also been shown to bind pericentromeric 359 nt repeats in interphase nuclei.^61,62 Furthermore, Top2, and the associated DNA supercoiling factor SCF, both influence dosage compensation.^63,64 Top2 and SCF control DNA topology, suggesting that topological constraints on transcriptional initiation or elongation could be a factor in dosage compensation. For example, topological alterations might favor MSL complex association or spreading along the X chromosome.

Other aspects of nuclear organization may also be modulated by the 1.688^X repeats. The male X chromosome is reported to associate with the nuclear envelope.⁶⁵ Some nuclear pore proteins are enriched in regions of active expression, and these are particularly prominent on the male X chromosome, overlapping MSL-bound genes.⁶⁶ One possibility is that 1.688^X repeats mediate association of the X chromosome with the nuclear periphery. Long-range interactions between 1.688^X repeats at different locations on the X chromosome could also be instrumental in establishing a chromosome-specific interphase architecture. In support of this idea, the organization of the X chromosome is discernably different in males and females, and compensated genes are closer together in the interphase male nucleus than non-compensated regions.⁶⁷ An intriguing possibility is that siRNA-directed chromatin modifications regulate long-range interactions between 1.688^X repeats, or between repeats and the nuclear envelope. If 1.688^X repeats do participate in long-range interactions, this would bring roX1, immediately adjacent to 1.688^3F, close to distant regions of the X chromosome with 1.688^X repeats, providing a mechanism for MSL complexes assembled at roX1 to rapidly access distant regions of the X chromosome. The possibility of long range interactions between distant 1.688^X repeats could be tested by Chromosome Conformation Capture (3C), a molecular method to identify chromosome loops.⁶⁸

Insulators are another potential link between nuclear organization and X chromosome dosage compensation.^69,70 Insulators regulate enhancer access to promoters and govern folding of the genome.⁷⁰ Examples of repeats that act as insulators include long tandem repeats that mark the heterochromatin/euchromatin boundary on the Drosophila X chromosome, the human DXZ4 macrosatellite repeat that regulates the nuclear organization of the inactive X chromosome and the sub-telomeric D4Z4 repeats that oppose position effects.^69,71-74 Interestingly, both RNA and RNAi proteins, including Ago2, are implicated in the function of fly insulators.^56,75 The capacity of 1.688^X repeats to act as insulators could be determined using a well-established genetic assay.⁷⁶

Emerging functions of repetitive sequences

Repetitive sequences make up large portions of eukaryotic genomes, yet the known functions of these repeats are largely limited to their participation in chromosome compaction and segregation. Repeats are central to centromere formation, and rapidly evolving repeats are features of meiotic drive systems.⁷⁷ Indeed, sequences closely related to the euchromatic 1.688^X repeats have been implicated in both processes.^78,79 In contrast to these mitotic or meiotic roles, the ideas presented in this article highlight the potential for repetitive regions to act as regulators of gene expression in the interphase nucleus. The involvement of satellite repeats in X chromosome dosage compensation presents a powerful experimental system in which to dissect the regulatory potential of repetitive DNA.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Funding

The authors wish to thank the Wayne State University Graduate Enhancement Research Award, Graduate Enhancement Research Assistantship and a Thomas C. Rumble University Graduate Fellowship for support of DUM. VHM is supported by a Wayne State University Career Development award and NIGMS (GM 093110).