Chromosoma

, Volume 119, Issue 1, pp 99–113

Widespread regulation of gene expression in the Drosophila genome by the histone acetyltransferase dTip60

Authors

  • Corinna Schirling
    • Abteilung für Genetik, Zentrum für Medizinische BiotechnologieUniversität Duisburg-Essen
  • Christiane Heseding
    • Abteilung für Genetik, Zentrum für Medizinische BiotechnologieUniversität Duisburg-Essen
  • Franziska Heise
    • Abteilung für Genetik, Zentrum für Medizinische BiotechnologieUniversität Duisburg-Essen
  • Dörthe Kesper
    • Abteilung für Entwicklungsbiologie, Zentrum für Medizinische BiotechnologieUniversität Duisburg-Essen
  • Ansgar Klebes
    • Institut für Biologie-GenetikFreie Universität Berlin
  • Ludger Klein-Hitpass
    • Institut für ZellbiologieUniversitätsklinikum
  • Andrea Vortkamp
    • Abteilung für Entwicklungsbiologie, Zentrum für Medizinische BiotechnologieUniversität Duisburg-Essen
  • Daniel Hoffmann
    • Abteilung für Bioinformatik, Zentrum für Medizinische BiotechnologieUniversität Duisburg-Essen
  • Harald Saumweber
    • Abteilung Zytogenetik, Institut für BiologieHumboldt Universität Berlin
    • Abteilung für Genetik, Zentrum für Medizinische BiotechnologieUniversität Duisburg-Essen
Research Article

DOI: 10.1007/s00412-009-0247-z

Cite this article as:
Schirling, C., Heseding, C., Heise, F. et al. Chromosoma (2010) 119: 99. doi:10.1007/s00412-009-0247-z

Abstract

The MYST histone acetyltransferase (HAT) dTip60 is part of a multimeric protein complex that unites both HAT and chromatin remodeling activities. Here, we sought to gain insight into the biological functions of dTip60. Strong ubiquitous dTip60 knock-down in flies was lethal, whereas knock-down in the wing imaginal disk led to developmental defects in the wing. dTip60 localized to the nucleus in early embryos and was present in a large number of interbands on polytene chromosomes. Genome-wide expression analysis upon depletion of dTip60 in cell culture showed that it regulated a large number of genes in Drosophila, among which those with chromatin-related functions were highly enriched. Surprisingly, a significant portion of these genes were upregulated upon dTip60 loss, indicating that dTip60 has repressive as well as activating functions. dTip60 protein was directly located at promoter regions of a subset of repressed genes, suggesting a direct role in gene repression. Comparison of the gene expression signature of dTip60 downregulation with that of histone deacetylase inhibition with trichostatin A revealed a significant correlation, suggesting that the dTip60 complex recruits an HDAC-containing complex to regulate gene expression in the Drosophila genome.

Abbreviations

TSA

trichostatin A

HAT

histone acetyltransferase

HDAC

histone deacetylase

Introduction

The ability to temporally and spatially regulate gene expression is a prerequisite for development and differentiation in metazoan organisms. Gene regulation is a complex process that entails modulating accessibility of genes within chromatin to the transcription machinery, which is achieved by multiprotein complexes via chromatin modifications and nucleosome positioning. Chromatin structure is highly dynamic and is regulated by a variety of histone modifications, including acetylation, methylation, phosphorylation, and ubiquitination (Ehrenhofer-Murray 2004).

Histone acetylation by histone acetyltransferases (HATs) plays an important role in gene regulation. There are five families of HATs: Gcn5-related HATs, p300/CBP HATs, the general transcription factor HATs, nuclear hormone receptor-related HATs, and the MYST family of HATs, to which Tip60 belongs (Carrozza et al. 2003). The MYST family is named for its founding members MOZ, Ybf2/Sas3, Sas2, and Tip60. These proteins are highly conserved from yeast to humans and share a structurally similar catalytic HAT domain (Sanjuan and Marin 2001). The MYST family can be divided into three subgroups (Utley and Cote 2003): the first contains proteins carrying a PHD finger (MOZ and MORF), the second contains a zinc finger motif (HBO1, Sas2), and the third, comprising Esa1, dMOF, and Tip60, contains a chromodomain, which in dMOF has RNA-binding capacity (Akhtar et al. 2000).

Tip60 was first identified as a protein interacting with the Tat protein from HIV, and hence was named Tat interactive protein, 60 kDa (Kamine et al. 1996). Like many other HATs, Tip60 is part of a multiprotein complex whose basic composition is evolutionarily conserved (Doyon et al. 2004). Remarkably, the Tip60 complex in human and Drosophila combines two distinct chromatin-modifying activities, histone acetylation and chromatin remodeling (Sapountzi et al. 2006; Squatrito et al. 2006). These activities are found in two separate complexes in the yeast Saccharomyces cerevisiae, the HAT complex NuA4 and the ATP-dependent chromatin remodeler SWR1. In vitro, the Tip60 complex preferentially acetylates nucleosomal H2A and H4 (Ikura et al. 2000). Next to the Tip60 enzyme itself, the NuA4 part of the Tip60 complex contains the scaffolding protein TRRAP as well as several proteins containing motifs common to many chromatin proteins. All components have equivalents in the respective fly and yeast complexes (Sapountzi et al. 2006). A similar complex has been identified genetically in Caenorhabditis elegans (Ceol and Horvitz 2004). The chromatin remodeler portion of the Tip60 complex comprises, among others, the protein p400/Domino, which is a SWI2/SNF2-like ATPase, and the RuvB-like helicase proteins RuvBL1 and RuvBL2 (Ikura et al. 2000), whose homologs in Drosophila are called dPontin and dReptin, respectively (Gallant 2007).

Notably, the Tip60 complex from Drosophila (dTip60) co-purifies with H2B and the histone variant H2A.v (Kusch et al. 2004), which shares functions of both H2A.Z and H2A.X and becomes phosphorylated at sites of DNA damage. When incorporated into nucleosomes in vitro, phospho-H2A.v is the preferred substrate of the dTip60 complex and is acetylated at lysine 5 (K5). The dTip60 complex catalyzes replacement of acetylated phospho-H2A.v by unmodified H2A.v via its chromatin remodeler moiety (Kusch et al. 2004). Thus, the dTip60 complex is involved in a concerted acetylation-remodeling reaction that may be required to reverse the phospho-H2A.v mark imposed by DNA damage. In agreement with this, Tip60 complexes play a role in DNA damage repair in several organisms (Squatrito et al. 2006). More generally, this acetylation-remodeling activity may reflect the general activity of the Tip60 complex in transcriptional activation.

Several of the Tip60 complex subunits are part of other complexes or have functions independent of Tip60. For instance, dPontin and dReptin are also found in the Ino80 chromatin remodeling complex and the Uri complex (Gallant 2007). Furthermore, Reptin in Drosophila is a component of the PRC1 complex (Qi et al. 2006), whereas Pontin purifies with the Brahma chromatin remodeling complex (Diop et al. 2008). Their residence in several complexes may explain their antagonistic effects on Wingless signaling (Bauer et al. 2000).

Among the many roles of histone acetyltransferases, gene activation is probably the most widely recognized. HATs are generally thought to activate gene transcription by loosening chromatin compaction and by creating binding sites for the bromodomain moieties of chromatin proteins or transcription factor complexes through lysine acetylation, therefore facilitating access of the transcriptional machinery to the DNA (Ehrenhofer-Murray 2004). In agreement with this notion, the Tip60 complex is recruited to promoters by DNA-binding factors and acts as a transcriptional co-activator in several contexts. For example, Tip60 serves as a co-activator for nuclear hormone receptors (Gaughan et al. 2001), c-myc (Patel et al. 2004) and NF-κB (Bauer et al. 2000). The co-activator function in some instances is mediated by histone acetylation within the promoter region (Taubert et al. 2004), whereas in other cases, Tip60 acetylates the transcription factors and thus modulates their activity, resulting in context-dependent transcriptional outcomes (Sapountzi et al. 2006). As an example, Tip60 acetylates p53, which helps to distinguish between the cell-cycle arrest and apoptotic functions of p53 (Legube et al. 2004; Sykes et al. 2006; Tang et al. 2006). Tip60 also interacts with the Notch intracellular domain NICD (Gause et al. 2006). The cooperation between Tip60 and NICD might be mediated by the subunit TRRAP, and there is evidence that another HAT, dGCN5, is recruited (Gause et al. 2006). Again, Tip60 function is highly context dependent, since other data indicate that Tip60 inhibits Notch signaling (Kim et al. 2007).

Interestingly, Tip60 is not exclusively involved in gene activation, but also has a role in gene regulation as a co-repressor. While this may not require its HAT activity, Tip60 interaction with the DNA-binding factors STAT3 and KLF4 serves to recruit the histone deacetylase HDAC7 to gene promoters, which causes gene repression (Ai et al. 2007; Xiao et al. 2003).

The cooperation of Tip60 with some transcription factors seems to be highly dependent on the surrounding context in that it serves as a co-activator in some cases, whereas in others it is a transcriptional repressor. Myc recruits Tip60 to chromatin via the TRRAP subunit (Frank et al. 2003) and interacts with Pontin, thus regulating cellular growth and proliferation by gene repression (Bellosta et al. 2005). However, Tip60 not only interacts with myc, but may also counteract myc as a haplo-insufficient tumor suppressor by modulating oncogene-induced DNA-damage response during lymphomagenesis in both mouse and humans (Gorrini et al. 2007). These findings imply complex, context-dependent roles of the Tip60 complex in transcription, precluding a simple prediction for Tip60 function.

As a result of its roles in transcriptional regulation and double-strand break repair, Tip60 is involved in several physiological processes on an organismal level. In Drosophila, the dTip60 complex components Reptin, Domino, Mrg15, and Epc1 promote the generation of silent chromatin in that mutations in these genes suppress position effect variegation (Qi et al. 2006). In C. elegans, a genetically defined Tip60 complex regulates Ras-mediated vulval induction in cooperation with a chromatin remodeling complex carrying histone deacetylase activity (Ceol and Horvitz 2004).

In this study, we sought to investigate the function of dTip60 in flies. Strong dTip60 knock-down was lethal, and reduced dTip60 expression led to developmental defects in the wing. We characterized the localization of dTip60 in early embryos and larvae. Furthermore, we found dTip60 to be localized to interbands of salivary gland polytene chromosomes. Genome-wide expression analysis showed that dTip60 significantly influenced global gene expression in SL2 cells. Detailed analysis of dTip60 targets suggested a repressive in addition to an activating function of dTip60. Indeed, repression coincided with the direct physical presence of dTip60 at the promoter region of its target genes. Further analysis suggested that dTip60 exercised this repressive function in cooperation with a trichostatin-A-sensitive HDAC complex.

Results

Phenotypes of dTip60 knock-down in Drosophila

The observation that dTip60 is a HAT suggests a function in gene regulation. In order to investigate dTip60 function in the organismal context, we sought to characterize the consequences of the loss of dTip60 in flies. To this end, we generated an inducible dTip60-RNAi line using the GAL4/UAS system (Brand and Perrimon 1993). Flies were transformed with a UAS-dTip60 RNAi construct, and one transgenic line was obtained. This line, which carried the P-element insertion on the X chromosome, was further used to obtain eight autosomal P-element insertions through mobilization of this P-element by crossing it to a line expressing transposase. All lines showed the same phenotypes described here, indicating that the effects of dTip60 knock-down were independent of the chromosomal localization of the P-element. Flies with the UAS-dTip60 RNAi P-element were homozygously or hemizygously viable, fertile, and did not show signs of abnormal development.

In order to induce dTip60 RNAi expression, we crossed the insertion line to different GAL4 driver lines (Brand and Perrimon 1993). Knock-down of dTip60 caused pupal lethality under the control of either the daughterless promoter (da-GAL4) or the T80-GAL4 driver line at 29°C (Table 1), which was consistent with previous work (Zhu et al. 2007). At 27°C, only little lethality was observed in the T80 cross, and more than 90% of the survivors showed ectopic veins parallel to L5 (Fig. 1a). These ectopic veins varied slightly in length, but often occurred at a similar position close to the posterior crossvein. This effect was also observed using a GAL4-driver line under the control of the engrailed promoter (en-GAL4) at 29°C, although with a slightly reduced frequency (data not shown). Furthermore, this wing phenotype was enhanced when combining dTip60 knock-down with GAL4-driven expression of Dicer, a component of the RNAi machinery (Hammond et al. 2000). Here, en-GAL4 occasionally led to a notched wing phenotype (data not shown), which was consistent with dTip60 playing a role in Notch signaling (Gause et al. 2006). Occasionally, the formation of ectopic veins in other regions of the wing was observed, e.g. parallel to L1 and L2. A further decrease of temperature led to reduced lethality in da-GAL4 driven UAS-dTip60-RNAi flies and reduced the formation of ectopic veins in T80-GAL4 driven UAS-dTip60-RNAi flies. The observation of a temperature dependence of dTip60 RNAi phenotypes was in line with previous reports of temperature-dependent effects in GAL4 driver lines (Hayward et al. 2005; Ni et al. 2008).
Table 1

Phenotypes of dTip60 RNAi knock-down in flies

Cross (F0)

Gal4 expression line

Phenotype (F1)

da-Gal4 × dTip60-RNAi

da = daughterless promoter (early and ubiquitous expression)

29°C: 100% lethal

25°C: few escapers

22°C: no effect

 

T80-Gal4 × dTip60-RNAi

T80 promoter (late and ubiquitous expression)

29°C: >90% lethality

27°C: little lethality survivors ∼100% ectopic wing veins

25°C: no lethality ca. 50% ectopic wing veins

 

en-Gal4 × dTip60-RNAi

en = engrailed promoter (early expression in part of the posterior imaginal wing disk)

29°C: no lethality ca. 50% ectopic wing veins

https://static-content.springer.com/image/art%3A10.1007%2Fs00412-009-0247-z/MediaObjects/412_2009_247_Fig1_HTML.gif
Fig. 1

a Downregulation of dTip60 by RNAi resulted in defects in the Drosophila wing. a Wild-type wing with longitudinal veins L1–L5 indicated and b wing from a representative fly with dTip60-RNAi driven by T80-GAL4. Magnification of the area around L5 in wild-type (c) and dTip60-RNAi (d). b Effects of en-GAL4-driven dTip60 downregulation in the imaginal wing disk. a En expression in the posterior part of the imaginal wing disk. ben-GAL4-driven dTip60-RNAi caused reduced levels of dTip60 in the posterior part of imaginal wing disk. c Reduction of dTip60 protein level led to increased cell death as visualized by acridine orange staining

We further evaluated the effect of en-GAL4 driven dTip60 reduction in the wing imaginal disk. En is expressed in the posterior part of the wing imaginal disk (Fig. 1b), such that en-GAL4-driven dTip60 RNAi resulted in a weak, but visible reduction of dTip60 levels in the knock-down part of the wing disk as compared to the control wing compartment. This caused increased apoptosis as measured by staining with acridine orange (Fig. 1b). Quantitation of dTip60 levels using another imaginal wing disk-specific GAL4 driver line, C765, showed a reduction of dTip60 mRNA levels to 75% of the wild-type as measured by reverse transcription and quantitative polymerase chain reaction (PCR; data not shown). dTip60 was not quantified when reduced under en-GAL4, because reduction could at most be expected in one half of the wing disk. The relatively modest downregulation of dTip60 using these wing imaginal disk drivers might explain the mild wing phenotypes observed in adult flies.

Taken together, these findings showed that dTip60 was essential, which was in agreement with previous findings (Zhu et al. 2007). Furthermore, our results suggested a role for dTip60 during wing development. However, we note that off-target effects cannot be excluded at this point, since the in vivo knock-down was performed with only one UAS-dTip60 RNAi construct.

Nuclear localization of dTip60 during embryonic development

We were interested in evaluating the localization of dTip60 during embryonic development. The HAT function of dTip60 as well as its potential function in gene regulation might suggest a nuclear localization. In order to test this, an antibody was raised against a fragment spanning the N-terminal amino acids 29–304 of dTip60. Western blotting of Kc nuclear cell extracts showed a single band for dTip60 (Fig. 2a), and the anti-dTip60 signal was strongly reduced on polytene chromosomes as well as in wing imaginal disks of UAS-dTip60-RNAi larvae (data not shown, Fig. 1b), suggesting that the antibody was specific to dTip60.
https://static-content.springer.com/image/art%3A10.1007%2Fs00412-009-0247-z/MediaObjects/412_2009_247_Fig2_HTML.gif
Fig. 2

Localization of the dTip60 protein in Drosophila embryos. a Immunoblot of Kc nuclear extracts with an antibody against dTip60 or preimmune serum (PI). b Early embryos from wild-type flies were double labeled with DAPI (blue, a, d, g) and the dTip60 antibody (red, b, e, h). dTip60 was present in embryos after stage 3 (ac) and mainly stained the nuclei of yolk cells and the blastoderm (stage 4 and 5, di). c and d Nuclear localization of dTip60 in the blastoderm embryo. Heterochromatic regions showed a bright dot of DAPI staining (open arrow), whereas dTip60 staining localized to the nucleus and showed an even distribution indicative of a euchromatic localization (filled arrow)

In order to determine when and where dTip60 is present during embryonic development, we performed whole-mount fluorescent dTip60 antibody staining of Drosophila embryos. Significantly, dTip60 showed nuclear localization in embryos by stage 3 (Fig. 2b, (a–c)). dTip60 also localized to the nuclei of blastoderm embryos (Stage 4 and 5, Fig. 2b (d–f and g–i)). It displayed an even distribution on the DNA and was not enriched in peripheral heterochromatic chromosomal areas which were more strongly stained with DAPI (Fig. 2c and d, indicated by arrows). In situ staining of embryos at age 20–24 h showed ubiquitous expression of dTip60 RNA (supplementary Fig. 1). Altogether, these findings indicated a nuclear localization of dTip60 beginning at early stages of embryonic development.

dTip60 localized to interbands on polytene chromosomes

In order to gain further insight into both the chromosomal localization of dTip60 and its possible role in gene regulation, we stained polytene chromosomes in the salivary glands of Drosophila L3 larvae. Polytene chromosomes are organized in bands and interbands, which are the result of differential chromatin compaction. Inactive genes are frequently located in the compacted bands, whereas actively transcribed genes are thought to be located in the less compacted bands, interbands, and puffed regions (Ebert et al. 2006).

Significantly, staining of dTip60 on polytene chromosomes showed that it localized to a subset of interbands (Fig. 3), whereas the respective preimmune serum showed only background staining (supplementary Fig. 2). Overall, dTip60 was present in roughly 50% of the interbands, where it was either localized in compact dots within the interband, or in a narrow stripe next to a band. This localization pattern was consistent with a role for dTip60 in gene regulation.
https://static-content.springer.com/image/art%3A10.1007%2Fs00412-009-0247-z/MediaObjects/412_2009_247_Fig3_HTML.gif
Fig. 3

dTip60 localized to interbands of polytene chromosomes. Four different views of chromosome 3L showing the distal area with region 61 stained for DNA with Hoechst (blue, a, d, g, j) and for dTip60 (red, b, e, h, k)

dTip60-regulated genes were involved in chromatin-related processes

Next, we sought to identify genes that are regulated by dTip60. A priori, one might expect dTip60 to activate transcription of target genes because of its function as a histone acetyltransferase. To test this, we used RNAi to downregulate dTip60 in SL2 cells and compared the genome-wide expression profiles to those of mock-treated cells. The RNAi treatment reduced the level of dTip60 mRNA to 8%, as determined by semiquantitative PCR and real-time PCR (data not shown), inhibited cell proliferation and caused increased levels of cell death, because cell numbers did not increase in dTip60-depleted cultures (data not shown). This likely reflected the fact that dTip60 was essential in SL2 cells (data not shown, (Kusch et al. 2004)). Microarray-based expression analysis was performed after 5 days of dTip60 knock-down using three different RNAi constructs against dTip60 in order to minimize off-target effects (supplementary Fig. 3), and two biological replicates each were analyzed. We identified 1,064 reliable measured targets that showed consistent changes of expression levels upon downregulation of dTip60 (see Materials and methods for details). This indicated that a large proportion of genes in Drosophila SL2 cells were directly or indirectly regulated by a dTip60-containing complex, which emphasized the central role of dTip60 in gene regulation. Of those 1,064 genes, 601 were downregulated, indicating that dTip60 was required for their activation. Surprisingly, 463 genes were upregulated in the absence of dTip60, indicating either that dTip60 has a repressive function, or that downregulation of these genes was due to secondary effects of dTip60 loss.

We were interested in the biological function of the genes regulated by dTip60 and therefore performed an analysis of the Gene Ontology (GO) annotations of these potential target genes. Interestingly, there was a strong enrichment of genes with chromatin-related annotations among the repressed genes, i.e. the genes upregulated as a result of dTip60 loss (Table 2). This suggested that dTip60 plays an important role in the regulation of genes with chromatin-associated functions.
Table 2

GO category association of genes repressed by dTip60

GO category

Total number of genes in category

Number of dTip60-regulated genes in category

p value

Biological function

0006325/0006323

302

112

<10−100

Establishment and/or maintenance of chromatin architecture/DNA packaging

0006333

240

111

<10−100

Chromatin assembly or disassembly

0051276

391

116

<10−100

Chromosome organization and biogenesis

dTip60 displayed activating as well as repressive function at distinct genes

We further validated the effect of dTip60 on gene regulation by determining how the expression of the affected genes changed during the course of dTip60 depletion. RNAi-treated cells were harvested immediately and in intervals of 24 h, and the expression of dTip60 and selected genes that served as representative examples from the microarray analysis were tested by reverse transcription and quantitative PCR. Within 24 h of RNAi depletion, the level of dTip60 mRNA dropped to approximately 14% (Fig. 4a). In the next 2 days, the dTip60 mRNA level was further reduced, and this level remained stable for the last 2 days of the time course experiment. The maximum level of dTip60 mRNA reduction occurred after 3 days, with 7% of dTip60 mRNA remaining, which was in good agreement with the depletion experiments used for the microarray analysis.
https://static-content.springer.com/image/art%3A10.1007%2Fs00412-009-0247-z/MediaObjects/412_2009_247_Fig4_HTML.gif
Fig. 4

Changes in gene expression in SL2 cells after zero to 5 days of dTip60 RNAi treatment. Expression differences were measured by quantitative RT-PCR and are indicated as log2 (dTip60 RNAi/mock RNAi). Error bars indicate standard deviations from three independent experiments. a Analysis of genes identified as downregulated upon dTip60 RNAi in the microarray-based expression analysis. b Analysis of genes identified as upregulated in the microarray-based expression analysis

First, we evaluated the expression of genes that were identified as downregulated upon loss of dTip60 in the microarray analysis, suggesting an activating role for dTip60 at these genes. Candidate genes were selected based on the magnitude of their change in gene expression and on whether they were present upon RNAi treatment against dTip60 or mock treatment. Ac3 displayed a moderate downregulation in the microarray experiments, with its log2 (dTip60 RNAi/mock RNAi) being below −0.5. All other selected genes (His4r, CG16888, nvy, and CG11899) showed a more pronounced downregulation with a log2 (dTip60 RNAi/mock RNAi) below −1.5 in all three microarray experiments.

The analysis of these candidates in the time course experiment gave a mixed picture. All genes displayed downregulation after 4 days of RNAi treatment (Fig. 4a). Expression levels of some genes, e. g. His4r and nvy, showed fluctuation in their expression for reasons that were not apparent.

We next investigated the expression levels of genes that showed upregulation upon loss of dTip60 in the microarray analysis, suggesting a possible repressive role for dTip60 at those genes. A representative sample of seven genes was selected. In the microarray analysis, mthl2 displayed a prominent upregulation with the log2 (dTip60 RNAi/mock RNAi) above 4.85 in all three microarray experiments. CG10131 showed a moderate upregulation (log2 (dTip60 RNAi/mock RNAi) above 0.37). Fz2 was upregulated by a log2 (dTip60 RNAi/mock RNAi) of more than 1.3. The other selected genes (CG14273, CG8942, CG5397, and drip) displayed an intermediate upregulation, with their log2 (dTip60 RNAi/mock RNAi) being above 2. All seven genes were present in both dTip60 knock-down and mock-treated cells and showed robust expression.

In the time course depletion of dTip60, six of these seven genes showed upregulation after 2 days of RNAi treatment (CG14273, CG8942, CG5397, drip, CG10131, and mthl2), and their expression further increased up to 5 days of treatment (Fig. 4b). Of these, mthl2 showed the strongest upregulation, which is in accordance with the findings of the gene expression analysis. This suggested that these genes might be direct targets of dTip60. In contrast, the gene Fz2 showed only moderate and delayed upregulation, suggesting that it might be indirectly affected by dTip60 loss.

Taken together, this analysis showed that the expression of six genes—CG14273, CG8942, CG5397, CG10131, drip, and mthl2—increased as a result of dTip60 loss, suggesting a direct role for dTip60 in their repression. In addition, the majority of genes downregulated in the absence of dTip60 showed a moderate downregulation, thus establishing them as potential direct targets of dTip60.

dTip60 was physically located at activated as well as at repressed genes

If dTip60 directly activates or represses target genes, as suggested by the above observation, then one would expect dTip60 to be physically associated with the promoter region of these potential target genes. In order to test this, we measured the association of dTip60 with its target genes by chromatin immunoprecipitation (ChIP) analysis on three independent biological replica. First, we evaluated whether dTip60 was present at promoter regions of activated genes. Indeed, dTip60 was enriched on the promoter sequences of CG11899, Ac3, nvy, CG16888, and His4r, whereas no enrichment was observed on an intergenic control region (Fig. 5). Next, we asked whether dTip60 was associated with promoter regions of repressed genes. Interestingly, we found that dTip60 was enriched at G5397, drip, CG8942, mthl2, Fz2, CG14273, and CG10131. Taken together, this showed that dTip60 was directly located at promoter regions of all investigated activated as well as repressed genes, indicating that dTip60 might be directly involved in transcriptional repression next to its role in gene activation.
https://static-content.springer.com/image/art%3A10.1007%2Fs00412-009-0247-z/MediaObjects/412_2009_247_Fig5_HTML.gif
Fig. 5

Localization of dTip60 protein at promoter regions of dTip60-regulated genes by ChIP analysis. dTip60 binding is shown as enrichment in ChIP experiments with the dTip60 antibody or preimmune serum (PI) relative to the input DNA. Error bars indicate standard deviations of three biological replica. Control, enrichment on a fragment in the intergenic region downstream of the 3′ end of the Hsp70 gene located at 87C

Cooperation of dTip60 and a TSA-sensitive HDAC-containing complex in gene regulation

HATs are generally thought to be involved in gene activation. However, our results suggested that dTip60 also played a direct role in the repression of a number of genes in the embryonic SL2 cell line. In humans, there is evidence that hTip60 acts in a complex with a histone deacetylase, which mediates a repressive function (Ai et al. 2007; Xiao et al. 2003). Therefore, we hypothesized that in Drosophila, dTip60 might also work with an HDAC to repress distinct genes.

In order to test this hypothesis, we made use of the work of Foglietti and colleagues, who performed genome-wide expression analysis of SL2 cells treated with Trichostatin A (TSA), which inhibits class I and II, but not class III HDACs (Foglietti et al. 2006). We compared the genes upregulated (n = 463) or downregulated (n = 601) upon dTip60 loss to those upregulated (n = 348 to 651, depending on the length of treatment) or downregulated (n = 234 to 377) as a result of treatment with TSA. Importantly, there was a significant overlap between those groups of genes (p < 0.000245), and this overlap increased over the time of TSA treatment. Eighty-seven genes that appeared upregulated after 5 days of TSA treatment were also identified as upregulated upon dTip60 loss (Table 3). For downregulated genes, there was an overlap of 90 genes (Table 4). The respective p values (8.51 × 10−29 for upregulated genes, and 5.44 × 10−41 for downregulated genes) indicate that these findings are statistically highly significant. In summary, these results suggested that dTip60 cooperates with a TSA-sensitive HDAC to regulate gene expression.
Table 3

Comparison of genes upregulated upon dTip60 depletion with TSA treatment as determined by Foglietti et al.

Genes upregulated upon dTip60 loss (n = 463)

 

Total genes upregulated by TSA treatment

Overlap with dTip60

p value

TSA 12 h

348

26

0.000245

TSA 24 h

308

35

5.24 × 10−10

TSA 3 days

403

48

4.39 × 10−14

TSA 5 days

651

87

8.51 × 10−29

Table 4

Comparison of genes downregulated upon dTip60 depletion with TSA treatment as determined by Foglietti et al.

Genes downregulated upon dTip60 loss (n = 601)

 

Total genes downregulated by TSA treatment

Overlap with dTip60

p value

TSA 12 h

234

37

2.15 × 10−11

TSA 24 h

253

57

1.08 × 10−24

TSA 3 days

214

61

1.59 × 10−32

TSA 5 days

377

90

5.44 × 10−41

One possible explanation for the overlap between the effect of dTip60 knock-down and TSA treatment is that TSA treatment would downregulate dTip60 expression. However, this was not the case. Alternatively, it was possible that dTip60 knock-down would repress the expression of one or several HDACs. In fact, HDAC1 showed a slight repression upon dTip60 depletion in that it was expressed at 64% of the wild-type level, and the other HDACs showed no regulation by dTip60. Since this constitutes a minor change in HDAC1 levels, it seems unlikely that this degree of HDAC1 repression would explain the significant overlap between the two data sets. Therefore, in summary, we favor the interpretation that dTip60 cooperates with HDACs to regulate gene expression in the Drosophila genome.

Discussion

In this study, we have investigated the function of the histone acetyltransferase dTip60 in Drosophila melanogaster. Strong reduction of dTip60 by RNAi was lethal, whereas a moderate knock-down led to defects in wing development, and reduction of dTip60 in the wing imaginal disk caused increased apoptosis. Furthermore, dTip60 was present in the nucleus of Drosophila embryos and was localized to about 50% of interbands on polytene chromosomes. Genome-wide expression analysis showed that dTip60 regulated over 1,000 genes in Drosophila SL2 cells, and that many of these genes had chromatin-related functions. Interestingly, reduction of dTip60 levels caused repression at some genes, but activation at others, implying a role for dTip60 both in gene activation and repression. Furthermore, dTip60 was physically present at repressed as well as activated genes, suggesting a direct role in their regulation. Statistical analysis of expression signatures indicated that dTip60 worked in cooperation with a TSA-sensitive HDAC complex in gene regulation.

dTip60 is part of a dual-activity HAT/chromatin remodeler complex (Squatrito et al. 2006), and one therefore might posit a role for dTip60 in activating genes. Indeed, our analysis shows that dTip60 is required for the activation of a large number of genes in the Drosophila genome. This activation may be due to the dTip60 complex performing histone acetylation and chromatin remodeling in the gene promoters and thus helping in the recruitment of the transcriptional machinery, but may also be independent of HAT activity (Gavaravarapu and Kamine 2000). In addition, our data also reveal a substantial contribution of dTip60 to gene repression, and we show physical presence of dTip60 at a number of repressed genes. Despite the prominent role of HATs in gene activation, several studies in humans and mice indicate that Tip60 has a repressive function. Tip60 acts as a co-repressor for STAT3 by recruitment of HDAC7 (Xiao et al. 2003), and it represses KLF4-dependent promoter activity together with HDAC7 (Ai et al. 2007). Furthermore, mutations in components encoding dTip60 complex subunits suppress position effect variegation in flies (Qi et al. 2006), indicating a role for the complex in heterochromatin formation and gene repression.

How then can a HAT have repressive function? The dTip60 complex acetylates phosphorylated H2A.v and replaces it with the unmodified form (Kusch et al. 2004). H2A.v in Drosophila is present at heterochromatin and is involved in the formation of repressed chromatin. Since phospho-H2A.v is not only associated with DNA damage repair, but also with general transcriptional activation in flies (Swaminathan et al. 2005), one might then speculate that the dTip60 complex is globally required to deposit H2A.v on repressed genes. Loss of dTip60 would therefore result in a loss of H2A.v and hence a loss of gene repression.

Notably, one homolog of dTip60, Sas2 in S. cerevisiae, has also been implicated in gene silencing at yeast telomeres and the silent mating-type loci (Ehrenhofer-Murray et al. 1997; Meijsing and Ehrenhofer-Murray 2001; Reifsnyder et al. 1996). However, Sas2’s telomeric silencing function stems from its role as a boundary factor in preventing the spreading of heterochromatin into euchromatic regions (Kimura et al. 2002; Suka et al. 2002), rather than from direct promoter histone acetylation. In addition, the boundary function of Sas2 may entail the deposition of H2A.Z in subtelomeric regions (Shia et al. 2006). Therefore, by analogy, dTip60 may also act in gene repression in flies as a boundary factor by restricting heterochromatin to its designated genomic region. In line with this, dTip60 localized to the border of interbands on polytene chromosomes, which is compatible with a barrier function for dTip60. However, our ChIP analysis showed localization of dTip60 at repressed genes, arguing for a more direct role of dTip60 in gene repression. The interband localization of dTip60 is also consistent with its role in gene activation.

Another possibility for dTip60’s role in gene repression is that it is not the HAT function of dTip60 that is relevant for repression. Notably, Tip60 can acetylate histones in the nucleosomal context only when it is within a protein complex, but not as a recombinant protein (Ikura et al. 2000), suggesting that other accessory proteins (and perhaps chromatin remodeling) are required for its activity. In consequence, the activity of these factors may be more important for the repressive activity of the dTip60 complex than the HAT activity alone, for instance by moving nucleosomes over the promoter of genes to a position repressive to transcription. Also, dTip60-interacting factors may inhibit its HAT activity and thus may predominate in the epigenetic mark the complex leaves on chromatin.

As an extension of this model, our data indicate that dTip60 cooperates with a TSA-sensitive HDAC complex in gene regulation at a subset of genes. Interestingly, we see this correlation for genes both up- and downregulated by dTip60 and TSA treatment, implying that they cooperate in gene activation as well as in repression. Therefore, the dTip60 complex may recruit an HDAC to the promoters to perform histone deacetylation, which then causes gene regulation. The final transcriptional output of a gene may then result from the sum of (histone) acetylation by dTip60 and deacetylation by this HDAC. In support of this cooperation of dTip60 and an HDAC-containing complex, mutations in HDAC1, like mutations in dTip60 components, cause defects in position effect variegation (Mottus et al. 2000). Furthermore, genes that are weakly activated by dTip60/HDAC may be in a “poised” state of gene expression and carry both transcription-activating and transcription-repressing chromatin marks. Such genes could then be rapidly and strongly activated under inducing conditions.

An alternative explanation for dTip60’s role in gene repression is that dTip60 acetylates targets other than histones, and that these are relevant for dTip60’s repressive function, whereas histone acetylation might be important for gene activation. Tip60 is known to acetylate other factors, for instance p53, which modulates its transcriptional output (Sykes et al. 2006; Tang et al. 2006).

For the regulation of genes by dTip60 that are insensitive to TSA treatment, dTip60 most likely cooperates with other co-repressors, for instance TSA-resistant HDACs, as well as with other, as yet unidentified co-repressors to control gene expression.

dTip60 joins a growing number of chromatin regulators that have a context-dependent influence on gene expression in that they contribute both to gene activation and repression, although it remains to be determined how many of the dTip60 targets identified here are directly regulated by dTip60. Notably, the Drosophila MYST HAT Chameau stimulates transcription that depends on the transcription factor AP-1, and this coactivation is accompanied by increased histone acetylation at the gene promoter (Miotto et al. 2006). Conversely, mutations in chameau cause defects in position effect variegation and in the maintenance of Polycomb-dependent gene silencing, suggesting a parallel role for Chameau in gene repression (Grienenberger et al. 2002). Similarly, the dTip60 homolog MOF is best known for upregulating expression of X-linked genes in male flies through acetylation of H4 K16 (Mendjan and Akhtar 2007). However, MOF is also required for the downregulation of a substantial number of genes on the autosomes of male as well as female flies (Kind et al. 2008). Furthermore, pericentric heterochromatin in Drosophila contains H4 K12 acetylation that depends on the presence of H2A.v, suggesting a role for this modification in establishing pericentric gene silencing (Swaminathan et al. 2005). Thus, there seems to be no simple correlation between the enzymatic activity of a regulator and its effect on gene expression at an individual genomic location. The overall transcriptional outcome likely depends on which cofactors interact with dTip60 at a given promoter, and what the sum of epigenetic modifications is.

In summary, with this study we have expanded the view of dTip60’s function and found it to cause widespread repression in addition to gene activation. Further work will be required to determine the molecular basis for this and to identify the co-regulators it cooperates with to exert these divergent functions.

Material and methods

Drosophila stocks

Fly strains were maintained at room temperature or as indicated on instant medium Formula 24-4 (Carolina Biological Supply Company). Oregon R was used as wild-type stock. Ectopic expression was achieved using the GAL4/UAS system (Brand and Perrimon 1993) with the GAL4-lines Da-GAL4 (w1118; P{da-GAL4.w-}3), T80-GAL4 (w*; P{GawB}T80/CyO) (Gortchakov et al. 2005) and en-GAL4 (y1 w*; P{en2.4-GAL4}e16E P{UAS-FLP1.D}JD1, Negeri et al. 2002).

For dTip60-RNAi in flies, two dTip60 fragments were amplified from genomic DNA. The sense fragment spanned the base pairs 191–1,222, while the antisense fragment spanned the base pairs 250–1,275 (Supplementary Fig. 3). The fragments were cloned in the P-element transformation vector pUAST (Brand and Perrimon 1993) in sense and antisense orientation to generate the vector pUAST-Tip60-RNAi, which was injected into w1118 embryos to establish a homozygous transgenic line (named UAS-dTip60-RNAi No. 1). Germline transformation was performed as described previously (Gortchakov et al. 2005). As a genomic source of transposase for P-element mobilization, the strain P[ry+ ∆2-3]99B was used (Robertson et al. 1988).

Wings were briefly fixed in ethanol after dissection and mounted in 87% glycerol with 2% n-propylgalleat.

Antibody generation and immunohistochemistry

A full-length dTip60 cDNA vector (LD31064) was obtained from Open Biosystems. A dTip60 fragment corresponding to amino acids 29–341 was amplified from LD31064 and cloned into pET15b (Novagen). The recombinant His-tagged dTip60 fusion protein was purified using Talon beads (Clontech) according to manufacturer’s instruction. Polyclonal dTip60-specific antibodies were raised in guinea pig.

Western blot analysis on Kc nuclear extracts was performed as described previously (Eggert et al. 2004). Proteins were separated on 10% sodium dodecyl sulfate (SDS)-polyacrylamide gels and blotted to nitrocellulose membranes, blocked, and incubated with anti-dTip60 (1:5,000) or preimmune serum (1:200) as primary and goat anti-guinea pig IgG (peroxidase conjugated, 1:200, Sigma) as secondary antibody.

Wing imaginal disks were dissected from wandering third instar larvae (w1118—control, Fig. 2b (a) or engrailed-GAL4, UAS-dTip60-RNAi, Fig. 2b (b–d)) and fixed with 4% paraformaldehyde for immuno-labeling following standard procedures. Primary antibodies: guinea pig anti-dTip60 (1:400), mouse anti-engrailed/invected 4D9 (Developmental Studies Hybridoma Bank, University of Iowa). Secondary antibodies: goat anti-guinea pig Cy3 and goat anti-mouse (both 1:500, Dianova). Cell death was analyzed by acridine orange (Invitrogen, USA) vital staining as described (Kramer and Staveley 2003).

Whole-mount fluorescent staining of Drosophila embryos was performed as described previously (Kesper et al. 2007) with anti-dTip60 in 1:400 dilution and goat anti-guinea pig Cy3 (1:200, Dianova). Images were taken using a Zeiss Axiovert 200 M microscope and electronically processed with the MetaMorph program (Molecular Devices).

Whole-mount in situ hybridization was carried out essentially as described (Tautz and Pfeifle 1989). Digoxigenin-labeled probes were generated by digesting the dTip60 cDNA vector LD31064 with XhoI or EcoRI followed by transcription with T7 (antisense probe) or Sp6 polymerase (control), respectively.

Polytene chromosomes were prepared from wandering third instar larvae and were fixed and stained as described (Eggert et al. 2004) with anti-dTip60 (1:5,000) or preimmune serum (1:200) and goat anti-guinea pig Cy3 (1:200). DNA on chromosomes was counterstained using Hoechst or DAPI dye, as indicated. For Fig. 3, dual wavelength images were taken on a Deltavision Spectris optical sectioning microscope (Applied Precision, USA) and processed using the Softworx iterative deconvolution software. Images shown are ∼400 nm optical sections. For supplementary Fig. 2, images were taken on a Zeiss Axiovert 200 M microscope using the MetaMorph software (Molecular Devices). Slides were prepared in the same batch, and images were taken at identical exposition. No further image processing was applied to ensure comparability of images.

dTip60 knock-down in SL2 cells

SL2 cells were cultured in InsectExpress Sf9-S2 medium (PAA the cell culture company) with 10% fetal calf serum (FCS) at 26°C. Three double-stranded RNA (dsRNA) corresponding to base pairs 190–807 (dsRNA1), 603–1,116 (dsRNA2), and 1,321–1,618 (dsRNA3) of dTip60 cDNA was synthesized with the MEGAscript Kit (Ambion) following the manufacturer’s instruction and precipitated with LiCl. The position of these dsRNAs relative to the dTip60 cDNA obtained from the cDNA clone LD31064 is indicated in supplementary Fig. 3. A dsRNA for enhanced green fluorescent proteins (eGFP) was synthesized as a control as described (Muller et al. 2006). SL2 cells, 106, were incubated with 10 µg dsRNA in 1 ml serum-free medium with Pen/Strep with agitation at room temperature for 10 min and for 50 min at 26°C without agitation. Two milliliters medium with Pen/Strep and FCS to a final concentration of 10% were added. For microarray analysis, cells were harvested after 5 days. For the time course experiment, cells were harvested at the indicated timepoints. RNA was extracted using the NucleoSpin RNA II kit (Macherey–Nagel).

For detailed analysis of selected genes in Fig. 4, dsRNA3 was used for knock-down of dTip60. Genome-wide gene expression profiling had demonstrated that the selected genes belonged to the large group of genes who responded equally well to knock-down of dTip60 by all three dsRNA constructs.

Genome-wide gene expression profiling

Expression profiling was performed on knock-downs of dTip60 with three different RNAi constructs and eGFP-treated samples (as a control) in two biological replicates each and two technical replicates of eGFP alone using Affymetrix GeneChip Drosophila 2.0 arrays. RNA was amplified, labeled, and hybridized according to the manufacturer’s instructions and scanned in a GeneChip 3000 scanner with G7 update. Array images (CEL files) were processed to determine signals and detection calls (Present, Absent, Marginal) for each probeset using the Affymetrix GCOS1.4 software (MAS 5.0 statistical algorithm). Scaling across all probesets to an average intensity of 1,000 was used to compensate for variations in the amount and quality of the cRNA samples and other experimental variables of non-biological origin. Pairwise comparisons of treated versus control samples was carried out with GCOS1.4, which calculates the change as a signal log ratio (basis 2) and the significance of each change in gene expression (change p value) based on a Wilcoxon ranking test.

Probesets exhibiting a significant increase or decrease (p ≤ 0.002) were identified by filtering using the Affymetrix Data Mining Tool 3.0. To limit the number of false positives, further target identification was restricted to those probesets that received at least one present detection call in the treated/control pair. A total of 12 target lists obtained in 12 cross-comparisons were merged [batch1: eGFP versus Tip60-RNAi (bp 1,321–1,618; dsRNA3), batch two: eGFP versus Tip60 RNAi (bp 190–807; dsRNA1), batch2: eGFP versus Tip60 RNAi (bp 603–1,116; dsRNA2)]. Genes were determined as regulated by dTip60 only if they appeared regulated in at least 10 of these 12 comparisons. Of the 463 genes upregulated by dTip60 loss, 313 were present in 12 of 12 comparisons, and 76 and 74 were present in 11 or 10 of 12 comparisons, respectively. Of the 601 downregulated genes, 359, 124, and 118 were present in 12, 11, or 10 comparisons, respectively. Therefore, in summary, 1,117 probesets representing 1,064 genes that exhibited increased or decreased change calls in ≥10 of these comparisons were identified and defined as consistently regulated targets.

Microarray data is accessible at NCBI GEO, GEO accession number GSE 13878.

Quantitative real-time PCR

cDNA was synthesized from total RNA using the SuperScript III kit (Invitrogen). Real-time PCR was performed using real master mix (5Prime) containing Sybr Green and analyzed in a Rotor Gene 3000 (Corbett). Samples were cycled 45 times for 15 s at 94°C, 30 s at 56°C, and 40 s at 68°C, and the Ct value for each reaction was determined. For expression analysis, the difference in Ct was determined with normalization to actin, using the following equation: ∆∆Ct = ∆Ct (dTip60 − RNAi) − ∆Ct (mock-treated). For ChIP analysis, a standard curve of input samples was used to calculate the amount of DNA precipitated relative to the input DNA. Primer sequences are available from the authors upon request. The position of the sequence used for detection of dTip60 expression levels is indicated in supplementary Fig. 3. For ChIP analysis, primer sequences usually spanned 250–300 bp, with the forward primer about 200 bp upstream of the ORF of the analyzed gene.

Chromatin immunoprecipitation

ChIP was performed according to standard protocols (Wu et al. 2003). Briefly, 108 SL2 cells were crosslinked with 1% fomaldehyde for 10 min at room temperature. Cells were harvested by centrifugation and incubated for 10 min in SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris–HCl, pH 8.1). ChIP was performed using 6 µl dTip60 antiserum or preimmune serum.

Bioinformatics

For GOstat analyses of differentially regulated gene products, the software at http://gostat.wehi.edu.au (Beissbarth and Speed 2004) was used. For comparison of gene expression signatures of dTip60 RNAi and TSA-treated samples, p values were determined using a two-sided Fisher’s exact test as implemented in the statistical software R, version 2.7.1 (R Development Core Team 2008). Bonferroni correction for multiple testing led to significance levels of α = 0.0028 instead of 0.05.

Acknowledgements

We are indebted to Peter B. Becker and Gunter Reuter for reagents. Bodo M. H. Lange and Hannah Müller are gratefully acknowledged for advice concerning RNAi in Drosophila cultured cells. Lothar Vassen is thanked for advice on the analysis of microarray experiments. We further thank Irina Passow, Renate Gienap, and Christiane Vole for excellent technical assistance and the Ehrenhofer-Murray lab for many helpful discussions. This project was funded by the Max-Planck-Society and the University of Duisburg-Essen.

Supplementary material

412_2009_247_Fig1_ESM.jpg (86 kb)
Supplementary Fig. 1

Ubiquitous expression of dTip60 RNA during embryogenesis. Lateral views of embryos hybridized with digoxigenin-labeled RNA probes of dTip60 in antisense (a, b) and, as a control, in sense (c, d) orientation. Enrichment of staining at mesoderm and endoderm is likely to be due to the thickness of tissue in these regions (JPEG 85 kb)

412_2009_247_Fig1_ESM.tif (7.1 mb)
High resolution (TIFF 7291 kb)
412_2009_247_Fig2_ESM.jpg (50 kb)
Supplementary Fig. 1

Ubiquitous expression of dTip60 RNA during embryogenesis. Lateral views of embryos hybridized with digoxigenin-labeled RNA probes of dTip60 in antisense (a, b) and, as a control, in sense (c, d) orientation. Enrichment of staining at mesoderm and endoderm is likely to be due to the thickness of tissue in these regions (JPEG 85 kb)

412_2009_247_Fig2_ESM.tif (911 kb)
High resolution (TIFF 911 kb)
412_2009_247_Fig3_ESM.jpg (86 kb)
Supplementary Fig. 1

Ubiquitous expression of dTip60 RNA during embryogenesis. Lateral views of embryos hybridized with digoxigenin-labeled RNA probes of dTip60 in antisense (a, b) and, as a control, in sense (c, d) orientation. Enrichment of staining at mesoderm and endoderm is likely to be due to the thickness of tissue in these regions (JPEG 85 kb)

412_2009_247_Fig3_ESM.tif (7 mb)
High resolution (TIFF 7159 kb)

Copyright information

© Springer-Verlag 2009