@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix skos: . vivo:departmentOrSchool "Science, Faculty of"@en, "Zoology, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:identifierCitation "Epigenetics & Chromatin. 2015 May 07;8(1):18"@en ; ns0:rightsCopyright "Dupont et al.; licensee BioMed Central."@en ; dcterms:creator "Dupont, Camille A"@en, "Dardalhon-Cuménal, Delphine"@en, "Kyba, Michael"@en, "Brock, Hugh W"@en, "Randsholt, Neel B"@en, "Peronnet, Frédérique"@en ; dcterms:issued "2016-01-13T18:21:38Z"@*, "2015-05-07"@en ; dcterms:description """Background: Cyclins and cyclin-dependent kinases (CDKs) are essential for cell cycle regulation and are functionally associated with proteins involved in epigenetic maintenance of transcriptional patterns in various developmental or cellular contexts. Epigenetic maintenance of transcription patterns, notably of Hox genes, requires the conserved Polycomb-group (PcG), Trithorax-group (TrxG), and Enhancer of Trithorax and Polycomb (ETP) proteins, particularly well studied in Drosophila. These proteins form large multimeric complexes that bind chromatin and appose or recognize histone post-translational modifications. PcG genes act as repressors, counteracted by trxG genes that maintain gene activation, while ETPs interact with both, behaving alternatively as repressors or activators. Drosophila Cyclin G negatively regulates cell growth and cell cycle progression, binds and co-localizes with the ETP Corto on chromatin, and participates with Corto in Abdominal-B Hox gene regulation. Here, we address further implications of Cyclin G in epigenetic maintenance of gene expression. Results: We show that Cyclin G physically interacts and extensively co-localizes on chromatin with the conserved ETP Additional sex combs (ASX), belonging to the repressive PR-DUB complex that participates in H2A deubiquitination and Hox gene silencing. Furthermore, Cyclin G mainly co-localizes with RNA polymerase II phosphorylated on serine 2 that is specific to productive transcription. CycG interacts with Asx, PcG, and trxG genes in Hox gene maintenance, and behaves as a PcG gene. These interactions correlate with modified ectopic Hox protein domains in imaginal discs, consistent with a role for Cyclin G in PcG-mediated Hox gene repression. Conclusions: We show here that Drosophila CycG is a Polycomb-group gene enhancer, acting in epigenetic maintenance of the Hox genes Sex combs reduced (Scr) and Ultrabithorax (Ubx). However, our data suggest that Cyclin G acts alternatively as a transcriptional activator or repressor depending on the developmental stage, the tissue or the target gene. Interestingly, since Cyclin G interacts with several CDKs, Cyclin G binding to the ETPs ASX or Corto suggests that their activity could depend on Cyclin G-mediated phosphorylation. We discuss whether Cyclin G fine-tunes transcription by controlling H2A ubiquitination and transcriptional elongation via interaction with the ASX subunit of PR-DUB."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/56474?expand=metadata"@en ; skos:note "RESEARCHDrosophila Cyclin G and eof gene expression duringchBackground studies of cell cycle regulators, have revealed that they areDupont et al. Epigenetics & Chromatin (2015) 8:18 DOI 10.1186/s13072-015-0008-6serine 2 by the positive transcription elongation factor b(P-TEFb), formed by CDK9 and Cyclin T or K [5-7].F-75005 Paris, FranceFull list of author information is available at the end of the articleCyclins and cyclin-dependent kinases (CDKs) are classic-ally described as essential regulators of the cell cycle. How-ever, the discovery of new cyclins, along with extensiveinvolved in diverse processes ranging from metabolism tostem cell self-renewal, and play key roles in regulationof transcription [1,2]. Indeed, cyclin/CDK complexesorchestrate the transcription cycle by dynamically phos-phorylating the heptapeptide repeats which form the C-terminal domain (CTD) of the largest RNA polymeraseII (RNA Pol II) subunit [3,4]. Notably, transcriptionalelongation requires phosphorylation of the CTD on* Correspondence: neel.randsholt@upmc.fr; frederique.peronnet@upmc.fr1Sorbonne Universités, UPMC Univ Paris 06, Institut de Biologie Paris-Seine(IBPS), UMR 7622, Developmental Biology, 9, quai Saint-Bernard, F-75005Paris, France2CNRS, IBPS, UMR 7622, Developmental Biology, 9, quai Saint-Bernard,AbstractBackground: Cyclins and cyclin-dependent kinases (CDKs) are essential for cell cycle regulation and are functionallyassociated with proteins involved in epigenetic maintenance of transcriptional patterns in various developmental orcellular contexts. Epigenetic maintenance of transcription patterns, notably of Hox genes, requires the conservedPolycomb-group (PcG), Trithorax-group (TrxG), and Enhancer of Trithorax and Polycomb (ETP) proteins, particularly wellstudied in Drosophila. These proteins form large multimeric complexes that bind chromatin and appose orrecognize histone post-translational modifications. PcG genes act as repressors, counteracted by trxG genes thatmaintain gene activation, while ETPs interact with both, behaving alternatively as repressors or activators. DrosophilaCyclin G negatively regulates cell growth and cell cycle progression, binds and co-localizes with the ETP Corto onchromatin, and participates with Corto in Abdominal-B Hox gene regulation. Here, we address further implicationsof Cyclin G in epigenetic maintenance of gene expression.Results: We show that Cyclin G physically interacts and extensively co-localizes on chromatin with the conserved ETPAdditional sex combs (ASX), belonging to the repressive PR-DUB complex that participates in H2A deubiquitination andHox gene silencing. Furthermore, Cyclin G mainly co-localizes with RNA polymerase II phosphorylated on serine 2 that isspecific to productive transcription. CycG interacts with Asx, PcG, and trxG genes in Hox gene maintenance, and behavesas a PcG gene. These interactions correlate with modified ectopic Hox protein domains in imaginal discs, consistent witha role for Cyclin G in PcG-mediated Hox gene repression.Conclusions: We show here that Drosophila CycG is a Polycomb-group gene enhancer, acting in epigeneticmaintenance of the Hox genes Sex combs reduced (Scr) and Ultrabithorax (Ubx). However, our data suggest thatCyclin G acts alternatively as a transcriptional activator or repressor depending on the developmental stage, thetissue or the target gene. Interestingly, since Cyclin G interacts with several CDKs, Cyclin G binding to the ETPs ASXor Corto suggests that their activity could depend on Cyclin G-mediated phosphorylation. We discuss whetherCyclin G fine-tunes transcription by controlling H2A ubiquitination and transcriptional elongation via interactionwith the ASX subunit of PR-DUB.Keywords: Polycomb, Trithorax, Cyclin G, HomeoticCamille A Dupont1,2, Delphine Dardalhon-Cuménal1,2, Miand Frédérique Peronnet1,2*© 2015 Dupont et al.; licensee BioMed CentraCommons Attribution License (http://creativecreproduction in any medium, provided the orDedication waiver (http://creativecommons.orunless otherwise stated.Open Accesspigenetic maintenancedevelopmentael Kyba3, Hugh W Brock4, Neel B Randsholt1,2*l. This is an Open Access article distributed under the terms of the Creativeommons.org/licenses/by/4.0), which permits unrestricted use, distribution, andiginal work is properly credited. The Creative Commons Public Domaing/publicdomain/zero/1.0/) applies to the data made available in this article,Dupont et al. Epigenetics & Chromatin (2015) 8:18 Page 2 of 17Cyclins and CDKs can also directly alter gene expres-sion. For example, mammalian Cyclin D1 binds severalcore transcription factors in regions close to transcrip-tion start sites [8], thus participating in regulation ofadjacent genes. Furthermore, a cyclin/CDK complex de-termines the activity of the multi-subunit complex medi-ator that plays a fundamental role in eukaryotic generegulation. Indeed, the mediator kinase subunit, com-posed of Cyclin C/CDK8, MED12, and MED13 in Dros-ophila, promotes either transcriptional activation orrepression depending on the context [9-12]. Interest-ingly, MED12 and MED13, encoded by kohtalo (kto) andskuld (skd), were initially identified as suppressors ofhomeotic phenotypes induced by mutation of Polycomb(Pc), an epigenetic repressor of transcription [13]. How-ever, kto and skd are also involved in epigenetic repres-sion of the Hox gene Ubx during development [14].These data link the mediator kinase subunit to the epi-genetic mechanisms that ensure faithful transmission ofchromatin states from mother to daughter cells.During development, epigenetic maintenance of gene ex-pression patterns is under control of evolutionary con-served proteins encoded by the Polycomb-group (PcG),trithorax-group (trxG), and Enhancer of Trithorax andPolycomb (ETP) genes, whose roles have been particularlywell studied for homeotic (Hox) gene regulation in Dros-ophila [15-18]. PcG genes are involved in long-term generepression, whereas trxG genes participate in maintenanceof gene activation and counteract PcG action. ETP geneshave been mainly characterized in Drosophila. They inter-act with both trxG and PcG genes, and behave alternativelyas repressors or activators of target genes [17-19].The Trithorax-group (TrxG) and PcG proteins formlarge multimeric complexes that bind chromatin and ap-pose or recognize histone post-translational modifica-tions. TrxG complexes are mostly involved in geneactivation (for a review see [20]). They comprise histonemodifying complexes, such as trithorax activating com-plex 1 (TAC1) containing the SET-domain histonemethyltransferase TRX that trimethylates lysine 4 of his-tone H3 (H3K4me3) [21], or the ASH1 complex con-taining the histone methyltransferase absent, small, orhomeotic discs 1 (ASH1), which methylates histone H3and histone H4 [22-24]. TrxG complexes also includeATP-dependent chromatin remodeling complexes, suchas BAP that contains the ATPase Brahma (BRM) [25].On the contrary, PcG complexes are involved in epigen-etic maintenance of gene silencing (for a review see[15]). The conserved polycomb repressive complex 2(PRC2) contains several PcG proteins, including extrasex combs (ESC) and enhancer of zeste [E(Z)], a SET-domain histone methyltransferase that trimethylates his-tone 3 on lysine 27 (H3K27me3). A second repressivecomplex, PRC1, comprises the PcG proteins polycomb(PC) and polyhomeotic (PH). PRC1 silences genes throughubiquitination of histone H2A on lysine 119 and chroma-tin compaction [26,27]. Other PcG complexes include therecently identified polycomb repressive deubiquitinase(PR-DUB) complex consisting of the deubiquitinase Ca-lypso and the ETP Additional sex comb (ASX) in Drosoph-ila [28]. Mammalian PR-DUB contains BAP1, homolog toCalypso, ASXL1 and ASXL2, two ASX homologs, as wellas several additional partners [29,30]. PR-DUB catalyzesdeubiquitination of H2AK119, binds PcG targets and is es-sential for promoter silencing [28-31]. However, Drosoph-ila ASX, as well as its murine homologs, are required forboth activation and repression of Hox genes, which makesthem genuine ETPs [32-34].ETPs also comprise the evolutionary conserved GAGAfactor that interacts with both BAP and PRC complexesin the regulation of Hox genes [35] and the proteinsCorto and dorsal switch protein 1 (DSP1) in Drosophila[36,37]. Corto interacts physically with GAGA and DSP1which are both involved in the recruitment of PcG com-plexes to Polycomb Response Elements (PRE) [38]. Asseveral ETPs are differently recruited to PREs dependingon tissues or developmental stages, it has been proposedthat different combinations of ETPs could favor the re-cruitment of either PcG or TrxG complexes, thus partici-pating in maintenance of silenced or active states [37].Drosophila Cyclin G was first isolated during a two-hybrid screen using the ETP Corto as bait [39]. Cyclin Ghas two mammalian homologs, Cyclin G1 and G2,whose functions remain elusive. CCNG1 that encodesCyclin G1 is a direct target of the tumor-suppressor p53[40]. Its induction following γ-irradiation leads to cellcycle arrest at the G2/M transition, thus allowing DNAdamage repair [41]. On the contrary, overexpression ofCCNG1 activates the proliferation of colon carcinomacells [42]. Overexpression of CCNG2 that encodes Cyc-lin G2 induces cell cycle arrest at the G1/S transition[43]. Like Cyclin G2 [44], Drosophila Cyclin G bears aPEST sequence at its C-terminal extremity. Furthermore,Cyclin G negatively regulates both cell growth and cellcycle progression, preventing G1/S transition and slow-ing down the S phase [45,46]. Collectively, these charac-teristics suggest that Drosophila Cyclin G behaves morelike Cyclin G2. Cyclin G co-localizes at many sites withthe ETP Corto as well as with the PcG protein PH onlarval polytene chromosomes, suggesting that it is in-volved in the control of gene expression [39]. Althougha direct role of mammalian G-type cyclins in gene expres-sion has not been reported, overexpression of CCNG1 inhuman cells induces chromatin relaxation [47]. Geneticinteractions between CycG (encoding Drosophila CyclinG) and corto showed that both genes are involved in regu-lating expression of the homeotic gene Abdominal-B(Abd-B) in the early pupal epithelium, corto acting as aDupont et al. Epigenetics & Chromatin (2015) 8:18 Page 3 of 17repressor and CycG as an activator of Abd-B [39]. Further-more, Cyclin G and Corto bind the iab-7 cis-regulatoryelement as well as the promoter of Abd-B in embryos[39]. Altogether, these data strongly suggest a role forCyclin G in regulation of Hox gene expression duringdevelopment.In this work, we address the involvement of Cyclin Gin epigenetic maintenance of Hox gene expression inDrosophila. We first demonstrate that Cyclin G also in-teracts with the ETP ASX, and extensively co-localizeswith ASX on polytene chromosomes. We next show thatCycG genetically interacts with Asx in maintenance ofHox genes. Interestingly, CycG also interacts with severalPcG and trxG genes and behaves genetically as a PcG,since loss of CycG enhances PcG-mediated homeoticphenotypes and suppresses trxG-mediated ones, whereasoverexpression of CycG has the opposite effect. Thesegenetic interactions were correlated with modificationsof ectopic Hox protein domains in imaginal discs, sug-gesting a role for Cyclin G in PcG-mediated Hox generepression during development. Surprisingly, we foundthat Cyclin G largely co-localizes with C-terminal do-main of RNA polymerase II (RNA Pol II CTD) phos-phorylated on serine 2, suggesting a role in productivetranscription. We propose that Cyclin G acts as an acti-vator or a repressor of transcription depending on thedevelopmental stage, the tissue, or the target gene.ResultsDrosophila Cyclin G interacts with the enhancer ofTrithorax and Polycomb ASXCyclin G physically interacts with the ETP Corto in vivo,and the Cyclin G N-terminal domain (amino-acids 1 to130) is necessary and sufficient for this interaction [39].Interestingly, a yeast two-hybrid screen to isolate inter-actors of the ETP ASX also identified Cyclin G as a po-tential partner. In this screen, the ASX-C terminaldomain (ASX-C, amino-acids 1139 to 1668), containinga PHD domain (amino-acids 1634 to 1665), was used asbait against a two-hybrid library of 0 to 12 h embryoniccomplementary DNAs (cDNAs) [48]. A deleted form ofCyclin G containing only the 275 N-terminal residuesinteracted strongly with ASX-C in two-hybrid assays(data not shown). These results showed that Cyclin Gresidues 1 to 275, N-terminal to the cyclin box, weresufficient for the interaction with ASX.To further characterize the interaction between CyclinG and ASX, we co-transfected S2 cells with pAct::Myc-AsxC and pAct::FLAG-CycG and performed immunopre-cipitation using anti-Myc and anti-FLAG antibodies.FLAG-Cyclin G co-immunoprecipitated with Myc-ASX-C (Figure 1B), confirming the interaction between CyclinG and ASX. To determine which Cyclin G region medi-ates this interaction, we next constructed different vectorsexpressing FLAG-tagged, truncated forms of Cyclin G(Figure 1A). Myc-ASX-C co-immunoprecipitated with the130 N-terminal residues of Cyclin G (Figure 1C). Hencethis domain, which also mediates the interaction with Corto,was sufficient to bind ASX. However, Cyclin G deleted ofthis N-terminal domain also co-immunoprecipitated withASX-C (Figure 1D), indicating that other parts of CyclinG interact with ASX-C.To test the interaction between Cyclin G and ASXin more native conditions, we performed immunopre-cipitation using protein extracts of third instar larvaeoverexpressing a Myc-tagged form of Cyclin G (da >Myc-CycGΔP) using anti-Myc and anti-ASX antibodies.Myc-Cyclin G co-immunoprecipitated with ASX (Figure 1E),demonstrating that Cyclin G interacts with endogenousASX in vivo.Taken together, these results confirm that Cyclin G in-teracts with ASX, and show that the Corto-interactingdomain, that is, amino-acids 1 to 130 of Cyclin G, is suf-ficient for this interaction.Cyclin G mainly localizes on active chromatinSince Cyclin G physically interacts with two ETPs (Cortoand ASX), and since we previously established that Cyc-lin G co-localizes with Corto at many loci on polytenechromosomes of larval salivary glands [39], we testedwhether Cyclin G also shares binding sites with ASX onchromatin. We co-immunostained salivary gland poly-tene chromosomes of wild-type larvae with anti-CyclinG and anti-ASX antibodies. We detected an importantoverlap of Cyclin G and ASX binding sites (Figure 2A),indicating that interaction of Cyclin G with ASX couldtake place on chromatin.Furthermore, Cyclin G preferentially localized at DAPIinterbands (Figure 2B), suggesting that this proteinmainly binds open chromatin. This observation led us totest whether Cyclin G binding was preferentially coupledwith an active or a repressed chromatin state, by exam-ining its co-localization with specific marks on polytenechromosomes. Comparison of Cyclin G binding siteswith the distribution of the repressive histone markH3K27me3 showed that very few Cyclin G-bound lociwere positive for H3K27me3 (Figure 2C). Conversely,Cyclin G co-localized largely with RNA Pol II phosphor-ylated on serine 2 (Figure 2D). These results suggest thatCyclin G binds mainly open chromatin and might be as-sociated with genes undergoing transcription.CycG genetically interacts with AsxThe physical interaction between Cyclin G and ASX aswell as their extensive co-localization on chromatin ledus to examine genetic interactions between CycG andAsx, using two loss of function alleles of Asx, AsxXF23,and Asx3 [49,50]. To address the effects of CycGDupont et al. Epigenetics & Chromatin (2015) 8:18 Page 4 of 17misregulation on Asx phenotypes, we combined theseAsx mutations with CycG loss of function [ubiquitous in-activation by RNA interference using the UAS::dsCycG2transgene driven by da::Gal4 (da >CycG RNAi)] or gainof function [ubiquitous overexpression of CycG using theUAS::CycGΔP transgene (da >CycGΔP encoding Cyclin GFigure 1 Cyclin G physically interacts with ASX in vivo. (A) Schematic representand CycG 130 to 566 corresponding to amino-acids 1 to 130 and 130 to 566, reG interacts with ASX-C (amino-acids 1139 to 1668) in S2 cells; black arrowheads(D) co-immunoprecipitate with Myc-ASX-C. Note that FLAG-CycG 130 to 566 cowith FLAG-CycG 1 to 130. Immunoprecipitations were performed with anti-Mycwere revealed by Western blot, using anti-Myc (top panel) or anti-FLAG antibodsupernatant after immunoprecipitation; IP: protein G-agarose beads. Five percenloaded onto the gels. (E) Myc-Cyclin G co-immunoprecipitates with endogenouwere revealed by Western blot, using anti-Myc antibody.deleted of amino-acids 542–566)]. We focused on the ef-fect of CycG misregulation on PcG-like homeotic transfor-mations induced by Asx mutations. Importantly, thesehomeotic phenotypes were never detected in CycG loss offunction or gain of function flies. All genetic interactiondata are shown in Additional file 1.ation of Cyclin G full-length (CycG) and truncated forms (CycG 1 to 130spectively). Grey box: Cyclin domain; red box: PEST sequence. (B-D) Cyclinindicate co-immunoprecipitation. FLAG-CycG (B) and FLAG-CycG 130–566-migrates with IgG heavy chains. (C) Myc-ASX-C co-immunoprecipitates, anti-FLAG, or anti-HA (Mock) antibodies. Immunoprecipitated proteinsies (bottom panel). In (B) and (D), asterisks indicate IgG heavy chains. S:t of the input or supernatant and 50% of the immunoprecipitate weres ASX in da>Myc-CycGΔP third instar larvae. Immunoprecipitated proteinsDupont et al. Epigenetics & Chromatin (2015) 8:18 Page 5 of 17PcG genes derived their name from the most con-spicuous common phenotype of adult males carryingPcG mutations: ectopic sex combs on posterior legs. In-deed, in Drosophila melanogaster, this male-specificstructure, composed of specialized bristles called teeth,is specific to the first tarsal segment of prothoracic legs(L1). The occurrence of sex comb teeth on mesothoracic(L2) and metathoracic legs (L3) thus indicates a partialtransformation of L2 and L3 into L1. Males heterozy-gous for AsxXF23 presented ectopic sex combs on L2with a lower penetrance than Asx3 heterozygous males.CycG loss of function significantly enhanced penetranceof L2 sex combs induced by AsxXF23 (Figure 3A) but didnot modify that of Asx3 (Figure 3B). Furthermore, CycGFigure 2 Cyclin G co-localizes with ASX and binds mainly active chromatinlarvae. DNA was stained with DAPI (blue). (A) Cyclin G (green) and ASX (redby a white rectangle. (B) Cyclin G (green) binds DAPI interbands. Right: clo(green) and H3K27me3 (red) are almost completely exclusive. Rare co-localat many sites with RNA Pol II phosphorylated on serine 2 (red). Bottom: clogain of function significantly suppressed L2 sex combsinduced by Asx3 (Figure 3B). To summarize, CycG gainof function opposed an Asx-induced PcG phenotype onL2, whereas CycG loss of function enhanced this pheno-type. Therefore, CycG behaved as an enhancer of Asx re-garding homeotic transformation of L2 into L1. Thissuggests that CycG participates with Asx in the mainten-ance of mesothoracic leg identity.Maintenance of abdominal segment identity also relieson PcG, trxG, and ETP genes. In wild-type Drosophilamales, tergites of the fifth (A5) and sixth (A6) abdominalsegments present a uniform dark pigmentation, whereastergites of more anterior abdominal segments show onlya posterior stripe of dark pigmentation. Thirty percent. Immunostaining of polytene chromosomes from w1118 third instar) co-localize at many sites. Bottom: close-up view of the region framedse-up view of the region framed by a white rectangle. (C) Cyclin Gizations are shown with white arrows. (D) Cyclin G (green) co-localizesse-up view of the region framed by a white rectangle.Dupont et al. Epigenetics & Chromatin (2015) 8:18 Page 6 of 17of Asx3/+ males presented darkly pigmented patches onthe anterior part of the fourth abdominal tergite (A4),indicating a partial transformation of A4 into a moreposterior segment, a phenotype observed in some PcGmutants (Figure 3E). As shown on Figure 3C, thisphenotype was not affected by CycG loss of function,Figure 3 CycG genetically interacts with Asx. (A, B) Effect of CycG misregul(A) AsxXF23 and (B) Asx3. (C, D) Penetrance of A4 to A5 (C) and A5 to A4 (Dand Asx3/+;da > CycGΔP/+ males. (E) Posterior abdomens of adult males showAsx3/+ male (black arrowhead), and A5 to A4 transformation in Asx3/+;da> CycG(n≥ 30). L2: mesothoracic leg. A4, A5, A6: abdominal segments 4, 5, and 6.but was completely suppressed by CycG gain of function.Interestingly, 70% of Asx3/+; da >CycGΔP males pre-sented small patches lacking dark pigmentation on ter-gite A5 (Figure 3D,E). This corresponds to a partialtransformation of A5 into segment A4, which is a clas-sical trxG homeotic transformation. Hence, in theation on the penetrance of ectopic sex combs on L2 induced by) abdominal segment transformations in Asx3/+, Asx3/+;da > CycG RNAiing wild-type male pigmentation in A4 to A6, A4 to A5 transformation inΔP/+ male (white arrowhead). Fisher’s exact test, *P< 0.01 **P< 0.0001abdomen, CycG gain of function not only suppressed thetypical PcG phenotype induced by Asx3 but also shiftedit toward a trxG phenotype.Together, these results reveal that CycG and Asx areengaged in complex genetic interactions and suggest arole for CycG similar to that of a PcG gene in mainten-ance of segmental identities during development.CycG loss of function enhances PcG-induced homeotictransformationsThe genetic interactions reported above, the physicalinteraction between Cyclin G and the ETPs ASX andCorto, and the co-localization on chromatin of Cyclin Gwith ASX as well as with Corto and the PcG proteinPH, led us to further examine genetic interactions be-tween CycG and PcG, trxG, or ETP genes. We associatedCycG misregulation with alleles of these genes (listed inAdditional file 2) reported to induce a visible dose-sensitive phenotype, alone or in combination with otherPcG, trxG, or ETP mutations. All genetic interactiondata are shown in Additional file 1.Contrary to Asx mutations, mutant alleles of the ETPscorto and Dsp1 induced no PcG-like leg transformation,whether alone or in combination with CycG misregula-tion. We next analyzed interactions with mxc (multi sexcombs) and crm (cramped), genetically classified as PcGgenes [51,52], although their products have not beenfound in PcG complexes so far. mxcG46/Y males presentsex comb teeth on posterior legs [52], indicating a partialtransformation into L1 (Figure 4A). Indeed, 66% of L2and 11% of L3 (n = 56) carried at least one sex combtooth (1.4 ± 1.3 teeth on L2 and 0.2 ± 0.5 on L3). Pharatemxb ipepeDupont et al. Epigenetics & Chromatin (2015) 8:18 Page 7 of 17Figure 4 CycG misregulation alters ectopic sex comb phenotypes of anleg (L3) transformations into prothoracic leg (L1). The wild-type sex comcomb teeth on distal L1, on L2 and L3 (black arrowheads). Only phenotyCycG loss of function. (B) Penetrance of the ectopic sex combs phenotysex combs on L2 and L3, whereas CycG gain of function suppresses ectopi(n ≥ 50).c mutant. (A) Expressivity of mesothoracic leg (L2) and metathoracics marked by a white arrowhead. mxcG46/Y males present ectopic sexs affecting L2 and L3 were rated. These phenotypes are enhanced byon male legs. CycG loss of function enhances mxcG46 induced ectopicc sex combs on both legs. Fisher’s exact test, *P < 0.05 **P < 0.0001Dupont et al. Epigenetics & Chromatin (2015) 8:18 Page 8 of 17mxcG46/Y males presented similar phenotypes (Additionalfile 1). As both gain and loss of function of CycG inducedmxcG46 male lethality just prior to adult emergence, ec-topic sex combs were scored in pharates. CycG loss offunction enhanced the expressivity of this phenotype (3.7 ±2.3 teeth on L2 and 1.3 ± 2.0 on L3, n ≥ 56; t-test, P <0.0001) (Figure 4A). Penetrance of the phenotype was alsoenhanced by CycG loss of function (Figure 4B). Con-versely, CycG gain of function completely suppressed L2and L3 sex combs of mxcG46/Y males (Figure 4B). CycGmisregulation mxcG46 males presented no other pheno-types than modified ectopic extra sex combs. These ani-mals died prior to full pigmentation of the abdomen,preventing evaluation of phenotypes observed in adultmxcG46 males [52]. Interactions between CycG and crmgave similar results. Eighty percent of crm7/Y flies pre-sented sex combs on L2 and 35% on L3. Males combiningcrm7/Y and CycG loss of function died before the pupalstage, preventing analysis. In the few pharate escapers com-bining crm7/Y and CycG overexpression, leg transforma-tions were significantly suppressed (Additional file 1).Next, we tested several genes encoding members ofPRC2. Neither E(z)63/+ nor esc21/+ males presented ec-topic sex combs, and this was not significantly modifiedby CycG misregulation. On the other hand, mutants forpolyhomeotic (ph) and Polycomb (Pc), encoding PRC1 sub-units, both present ectopic sex combs. In ph-p410/Y males,100% of L2 and L3 carried combs, and these phenotypeswere significantly suppressed by CycG gain of function.Furthermore, the 20% of L2 and 5% of L3 sex combs ofph220/Y males were totally suppressed by CycG gain offunction, and significantly enhanced by CycG loss of func-tion (see Additional file 1). Finally, Pc3/+ males exhibit sexcombs on 100% of L2 and 98% of L3 which were drastic-ally suppressed by CycG gain of function (Figure 5A).We also examined other PcG-induced homeotic phe-notypes, focusing first on wing to haltere transforma-tions. Ninety-three percent of female and 96% of malePc3/+ wings exhibit a partial wing into haltere trans-formation of the area close to the posterior wing, indu-cing deformation of the wing (Figure 5B,C). CycG loss offunction had no effect on this Pc3 phenotype, whereasCycG gain of function reduced both its strength and itspenetrance (Figure 5B,C).Similar to Asx3/+ males, several PcG and ETP mutantspresent a partial transformation of A4 into A5. A fewesc21/+ or Dsp11/Y males presented this phenotype,which was not significantly affected by CycG misregula-tion (Additional file 1). However, 15% of Pc3/+ malespresented such A4 transformations (Figure 5D,E), whosestrength and penetrance were significantly enhanced byCycG loss of function and completely suppressed by CycGgain of function (Figure 5D,E). Ninety-seven percent ofph-p410 males exhibited A4 into A5 transformations thatwere significantly suppressed by CycG gain of function(Additional file 1). The effect of CycG loss of function wasnot analyzed in ph-p410 males because pharates died priorto abdominal pigmentation. In ph220/Y males, posteriorA4 transformations were not modified by CycG loss offunction but significantly suppressed by CycG gain offunction (Additional file 1).All these genetic data point toward a role of CycGsimilar to that of a PcG gene as PcG loss of functionphenotypes were enhanced by CycG loss of function butsuppressed by CycG gain of function.CycG gain of function enhances trxG-induced homeotictransformationsWe next examined the effects of CycG misregulation onmutant phenotypes induced by loss of function of trxGgenes. We studied mutations in brm, ash1, and trx, focus-ing on posterior abdominal pigmentation patterns, in par-ticular A5 depigmentation indicating transformation ofsegment A5 into A4 [53]. Males brm2/+ did not present A5toward A4 transformations, whereas 97% of males combin-ing brm2/+ and CycG gain of function (n = 30) presentedthis phenotype (Figure 6A). ash1B1/+ males exhibited A5 toA4 transformations with a low penetrance (3%, n = 31).This phenotype was increased by CycG gain of function, al-though not to a significant level (13%, n = 24). Finally, 10%of trxE2/+ males presented partial transformation of A5 intoA4 (Figure 6B) (n = 31). This transformation was decreased,though not significantly, by CycG loss of function (3%; n =30), whereas both its strength and penetrance were drastic-ally enhanced by CycG gain of function (100%, n = 30;Fisher’s exact test, P < 0.0001) (Figure 6C,D).Together, these data show that CycG loss of functiontended to decrease segmental identity transformations oftrxG loss of function mutants, whereas CycG overex-pression increased them. Hence, CycG antagonizes trxGgenes in the posterior abdomen, behaving again as a PcGgene.CycG participates in PcG-dependent repression of Scr inleg imaginal discsEctopic sex combs of PcG mutants are due to loss of si-lencing of the homeotic gene Sex combs reduced (Scr) inthe second and third pairs of leg imaginal discs duringlarval development, leading to acquisition of a partialfirst-leg identity [54]. We therefore investigated the ef-fect of CycG misregulation on expression of the Scr genethat specifies identity of the first thoracic segment inDrosophila melanogaster [55,56]. We monitored the pat-tern of SCR by immunostaining of leg imaginal discs,the larval structures that differentiate into legs duringmetamorphosis. In wild-type third instar larvae, cellswith high levels of SCR form two semicircles in the L1imaginal disc territory that gives rise to the anterior tibiaDupont et al. Epigenetics & Chromatin (2015) 8:18 Page 9 of 17and first tarsal segment (Figure 7A), whereas L2 and L3imaginal discs present no SCR positive cells. Import-antly, neither CycG loss of function nor CycG gain offunction male larvae exhibited any detectable alterationin the spatial pattern of SCR (data not shown).We next looked at SCR distribution in third instar legdiscs from three different PcG mutant males (mxcG46,Pc3, ph-p410). We observed ectopic SCR in the anteriorcompartment of L2 and L3 imaginal discs in all threeFigure 5 CycG misregulation alters homeotic transformations of a Pc mutaCycG RNAi and Pc3/da > CycGΔP males. Fisher’s exact test, **P < 0.0001 (n ≥ 5da > CycG RNAi and Pc3/da > CycGΔP females and males. Fisher’s exact test,posterior deformation corresponding to a partial transformation into haltersuppressed haltere to wing transformation. (D) Penetrance of A4 to A5 abdPc3/da > CycGΔP males. Fisher’s exact test, *P < 0.05, **P < 0.0001 (n ≥ 29). (EA4 to A5 transformations (arrowheads). A4, A5, A6: abdominal segments 4,contexts (Figure 7A, Additional file 3). In male larvaecombining mxcG46 and CycG loss of function, size of thisectopic SCR domain in both L2 and L3 discs was signifi-cantly increased (Figure 7A; Additional file 4). Con-versely, in male larvae combining mxcG46 and CycG gainof function, the ectopic SCR domain was almost com-pletely suppressed (Additional file 4). Ectopic SCR do-mains were also significantly enlarged in L2 and L3 discsof male larvae combining either Pc3 or ph-p410 withnt. (A) Penetrance of L2 and L3 ectopic sex combs in Pc3/+, Pc3/+;da >8). (B) Penetrance of wing to haltere transformations in Pc3/+, Pc3/+;**P < 0.0001 (n≥ 31). (C) Representative Pc3/+ adult female wing withe tissue (arrowhead). Representative Pc3/da > CycGΔP female wing withominal segment transformations in Pc3/+, Pc3/+;da > CycG RNAi and) Abdomens of Pc3 and Pc3/+;da > CycG RNAi males with representative5, and 6.trxindctivk aDupont et al. Epigenetics & Chromatin (2015) 8:18 Page 10 of 17CycG loss of function and decreased in those combiningPc3 or ph-p410 with CycG gain of function (AdditionalFigure 6 CycG gain of function enhances abdominal transformations ofmale cuticles. Effect of CycG misregulation on brm2-induced (A) or trxE2-Representative trxE2/+ (B) and trxE2/da > CycGΔP (C) males present, respepartial transformation of abdominal segment A5 into A4 (white and blacabdominal segments 4, 5, and 6.files 3 and 4).These results indicate that modulation of PcG-inducedectopic sex comb phenotypes by CycG reflects modula-tion of the SCR pattern in imaginal discs. Interestingly,CycG loss of function enlarged the SCR domain even inph410 and Pc3 individuals who presented a quasi com-plete penetrance of the ectopic sex comb phenotype inadults. Together, these data strongly suggest that CycGcooperates with PcG genes in epigenetic Scr repression.CycG represses Pc3-induced ectopic Ultrabithorax in wingimaginal discsCycG gain of function suppressed Pc3-induced wing tohaltere transformations. This homeotic transformation iscaused by ectopic expression of the homeotic geneUltrabithorax (Ubx) that specifies haltere identity [57].We therefore analyzed the effect of CycG misregulationon the UBX profile in wing imaginal discs. In wild-typethird instar larvae, we observed a high level of UBX inhaltere and third leg imaginal discs, whereas no UBXwas detected in wing discs (Figure 7B). This pattern wasnot altered by CycG misregulation (data not shown). Weobserved patches of ectopic UBX in the posterior com-partment of Pc3/+ wing imaginal discs (Figure 7B), thatis, in the region that will be partly transformed into hal-tere in adults. CycG loss of function had no significanteffect on this ectopic UBX profile, whereas CycG gain offunction almost completely suppressed it (Figure 7B,Additional file 4).G mutants. (A, D) Penetrance of A5 to A4 transformations in adultuced (D) A5 to A4 transformations. (B, C) Male abdominal cuticles.ely, small and large light-pigmented cuticle patches on A5, denoting arrowheads). Fisher’s exact test, **P < 0.0001 (n ≥ 30). A4, A5, A6:These data show that the suppression of Pc3-inducedwing to haltere transformation by CycG gain of functioncorrelates with a suppression of ectopic UBX in wingimaginal discs, suggesting a role for CycG in PcG-dependent repression of Ubx in this tissue.DiscussionDrosophila Cyclin G participates in control of the cellcycle and in transcriptional regulation [39,45,46,58].Physical interaction between Cyclin G and the ETPCorto suggests that Cyclin G might be related to PcGand trxG genes involved in maintenance of gene silen-cing and gene activation, respectively. The present workstrengthens this hypothesis, since we demonstrate thatCyclin G directly binds the ETP ASX as well, co-localizes extensively with ASX on chromatin, and genet-ically interacts with Asx. Furthermore, Cyclin G mostlybinds open chromatin in which genes are undergoingtranscription. We also addressed the connection be-tween Cyclin G and the PcG/TrxG system by investigat-ing genetic interactions with a broad range of PcG andtrxG alleles. We describe strong interactions betweenCycG misregulation and several PcG and trxG genes, to-gether with modification of ectopic Hox protein expres-sion induced by PcG mutations. Our results link CycGto maintenance of Hox gene expression patterns duringdevelopment, and sustain a role for Cyclin G inDupont et al. Epigenetics & Chromatin (2015) 8:18 Page 11 of 17epigenetic maintenance of transcription, a mechanismthat is essential for transmission of gene expression pat-terns to daughter cells.Cyclin G, a modulator of Enhancers of Polycomb andTrithorax activity?Drosophila Cyclin G was shown previously to physicallyinteract with the ETP Corto [39]. Here, we demonstratethat this cyclin also binds the ETP ASX. Unlike Corto,which to date has only been detected in arthropods,ASX proteins are conserved in mammals. These proteinsexhibit two remarkable domains, the ASX homologydomain (ASXH), an amino terminal region containingtwo consensus sequences for nuclear receptor bindingneeded for ASX binding to the repressive PR-DUBFigure 7 CycG misregulation modulates Hox protein profiles of PcG mutanfrom wild-type (wt), mxcG46/Y, mxcG46/Y;da > CycG RNAi or mxcG46/Y;da > Cywing (W), and prothoracic leg (L3) imaginal discs from a wild-type female lRNAi or Pc3/da > CycGΔP females. Note the reduced ectopic staining in thecomplex [28], and a PHD domain located at the carboxy-terminal end that interacts with DNA. Mouse and humanASXL1 also contain a number of cyclin recognition sitesand motifs for phosphorylation by CDKs scattered alongthe protein, suggesting that they are phosphorylated by aCyclin/CDK dimer [59]. Indeed, the presence of a cyclinbinding motif close to a CDK target sequence is essentialfor optimal phosphorylation of cyclin/CDK targets [60].We show here that CycG binds the C-terminal region ofASX (amino-acids 1139 to 1668), a region which containsa cyclin interaction domain (amino-acids 1210 to 1214) aswell as a substrate motif for phosphorylation by CDKs(amino-acids 1261 to 1267) [61]. Hence, a Cyclin G/CDKcomplex might phosphorylate ASX and modulate its ac-tivity. Nuclear Corto is highly phosphorylated [62] andts. (A) Anti-SCR immunostainings of third instar larval leg imaginal discscGΔP males. (B) Anti-UBX immunostaining of third instar haltere (H),arva, and anti-UBX staining of wing discs from Pc3/+, Pc3/+;da > CycGPc3/da > CycGΔP female wing disc (arrowhead).Dupont et al. Epigenetics & Chromatin (2015) 8:18 Page 12 of 17although Corto lacks a canonical cyclin recognition motif,it contains a substrate motif for phosphorylation by CDKsin the Cyclin G interacting region [39]. Interestingly, thisCDK target sequence is located in the chromodomain, in-volved in chromatin binding [63], suggesting that associ-ation of Corto with chromatin might be regulated byphosphorylation by a Cyclin G/CDK complex.Phosphorylations are of paramount importance toregulate PcG and TrxG protein activity through the cellcycle. For example, in mammals, binding of the PcGprotein BMI1 to chromatin is cell-cycle regulated andcorrelates with its phosphorylation status [64]. Further-more, phosphorylation of PcG proteins EZH2 andSCML2 by cyclin/CDK complexes is regulated throughthe cell cycle [65,66]. These findings highlight a directcrosstalk between the Polycomb system of cellular mem-ory and the cell-cycle machinery in mammals. Interest-ingly, fly Cyclin G controls the G1/S phase transition ofthe cell cycle and interacts with several CDKs (that is,CDK1, 2, 4, 5) [45,46,67,68]. Further investigations willbe needed to determine whether a Cyclin G/CDK dimerphosphorylates the ETPs ASX and Corto and modulatestheir activity during the cell cycle.Cyclin G interacts with ETP, PcG, and trxG genes in theregulation of homeotic genesASX is involved in both activation and repression ofhomeotic genes [32], and this role is conserved by itsmammalian homologs ASXL1 and ASXL2 [33,34]. Ac-cordingly, some Asx mutants present both PcG-like andtrxG-like transformations. For example, male AsxP1 fliesbear partial transformation of abdominal segment A4into A5, revealed by patches of dark pigmentation onA4, as well as partial transformation of A5 into A4, re-vealed by patches of unpigmented cuticle into A5 [69].In a mouse Asxl2 mutant, vertebras present both poster-ior and anterior transformations corresponding to PcGand trxG phenotypes, respectively [33]. We show herethat deregulation of CycG impacts on homeotic pheno-types of an Asx mutant that presents only PcG pheno-types, that is, ectopic sex combs and transformation ofA4 into A5. Importantly, neither inactivation nor over-expression of CycG induces PcG or trxG phenotypes perse. Nevertheless, CycG inactivation enhances ectopic sexcombs induced by Asx, while CycG overexpression sup-presses both PcG-like phenotypes. Combining the Asxmutation with CycG overexpression even leads to atrxG-like transformation of A5 into A4. In agreementwith these results, homeotic phenotypes of mutants forPcG genes crm, mxc, Pc, and ph are all enhanced byCycG inactivation and suppressed by CycG overexpres-sion. On the other hand, transformation of abdominalsegment A5 into A4 observed in trxG mutants was en-hanced (ash1, trx) and even induced (brm) by CycGoverexpression. Hence, in both posterior legs and ab-dominal segments A4 and A5, CycG behaves as an en-hancer of PcG genes.CycG overexpression, as well as CycG inactivation, in-duce developmental delay (unpublished data) and fliesoverexpressing CycG suffer from the Minute syndrome[45,46]. In Drosophila melanogaster, the Minute syn-drome has been described as a dominant, haploinsuffi-cient phenotype that includes delayed development,short and thin bristles together with poor fertility andlongevity [70]. More than 50 Minute loci were genetic-ally identified; 15 of them were characterized molecu-larly and contain genes encoding ribosomal proteins[71]. Genetic screens designed to isolate new PcG andtrxG genes in flies have frequently identified Minute mu-tants as suppressors of PcG mutations [19,72]. This ef-fect has been considered to be unspecific, since otherfactors resulting in developmental delays, that is, lowtemperature, also suppress the ectopic sex comb pheno-type of PcG mutants [73]. Although homeotic phenotypesof PcG mutants are suppressed by CycG overexpression,the fact that CycG inactivation associates developmentaldelay and enhancement of PcG transformations indicatesthat CycG acts as a specific modifier of these phenotypes.Hence, our data characterize CycG as a bona fide enhan-cer of Polycomb-group genes, involved in Hox generegulation.Cyclin G, a transcriptional activator or repressor?In agreement with our genetic analyses, we observedthat CycG inactivation enlarges ectopic SCR domains inPcG mutant leg imaginal discs, whereas CycG overex-pression eliminates ectopic SCR and UBX domains inleg and wing imaginal discs. This suggests that Cyclin Gfacilitates the maintenance of Scr and Ubx silencing byPcG complexes in these imaginal tissues. Nevertheless,whether this effect is direct or not remains to be deter-mined. Paradoxically, we previously observed that CyclinG is required for maintenance of Abd-B expression inthe epithelium of abdominal segments A5 and A6 inyoung female pupae [58]. This effect might be direct asCyclin G was shown to bind the Abd-B promoter andthe iab-7 polycomb response element in embryos [39].Altogether, these data show that Cyclin G is involved inepigenetic regulation of Hox gene expression, acting as arepressor or an activator depending on the tissue, thedevelopmental stage, and the target gene. CycG is thussimilar to many other genes encoding maintenance pro-teins that affect transcription differently depending onthe context. For example, Drosophila E(z) is classified asa PcG repressor, but behaves genetically as an ETP[17,19]. Furthermore, brahma (brm) as well as mostgenes encoding members of the SWI/SNF chromatin re-modeling complex Brahma-associated protein (BAP) areDupont et al. Epigenetics & Chromatin (2015) 8:18 Page 13 of 17classified as trxG activators, but snr1, that encodes a con-served subunit of BAP, participates in cell-type specifictranscriptional repression in the developing Drosophilawing [74].To tackle the molecular mechanisms by which CyclinG controls transcription, we analyzed its binding topolytene chromosomes in larval salivary glands. CyclinG binds DAPI interbands indicative of open chromatin,co-localizes largely with RNA Pol II CTD phosphory-lated on serine 2, and shows very few overlaps withH3K27me3. All these data point to the presence of Cyc-lin G on actively transcribed genes. Deregulation ofCycG does not induce homeotic transformations per se,but modulates those due to PcG or trxG mutations. Cyc-lin G might thus preferentially affect ‘destabilized’ genes.Since Cyclin G co-localizes with DAPI interbands andwith RNA Pol II phosphorylated on serine 2, and be-haves as a PcG enhancer, its role could possibly be tomoderate the expression of active genes.We also show that Cyclin G extensively overlaps withASX. As Cyclin G and ASX co-immunoprecipitate, Cyc-lin G and ASX can be assumed to interact on chromatin.ASX belongs to the repressive PR-DUB complex, whichcontains the histone deubiquitinase BAP1, also calledCalypso in Drosophila [28]. PRC1 ubiquitinates H2A onlysine 119, and this ubiquitin residue can be removed byPR-DUB. Surprisingly, disruption of PR-DUB enzymaticfunction led to impaired Hox gene repression, as does ashortage of PRC1. Hence, PcG silencing has been pro-posed to depend on a dynamic equilibrium betweenH2A ubiquitination by PRC1 and deubiquitination byPR-DUB [28]. An interesting possibility could then bethat Cyclin G influences this equilibrium by modulatingASX activity.In embryonic stem cells, some key developmental genes,called bivalent, are simultaneously stamped by both re-pressive (H3K27me3) and activating (H3K4me3) histonemarks [75]. Although these genes are associated withRNA Pol II CTD phosphorylated on serine 5, they aretranscribed at a low level [76]. Ubiquitination of H2A byPRC1 controls this process, and has therefore been sug-gested to control transcriptional elongation [76,77]. Thesedata raise the exciting possibility that Cyclin G fine-tunestranscription by controlling H2A ubiquitination via inter-action with the ASX subunit of PR-DUB.ConclusionsOur findings highlight a crosstalk between the Polycombsystem and Drosophila Cyclin G. The importance andcomplexity of the interaction between Cyclin G andASX warrant further investigation. It is tempting tospeculate that this interaction regulates transcriptionalelongation. Specific points to be explored in the near fu-ture include interaction between Cyclin G and ASX inthe context of PR-DUB, involvement of Cyclin G in thebalance between ubiquitinated and unubiquitinatedH2A, and regulation of PR-DUB activity by Cyclin Gthrough the cell cycle.MethodsDrosophila strains and genetic analysesDrosophila melanogaster stocks were raised on standardyeast-cornmeal medium at 25°C. Transgenic lines UAS::dsCycG2 (referred to as CycG RNAi) [39] and UAS::CycGΔP (line RCG76, allowing expression of Cyclin Gdeleted of the 25 C-terminal amino-acids, containing aputative PEST sequence [45,46]) were used to inactivateCycG by RNA interference (CycG loss of function) oroverexpress CycG (CycG gain of function), respectively.CycG misregulations were induced using the ubiquitousdriver daughterless (da::Gal4). The third chromosometransgenes da::Gal4 and UAS::CycGΔP were recombinedand gave rise to chromosome da::Gal4,UAS::CycGΔP,called da >CycGΔP. Chromosome da >CycGΔP was main-tained in males at 18°C to overcome female sterility andhigh lethality associated with CycG overexpression at 25°C[45]. For co-immunoprecipitations, a new transgenic linecontaining a UAS::Myc-CycGΔP construct was obtainedby PhiC31 integrase-mediated insertion of pUASP-Myc-CycGΔP-attB at 51C (stock BL-24482) [78].Alleles of PcG, trxG, or ETP genes used in this studyare listed in Additional file 2, and their characteristicsare described in [79]. Genetic interactions between thesegenes and CycG were assessed in trans-heterozygousflies obtained by crossing females heterozygous for a bal-anced PcG, trxG, or ETP mutation with males eitherda >CycG RNAi, da >CycGΔP, or da::Gal4 as a control.All crosses were performed at 25°C, and parents weretransferred to new vials every 3 days. Penetrance ofhomeotic phenotypes affecting legs, wings, or abdomenwas determined among the progeny. Phenotypes wereassessed by examining 30 flies for each genotype (when-ever possible) under a dissecting microscope. Wings andmale legs were mounted in Hoyer’s medium. Sex combswere counted under a microscope at × 100 magnifica-tion. Statistical significance of results was evaluatedusing t-test and Fisher’s exact test on GraphPad Quick-Calcs Web site: http://www.graphpad.com/quickcalcs/contingency1/ (accessed April 2014).Plasmid constructsThe 3′ sequence of Asx (bp 3415 to stop codon, Dmel_CG8787) was amplified from embryonic cDNAs usingprimers AsxC_F (5′-caccgccgccatgacgcgtcctgccaatgcatcacc-3′) and AsxC_R (5′-tcatcatctaatcacacaggcgacacacagc-3′). The full-length CycG cDNA was amplifiedusing primers CycGnF 5′-cacctctgtccctgtacgctactcc-3′and CycGnR 5′-ctaacattgttcgaaaattggaattatggg-3′. cDNAsDupont et al. Epigenetics & Chromatin (2015) 8:18 Page 14 of 17encoding truncated forms of Cyclin G (Cyclin G 1 to 130and Cyclin G 130 to 566) were amplified using primersCycGnF and CycG1-130R 5′-ctaggcagcctgggccgaagtcgagggctg-3′, and CycG130-566 F 5′-caccgccgctgctgccgcatcc-3′ and CycGnR, respectively. PCR products were clonedinto pENTR/D-TOPO® (Invitrogen), then transferred intoGateway® vector pAMW (Invitrogen, a gift from T. Mur-phy) to produce the Myc-ASX-C fusion protein undercontrol of the actin5C promoter, or pAFW (Gateway®,Invitrogen, Carlsbad, CA, USA) to produce FLAG-taggedCyclin G fusion proteins under control of the samepromoter.Cell transfection and protein-protein interactionsThe yeast two-hybrid screen performed to find interac-tors of the ASX-C terminal domain (ASX-C, residues1139 to 1668) has been described previously [48,80].Drosophila S2 cells were cultivated at 25°C in Schnei-der’s Drosophila medium supplemented with 10% fetalcalf serum and antibiotics. Then 5.106 cells were trans-fected into 25 cm2 flasks using Effecten® Transfection re-agent kit at a 1/10 DNA-Effecten® ratio according to themanufacturer’s instructions (Effecten®; QIAGEN, Venlo,Limburg, The Netherlands). Cells were collected 36 or72 h after transfection depending on the constructs, andtotal protein extracts were prepared as previously de-scribed [45]. Cross-linking was performed by treatmentof cells with 1% paraformaldehyde for 10 min on iceprior to protein extraction. For co-immunoprecipitation,500 μg of total cell extracts were incubated overnight at4°C with 3 μg of either goat anti-Myc antibody (ab9132,Abcam, Cambridge, UK), mouse monoclonal anti-FLAGantibody (F3165, Sigma, St. Louis, MO, USA), or goatanti-HA as mock antibody (sc-805, Santa Cruz Biotech-nology, Santa Cruz, CA, USA). Then, 30 μl of Bio-Adembeads Protein G (Ademtech, Westbury, New York,USA) were then incubated with the cell lysate for 3 h at4°C. The beads were washed three times in ELB buffer[45] and resuspended in 30 μl of the same buffer. Fur-thermore, 20 μl of input, 20 μl of supernatant, and halfof the beads were used for Western analysis. Immuno-precipitates were detected with rabbit polyclonal anti-Myc antibody (1:5,000; A00172, GenScript) and mousemonoclonal anti-FLAG antibody (1:2,000).For co-immunoprecipitation in larvae, protein extractswere prepared from da >Myc-CycGΔP third instar larvae,previously cleared of gut and fat body and treated with1% paraformaldehyde for 10 min at room temperature.Five micrograms of either goat anti-Myc antibody, sheepanti-ASX N-ter antibody (described in [81]), or goatanti-HA as mock antibody were incubated with 50 μl ofDynabeads® Protein G (Life Technologies, Carlsbad, CA,USA) for 3 h at 4°C. The bead-antibody complexes werewashed in ELB buffer and incubated with 1 mg of proteinextracts overnight at 4°C. The beads were washed threetimes in ELB buffer before Western analysis. Immunopre-cipitates were detected with goat anti-Myc antibody(1:5,000).Immunolocalization on polytene chromosomesSquashes of w1118 third instar larval salivary glands andimmunostainings were performed as described previ-ously [36] using guinea pig anti-Cyclin G (1:40), sheepanti-ASX N-ter (1:20) (described in [39] and [81], re-spectively), rabbit anti-H3K27me3 (1:40; pAb-069-050,Diagenode, Denville, NJ, USA), or rabbit anti-RNA poly-merase II CTD phosphorylated on serine 2 (1:200;ab5095, Abcam, Cambridge, UK) antibodies. Secondaryantibodies (Alexa Fluor® 488 goat anti-guinea pig, AlexaFluor® 594 goat anti-rabbit IgG and Alexa Fluor® 680donkey anti-sheep IgG, Molecular Probes, Eugene, OR,USA) were used at a 1:1,000 dilution.Immunostaining of imaginal discsFor each genotype, at least 15 third instar wandering larvaewere dissected and fixed in 3.7% paraformaldehyde for20 min at room temperature, then immunostained accord-ing to [82,83] using rat polyclonal anti-SCR antibody (1:100[82]) or mouse anti-UBX monoclonal antibody (1:20;FP3.38 [84]). The universal biotinylated antibody (VectorLaboratories, CA, USA) was used at a 1:200 dilution. Stain-ing was performed with VECTASTAIN Elite ABC system(Vector Laboratories, CA, USA) using DAB as substrate(D4418, Sigma, St. Louis, MO, USA). Note that for a givenantibody, discs of all genotypes were incubated for the samelength of time in DAB. Imaginal discs were mounted inPBS:glycerol (50:50). All pictures were acquired with aQICAM Fast 1394 digital camera, at × 100 magnification.Staining was quantified by calculating the percentage ofstained area in the discs using Image J. SCR positive areawas measured as a percentage of the total leg disc area, andUBX positive area was evaluated relative to the presumptivewing blade and hinge area of the wing disc. Statistical sig-nificance of results was evaluated using t-test.Additional filesAdditional file 1: Table S1. Effect of CycG misregulation on PcG-induced homeotic transformations. L1, L2, and L3: prothoracic,mesothoracic and metathoracic legs. A4 and A5: abdominal segments 4and 5. Fisher’s exact test *P < 0.05 **P < 0.0001. NA, not analyzed. a n = 81.Additional file 2: Table S2. trxG, PcG and ETP alleles used in this study.These alleles are described in [79].Additional file 3: Figure S1. CycG misregulation modulates Hoxprotein profiles of Pc3 and ph-p410. Anti-SCR immunostainings of thirdinstar larval leg imaginal discs from (A) Pc3/+, Pc3/+; da > CycG RNAi, andPc3/da > CycGΔP males, and (B) ph-p410/Y, ph-p410/Y;da > CycG RNAi andph-p410/Y;da > CycGΔP males. L2, L3: mesothoracic and metathoracic legimaginal discs. H, haltere imaginal disc.Dupont et al. Epigenetics & Chromatin (2015) 8:18 Page 15 of 17Additional file 4: Table S3. Size of ectopic SCR and UBX domains inimaginal discs of PcG mutants. n: number of imaginal discs analyzed. L2,L3: mesothoracic and metathoracic leg imaginal discs. t-test *P < 0.05,**P < 0.005, and ***P < 0.0005. NA, not analyzed.AbbreviationsA4: A5, A6, abdominal segment 4, 5, 6; Abd-B: abdominal B; ASH1: absent,small, or homeotic discs 1; ASX: Additional sex combs; BRM: Brahma;CCNG1: Cyclin G1; CCNG2: Cyclin G2; CDK: cyclin-dependent kinase;crm: cramped; DSP1: dorsal switch protein 1; ESC: extra sex comb;ETP: enhancer of trithorax and polycomb; E(Z): enhancer of zeste;H3K27me3: tri-methylated histone H3 lysine 27; L1: prothoracic leg;L2: mesothoracic leg; L3: metathoracic leg; mxc: multi sex combs;PC: polycomb; PcG: polycomb-group; PH: polyhomeotic; PR-DUB: polycombrepressive deubiquitinase; PRC: polycomb repressive complex; PRE: PolycombResponse Element; RNA Pol II CTD: C-terminal domain of RNA polymerase II;Scr: Sex combs reduced; TAC1: trithorax activating complex 1; TrxG: trithorax-group; Ubx: Ultrabithorax.Competing interestsThe authors declare that they have no competing interests.Authors’ contributionsCAD performed genetic analyses, molecular cloning, co-immunoprecipitations,and antibody labeling. DDC performed immunolocalizations on polytenechromosome. MK and HWB conceived and carried out two-hybridscreens for ASX interactors. NBR participated in genetic analyses andantibody labeling. NBR and FP conceived and coordinated the study, andwrote the paper with the help of CAD. All authors read and approvedthe final manuscript.AcknowledgementsWe thank V. Ribeiro for excellent technical assistance, S. Bloyer, J-M. Gibert, and E.Mouchel-Vielh for stimulating discussions, and J-M. Gibert, E. Mouchel-Vielh, and J.Deraze for critical reading of the manuscript. We thank the Bloomington StockCenter for fly stocks. This work was funded by the Centre National de laRecherche Scientifique (CNRS), by the Université Pierre et Marie Curie - Paris 06,and by a Ligue Nationale Contre le Cancer (Comité Val d’Oise) grant to FP. CADwas funded by a doctoral fellowship from the MESR (Ministère de l’EnseignementSupérieur et de la Recherche).Author details1Sorbonne Universités, UPMC Univ Paris 06, Institut de Biologie Paris-Seine(IBPS), UMR 7622, Developmental Biology, 9, quai Saint-Bernard, F-75005Paris, France. 2CNRS, IBPS, UMR 7622, Developmental Biology, 9, quaiSaint-Bernard, F-75005 Paris, France. 3Lillehei Heart Institute and Departmentof Pediatrics, University of Minnesota, 2231 6th Street SE, Minneapolis, MN55455, USA. 4Department of Zoology, University of British Columbia, 6270University Boulevard, V6T 1Z4 Vancouver, BC, Canada.Received: 14 October 2014 Accepted: 1 April 2015References1. Gopinathan L, Ratnacaram CK, Kaldis P: Established and novel Cdk/cyclincomplexes regulating the cell cycle and development. In Results andproblems in cell differentiation. Volume 53. Jacek Z. Kubiak; 2011:365–389.2. Lim S, Kaldis P. Cdks, cyclins and CKIs: roles beyond cell cycle regulation.Development. 2013;140:3079–93.3. Egloff S, Dienstbier M, Murphy S. Updating the RNA polymerase CTD code:adding gene-specific layers. Trends Genet. 2012;28:333–41.4. Heidemann M, Hintermair C, Voß K, Eick D. Dynamic phosphorylationpatterns of RNA polymerase II CTD during transcription. Biochim BiophysActa. 2013;1829:55–62.5. Garriga J, Graña X. Cellular control of gene expression by T-type cyclin/CDK9 complexes. Gene. 2004;337:15–23.6. Zhou Q, Li T, Price DH. RNA polymerase II elongation control. Annu RevBiochem. 2012;81:119–43.7. Kwak H, Lis JT. Control of transcriptional elongation. Annu Rev Genet.2013;47:483–508.8. Bienvenu F, Jirawatnotai S, Elias JE, Meyer CA, Mizeracka K, Marson A, et al.Transcriptional role of cyclin D1 in development revealed by a genetic-proteomicscreen. Nature. 2010;463:374–8.9. Loncle N, Boube M, Joulia L, Boschiero C, Werner M, Cribbs DL, et al.Distinct roles for Mediator Cdk8 module subunits in Drosophiladevelopment. EMBO J. 2007;26:1045–54.10. Conaway RC, Conaway JW. The Mediator complex and transcriptionelongation. Biochim Biophys Acta. 2013;1829:69–75.11. Knuesel MT, Meyer KD, Bernecky C, Taatjes DJ. The human CDK8subcomplex is a molecular switch that controls Mediator coactivatorfunction. Genes Dev. 2009;23:439–51.12. Kumafuji M, Umemura H, Furumoto T, Fukasawa R, Tanaka A, Ohkuma Y.Mediator MED18 subunit plays a negative role in transcription via the CDK/cyclin module. Genes Cells. 2014;19:582–93.13. Kennison JA, Tamkun JW. Dosage-dependent modifiers of polycomb andantennapedia mutations in Drosophila. Proc Natl Acad Sci U S A.1988;85:8136–40.14. Gaytán De Ayala Alonso A, Gutiérrez L, Fritsch C, Papp B, Beuchle D, MüllerJ. A genetic screen identifies novel polycomb group genes in Drosophila.Genetics. 2007;176:2099–108.15. Di Croce L, Helin K. Transcriptional regulation by Polycomb group proteins.Nat Struct Mol Biol. 2013;20:1147–55.16. Schuettengruber B, Cavalli G. Recruitment of polycomb group complexesand their role in the dynamic regulation of cell fate choice. Development.2009;136:3531–42.17. Grimaud C, Nègre N, Cavalli G. From genetics to epigenetics: the tale ofPolycomb group and trithorax group genes. Chromosome Res.2006;14:363–75.18. Beck S, Faradji F, Brock H, Peronnet F. Maintenance of Hox gene expressionpatterns. Adv Exp Med Biol. 2010;689:41–62.19. Gildea JJ, Lopez R, Shearn A. A screen for new trithorax group genesidentified little imaginal discs, the Drosophila melanogaster homologue ofhuman retinoblastoma binding protein 2. Genetics. 2000;156:645–63.20. Schuettengruber B, Martinez A-M, Iovino N, Cavalli G. Trithorax group proteins:switching genes on and keeping them active. Nat Rev Mol Cell Biol.2011;12:799–814.21. Smith ST, Petruk S, Sedkov Y, Cho E, Tillib S, Canaani E, et al. Modulation ofheat shock gene expression by the TAC1 chromatin-modifying complex.Nat Cell Biol. 2004;6:162–7.22. Beisel C, Imhof A, Greene J, Kremmer E, Sauer F. Histone methylation by theDrosophila epigenetic transcriptional regulator Ash1. Nature. 2002;419:857–62.23. Tanaka Y, Katagiri Z, Kawahashi K, Kioussis D, Kitajima S. Trithorax-groupprotein ASH1 methylates histone H3 lysine 36. Gene. 2007;397:161–8.24. Yuan W, Xu M, Huang C, Liu N, Chen S, Zhu B. H3K36 methylationantagonizes PRC2-mediated H3K27 methylation. J Biol Chem.2011;286:7983–9.25. Mohrmann L, Verrijzer CP. Composition and functional specificity of SWI2/SNF2 class chromatin remodeling complexes. Biochim Biophys Acta.2005;1681:59–73.26. De Napoles M, Mermoud JE, Wakao R, Tang YA, Endoh M, Appanah R, et al.Polycomb group proteins Ring1A/B link ubiquitylation of histone H2A toheritable gene silencing and X inactivation. Dev Cell. 2004;7:663–76.27. Grau DJ, Chapman BA, Garlick JD, Borowsky M, Francis NJ, Kingston RE.Compaction of chromatin by diverse Polycomb group proteins requireslocalized regions of high charge. Genes Dev. 2011;25:2210–21.28. Scheuermann JC, de Ayala Alonso AG, Oktaba K, Ly-Hartig N, McGintyRK, Fraterman S, et al. Histone H2A deubiquitinase activity of thePolycombrepressive complex PR-DUB. Nature. 2010;465:243–7.29. Dey A, Seshasayee D, Noubade R, French DM, Liu J, Chaurushiya MS, et al.Loss of the tumor suppressor BAP1 causes myeloid transformation. Science.2012;337:1541–6.30. Baymaz HI, Fournier A, Laget S, Ji Z, Jansen PWTC, Smits AH, et al. MBD5and MBD6 interact with the human PR-DUB complex through their methyl-CpG-binding domain. Proteomics. 2014;14:2179–89.31. Scheuermann JC, Gutiérrez L, Müller J. Histone H2A monoubiquitination andPolycomb repression: the missing pieces of the puzzle. Fly. 2012;6:162–8.32. Milne TA, Sinclair DA, Brock HW. The Additional sex combs gene ofDrosophila is required for activation and repression of homeotic loci, andinteracts specifically with Polycomb and super sex combs. Mol Gen Genet.1999;261:753–61.Dupont et al. Epigenetics & Chromatin (2015) 8:18 Page 16 of 1733. Baskind HA, Na L, Ma Q, Patel MP, Geenen DL, Wang QT. Functionalconservation of Asxl2, a murine homolog for the Drosophila Enhancer ofTrithorax and Polycomb group gene Asx. PLoS One. 2009;4:e4750.34. Fisher CL, Lee I, Bloyer S, Bozza S, Chevalier J, Dahl A, et al. Additional sexcombs-like 1 belongs to the Enhancer of Trithorax and Polycomb groupand genetically interacts with Cbx2 in mice. Dev Biol. 2010;337:9–15.35. Berger N, Dubreucq B. Evolution goes GAGA: GAGA binding proteins acrosskingdoms. Biochim Biophys Acta. 2012;1819:863–8.36. Salvaing J, Lopez A, Boivin A, Deutsch JS, Peronnet F. The Drosophila Cortoprotein interacts with Polycomb-group proteins and the GAGA factor. NucleicAcids Res. 2003;31:2873–82.37. Salvaing J, Decoville M, Mouchel-Vielh E, Bussière M, Daulny A, Boldyreva L, et al.Corto and DSP1 interact and bind to a maintenance element of the Scr Hoxgene: understanding the role of Enhancers of trithorax and Polycomb. BMC Biol.2006;4:9.38. Déjardin J, Rappailles A, Cuvier O, Grimaud C, Decoville M, Locker D, et al.Recruitment of Drosophila Polycomb group proteins to chromatin by DSP1.Nature. 2005;434:533–8.39. Salvaing J, Nagel AC, Mouchel-Vielh E, Bloyer S, Maier D, Preiss A, et al.The Enhancer of Trithorax and Polycomb Corto interacts with Cyclin Gin Drosophila. PLoS One. 2008;3:e1658.40. Okamoto K, Beach D. Cyclin G is a transcriptional target of the p53 tumorsuppressor protein. EMBO J. 1994;13:4816.41. Kimura SH, Ikawa M, Ito A, Okabe M, Nojima H. Cyclin G1 is involved in G2/M arrest in response to DNA damage and in growth control after damagerecovery. Oncogene. 2001;20:3290–300.42. Smith ML, Kontny HU, Bortnick R, Fornace AJ. The p53-regulated Cyclin G genepromotes cell growth: p53 downstream effectors Cyclin G and Gadd45 exertdifferent effects on cisplatin chemosensitivity. Exp Cell Res. 1997;230:61–8.43. Bennin DA, Don ASA, Brake T, McKenzie JL, Rosenbaum H, Ortiz L, et al.Cyclin G2 associates with protein phosphatase 2A catalytic andregulatory B’ subunits in active complexes and induces nuclearaberrations and a G1/S phase cell cycle arrest. J Biol Chem.2002;277:27449–67.44. Horne MC, Goolsby GL, Donaldson KL, Tran D, Neubauer M, Wahl AF.Cyclin G1 and Cyclin G2 comprise a new family of cyclins withcontrasting tissue-specific and cell cycle-regulated expression. J BiolChem. 1996;271:6050–61.45. Faradji F, Bloyer S, Dardalhon-Cuménal D, Randsholt NB, Peronnet F. Drosophilamelanogaster Cyclin G coordinates cell growth and cell proliferation. Cell Cycle.2011;10:805–18.46. Faradji F, Bloyer S, Dardalhon-Cuménal D, Randsholt NB, Peronnet F. Erratumto Faradji F, et al. Cell Cycle Volume 10, Issue 5; pp. 805–818. Cell Cycle.2014;13:2480.47. Smith ML, Bortnick RA, Sheikh MS, Fornace AJ. Chromatin relaxation byoverexpression of mutant p53, HPV16-E6, or Cyclin G transgenes. Exp CellRes. 1998;242:235–43.48. Dietrich BH, Moore J, Kyba M, dos Santos G, McCloskey F, Milne TA, et al.Tantalus, a novel ASX-interacting protein with tissue-specific functions. DevBiol. 2001;234:441–53.49. Simon J, Chiang A, Bender W. Ten different Polycomb group genes arerequired for spatial control of the abdA and AbdB homeotic products.Development. 1992;114:493–505.50. Beuchle D, Struhl G, Müller J. Polycomb group proteins and heritablesilencing of Drosophila Hox genes. Development. 2001;128:993–1004.51. Yamamoto Y, Girard F, Bello B, Affolter M, Gehring WJ. The cramped gene ofDrosophila is a member of the Polycomb-group, and interacts with mus209,the gene encoding Proliferating Cell Nuclear Antigen. Development.1997;124:3385–94.52. Santamaría P, Randsholt NB. Characterization of a region of the Xchromosome of Drosophila including multi sex combs (mxc), a Polycombgroup gene which also functions as a tumour suppressor. Mol Gen Genet.1995;246:282–90.53. Ingham P, Whittle R. Trithorax: a new homoeotic mutation of Drosophilamelanogaster causing transformations of abdominal and thoracic imaginalsegments. Mol Gen Genet. 1980;179:607–14.54. Kennison JA. The Polycomb and Trithorax group proteins of Drosophila:trans-regulators of homeotic gene function. Annu Rev Genet.1995;29:289–303.55. Struhl G. Genes controlling segmental specification in the Drosophilathorax. Proc Natl Acad Sci U S A. 1982;79:7380–4.56. Pattatucci AM, Kaufman TC. The homeotic gene Sex combs reduced ofDrosophila melanogaster is differentially regulated in the embryonic andimaginal stages of development. Genetics. 1991;129:443–61.57. Lewis EB. A gene complex controlling segmentation in Drosophila. Nature.1978;276:565–70.58. Salvaing J, Mouchel-Vielh E, Bloyer S, Preiss A, Peronnet F. Regulation ofAbd-B expression by Cyclin G and Corto in the abdominal epithelium ofDrosophila. Hereditas. 2008;145:138–46.59. Fisher CL, Randazzo F, Humphries RK, Brock HW. Characterization of Asxl1, amurine homolog of Additional sex combs, and analysis of the Asx-like genefamily. Gene. 2006;369:109–18.60. Takeda DY, Wohlschlegel JA, Dutta A. A bipartite substrate recognitionmotif for cyclin-dependent kinases. J Biol Chem. 2001;276:1993–7.61. Weatheritt RJ, Jehl P, Dinkel H, Gibson TJ. iELM - a web server to exploreshort linear motif-mediated interactions. Nucleic Acids Res. 2012;40(Web Server issue):W364–9.62. Mouchel-Vielh E, Rougeot J, Decoville M, Peronnet F. The MAP kinase ERKand its scaffold protein MP1 interact with the chromatin regulator Cortoduring Drosophila wing tissue development. BMC Dev Biol. 2011;11:17.63. Coléno-Costes A, Jang SM, de Vanssay A, Rougeot J, Bouceba T, RandsholtNB, et al. New partners in regulation of gene expression: the Enhancer ofTrithorax and Polycomb Corto interacts with methylated Ribosomal ProteinL12 via its chromodomain. PLoS Genet. 2012;8:e1003006.64. Voncken JW, Schweizer D, Aagaard L, Sattler L, Jantsch MF, van Lohuizen M.Chromatin-association of the Polycomb group protein BMI1 is cell cycle-regulated and correlates with its phosphorylation status. J Cell Sci.1999;112(Pt 24):4627–39.65. Wu SC, Zhang Y. Cyclin-dependent kinase 1 (CDK1)-mediated phosphorylationof Enhancer of zeste 2 (Ezh2) regulates its stability. J Biol Chem.2011;286:28511–9.66. Lecona E, Rojas LA, Bonasio R, Johnston A, Fernández-Capetillo O,Reinberg D. Polycomb protein SCML2 regulates the cell cycle bybinding and modulating CDK/Cyclin/p21 complexes. PLoS Biol.2013;11:e1001737.67. Chatr-Aryamontri A, Breitkreutz B-J, Heinicke S, Boucher L, Winter A, Stark C,et al. The BioGRID interaction database: 2013 update. Nucleic Acids Res.2013;41(Database issue):D816–23.68. Murali T, Pacifico S, Yu J, Guest S, Roberts GG, Finley RL. DroID 2011: acomprehensive, integrated resource for protein, transcription factor,RNA and gene interactions for Drosophila. Nucleic Acids Res.2011;39(Database issue):D736–43.69. Sinclair DA, Campbell RB, Nicholls F, Slade E, Brock HW. Genetic analysis of theAdditional sex combs locus of Drosophila melanogaster. Genetics. 1992;130:817–25.70. Saebøe-Larssen S, Lyamouri M, Merriam J, Oksvold MP, Lambertsson A.Ribosomal protein insufficiency and the minute syndrome in Drosophila: adose–response relationship. Genetics. 1998;148:1215–24.71. Marygold SJ, Roote J, Reuter G, Lambertsson A, Ashburner M, Millburn GH,et al. The ribosomal protein genes and Minute loci of Drosophilamelanogaster. Genome Biol. 2007;8:R216.72. Fauvarque MO, Laurenti P, Boivin A, Bloyer S, Griffin-Shea R, Bourbon HM,et al. Dominant modifiers of the polyhomeotic extra-sex-combs phenotypeinduced by marked P element insertional mutagenesis in Drosophila. GenetRes. 2001;78:137–48.73. Kennison JA, Russell MA. Dosage-dependent modifiers of homeoticmutations in Drosophila melanogaster. Genetics. 1987;116:75–86.74. Marenda DR, Zraly CB, Dingwall AK. The Drosophila Brahma (SWI/SNF)chromatin remodeling complex exhibits cell-type specific activation andrepression functions. Dev Biol. 2004;267:279–93.75. Bernstein BE, Mikkelsen TS, Xie X, Kamal M, Huebert DJ, Cuff J, et al. Abivalent chromatin structure marks key developmental genes in embryonicstem cells. Cell. 2006;125:315–26.76. Stock JK, Giadrossi S, Casanova M, Brookes E, Vidal M, Koseki H, et al.Ring1-mediated ubiquitination of H2A restrains poised RNA polymeraseII at bivalent genes in mouse ES cells. Nat Cell Biol. 2007;9:1428–35.77. Schuettengruber B, Cavalli G. The DUBle life of polycomb complexes. DevCell. 2010;18:878–80.78. Bateman JR, Lee AM, Wu CT. Site-specific transformation of Drosophila viaphiC31 integrase-mediated cassette exchange. Genetics. 2006;173:769–77.79. St Pierre SE, Ponting L, Stefancsik R, McQuilton P. the FlyBase Consortium:FlyBase 102–advanced approaches to interrogating FlyBase. Nucleic AcidsRes. 2014;42:D780–8.80. Kyba M, Brock HW. The Drosophila Polycomb group protein Psc contactsph and Pc through specific conserved domains. Mol Cell Biol.1998;18:2712–20.81. Sinclair DA, Milne TA, Hodgson JW, Shellard J, Salinas CA, Kyba M, et al. TheAdditional sex combs gene of Drosophila encodes a chromatin protein thatbinds to shared and unique Polycomb group sites on polytenechromosomes. Development. 1998;125:1207–16.82. Randsholt NB, Santamaria P. How Drosophila change their combs: the Hoxgene Sex combs reduced and sex comb variation among Sophophoraspecies. Evol Dev. 2008;10:121–33.83. LaJeunesse D, Shearn A. E(z): a polycomb group gene or a trithorax groupgene? Development. 1996;122:2189–97.84. White RA, Wilcox M. Protein products of the bithorax complex in Drosophila.Cell. 1984;39:163–71.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionDupont et al. Epigenetics & Chromatin (2015) 8:18 Page 17 of 17Submit your manuscript at www.biomedcentral.com/submit"@en ; edm:hasType "Article"@en ; edm:isShownAt "10.14288/1.0223369"@en ; dcterms:language "eng"@en ; ns0:peerReviewStatus "Reviewed"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "BioMed Central"@en ; ns0:publisherDOI "10.1186/s13072-015-0008-6"@en ; dcterms:rights "Attribution 4.0 International (CC BY 4.0)"@en ; ns0:rightsURI "http://creativecommons.org/licenses/by/4.0/"@en ; ns0:scholarLevel "Faculty"@en ; dcterms:subject "Polycomb"@en, "Trithorax"@en, "Cyclin G"@en, "Homeotic"@en ; dcterms:title "Drosophila Cyclin G and epigenetic maintenance of gene expression during development"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/56474"@en .