A Novel Component of the Drosophila Histone Locus Body Required for Muscle Integrity

Research Article 2697 Muscle wasted: a novel component of the Drosophila histone locus body required for muscle integrity

Sarada Bulchand*, Sree Devi Menon‡, Simi Elizabeth George§ and William Chia Temasek Lifesciences Laboratory, National University of Singapore, 1 Research Link, 117604, Singapore *Author for correspondence ([email protected]) ‡Present address: Ministry of Education, 248922, Singapore §Present address: Neuroscience and Behavioural Disorders Programme, Duke-NUS Graduate Medical School, 8 College Road, 169857, Singapore

Summary Skeletal muscles arise by cellular differentiation and regulated gene expression. Terminal differentiation programmes such as muscle growth, extension and attachment to the epidermis, lead to maturation of the muscles. These events require changes in chromatin organization as genes are differentially regulated. Here, we identify and characterise muscle wasted (mute), a novel component of the Drosophila histone locus body (HLB). We demonstrate that a mutation in mute leads to severe loss of muscle mass and an increase in levels of normal histone transcripts. Importantly, Drosophila Myocyte enhancer factor 2 (Mef2), a central myogenic differentiation factor, and how, an RNA binding protein required for muscle and tendon cell differentiation, are downregulated. Mef2 targets are, in turn, misregulated. Notably, the degenerating muscles in mute mutants show aberrant localisation of heterochromatin protein 1 (HP1). We further show a genetic interaction between mute and the Stem-loop binding protein (Slbp) and a loss of muscle striations in Lsm11 mutants. These data demonstrate a novel role of HLB components and histone processing factors in the maintenance of muscle integrity. We speculate that mute regulates terminal muscle differentiation possibly through heterochromatic reorganisation.

Key words: Muscle differentiation, Histone locus body, Heterochromatin

Introduction 2001; Harris et al., 1991; Mitra et al., 2003; Stauber and Schumperli, Drosophila melanogaster provides a simple model system to study 1988; Zhao et al., 2000). Unlike polyadenylated mRNAs, the 3Јend the development of the body wall muscles (analogous to vertebrate of metazoan histone mRNAs terminate in a stem loop (Marzluff, skeletal muscles). A stereotypic pattern of ~30 unique muscles 1992). Several trans-acting factors (Wang et al., 2006; Martin et in every embryonic hemisegment develop during mid-late al., 1997; Galli et al., 1983; Mowry and Steitz, 1987) recruit a

Journal of Cell Science embryogenesis, contract during late embryogenesis and in the complex that triggers cleavage between the stem loop and the developing larva where they are critical for motility (Bate, 1993; histone downstream element (HDE), forming the mature histone Broadie and Bate, 1993). Several muscle-specific genes are mRNA (Dominski et al., 2005; Kolev and Steitz, 2005; Wagner et conserved with those of vertebrates and some principles of muscle al., 2007). Drosophila Stem-loop binding protein (SLBP) (Sullivan development are similar (Abmayr et al., 2003; Baylies and et al., 2001), U7snRNA- (Dominski et al., 2003) and the U7snRNP- Michelson, 2001; Taylor, 1998). specific proteins Lsm10 and Lsm11 (Azzouz and Schumperli, Early differentiation programmes, around embryonic stages 8- 2003) have been identified as factors required for histone pre- 9, lead to the commitment of mesodermal progenitors that specify mRNA processing (Godfrey et al., 2006; Godfrey et al., 2009; founder cells (FC), which fuse with several fusion-competent Sullivan et al., 2001). U7snRNP particles have been shown to myoblasts (FCM) to form syncitial myotubes (Bate, 1990). The localise to the histone locus body (HLB), a distinct nuclear body activation of subsequent differentiation programmes between stages that is often in close proximity to the Cajal body (CB) (Liu et al., 15-17, including muscle growth, extension and attachment to 2006). The HLB is associated with the histone gene cluster and has specialised epithelial cells lead to muscle maturation. been speculated to play multiple roles in assembly, modification, Drosophila Myocyte enhancer factor 2 (Mef2) appears to be storage and delivery of the histone pre-mRNA processing central to the myogenic differentiation process and has diverse machinery (Liu et al., 2006). functions throughout myogenesis that include the temporal We present the isolation and characterisation of muscle wasted regulation of muscle genes (Elgar et al., 2008; Sandmann et al., (mute) that appears to be a novel component of the HLB. Animals 2006). Several studies have shown that chromatin states change in that lack mute function show the progressive loss of muscle mass response to cellular signalling and gene activity (Baxter et al., and detached muscles during late embryonic development. 2004; Haaf et al., 1990). Chromatin remodelling at muscle-specific Although early differentiation of myoblasts is not affected, terminal loci is required to regulate muscle differentiation (Yahi et al., differentiation appears to be defective. We show that Mef2 and 2006). Histones and other chromatin remodelling complexes are other terminal-muscle-differentiation genes are downregulated. In components of intricate epigenetic mechanisms that organise addition, levels of the predicted Drosophila homologue of atrogin- genomes into discrete chromatin domains to regulate gene 1 are increased, similar to that observed in vertebrate models of expression (Boyer et al., 2004; Quina et al., 2006). Histone levels skeletal muscle atrophy (Glass, 2003; Gomes et al., 2001). Mute are primarily controlled by regulating histone transcript abundance appears to regulate histone gene expression since levels of histone at the transcriptional and post-transcriptional levels (Berloco et al., transcripts are significantly elevated and small amounts of 2698 Journal of Cell Science 123 (16)

misprocessed histone transcripts are also present in mute mutants. chromosome 2 (Fig. 1C). A base change from C to T was detected The aberrant localisation of Heterochromatin protein 1 (HP1) has at position 559 of the coding sequence. CG34415 spans ~6.8 kb of led us to speculate that changes in heterochromatin may lead to the genome and is predicted to encode two isoforms that we genome instability and the misregulation of muscle genes. We also verified by northern analysis using total RNA from staged wild- show genetic interactions between mute and Slbp, and the loss of type (WT) embryos. Both isoforms, mute-L (long) and mute-S striations in larval muscles of Lsm11 mutants. To our knowledge, (short), are expressed from 0-5 hours after egg laying (AEL) and this is the first report supporting the role of these factors in are not detectable at later stages of embryogenesis. mute-L migrates differentiated muscles. at ~5.5 kb and mute-S at ~3.3 kb (Fig. 1E). The mutation in mute1281 changed amino acid Q187, producing Results a premature stop codon, TAA, which is predicted to produce a Identification and molecular cloning of mute truncated Mute-L protein (Fig. 1C, red triangle). A P-element To identify novel genes required for myogenesis, we performed an insertion P{GawB}NP1160, 90 bases upstream of the ATG of EMS-mutagenesis screen of the second chromosome. Embryos Mute-L also resulted in a similar degenerative muscle phenotype homozygous for an allele 1281, showed a severe loss of muscle when present in trans with mute1281 (Fig. 1C, green triangle, data mass during late embryogenesis (Fig. 1A,B). Using deficiency not shown). mute-L and mute-S encode predicted proteins of 1739 mapping, we found Df(2R)Exel 6065 (breakpoints 53D14-53F8) and 665 amino acids in length, respectively (Fig. 1D). Simple produced the degenerative muscle phenotype when in trans with modular architecture research tool (SMART) homology searches 1281. We named this locus muscle wasted (mute) and the EMS revealed that the N-terminal region of Mute-L [amino acid (aa) allele as mute1281. Upon sequencing, we identified a mutation in 216-1088] is weakly homologous to the Atrophin-1 protein family. CG34415 that lies at the 53E2 locus, on the right arm of The C terminal region (aa 1680-1732) is weakly homologous to the SANT (Swi3, Ada2, N-CoR and TFIIIB) DNA-binding domain. InterPro predicts the presence of a homeodomain-like region (aa 1657-1735) and two paired amphipathic helix (PAH) repeats (aa 1121-1198 and 1204-1279) in Mute-L, regions also common to Mute-S (Fig. 1D). Some PAH-repeat-containing proteins are known to function as components of co-repressor complexes in transcriptional silencing (Wang et al., 1990). SANT and homeodomains function in the regulation of transcription (Boyer et al., 2004; Hueber and Lohmann, 2008). Interestingly, BLAST searches revealed that the PAH repeats of Mute-L are 28% identical to that of zebrafish Ugly duckling (Udu), a nuclear protein required for somite development (Lim et al., 2009; Liu et al., 2007) and 27% identical to that of human GON4L or YARP (YY1AP-related protein 1), a nuclear protein with suggested transcriptional-

Journal of Cell Science regulator-like functions (supplementary material Fig. S1).

mute is required for maintenance of muscle mass and integrity Somatic muscles were analysed from stage 14, during the peak of myogenesis, up to the end of embryogenesis at stage 17, when the muscles are fully developed. The phenotype in mute1281 homozygotes was less severe compared with mute1281/Df (2R)Exel 6065 Fig. 1. Identification and molecular cloning of mute. (A)Wild-type (WT) embryos in ~10% of mutant embryos analysed (n4/40). Hence, 1281/1281 and (B) mute stage 16 embryos. Somatic muscles were labelled with mute1281/Df (2R)Exel 6065 embryos were analysed in all subsequent anti-Myosin heavy chain (MHC; red). Muscles (outlined) are thinner in 1281/1281 experiments, to ensure the removal of one copy of the entire mute mute embryos. Scale bar: 20m. (C)The mute locus is at 53E2 on the locus. The size and shape of the dorsal muscles DO (outlined in right arm of chromosome 2. CG34415 encodes mute-L and mute-S. mute-L has a unique N-terminal region (blue). The C-terminal region is the same in both dashed lines in Fig. 2) and DA (outlined in solid lines), were isoforms (black). The EMS allele has a C to T change at position 559 of the compared. At stage 14, no significant difference in muscle 1281/Df coding sequence leading to a premature stop codon, TAA (red triangle), at morphology or mass was observed between mute and WT amino acid (aa) 187, depicted as mute1281. The P element P{Gaw}NP1160 embryos (Fig. 2A,B). By stage 15, some muscles appeared thinner (green triangle) is inserted 90 bases upstream of the ATG of mute-L. White than their WT counterparts (Fig. 2C,D). Also, some muscles were boxes indicate UTRs. Coloured bars under mute-L indicate domains depicted no longer detectable in mute1281/Df embryos (Fig. 2D, asterisks). in D. (D)Protein structure of the 1739 aa Mute-L and 665 aa Mute-S. The Although only dorsal muscles are shown at these stages, it should region of aa 216-1088 of Mute-L is weakly homologous to Atrophin-1. Two be noted that most somatic muscles showed similar defects. At paired amphipathic helix (PAH) repeats extend from aa 1121-1198 and 1204- stage 16 many muscles appeared significantly thinner in mute1281/Df 1279. A C-terminal homeodomain-like (aa 1657-1735)/SANT DNA binding embryos compared with their WT counterparts (Fig. 2E-J, outlined). domain (aa1680-1732) is also predicted. (E)Northern blot using total RNA from WT embryos. The probe detects the C-terminus of mute-L and mute-S. There was no bias towards any specific muscle type. Some appeared These two isoforms are expressed at early stages after egg laying (AEL) and to be detached at least at one end (Fig. 2F,H,J, arrows). Owing to show the presence of a maternal component in 0- to 1-hour embryos. No the reduced muscle mass, gaps between muscle subsets were larger, expression is detected at later stages. Sizes are approximations (in kb). rp49 is making the underlying visceral mesoderm more visible (Fig. 2F, the loading control. v). At stage 16 some muscles could not be detected (Fig. 2F,J, mute is required for muscle integrity 2699

Fig. 2. Progressive loss of muscle mass in mute1281 mutants. (A,C,E,G,I) WT and (B,D,F,H,J) mute1281/Df somatic muscles labelled with anti-MHC (red). DO and DA muscles are outlined with dashes and solid lines, respectively. (A,B)At stage 14 muscle mass in mute1281/Df is comparable to that in WT. (C,D)By stage 15 some muscles appear thinner. Gaps between muscles begin to appear (D, asterisk). (E-J)At stage 16 muscles are significantly thinner. Dorsal, lateral and ventral groups are equally affected. Some muscles appear to be missing (F,J, asterisk). The underlying visceral mesoderm is more visible because of wide gaps (F,‘v’). (K,L) Stage17 WT and mute1281/Df somatic muscles labelled with an MHC-GFP reporter. mute1281 muscles are significantly thinner and have abnormal morphologies. Many are detached and rounded (asterisk), eventually forming blebs (L, arrow). Some are extremely thin (L, arrowhead). (M)Schematic representation of larval and embryonic somatic muscles. (N)Schematic representation of UAS-mute constructs used to rescue mute1281. Numbers indicate length (as the number of amino acids). The percentage of embryos rescued is indicated. (O-R)Dorsal muscles of stage 16 embryos labelled with anti-MHC. All constructs are expressed in mute1281/Df embryos using 24B GAL4 at 25°C. UAS- mute-L rescues the muscle wasting defect (O) to a level comparable to that observed in WT embryos (E). UAS-mute-S fails to rescue the muscle defect, as shown by the persistence of thin muscles (P, arrows). UAS-mute-LC is able to rescue the muscle defect (Q). UAS-muteC Atro fails to rescue the muscle defect (R). Similar results were obtained with multiple

Journal of Cell Science independent transgenic lines. Anterior is left, dorsal is up. Vertical white bars demarcate segment boundaries. Scale bar: 20m.

asterisk). To observe the musculature after the deposition of the 24BGAL4 and rp298GAL4 which carry the yeast GAL4 under the cuticle, a cytoplasmic GFP reporter was expressed under the control how (held out wings) and duf (dumbfounded) promoters, of the Myosin heavy chain (Mhc) promoter (Fig. 2K,L). By stage respectively. The results are summarised in Fig. 2N. Expression 17, the muscles degenerated to an extent where the WT pattern of mute-L in mute1281/Df embryos in the presence of 24B-GAL4 or (Fig. 2K,M) was no longer recognisable. All the muscles lost their rp298-GAL4 completely rescued the muscle-wasting phenotype characteristic WT morphology and appeared spherical in shape (n10/10; Fig. 2N,O). mute-S rescued the muscle defect in only (Fig. 2L, asterisk) or became extremely thin (Fig. 2L, arrowhead). 18% of the embryos (n2/11; Fig. 2N,P). In order to test the More than half of the muscles eventually fragmented into small possible role of the SANT DNA-binding and/or homeodomain- blebs (Fig. 2L, arrow) and the embryos died prior to hatching like region, a construct that deleted this region (mute-LC) was (n200). These morphologies are characteristic of degenerating similarly introduced into the mute1281/Df background. mute-LC muscles (Carrasco-Rando and Ruiz-Gomez, 2008; Nguyen et al., was completely able to rescue the muscle defect (n10/10; Fig. 2007). These results demonstrate that mute loss of function results 2N,Q). A truncated construct with a premature stop in the Atrophin- in progressive muscle degeneration late in embryonic development 1 like region (mute-LC Atro), failed to rescue the muscle defect and that mute is required for maintenance of muscle integrity. (n0/10; Fig. 2N,R). These results demonstrate that the observed Mute transcripts are detected at 0-1 hour after egg laying (AEL) muscle defect is due to the mutation in CG34415 that affects indicative of a maternal component (Fig. 1E). It is possible that mute-L. mute-S may be able to compensate weakly for the function this masks an early phenotype. We were unable to ascertain this of mute-L although in the majority of mutant embryos expression since mute1281germline clones do not lay eggs. of mute-S was unable to rescue the muscle defect. The To confirm that the observed muscle defects were due to a homeodomain-like and/or SANT DNA-binding region does not mutation in CG34415, transgenic lines carrying mute-L or mute- seem to be important for the function of mute in the somatic S under the UAS promoter, were generated. All UAS constructs muscles. It is possible that this region is required for its function were expressed under the control of the mesodermal drivers in other tissues. 2700 Journal of Cell Science 123 (16)

Late differentiation is defective in mute1281 reduction in Eve-positive nuclei (Fig. 3A,B, arrow). Importantly, In order to determine if the late muscle-wasting phenotype is the expression of Mef2 was reduced in the nuclei of mature caused by defects in early or late myogenic events, we tested mute1281/Df myotubes (Fig. 3C,D). As myogenesis proceeds, muscles stages of myogenesis from early specification to late differentiation. extend and attach to specialised epithelial cells, the tendon cells. The overall numbers of muscle progenitors in the early mesoderm D-Titin labels the mature myotube and is enriched at sites of was unaffected, as shown by the expression of Mef2 (supplementary muscle attachment at late embryonic stages. D-Titin was unevenly material Fig. S2A,B) and there was no significant difference in FC expressed in the cytoplasm (Fig. 3E,F, arrowheads) and reduced at numbers using the enhancer trap line, rp298-lacZ (supplementary the myotube attachment sites in mute1281/Df embryos (Fig. 3E,F, material Fig. S2C,D). The number of FCMs was similar to that of arrows). The reporter line, Kettin-GFP, another attachment-site the WT, as shown by the expression of D-Titin (also known as marker, also showed a reduction in expression at muscle attachment sallimus) (supplementary material Fig. S2E,F). Also, muscle sites (Fig. 3G,H, arrows). We did not observe any defects in the specification was unaffected as shown by the expression of Even general organisation of the epithelium when muscles were thinner skipped (Eve) in the muscle nuclei at stage 13 (supplementary than normal (data not shown). material Fig. S2G,H). Similar results were obtained using the DO1 Thus, the specification of myoblasts and muscles is unaffected marker, Kruppel (data not shown). in mute1281. The thinning muscles continue to express myotube We next examined whether late specification and differentiation specification markers. Hence the loss in muscle mass is due to were affected. Even as muscles started to lose their mass at stage neither defective myoblast specification nor the inability to maintain 16, Eve expression was sustained and there was no significant the expression of specification genes in myonuclei. Late decrease in nuclear number. A small fraction (one in seven DA1 differentiation is affected as observed by a reduction in Mef2 muscles) of abnormally shaped muscles showed a more drastic expression and attachment site markers concurrently with detached muscles. It is possible that these defects result in the inability of mature muscles to maintain their integrity.

mute1281 muscles do not undergo apoptosis Previously it has been shown that apoptotic markers are upregulated in degenerating muscles (Nguyen et al., 2007). TUNEL stainings and anti-caspase 3 antibodies, apoptotic markers, labelled cells in both WT and mute1281/Df embryos, below the plane of the myotubes. These are probably fusion-competent myoblasts, given their location (data not shown and supplementary material Fig. S3A,B, arrowheads). In addition, we also tested whether the muscle- wasting defect was rescued by removing grim, reaper (rpr) and head involution defective (hid), key regulators of apoptosis in Drosophila, using the chromosomal deficiency H99. Double 1281

Journal of Cell Science mutants homozygous for mute and H99 continued to show thin muscles characteristic of mute1281/Df (supplementary material Fig. S3C,D). This suggests that the muscle-wasting defect is not due to apoptosis in the myotubes.

Mute is a nuclear protein Antibodies were generated against the N- (anti-Mute-L, aa 3-164) and C-termini (anti-Mute-LS, aa 1093-1291) of the mute coding sequence and used to study Mute expression pattern in embryos and larvae. Anti-Mute-L, which recognises the N-terminus of Mute- L, detects a nuclear protein that is expressed as a prominent nuclear focus in all cells of the embryo. A single nuclear focus was clearly observed in WT somatic muscles (Fig. 4A, arrow), which was undetectable in mute1281/Df embryos (Fig. 4E, arrow). A similar pattern was detected using anti-Mute-LS antibodies, which recognise the C-terminus that is common to both Mute-L and Mute-S (Fig. 4B, arrow). Expression levels were drastically reduced in mute1281/Df embryos (Fig. 4F, arrow). It is probable that this Fig. 3. Late muscle differentiation is defective in mute1281embryos. weak expression is that of Mute-S, which is also nuclear but is 1281/Df (A,C,E,G) WT and (B,D,F,H) mute late stage 16 embryos. Dorsal insufficient for normal function. The anti-Mute-L and anti-Mute- muscles labelled with anti-MHC (red; A-D, G-H) or D-Titin (red; E,F). LS-positive foci overlapped in both muscle (Fig. 4C, arrow) and (A,B) Mature DA1 myotubes continue to express Eve (outlined). Occasionally  non muscle (Fig. 4C, arrowhead) nuclei. Protein expression in the abnormally shaped muscles show a reduction in Eve-positive nuclei (B, arrow). (C,D)Mef2 (green) expression is reduced in mute1281/Df (D) compared mutant was restored by rescue with UAS-mute-L expressed under with WT (C) embryos. (E,F)The enrichment of D-Titin at attachment sites is 24B-GAL4, which drives expression in muscle and tendon cells reduced in mute1281/Df embryos (E,F, arrow) and the cytoplasmic stain is (Fig. 4G, arrows indicate muscle, arrowhead indicates tendon cell). uneven (E, F, arrowheads). (G,H)Kettin-GFP (green) at the attachment sites is The diffuse nuclear stain is probably due to overexpression of the markedly reduced in mute1281embryos (G,H). Scale bar: 20m. protein. mute is required for muscle integrity 2701 Journal of Cell Science Fig. 4. Mute is a nuclear protein and colocalises with histone locus body markers. (A-C,E-F) Stage 16 dorsal muscles labelled with anti-MHC (red) and DNA labelled with Hoechst (blue). Insets are magnified regions of the same. WT embryos labelled with anti-Mute-L (A, pink) and anti-Mute-LS (B, green) show the expression of Mute in a discrete nuclear body in all nuclei (arrow). Both antibodies label the same structure (C, yellow; arrow indicates muscle; arrowhead indicates non muscle). This expression using anti-Mute-L is undetectable (E, arrow) and using anti-Mute-LS is highly reduced (F, arrow) in mute1281/Df embryos. This expression is restored in muscle (G, arrow) and tendon (G, arrowhead) cell nuclei in mutant embryos rescued with UAS-mute-L driven by 24B-GAL4. (D)WT third instar larval salivary gland nucleus labelled with anti-Mute-LS (green) and Lamin (red) shows the presence of a single Mute-positive locus associated with the polytene chromosomes (arrow). (I-K and N-P) Stage 16 WT embryos. Somatic muscles in one hemisegment labelled with anti-MHC (pink), anti-Mute-LS (green), anti-Lsm10 (J, red) and anti-Lsm11 (O, red). Mute colocalises with Lsm10 (I-K, arrow) and Lsm11 (N-P, arrow). (L,Q)Salivary gland nuclei labelled with anti- Mute-LS (green), anti-Lamin (red outline), Lsm10 (L, red) and Lsm11 (Q, red). Mute colocalises with Lsm10 and Lsm11 on the polytene chromosome (L,Q, arrow). (M,R)Stage 16 mute1281/Df embryos labelled with anti-MHC (red), Lsm10 (M, green) and Lsm11 (R, green). Lsm10 and Lsm11 foci are present in mute1281/Df somatic muscles (arrows). (H)HeLa cells labelled with anti-FLASH (red), anti-Mute-LS (green) and Hoechst (blue). Mute colocalises with FLASH (arrows). Scale bars: 10m.

To assess whether Mute associates with DNA, we tested its investigated whether Mute colocalised with markers for these localisation in the larger salivary gland nuclei of WT third instar nuclear structures, such as Lsm10 and Lsm11, components of the larvae. We found that Mute was expressed at a single locus HLB (Liu et al., 2006). We found that Mute colocalised with both associated with the large polytene chromosomes (Fig. 4D, Lsm10 and Lsm11 in the somatic muscles (Fig. 4I-K and 4N-P, arrowhead). Thus, the antibodies generated against Mute label a arrows) as well as other cells of the embryo (not shown). specific nuclear organelle or body and Mute appears to associate Colocalisation was also apparent in salivary gland nuclei (Fig. with a specific chromosomal locus. 4L,Q, arrows). The localisation of Lsm10 and Lsm11 was unaffected in the somatic muscles of mute1281/Df embryos (Fig. Mute is a novel component of the histone locus body 4M,R, arrows). The nuclear expression of Mute was reminiscent of the localisation Thus, Mute may be a novel component of the Drosophila HLB patterns of components of two nuclear bodies, the Cajal body (CB) that is not required for the localisation of Lsm10 and Lsm11. This and the histone locus body (HLB). In order to determine whether suggests that mute is not required for the organisation of the HLB. Mute was a component of either of these nuclear bodies, we Interestingly anti-Mute-LS also recognised a sub nuclear body in 2702 Journal of Cell Science 123 (16)

HeLa cells that colocalised with FLASH (FLICE-associated huge misprocessed histone transcripts (Godfrey et al., 2006; Sullivan et protein; Fig. 4H, arrows). FLASH-positive nuclear bodies have al., 2001), we examined the relative changes in polyadenylated been suggested to be similar to the Drosophila HLB and have been histone transcripts in mute1281 versus Slbp15 and U720. Although associated with histone transcription (Barcaroli et al., 2006; we found higher levels of poly(A) histone H3 (7.14±2.9 fold) and Bongiorno-Borbone et al., 2008). These results suggest that the histone H4 (7.41±1.77 fold) mRNAs in mute1281/Df embryos at epitope recognised by anti-Mute-LS might be conserved, and a early stage 17 by quantitative RT-PCR (qPCR), the levels of structural or functional homologue of Mute might exist in misprocessed histone H3 and H4 transcripts were ~5 to 6-fold less vertebrates. than that observed in stage 17 Slbp15/15 embryos and U720/20 third instar larvae compared with stage matched WT controls (Fig. 5A). mute regulates histone mRNA transcription Interestingly, northern analysis showed higher levels of normal H3 It has previously been shown that mutations in components of the and H4 transcripts in mute1281/Df embryos unlike Slbp15 and U720 histone pre-mRNA processing machinery results in the production mutants, which had high levels of mostly misprocessed transcripts of aberrantly long mRNAs with polyadenylated tails because of (Fig. 5D), as has been reported previously (Godfrey et al., 2006; the presence of cryptic polyadenylation signals downstream of the Sullivan et al., 2001). Similarly, qPCR using histone-specific cleavage site in the histone gene (Godfrey et al., 2006; Godfrey et primers for the reverse transcription, also showed significantly al., 2009; Lanzotti et al., 2002). Since U720 and Slbp15 mutants higher total levels of H3 and H4 transcripts, in mute1281/Df embryos have previously been shown to accumulate high levels of [compare Fig. 5A poly(A) fraction to 5C-total transcript].

Fig. 5. Histone pre-mRNAs are misprocessed and HP1 localisation is aberrant in mute1281. (A)Quantitative RT-PCR (qPCR) using oligo(dT) on total RNA extracted from early stage 17 Journal of Cell Science embryos. All values on the y-axis are normalized against the ribosomal gene rp49. H3 and H4 poly(A) transcripts were higher by 7.14±2.9, 37.73±4.28, 41.51±5.65 and 7.41±1.77, 17.24±0.43, 18.91±2.4 fold in mute1281/Df, Slbp15/15 and U720 mutants, respectively (Student’s t-test P<0.05). Error bars show standard deviation from three biological replicates. (B)Western blot of acid- extracted histones and heterochromatin protein 1 (HP1) from early stage 17 embryos. 5g protein was loaded per lane. Levels of Ac H3K9 were lower (arrow) in mute1281/Df compared with WT embryos. (C)qPCR using histone-specific primers shows more total H3 and H4 transcripts in mute1281/Df compared with WT embryos, by 13.73±1.69 and 13.69±2.26 fold, respectively (P<0.05). Error bars show standard deviation of three biological replicates. (D)Northern analysis on 1g total RNA shows the presence of misprocessed H3 and H4 in Slbp15/15 and U720 mutants and more total H3 and H4 transcripts in mute1281/Df embryos. (E-L)Dorsal muscles (outlined in H and L) of late stage 16 embryos labelled with Hoechst (blue), anti-HP1 (red) and anti-MHC (green). HP1 encircles the heterochromatin in WT muscles (F, arrow and inset). The distribution of HP1 is more punctate than circular in mute1281/Df myonuclei (J, arrow and inset). In some cases the area encircled by HP1 is smaller (J, arrowhead). Scale bar: 10m. mute is required for muscle integrity 2703

To test whether levels of histone proteins were affected, histones extracted from early stage 17 embryos were analysed by western blotting. Surprisingly, we were unable to detect any differences in the overall levels of histones H3 and H4 (Fig. 5B). The modification of histone tails is known to affect higher order chromatin structure by affecting the stability and packaging of chromatin and also the transcriptional status of genes. During muscle differentiation there is large scale reorganisation of heterochromatin (Agarwal et al., 2007). Defects in establishing and maintaining heterochromatin has deleterious effects, often leading to disease (Delcuve et al., 2009; Zhang and Adams, 2007). We, therefore, tested the levels of modified histones H3 and H4. We found that di- and trimethyl-H3- (Lys)9 (di/tri-met H3K9) levels were unaffected in mute1281/Df embryos. Levels of acetyl-H3-(Lys)9 (Ac H3K9) were slightly lower but we were unable to detect any muscle-specific differences by immunofluorescence (Fig. 5B, arrow and data not shown). The histone variant H2Av has been shown to regulate the formation of heterochromatin and to also function in the silencing of euchromatic genes (Swaminathan et al., 2005). The levels of H2Av were unaffected (Fig. 5B). In order to test for other detectable changes in the heterochromatin of myonuclei, we studied the localisation of the non-histone protein, Heterochromatin protein 1 (HP1). HP1 functions as a gene repressor and localises to heterochromatic regions via interaction with methylated H3K9 and has recently been shown to localise to transcriptionally active regions as well (Ayyanathan et al., 2003). We found that in WT muscles HP1 Fig. 6. Muscle-specific genes are misregulated in mute1281embryos. appeared to surround the heterochromatin and was present as a (A)qPCR on total RNA extracted from early stage 17 WT and mute1281 ring (Fig. 5E-H, arrows and inset). Interestingly this pattern was embryos. Fold change over WT is normalised against rp49 and is indicated on reduced to a punctate distribution in many of the thin muscles in the y-axis. Horizontal dashed line indicates WT level1. Mef2, how, Mhc, Mlp60A, Mlc1 and Mlc2 are downregulated (P<0.05 for Mef2 and how; for the mute1281/Df embryos (Fig. 5I-L, arrows and inset). Some muscles rest P<0.1). CG6972 is upregulated (P<0.05). Levels of act57B are not did show the circular HP1 stain but the area encircled appeared to significantly lower (P0.2). CG11658 is upregulated (P0.17). Error bars be much smaller compared with WT (Fig. 5F,J, arrowhead). The show standard deviation of three biological replicates. (B)Western blot for 1281 Df overall levels of HP1 were unaffected in mute / embryos (Fig. muscle proteins using extracts from early stage 17 WT and mute1281/Df

Journal of Cell Science 5B). embryos. 10g of protein was loaded per well. The level of How(L) was the Thus, mute appears to regulate histone gene expression, possibly same but How(S) (arrowhead) and MHC (arrow) are reduced in by functioning as a transcriptional repressor. The higher amounts mute1281embryos. (C)Quantification of the data from the western blot in B. of normal histone transcripts in mute1281 does not affect the overall Levels are normalised against tubulin. levels of histone proteins. The localization of HP1 is affected in mute1281, which suggests that mute may play a role in the organisation of heterochromatin. It is possible that by some, as yet reduced muscle mass in the mutant (Fig. 6B, arrow; quantified in unknown, mechanism, or some other function of mute the stability Fig. 6C). There was no significant difference in the transcript of chromatin is affected in terminally differentiated muscle cells. levels of actin57B (act57B; Fig. 6A). CG6972, was significantly upregulated by approximately threefold compared with WT (Fig. Muscle genes are misregulated in mute1281 embryos 6A; P<0.1). CG6972, an uncharacterised gene, has been proposed The loss in muscle mass and change in HP1 localisation observed to be positively regulated by Mef2 and is expressed specifically in in mute1281 embryos might be concurrent with the misregulation of the somatic muscles (Elgar et al., 2008; Sandmann et al., 2006) but muscle-specific genes. In order to test this, muscle genes expressed has also been shown to be upregulated in response to hypertrophy- late in embryonic development (i.e. early stage 17) were chosen induced muscular degeneration (Montana and Littleton, 2006). It for analysis by qPCR, in both WT and mute1281/Df embryos. This is possible that the higher levels of CG6972 observed in mute1281 data is summarised in Table S1 in supplementary material. We is due to the muscle-wasting defect. found that the levels of Mef2 RNA, a key regulator of muscle gene We also examined the expression levels of the RNA binding expression was downregulated by ~20% compared with the WT protein How that is expressed in both somatic muscles and tendon level (0.79±0.08 fold; Student’s t-test P<0.05; Fig. 6A). This might cells and is required during terminal differentiation, when myotubes explain the reduction in Mef2 protein (Fig. 3C,D). This could in begin to attach to their epidermal attachment sites (Baehrecke, turn cause the misregulation of Mef2 target genes. We quantified 1997). how transcript levels were ~20% lower in mute1281/Df the levels of some Mef2 targets and found that the levels of Myosin compared with WT embryos (Fig. 6A; 0.79±0.05 fold; P<0.05). alkali light chain l (Mlc1), Myosin alkali light chain 2 (Mlc2), The how locus encodes two isoforms, the long isoform [how(L)], Muscle LIM protein 60A (Mlp60A) and Mhc were lower in the which is expressed throughout embryonic development, and the mutant (Fig. 6A; P<0.1). Concurrently, we also observed that short [how(S)], which is expressed only at later stages of levels of MHC protein were lower. This could be attributed to the embryogenesis (Nabel-Rosen et al., 1999). We further tested the 2704 Journal of Cell Science 123 (16)

protein levels of How. The levels of How(L) were comparable to Some muscles were detached (Fig. 7B, arrow) while others that observed in WT but the levels of How(S) were significantly appeared to be missing (Fig. 7B, asterisk), similar to the mute1281/Df lower in mute1281/Df embryos (Fig. 6B, arrowhead; quantified in C). phenotype. Thus, mute appears to genetically interact with Slbp to how has been shown to regulate muscle activity as well as tendon affect the somatic muscles. cell differentiation. It is known that levels of How(S) are There was no observable defect in the somatic muscles of upregulated in tendon cells during terminal differentiation (Nabel- Lsm11c02047 homozygous mutant embryos. Hence we analysed Rosen et al., 1999). It is possible that in mute1281, How(S) their somatic muscles at the third instar larval stage since they die expression in both muscles and tendon cells is affected. as non pharate pupae (Godfrey et al., 2009). There was a loss in Interestingly, we also found that levels of CG11658, the predicted muscle striations in 66.66% (n4/6) of the Lsm11c02047 homozygous fly homologue of vertebrate atrogin-1, an F-box protein and mutant larvae compared with the regularly spaced striations in WT proposed ubiquitinylation enzyme, were ~30% higher (Fig. 6A, larval muscles (Fig. 7C,D), although the number of muscles affected P0.17). cDNA microarrays have previously shown that atrogin-1 and the extent of the area affected within a muscle were highly levels are upregulated specifically in skeletal muscles when they variable. undergo atrophy (Gomes et al., 2001). CG11658 is 26% identical Thus, other processing factors and components of the HLB to mouse atrogin-1 which in turn is 96% identical to human appear to regulate the terminal differentiation and integrity of Atrogin-1 (Gomes et al., 2001). myotubes to various extents. The observed lethality in these mutants These results show that the expression of a subset of important might be due to defects in tissues that have not yet been analysed. muscle-specific genes are affected in mute1281. We speculate that even minor changes in gene expression might have cumulative Discussion effects on muscle differentiation. Misregulation of these and other Through a genetic screen for regulators of muscle development we untested genes, might be due to changes in chromatin organisation. have identified mute, the function of which appears to be required for the maintenance of muscle mass and integrity. We have Other processing mutants also display defective muscles attempted to characterize the muscle phenotype when the mute Mutations in Slbp, U7 snRNA, Lsm10 and Lsm11 have been shown locus is disrupted. Although early myogenic events are unaffected, to affect the cell cycle and oogenesis in Drosophila (Godfrey et al., terminal differentiation of muscle and/or tendon cell is affected. 2006; Godfrey et al., 2009; Sullivan et al., 2001). Thus far, none Terminal differentiation of muscles involves the expression of of these genes have been investigated for a possible role in muscle structural proteins and the attachment of muscles to tendon cells, morphology. We analysed the somatic muscles of Slbp15 mutant through the coordinated action of several embryonic genes, embryos at late stage 16 and Lsm11c02047 mutant third instar larvae. including Mef2, resulting in a functional contractile larval muscle. Homozygous Slbp15 embryos do not display any defects in their Reciprocal signalling between the mesoderm and the overlying musculature (Fig. 7A), but interestingly, in the absence of one ectoderm is essential, as muscle growth cones seek out their proper copy of mute (mute1281/+, Slbp15/15) over half of this genotype attachment sites (Bate, 1992). This leads to tendon cell (n9/15) had somatic muscles that appeared significantly thinner differentiation and successful myotube attachment (Becker et al., than those of WT (compare Fig. 7A,B, dorsal muscles outlined). 1997). Vein protein secreted by the muscles upregulates the tendon

Journal of Cell Science cell differentiation factor How(S), through the EGF signalling pathway (Nabel-Rosen et al., 1999). We have shown that levels of How(S) is significantly reduced in mute1281cells. It is possible that How is affected in both muscle and tendon cells or the muscles do not secrete sufficient Vein. This could have drastic consequences, as has been shown in embryonic lethal how mutants that have defects in muscle migration and attachment (Baehrecke, 1997). Also, Kettin and D-Titin are significantly reduced at attachment sites, signifying defects in muscle and/or tendon cells in mute1281 embryos. Kettin and D-Titin are essential in organising myofilaments into highly ordered arrays and they also interact with several factors at the attachment sites. D-titin mutants lack visible striations, the muscles are thin and F-actin is disorganised (Zhang et al., 2000). Earlier defects in mute1281 embryos could not be investigated because of a possible maternal contribution of mute and the absence of egg laying in germline clones. It is possible that mute also functions in aspects of oogenesis similar to what has Fig. 7. Other HLB mutants display muscle defects. (A,B)Stage 16 embryos been shown for other components of the histone pre-mRNA labelled with anti-MHC (red). Homozygous Slbp15 embryos do not display any processing machinery, Slbp and U7 (Godfrey et al., 2006; Lanzotti muscle defects at this stage (A) but additional removal of one copy of mute et al., 2002; Sullivan et al., 2001). 1281/+ 15 15 leads to a muscle defect in 9 out of 15 (mute ;Slbp /Slbp ) embryos. The mechanism of muscle degeneration in mute1281 mutants Some muscles are thinner than normal (outlined) whereas others are detached is unknown. It has been suggested that the gross loss of muscle (arrow). Gaps between subsets are wide and muscles appear to be missing (asterisk). (C,D)Third instar larvae labelled with anti-MHC (red). WT larval mass during vertebrate skeletal muscle atrophy is largely muscles exhibit regularly spaced striations (C) but Lsm11c02047 homozygous mediated by the ubiquitin-proteasome system or autophagy- mutant muscles exhibit fused striations in 4 out of 6 larvae observed (D). lysosome pathway (Schwartz, 2008). atrogin-1, a ubiquitin ligase Numbers of muscles, and regions within a particular affected muscle, vary has been shown to be upregulated in several models of skeletal greatly. Scale bars: 10m. muscle atrophy (Glass, 2003; Gomes et al., 2001). CG11658 has mute is required for muscle integrity 2705

been proposed by sequence homology to be the Drosophila stabilise transcriptional repression during differentiation (Agarwal Atrogin-1 homologue (Gomes et al., 2001). It is interesting that et al., 2007). we observe an upregulation of CG11658 transcripts in mute1281 Several genetic studies have highlighted the regulatory role of embryos and it is thus tempting to speculate that the ubiquitin Mef2 throughout muscle development. Subtle changes in Mef2 proteasome pathway might be activated in the atrophying muscles transcripts result in stringent regulation of muscle genes (Elgar et of mute1281. al., 2008; Sandmann et al., 2006). We have shown that Mef2 We have shown that Mute colocalises with Lsm10 and Lsm11, transcript and protein is downregulated in mute1281embryos. Mute two components of the Drosophila HLB, but does not appear to be might regulate the expression of genes such as Mef2 through its required for their localization. The presence of misprocessed histone action on heterochromatin, perhaps on Mef2 regulatory regions. It transcripts in mute mutants is unlikely to be the cause of the late is possible that a large array of untested muscle and tendon genes embryonic lethality and muscle degeneration since Slbp15 and U720 are misregulated. We speculate that even minor changes in the mutants have been shown to have an approximate five- to six-fold levels of these transcripts could lead to sufficient misregulation of higher level of misprocessed histone transcripts compared with muscle and tendon proteins (e.g. How), resulting in cumulative mute1281, yet Slbp15 mutants show visible muscle defects only deleterious effects on processes of terminal muscle differentiation when in combination with mute1281 and U7 mutants are viable in the embryo, as we have shown here in the case of muscle though sterile (Godfrey et al., 2006). Interestingly, we find an attachment. overall higher level of normal histone transcripts in mute mutants, To our knowledge, this is the first report implicating a histone an observation that has not been reported for the other characterised gene regulatory factor in the maintenance of muscle integrity. histone-processing mutants such as Slbp, U7, Lsm10 and Lsm11. Perhaps the identification of mute can provide some insight into Mute might function in a manner similar to the abnormal oocyte the role of the HLB in differentiated tissue. It could also be used (abo) gene product, which localises to the histone gene cluster and to investigate the regulatory machinery of histone expression by is a negative regulator of histone gene expression (Berloco et al., studying mute-interacting molecules. It may well be that mute has 2001). These could lead to tissue-specific changes in histone protein other gene regulatory functions possibly through the organisation levels, which are difficult to detect. Given the proximity of the of heterochromatin, that might also contribute to the severe muscle HLB to the histone gene cluster it is possible that transcriptional degeneration. Interestingly we also find a Mute-positive nuclear and post-transcriptional activities are closely associated and that body in HeLa cells that colocalises with FLASH, suggesting the HLB serves as an integrated machinery for histone gene evolutionary conservation. Mutations in survival motor neuron expression. These and possibly other undiscovered functions of (smn), a CB marker, and a factor implicated in snRNP biogenesis mute could be responsible for the observed embryonic lethality (Fischer et al., 1997; Liu et al., 1997) leads to spinal muscular and muscle defects. atrophy in humans. It is possible that factors involved in the Mute is predicted to have PAH repeats, a SANT-DNA-binding assembly of processing factors or more directly in histone domain and an Atrophin-1-like region which have been implicated transcription, might have functions in tissues such as muscle that in the association with, and the function of, chromatin remodelling undergo extensive chromatin reorganisation because of cell fusion. complexes and the control of gene expression (Boyer et al., 2002;

Journal of Cell Science Boyer et al., 2004; Heinzel et al., 1997; Nagy et al., 1997), although Materials and Methods we have shown that the predicted SANT-DNA-binding domain is Plasmids and cloning pUAST-Mute-S was generated from the cDNA clone RE27864 (Drosophila Genomic not required for maintenance of muscle integrity. The modification Resource Centre). pUAST-Mute-L was generated by ligating a 2.7 kb fragment from of histone tails has been implicated in muscle differentiation RE27864 to a 3.7 kb genomic DNA fragment. pUAST-Mute-LC and pUAST- (McKinsey et al., 2002; Yahi et al., 2006) and deacetylation of MuteC-Atro were generated by introducing a stop codon at position 5038 and 2356 histones has been implicated in gene silencing (Park et al., 1999). of the coding sequence, respectively. 1281 The defects in mute mutants are observed during the late stages Drosophila strains of development when most cells, including myoblasts, have exited All crosses were maintained at 25°C. Stocks used were: yw, MHC-tau-GFP (E. the cell cycle. Muscles form a unique tissue in that they are formed Chen and E. Olson), P{GawB}NP1160 (Bloomington), Df H99 (Bloomington), Kettin-GFP (Mar Ruiz-Gomez), U720, Slbp15, Lsm11c02047 (Godfrey et al., 2006; by the fusion of differentiated myoblasts. Upon fusion, FCM nuclei Godfrey et al., 2009; Sullivan et al., 2001), rp298-lacZ (A. Nose), rp289-Gal4 and are entrained by the FC nuclei to express FC specific genes (Baylies 24B Gal4 (Zaffran et al., 1997), Df(2R)Exel 6065 (Bloomington). Homozygous et al., 1998; Beckett and Baylies, 2006). Hence nuclear re- mutants were identified by the absence of -galactosidase or GFP. Transgenic flies were generated as described previously (Spradling, 1986). Rescues were performed programming is critical in the formation of the mature using single copies of the UAS transgene and 24B-Gal4 or rp298-Gal4 in mute1281/Df multinucleated muscle. Differentiation and fusion of C2C12 mutant embryos at 25°C. Results are representative of two independent insertions myocytes is accompanied by remarkable reorganisation of for each transgene.

constitutive heterochromatin domains (Shen et al., 2003; Terranova Genetic screen et al., 2005). Non-histone proteins such as HP1 also function in Wild-type males were mutagenised with EMS. ~4000 stocks carrying the mutagenised cell differentiation by associating with transcriptional repressors chromosome over a Cy-Actin-GFP balancer were established. F2 embryos of ~2000 thus providing a mechanism for widespread silencing of gene embryonic lethal stocks were analysed for muscle defects using anti-MHC. expression (Lechner et al., 2000; McKinsey et al., 2000). We have Antibody generation shown that the presentation of HP1 is aberrant in the thinner Regions between aa 3-164 and 1093-1291 were used to generate anti-Mute-L and muscles of mute1281embryos. Most studies have focused on the anti-Mute-LS antibodies respectively. His-tagged antigens were bacterially produced and antibodies raised in guinea pigs and rabbits. role of HP1 in the organisation of heterochromatin, but more recently it has been implicated in the active transcription of Embryo fixation and immunostaining euchromatic cell cycle genes (De Lucia et al., 2005). Interestingly For anti-Mute antibodies, embryos were fixed for 45 minutes and manually devitellinated. For the rest embryos were fixed using standard procedures. Primary it has also been shown that the association of HP1 gamma increases antibodies used: guinea pig anti-Mute-LS, rabbit anti-Mute-L, mouse anti-MHC as vertebrate myoblasts differentiate to form myotubes, possibly to (D. Kiehart), rabbit anti-Eve (M. Frasch), rabbit anti--galactosidase (Cappel), 2706 Journal of Cell Science 123 (16)

rabbit anti-caspase 3 (Cell Signaling Technology) rabbit anti-Mef2 (Bruce Supplementary material available online at Paterson), guinea pig anti-D-Titin, chicken anti-GFP (Abcam), mouse anti-Lamin http://jcs.biologists.org/cgi/content/full/123/16/2697/DC1 (DSHB), rabbit anti-Lsm10 and anti-Lsm11 (J. Gall), mouse anti-HP1 (DSHB), mouse anti-Dlg (DSHB) and Hoechst 3342 (Invitrogen). Secondary antibodies References were conjugated to Cy3 (Jackson ImmunoResearch Laboratories) or Alexa Fluor Abmayr, S. M., Balagopalan, L., Galletta, B. J. and Hong, S. J. (2003). Cell and 488 (Molecular Probes). Samples were analysed under a Zeiss LSM 5 Exciter molecular biology of myoblast fusion. Int. Rev. Cytol. 225, 33-89. confocal microscope. Agarwal, N., Hardt, T., Brero, A., Nowak, D., Rothbauer, U., Becker, A., Leonhardt, H. and Cardoso, M. C. (2007). MeCP2 interacts with HP1 and modulates its Immunostaining of HeLa cells heterochromatin association during myogenic differentiation. Nucleic Acids Res. 35, HeLa cells were seeded onto polylysine-coated coverslips (Iwaki) and fixed in 3% 5402-5408. paraformaldehyde. Primary antibodies used were: guinea pig anti-Mute-LS, rabbit Ayyanathan, K., Lechner, M. S., Bell, P., Maul, G. G., Schultz, D. C., Yamada, Y., anti-FLASH (Santa Cruz Biotehcnology, Inc.). Secondary antibodies and imaging as Tanaka, K., Torigoe, K. and Rauscher, F. J., 3rd (2003). Regulated recruitment of above. HP1 to a euchromatic gene induces mitotically heritable, epigenetic gene silencing: a mammalian cell culture model of gene variegation. Genes Dev. 17, 1855-1869. Northern blotting Azzouz, T. N. and Schumperli, D. (2003). Evolutionary conservation of the U7 small Total RNA was run on a formaldehyde-agarose gel. Northern blotting was performed nuclear ribonucleoprotein in Drosophila melanogaster. RNA 9, 1532-1541. as per the Roche-DIG protocol. Probe was identical to the anti-Mute-LS antigenic Baehrecke, E. H. (1997). who encodes a KH RNA binding protein that functions in region, H3 and H4 coding regions. muscle development. Development 124, 1323-1332. Barcaroli, D., Bongiorno-Borbone, L., Terrinoni, A., Hofmann, T. G., Rossi, M., Knight, R. A., Matera, A. G., Melino, G. and De Laurenzi, V. (2006). FLASH is Embryo collections 1281/Df required for histone transcription and S-phase progression. Proc. Natl. Acad. Sci. USA Two-hour embryo collections of yw and mute were aged together at 25°C up 103, 14808-14812. to early embryonic stage 17. Homozygous mutant embryos were selected on the Bate, M. (1990). The embryonic development of larval muscles in Drosophila. Development basis of the absence of Actin-GFP. 200 embryos of each genotype were used for acid 110, 791-804. extraction of histones, pPCR and total protein extraction. Bate, M. (1992). Mechanisms of muscle patterning in Drosophila. Semin. Dev. Biol. 3, 267-275. Acid extraction of histones Bate, M. (1993). The mesoderm and its derivatives. In The Development of Drosophila Embryos were homogenised in lysis buffer (0.25 M sucrose, 60 mM KCl, 15 mM melanogaster (ed. M. Bate and A. Martinez-Arias). New York: Cold Spring Harbor NaCl, 5 mM MgCl2, 1 mM CaCl2, 15 mM Pipes pH 6.8, 0.8% Triton X-100, 0.1 Laboratory Press. mM Pefabloc and protease inhibitor cocktail (Roche). Homogenate was spun at Baxter, J., Sauer, S., Peters, A., John, R., Williams, R., Caparros, M. L., Arney, K., 10,000 g for 20 minutes. Samples were incubated for 1 hour in 0.4 M H2SO4 and Otte, A., Jenuwein, T., Merkenschlager, M. et al. (2004). Histone hypomethylation spun at 12,000 g for 5 minutes. Proteins were precipitated in acetone at –20°C. is an indicator of epigenetic plasticity in quiescent lymphocytes. EMBO J. 23, 4462- Samples were spun at 10,000 g for 15 minutes. Pellet was resuspended in 4 M urea. 4472. Total protein was quantified. 5 g protein was loaded per well. Baylies, M. K. and Michelson, A. M. (2001). Invertebrate myogenesis: looking back to the future of muscle development. Curr. Opin. Genet. Dev. 11, 431-439. Western blotting Baylies, M. K., Bate, M. and Ruiz Gomez, M. (1998). Myogenesis: a view from Samples were run on an SDS-PAGE gel and transferred onto a PVDF membrane Drosophila. Cell 93, 921-927. Becker, S., Pasca, G., Strumpf, D., Min, L. and Volk, T. (1997). Reciprocal signaling (Immobilon-PSQ, Millipore). Primary antibodies used were: mouse anti-MHC, rabbit between Drosophila epidermal muscle attachment cells and their corresponding muscles. anti-How (T. Volk), mouse anti-Tubulin (Sigma), rabbit anti-H3 (Santa Cruz Development 124, 2615-2622. Biotechnology), rabbit anti-H4 (Abcam), rabbit anti-trimet H3K9 (Abcam), rabbit Beckett, K. and Baylies, M. K. (2006). The development of the Drosophila larval body anti-dimet H3K9 and rabbit anti-Ac H3K9 (Millipore), mouse anti-HP1 (DSHB), wall muscles. Int. Rev. Neurobiol. 75, 55-70. rabbit anti-H2aV (Robert Glaser). Secondary antibodies were anti-mouse or anti- Berloco, M., Fanti, L., Breiling, A., Orlando, V. and Pimpinelli, S. (2001). The maternal rabbit-HRP (Roche). Proteins were detected using Luminol, coumaric acid (Sigma) effect gene, abnormal oocyte (abo), of Drosophila melanogaster encodes a specific and Amersham HyperfilmECL. negative regulator of histones. Proc. Natl. Acad. Sci. USA 98, 12126-12131. Bongiorno-Borbone, L., De Cola, A., Vernole, P., Finos, L., Barcaroli, D., Knight, R.

Journal of Cell Science Quantitative RT-PCR A., Melino, G. and De Laurenzi, V. (2008). FLASH and NPAT positive but not Coilin Total RNA was extracted using the Trizol method. Genomic DNA was removed with positive Cajal Bodies correlate with cell ploidy. Cell Cycle 7, 2357-2367. DNase I. Equal amounts of RNA were used for cDNA synthesis. First strand cDNA Boyer, L. A., Langer, M. R., Crowley, K. A., Tan, S., Denu, J. M. and Peterson, C. L. was synthesized using oligo(dT)- or H3/H4-specific primers and SuperScript RT-III (2002). Essential role for the SANT domain in the functioning of multiple chromatin (Invitrogen). Samples were treated with RNase-H and used for qPCR. SYBR Green remodeling enzymes. Mol. Cell 10, 935-942. PCR master mix, qPCR machine and software from Applied Biosystems was used. Boyer, L. A., Latek, R. R. and Peterson, C. L. (2004). The SANT domain: a unique MS Exel was used for statistical analysis. histone-tail-binding module? Nat. Rev. Mol. Cell Biol. 5, 158-163. Broadie, K. S. and Bate, M. (1993). Development of larval muscle properties in the Primers were designed using Oligotech software: rp49-F, 5Ј-GCTAAG - embryonic myotubes of Drosophila melanogaster. J. Neurosci. 13, 167-180. CTGTCGCACAAA-3Ј; rp49-R, 5Ј-TCCGGTGGGCAGCATGTG-3Ј; H3-F1, 5Ј- Ј Ј Carrasco-Rando, M. and Ruiz-Gomez, M. (2008). Mind bomb 2, a founder myoblast- ACCGAGCTTCTAATCCGCAAG-3 ; H3-R1, 5 -ACCAACCAGGTAGGCTTCGC- specific protein, regulates myoblast fusion and muscle stability. Development 135, 849- Ј Ј Ј Ј 3 ; H4-F2, 5 -TGGCGTTCTGAAGGTTTTCTTG-3 ; H4-R2, 5 -AACCGC- 857. CAAATCCGTAGAGG-3Ј; Mef2-F, 5Ј-ACAACGAGCCCCACGAGTCC-3Ј; Delcuve, G. P., Rastegar, M. and Davie, J. R. (2009). Epigenetic control. J. Cell Physiol. Mef2-R, 5Ј-GAGTGAGTGTGTAGTCCGTTTC-3Ј; How-F1, 5Ј-ATGGCACT- 219, 243-250. TATAGGGACACAAC-3Ј; How-R1, 5Ј-TGGATGTCAGCAGGCGGCTA-3Ј; De Lucia, F., Ni, J. Q., Vaillant, C. and Sun, F. L. (2005). HP1 modulates the transcription CG6972-F1, 5Ј-AAGGAAGTACCACGCGAACCG-3Ј; CG6972-R1, 5Ј-TCCTC - of cell-cycle regulators in Drosophila melanogaster. Nucleic Acids Res. 33, 2852-2858. CTTGTAGATCTTGCCATC-3Ј; MHC-F2, 5Ј-AGATCGAGGAGGCTG AG GAA - Dominski, Z., Yang, X. C., Purdy, M. and Marzluff, W. F. (2003). Cloning and ATC-3Ј; MHC-R2, 5Ј-ACAGAACCGGCACGTCCCTTG-3Ј; Act57B-F1, 5Ј- characterization of the Drosophila U7 small nuclear RNA. Proc. Natl. Acad. Sci. USA TCCACGAGACCGTCTACAACTC-3Ј; Act57B-F1, 5Ј-ATGGGGCCAGGGAG- 100, 9422-9427. GTGATC-3Ј; Mlp60A-F1, 5Ј- AGGCTACAAATTCCACAAGACC-3Ј; Mlp60A-R1, Dominski, Z., Yang, X. C. and Marzluff, W. F. (2005). The polyadenylation factor 5Ј-GGGACCGTATTTGCGACCATG-3Ј; Mlc1-F1, 5Ј-AAGCTCTACGACAAG- CPSF-73 is involved in histone-pre-mRNA processing. Cell 123, 37-48. GAGGAG-3Ј, Mlc1-R1, 5Ј-ATCCTCGGGATCCATGCAGTC-3Ј; Mlc2-F1, 5Ј- Elgar, S. J., Han, J. and Taylor, M. V. (2008). mef2 activity levels differentially affect gene expression during Drosophila muscle development. Proc. Natl. Acad. Sci. USA AGAAGCAGATCGCCGAGTTCA-3Ј; Mlc2-R1, 5Ј-AGCATGGCGTCCAACTCC - 105, 918-923. TTGT-3Ј, CG11658-F1, 5Ј-ACCATCAAGCGCATCCAGC-3Ј; CG11658-R1, 5Ј- Ј Fischer, U., Liu, Q. and Dreyfuss, G. (1997). The SMN-SIP1 complex has an essential ATGGTCGGCTATGCACAG-3 . role in spliceosomal snRNP biogenesis. Cell 90, 1023-1029. Galli, G., Hofstetter, H., Stunnenberg, H. G. and Birnstiel, M. L. (1983). Biochemical We are grateful to Mar Ruiz-Gomez, Talila Volk, Bruce Patterson, complementation with RNA in the Xenopus oocyte: a small RNA is required for the Hanh Nguyen, Elizabeth Chen, Eric Olson, Akinao Nose, Daniel generation of 3Ј histone mRNA termini. Cell 34, 823-828. Glass, D. J. (2003). Signalling pathways that mediate skeletal muscle hypertrophy and Keihart, Manfred Frasch, Robert Glaser, DSHB and Bloomington atrophy. Nat. Cell Biol. 5, 87-90. Stock Center for antibodies and fly stocks. We thank Robert Duronio Godfrey, A. C., Kupsco, J. M., Burch, B. D., Zimmerman, R. M., Dominski, Z., and Joseph Gall for fly stocks, antibodies and fruitful discussions. We Marzluff, W. F. and Duronio, R. J. (2006). U7 snRNA mutations in Drosophila block also thank Karuna Sampath and Jishy Varghese for critical comments histone pre-mRNA processing and disrupt oogenesis. RNA 12, 396-409. Godfrey, A. C., White, A. E., Tatomer, D. C., Marzluff, W. F. and Duronio, R. J. on the manuscript. This work was supported by funding from Temasek (2009). The Drosophila U7 snRNP proteins Lsm10 and Lsm11 are required for histone Life Sciences Laboratory and the Singapore Millennium Foundation. pre-mRNA processing and play an essential role in development. RNA 15, 1661-1672. mute is required for muscle integrity 2707

Gomes, M. D., Lecker, S. H., Jagoe, R. T., Navon, A. and Goldberg, A. L. (2001). Nagy, L., Kao, H. Y., Chakravarti, D., Lin, R. J., Hassig, C. A., Ayer, D. E., Schreiber, Atrogin-1, a muscle-specific F-box protein highly expressed during muscle atrophy. S. L. and Evans, R. M. (1997). Nuclear receptor repression mediated by a complex Proc. Natl. Acad. Sci. USA 98, 14440-14445. containing SMRT, mSin3A, and histone deacetylase. Cell 89, 373-380. Haaf, T., Steinlein, C. and Schmid, M. (1990). Nucleolar transcriptional activity in Nguyen, H. T., Voza, F., Ezzeddine, N. and Frasch, M. (2007). Drosophila mind bomb2 mouse Sertoli cells is dependent on centromere arrangement. Exp. Cell Res. 191, 157- is required for maintaining muscle integrity and survival. J. Cell Biol. 179, 219-227. 160. Park, E. J., Schroen, D. J., Yang, M., Li, H., Li, L. and Chen, J. D. (1999). SMRTe, a Harris, M. E., Bohni, R., Schneiderman, M. H., Ramamurthy, L., Schumperli, D. and silencing mediator for retinoid and thyroid hormone receptors-extended isoform that is Marzluff, W. F. (1991). Regulation of histone mRNA in the unperturbed cell cycle: more related to the nuclear receptor corepressor. Proc. Natl. Acad. Sci. USA 96, 3519- evidence suggesting control at two posttranscriptional steps. Mol. Cell Biol. 11, 2416- 3524. 2424. Quina, A. S., Buschbeck, M. and Di Croce, L. (2006). Chromatin structure and Heinzel, T., Lavinsky, R. M., Mullen, T. M., Soderstrom, M., Laherty, C. D., Torchia, epigenetics. Biochem. Pharmacol. 72, 1563-1569. J., Yang, W. M., Brard, G., Ngo, S. D., Davie, J. R. et al. (1997). A complex Sandmann, T., Jensen, L. J., Jakobsen, J. S., Karzynski, M. M., Eichenlaub, M. P., containing N-CoR, mSin3 and histone deacetylase mediates transcriptional repression. Bork, P. and Furlong, E. E. (2006). A temporal map of transcription factor activity: Nature 387, 43-48. mef2 directly regulates target genes at all stages of muscle development. Dev. Cell 10, Hueber, S. D. and Lohmann, I. (2008). Shaping segments: Hox gene function in the 797-807. genomic age. BioEssays 30, 965-979. Schwartz, L. M. (2008). Atrophy and programmed cell death of skeletal muscle. Cell Kolev, N. G. and Steitz, J. A. (2005). Symplekin and multiple other polyadenylation Death Differ. 15, 1163-1169. factors participate in 3Ј-end maturation of histone mRNAs. Genes Dev. 19, 2583-2592. Shen, X., Collier, J. M., Hlaing, M., Zhang, L., Delshad, E. H., Bristow, J. and Lanzotti, D. J., Kaygun, H., Yang, X., Duronio, R. J. and Marzluff, W. F. (2002). Bernstein, H. S. (2003). Genome-wide examination of myoblast cell cycle withdrawal Developmental control of histone mRNA and dSLBP synthesis during Drosophila during differentiation. Dev. Dyn. 226, 128-138. embryogenesis and the role of dSLBP in histone mRNA 3Ј end processing in vivo. Mol. Spradling, A. C. (1986). P element-mediated transformation. In Drosophila: A Practical Cell. Biol. 22, 2267-2282. Approach (ed. D. B. Roberts), pp. 175-197. Oxford: IRL Press. Lechner, M. S., Begg, G. E., Speicher, D. W. and Rauscher, F. J., 3rd. (2000). Molecular Stauber, C. and Schumperli, D. (1988). 3Ј processing of pre-mRNA plays a major role determinants for targeting heterochromatin protein 1-mediated gene silencing: direct in proliferation-dependent regulation of histone gene expression. Nucleic Acids Res. 16, chromoshadow domain-KAP-1 corepressor interaction is essential. Mol. Cell. Biol. 20, 9399-9414. 6449-6465. Sullivan, E., Santiago, C., Parker, E. D., Dominski, Z., Yang, X., Lanzotti, D. J., Lim, C. H., Chong, S. W. and Jiang, Y. J. (2009). Udu deficiency activates DNA damage Ingledue, T. C., Marzluff, W. F. and Duronio, R. J. (2001). Drosophila stem loop checkpoint. Mol. Biol. Cell 20, 4183-4193. binding protein coordinates accumulation of mature histone mRNA with cell cycle Liu, J. L., Murphy, C., Buszczak, M., Clatterbuck, S., Goodman, R. and Gall, J. G. progression. Genes Dev. 15, 173-187. (2006). The Drosophila melanogaster Cajal body. J. Cell Biol. 172, 875-884. Swaminathan, J., Baxter, E. M. and Corces, V. G. (2005). The role of histone H2Av Liu, Q., Fischer, U., Wang, F. and Dreyfuss, G. (1997). The spinal muscular atrophy variant replacement and histone H4 acetylation in the establishment of Drosophila disease gene product, SMN, and its associated protein SIP1 are in a complex with heterochromatin. Genes Dev. 19, 65-76. spliceosomal snRNP proteins. Cell 90, 1013-1021. Taylor, M. V. (1998). Muscle development: a transcriptional pathway in myogenesis. Curr. Liu, Y., Du, L., Osato, M., Teo, E. H., Qian, F., Jin, H., Zhen, F., Xu, J., Guo, L., Biol. 8, R356-R358. Huang, H. et al. (2007). The zebrafish udu gene encodes a novel nuclear factor and is Terranova, R., Sauer, S., Merkenschlager, M. and Fisher, A. G. (2005). The essential for primitive erythroid cell development. Blood 110, 99-106. reorganisation of constitutive heterochromatin in differentiating muscle requires HDAC Martin, F., Schaller, A., Eglite, S., Schumperli, D. and Muller, B. (1997). The gene for activity. Exp. Cell Res. 310, 344-356. histone RNA hairpin binding protein is located on human chromosome 4 and encodes Wagner, E. J., Burch, B. D., Godfrey, A. C., Salzler, H. R., Duronio, R. J. and a novel type of RNA binding protein. EMBO J. 16, 769-778. Marzluff, W. F. (2007). A genome-wide RNA interference screen reveals that variant Marzluff, W. F. (1992). Histone 3Ј ends: essential and regulatory functions. Gene Expr. histones are necessary for replication-dependent histone pre-mRNA processing. Mol. 2, 93-97. Cell 28, 692-699. McKinsey, T. A., Zhang, C. L. and Olson, E. N. (2000). Activation of the myocyte Wang, H., Clark, I., Nicholson, P. R., Herskowitz, I. and Stillman, D. J. (1990). The enhancer factor-2 transcription factor by calcium/calmodulin-dependent protein kinase- Saccharomyces cerevisiae SIN3 gene, a negative regulator of HO, contains four paired stimulated binding of 14-3-3 to histone deacetylase 5. Proc. Natl. Acad. Sci. USA 97, amphipathic helix motifs. Mol. Cell. Biol. 10, 5927-5936. 14400-14405. Wang, P., Palfi, Z., Preusser, C., Lucke, S., Lane, W. S., Kambach, C. and Bindereif, McKinsey, T. A., Zhang, C. L. and Olson, E. N. (2002). Signaling chromatin to make A. (2006). Sm core variation in spliceosomal small nuclear ribonucleoproteins from muscle. Curr. Opin. Cell Biol. 14, 763-772. Trypanosoma brucei. EMBO J. 25, 4513-4523.

Journal of Cell Science Mitra, P., Xie, R. L., Medina, R., Hovhannisyan, H., Zaidi, S. K., Wei, Y., Harper, J. Yahi, H., Philipot, O., Guasconi, V., Fritsch, L. and Ait-Si-Ali, S. (2006). Chromatin W., Stein, J. L., van Wijnen, A. J. and Stein, G. S. (2003). Identification of HiNF-P, modification and muscle differentiation. Expert Opin. Ther. Targets 10, 923-934. a key activator of cell cycle-controlled histone H4 genes at the onset of S phase. Mol. Zaffran, S., Astier, M., Gratecos, D. and Semeriva, M. (1997). The held out wings Cell. Biol. 23, 8110-8123. (how) Drosophila gene encodes a putative RNA-binding protein involved in the control Montana, E. S. and Littleton, J. T. (2006). Expression profiling of a hypercontraction- of muscular and cardiac activity. Development 124, 2087-2098. induced myopathy in Drosophila suggests a compensatory cytoskeletal remodeling Zhang, R. and Adams, P. D. (2007). Heterochromatin and its relationship to cell response. J. Biol. Chem. 281, 8100-8109. senescence and cancer therapy. Cell Cycle 6, 784-789. Mowry, K. L. and Steitz, J. A. (1987). Identification of the human U7 snRNP as one of Zhang, Y., Featherstone, D., Davis, W., Rushton, E. and Broadie, K. (2000). Drosophila several factors involved in the 3Ј end maturation of histone premessenger RNA’s. D-titin is required for myoblast fusion and skeletal muscle striation. J. Cell Sci. 113, Science 238, 1682-1687. 3103-3115. Nabel-Rosen, H., Dorevitch, N., Reuveny, A. and Volk, T. (1999). The balance between Zhao, J., Kennedy, B. K., Lawrence, B. D., Barbie, D. A., Matera, A. G., Fletcher, J. two isoforms of the Drosophila RNA-binding protein how controls tendon cell A. and Harlow, E. (2000). NPAT links cyclin E-Cdk2 to the regulation of replication- differentiation. Mol. Cell 4, 573-584. dependent histone gene transcription. Genes Dev. 14, 2283-2297.