The bromodomain-containing protein tBRD-1 is specifically expressed in spermatocytes and is essential for male fertility

Cell type specific transcription programs are a major feature of metazoans prerequisite for the specialisation of different cell types during development and tissue maintenance. In addition, translational control programs play important roles in events like oocyte maturation, very early embryogenesis, and spermatogenesis, where a series of temporally controlled events must take place in the absence of transcription. In Drosophila spermatogenesis, most transcription ceases with entry into meiotic divisions. Therefore, many genes encoding proteins required for spermatid differentiation are transcribed in primary spermatocytes but translationally repressed until the appropriate time later in gamete development (reviewed by Fuller, 1993; Renkawitz-Pohl et al., 2005; White-Cooper, 2009). In primary spermatocytes, more than 2000 testis-specific or enriched transcripts are synthesised (Doggett et al., 2011; reviewed by White-Cooper, 2010). Gene transcription in primary spermatocytes depends on a group of genes collectively named “meiotic arrest” loci (Ayyar et al., 2003; Doggett et al., 2011; Jiang et al., 2007; Jiang and White-Cooper, 2003; Lin et al., 1996; Perezgasga et al., 2004; Wang and Mann, 2003; White-Cooper et al., 2000; White-Cooper et al., 1998). Two types of meiotic arrest genes are described: the aly-class and the can-class (White-Cooper et al., 1998). The aly-class genes (aly, comr, tomb, topi and achi/vis) are required for expression of a broader range of target genes than the can-class genes (can, mia, nht, rye and sa). Proteins of the aly-class together with other proteins form the testis meiotic arrest complex (tMAC) (Beall et al., 2007). The can-class proteins are homologs of TATA box binding protein-associated factors (TAFs) and are expressed only in testis: Cannonball (Can; dTAF5 homolog), No hitter (Nht; dTAF4 homolog), Meiosis I arrest (Mia; dTAF6 homolog), Spermatocyte arrest (Sa; dTAF8 homolog) and Ryan express (Rye; dTAF12 homolog) (Hiller et al., 2001; Hiller et al., 2004). In males mutant for any of these tTAF genes transcription of several spermatid differentiation relevant genes, such as Mst87F, dj or dj like, and fzo, is greatly reduced (Hempel et al., 2006; Hiller et al., 2004; White-Cooper et al., 1998). In addition, tTAFs are required for cell cycle progression and mutant males show a meiotic arrest phenotype (reviewed by White-Cooper, 2010). Previously, it was shown that tTAFs are concentrated in a Fibrillarin-deficient sub-compartment within the spermatocyte nucleolus, along with components of the Polycomb Repression Complex 1 (PRC1), such as Polycomb (Pc) (Chen et al., 2005). Localisation of Pc to the spermatocyte nucleolus depended on tTAF activity and it has been proposed that tTAFs promote displacement of PRC1 from promoters of tTAF target genes to allow robust transcription (Chen et al., 2005). Beside tTAFs also a TAF1 isoform, presumably TAF1-2, localises to the nucleolus in primary spermatocytes in a tTAF dependent manner (Metcalf and Wassarman, 2007). Thus far, it is not known how tTAFs are recruited to their target genes and also hints for tTAF effector molecules are missing.

The activation of transcription in eukaryotes requires modifications to open the chromatin-packaged DNA (Jenuwein and Allis, 2001). Covalent modifications of N-terminal histone tails, such as acetylation, methylation, phosphorylation and ubiquitination can control patterns of gene expression (Strahl and Allis, 2000). Acetylation of histone tails is connected to gene activation and these residues can be recognised by bromodomain containing proteins (Dhalluin et al., 1999; Jacobson et al., 2000; Owen et al., 2000). Bromodomains were first discovered in the Drosophila protein Brahma which is required for activation of many homeotic genes (Kennison and Tamkun, 1988; Tamkun et al., 1992). The sequence of bromodomains is highly conserved between yeast, Drosophila and humans (Haynes et al., 1992).

Here, we identify a novel bromodomain-containing protein tBRD-1, expressed specifically in primary spermatocytes. tBRD-1 partially colocalises with tTAFs, TAF1 and Polycomb to a Fibrillarin-deficient region within the nucleolus. Nucleolar localisation of tBRD-1 depends on tTAF function as well as on the acetylation status of primary spermatocytes. In addition, ectopically expressed tBRD-1 is able to bind euchromatic interbands on polytene chromosomes. Function of tBRD-1 is required for proper differentiation of spermatids: tbrd-1 mutants are male sterile, although tBRD-1 function is not required for progression into the meiotic divisions or for transcription of the three thus far known direct tTAF target genes. Here, with tBRD-1 we propose for the first time a promising candidate who could act as a cofactor and/or effector of tTAFs.

The Drosophila CG13597 gene encodes a 513 amino acid protein of 59.2 kDa (FlyBase) (Tweedie et al., 2009) with two widely spaced bromodomains (amino acid 55 to 127 and 336 to 409) (PROSITE database) (Sigrist et al., 2010) (Fig. 1C). RT-PCR experiments revealed that transcripts are specific to male gonads (data not shown), consistent with Affymetrix expression data also indicating a strong enrichment of CG13597 transcripts in the testis (FlyAtlas: the Drosophila gene expression atlas) (Chintapalli et al., 2007). Very low levels of tbrd-1 expression in other tissues were detected by RNA-seq (modENCODE Project led by Sue Celniker, visible on FlyBase via GBrowse). Accordingly, we named CG13597 tBRD-1 (testis-specifically expressed bromodomain containing protein-1) and the corresponding gene tbrd-1. A tbrd-1 mutant allele (tbrd-1¹) was generated by remobilisation and subsequent integration of P(EPgy2)^EY02323 into the tbrd-1 gene (Fig. 1A,C′; Materials and Methods). This led to a premature translational stop codon after amino acid 281 (262 amino acids of tBRD-1 and 19 amino acids encoded by the P-element) (Fig. 1C′). We raised a peptide antibody (aa 249 to aa 267) against tBRD-1 (Fig. 1C). By Western blot analysis neither the full length nor a truncated tBRD-1 protein could be detected in protein extracts of homozygous tbrd-1¹ mutant testes (Fig. 1D). After strong overexposure a protein is visible at about 35 kDa, which might represent a truncated tBRD-1 protein (predicted molecular mass: 32 kDa) (Fig. 1D′, arrowhead). Analysis of tbrd-1 mutants revealed that tBRD-1 is required for male but not for female fertility. Fertility tests with both tbrd-1¹ homozygotes and tbrd-1¹/Df(3R)ED10893 or tbrd-1¹/Df(3R)Exel9014 trans-heterozygous males demonstrated that mutation of tbrd-1 leads to complete male sterility. Seminal vesicles of 5 days old tbrd-1 mutant males were devoid of sperm also when males were kept isolated from females. A tbrd-1-eGFP transgene made from the genomic region (Fig. 1B; Materials and Methods) rescued the sterility of homozygous tbrd-1¹ males demonstrating that the male infertility was due to the compromised activity of tBRD-1. Whole mount testis and squashed preparations of homozygous tbrd-1¹ (Fig. 2A,D) and tbrd-1¹/Df(3R)ED10893 or tbrd-1¹/Df(3R)Exel9014 transheterozygous (data not shown) males demonstrated testis tubes filled with elongated spermatids (Fig. 2A,D, arrows) indicating substantial differentiation of post-meiotic stages. Nevertheless, first effects were already detectable in early round spermatids (Fig. 2B). In wild-type males the phase-dark Nebenkern formed by the two mitochondrial derivatives and the phase-light nucleus have nearly the same size and are arranged side by side in early round spermatids. During spermatid differentiation the derivatives elongate beside the growing flagellar axoneme (reviewed by Fuller, 1993). In homozygous tbrd-1¹ mutants (Fig. 2B) the Nebenkern (double arrow) was larger than the nucleus (arrowhead) and was surrounded by several nuclei varying in size. Introduction of the tbrd-1-eGFP genomic transgene restored normal morphology of round spermatids (Fig. 2C). Spermatid differentiation is accompanied by an extensive reorganisation of the nucleus. The nuclei develop from a round shape in early spermatids (Fig. 2C, arrowhead) to a very thin needle shape in mature sperm. Visualising DNA by Hoechst revealed that spermatid nuclei in homozygous tbrd-1¹ mutant testis become very small, however, the nuclei remained round (Fig. 2D,D′, arrowheads). In addition, spermatid nuclei were scattered throughout the elongated bundles of flagella (Fig. 2D, arrowhead) instead of being clustered at one end as in the wild-type. At the end of spermatid differentiation mature sperm become individualised by the individualisation complex. This complex is a coordinated array of discrete investment cones which can be visualised by phalloidin staining. Each investment cone individualises a single spermatid (Fabrizio et al., 1998). In homozygous tbrd-1¹ mutants, investment cones are formed (Fig. 2F, arrow) but individualisation complexes were never observed. Introduction of the tbrd-1-eGFP genomic transgene restored normal nuclear shaping and clustering as well as individualisation complex formation (Fig. 2E,E′, arrowheads, Fig. 2G, arrow).

The extensive reorganisation of the nucleus during spermatid differentiation is accompanied by a compaction of the chromatin. In mammals, the testis-specific bromodomain protein BRDT is involved in chromatin reorganisation and is essential for male germ cell differentiation (Pivot-Pajot et al., 2003; Shang et al., 2007). A common feature of mammalian and Drosophila spermatid differentiation is the dramatic reorganisation of chromatin due to replacement of histones by protamines. In Drosophila, Protamine A (Mst35Ba, ProtA), Protamine B (Mst35Bb, ProtB) and Mst77F are major chromatin components of the mature sperm (Jayaramaiah Raja and Renkawitz-Pohl, 2005; Rathke et al., 2010). In situ hybridisations as well as immunofluorescence stainings demonstrated that expression of Mst77F and protB was unaffected in homozygous tbrd-1¹ mutant testis (Fig. 3A,C,E,G). In addition, the histone to protamine transition is accompanied by the occurrence of many DNA breaks and it was shown that hyper-acetylation of histone H4 is essential for chromatin reorganisation (Awe and Renkawitz-Pohl, 2010; Rathke et al., 2007). In homozygous tbrd-1¹ mutants, hyper-acetylation of histone H4, disappearance of histones as well as the occurrence of DNA breaks was not obviously altered (data not shown). Apparently, tBRD-1 is not required for main features of post-meiotic chromatin reorganisation. Nevertheless, many other reorganisations, besides chromatin remodelling, accompany post-meiotic spermatid differentiation.

In adult testis, different staged germ cells are arranged in a spatially ordered manner from germline stem cells at the apical tip to mature sperm in the seminal vesicles at the basal end (reviewed by Fuller, 1993). Analysis of whole mount testis of flies bearing the tbrd-1-eGFP genomic rescue transgene showed expression of tBRD-1-eGFP starting with the onset of the spermatocyte stage (Fig. 4A, arrowhead). Stem cells and spermatogonia in the testis tip as well as post-meiotic stages were free of eGFP signal (Fig. 4A, arrow, double arrow). This is in agreement with in situ hybridisation analyses showing that tbrd-1 transcripts were restricted to spermatocytes (Fig. 5A). Examination of mature spermatocytes expressing tBRD-1-eGFP or immunostained with an anti-tBRD-1 antibody at higher magnification revealed strong localisation of tBRD-1 to the spermatocyte nucleolus, with lower intensity signal distributed over the partially condensed chromosomes (Fig. 4A′,A″, arrowheads, double arrows). In addition, within the nucleoplasm tBRD-1 localises in a set of nuclear speckles (Fig. 4A′,A″, arrows). All tBRD-1 expression vanished with the breakdown of the nucleolus during the G₂/M transition of meiosis I (data not shown). Thus, expression of tBRD-1 and first spermatogenesis defects in homozygous tbrd-1¹ mutants do not coincident in time. This indicates that mutant testis show secondary effects in post-meiotic stages due to loss of tBRD-1 expression in pre-meiotic and meiotic stages. In young spermatocytes tBRD-1 became visible first at the periphery of the nucleolus (Fig. 4B). When spermatocytes grew concurrently the tBRD-1 signal accumulated at the nucleolus and speckles within the nucleoplasm appeared (compare Fig. 4B, Fig. 4E, Fig. 4H). Analysis of testis of flies bearing a further transgene, tbrd-1Δc-eGFP, made from the genomic region but lacking the C-terminal part of the gene (Fig. 1B′; Materials and Methods) showed exactly the same expression pattern like testis from tbrd-1-eGFP (data not shown). The missing C-terminal part encodes amino acid 349-513 containing the second bromodomain. Obviously, the second bromodomain is not essential for proper localisation of tBRD-1. In addition, flies expressing two copies of tbrd-1Δc-eGFP showed normal spermatogenesis and were fertile. The tbrd-1Δc-eGFP transgene was not able to rescue the sterility of homozygous tbrd-1¹ males (data not shown). Immunofluorescence staining of testis of flies bearing tbrd-1-eGFP showed that tBRD-1-eGFP partially colocalises with the tTAF Sa, the general TAF TAF1 and Polycomb in the nucleolus of primary spermatocytes (Fig. 4B–D,E–G, arrowheads; data not shown). Like tBRD-1 also Sa, TAF1 and Polycomb showed a stepwise accumulation at the nucleolus. The tBRD-1 localised to the Fibrillarin-deficient regions of the nucleolus (Fig. 4H–J). Fibrillarin is a component of the nucleolus involved in ribosomal RNA processing (Girard et al., 1993; Jansen et al., 1991). As tBRD-1 and Sa showed a highly similar expression pattern we analysed the three thus far known direct tTAF target genes fzo, dj and Mst87F in homozygous tbrd-1¹ mutant testis. By RT-PCR we were able to detect all three transcripts in homozygous tbrd-1¹ mutant testis (data not shown).

In situ hybridisation and Western blot experiments using homozygous can¹² mutant testis revealed that expression of tBRD-1 was independent of tTAFs (Fig. 5B; data not shown). Immunofluorescence staining of homozygous can¹² and sa² mutant testis showed that localisation of tBRD-1 to the nucleolus required wild-type function of tTAFs (Fig. 5E,F; data not shown). The prominent localisation of tBRD-1 to the nucleolus in wild-type (Fig. 5C, arrow) was strongly reduced in homozygous can¹² and sa² mutant spermatocytes (Fig. 5E, arrow; data not shown). A reduced tBRD-1 localisation to the nucleolus was also observed in homozygous can¹² and sa² mutants, which express the tBRD-1-eGFP fusion protein (data not shown). In addition, an increased tBRD-1 signal was visible within chromosome territories in homozygous can¹² and sa² mutants while hardly any nuclear speckles were detectable (Fig. 5E; data not shown). Analyses of homozygous tbrd-1¹ mutant spermatocytes demonstrated that localisation of tTAFs, TAF1 and Polycomb was independent of tBRD-1. Localisation of tTAFs, TAF1 and Polycomb to the nucleolus could still be detected in homozygous tbrd-1¹ mutant spermatocytes (Fig. 5G,I, arrows; data not shown).

Pupal testis of tBRD-1-eGFP transgenic flies dissected 24 hours after puparium formation (APF) were treated with 50 µM of the histone deacetylase (HDAC) inhibitor trichostatin A (TSA) or 150 µM of the histone acetyltransferase (HAT) inhibitor anacardic acid (AA) for about 24 hours in culture (for establishment of culture conditions, see Awe and Renkawitz-Pohl, 2010). This affected the localisation of tBRD-1 within spermatocyte nuclei (Fig. 6). Immunofluorescence staining with an antibody against acetylated histone H4 revealed a strong increase in H4 acetylation upon TSA-treatment (Fig. 6F), while AA-treatment blocked acetylation nearly completely (Fig. 6J). In contrast to the untreated control (Fig. 6A–D), in TSA-treated testis (Fig. 6E–H) an increased tBRD-1-eGFP signal was visible within chromosome territories (Fig. 6E, arrow). No obvious consequence of TSA-treatment was observed for tBRD-1-eGFP concentrations in the nucleolus (Fig. 6E, arrowhead). Conversely, testis treated with AA showed strong reduction of tBRD-1-eGFP within the nucleolus (Fig. 6I, arrowhead). Similar results were obtained using anti-tBRD-1 to stain for endogenous tBRD-1 in non-transgenic flies (data not shown).

To investigate if tBRD-1 is able to bind chromatin we isolated salivary glands of larvae and prepared polytene squashes. As tBRD-1 is normally not expressed in salivary glands UAS-tbrd-1-eGFP was driven by Sgs58AB for ectopic expression. Immunofluorescence staining using anti-GFP antibody and Hoechst showed that ectopically expressed tBRD-1-eGFP binds at multiple sites along polytene chromosomes (Fig. 7A). While the chromocenter showed no significant binding of tBRD-1 the puffs are highly stained (Fig. 7A, double arrow, arrows). Deeper analysis showed colocalisation of tBRD-1-eGFP with the interbands of polytene chromosomes (Fig. 7B–D, arrowheads show one interband as example).

With tBRD-1 we have identified here a bromodomain-containing protein specifically expressed in the testis in the nuclei of primary spermatocytes. Bromodomain modules are part of many chromatin-associated proteins including histone acetyltransferases (HATs), ATP-dependent chromatin-remodelling factors and the BET family of nuclear proteins, such as Brd2, Brd4 and Bdf1 (Jeanmougin et al., 1997). Moreover, bromodomains can also bind acetylated lysines of non-histone proteins like p53 or c-Myb (Barlev et al., 2001; Sano and Ishii, 2001). tBRD-1 shows similarity to proteins of the BET family (predicted by Ensembl) (Flicek et al., 2011). Function of tBRD-1 is required for spermatid differentiation as mutant males are sterile and exhibit partially disturbed spermatid differentiation. Spermatid nuclei become condensed but remain round and are randomly distributed within spermatid bundles. The typical features of chromatin remodelling during spermatid differentiation are not disturbed in homozygous tbrd-1¹ mutants. Nevertheless, many other processes and different genes are necessary for this dramatic reorganisation of germ cells. About 350 protein components of mature sperm were identified in whole sperm proteomics (Dorus et al., 2006; reviewed by White-Cooper, 2010). A testis-specifically expressed member of the BET-family in mammals is BRDT which is also required for spermatogenesis. Lacking of the first bromodomain of BRDT in mice leads to production of malformed sperm and male sterility despite unaffected protamine expression (Shang et al., 2007). Previously, in Drosophila, the bromodomain-related protein Mtsh was identified. In mtsh mutant males meiosis and spermiogenesis proceed though with lack of proper coordination. Mtsh is proposed to participate in transcriptional regulation of spermatogenesis-specific genes (Bergner et al., 2010).

When we analysed the sub-cellular localisation of tBRD-1 within the testis, we observed that expression is restricted to primary spermatocytes. This is in clear contrast to the situation in homozygous tbrd-1¹ mutants which show first defects of spermatogenesis in post-meiotic stages when expression of tBRD-1 has already vanished. Within primary spermatocytes, we observed a strong localisation of tBRD-1 to the nucleolus as well as localisation over the chromosomes and to several nuclear speckles. In 2009 it was published that CG13597 (tBRD-1) was found as putative component of spliceosomal complexes formed in Kc cell nuclear extracts. However, the authors note that this might simply be a contamination (Herold et al., 2009). As the nuclear speckles tBRD-1 localises to where devoid of active chromatin marks as active RNA-Polymerase II or acetylated histones (data not shown) we focused on the nucleolar and chromosomal localisation of tBRD-1. Fibrillarin is involved in processing of ribosomal RNA (Girard et al., 1993; Jansen et al., 1991). As tBRD-1 proteins largely localised to Fibrillarin-deficient regions within the nucleolus this might argue against a role of tBRD-1 for ribosomal RNA processing. At the same time also tTAFs, TAF1 and Polycomb localised in a Fibrillarin-deficient region within the nucleolus and immunofluorescence stainings showed a partial overlap of tBRD-1 with the tTAF Sa, TAF1 and Polycomb. Only low amounts of tBRD-1 were detectable within chromosome territories, which is also true for tTAFs. Two functions for tTAFs in primary spermatocytes have been described so far: tTAFs directly bind to promoters of several spermatid differentiation genes and they also recruit PRC1 components to the nucleolus (Chen et al., 2005). Thus far, it is not known how tTAFs are recruited to their target genes and also hints for tTAF effector molecules are missing. Analyses of tTAF mutants revealed that tBRD-1 required tTAF function for nucleolar localisation. This was also true for Polycomb and TAF1 (Chen et al., 2005; Metcalf and Wassarman, 2007). In addition, tTAF mutants showed an increased tBRD-1 signal within chromosome territories when compared to the wild-type situation. While tBRD-1 required tTAF function for nucleolar localisation, localisation of tTAFs, TAF1 and Polycomb to the nucleolus was not visibly altered in homozygous tbrd-1¹ mutants when analysed by immunofluorescence. Function of tBRD-1 is required for spermatid differentiation and protein expression is limited to spermatocytes. This holds also true for tTAFs. However, unlike tTAF mutants, which arrest spermatocytes at the G2/M transition and show a complete absence of spermatid differentiation (Hiller et al., 2001; Hiller et al., 2004; Lin et al., 1996), homozygous tbrd-1¹ mutants do not show a meiotic arrest phenotype. Also expression of spermatid differentiation relevant and tTAF dependent genes fzo, dj and Mst87F was not obviously changed in homozygous tbrd-1¹ mutant testis. Nevertheless, thus far, only these three direct target genes of tTAFs are known (Chen et al., 2005) and all three gene products fall into three different classes of proteins. The GTPase Fzo, required for mitochondrial fusion, is expressed early after meiosis (Hales and Fuller, 1997). Thus, the mRNA is under translational repression only for a very short time in comparison to that of Mst87F, which encode a protein of the sperm tail, expressed very late during spermatogenesis (Kuhn et al., 1988). DJ shows a dual expression as a chromatin component until the time of histone degradation and as flagellar protein in later stages of spermiogenesis and in mature sperm (Rathke et al., 2007; Santel et al., 1998). All three proteins have different functions and their mRNAs are released from repression at completely different stages. Considering the high amount of gene products required for proper spermatid differentiation many different mRNAs, encode for many different classes of proteins, have to be synthesised in primary spermatocytes. We propose that a set of tTAF target genes exist which is dependent on tBRD-1 function while other tTAF target genes, like fzo, dj and Mst87F, are independent on tBRD-1 function. In humans, for instance, the bromodomain protein BRD7 is required for efficient p53-mediated transcription of a subset of target genes (Drost et al., 2010; Mantovani et al., 2010).

Inhibitor treatment of cultured testis showed that the acetylation status of primary spermatocytes influenced tBRD-1 localisation within the cell. Reduced acetylation disturbed the normal sub-cellular localisation of tBRD-1 and localisation to the nucleolus was strongly reduced, while induced hyperacetylation led to an enhanced localisation of tBRD-1 to the chromosome territories. Obviously, like other bromodomain proteins, also tBRD-1 is able to recognise and bind acetylated histones and/or non-histone proteins. Ectopically expressed tBRD-1 is per se able to bind polytene chromosomes and localises to euchromatic interbands. This supports the idea that tBRD-1 might regulate transcription.

We hypothesise, that tBRD-1 acts as a reader of acetylated residues of histones and/or non-histone proteins at the promoters of a special set of yet unknown tTAF target genes. Thereby, tBRD-1 may facilitate binding of tTAFs to the promoters of certain genes relevant for spermatid differentiation. In addition, tBRD-1 may support detachment of Polycomb from these promoters. The fact that tTAFs and Polycomb are still detectable within the nucleolus in tbrd-1 mutants is not mutually exclusive with this theory. Indeed it strengthens this idea because transcription of some tTAF target genes, like fzo, dj and Mst87F, is independent of tBRD-1 function and Polycomb is recruited from the promoters of these genes to the nucleolus.

Drosophila melanogaster strains were maintained on standard medium at 25°C. w¹ (Klemenz et al., 1987) and w¹¹¹⁸ were used as wild-type strains. P(EPgy2)^EY02323 (BL15415), Df(3R)ED10893 (BL28827), Df(3R)Exel9014 (BL7992), ZH-86Fb (BL24749) and BL25709 were obtained from the Bloomington Stock Center. Pc-GFP flies were kindly provided by R. Paro (Dietzel et al., 1999). can¹² and sa² mutants (Hiller et al., 2001) were kindly provided by M.T. Fuller (Palo Alto). can¹² were used for in situ hybridisations, immunofluorescence stainings and Western blot analyses. The sa² mutant strain (Hiller et al., 2004) was used for immunofluorescence stainings. Sgs58AB (GAL4 strain under control of the regulatory regions of sgs4) was kindly provided by A. Hofmann and M. Lehmann (Berlin; unpublished).

Whole mount in situ hybridisation of adult testis was performed with modifications according to White-Cooper et al. (White-Cooper et al., 1998). DIG labelled RNA probes were generated using 500 to 800 bp fragments of the corresponding ORFs amplified by PCR on genomic DNA and cloned into pCR®II-TOPO® Vector (Invitrogen).

The insertion in P(EPgy2)^EY02323 lies 635 bp downstream of the translational stop codon of tbrd-1, 115 bp downstream of the poly(A) signal (Bellen et al., 2004). This strain is homozygous viable and fertile. Before remobilisation the P insertion site was proven by PCR strategy. Genomic DNA of P(EPgy2)^EY02323 was isolated and used as template DNA in standard PCR reactions. One primer was chosen out of the insertion element and the secondary primer out of the neighbouring genomic region. The insertion was analysed from the 5′ as well as from 3′ end. PCR products were analysed by sequencing. P(EPgy2)^EY02323 was remobilised by using the transposase source of line w; Δ2-3 Ki/TM3, Sb (C. Klämbt, Münster). 80 single jumpstarter males were crossed with females of the balancer line w; Dr/TM3, Sb Dfd-lacZ. The P-element was followed by monitoring eye colour. Jump-out of the P-element was indicated by loss of white eye marker. Jump-in of the P-element at a new position was indicated by altered red eye colour in comparison to the red eye colour of the original P(EPgy2)^EY02323 insertion. Individual white-eyed P(EPgy2)^EY02323-jo jump-out lines as well as P(EPgy2)^EY02323-ji jump-in lines were isogenised and analysed with regards to lethality and male sterility. Loss of tBRD-1 full length expression was proven by immunofluorescence staining and Western blot analyses using anti-tBRD-1. The molecular analysis of tbrd-1¹ was done using standard PCR experiments and subsequent sequencing.

20 young adult males (wild-type or tbrd-1-eGFP/+; tbrd-1¹/tbrd-1¹ or tbrd-1-eGFP/tbrd-1-eGFP; tbrd-1¹/tbrd-1¹ or tbrd-1Δc-eGFP/tbrd-1Δc-eGFP; tbrd-1¹/tbrd-1¹) were placed individually with three wild-type virgin females in separate vials at 25°C. After 5 days the parental generation was removed. The number of offspring in every vial was counted after two weeks.

Total RNA was extracted from wild-type testis, carcass males (testis were removed by dissection), embryos (0–24 hours), larvae (mixture of male and female) and whole bodies of females by using TRIzol® (Invitrogen). We used the OneStep RT-PCR Kit (Qiagen) to amplify a 332 bp cDNA fragment from the open reading frame (ORF) of tbrd-1. The chosen primer pair spans an intron of 61 bp to distinguish between PCR products based on cDNA template and those from genomic DNA contamination.

To generate a tbrd-1 rescue construct, the open reading frame (ORF) of tbrd-1 gene together with a 531 bp sequence upstream of the ATG translational start was PCR amplified using genomic DNA and primers with linked EcoRI and SpeI restriction sites. The PCR fragment was inserted into pBSIIKS⁺eGFP, which was opened with EcoRI and XbaI, in frame with the eGFP. This clone was digested with EcoRI and NotI and the resulting tbrd-1-eGFP fragment was cloned into the germline transformation vector pChabΔsal (Thummel et al., 1988) (lacZ sequences were removed). Transgenic fly strains were established by injection into w¹ as described by Michiels et al. (Michiels et al., 1993).

To generate tbrd-1Δc-eGFP the N-terminal part of tbrd-1 gene together with a 531 bp sequence upstream of the AUG translational start was PCR amplified using genomic DNA and the primer pair: 5′CACCACTGGGACTCCGCCTTATAGCC3′ and 5′GGAAAAGCGCAAGAGAAAGGCTACT3′. The PCR fragment tbrd-1Δc was inserted into pENTR^TM/D-TOPO® (Invitrogen^TM). tbrd-1Δc was subsequently inserted into the transformation vector pUAST containing the attR cassettes for Gateway® recombination cloning technology (Invitrogen^TM), the attB recognition site for phiC31 mediated integration at attP destination sites in the genome as well as the coding sequence for the C-terminal tag eGFP (pUAST-attB-rfa-eGFP; kindly provided by S. Bogdan, Münster; unpublished). Recombination reaction was catalysed by using the Gateway® LR Clonase® II enzyme mix (Invitrogen^TM). Transgenic fly strains were established by injection into ZH-86Fb (BL24749) and BL25709 (Bischof et al., 2007; Markstein et al., 2008).

To generate UAS-tbrd-1-eGFP the ORF of tbrd-1 gene was PCR amplified using genomic DNA and primers with linked EcoRI and SacII restriction sites. The PCR fragment was inserted into pUASTgreen, which was opened with EcoRI and SacII, in frame with the eGFP. Transgenic fly strains were established by injection into w¹¹¹⁸ as described by Michiels et al. (Michiels et al., 1993). pUASTgreen was generated by transferring the eGFP and the MCS of pEGFP-N1 (Clontech) into pUAST (Brand and Perrimon, 1993) using BglII and XbaI restriction sites (kindly provided by M. Schäfer, Kassel; unpublished).

Hoechst staining was used to visualise chromatin. All antibodies were used in immunofluorescence stainings of squashed testis carried out essentially as described in Hime et al. (Hime et al., 1996) and Rathke et al. (Rathke et al., 2007). We raised a peptide antibody (aa 249 to aa 267) against tBRD-1 in rabbit and applied the affinity-purified antibody in a dilution 1:5000 (Pineda- Antibody-Service; http://www.pineda-abservice.de). Other antibodies were used at the following dilutions: anti-Sa 1:500 (guinea pig) and anti-Fibrillarin straight; from M.T. Fuller (Palo Alto) (Chen et al., 2005), anti-TAF1-C 1:800 (rabbit; kindly provided by D. Wassarman (Madison) (Maile et al., 2004), anti-acetyl-histone H4 1:300 (rabbit; Upstate, Cat#06-598), anti-Mst77F 1:500 (Rathke et al., 2010). Cy3-conjugated anti-rabbit (Dianova; 1:100), Cy3-conjugated anti-guinea pig (Dianova; 1:100) and Cy5-conjugated anti-mouse (Dianova; 1:100) were used as secondary antibodies. TUNEL staining was done essentially as described in Rathke et al., 2007. Polytene chromosomes were prepared and stained as described in (Murawska and Brehm, 2012). The GFP-antibody (rabbit, Rockland Inc.) was applied in a 1:500 dilution. Immunofluorescence, eGFP and Hoechst signals were examined using a Zeiss microscope (AxioPlan2) equipped with appropriate fluorescence filters. Images were individually recorded and processed with Adobe Photoshop 7.0.

Western blots were performed using standard methods. Protein extracts were made from wild-type, homozygous can¹² and homozygous tbrd-1¹ mutant testis. We used 20 testis per protein extract. Dissected testis were homogenised in 20 µl 2×SDS sample buffer by sonication for 30 minutes and incubated afterwards for 5 minutes at 37°C. The whole testis extract was applied to a 10% SDS-gel. Anti-tBRD-1 was used at 1:1000 in 5% dry milk in 1×TBS. Anti-Actin (Biomeda) was used at 1:1000. POD-conjugated anti-rabbit and anti-mouse antibodies were subsequently applied at 1:5000 (Jackson Immunology). ECL reagents (Amersham Pharmacia) were used according to manufactureŕs recommendation to detect the signals.

Pupal testis were dissected, cultured and treated with inhibitors as previously described (Awe and Renkawitz-Pohl, 2010). Pupal testis (one day after puparium formation) were dissected in Shields and Sang M3 insect culture medium (Sigma-Aldrich Cat#S8398) supplemented with 10% fetal bovine serum (heat inactivated, insect culture tested, Sigma-Aldrich Cat#F3018), 100 U/ml penicillin and 100 µg/ml streptomycin (Gibco-Invitrogen Cat#15140-148). Cultures were incubated at 25°C and treated with inhibitors for about 24 hours before fixation. Anacardic acid (Merck Biosciences Cat#172050) was dissolved in DMSO to obtain a 28.69 mM stock solution. Trichostatin A (Cell Signalling Tech. Cat#9950) was dissolved in ethanol to obtain a 4 mM stock solution. For treatment, inhibitors were diluted appropriately in culture medium and added to the culture chambers. Immunofluorescence staining procedure is the same as described above.

We thank Margaret T. Fuller, Renato Paro, David Wassarman and Michael Lehmann for antibodies and fly strains, Mireille Schäfer and Sven Bogdan for plasmids, Xin Chen and Margaret T. Fuller for helpful discussions, Ruth Hyland for excellent technical assistance and Katja Gessner for competent secretarial assistance. This work was supported by the Deutsche Forschungsgemeinschaft within the International Graduate School GRK767, the FOR531, the TRR81 (to R.R.-P.) and the KFO181 (to C.R. and R.R.-P.) and the foundation of the Philipps-Universität Marburg (to C.R.).