Nucleolar colocalization of TAF1 and testis-specific TAFs during Drosophila spermatogenesis

Abstract

In Drosophila, testis-specific TBP-associated factors (tTAFs) predominantly localize to spermatocyte nucleoli and regulate the transcription of genes necessary for spermatocyte entry into meiosis. tTAFs are paralogs of generally expressed TAF subunits of transcription factor IID (TFIID). Our recent observation that the generally expressed TAF1 isoform TAF1-2 is greatly enriched in testes prompted us to explore the functional relationship between general TAFs and tTAFs during spermatogenesis. Analysis by immunofluorescence microscopy revealed that among the general TFIID subunits examined (TATA-box binding protein [TBP], TAF1, TAF4, TAF5, and TAF9), only TAF1 colocalized with the tTAF Mia in spermatocyte nucleoli. Nucleolar localization of TAF1, but not Mia, was disrupted in tTAF mutant flies, and TAF1 dissociated from DNA prior to Mia as spermatocytes entered meiosis. Taken together, our results suggest stepwise assembly of a testis-specific TFIID complex (tTFIID) whereby a TAF1 isoform, presumably TAF1-2, is recruited to a core subassembly of tTAFs in spermatocyte nucleoli. Developmental Dynamics 236:2836–2843, 2007. © 2007 Wiley-Liss, Inc.

INTRODUCTION

Transcription initiation by RNA polymerase II is directed by a host of protein factors, including the general transcription factor TFIID (Walker et al.,2001; Chen and Hampsey,2002; Matangkasombut et al.,2004). TFIID is a multiprotein complex composed of TATA-box binding protein (TBP) and ∼12 TBP-associated factors (TAFs), which function as core promoter selectivity factors and as transcription coactivators (Dynlacht et al.,1991; Verrijzer et al.,1995; Aoyagi and Wassarman,2000; Matangkasombut et al.,2004). All eukaryotic organisms encode a common set of TAFs that appear to be generally expressed (Walker et al.,2001; Sanders et al.,2002; Tora,2002). In yeast, Drosophila, and C. elegans, the general TAFs are essential for viability (Wassarman et al.,2000; Aoyagi and Wassarman,2001; Walker et al.,2001; Matangkasombut et al.,2004).

In addition, metazoan organisms have evolved unique TAF paralogs that are expressed in a subset of cells and are required for transcription of cell type–specific genes. In mammals, paralogs of TAF1 (TAF1L), TAF4 (TAF4b), and TAF7 (TAF7L) are highly expressed in testes and, at least in the case of TAF4b and TAF7L, are required for normal levels of sperm production (Wang and Page,2002; Falender et al.,2005; Cheng et al.,2007). Similarly, in Drosophila, No hitter (Nht), Cannonball (Can), Meiosis I arrest (Mia), Spermatocyte arrest (Sa), and Ryan express (Rye), which are paralogs of generally expressed TAF4, TAF5, TAF6, TAF8, and TAF12, respectively, are exclusively expressed in testes (Hiller et al.,2001,2004). Flies homozygous mutant for any one of the testis-specific TAF (tTAF) genes are viable but males are infertile, demonstrating the limited requirement for tTAFs (Lin et al.,1996; Hiller et al.,2001,2004). Furthermore, mutation of any one of the tTAF genes causes similar cellular and molecular phenotypes: arrest in spermatogenesis at the mature spermatocyte stage and reduced transcription of genes required for entry into meiosis.

These observations and the three observations outlined below strongly suggest that tTAFs are components of a testis-specific TFIID (tTFIID) complex. tTAF proteins colocalize within the nucleoplasm and at high levels in the nucleoli of spermatocytes (Chen et al.,2005). TAF4, TAF5, TAF6, and TAF12 have been shown to form a stable core TFIID complex in Drosophila S2 cultured cells, suggesting that tTAF paralogs of these TAFs form a stable core tTFIID complex (Wright et al.,2006). Finally, recombinant Nht and Rye proteins interact through their histone fold domain dimerization motifs, as do their paralogs TAF4 and TAF12, but Nht does not interact with TAF12 and Rye does not interact with TAF4 (Hiller et al.,2004). Thus, tTAFs and general TAFs are likely to be components of distinct TFIID complexes.

To examine this proposal, we have used immunofluorescence microscopy to determine the localization of generally expressed TFIID subunits, namely TBP, TAF1, TAF4, TAF5, and TAF9, and the tTAF Mia in Drosophila testes. We were particularly interested in the localization of TAF1 relative to tTAFs because we have found that TAF1-2, one of four TAF1 mRNA isoforms generated by alternative splicing, is more highly expressed in testes than other tissues and is the most abundant TAF1 mRNA isoform in testes, constituting ∼45% of total TAF1 mRNA (Fig. 1) (Katzenberger et al.,2006). For comparison, in whole adult male flies, TAF1-2 mRNA constitutes ∼10% of total TAF1 mRNA. Furthermore, TAF1-2 contains a DNA binding domain composed of two AT-hook motifs, which is required for binding testes-specific gene promoters in vitro (Metcalf and Wassarman,2006).

Figure 1. A schematic diagram of the Drosophila TAF1 gene and mature TAF1 mRNAs generated by alternative splicing. Details are provided in Katzenberger et al. (2006). Exons are depicted as rectangles and introns as lines. Exons colored gray are constitutively included in the mature mRNA and exons 12a and 13a (colored in red and green, respectively) are alternatively included in the mature mRNA to produce four distinct mRNAs, TAF1-1, TAF1-2, TAF1-3, and TAF1-4. For simplicity, only exons 12–14 are shown for the TAF1 mRNA isoforms.

RESULTS AND DISCUSSION

TAF1 Is Expressed in Nuclei of Transcriptionally Active Sperm Cells

We used immunofluorescence microscopy to determine the cellular distribution of TAF1 in Drosophila testes. A whole mounted testis reveals a developmental timecourse of spermatogenesis in succession from the apical tip to the seminal vesicle: germ cells and the mitotically proliferating spermatagonia, early spermatocytes, mature spermatocytes, round spermatids, elongated spermatids, and mature sperm (Fuller,1993). Three independently generated antibodies that specifically recognize TAF1 revealed identical expression patterns in whole testes (Fig. 2A; see Supplemental Fig. 1, which can be viewed at www.interscience.wiley.com/jpages/1058-8388/suppmat, and data not shown). During sperm cell differentiation, TAF1 was first expressed at a low level in nuclei of mitotically proliferating spermatogonia at the apical tip of the testis (Fig. 2B). TAF1 expression increased in spermatocytes and reached its highest level in nuclei of spermatocytes that were undergoing growth during an extended pre-meiotic G2 phase. Upon entry into meiosis I, TAF1 expression was reduced and TAF1 was dispersed throughout the cell. TAF1 protein was not observed in cells that had entered meiosis II or transcriptionally quiescent round or elongating spermatids. Thus, TAF1 expression is restricted to transcriptionally active pre-meiotic cells, consistent with its role as a component of TFIID. Additionally, TAF1 expression peaked during the highly transcriptionally active spermatocyte growth phase, as was observed for tTAFs, suggesting that TAF1 collaborates with tTAFs to regulate transcription in spermatocytes (Chen et al.,2005).

Figure 2. Immunoflourescence microscopy of Drosophila testes with α-TAF1-C antibody revealed accumulation of TAF1 in spermatocyte nuclei. A: A whole mount wild type testis stained with DAPI and α-TAF1-C antibody. Labels in the lower right of each panel are colored to match the channel colors in the merged image (Merge) on the right. B: Timecourse of α-TAF1-C staining in male germ cells throughout development. Dotted lines indicate the transition points between developmental stages. Labels in the top of each row of panels are colored to match the channel colors in the merged image (Merge) on the right.

TAF1 Localizes to Fibrillarin-Deficient Regions of Spermatocyte Nucleoli

To further characterize TAF1 expression, we examined its subnuclear distribution. In spermatocyte nuclei, phase contrast microscopy and DAPI staining revealed the association of chromosomes in three domains, which are thought to correspond to the paired second, third, and X and Y chromosomes (Fig. 3A) (Cenci et al.,1994). Costaining with DAPI and an α-TAF1 antibody showed that TAF1 colocalized with one of the domains, which, based on morphology, corresponded to the paired X and Y chromosomes. Since the X and Y chromosomes contain the ribosomal DNA repeats that define the nucleolus, we compared TAF1 localization to that of the nucleolar marker Fibrillarin (Ochs et al.,1985; McKee and Karpen,1990). Strong Fibrillarin staining was observed in spermatogonia and early spermatocytes while staining was reduced in mature spermatocytes (Fig. 3B). In spermatogonia, TAF1 and Fibrillarin localization did not overlap, indicating that TAF1 was exclusively nucleoplasmic (Fig. 3C). In contrast, in spermatocytes, TAF1 and Fibrillarin appeared to colocalize in nucleoli (Fig. 3D). However, high-magnification images of individual nucleoli revealed that TAF1 and Fibrillarin occupied distinct domains in the nucleolus and that TAF1 protein was primarily at the nucleolar periphery (Fig. 3E). A similar complementary pattern of localization was observed with tTAFs and Fibrillarin (Chen et al.,2005). These data indicate that TAF1 localization within the nucleus is dynamically regulated during spermatogenesis and that, similar to tTAFs, TAF1 predominantly localizes to spermatocyte nucleoli.

Figure 3. TAF1 localizes to spermatocyte nucleoli. A: Images of spermatocytes by phase contrast (left) and immunofluorescence (DAPI in the middle and α-TAF1-C on the right). Arrowheads indicate presumptive paired second or third chromosomes and arrows indicates the paired X and Y chromosomes. B–E: α-TAF1-C staining is shown in red on the left, α-Fibrillarin staining is shown in green in the middle, and a merged image (Merge) is shown on the right. Arrows indicate the position of the nucleolus. B: A whole-mount testis. C: Spermatogonia. D: Spermatocytes. E: High magnification images of four spermatocyte nucleoli.

Localization of TAF1 to spermatocyte nucleoli raised that possibility that TAF1 regulates transcription from the Y-chromosome. In addition to serving as the site of ribosome biogenesis, Drosophila spermatocyte nucleoli are sites from which Y-chromosome loops originate (Bonaccorsi et al.,1988,1990). Y-chromosome loops encode fertility genes that are highly transcribed and are essential for spermatocytes to enter meiosis (Lifschytz and Hareven,1977; Bonaccorsi et al.,1990). To determine whether localization of TAF1 to spermatocyte nucleoli requires the Y-chromosome, we examined TAF1 localization in X0 male flies. This analysis revealed that TAF1 localization in primary spermatocytes of X0 males was indistinguishable from that of wild type males (data not shown), indicating that association of TAF1 with spermatocyte nucleoli does not require the Y-chromosome.

Colocalization of TAF1 and Mia Is Restricted to Pre-Meiotic Spermatocytes

Similarities between the localization pattern of TAF1 and the described localization pattern of tTAFs during spermatogenesis suggested that TAF1 colocalizes with tTAFs (Chen et al.,2005). To directly test this proposal, we costained testes with α-TAF1 and α-Mia antibodies. This analysis revealed that TAF1 and Mia colocalized in spermatocyte nucleoli (Fig. 4A). High-magnification images of individual nucleoli revealed that both proteins were distributed throughout the nucleolus and TAF1 was enriched at the outer edge of the nucleolus (Fig. 4B). This colocalization was strikingly different from that of TAF1 and Fibrillarin (Fig. 3E), suggesting that TAF1 and Mia localize to the same Fibrillarin-deficient regions of nucleoli. Interestingly, the subcellular localization patterns of TAF1 and Mia were not coincident at all stages of spermatogenesis. In spermatocytes after breakdown of the nucleus in meiotic prophase I, Mia remained localized to the condensed DNA while TAF1 was excluded from the DNA (Fig. 4C). These findings provide support for the hypothesis that TAF1 and tTAFs are components of a tTFIID complex within spermatocytes and suggest that association of TAF1 with tTAFs is a developmentally regulated event.

Figure 4. Colocalization of TAF1 and Mia is restricted to pre-meiotic spermatocytes. A: Spermatocytes stained with DAPI and α-mTAF1-C and α-Mia antibodies. B: High magnification images of four spermatocytes nucleoli stained with α-mTAF1-C and α-Mia antibodies. C: Cells in meiosis I stained with DAPI and α-mTAF1-C and α-Mia antibodies. Arrows indicate the same cell. Labels in the lower right of each panel are colored to match the channel colors in the merged image (Merge) on the right.

Generally Expressed TFIID Subunits Do Not Localize to Spermatocyte Nucleoli

For comparison to TAF1 and tTAFs, we examined the localization of generally expressed TFIID subunits during spermatogenesis. Testes were costained with antibodies to TAF1 and TBP, TAF4, TAF5, or TAF9. In spermatogonia, TBP, TAF4, and TAF5 levels were substantially higher than TAF1 or TAF9, but TBP and all of the generally expressed TAFs were localized to the nucleoplasm (Fig. 5 and data not shown). Similarly, in the same testis preparations, TBP and all of the generally expressed TAFs were localized to the nucleoplasm of somatic cells in the accessory gland and seminal vesicle (see Supplementary Fig. 2). These findings are consistent with the described role of TBP and generally expressed TAFs as TFIID subunits. In spermatocytes, TAF1, TBP, and TAF9 levels were substantially higher than TAF4 or TAF5, which were not detectable. However, unlike TAF1, TBP and TAF9 were not concentrated in nucleoli and, in fact, TAF9 was excluded from nucleoli (Fig. 5). These data are consistent with a model in which TAF4 and TAF5 are replaced by the tTAF paralogs Nht and Can, respectively, in a spermatocyte tTFIID complex that is distinct from spermatogonia or somatic cell TFIID complexes.

Figure 5. Not all TFIID subunits localize to spermatocyte nucleoli. A: An apical tip of a testis stained with DAPI and α-TAF1-C and α-TBP antibodies. B: An apical tip of a testis stained with DAPI and α-TAF1-C and α-TAF4 antibodies. C: An apical tip of a testis stained DAPI and α-TAF1-C and α-TAF9 antibodies. D: High magnification images of primary spermatocytes stained with DAPI and α-TAF1-C and α-TAF9 antibodies. Labels in the lower right of each panel are colored to match the channel colors in the merged image (Merge) on the right.

TAF1 Requires tTAFs for Nucleolar Localization in Spermatocytes

To directly determine whether TAF1 localization to spermatocyte nucleoli depends on tTAFs, we used immunofluorescence microscopy to examine TAF1 localization in flies homozygous mutant for the tTAFs sa, mia, or rye. Since spermatocytes are unable to enter meiosis in tTAF mutants, they accumulate and occupy a large portion of the testis, as observed in sa mutant testis stained with an α-TAF1 antibody (Fig. 6A) (Lin et al.,1996). In sa, mia, or rye mutant testes, TAF1 was not concentrated in the spermatocyte nucleoli, but instead was dispersed throughout the nucleoplasm (Fig. 6B–D). In contrast, in sa or rye mutant testes, Mia localization to spermatocyte nucleoli was largely unperturbed (Fig. 6E and F). These data indicate that recruitment of TAF1 to spermatocyte nucleoli requires tTAFs and suggests that TAF1 is recruited to a core tTFIID complex composed of tTAFs whose integrity is not disrupted by the absence of any individual tTAF.

Figure 6. TAF1 does not localize to spermatocyte nucleoli in sa, mia, or rye mutant flies. A: Whole mount testis from a fly homozygous mutant for sa stained with DAPI and an α-TAF1-C antibody. B: Homozygous sa¹ mutant spermatocytes stained with DAPI and an α-TAF1-C antibody. C: Transheterozygous mia^EY07883/mia^B560 mutant spermatocytes stained with DAPI and an α-TAF1-C antibody. D: Homozygous rye^KG00946 mutant spermatocytes stained with DAPI and an α-TAF1-C antibody. E: Homozygous sa¹ mutant spermatocytes stained with DAPI and α-mTAF1-C and α-Mia antibodies. F: Homozygous rye^KG00946 mutant spermatocytes stained with DAPI and α-mTAF1-C and α-Mia antibodies. Labels in the lower right of each panel are colored to match the channel colors in the merged image (Merge) on the right. Labels in the upper right of each panel indicate the fly genotype.

Aly and Twe Are Not Required for Localization of TAF1 to Spermatocyte Nucleoli

To determine whether disruption of the spermatogenesis transcription program or blocking meiotic cell cycle entry is sufficient to disrupt TAF1 nucleolar localization in spermatocytes, we used immunofluorescence to examine TAF1 localization in flies homozygous mutant for always early (aly) or twine (twe). Similar to tTAF mutants, aly and twe mutants cause meiotic arrest resulting in the accumulation of spermatocytes (Alphey et al.,1992; Lin et al.,1996). Aly acts upstream of tTAFs and aly mutants fail to transcribe key genes involved in cell cycle control and spermatid differentiation (White-Cooper et al.,1998). Aly may regulate chromatin structure in spermatocytes through activation of a chromatin-remodeling complex (Perezgasga et al.,2004). Twe acts downstream of Aly and the tTAFs and is a germline-specific paralog of the cell cycle regulator CDC25, which stimulates meiotic entry of spermatocytes (Alphey et al.,1992; Lin et al.,1996). As shown in Figure 6A and B, TAF1 localization to spermatocyte nucleoli was not disrupted in aly or twe mutant flies, indicating that Aly and Twe are not necessary for TAF1 nucleolar localization. Furthermore, these data indicate that neither perturbation of the testis-specific transcription program nor resulting meiotic arrest is sufficient to block TAF1 accumulation in spermatocyte nucleoli. However, TAF1 localization to nucleoli of aly mutant spermatocytes was morphologically distinct from that of wild type spermatocytes. Specifically, TAF1 localized to several amorphously shaped domains at the outer edge of the nucleolus. To determine whether the altered localization was due to a change in nucleolus morphology, we costained aly mutant testis with α-Fibrillarin and α-TAF1 antibodies. High magnification analysis of individual nucleoli revealed that Fibrillarin localization in aly testes was comparable but not identical to wild type testes, which showed a less dispersed distribution (Fig. 7C). These data indicate that Aly-mediated chromatin remodeling is generally required for proper localization of nucleolar proteins.

Figure 7. TAF1 localizes to spermatocyte nucleoli in flies harboring the meiotic arrest mutations aly and twe. A: Spermatocytes mutant for aly¹ stained with DAPI and an α-TAF1-C antibody. B: Spermatocytes mutant for twe¹ stained with DAPI and an α-TAF1-C antibody. C: High magnification images of four spermatocytes nucleoli from aly¹ mutant flies stained with α-TAF1-C and α-Fibrillarin antibodies. Labels in the lower right of each panel are colored to match the channel colors in the merged image (Merge) on the right. Labels in the upper right of each panel indicate the genotype.

Summary

We observed that the localization pattern of TAF1, but not other generally expressed TAFs or TBP, was similar to that of the tTAF Mia. TAF1 and Mia expression was most prominent in spermatocytes, where they were found to colocalize in nucleoli. In contrast, TBP, TAF4, and TAF5 expression was highest in the nucleoplasm of spermatogonia, and TAF9 expression was highest in the nucleoplasm of spermatocytes. Thus, there are at least three distinct patterns of expression of TFIID subunits during spermatogenesis. TAF1 localization to nucleoli appears to be dependent on a complete core tTFIID complex, as TAF1 predominantly localized to the nucleoplasm in spermatocytes of sa, mia, or rye mutant testes. However, Mia localization to nucleoli does not appear to require a complete core tTFIID complex, as Mia remained localized to spermatocyte nucleoli of sa or rye mutant testes. These data extend the observations of Chen et al. (2005) and Wright et al. (2006) and suggest a stepwise assembly pathway for a tTFIID complex in which a TAF1 isoform, most likely TAF1-2, is recruited to a core tTFIID complex composed of tTAFs. Finally, these results highlight the specialized role of the nucleolus during spermatogenesis. Further characterization of the requirements for TAF1-2 recruitment to the nucleolus and the role of TAF1-2 AT-hook-dependent DNA binding in this process should provide insight into the function of the putative tTFIID complex and the nucleolus during spermatogenesis.

EXPERIMENTAL PROCEDURES

Drosophila Stocks

Drosophila melanogaster stocks were maintained and crosses performed according to standard procedures. w¹¹¹⁸ flies were used as the wild type strain. The C(1;Y)2, y⁺/0 flies used to generate X0 males and stocks carrying the sa¹, mia^EY07883, mia^B560, aly¹, twe¹, and TAF12L^KG00946 (rye) mutations were obtained from the Bloomington Stock Center (Bloomington, IN). X0 males were generated by crossing C(1;Y)2, y⁺/0 males to w¹¹¹⁸ females.

Primary Antibodies

Affinity purified TAF1 rabbit polyclonal antibody TAF1-C was previously described (Maile et al.,2004). A mouse polyclonal α-mTAF1-C antibody was raised to a recombinant polypeptide encoded by exons 12–14 of TAF1-4 (Katzenberger et al.,2006). Affinity purified rabbit polyclonal α-Mia antibody was raised against amino acids 1–148 of recombinant Mia protein and was characterized by Chen et al. (2005). The mouse monoclonal α-Fibrillarin antibody was obtained from Cytoskeleton Inc. Antibodies to Drosophila TBP, TAF4, TAF5, and TAF9 were kindly provided by R. Tjian (University of California-Berkeley) (Wright et al.,2006). Antibodies were used at the indicated dilutions: mouse Fibrillarin (1:200), rabbit TAF1-C (1:800), mTAF1-C (1:200), Mia (1:5), TBP (1:10), TAF4 (1:5), TAF5 (1:5), and TAF9 (1:5).

Testis Staining and Immunofluorescence

Testes from five 1-day-old males were manually dissected in Ringers buffer, transferred to 10 μL of Ringers buffer on a Superfrost/Plus slide (Fisher), squashed, and fixed with methanol-acetone according to the method of Pisano et al. (1993). Fixed preparations were washed in Ringers buffer and blocked with Ringers buffer containing 5% normal goat serum (Sigma) for 1 h at room temperature. Primary antibody incubation was performed overnight (10–14 h) at 4°C in a humid chamber. Incubation with FITC- or rhodamine-conjugated α-rabbit or α-mouse IgG secondary antibodies (Jackson Laboratories) was performed for 1 h at room temperature. Immunostained samples were washed three times in Ringers buffer with the second wash containing 1 μg/ml 4′,6′-diamidino-2-phenylindole (DAPI). Samples were mounted in Vectashield (Vector Labs) under a coverslip are sealed with fingernail polish. Phase and immunoflourescence microscopy was performed with a Axiovert 200M microscope (Zeiss). Micrographs were recorded with an AxioCam digital camera (Zeiss).

Acknowledgements

We thank M. Marengo for producing the α-mTAF1-C antibody, N. Aoyagi for producing the α-Mia antibody, R. Tjian for providing antibodies to general TFIID subunits, and S. Rimkus, R. Katzenberger, and M. Marengo for many helpful discussions.