The testis-specific proteasome subunit Prosα6T of D. melanogaster is required for individualization and nuclear maturation during spermatogenesis

The proteasome is the protein degrading machine of the ubiquitin-mediated proteolytic pathway, which has been implicated in many cellular processes,including cell cycle progression, transcriptional regulation, signal transduction, and cell fate determination(Glickman and Ciechanover,2002). This pathway also carries out an important quality control function by ridding cells of defective polypeptides. The highly conserved 26S proteasome is composed of two subcomplexes: a 20S catalytic core particle (CP)and 19S regulatory particles (RP) capping each end. The CP consists of two identical outer α-rings and two identical inner β-rings, each composed of seven evolutionarily related subunits (α1-7 and β1-7)that stack to form a hollow cylinder(Gröll et al., 2000; Unno et al., 2002). The RP includes a ring of six homologous AAA-family ATPase subunits that function to unfold protein substrates and translocate them into the CPs inner chamber. The functions of most of the twelve other RP subunits are less well understood but they include binding and removing polyubiquitin chains from proteins targeted for destruction (Glickman et al.,1998).

In addition to these integral proteasomal subunits, a number of more loosely, or transiently, associated proteins have recently been identified(Glickman and Raveh, 2005). These auxiliary proteins include ubiquitin-binding proteins that facilitate the delivery of substrates to the proteasome, ubiquitin ligases that build polyubiquitin chains onto the target proteins, and deubiquitinating enzymes that remove these tags (Schmidt et al.,2005).

The variety of proteasome-associated proteins suggests that proteasomes may be dynamic complexes with some diversity in composition. There is, however,another level of compositional diversity, as represented by the structurally heterogeneous proteasomes found in some species(Belote and Zhong, 2005). For example, in mammals, during the anti-viral immune response, γ-interferon induces the synthesis of three new β-type subunits which take the places of their corresponding conventional subunits to form `immunoproteasomes'(Kloetzel, 2004) that are more efficient at producing peptide antigens for MHC class-I mediated antigen presentation. In Arabidopsis and rice, many of the proteasome subunits are represented by more than one gene, suggesting a high degree of structural heterogeneity in which different proteasome complexes are composed of different isoforms of several of the subunits(Fu et al., 1999; Yang et al., 2004; Oguchi et al., 2001; Shibahara et al., 2004). In this case, however, the functional significance of this proposed heterogeneity is unknown.

One of the most striking examples of this phenomenon is seen in Drosophila melanogaster, where about a third of the 26S proteasome subunits have two or three alternative isoforms encoded by paralogous genes(Yuan et al., 1996; Ma et al., 2002; Belote and Zhong, 2005). Molecular analyses have shown that, in every case, one form of each subunit is expressed in both sexes at all developmental stages examined (these will be referred to as conventional proteasome subunits), whereas all of the additional isoforms are testis specific. This represents a unique example of developmental regulation of alternative proteasome subunit expression, and suggests that there might be a specialization of proteasome function during spermatogenesis. Expression pattern studies of the testis-specific subunit Prosα3T, using GFP-tagged reporters, found that this subunit is first detected during meiosis and becomes conspicuous during the late stages of spermiogenesis (Ma et al.,2002). By contrast, Prosα3 (also known as Pros29 - FlyBase),representing the conventional counterpart of Prosα3T, is expressed somatically and in the germline at the early stages of spermatogenesis, but gradually disappears after meiosis, and is not detected in mature sperm(Ma et al., 2002). Although the expression patterns of all testis-specific proteasome subunit genes have not been analyzed in detail, preliminary studies of a few of them suggest that they are expressed in a similar way (Yuan et al., 1996) (L.Z., unpublished; X. Li and J.M.B., unpublished). The large number of proteasome subunit isoform genes, and the collective shift of the expression patterns between the conventional subunit genes and their testis-specific counterparts, suggest that there might be a testis-specific proteasome that is dynamically assembled during Drosophilaspermatogenesis. If so, what is its functional significance?

Here we characterize in detail the spermatogenic expression of theα-type subunit Prosα6 (also known as Pros35 - FlyBase) and its testis-specific isoform Prosα6T. We also address the question of the functional significance of testis-specific proteasome subunits by studying the phenotypes of a knockout mutant of Prosα6T. Mutants are male-sterile, with disruption of actin cone movement during sperm individualization, and abnormal nuclear maturation and morphology. These findings establish a necessary role of at least one of the testis-specific proteasome subunits in Drosophila spermatogenesis.

The Df(2L)ED784 and His2AvGFP stocks were obtained from the Bloomington Stock Center (stock nos 7421 and 5941, respectively). The y w; [ry⁺ 70FLP]4 [v⁺, 70 I-SceI]2B Sco/S² CyO and w¹¹¹⁸; [ry⁺, 70FLP]10 lines were provided by Kent Golic (University of Utah).

The construction of the Prosα6T-GFP reporter gene and generation of transgenic flies were previously described by Ma et al.(Ma et al., 2002). For the construction of the Prosα6-GFP reporter gene, a recombinant λ-phage containing the Pros35 gene (encoding the Prosα6 proteasome subunit) was isolated as described by Ma et al.(Ma et al., 2002), and an 8.0 kb BamHI fragment was cut out of this and subcloned into pGEM4(Promega) to give pGEM4/Prosα6-8.0B. Site-directed mutagenesis, with primers Pros35Not5-(5′-AGCAGCGTCCAGCGGCCGCTAGGCATTTATAG-3′) and Pros35Not3-(5′-CTATAAATGCCTAGCGGCCGCTGGACGCTGCT-3′), was carried out to create a NotI site at the 3′ end of the Pros35open reading frame (ORF) using the Quik Change Site-directed Mutagenesis Kit(Stratagene), yielding pGEM4/Prosα6-8.0B-Not. A NotI GFP cassette was made by PCR amplification of pEGFP (Invitrogen) using the following primers: EGFP Not5′-(GCGCGGCCGCCATGGTGAGCAAGGGCGAGGAGCTGTTCAC-3′) and EGFP Not3-(5′-AAGGGCCCGTACGGCCGACTAGTAGGCCTA-3′). This PCR product was cloned into pGEM-T Easy and then cut out with NotI and ligated into the newly created NotI site of pGEM4/Prosα6-8.0B-Not. A 4.4 kb PstI-EcoRV fragment containing the GFP-tagged Pros35 gene was shuttled into pBlueScript KS+ (Stratagene), and then removed using PstI and XhoI and subcloned into the transformation vector pW8 to give pW8/Prosα6-GFP.

For the generation of flies expressing fluorescently tagged protamine, a BAC clone (RPCI-98 33.C.24) of the genomic region containing the Mst35Ba (Protamine A) gene was obtained(Hoskins et al., 2000), and a 6.0 kb NruI-Asp700 fragment, containing the ProtAgene, was subcloned into pBlueScript to give pBS/ProtA6.0NA. To create the GFP fusion, site-directed mutagenesis was done to create a unique NdeI site at the C terminus of the ProtA coding region yielding pBS/ProtA-Nde. A GFP NdeI cassette was created by PCR amplification of pEGFP. This was inserted in-frame into the NdeI site of pBS/ProtA-Nde and the fusion construct then subcloned into pW8 to give pW8/ProtA-GFP.

A testis-specific expression vector, pW8/TS1, was created by modifying a genomic clone, pW8/Prosα3T-2.8, containing the testis-specific Prosα3T gene coding region flanked by 0.8 kb of upstream sequences and 1.1 kb of downstream sequences(Ma et al., 2002). Using site-directed mutagenesis, a KpnI site was inserted just before the start codon, and a NotI site was inserted just before the stop codon. The Prosα3T coding region was then replaced by sequences between the KpnI and NotI sites of the multiple cloning region of pBlueScript KS+, yielding pW8/TS1. The Prosα6 coding region was PCR amplified from genomic DNA and inserted into the NotI site of pW8/TS1 to give pW8/TS1-Prosα6.

The various pW8 constructs were introduced into the genome by P-element transformation methods, with w¹¹¹⁸ as the host, and several transgenic lines established for all constructs.

For the construction of the Prosα6T targeting donor plasmid, two oligonucleotides, 5′-GGCCTAGGGATAACAGGGTAAT-3′ and 5′-GGCCATTACCCTGTTATCCCTA-3′, were annealed to create an I-SceI cutting site flanked by 5′-GGCC-3′ overhangs. This was used to replace the GFP-NotI cassette in pBS/Prosα6T-GFP. To eliminate two unwanted PstI sites flanking the Prosα6T gene, this plasmid was digested with HincII, ligated, digested with EcoRV and SmaI, and ligated. To create a frameshift in the 5′ region of the Prosα6T ORF, the resulting plasmid was digested with PstI, the 3′ overhangs polished using the Strategene Polishing Kit (Strategene), and treated with ligase to yield plasmid pBS/Prosα6T-KO-SceI. The insert was cut out with HindIII and XhoI and cloned into the corresponding sites of plasmid pBS(-Not),provided by Kent Golic. The insert was then cut out with NotI and cloned into the transformation vector pTV2, also provided by Kent Golic. The resulting plasmid, pTV2-Prosα6T-KO-SceI, was then introduced into the fly genome by P-element germline transformation.

Flies with an X-linked TV2-Prosα6T-KO-SceItargeting donor transgene were crossed to y w; [ry⁺70FLP]4 [v⁺, 70 I-SceI]2B Sco/S² CyO mates and 0-to 3-day-old F1 larvae were heat-shocked at 38°C for 90 minutes. Mosaic-eyed daughters were then crossed to w¹¹¹⁸;[ry⁺, 70FLP]10 males, and non-mosaic red-eyed F2 offspring,potentially representing homologous recombinants, were selected and crossed to w; Bl/CyO to establish balanced lines. Of the 79 candidate F2 flies,only one had the w^hs marker on chromosome 2, as expected for the homologous recombinant. Curly-winged flies from this line were crossed with w¹¹¹⁸; Bl/CyO; P{v[+t1.8]=hs-I-CreI.R}1A Sb¹/TM6 flies and F1 larvae heat-shocked at 36°C for 60 minutes. Resulting mosaic eyed, Curly-winged females were then crossed with w¹¹¹⁸; Bl/CyO males and selected for white-eyed,Curly-winged, non-Bristle offspring, which were the potential Prosα6T mutants. Separate lines were established for each candidate mutant, balanced over CyO, and the Prosα6T gene PCR amplified and checked by PstI digestion and DNA sequencing. Out of 11 candidate mutant lines,only one was found to have the desired mutation.

Individual males were placed with two virgin w¹¹¹⁸females in culture vials for 7 days at 25°C. The flies were then removed and all resulting F1 progeny were counted. At least six vials were set up and scored for each genotype tested.

Alexa Fluor 635 phalloidin, Alexa Fluor 532 phalloidin, Alexa Fluor 488 phalloidin, Alexa Fluor 633 rabbit anti-mouse IgG, Alexa Fluor 532 rabbit anti-mouse IgG, Alexa Fluor 488 rabbit anti-mouse IgG, and TO-PRO-3 were obtained from Molecular Probes (Invitrogen). The anti-Prosα7 monoclonal(clone MCP72) antibody was from Biomol (Exeter, UK). The anti-active Dronc(also known as Nedd-2 like caspase - FlyBase) antibody was generously provided by Bruce Hay (California Institute of Technology), and the CM1 antibody was a gift from Maya Bader and Hermann Steller (The Rockefeller University). For DNA and actin double staining, testes from newly eclosed males were dissected in phosphate-buffered saline (PBS, 130 mM NaCl, 7 mM Na₂HPO₄, 3 mM NaH₂PO₄, pH 7.2) and then fixed with EM grade 4% paraformaldehyde (Electron Microscope Sciences) in PBST (PBS+0.3% Triton X-100) for 1 hour at room temperature. After washing with PBST, the testes were blocked with 10% fetal bovine serum (FBS) and 0.5 mg/ml RNaseA in PBST at room temperature for 1 hour. Then the testes were washed once in PBST, and twice in PBS, followed by incubation with 1 μM TO-PRO3 and 4 units/ml Alexa Fluor phalloidin in PBS for 30 minutes in the dark. They were then rinsed three times in PBS, and mounted in ProLong mounting solution (Molecular Probes). For anti-Prosα7, active Dronc and CM1 antibody staining, after fixing and washing as described above, the testes were blocked with 10% FBS in PBST at room temperature for 1 hour, and then incubated with anti-Prosα7 antibody, diluted 1:500 in PBST + 10% FBS, at 4°C overnight. After washing in PBST, the testes were incubated with secondary antibody (Alexa Fluor 633 or 532 anti-mouse IgG, 1:1000 dilution in PBST + 10% FBS) at room temperature for 4 hours, rinsed three times in PBST,three times in PBS, and then mounted in ProLong. A Zeiss LSM5 Pascal confocal microscope was used for fluorescence imaging.

To investigate the roles of the testis-specific proteasome subunit Prosα6T, we analyzed in detail its expression and subcellular localization patterns during spermatogenesis. In normal spermatogenesis, stem cells located at the apical tip of the coiled testis divide to produce spermatogonial cells that undergo four cycles of mitosis, each yielding sixteen primary spermatocytes that are interconnected, as a result of incomplete cytokinesis, and surrounded by two somatic cyst cells(Lindsley and Tokuyasu, 1980; Fuller, 1993). Following a growth phase in which the cyst increases in size 25-fold, the 16 primary spermatocytes of each cyst undergo meiosis, resulting in a cyst of 64 spherical, haploid spermatids, each containing a prominent round nucleus and a phase-dark structure, the nebenkern, consisting of fused mitochondria.

In the later stages of sperm development, known as spermiogenesis, the spermatids undergo tremendous elongation. The nuclei condense and experience a dramatic shape change to assume dense, needle-shaped configurations. The spermatids undergo individualization, during which the syncytial spermatid bundle is resolved into 64 separate sperm cells, and excess cytoplasmic contents are removed and accumulated into a `waste bag' that is subsequently degraded and reabsorbed. In the final stage of spermiogenesis, the sperm coil up, the cyst ruptures, and the mature sperm move into the seminal vesicle for storage.

In the present study, we have examined the spermatogenic expression patterns of Prosα6 and Prosα6T in detail, using GFP-tagged reporter transgenes. Similar to the expression pattern of Prosα3-GFP,Prosα6-GFP is seen in both cytoplasm and nuclei of spermatogonia,spermatocytes and early stage spermatids. During spermiogenesis, the Prosα6-GFP signal begins to fade in elongated sperm bundles and condensed nuclei and is undetected in individualized, mature sperm(Fig. 1A-H). By contrast,Prosα6T-GFP is not observed in pre-meiotic stages, but becomes prominent in the cytoplasm and nuclei of 64-cell spermatid cysts. The GFP fluorescent signal persists in the elongated and condensed spermatid nuclei and also in spermatid tail bundles and can be seen in the mature sperm heads and tails(Fig. 1A′-H′).

In addition to being expressed in elongated nuclei and distributed along spermatid tail bundles, Prosα6T-GFP also aggregates as `speckles' at certain positions along the sperm bundles(Fig. 2A). To better discern the precise localization of these speckles, Prosα6T-GFP testes were fixed and incubated with Alexa-Fluor-phalloidin and TO-PRO3 to stain actin and nuclei, respectively (Fig. 2B). In the elongated spermatid bundles, Prosα6T-GFP speckles were found near the individualization complex (IC), a cytoskeletal-membranous complex containing a cluster of 64 actin-rich structures, called actin cones, that mediate sperm individualization. During this process, the IC moves caudally down the bundle, eliminating all cytoplasmic bridges and pushing the excess cytoplasm and organelles into a `cystic bulge' that eventually becomes the waste bag. As the IC progresses down the bundle, each spermatid becomes sheathed in its own plasma membrane(Tokuyasu et al., 1972; Noguchi and Miller, 2003). The relative positions of the speckles and the actin cones depend on the position of the IC along the bundle. When the IC is first assembled around the elongated and condensed nuclei, the speckles are seen in front of the actin cones. As the IC moves away from the nuclei, the speckles first overlap the actin cones and then lag behind them. Similar speckles are also found in Prosα3-GFP, Prosα3T-GFP and Prosα6-GFP sperm bundles,although speckles associated with the conventional subunits, Prosα3-GFP and Prosα6-GFP, are fainter than those of their testis-specific counterparts (Fig. 2C-E).

To test whether these speckles are really proteasome-related structures,and not merely artifactual aggregates of GFP-tagged proteins, testes were immunostained with an antibody directed against Prosα7 (also known as Prosalpha7 - FlyBase), a conventional proteasome subunit that has no testis-specific isoform. As shown in Fig. 2F, similar speckles were detected. In addition, immunostaining testes of Prosα3T-GFP (not shown) or Prosα6T-GFP transgenic flies with anti-Prosα7 antiserum shows colocalization of the GFP and antibody signals (Fig. 2G-I), confirming that these speckles are indeed proteasome-related structures. The subcellular nature of these structures is not known, and there are no obvious correlative structures in the published transmission electron micrographs of individualization complexes (Tokuyasu et al., 1972).

The expression patterns of Prosα6T and other testis-specific proteasome subunits suggest that there might be crucial roles for the proteasome during the late stages of sperm development. To investigate this,we generated a knockout mutant of the Prosα6T gene(also known as CG5648, map position 34A11), using targeted homologous recombination (Rong et al.,2002). A four base pair deletion was generated in the Prosα6T coding region to produce a frame-shift mutation at codon 115 (out of 289), with a translation stop occurring 41 codons later (Fig. 3A). Candidate mutation lines were examined by PCR amplification, restriction analysis and sequencing and one was shown to carry the desired mutant allele,which was named Prosα6T¹.

Because Prosα6T is expressed only in testes it was expected that the Prosα6T¹ mutant would be homozygous viable, but that it might affect male fertility. Indeed, this was the case. Homozygous mutants exhibited normal viability, but the males were sterile (Fig. 3B). That this male-sterility is due to Prosα6T¹, and not caused by some inadvertently induced mutation, is shown by the observations that (1) the deficiency chromosome Df(2L)ED784, a deletion of the Prosα6T gene region, fails to complement this mutant,and (2) the male sterility can be completely rescued by the Prosα6T-GFP transgene(Fig. 3B).

Homozygous Prosα6T¹ mutant testes are of normal size and shape, but there are no mature sperm in the seminal vesicle(Fig. 4A-D). Phase-contrast microscopy of testis squashes showed that all stages up through spermatid elongation were indistinguishable in Prosα6T¹ mutants and wild type(Fig. 4E-L). However, there appeared to be defects in sperm individualization, and few, if any, mature,motile sperm were evident. The phenotype of Prosα6T¹/Df(2L)ED784 hemizygotes is indistinguishable from that of Prosα6T¹homozygotes, suggesting that this mutant is a strong loss-of-function allele(not shown).

The presence of proteasome-associated speckles in the IC of elongated sperm bundles, together with the observation that the Prosα6T¹ mutant testes are unable to process elongated sperm bundles to mature sperm, suggest that at least one major function of the proteasome during spermatogenesis is to facilitate sperm individualization. To investigate this possibility, we stained wild-type and Prosα6T¹ testes with Alexa Fluor phalloidin to visualize the actin cones in the IC.

In wild-type testes, actin cones have a characteristic triangular shape and they move in a coordinated fashion (Fig. 5A,B), with a few actin cones only occasionally falling out of synchronization (Fig. 5B,arrow). This fine coordination is disrupted in the Prosα6T¹ mutants, and becomes more extreme as the actin cones move further down the spermatid bundles. The initial assembly of the actin cones, and their triangular shape are not affected in the mutants, indicating that the movement, and not the formation, of the IC,is malfunctioning (Fig. 5C,D). In addition, the proteasome-related speckles that are seen in the IC's of wild-type testes are absent in the Prosα6T¹mutants (Fig. 5E,F).

The pattern of proteasome speckles associated with the IC is very reminiscent of similar speckles detected when testes are immunostained for the active form of the apoptotic protein Dronc (see Huh et al., 2004). In that study, the punctate pattern of active Dronc staining was found localized within the IC, with the speckles first appearing ahead of the actin cones, and then transitioning to a position behind them as the IC progresses caudally along the spermatid bundle during individualization. This dynamic is very similar to what we see here for the proteasomal speckles.

In an attempt to confirm if the proteasome-related speckles are the same as the previously described Dronc speckles, and to test the hypothesis that the activation of Dronc requires the function of testis-specific proteasomes, we used anti-active Dronc antiserum to immunostain testes from wild type, Prosα3T-GFP, Prosα6T-GFP and Prosα6T¹ flies. In wild-type testes,although we could not reproduce the speckled pattern of immunostaining as found by Huh et al. (Huh et al.,2004), probably because of the loss of the effectiveness of the anti-active Dronc antiserum over time in storage (B. Hay and J. Huh, personal communication), we did observe active Dronc staining just ahead of the IC in sperm bundles undergoing individualization(Fig. 6A). Importantly, this staining is absent in the Prosα6T¹ testes,suggesting that the testis-specific proteasome is required for the activation of Dronc (Fig. 6B). In wild-type testes, active Dronc is responsible for activating the downstream effector caspase Ice (Huh et al.,2004). To test whether Prosα6T¹is required for Ice activation, we immunostained wild-type and mutant testes with the CM1 antibody, which is specific for the active form of Ice(Dorstyn et al., 2002). As expected from the previous results of Arama et al.(Arama et al., 2003), in wild-type testes, active Ice is prominently observed in the cystic bulge and uniformly along non-individualized spermatid bundles(Fig. 6C). By contrast, the signal is significantly reduced, although not eliminated, in mutant Prosα6T¹ testes(Fig. 6D).

The prominent Prosα6T-GFP and Prosα3T-GFP signals in spermatid nuclei suggest that the proteasome might also play a role in the dramatic nuclear morphological changes during spermiogenesis. Normally, the spherical spermatid nuclei elongate and condense to form long, needle-shaped structures during and after spermatid bundle elongation(Lindsley and Tokuyasu, 1980). During this shape change and the subsequent individualization process, nuclei in the same bundle tend to remain in register(Fig. 7A). However, in Prosα6T¹ mutants, some nuclei do not fully condense, some are scattered, and some lose the rigid, needle-like shape and appear curled, and in many cases, contorted into circles(Fig. 7B,C). Although there are a few normal looking sperm nuclear bundles in the Prosα6T¹ mutant testes, their number decreases dramatically in older males. In accordance with this, the number of scattered nuclei increases with time (Fig. 7D).

In addition to the morphological changes of nuclei during spermiogenesis,the spermatid chromatin undergoes dramatic molecular changes as well. Some, if not all, of the somatic histones become replaced by the small, highly basic protamines, ProtA and ProtB (also known as Mst35Ba and Mst35Bb, respectively -FlyBase), and a sperm-specific linker histone-related protein Mst77F, during the process of sperm nuclear elongation and condensation(Jayaramaiah Raja and Renkawitz-Pohl,2005; Rathke et al.,2007). The timing of the histone-protamine transition overlaps the expression patterns of Prosα3T and Prosα6T, raising the possibility that the somatic histones are degraded by testis-specific proteasomes to set the stage for recruiting the protamines and Mst77F to the sperm chromosomes. To investigate whether this transition is disrupted in the Prosα6T¹ mutant, a GFP-tagged histone His2AvD transgene (Clarkson and Saint,1999) was crossed into the Prosα6T¹ mutant background. In testes of heterozygous wild-type controls, the H2AvD-GFP signal is prominent in spermatocyte and spermatid nuclei of pre-elongation stages of spermatogenesis,but fades away in the elongating and condensing spermatid heads, and completely disappears in fully condensed sperm nuclei(Fig. 8A-F). The overall pattern of H2AvD-GFP expression in homozygous mutants is similar to that in the wild type, indicating that there is no dramatic alteration in histone replacement. However, there is a slight, but noticeable, delay in the disappearance of the histone signal in the mutant. That is, the H2AvD-GFP signal is seen to persist longer in the elongating and partially condensed nuclei of the mutant (compare Fig. 8D with J), although it does eventually disappear as the nuclei reach full condensation, and it is not seen at all in the abnormal curled nuclei(Fig. 8K-L).

Does this delay in histone removal affect the integration of protamines into the chromosomes? To address this question, a ProtA-GFP transgene was introduced into the Prosα6T¹ mutant background. As expected, ProtA-GFP in control heterozygotes is first detected in the elongating nuclei, and it remains prominently visible in the individualized, mature sperm nuclei (Fig. 8M-R). In homozygous Prosα6T¹mutants, the ProtA-GFP signal is also first seen in the elongating nuclei, and it remains in the scattered and the curled nuclei(Fig. 8S-X). Because of the delay in histone disappearance, there is a more extensive overlap of expression of H2AvD-GFP and ProtA-GFP in mutant testes, but the Prosα6T¹ mutant has no significant effect on the timing and degree of protamine incorporation. Given these results,there does not appear to be a strong link between the sperm nuclear shape defects of the Prosα6T¹ mutant and the histone-protamine transition.

Given the different expression patterns of the Prosα6 and Prosα6T subunits, one question that arises is: does the testis-specific subunit result in proteasomes that are functionally distinct from proteasomes containing the conventional isoform? To address this, transgenic flies were produced that ectopically express the Prosα6 subunit in late spermatogenesis, in flies that lack the testis-specific Prosα6T subunit. The testis-specific Prosα6 transgene construct(pW8/TS1-Prosα6) was created by replacing the coding region of the cloned testis-specific proteasome gene Prosα3T with the coding sequence of Prosα6 (see Materials and methods), and the transgene was introduced into the Prosα6T¹ mutant background by P-element transformation. It was found that driving testis-specific expression of Prosα6 could completely rescue the sterility of Prosα6T¹ mutant males, and microscopic analysis of spermatogenesis in such flies showed no differences when compared to that of wild type. This result demonstrates the Prosα6 and Prosα6T subunits are functionally similar.

One previously unanswered question concerns the functional significance of testis-specific proteasomes. The present study demonstrates, at least for Prosα6T, the testis-specific subunit is required for normal sperm development. Examination of the phenotypic effects of the knockout allele, Prosα6T¹, demonstrates essential roles for this gene in post-meiotic sperm cell differentiation, most notably in the processes of spermatid individualization and nuclear differentiation,consistent with the observed expression pattern of this subunit. Prosα6-GFP can be detected throughout spermatogenesis, from early spermatogonia to elongating spermatids, but it begins fading away after meiosis, and completely disappears by the end of spermiogenesis. By contrast,the testis-specific subunit Prosα6T-GFP is not detected in early stages until meiosis, after which it becomes prominent in both nuclei and cytoplasm. In elongated spermatids, it persists in the bundled sperm heads and tails,although the GFP signal in nuclei is reduced once nuclear elongation is complete and individualization begins.

The role of the proteasome in sperm individualization may be related to its localization to numerous small `speckles' within the IC, trailing the actin cones as they move along the bundle. We suggest there is a complex of protein degrading machinery that helps in the removal of unneeded cytosolic material,or assists in the release of actin so that movement of the IC can proceed. In Prosα6T¹ mutants, where this does not occur efficiently, the actin cones soon lose their synchronous movement and the IC breaks down, thereby stalling individualization.

In addition to this general requirement of proteasome-mediated protein degradation during sperm individualization, there might be a specific role of proteasomes in mediating the activation of caspases to trigger the spermatogenesis-specific apoptotic pathway that facilitates the individualization process. It has been reported that the active form of caspase Dronc shows a punctate staining pattern associated with the IC, very similar to the proteasome speckles we have observed(Huh et al., 2004). Dronc is activated when bound to Dark, a homolog of the mammalian Apaf1 protein,triggering its autoprocessing (Muro et al., 2004). Normally, its activity is held in check by binding to the inhibitory protein Diap1 [also known as Thread (Th) - FlyBase]. During apoptosis, Diap1 is autoubiquitinylated, through the action of RHG [Rpr, Hid(also known as Wrinkled) and Grim] proteins, and is presumably degraded via the ubiquitin-proteasome pathway (Hays et al., 2002; Holley et al.,2002; Ryoo et al.,2002; Wing et al.,2002; Yoo et al.,2002). Dronc is thus released from the tether of Diap1 and able to cleave and activate the downstream caspases such as the caspase-3-like effector caspase Ice.

In recent years, increasing evidence has indicated that spermatogenesis borrows the apoptotic machinery to achieve non-apoptotic results(Cagan, 2003). Key apoptotic components such as active Dronc, active Ice, and Hid proteins have been found in elongated spermatids (Arama et al.,2003; Huh et al.,2004), and a Ice deletion mutant is defective in spermatid individualization (Muro et al.,2006). It has been shown that Dark- and Hid-dependent activation of Dronc happens in the cystic bulges where individualization takes place(Huh et al., 2004). Furthermore, when Diap1 or dominant-negative (Dn) Dronc is overexpressed using the testis-specific β₂-tubulin promoter, the active Dronc speckles disappear, individualization is defective, and the synchronized movement of the actin cones is disrupted, producing a phenotype that is very similar to what is observed here with the Prosα6T¹ mutant(Huh et al., 2004). These observations, along with the fact that Diap1 is ubiquitinylated during apoptosis (Wilson et al.,2002; Yoo et al.,2002), raise the possibility that testis-specific proteasomes play an important role in degrading DIAP1, which, in turn, triggers Dronc to activate downstream caspases, thus initiating the individualization process. Our observations that in the Prosα6T¹mutant, active Dronc disappears from the IC and active Ice is significantly reduced is consistent with this model. The fact that active Ice is not completely eliminated suggests that there might be an additional Dronc-independent pathway for Ice activation, as has been suggested by Huh et al. (Huh et al., 2004).

In Prosα6T¹ mutant testes, coordinated actin cone movement is disrupted. A similar phenotype has been documented for a number of cytoskeletal protein mutations, including jar¹(myosin VI), dtctex-1¹ (Dlc90F - FlyBase; Drosophila 14 kDa dynein light chain), ddlc1^ins(ctp - FlyBase; 8 kDa Drosophila dynein light chain 1), sw¹ (74 kDa intermediate chain) and MyoV^Q1052st (didum - FlyBase; myosin V)(Hicks et al., 1999; Li et al., 2004; Ghosh-Roy et al., 2005; Mermall et al., 2005). The spermatid nuclear bundles in these mutants are also disrupted. Although the mechanism of the actin cone movement has not been fully elucidated, it has been proposed that actin polymerization and depolymerization are the driving force (Noguchi and Miller,2003). Myosin VI seems to be an actin cone structure stabilizer(Noguchi et al., 2006), and myosin V is required for actin cone formation(Mermall et al., 2005). Here,we demonstrate that the ubiquitin-proteasome pathway is also important in IC movement. It is unclear at this time how proteasomes affect actin cone movement, but it is possible that they facilitate the movement by regulating the protein levels of some key components of the actin cone movement machinery. Alternatively, they could achieve this goal through regulation of the apoptosis pathway, as discussed above. Evidence for the latter possibility is that Dcp1, a caspase activated by Dronc, is required for the cytoskeletal reorganization that occurs during the dumping process of Drosophilaoogenesis (McCall and Steller,1998). Another possibility is that proteasome-mediated degradation of numerous cytosolic proteins might achieve the breakdown of cytoplasmic structure within the IC and facilitate the movement of actin cones as they push the digested material into the growing waste bag.

We observed that in Prosα6T¹ mutants,whereas the histone H2AvD does eventually disappear in the scattered and curled spermatid nuclei, this displacement is slightly delayed. One possible explanation for this delay is that during spermatogenesis the histones may be degraded via the Ub-proteasome pathway to facilitate their removal and replacement by the protamines. In the Prosα6T¹ mutant, the Ub-proteasome pathway may be compromised at this late stage of spermatogenesis, thus inhibiting histone replacement. One possibility is that the mutation affects the removal of the histones from the chromatin. Alternatively, the histones might be removed normally, but then not efficiently degraded.

The incorporation of protamines does not appear to be affected in the Prosα6T¹ mutant nuclei, indicating that the malformed sperm nuclei are not the result of a gross disruption of the histone-protamine transition. From these results, it is not clear why the sperm nuclei exhibit aberrant shapes in the mature spermatid bundles. One possibility is that the abnormal nuclear morphology arises at the end of spermiogenesis, when sperm undergo coiling and abnormal sperm are screened for elimination (Fuller, 1993). In this case, there would not be a direct link between this phenotype and reduced proteasome function in the nucleus.

The proteasome genes of D. melanogaster can be divided into those encoding subunits that are expressed throughout development and those whose expressions are limited to the late stages of spermatogenesis. The driving force for the evolution of this situation might be related to the unusual dynamics of gene expression during spermatogenesis(Fuller, 1993; Hiller et al., 2004). That is,because transcription of virtually all genes ceases after the primary spermatocyte stage, and because there is a very large amount of protein degradation occurring during the later stages of spermiogenesis, there might have been strong selective pressure to increase the levels of proteasome subunit expression in these cells. Gene duplication is one way to immediately increase the level of gene expression, and once it happens, one copy would be free to evolve a testis-specific expression pattern to ensure adequate levels of the subunit throughout spermatogenesis. The duplicated gene might then evolve to produce proteasome subunits that are better adapted to carry out the extensive protein breakdown that occurs as the spermatids eliminate their cytoplasm during the individualization process. Importantly, the observation reported here that driving spermatogenic expression of the conventional subunit Prosα6 can functionally substitute for Prosα6T rules out the possibility that the somatic and testis-specific forms of the proteasome have completely distinct functional properties. This situation is analogous to that of the spermatogenic-specific isoform of cytochrome C (Cyt-c-d), known as Blanks (Bln; FlyBase) (Arama et al.,2006). In that case, it was shown that whereas bln^- mutants are male sterile, the mutant phenotype can be rescued by driving testis-specific expression of the somatically expressed paralogous gene, Cyt-c-p.

We thank Jing Ma and Evan Katz for constructing the Prosα6T-GFP and Prosα6-GFPreporters, and Sheree McClear for her assistance in generating the fluorescently tagged protamine construct. We are indebted to Kent Golic for providing the targeting vector pTV2 and several of the stocks used for the mutagenesis procedure, and to Yikang Rong for advice. We thank Bruce Hay and Jun Huh for the anti-active Dronc antibody, and Maya Bader and Hermann Steller for the CM1 antibody. We also thank Xiazhen Li and Jing Dai for helpful discussions, and Brian Calvi for comments on the manuscript. We thank Changyu Shen for his help. This work was supported by a grant to J.M.B. from the National Science Foundation (MCB-0416647).