4, 687697, May, 2003, Copyright 2003 by Cell Press
Caspase Activity and a Specific Cytochrome C Are Required for Sperm Differentiation in Drosophila in spermatogenesis, cellular analyses have so far fo- cused on the early stages of sperm development (Knud- son et al., 1995; Rodriguez et al., 1997; Print et al., 1998; Eli Arama, 1 Julie Agapite, 2 and Hermann Steller 1, * 1 Howard Hughes Medical Institute Strang Laboratory of Cancer Research Honarpour et al., 2000; Furuchi et al., 1996). Therefore, The Rockefeller University little is known about the exact role of apoptosis during 1230 York Avenue sperm cell terminal differentiation. New York, New York 10021 Postmeiotic spermcell differentiation is characterized 2 Department of Biology by a sequence of conserved changes in the morphology Massachusetts Institute of Technology and structural organization of the spermatids. In Dro- 77 Massachusetts Avenue sophila, these changes include elongation of the flagel- Cambridge, Massachusetts 02139 lum, fusion and subsequent elongation of the mitochon- dria to the entire length of the spermatid, nuclear condensation, and the expulsion of most of the sperma- Summary tid cytoplasm (Fuller, 1993). Although spermatid bulk cytoplasm elimination was already described more than The final stage of spermatid terminal differentiation three decades ago, the molecular and cellular mecha- involves the removal of their bulk cytoplasm in a pro- nisms that drive this process are still unknown (Fawcett cess known as spermatid individualization. Here we and Phillips, 1969; Tokuyasu et al., 1972). show that apoptotic proteins play an essential role We have investigated the role of apoptotic proteins during spermatid individualization in Drosophila mela- during Drosophila spermatogenesis. We find that cas- nogaster. Several aspects of spermterminal differenti- pase activity is required for spermatid individualization. ation, including the activation of caspases, are remi- More generally, cystic bulges and waste bags, which niscent of apoptosis. Notably, caspase inhibitors contain cytoplasmexpelledfromdifferentiatingsperma- tids, display many features of apoptotic corpses and prevent the removal of bulk cytoplasm in spermatids contain all the components required for a functional and block spermmaturation in vivo, causing male ste- apoptosome. Activation of the effector caspase drICE rility. We further identified loss-of-function mutations in spermatids requires one of the two cytochrome c in one of the two Drosophila cyt-c genes, cyt-c-d, which genes of Drosophila, cyt-c-d. Interestingly, loss-of-func- block caspase activation and subsequent spermatid tion mutants for cyt-c-d are male-sterile but viable and terminal differentiation. Finally, a giant ubiquitin-con- have no other apparent defects in caspase activation, jugating enzyme, dBruce, is required to protect the apoptosis, cell vitality, and development. Therefore, cyt- spermnucleus against hypercondensationanddegen- c-d appears to be specifically required for caspase acti- eration. These observations suggest that an apopto- vation during spermatogenesis. Finally, loss-of-function sis-like mechanism is required for spermatid differen- mutants in dBruce, a giant ubiquitin-conjugating en- tiation in Drosophila. zyme (Chen et al., 1999; Hauser et al., 1998; Vernooy et al., 2002), have spermatid nuclei that are hyperconde- nsed and degenerate, leading to male sterility. These Introduction findings are consistent with the idea that dBruce is re- quired to protect the spermnucleus fromexcessive cas- Apoptosis is amorphologically distinct formof cell death pase activity and degeneration. that usually serves to remove unwanted and potentially dangerous cells (Meier et al., 2000a; Hengartner, 2000). Results A key event during apoptosis is the activation of a spe- cific class of cysteine proteases, termed caspases Spermatid Individualization (Thornberry and Lazebnik, 1998; Cryns and Yuan, 1998). Spermatogenesis in Drosophila melanogaster takes place Caspases are expressed as inactive zymogens in virtu- within individual units known as cysts. Each cyst con- ally all animal cells and are specifically activated in cells tains 64 spermatids that remain initially connected after destined to undergo apoptosis. On the other hand, there meiosis via cytoplasmicbridges anddifferentiate synchro- are some examples where apoptosis-like events do not nously. During terminal differentiation, the round-shaped lead to the death, but rather the terminal differentiation spermatids are transformed into thin, approximately 2 of certain cell types. For example, lens epithelial cells mm long spermatozoa with highly elongated, needle- and mammalian red blood cells lose their nucleus and shaped nuclei. In the final stage of spermatogenesis, other subcellular organelles during terminal differentia- termed individualization (Figure 1A), the cytoplasmic tion but continue to be metabolically active (Jacobson bridges are disconnected and most of the cytoplasm is et al., 1997). Likewise, spermcell terminal differentiation expelled, leading to individual sperm. The individualiza- involves complex changes in the cytoarchitecture, some tion process involves the assembly of a cytoskeletal- of which are reminiscent of apoptosis. Although mouse membrane complex, referred to as the individualization mutations in several apoptotic genes produce defects complex (IC), which contains actin as its major cy- toskeletal component. The IC can be detected by stain- ing with phalloidin, which binds to actin (Figures 1B and *Correspondence: [email protected] Developmental Cell 688 Figure 1. The Removal of Bulk Cytoplasm during Sperm Differentiation Shares Similari- ties with Apoptosis (A) Schematic diagramof spermatid individu- alization. In this diagram, a single cyst with four representative spermatids is depicted at increasingly advanced developmental stages (from top to bottom). The first cyst contains elongated spermatids prior to individualiza- tion. In the second cyst, assembly of the indi- vidualization complex (IC, red) has begun around the nuclei. The ICprogresses caudally from the nuclear region to the end of the tail (in this figure from left to right). The third cyst has a cystic bulge (CB, red), which contains the progressed IC and the cytoplasm of the postindividualized part of the spermatids. At this point, many vesicles and organelles are present in the CB. Finally, most of the sper- matids syncytial cytoplasm is removed into the waste bag (WB, red). The WB is pinched off from the base of the cyst and eventually degrades. Each sperm is now surrounded by its own membrane and is largely devoid of cytoplasm and organelles (with the excep- tion of a single giant mitochondrion). Adapted from Fabrizio et al., 1998. c.c., cyst cell; ind., postindividualized; preind., preindividu- alized. (BG) Both actin and lamin Dm 0 are compo- nents of the IC. In squashed testes prepara- tions, the IC can be visualized either by TRITC-phalloidin (red) that binds to actin, or by anti-lamin Dm 0 (green) that detects the ma- jor protein of the nuclear lamina. Bundles of wild-type elongated spermatids (blue) are de- picted either before IC assembly (B), associ- ated with an IC ([B and D], IC in pink), or after the IC translocation ([E], IC in green). The CBs size increases as the spermatid in- dividualization progresses; small CB is de- tected right after the detachment of the IC fromthe vicinity of the nuclei ([E], ICin green). As the IC progresses, the CB increases invol- ume ([C], IC in red, and [F], IC in green and red). Finally, the CB reaches the end of the spermatids, becoming a large WB ([G], IC in green). (HJ) Micrographs of live testes stained with the apoptotic marker acridine orange (AO, green). AO, a specific marker for apoptotic cells in Drosophila, specifically stained both CBs (H) and WBs (I). (J) Scattered AO-positive spots at the base of the testis represent degrading WBs (arrowheads). The scale bars represent 50 m. 1C, red; Fabrizio et al., 1998). In addition, we observed CB and WB with AO in spermatids undergoing individu- alization suggests that the apoptotic program is acti- that lamin Dm 0 leaves the vicinity of the nucleus and translocates as a component of the IC and thus can vated in the late stages of sperm differentiation, and that CB and WB resemble apoptotic corpses without also be a useful marker of the IC (Figures 1D1G, green). The IC is assembled at the nuclear end of the cyst and nuclei. subsequently translocates caudally along the entire length of the spermatid bundle, expelling most of the Activation of an Effector Caspase at the Onset of Spermatid Individualization cytoplasm in the process. The discarded cytoplasm ac- cumulates in a membrane-enclosed structure, termed A key feature of apoptosis is the activation of caspases. To investigate whether Drosophila effector caspases the waste bag (WB; Figures 1A, 1G, and 1I; Tokuyasu et al., 1972). The WBs eventually undergo fragmentation become activated during the individualization process, testis preparations were stained with the CM1 antibody. and subsequent degradation (Figure 1J). To investigate the possible occurrence of apoptosis CM1 was raised against the active form of mammalian caspase 3 but also crossreacts with the caspase-3-like during Drosophila sperm differentiation, we stained live wild-type testes with the vital dye acridine orange (AO), Drosophila effector caspase drICE (Baker and Yu, 2001; Srinivasan et al., 1998; Yu et al., 2002; Ryoo et al., 2002). which specifically detects apoptotic cells (Abrams et al., 1993). AO staining was observed in intact cystic bulges In addition, we also used an antibody that was raised specifically against the active form of drICE (Dorstyn et (CBs) and WBs (Figures 1H and 1I). The staining of the Role of Apoptotic Factors in Spermatogenesis 689 Figure 2. Activation of the Effector Caspase drICE at the Onset of Individualization (A) and its enlargement in (B). The assembly of the IC (phalloidin, red) at the nuclear end (DAPI, blue) of a cyst, which marks the begin- ning of the process of spermatid individual- ization, is followed by a steep gradient of drICE activation (CM1, green). Active drICE accumulates in the preindividualized part of a cyst (arrows) and within the CB. After the caudal translocationof the CB, active drICEis no longer detectable in the postindividualized part of the spermatids (arrowheads). C, elon- gated spermatid cysts. (CF) The activation of drICE during sperma- tid individualization was detected either by the anti-active mouse caspase 3 (CM1) anti- body that crossreacts with drICE ([C and D], green), or by the anti-active drICE antibody ([E and F], green). Both antibodies gave rise to intense staining in the cytoplasmic parts of the spermatids, namely the CB and distal parts of the cell. Once all the cytoplasm is stripped away, active drICEis detectable only in the WB. The scale bars represent 50 m. al., 2002; Yoo et al., 2002). Using either antibody, active appeared flat (Figure 3B) as compared to the full, oval shape of CBs in the control (Figure 3A). This suggests drICE became detectable immediately after a mature IC was formed (Figures 2A and 2B). During the caudal that caspase activity is required for the efficient collec- tion of cytoplasminto the CBs. Furthermore, the translo- translocation of the IC, active drICE became completely depleted from the newly individualized portion of the cation of the ICs appearedtobedefective inthe absence of caspase activity. In controls, only a fewearly ICs were spermatids. The staining remained abundant, however, in the preindividualized portion of the spermatids, with associated with the nuclear complexes because the IC translocates caudally in the process of spermatid differ- the highest levels within the CB (Figures 2B, 2C, and 2E). At the end of this process, the newly formed WBs entiation (Figure 3C). In contrast, in the presence of a caspase inhibitor, most of the ICs remained in the vicin- contained high levels of active drICE (Figures 2D and 2F). These observations indicate that caspase activation ity of the nuclei (Figure 3D), indicative of a defect in IC translocation. Therefore, caspase activity appears to be and apoptosis-like events are intimately associated with the terminal differentiation of spermatids in Drosophila. required for proper IC movement and the removal of bulk cytoplasm from differentiating spermatids. In order to extend these observations to spermato- Caspase Activity Is Required for Spermatid Individualization genesis in vivo, we expressed the baculoviral broad- spectrumcaspase inhibitor p35 (Hay et al., 1994) in male In order to examine whether caspase activation is nec- essary for spermatid differentiation, we inhibited cas- gonads using the GAL4-UAS system (Brand and Perri- mon, 1993). Transgenic flies were generated that carry pase activity both in cultured testes and in vivo. Under in vitro culture conditions, each cyst accomplishes indi- the GAL4 gene fused downstream of the Drosophila Hsp83 promoter, which directs strong expression in the vidualization within 12 hr (Noguchi and Miller, 2003). To test the effect of caspase inhibition on the caudal germ cells throughout spermatogenesis (Yue et al., 1999). The Hsp83-Gal4 driver line was crossed to the movement of the IC in vitro, testes from young adults were isolated. One testis of a pair fromthe same individ- UAS-p35 line and male progeny were subjected to a fertility test. Whereas the parental lines that bear two ual was cultured in the presence of the caspase inhibitor Z-VAD, while the other testis served as a control. Z-VAD copies of either construct alone were fertile, males with both the driver and the UAS-p35 construct were sterile. was previously shown to block the activity of drICE in vitro and drICE-inducedcell death of Drosophila S2 cells Cytological analyses of these males revealed that the translocation of the IC was severely impaired (Figure (Fraser and Evan, 1997). After 26 hr of culture, testes were fixed and analyzed. In testes that were cultured 3E). Although the initial assembly of the IC appeared normal (datanot shown), therewas afailureof cytoplasm with the caspase inhibitor, the already advanced CBs Developmental Cell 690 to collect into a normal CB (Figure 3E). These observa- tions demonstrate that caspase activity is required for proper individualization, cytoplasmic exclusion, and male fertility. Caspase Activation at the Onset of Spermatid Individualization Is Independent of Terminal Differentiation To distinguish whether caspase activation is the cause or the consequence of sperm cell differentiation, we examined mutants that disrupt the individualization pro- cess. In purity of essence (poe 1 ) spermatids, the ICs are significantly reduced (arrowheads in Figure 4A), and oftenare not detectedat all (arrows inFigure 4A; Fabrizio et al., 1998). However, active caspases are clearly de- tected in elongated spermatids that lack the IC, sug- gesting that the activation of caspases at the onset of individualization is independent of IC assembly. In jag- uar (jar 1 ) male flies, IC assembly and movement are impaired due to a mutation in a class VI myosin (Hicks et al., 1999). Interestingly, in jar 1 flies, active drICE could only be detected in mature cysts, which contain needle- shaped and condensed nuclei (Figure 4B). Therefore, the completion of spermatid morphogenesis rather than the assembly of a functional ICtriggers drICE activation. To test whether caspase activation is dependent on the completion of nuclear morphogenesis, we examined testes of fuzzy onions (fzo) mutant flies. fzo flies are defective in the process of mitochondrial fusion during spermatogenesis (Hales and Fuller, 1997), and hence they arrest at mid-spermatid elongation stage. Active drICEaccumulatedin several immature elongatingsper- matid cysts containing undifferentiated nuclei (Figure 4C). Thus, we conclude that drICE activation is indepen- Figure 3. Inhibition of Caspase Activity Blocks Individualization In dent of nuclear maturation. Vitro and In Vivo Under in vitro culture conditions, spermatids become individualized within 12 hr. Two testes from the same wild-type individual were Components of the Apoptosome Are Expressed cultured for 26 hr under identical conditions, except that the caspase during Spermatogenesis inhibitor Z-VAD was added to one of them. Nuclei were visualized In mammals, mitochondria play an important role in the with DAP1 (blue), ICs with phalloidin (red), and active drICE with the activation of the apoptosome, a multiprotein complex CM1 antibody (green). Whereas testes that were cultured without Z-VAD displayed the typical oval appearance of CBs (A), the treated that includes caspases, Apaf-1, and cytochrome c testes displayed flat, irregular CBs, indicating that the bulk cyto- (Salvesen and Renatus, 2002). Because orthologs of plasm was not collected properly in the CBs (B). caspase-9, Apaf-1, and cytochrome c have been identi- (C) After 26 hr of culture, most of the mature cysts underwent individ- fied in Drosophila (Meier et al., 2000b; Zhou et al., 1999), ualization in the absence of Z-VAD (note that only a few ICs remain we examined whether they are expressed during Dro- associated with nuclei at the base of the testis). sophila spermatogenesis. Using an antibody against (D) Testes that were cultured in the presence of Z-VAD displayed many early stages of IC assembly, but no evidence of IC transloca- full-length Dronc, a caspase-9 homolog (Meier et al., tion, indicating that caspase activity is necessary for the movement 2000b), revealed that it was expressed along the length of the IC. of elongated spermatids (arrowheads in Figures 5A and (E) Ectopic expression of the baculoviral broad-spectrum caspase inhibitor p35 blocks individualization. Flies containing two copies of HSP83-Gal4 and at least one copy of UAS-p35 were sterile. Stain- ing of their testes for ICs (red) andactive drICE(CM1, green) revealed severe defects in individualization, including scattered ICcones and called WB, whereas in mammals the spermatid cytoplasm ends up flat CBs and WBs. Note that inhibition of drICE by p35 does not in a structure named residual body (RB). On the right is a scanning block CM1 staining, as previously reported (Baker and Yu, 2001; Yu electron micrograph (SEM) of a human spermatozoon that displays et al., 2002; Ryoo et al., 2002). a cytoplasmic droplet (CD) morphological abnormality. This abnor- (F) Overall similarities of sperm differentiation in Drosophila and mality is characterized by the failure to eliminate spermatid cyto- mammals. On the left is a schematic presentation of the last step plasm, a phenotype that is very similar to the individualization de- of the individualization process in Drosophila and mammals (rat). fects seen in (B), (D), and (E). Reproduced from Hollanders and Like in Drosophila, mammalian spermatids are also connected by Carver-Ward, 1996 with permission from Parthenon Publishing, UK. cytoplasmic bridges, which are eliminated together with most of N, nucleus; WB, waste bag; RB, residual body. the cytoplasm and organelles during terminal differentiation. In Dro- The scale bars in (A)(E) represent 50 m; in (F), the scale bar sophila, the cytoplasm of 64 spermatids accumulates in a structure represents 1 m. Role of Apoptotic Factors in Spermatogenesis 691 Figure 5. Expression of the Drosophila Orthologs of the Apopto- some Complex during Spermatogenesis Nuclei are stained blue, and the IC is either pink or yellow. (A) Dronc (green), a caspase-9 ortholog, accumulates along fully elongated spermatid cysts (arrowheads) and is depleted from indi- vidualized spermatozoa (arrow). Figure 4. Caspase Activation Is Independent of Nuclear Differentia- (B) As the ICprogresses caudally, Dronc expressionincreases within tion or IC Assembly the CB (arrowhead). (C) The Drosophila ortholog of Apaf-1, hac-1, is expressed in sper- Mutants that disrupt sperm differentiation retain CM1 staining. matocytes. Transcription of hac-1 was visualized using the (A) purity of essence (poe 1 ) spermatids contain significantly fewer l(2)k11502 line containing a P-lacZ insertion in the hac-1 promoter. ICs (arrowheads), and often the ICs cannot be detected at all In this line, lacZ expression (green) mimics the mRNA pattern of (arrows). Although caspase activation normally occurs only after IC hac-1 (Zhou et al., 1999). assembly, CM1 staining (green) can be readily detected in elongated (D) Full-length inactive drICE (green) is expressed uniformly in prein- spermatids that lack an IC. dividualizing spermatid cysts (arrowheads), demonstrating that pro- (B) In jaguar (jar 1 ) male flies, IC assembly and movement are im- drICE is expressed in early stages of spermatogenesis. However, paired. Active drICE (green) accumulates independently of func- drICE becomes activated by posttranslation modification only after tional ICassembly in jar 1 testes, but is correlated with the differentia- the assembly of the IC. The scale bar represents 50 m. tion stage of nuclei (blue). The black and white panels of nuclear bundles on the right correspond to the blue nuclei on the left; cas- pase activation was only detected in spermatid bundles with thin, fully differentiated nuclei (see asterisks). in testes of the enhancer-trap line l(2)k11502, which (C) fzo spermatids arrest at the mid-spermatid elongation stage and mimics the mRNA pattern of expression of hac-1 (Zhou do not complete nuclear condensation. In fzo 1 /fzo 2 mutants, active et al., 1999). As shown in Figure 5C, hac-1 accumulated drICE (green) accumulates in the more advanced elongating sper- in late primary spermatocytes. Finally, antibody staining matid cysts despite the immature nuclei (blue), demonstrating that for full-length drICE revealed that pro-drICE accumu- drICE activation can occur independently from nuclear differenti- lated in primary spermatocytes and was uniformly dis- ation. The scale bars in (A) and (C) represent 50 m; in (B), the scale bar tributed in preindividualized elongated spermatids (data represents 25 m. not shown and Figure 5D). Interestingly, a significant proportion of cellular Dronc and drICE appears to local- ize near the mitochondria (Dorstyn et al., 2002). These 5B). To investigate whether the Drosophila Apaf-1 ho- molog hac-1/dark/dapaf is expressed during spermato- observations suggest that Drosophila apoptosomes may form at the surface of mitochondria. However, the genesis, we examined the -galactosidase distribution Developmental Cell 692 Figure 6. Molecular Genetic Analysis of the Drosophila Cytochrome C Genes (A) Genomic organization of the two cyto- chrome c genes and transposon insertions. The map illustrates the cyt-c-d and cyt-c-p exons (thick lines in black and gray), introns (thin lines), and insertion points of the P ele- ment transposons in strains EP(2)2305, EP(2)2049, bln 1 , and l(2)k13905 (triangles). Three independent insertions were identified in the 5 UTR of cyt-c-d (exon-1). The p[Z] element of the bln 1 allele is inserted 2 bp up- stream of the last nucleotide of exon-1, while the EP 2305 and EP 2049 insertions are 83 bp and 77 bp upstream of the last nucleotide of exon-1. The predicted ORF for cyt-c-d (ATG- TAG within exon-2) encodes a protein 105 amino acids long. The cyt-c-p gene is located 241 bp downstreamof cyt-c-d. The predicted ORF for cyt-c-p (ATG-TAA within exon-2) en- codes a protein of 108 amino acids. The l(2)k13905 allele is a P-lacW insertion in the first nucleotide of the cyt-c-p intron. (B) Transcript analysis of the cyt-c-d gene. Northern blot analysis of adult male poly(A) RNA of wild-type and bln 1 mutant. A probe unique for the 3 UTR region of cyt-c-d de- tected a single band of the predicted size for cyt-c-d (0.87 kb) in wild-type (wt, right lane), but not in the bln 1 mutant (left lane). The Dro- sophila rp49 gene was used as a loading con- trol marker. role of cytochrome c for caspase activation in Dro- revealed that the bln 1 allele complemented the lethality sophila has remained unclear, as no release from mito- of the l(2)k13905 allele, and that the latter complemented chondria was seen in previous studies, and RNAi ex- the sterility of bln 1 . l(2)k13905/DF(2L)H20 trans-hetero- periments in SL-2 cells produced negative results as zygotes are also recessive embryonic lethal, demonstra- well (Dorstyn et al., 2002; Varkey et al., 1999; Wang, ting that this phenotype is due to the loss of cyt-c-p 2001; Zimmermann et al., 2002). function. Significantly, one of the cyt-c-d alleles, blanks (bln 1 ), was previously reported as a male-sterile mutant (Castrillon et al., 1993). Detailed phenotypic analyses Caspase Activation Requires a Specific (see below) revealed that these various insertions and Cytochrome C trans-heterozygous combinations can be arranged into To critically examine the requirement of cytochrome c an allelic series, with bln 1 behaving as a genetic null. for caspase activation and sperm cell differentiation in Furthermore, in order to verify that the mutant pheno- Drosophila, weexaminedthe phenotypeof loss-of-func- types are causedby thetransposon insertions, we mobi- tion mutants. Drosophila contains two closely linked but lized the P elements in all three lines and generated distinct cytochrome cgenes, termedcyt-c-dandcyt-c-p revertants (see Experimental Procedures). Finally, North- (Limbach and Wu, 1985). The cyt-c-p gene encodes the ern blot analysis with a probe directed against the unique major form of cytochrome c and is expressed at much 3 UTR of cyt-c-d identified one band of the predicted higher levels than cyt-c-d. Searching FlyBase for possi- size in wild-type, but no transcript was detected in bln 1 ble mutations, we identified three independent P ele- mutants (Figure 6B). Therefore, bln 1 appears to be a com- ment insertions located in cyt-c-d (bln 1 , EP 2305 , and plete loss-of-function allele for cyt-c-d. EP 2049 ), and one insertion in cyt-c-p (Figure 6A). The P Next, we stained mutant testes from all cyt-c-d alleles insertion in cyt-c-p was homozygous lethal, and all three as homozygotes andall thedifferent trans-heterozygous P insertions into the cyt-c-d gene were homozygous combinations for active drICE using the CM1 antibody. viable but male-sterile. All three cyt-c-dalleles produced Strikingly, there was no CM1 immunoreactivity in bln 1 male sterility when placed in trans with each other, and over the DF(2L)H20 deletion. Complementation analysis homozygotes and bln 1 /DF(2L)H20 flies (Figures 7B, 7D, Role of Apoptotic Factors in Spermatogenesis 693 Figure 7. Mutations in cyt-c-d Block Cas- pase Activation at the Onset of Spermatid Individualization (A and B) Visualization of active drICE with CM1 in wild-type (A) and bln 1 testes (B). (A) Whereas CM1-positive cysts at different individualization stages can be readily seen in wild-type testes (black arrows pointing at CBs and WBs), no CM1-positive cysts were detected in testes of flies homozygous for the bln 1 allele (B). (CF) Nuclei, blue; IC, red; active drICE, green. (C) CM1-positive spermatids were easily de- tected in wild-type individualizing cysts (green, white arrow). However, no CM1 stain- ing was detected in bln 1 mutant testis (D) and (F), although bln 1 spermatids showed noobvi- ous morphological defects. Note that the spermatids appeared elongated (F), that the nuclei differentiated and acquired a normal needle-like shape, and that the IC was as- sembled normally ([D], arrowheads). (E) Spermatids of the hephaestus (heph 2 ) mu- tant showed strong immunoreactivity with CM1 despite other severe defects in the indi- vidualization process. Similar results were obtained for ten additional different mutants that show defects in spermatid individualiza- tion process (data not shown). (G) mAB 2G8, which detects cytochrome c only in apoptotic cells, stains the mitochon- dria of round spermatids (arrowheads) and then acquires a punctate expression pattern in elongated spermatids typical of mitochon- drial protein expression at this stage (H). Note that the giant mitochondrion of round sper- matids can be easily visualized by DAPI due to staining of the mitochondrial DNA. The scale bars represent 50 m. and 7F), even though bln 1 testes contained elongated cyt-c-p. Because cyt-c-p has no apparent function in caspase activation/apoptosis but is expressed at much spermatids with an intact IC and with the characteristic needle-shaped nuclei that are typical for advanced higher levels than cyt-c-d, this may explain the previous difficulties in detecting cytochrome c release during stages of spermatid differentiation (arrowheads in Fig- ure 7D). Furthermore, whereas in wild-type the IC trans- apoptosis in Drosophila. Alternatively, it is possible that cytochrome c is only required for caspase activation locates caudally (arrows in Figures 7A and 7C), in bln 1 mutants the ICdoes not separate fromthe nuclei (Figure during spermatogenesis. We also examined 11 other mutants that displayed 7D). Very similar observations were made for the other allelic combinations. EP 2305 and EP 2049 homozygotes severe defects in individualization, and all of them had intense CM1 staining (see, for example, the hephaestus showed some occasional CM1 staining and more pro- gressed ICs, but all allelic combinations were male-ster- mutant in Figure 7E). These results demonstrate that the lack of CM1 staining is not simply the consequence ile, although we observed some occasional escapers in EP 2049 / and EP 2305 /EP 2049 flies (data not shown). All P of defective individualization, but rather reflects a spe- cific requirement of the minor cytochrome c protein element revertants that we generated were fertile and had normal CM1 staining and individualization, demon- in Drosophila. Therefore, we conclude that cyt-c-d is required for caspase activation and for the subsequent strating that the phenotypes reported here are caused by the transposon insertions into the cyt-c-d locus. IC translocation during spermatid individualization. Finally, we stained testes with the mAB2G8 anti-cyto- chrome c antibody (Varkey et al., 1999). As previously Role of dBruce during Sperm Terminal Differentiation Effector caspases, such as drICE, normally cleave a reported for apoptotic cells, we saw increased mAB2G8 immunoreactivity in elongated spermatids of both wild- variety of nuclear targets and thereby destroy the nu- cleus and chromosomal DNA (Hengartner, 2000). There- type and bln 1 mutants (Figures 7G and 7H). We attribute this to the expression of the major cytochrome c, fore, somehow the sperm nucleus must be protected Developmental Cell 694 Figure 8. Spermatids in dbruce / Mutants Have Hypercondensed Nuclei and Degenerate Nuclei are stained in blue (DAPI), and ICs are stained in red (phalloidin). (A) Wild-type cysts. Normally, the elongated spermatid nucleus acquires a highly elongated, needle shape morphology; each cyst contains a bundle of 64 needle-shaped nuclei located at one end. (B) dbruce E81 /cysts. dbruce /mutants are homozygous viable but male-sterile. Mutant spermatid nuclei are significantly more condensed (arrowheads) and rounded (arrows and arrowheads) and are scattered throughout the cyst. In addition, there are fewer nuclei, and many nuclei are stained faintly (yellow arrows and enlargement in the inset), presumably because they undergo degeneration. Although an IC forms, it is highly reduced and scattered (red). The scale bars represent 20 m. against this lethal activity of drICE. One candidate for apoptosis-like process. However, unlike in regular apoptosis, this process is restricted to the cytoplasmic such a protective function is dBruce, the Drosophila orthologof mammalian Bruce/Apollon (Chen et al., 1999; compartment. An apoptotic marker, acridine orange (AO), specifically stains the CBs and WBs, indicating Hauser et al., 1998; Vernooy et al., 2002). Bruce/Apollon proteins are giant E2 ubiquitin-conjugating enzymes that these structures resemble apoptotic corpses. Fur- thermore, we find that effector caspases, such as drICE, that are thought to inhibit apoptosis. The presence of the BIR domain, a domain also found in IAPs, suggests are activated during Drosophila spermatogenesis and are necessary for the removal of cytoplasm and the that these proteins may bind to caspases (Salvesen and Duckett, 2002). We identified many mutations in dBruce generation of functional sperm. Other key proapoptotic proteins are also expressed and become upregulated from a screen for genetic modifiers of Reaper-induced apoptosis (J.A., K. McCall, and H.S., unpublished re- during Drosophila spermatogenesis. Our observations suggest that an apoptosome-like complex is assembled sults). These alleles are homozygous viable but male- sterile, and they behave genetically like loss-of-function prior to individualization and is important for the removal alleles. In all dBruce / mutants, we saw evidence for of bulk cytoplasm from spermatids. nuclear hypercondensation and degeneration, indica- tive of excessive caspase activity (Figure 8). Whereas A Specific Cytochrome C Is Required nuclei in wild-type acquire a highly elongated, needle for Caspase Activation shape morphology, nuclei in dBruce / mutants ap- In mammals, mitochondria are an important organelle peared much more condensed and rounded, and were for the induction of apoptosis, and it has been shown scattered throughout the cyst. In addition, many nuclei that they can release several proapoptotic proteins into stained very faintly, and eventually they degenerated. the cytosol in response to apoptotic stimuli (Green and These observations are consistent with a role of dBruce Reed, 1998; Meier et al., 2000a; van Loo et al., 2002). in protecting spermatids against excessive caspase ac- The best-studied case is the release of cytochrome c, tivity and death. which binds to and activates Apaf-1, which in turn leads to the activation of caspase-9 (Wang, 2001). However, Discussion no comparable role of mitochondrial factors for caspase activation has yet been established in invertebrates. Removal of the Bulk Cytoplasm during Sperm Here, we present evidence that a specific form of cyto- Differentiation Resembles Apoptosis chrome c, encoded by the cyt-c-d gene, is required for Generation of functional sperm in all metazoan animals the activation of the effector caspase drICE at the onset requires the elimination of most of the cytoplasm to of spermatid individualization. Loss-of-function mutants generate a highly condensed, compact cell. The mole- for cyt-c-d are homozygous viable but male-sterile. Sig- cular and cellular mechanisms that drive this process nificantly, these mutants were defective in drICE activa- are poorly understood. In this study, we provide evi- tion and failed to exclude the bulk cytoplasm, producing dence that the elimination of the cytoplasmduringtermi- nal differentiation of elongated spermatids involves an phenotypes virtually identical to the ones resulting from Role of Apoptotic Factors in Spermatogenesis 695 the application/expression of caspase inhibitors. This the cytoplasmic compartment that will be eliminated. One plausible mechanism for locally restricting drICE provides compelling evidence for a role of the cyt-c-d gene for caspase activation during spermatogenesis in activation may be the local release of the minor cyto- chrome c from mitochondria, which are known to un- Drosophila. Interestingly, it was previously suggested that only cyt-c-p, but not cyt-c-d, functions in respiration dergo dramatic morphological changes only in the post- individualized portion of the cyst (Fuller, 1993). (Inoue et al., 1986). Consistent with an essential role of cyt-c-p in respiration, a P element insertion into this locus resulted in recessive lethality (data not shown). Similarities between Drosophila and Mammalian Likewise, targeted gene inactivation of the murine cyto- Sperm Differentiation chrome c gene causes very early embryonic lethality, Terminal differentiation of sperm shares many morpho- and this has precluded functional studies on the role logical and biochemical features with apoptosis. How- of cytochrome c for caspase activation during normal ever, rather than causing the death of the entire cell, development in mammals (Li et al., 2000). We propose in this case apoptotic proteins are used to specifically that the two cytochrome c genes in Drosophila fulfill eliminate cytoplasmic components, thereby producing distinct functions in respiration (cyt-c-p) and cas- a highly specialized living cell. Interestingly, a similar pase activation/apoptosis (cyt-c-d). Previous argu- phenomenon is observed in mammals. As in Drosophila, ments against a role of cytochrome c for caspase activa- intracellular bridges between spermatids and the bulk tion in Drosophila were largely based on the failure to of the spermatidcytoplasmneed to be eliminated during detect releaseof cytochromec frommitochondria. How- mammalian spermatogenesis. In mammals, the cyto- ever, because cyt-c-d is expressed at much lower levels plasm collects in the residual body (RB), which is func- than cyt-c-p(Figure 6B; Limbach andWu, 1985), it would tionally homologous to the WB in Drosophila (Figure be virtually impossible to detect the release of the rele- 3F). Consistent with this idea, mammalian RBs display vant protein in the absence of highly specific antibodies. several features of apoptosis (Blanco-Rodriguez and Furthermore, because cyt-c-d null flies are viable and, Martinez-Garcia, 1999). Although a role of caspases for apart from male sterility, have no obvious anatomical the removal of bulk cytoplasm during mammalian sper- defects, it is unlikely that this gene is broadly required matogenesis remains to be established, we have prelim- for the activation of apoptosis. A complete block of inary data showing that active caspase-3 is present in apoptosis in Drosophila interferes with normal em- RBs in the testes of mice (H. Kissel, E.A., and H.S., bryogenesis, and mutants with significantly reduced unpublished results). Consequently, there are both ana- apoptosis canbe viablebut arephenotypically abnormal tomical and biochemical similarities between insects (White et al., 1994; Grether et al., 1995; Peterson et al., and mammals that warrant more detailed studies. This 2002). Therefore, although we cannot rule out that the is not only of academic interest, as various types of loss of cyt-c-d function may affect and/or delay apopto- caspase inhibitors are being considered as drugs for sis in somatic tissues, the main function of this gene therapeutic purposes, and effects on human fertility appears to be in caspase activation during spermatid have not been studied. Furthermore, the abnormal sper- differentiation. matozoa with residual cytoplasmresultingfromcaspase inhibition in Drosophila (Figure 3E) bear a striking resem- blance to one of the most commonly seen abnormalities Protection of the Sperm Nucleus of human spermatozoa, known as cytoplasmic droplet Effector caspases, such as drICE, Dcp-1, and cas- sperm (Figure 3F, right panel; Hollanders and Carver- pase-3, normally can cleave a variety of nuclear targets, Ward, 1996). Therefore, it is possible that defects in including lamins, I-CAD, and PARP (Hengartner, 2000). proper caspase activation may be responsible for this Therefore, the spermnucleus must be protected against pathology, and further studies of apoptotic proteins may this potentially lethal activity of drICE. Our data indicate shed light on the etiology of some forms of human male that dBruce, which encodes a giant E2 ubiquitin-conju- infertility. gating enzyme, may exercise this function (Chen et al., 1999; Hauser et al., 1998; Vernooy et al., 2002). Loss of Experimental Procedures dBruce function results in nuclear hypercondensation, degeneration, and male sterility, consistent with a role of Fly Strains dBruce to restrain or limit caspase activity. Interestingly, Canton S and yw were used as wild-type controls. fuzzy onions (fzo) dBruce contains a BIR domain, a motif also found in mutant alleles were obtained from M.T. Fuller (Stanford University), jaguar (jar 1 ), purity of essence (poe 1 ), hephaestus (heph 2 ), blanks IAPs. This suggests that dBruce may bind to either cas- (bln 1 ), Df(2L)H20, l(2)k13905, and 2-3 jumpstart lines from the pases or Reaper/Hid/Grim-like (RHG) proteins (Salvesen Bloomington Stock Center, and EP(2)2305 and EP(2)2049 lines from and Duckett, 2002). Previous work has argued against Exelixis and Szeged Drosophila Stock Centre, respectively. dbruce RHG proteins as direct targets for dBruce (Vernooy et alleles were isolatedinour lab (J.A., K. McCall, andH.S., unpublished al., 2002). Therefore, it is attractive to speculate that results). HSP83-Gal4 lines were generated by using CasperR-HSP83 dBruce functions by directly binding to and degrading (Horabin and Schedl, 1993; Hicks et al., 1999) to drive GAL4 (Brand and Perrimon, 1993). caspases. Obviously, this proposedfunction wouldhave to be spatially restricted during spermatogenesis, for ex- Antibodies ample, by localizing dBruce to protected compartments, Primary antibodies used in this study were anti-lamin Dm 0 (mAb- or by spatially limiting its E2 activity. An additional possi- ADL84; P. Fisher; 1:100), rabbit CM1 antiserum (IDUN Pharmaceuti- bility is that drICE activation occurs only locally, in the cals; 1:1000), rabbit anti-FL-DRICE and rabbit anti-active DRICE affected compartment. Consistent with this idea, strong (B. Hay; 1:500 to 1:1000; Dorstyn et al., 2002), anti-cytochrome c (mAb2G8; R.J. Jemmerson; 22 g/ml; Varkey et al., 1999), guinea CM1 staining was only observed distal to the nuclei, in Developmental Cell 696 pig anti-Dronc (H.D. Ryoo and H.S., unpublished, 1:2000), and rabbit Acknowledgments anti--galactosidase (Cappell; 1:1000). All secondary antibodies were from Jackson Laboratories (1:500). We are grateful to M.T. Fuller, R. Jemmerson, B. Hay, K.G. Miller, the Bloomington Stock Center, Exelixis, and the Szeged Drosophila Stock Centre for providing stocks and reagents. We thank Steller Antibody Staining lab members for advice and criticism, T. Noguchi for advice on At least 20 testes were examined for each experiment. Testes of primary culture of cysts, T. Gorenc, B. Mollereau, H.D. Ryoo, S. young adults were dissected in testis buffer (TB; 10 mM Tris-HCl Sampath, S. Shaham, and A. Tang for critically reading the manu- [pH 6.8], 183 mM KCl, 47 mM NaCl, 1 mM EDTA, and 1 mM PMSF), script, and R. Cisse and A. Persaud for technical assistance. E.A. transferred to a 2.5 l drop of TB on a siliconized coverslip (GOLD is a postdoctoral associate and H.S. is an investigator with the SEAL), opened using thin forceps, and sandwiched with a poly-L- Howard Hughes Medical Institute. Part of this work was supported lysine-coated slide. The sandwich was frozen in liquid nitrogen, the by NIH grant RO1 GM60124. coverslip was removed with a razor blade, and the slide was placed in ice-cold absolute ethanol. The slides were drained and a hy- Received: January 23, 2003 drophobic ring surrounding the opaque tissue was drawn using a Revised: April 7, 2003 PAP PEN (Zymed Laboratories). The tissue was fixed in 4% formal- Accepted: April 7, 2003 dehyde in PBS for 20 min, rinsed twice with PBS for 5 min, incubated Published: May 5, 2003 in PBT (PBS 0.1% Triton X-100) for 30 min, and rinsed twice again. The fixed testes were then blocked with PBS/BSA (1% BSA in PBS) for 45 min, incubated with primary antibody (diluted in PBS/ References BSA) within the hydrophobic ring overnight at 4C inside a humid chamber, and rinsed twice for 5 min in PBS. Testes were incubated Abrams, J.M., White, K., Fessler, L.I., and Steller, H. (1993). Pro- with the secondary antibody (diluted in PBS) for 1 hr at roomtemper- grammed cell death during Drosophila embryogenesis. Develop- ature, rinsed twice in PBS for 5 min, incubated with 1 ng/l TRITC- ment 117, 2943. phalloidin (Sigma) in PBTw(PBS 0.1% Tween 20) for 1.5 hr in 37C Baker, N.E., and Yu, S.Y. (2001). The EGF receptor defines domains in a humid chamber, rinsed once for 15 min at room temperature, of cell cycle progression and survival to regulate cell number in the and mounted in Vectashield mounting medium with DAPI (Vector developing Drosophila eye. Cell 104, 699708. Laboratories). For AO staining, testes were dissected in TB, incu- Blanco-Rodriguez, J., and Martinez-Garcia, C. (1999). Apoptosis is bated for 5 min in fresh 1.6 10 6 M AO (Sigma) in TB, rinsed briefly physiologically restricted to a specialized cytoplasmic compartment in TB, and mounted in TB. All pictures were taken with a Zeiss in rat spermatids. Biol. Reprod. 61, 15411547. confocal microscope. Brand, A.H., and Perrimon, N. (1993). Targeted gene expression as The immunocytochemistry procedure was essentially as de- a means of altering cell fates and generating dominant phenotypes. scribed in Gonczy et al. (1992). Development 118, 401415. Castrillon, D.H., Gonczy, P., Alexander, S., Rawson, R., Eberhart, RNA Isolation and Northern Blotting C.G., Viswanathan, S., DiNardo, S., and Wasserman, S.A. (1993). Total RNA was extracted from 100 adult male flies by using the Toward a molecular genetic analysis of spermatogenesis in Dro- TRIZOL reagent according to the manufacturers instructions (In- sophila melanogaster: characterization of male-sterile mutants gen- vitrogen). Poly(A) RNA was isolated (PolyATract III; Promega), and erated by single P element mutagenesis. Genetics 135, 489505. 2.5 g poly(A) RNAwas loaded per lane. Northern blots were probed with a 200 bp fragment from the 3 UTR of cyt-c-d (BDGPs EST Chen, Z., Naito, M., Hori, S., Mashima, T., Yamori, T., and Tsuruo, clone LP05614 cleaved with BseRI and XhoI). T. (1999). A human IAP-family gene, apollon, expressed in human brain cancer cells. Biochem. Biophys. Res. Commun. 264, 847854. Testes Cultures Cryns, V., and Yuan, J. (1998). Proteases to die for. Genes Dev. 12, Testes were dissected in a drop of culture media (10% fetal calf 15511570. serum, modified Shields and Sang M3 medium [bicarbonate; Dorstyn, L., Read, S., Cakouros, D., Huh, J.R., Hay, B.A., and Kumar, Sigma], with 1 penicillin/streptomycin [GIBCO-BRL]). Each iso- S. (2002). The role of cytochrome c in caspase activation in Drosoph- lated testis was incubated at 25Cfor 2426 hr with or without 30 M ila melanogaster cells. J. Cell Biol. 156, 10891098. caspase inhibitor Z-VAD(OMe)-FMK (Enzyme Systems Products). Fabrizio, J.J., Hime, G., Lemmon, S.K., and Bazinet, C. (1998). Ge- netic dissection of spermindividualization in Drosophila melanogas- Fertility Test ter. Development 125, 18331843. Young adult males were placed individually with three wild-type Fawcett, D.W., and Phillips, D.M. (1969). Observation on the release virgin females in separate vials at 25C, and vials were scored for of spermatozoa and on changes in the head during passage through offspring after 1 week. the epididimis. J. Reprod. Fertil. 6, 405418. Fraser, A.G., and Evan, G.I. (1997). Identification of a Drosophila P Element Excisions and Isolation of Fertile Revertants melanogaster ICE/CED-3-related protease, drICE. EMBO J. 16, To generate revertants fromeither the bln 1 or the EP strains, the p[Z] 28052813. and EP transposons were excised using the transposase-producing 2-3 jumpstart strain (Robertson et al., 1988). For the bln 1 strain, Fuller, M.T. (1993). Spermatogenesis in Drosophila. In The Develop- we used the following excision scheme: bln 1 /bln 1 females were ment of Drosophila melanogaster, M. Bate and A.M. Arias, eds. crossed to Sco/CyO; 2-3, ki males, and progeny of the bln 1 /CyO; (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press), pp. 2-3, ki genotype were crossed to the original bln 1 /CyO strain. 71147. Groups of five males, which contain a potential excised allele over Furuchi, T., Masuko, K., Nishimune, Y., Obinata, M., and Matsui, Y. the bln 1 allele, were allowed to copulate with wild-type females for (1996). Inhibition of testicular germcell apoptosis and differentiation 4 days for a fertility test. Then, testes of fertile groups were dissected in mice misexpressing Bcl-2 in spermatogonia. Development 122, and analyzed for the presence of active drICE and proper individual- 17031709. ization structures. Gonczy, P., Viswanathan, S., and DiNardo, S. (1992). Probing sper- matogenesis in Drosophila with P-element enhancer detectors. De- Isolation and Characterization of dbruce Alleles velopment 114, 8998. In a genetic screen for cell death genes, we isolated 13 alleles of Green, D.R., and Reed, J.C. (1998). Mitochondria and apoptosis. the Drosophila IAP-related gene, dbruce. Most of our dBruce alleles Science 281, 13091312. dominantly enhance cell death induced by Reaper and Grim, but have no effect on Hid-induced apoptosis. dbruce E81 allele bears a Grether, M.E., Abrams, J.M., Agapite, J., White, K., and Steller, H. (1995). The head involutiondefective gene of Drosophila melanogas- deletion of 534 bp in-frame, which predicts to encode a truncated protein form that lacks the BIR domain (amino acids 234411). ter functions in programmed cell death. Genes Dev. 9, 16941708. Role of Apoptotic Factors in Spermatogenesis 697 Hales, K.G., and Fuller, M.T. (1997). Developmentally regulated mito- Salvesen, G.S., and Renatus, M. (2002). Apoptosome: the seven- spoked death machine. Dev. Cell 2, 256257. chondrial fusion mediated by a conserved, novel, predicted GTPase. Cell 90, 121129. Srinivasan, A., Roth, K.A., Sayers, R.O., Shindler, K.S., Wong, A.M., Fritz, L.C., and Tomaselli, K.J. (1998). In situ immunodetection of Hauser, H.P., Bardroff, M., Pyrowolakis, G., and Jentsch, S. (1998). Agiant ubiquitin-conjugating enzyme related to IAPapoptosis inhib- activated caspase-3 in apoptotic neurons in the developing nervous system. Cell Death Differ. 5, 10041016. itors. J. Cell Biol. 141, 14151422. Hay, B.A., Wolff, T., and Rubin, G.M. (1994). Expression of baculovi- Thornberry, N.A., and Lazebnik, Y. (1998). Caspases: enemies within. Science 281, 13121316. rus P35 prevents cell death in Drosophila. Development 120, 2121 2129. Tokuyasu, K.T., Peacock, W.J., and Hardy, R.W. (1972). Dynamics of spermiogenesis in Drosophila melanogaster. I. Individualization Hengartner, M.O. (2000). The biochemistry of apoptosis. Nature 407, 770776. process. Z. Zellforsch. Mikrosk. Anat. 124, 479506. van Loo, G., Saelens, X., van Gurp, M., MacFarlane, M., Martin, S.J., Hicks, J.L., Deng, W.M., Rogat, A.D., Miller, K.G., and Bownes, M. (1999). Class VI unconventional myosin is required for spermatogen- and Vandenabeele, P. (2002). The role of mitochondrial factors in apoptosis: a Russian roulette with more than one bullet. Cell Death esis in Drosophila. Mol. Biol. Cell 10, 43414353. Differ. 9, 10311042. Hollanders, J.M.G., and Carver-Ward, J.A. (1996). Male Infertility from A to Z: A Concise Encyclopedia (New York: The Parthenon Varkey, J., Chen, P., Jemmerson, R., and Abrams, J.M. (1999). Al- tered cytochrome c display precedes apoptotic cell death in Dro- Publishing Group). sophila. J. Cell Biol. 144, 701710. Honarpour, N., Du, C., Richardson, J.A., Hammer, R.E., Wang, X., and Herz, J. (2000). Adult Apaf-1-deficient mice exhibit male infertil- Vernooy, S.Y., Chow, V., Su, J., Verbrugghe, K., Yang, J., Cole, S., Olson, M.R., and Hay, B.A. (2002). Drosophila Bruce can potently ity. Dev. Biol. 218, 248258. suppress Rpr- and Grim-dependent but not Hid-dependent cell Horabin, J.I., and Schedl, P. (1993). Sex-lethal autoregulation re- death. Curr. Biol. 12, 11641168. quires multiple cis-acting elements upstream and downstream of the male exon and appears to depend largely on controlling the use Wang, X. (2001). The expanding role of mitochondria in apoptosis. Genes Dev. 15, 29222933. of the male exon 5 splice site. Mol. Cell. Biol. 13, 77347746. Inoue, S., Inoue, H., Hiroyoshi, T., Matsubara, H., and Yamanaka, White, K., Grether, M.E., Abrams, J.M., Young, L., Farrell, K., and Steller, H. (1994). Genetic control of programmed cell death in Dro- T. (1986). Developmental variation and amino acid sequences of cytochromes c of the fruit fly Drosophila melanogaster and the flesh sophila. Science 264, 677683. fly Boettcherisca peregrina. J. Biochem. (Tokyo) 100, 955965. Yoo, S.J., Huh, J.R., Muro, I., Yu, H., Wang, L., Wang, S.L., Feldman, R.M., Clem, R.J., Muller, H.A., and Hay, B.A. (2002). Hid, Rpr and Jacobson, M.D., Weil, M., and Raff, M.C. (1997). Programmed cell death in animal development. Cell 88, 347354. Grim negatively regulate DIAP1 levels through distinct mechanisms. Nat. Cell Biol. 4, 416424. Knudson, C.M., Tung, K.S., Tourtellotte, W.G., Brown, G.A., and Korsmeyer, S.J. (1995). Bax-deficient mice with lymphoid hyperpla- Yu, S.Y., Yoo, S.J., Yang, L., Zapata, C., Srinivasan, A., Hay, B.A., and Baker, N.E. (2002). A pathway of signals regulating effector and sia and male germ cell death. Science 270, 9699. initiator caspases in the developing Drosophila eye. Development Li, K., Li, Y., Shelton, J.M., Richardson, J.A., Spencer, E., Chen, Z.J., 129, 32693278. Wang, X., andWilliams, R.S. (2000). Cytochrome c deficiency causes embryonic lethality and attenuates stress-induced apoptosis. Cell Yue, L., Karr, T.L., Nathan, D.F., Swift, H., Srinivasan, S., and Lind- quist, S. (1999). Genetic analysis of viable Hsp90 alleles reveals a 101, 389399. critical role in Drosophila spermatogenesis. Genetics 151, 1065 Limbach, K.J., and Wu, R. (1985). Characterization of two Drosophila 1079. melanogaster cytochrome c genes and their transcripts. Nucleic Acids Res. 13, 631644. Zhou, L., Song, Z., Tittel, J., and Steller, H. (1999). HAC-1, a Drosoph- ila homolog of APAF-1 and CED-4, functions in developmental and Meier, P., Finch, A., and Evan, G. (2000a). Apoptosis in development. radiation-induced apoptosis. Mol. Cell 4, 745755. Nature 407, 796801. Zimmermann, K.C., Ricci, J.E., Droin, N.M., and Green, D.R. (2002). Meier, P., Silke, J., Leevers, S.J., and Evan, G.I. (2000b). The Dro- The role of ARK in stress-induced apoptosis in Drosophila cells. J. sophila caspase DRONC is regulated by DIAP1. EMBO J. 19, Cell Biol. 156, 10771087. 598611. Noguchi, T., and Miller, K.G. (2003). A role for actin dynamics in individualization during spermatogenesis in Drosophila melanogas- ter. Development 130, 18051816. Peterson, C., Carney, G.E., Taylor, B.J., and White, K. (2002). reaper is requiredfor neuroblast apoptosis duringDrosophila development. Development 129, 14671476. Print, C.G., Loveland, K.L., Gibson, L., Meehan, T., Stylianou, A., Wreford, N., de Kretser, D., Metcalf, D., Kontgen, F., Adams, J.M., and Cory, S. (1998). Apoptosis regulator bcl-w is essential for sper- matogenesis but appears otherwise redundant. Proc. Natl. Acad. Sci. USA 13, 1242412431. Robertson, H.M., Preston, C.R., Phillis, R.W., Johnson-Schlitz, D.M., Benz, W.K., and Engels, W.R. (1988). A stable genomic source of P element transposase in Drosophila melanogaster. Genetics 118, 461470. Rodriguez, I., Ody, C., Araki, K., Garcia, I., and Vassalli, P. (1997). An early and massive wave of germinal cell apoptosis is required for the development of functional spermatogenesis. EMBO J. 16, 22622270. Ryoo, H.D., Bergmann, A., Gonen, H., Ciechanover, A., and Steller, H. (2002). Regulation of Drosophila IAP1 degradation and apoptosis by reaper and ubcD1. Nat. Cell Biol. 4, 432438. Salvesen, G.S., and Duckett, C.S. (2002). IAP proteins: blocking the road to deaths door. Nat. Rev. Mol. Cell Biol. 3, 401410.