The Mer³ allele carries a missense mutation (Met¹⁷⁷→Ile), and male flies hemizygous for Mer³ are viable but sterile 8. Upon examination of the testis, we noted that seminal vesicles from the Mer³ males were smaller than those from the control FM6 flies (compare Figure 1B with Figure 1A). In addition, very few sperm were found in the Mer³ seminal vesicles, and they were immotile, in contrast to those seen in the control FM6 siblings obtained from the same cross. Acetic acid/orcein staining showed that the Mer³ testis had fewer sperm heads in each bundle than the control testis (compare Figure 1D with Figure 1C). Also, the sperm in the bundle were arranged in a more disorganized fashion.

Previously, LaJeunesse et al. 8 demonstrated that ectopic expression of a Mer⁺ or Mer³ transgene using a ubiquitously-expressed Gal4 driver (T80-Gal4) in a Mer⁴ (Gln¹⁷⁰→stop) background rescued Mer⁴ lethality. In addition, insertion of a cosmid construct (P{cosMer⁺}), carrying the entire D-Mer gene, was capable of rescuing various Mer mutations. We conducted a similar experiment to test whether a Mer⁺ transgene could rescue the sterility of the Mer⁴ allele. Table 1 shows that males carrying both Mer⁴ and P{cosMer⁺} were viable and fertile. To ensure germline expression of the Mer⁺ or Mer³transgene, we cloned the Mer⁺ and Mer³ sequences into the pUASP vector, and used them to transform embryos. We showed that both pUASP-Mer⁺ and pUASP-Mer³ could rescue the lethality of Mer⁴ mutation when ectopically activated by the Act5C-Gal4 driver (Table 1). However, only Mer⁺ over-expression restored the fertility of flies with the Mer⁴ mutation.

To understand the cause of sterility in the Merlin mutant flies, we studied the subcellular localization of the Merlin protein in the control FM6 and Mer³ testes at various stages of spermatogenesis. In the cysts from the control testis, Merlin expression was detected in the cellular cortex of spermatocytes (Figure 2A), as seen in somatic tissues and earlier germ cells 17. This cortical localization became more pronounced in spermatocytes during prometaphase and metaphase of meiosis (Figures 2B and 2C). Merlin was found to redistribute, and more intense staining was observed in the area associated with the microtubules of the central spindle in telophase (Figure 2D). During cytokinesis, intense Merlin staining remained associated with the microtubules but the intensity was less in the region of the contractile ring (Figure 2E). A similar Merlin distribution pattern was seen during the second meiotic division (data not shown). At the onion stage, Merlin was concentrated in the Nebenkern, a specialized structure formed by the fusion of mitochondria during spermatid differentiation (Figure 2F). This intense Merlin staining in the Nebenkern remained throughout the comet stage of spermatid elongation, during which the Nebenkern split into two parts (Figure 2G). Note that concentrated Merlin immunoreactivity was clearly seen in the two subunits of the Nebenkern in the spermatid (insert in Figure 2G). In the control cyst containing mature sperm, Merlin staining continued to be present in the elongated Nebenkern (Figure 2H). In addition, Merlin was seen as a bright punctate dot in the acrosomal region, a Golgi apparatus-derived structure developed over the anterior part of the sperm's head. We also performed a similar immunostaining on the Mer³ testis. Although we could detect Merlin staining in the Mer³ cyst at the comet stage, the Mer³ spermatid nuclei were scattered throughout the cyst, and the arrangement of spermatids appeared irregular (Figure 2I). The ability of the antibody to detect Merlin staining in the Mer³ cyst suggests that the missense mutation in Mer³ did not affect antibody recognition. Using the same antibody, LaJeunesse 8 previously detected a similar cortical localization of Merlin in both the wild-type and Mer³ imaginal discs. However, the Mer³ mutation clearly affects Merlin function as the spermatids in the Mer³ cyst were poorly arranged.

Next, we examined if there were any abnormalities during early steps of spermatogenesis in the Mer³ testis. We dissected testes from both the control and Mer³ mutant flies. Following staining with acetic acid/orcein, the testes were squashed and examined according to Ashburner 27. Although we did not find any abnormalities during mitosis of spermatogonia or spermatocyte growth, we observed three types of abnormalities during meiosis of spermatocytes from the Mer³ testis, as compared with the control testis. The first type of abnormality is shown in Figure 3A, demonstrating a spermatid containing two nuclei of equal size and two Nebenkerns. This type of abnormality was likely caused by cytokinesis failure during the second meiotic division. The second type of abnormality is tripolar spindle in a spermatocyte going through the second meiotic division (Figure 3B). This result suggests incomplete cytokinesis in the previous meiotic division. The third type of abnormality is four-polar spindle in a spermatocyte undergoing the second meiotic division (Figures 3C–E). Note that a secondary Mer³ spermatocyte contained two pairs of telophase nuclei (Figure 3D). Each pair of nuclei was situated with its own spindle, and two spindles shared a common mid zone (Figure 3C and 3E). This is in contrast with wild-type spermatocytes in telophase showing two daughter nuclei separate from each other along the central spindle (Figure 3F). The four-polar spindle represents another case of abnormal cytokinesis in the first meiotic division. It should be mentioned that we detected the first type of abnormality in about 5% of the mutant cysts, while the second and third types of abnormalities appeared less frequent.

Following meiosis, spermatid elongation ensues 1. Prior to spermatid elongation, spermatid nuclei group at a defined area of the cyst wall in a process referred to as cyst polarization 3. We noted that, at this stage, spermatid nuclei were detected as a group in the control cyst (Figure 4A). Intriguingly, spermatid nuclei in the Mer³ cyst were grouped in two locations (Figure 4B–E), and in some other cysts, nuclei appeared more scattered (Figure 4C–E). We also found a similar abnormality in nuclear grouping in the Mer⁴ cyst. Although Mer⁴ is larva-lethal 19, we isolated some rare hemizygous Mer⁴ male pupae from the y Mer⁴/Binsn stock, and a few of them were able to grow to pharates adults. When examining testis preparations from these Mer⁴ males, we detected spermatid nuclei arranged in two groups (Figures 4F and 4G) or in a scattered manner in all of the mutant cysts (Figure 4H). Also, the cyst polarization defects were more frequently seen in Mer⁴ (about 90%) than Mer³ (about 50%) cysts. These results indicate that Merlin mutations affect cyst polarization.

The final stage of spermatogenesis is the process of individualization, followed by sperm coiling 12. The individualization process is initiated at the head of the spermatid cyst, and involves the formation and movement of the actin cones from the head region of the spermatid bundle to the caudal end 4. Analogous to the previous finding, we detected the actin cones, which moved caudally as a bundle in the control cyst (Figures 5A–D). Note that the sperm heads were grouped in one end of the control cyst, consistent to that seen at the cyst polarization stage (Figures 5B and 5D). Also, the sperm heads became needle-shaped. However, we observed that the sperm nuclei in the Mer³ cyst appeared scattered (Figure 5E–H) and had variable morphology; some were round, while others were needle shape (Figure 5I). Unlike the control cyst, the Mer³ cyst had the actin cones located at multiple sites (Figure 5G). Dispersed actin cones together with scattered nuclei were also found in the Mer⁴ cyst (data not shown).

Since we detected intense Merlin staining in the Nebenkern at the onion stage, we employed electron microscopy to further examine Nebenkern transformation from the structure containing two mitochondrial derivatives at the late stage of spermatid elongation into a configuration filled by electron dense material, called the paracrystalline body, at the end of the individualization stage 4. Thin sections of testes from the control FM6, Mer³, and Mer⁴ males were analyzed under a transmission electron microscope. As shown in Figure 6A, we observed that at the late elongation stage, each spermatid in the control cyst contained one major and one minor mitochondrial derivative associated with one axoneme. A paracrystalline body could be seen within the major mitochondrial derivative. However, we found that some of the spermatids in the Mer³ cyst contained two paracrystalline bodies within the major mitochondrial derivative (Figure 6B). Also, some spermatids had two axonemes. Similarly, two paracrystalline bodies within the major mitochondrial derivative were frequently seen in the elongating spermatids of the Mer⁴ cyst (Figure 6C). It should be mentioned that, unlike the spermatids in the control cyst, which displayed cell-cell contact, the spermatids in the Mer³ cyst were loosely arranged (Figures 6B), and those in the Mer⁴ cyst were grossly disorganized (Figure 6C). In addition, some cytoplasmic shedding was present in the Mer³ cysts (Figure 6B), and excessive amount of cytoplasmic fragments was seen in the Mer⁴ cyst (Figure 6C).

At the individualization stage, each of the 64 spermatids in the control cyst contained one axoneme associated with the major and minor mitochondrial derivatives (Figure 6D). Furthermore, the paracrystalline body almost filled the entire major mitochondrial derivative. While some spermatids in the Mer³ cyst at a similar developmental stage had a paracrystalline body-filled major mitochondrial derivative with axoneme, others containing two paracrystalline body-filled mitochondrial derivatives with abnormal shape or three paracrystalline body-filled mitochondrial derivatives together with one axoneme were also seen (Figure 6E). In addition, spermatids having one or two paracrystalline-filled mitochondrial derivatives but lacking axoneme were observed. Also, the spermatids remained loosely contacted with each other in the Mer³ cyst. The paired arrangement of axoneme with the mitochondrial derivatives was often lost in the spermatids of the Mer⁴ cyst at the individualization stage (Figure 6F). Spermatids with multiple paracrystalline body-filled mitochondrial derivatives but without axoneme, or with two axonemes, were found. Also, cytoplasmic fragmentation together with condensed cytoplasmic remnants and gigantic cytoplasmic bodies were also seen in the Mer⁴ cyst.

Consistent with previous observations 124, we noted that mature spermatids in the control cyst had a substantial reduction in the amount of cytoplasm and a significant reduction in the size of the minor mitochondrial derivative at the end of the individualization stage. The major mitochondrial derivative is filled by a dark-staining paracrystalline material (Figure 6G). Each of the 64 individualized spermatids in the control cyst displayed a highly ordered axoneme-Nebenkern pair. In contrast, a gross disorganization in the arrangement of the individualized spermatids in the Mer³ cyst was detected (Figure 6H). Although some spermatids exhibited the axoneme-Nebenkern pair, others appeared to be fused together or connected by a thin cytoplasmic extension. In addition, the boundary between the cysts was not evident, and each cystic area contained less than 64 spermatids. The most dramatic alteration was observed in the Mer⁴ cyst (Figure 6I). Although axoneme and Nebenkern could be found, very little cytoplasmic material was seen, and the cell-cell boundary could not be easily identified. Insert in Figure 6I illustrates that the structure of axoneme appeared to be intact. Despite this dramatic alteration, the 9+2 microtubule-containing structure of axoneme appeared to be preserved in the Merlin mutant cysts, indicating that Merlin is not required for axoneme formation and elongation.

Taken together, our results show that Merlin mutations affect cytokinesis, cyst polarization, nuclear shaping, and spermatid individualization. The observation that Merlin is highly concentrated in the Nebenkern suggests that Merlin may pay a role in mitochondria formation and function during various stages of spermatogenesis.

Flies were maintained at 25°C in standard cornmeal yeast-agar medium. Various Merlin mutant strains were generously provided by Rick G. Fehon at University of Chicago, Chicago, IL 40. Hemizygous Mer³ males were taken from the strain w Mer³P{ry [+t7.2] = neoFRT}19A/FM6, y B. The FM6, y B siblings and the y w males from the stock y w Pim 19A-FRT/TM6, Tb (abbreviated here as Pim) were used as controls. The strain y w Mer⁴P{ry [+t7.2] = neoFRT}19A/FM7i, P{w [+mC] = ActGFP}JMR3 was used as a source of Mer⁴ mutant individuals. The w sn³l(1)18DEb³ P{ry [+t7.2] = neoFRT}19A; P{w [+mC] = cosMer⁺}3/+ strain with an insertion of a genomic fragment containing the entire Mer gene was also used. Transgenic strains carrying the pUAST-MycMer⁺ or pUAST-MycMer³ constructs have been described previously 8. The strain, containing an insertion of a transposable element carrying the green fluorescent protein (GFP) tag inserted into the CG8351 gene, was kindly provided by Alain Debec of Université Pierre et Marie Curie, Observatoire Océanologique, Villefranche-sur-mer, France. This strain allowed labeling the cytoplasm of all cells uniformly during spermatogenesis. The strain y w; Ki Delta2-3 carrying endogenous transposase activity was a gift from the laboratory of Igor Zhimulev, Institute of Cytology and Genetics, Novosibirsk, Russia.

The females w Mer³P{ry [+t7.2] = neoFRT}19A/FM6, y B were mated with the males FM6, y B/Y from the same stock. The resulting Mer³ male y w Mer³P{ry [+t7.2] = neoFRT}19A/Y and control FM6, y B/Y males were obtained and their testes were dissected. The squashed preparations of testes were performed according to Ashburner 27. Briefly, testes were dissected in Hanks balanced salt solution (HBSS) and stained in a 1:1 mixture of 1% acetic acid/orcein and 1% acetic acid/carmine for 2 hours at room temperature. Stained testes were examined under phase-contrast optics of an Axiovert-200 microscope (Carl-Zeiss). For DAPI staining, dissected testis tissues were fixed in 3.7% formaldehyde in Dulbecco's phosphate-buffered saline (PBS), pH 7.2, and stained with DAPI (1.5 μg/ml) prior to visualization under the epifluorescence optics of an Axiovert-200 microscope.

Genomic DNA were isolated from flies carrying the UAS-MycMer⁺ or UAS-MycMer³ insertion as described above and amplified by PCR to generate Merlin cDNAs as previously described 8. The Merlin cDNAs were inserted into the pUASP vector at the SacII and XbaI sites. The DNA inserts in the plasmids were confirmed by restriction digestions and direct sequencing. Embryos with the genetic constitution y w; Ki Delta2-3, carrying the endogenous transposase Delta2-3, were injected with the pUASP-MycMer⁺ or pUASP-MycMer³ DNA. After reaching the adult stage, the injected flies were mated with y w mating partners. The resulting w⁺ progeny were isolated and crossed to establish a strain carrying a transposition. A few independent insertions were obtained for each construct, and the presence of the Merlin coding sequence in the transposants was tested by PCR analysis of genomic DNA described above.

For the examination of sperm motility, seminal vesicles of male flies that had been alone for three days were isolated and checked for movement of sperm heads under Varel contrast optics of an Axiovert-200 microscope. For general spermatogenesis inspection, dissected testes were squashed in HBSS using coverslips as described by Fuller 2. The unfixed preparations of live cysts were examined for the coiling process according to Cross and Shellenbarger 3.

Dissected testes were placed onto poly-L-lysine coated slides. To isolate cysts, the dissected testes were pierced using a tungsten needle attached to a Narishigi micromanipulator. Slides with testis tissues attached were fixed in 3.7% formaldehyde in PBS, pH 7.2. After washing in PBS 10 min three times, fixed tissues were permeated with 1% Triton X-100 in PBS for 30 min and then pretreated with the blocking solution containing 1% non-fat dry milk in PBS. Pretreated tissues were incubated with a guinea pig anti-Merlin antibody (1:6000 dilution; 17) or a mouse anti-α-tubulin antibody (1:500 dilution; Sigma Chemicals) overnight. After washing with PBS three times, a secondary antibody conjugate (Alexa 488-conjugated anti-guinea pig IgG [1:700 dilution], Alexa 568-conjugated anti-mouse IgG [1:200 dilution, or a FITC-conjugated anti-mouse IgG [1:50 dilution]) was added for 2 hours at room temperature. To visualize actin filaments, FITC-conjugated phalloidin (1:50 dilution; Molecular Probe) was used. In some experiments, nuclei were stained with DAPI (1.5 μg/ml). After staining, the slides were mounted in Mowiol with 10% DABCO and examined under the epifluorescence or phase-contrast optics of an Axiovert-200 or an Olympus BX50 microscope.

Dissected testes were fixed in 2% glutaraldehyde in PBS, pH 7.4, for 2 hours and then treated with 1% osmium tetraoxide in PBS for 1 hour. Treated tissues were stained with 1% uranyl acetate at 4°C overnight. Following dehydration in ascending ethanol solutions, stained tissues were embedded in Agar-100, mounted in the block, and polymerized at 60°C for two to three days. Ultrathin sections were prepared and contrasted by incubating in 1% uranyl acetate and lead citrate, and examined using a Hitachi H7650 or a JEOL 1000SX transmission electron microscope.