nup154 Genetically Interacts With cup and Plays a Cell-Type-Specific Function During Drosophila melanogaster Egg-Chamber Development

Abstract

Nucleoporin Nup154 is a Drosophila component of the nuclear pore complex (NPC), evolutionarily conserved from yeast to humans. While functional studies carried out in both yeast and metazoan cells indicated that Nup154 homologs are key elements of the NPC framework, the striking phenotypic specificity displayed by nup154 hypomorphic mutant alleles suggested that Nup154 might play additional roles in the context of the NPC. Actually, genetic analyses demonstrated that mutant nurse-cell nuclei do not undergo a normal chromosome dispersal process, uncovering an essential requirement for nup154 gene function during oogenesis. In this report, we show that Nup154 interacts genetically and physically with Cup, a germline-specific protein implicated in multiple aspects of female gametogenesis, including the regulation of the nurse-cell chromosome structure. The two proteins colocalize in vivo and are co-immunoprecipitated from ovarian extracts. Moreover, cup, nup154 double mutants exhibit much stronger oogenesis defects than single mutants. Our findings delineate an intriguing scenario where an ubiquitous nucleoporin might directly influence specialized developmental events.

BOTH yeast and vertebrate nuclear pore complexes (NPCs) are composed of ∼30 nucleoporins. Two-thirds of them, to different extents, have been conserved in evolution, suggesting that they might share critical structural and/or functional features (reviewed in Suntharalingam and Wente 2003). One group of highly related nucleoporins comprises S. cerevisiae Nup170 and Nup157, Drosophila Nup154, and mammalian Nup155 (Radu et al. 1993; Aitchison et al. 1995; Gigliotti et al. 1998). Initially depicted as structural proteins, these nucleoporins have more recently revealed very interesting roles in the regulation of both NPC assembly and function. For example, it has been shown that Nup155 is an essential player of a checkpoint mechanism that links the interconnected processes of nuclear envelope and NPC formation (Franz et al. 2005). Nup155 is also able to interact with the mRNA export factor Gle1, and this interaction is required for Gle1 docking at the nuclear rim (Rayala et al. 2004). Moreover, Nup170 is able to prevent the inhibitory effect of Nup53 on Kap121-mediated nuclear transport processes (Makhnevych et al. 2003). These data indicate that the biological relevance of this group of evolutionarily conserved core elements of the NPC is greater than previously thought. The finding that Nup154 is essential for gametogenesis in both sexes not only supports this notion, but also raises the question of whether this nucleoporin might have acquired cell-type-specific functional features (Gigliotti et al. 1998; Kiger et al. 1999).

nup154 is a gene essential for viability: individuals carrying nup154 loss-of-function mutations fail to grow and die at larval stages. Interestingly, nup154 hypomorphic mutations affect egg-chamber development and result in female sterility. Phenotypic analyses of a large collection of nup154 hypomorphic alleles revealed characteristic defects in the chromatin organization of the germline-derived nurse cells, whose chromosomes retain a compact polytene morphology, instead of dissociating and decondensing to be uniformly distributed throughout the nucleus (Gigliotti et al. 1998; Kiger et al. 1999). This peculiar phenotype prompted us to start a functional characterization of the Nup154 nucleoporin during oogenesis.

Here we report that large Nup154 cytoplasmic aggregates are formed under experimental conditions, allowing the synthesis of a large excess of this nucleoporin. While not identifying ectopic pore complexes, these cytoplasmic sites of Nup154 accumulation are highly enriched in the germ-cell-specific Cup protein, which is also implicated in the regulation of the nurse-cell chromosome structure. It was previously reported that Cup is a cytoplasmic protein present throughout oogenesis in all germline cells. In the nurse cells, Cup displays a dynamic localization pattern with an early phase, culminating at stage 4, characterized by marked perinuclear distribution and followed by cytoplasmic dispersion and transfer to the cellular periphery, where most of the protein becomes repositioned by stage 10 (Keyes and Spradling 1997). We demonstrate that Nup154 and Cup colocalize at the nuclear rim of wild-type nurse cells in early oogenesis and associate in immunoprecipitation assays. Moreover, we show that an early developmental block is induced in egg chambers by double combinations of weak nup154 and cup mutant alleles. The strong enhancement of the ovarian phenotype in double mutants suggests that Nup154 and Cup affect common developmental processes during early oogenesis. Taken together, our data indicate that Nup154 interacts with Cup and provide the first evidence for a functional specialization of an essential structural component of the NPC in a distinct cell type.

MATERIALS AND METHODS

Drosophila strains:

Flies were raised on standard sucrose–cornmeal–yeast medium at 25°. The driver line nanos-Gal4:VP16 (Van Doren et al. 1998), used to induce germline-specific expression of the transgenes, was obtained from P. Rorth. nup154 and gfp-nup154 transgenic fly strains were generated by P-element-mediated transformation (Spradling 1986). Two independent lines for each transgene were tested for the ability to rescue the tlp² mutant phenotype. Transgenic lines showing high induction rates were selected and used to set up the overexpression experiments. In these experiments, two copies of either the nup154 or the gfp-nup154 transgene were combined with one copy of the driver transgene through appropriate genetic crosses.

To isolate a recombinant cup¹³⁵⁵, tlp² chromosome, 50 independent lines were generated from the cross of cup¹³⁵⁵/tlp²; ry/ry virgin females with Sp/SM6; ry/ry males. Genomic DNA from these lines was used as template in PCR reactions employing one P-element and two gene-specific primers. Since both mutant alleles are due to P-element insertion, analyses of the PCR products allowed the identification of the recombinant lines. Three of them have been characterized and all displayed the same mutant phenotype. Recombinant cup²¹, tlp¹ chromosomes were identified in a two-step analysis of 46 independent lines, established after the cross of cup²¹/tlp¹ females to Sp/CyO males. First, heterozygous males, possibly carrying the recombinant chromosome balanced with CyO, were crossed to cup²¹/CyO virgin females, and the resulting Cy⁺ females were tested for sterility. The lines that did not complement the cup phenotype were then analyzed by PCR for the presence of P-element insertion in the nup154 gene. Two recombinant lines have been characterized and were phenotypically indistinguishable.

Transgene construction:

A full-length nup154 cDNA was reconstructed from overlapping partial cDNAs. For preparing the gfp-nup154 transgene, the sequence coding for the GFP variant mGFP6 (derived from the UASmGFP6 plasmid provided by A. Brand) and the Nup154 coding sequence were amplified and then joined through an EcoRI site included in the PCR primers.

Both the nup154 cDNA and the gpf-nup154 fusion were inserted into the pUASp vector obtained from P. Rorth (Rorth 1998).

Immunostaining:

Immunostaining of hand-dissected ovaries was carried out as previously described (Gigliotti et al. 1998). Affinity-purified anti-Nup154 rabbit IgG's (obtained from a new batch of anti-Nup154 antiserum), anti-Cup rat polyclonal antiserum (Keyes and Spradling 1997) or mAb414 (Berkeley Antibody, Richmond, CA) were used as primary antibodies. BODIPY FL or Texas red conjugated goat anti-rabbit, anti-rat, or anti-mouse IgG's (Molecular Probes, Eugene, OR) were used as secondary antibodies according to the manufacturer's instructions. DNA staining was performed by 15 or 30 min of incubation in PBS containing 0,5 μg/ml DAPI or 2 μm TOTO-3 (Molecular Probes), followed by extensive washes in PBS. Stained egg chambers, mounted in Aquamount (Polysciences), were analyzed by conventional epifluorescence using a Zeiss Axioskop 2 microscope equipped with Plan-NEOFLUAR 40×/1,3 oil and 20×/0,50 objectives and a Zeiss AxioCam HRc under control of Axiovision 3.1 software or by laser-scanning confocal microscopy using a Leica TCS SP2 AOBS microscope equipped with a HCX PL APO 63×/1,32–0,6 oil objective. Image cropping and adjustment were accomplished using Photoshop (Adobe).

Immunoprecipitation–Western analyses:

Protein extracts were prepared by homogenizing hand-dissected ovaries in 150 mm NaCl, 1% NP-40, 50 mm Tris–HCl (pH 8), 5% glycerol, and a cocktail of protease inhibitors (10 μg/ml aprotinin, pepstatin, and leupeptin, 1 mm PMSF). Protein G Sepharose 4 Fast Flow beads (Amersham, Buckinghamshire, UK) were incubated overnight at 4° with rat anti-Cup polyclonal antiserum (kindly provided by A. Verrotti) in Tween lysis buffer (50 mm HEPES, pH 7.5, 150 mm NaCl, 2,5 mm EGTA, 1 mm EDTA, 0,1% Tween-20). The anti-Cup antibody-coated beads were washed five times and incubated with the ovarian protein extract for 3 hr at 4° in Tween lysis buffer supplemented with the cocktail of protease inhibitors reported above. After five washes, bound proteins were eluted by boiling in SDS sample buffer, separated by SDS–PAGE and transferred to a PVDF membrane (Amersham) by electroblot. A blocking step in TBST (Tris-buffered saline, 0.1% Tween-20) plus 5% ECL blocking agent (Amersham) was followed by overnight incubation at 4° with the anti-Nup154 affinity-purified antibody and washes and incubation with HRP-conjugated goat anti-rabbit secondary antibody (Amersham). After washing, the bands were visualized by the ECL Plus chemiluminescent system (Amersham).

RESULTS

Unbalanced Nup154 levels induce a lethal effect:

To gain new insights into the role(s) played by Nup154 in female germ cells, we chose an overexpression approach and generated several transgenic lines conditionally expressing either a wild-type nup154 transgene or a gfp-nup154 gene fusion. Upon germline-specific induction, GFP-Nup154 was correctly localized at the nuclear envelope of both nurse cells and oocyte (Figure 1, A–C). Furthermore, germline-specific expression of one copy of gfp-nup154 or wild-type nup154 in a nup154 mutant background led to an almost complete rescue of all the ovarian aspects of the mutant phenotype. Specifically, 95–98% of egg chambers and eggs displayed a wild-type morphology, demonstrating that both transgenes were functional. This result also indicated that the developmental defects displayed by nup154 mutant egg chambers are germline dependent.

Figure 1.—

GFP-Nup154 and Nup154 protein distribution in transgenic lines. (A–C) Single confocal section of a stage 8 egg chamber expressing one copy of the gfp-nup154 transgene in the germline. (A) GFP-Nup154 is correctly localized at the edge of the nurse-cell nuclei, highlighted in B by TOTO-3 staining. (C) The merged image shows that GFP-Nup154 also decorates the rim of the oocyte nucleus, where the tightly condensed chromosomes occupy a central position. (D–F) Single confocal section of a stage 8 egg chamber expressing two copies of the nup154 transgene in the germline. (D) Nup154 is localized as expected at the nuclear periphery, but an additional signal is detected in cytoplasmic particles localized in the proximity of the nurse-cell nuclei (arrows). (E) TOTO-3 staining. (F) Merged image.

After having set up appropriate conditions to enhance the production of the transgenic proteins, we investigated the morphological and/or functional consequences of their overexpression. Careful examination of egg-chamber development did not show any evident alterations in the germline or in somatic cells. Nevertheless, eggs laid by females overexpressing Nup154 (or GFP-Nup154) in their germline were not able to sustain embryo development beyond a few mitotic cycles. In contrast, Nup154 overexpression during spermatogenesis did not have any consequences on sperm function and did not affect male fertility. Taken together, these data suggest that an excess of Nup154 does not interfere with the normal developmental programs leading to gamete production in both males and females, but Nup154 accumulation in female germ cells is deleterious for early embryo development.

Nup154 overexpression does not trigger annulate lamellae formation:

Immunofluorescence experiments were performed to visualize the localization of the Nup154 protein produced in excess. Along with the expected staining at the nuclear envelope of all the germ cells, a novel signal was detected in the cytoplasm of the nurse cells. Large cytoplasmic aggregates containing Nup154 first appeared at stage 7–8 in proximity of the nuclei and gradually increased both in size and amount while moving toward the nurse-cell periphery (Figure 1, D–F, and Figure 2A). Characteristic cytoplasmic structures are formed in mammalian cells by overexpression of several NPC components (Imreh and Hallberg 2000; Daigle et al. 2001). These structures, named the annulate lamellae (AL), are tightly stacked layers of double membranes perforated with a high density of pore complexes ultrastructurally very similar to NPCs. To test the hypothesis that the aggregates containing Nup154 were AL, we performed double-labeling experiments, using the anti-Nup154 antibody in combination with the mAb414 antibody (Davis and Blobel 1986), which recognizes a group of phenylalanine–glycine (FG)-repeat-containing nucleoporins (Figure 2). The mAb414 signal was restricted to the nuclear envelope of the Nup154-overexpressing nurse cells (Figure 2B). The lack of mAb414 staining in the cytoplasmic particles formed by Nup154 (Figure 2C) indicated that the endogenous nucleoporins detected by the antibody are not recruited at these sites, which, therefore, do not contain ectopic NPCs.

Figure 2.—

Cytoplasmic accumulation of Nup154 does not trigger AL assembly. (A) At stage 10, abundant Nup154-positive particles are scattered in the subcortical region of Nup154-overexpressing nurse cells. (B) These cytoplasmic sites of Nup154 accumulation do not stain with the monoclonal antibody mAb414, which recognizes a subclass of repeat-containing nucleoporins. (C) The overlap between Nup154 and mAb414 signals is restricted to the nuclear envelope.

Cup colocalizes with ectopically accumulated Nup154:

The peculiar accumulation pattern shown by Nup154 in overexpression conditions raised the question of whether it simply reflected the tendency of this protein to self-associate when produced in large excess or correlated with its ability to specifically interact with cytoplasmic proteins. As a first step in distinguishing between these two alternatives, we searched for potential Nup154 partners. The protein product of the cup gene seemed to be an appealing candidate for several reasons: (1) cup and nup154 mutants share many ovarian phenotypes, suggesting that the two gene functions are involved in common developmental processes during oogenesis (Keyes and Spradling 1997; Gigliotti et al. 1998); (2) Cup forms large aggregates that are transiently localized around the nurse-cell nuclear envelope during early egg-chamber developmental stages (Keyes and Spradling 1997); and (3) Cup has been shown to be able to shuttle between the nucleus and the cytoplasm, at least in transfected Schneider cells (Zappavigna et al. 2004). To test whether Cup was able to interact with the cytoplasmic structures induced by Nup154 overexpression, we compared Nup154 and Cup protein localization in double-labeling experiments. As soon as the first cytoplasmic Nup154 aggregates appeared around stage 7–8, a simultaneous enrichment of the Cup protein was detected at the same sites. This colocalization persisted in the developmental stages that followed, leading, at stage 10, to an overall alteration of the Cup distribution pattern (Figure 3, D–F). At this stage, wild-type control egg chambers clearly localized Nup154 and Cup to different cell compartments (Figure 3, A–C). While Nup154 decorated the nurse-cell nuclear rim (Figure 3A), Cup was distributed along the subcortical cell surface in a smooth network of particulate structures (Figure 3B). In Nup154-overexpressing nurse cells, the even arrangement of the Cup-containing particles observed in the control was lost and the protein largely concentrated at the level of the Nup154 cytoplasmic foci (Figure 3, D–F).

Figure 3.—

Cup colocalizes with Nup154 in the cytoplasm of Nup154-overexpressing nurse cells. (A–C) Single confocal section of a wild-type stage 10 egg chamber stained with anti-Nup154 and anti-Cup antibodies. (A) Nup154 is confined to the nurse-cell nuclear envelope. (B) Cup accumulates in a network of particulate structures underneath the cell surface. (C) Merged image further attesting to the distinct compartmentalization of the two proteins. (D–F) Single confocal section of a stage 10 egg chamber from a female carrying two copies of the Nup154 transgene, stained with anti-Nup154 and anti-Cup antibodies. (D) Nup154 is ectopically accumulated in the nurse-cell cytoplasm. (E) Cup is distributed in big and dispersed particles. (F) These particles are largely coincident with the particles formed by Nup154. (G) Western blot probed for Nup154 after immunoprecipitation (IP) from ovarian lysate of flies overexpressing Nup154 in their germline, using different amounts of either control rat IgG or rat anti-Cup antibodies. Nup154 copurifies with Cup, but is not detected in the control lanes.

The high degree of overlap observed in the cytoplasmic staining for Nup154 and Cup in Nup154-overexpressing nurse cells suggested that the two proteins might be held together by physical interaction. To test this hypothesis, ovarian extracts from flies overexpressing Nup154 in their germline were immunoprecipitated by anti-Cup antibodies and the precipitates were analyzed by Western blotting. Figure 3G demonstrates that the anti-Cup antibodies coprecipitated Nup154, while control IgG's did not. These results indicated that Nup154 is able to associate with Cup in vivo. The alteration of the Cup distribution pattern observed when a large excess of Nup154 ectopically accumulated in the cytoplasm is therefore a likely direct consequence of the recruitment of Cup to Nup154-containing particles.

Nup154 interacts with Cup in wild-type egg chambers:

Since a special caution is necessary when considering the results of overexpression experiments, we asked whether Nup154 and Cup were able to interact also in physiological conditions. In Figure 4, A–C, the localization of Nup154 and Cup in a wild-type stage 4 egg chamber is documented by a double immunofluorescence experiment. In addition to a diffuse cytoplasmic distribution, at this developmental stage Cup displayed a marked perinuclear enrichment, in agreement with a previous report (Keyes and Spradling 1997). Interestingly, at the nurse-cell nuclear rim, the distribution patterns of Nup154 and Cup were largely coincident. This colocalization was restricted to early egg-chamber developmental stages and was no longer detectable after stage 5 (Figure 3, A–C, and data not shown). Therefore, immunoprecipitation experiments seeking a physical association between Cup and Nup154 in wild-type oogenesis were performed using ovarian extracts from newly eclosed females that contain only previtellogenic egg chambers. In these experiments, a clearly detectable, albeit small, Nup154 amount consistently and specifically coprecipitated with Cup (Figure 4D). On the basis of the immunolocalization experiments described above, this low recovery rate may be explained, considering that in wild-type conditions the amount of Cup localized to the nuclear envelope represents only a small fraction of the whole protein pool. In addition, since Nup154 is present in both somatic and germline cells, the comparison between total and Cup-precipitable Nup154 protein present in ovarian extracts is not readily informative of the fraction of Nup154 that is bound to Cup in the germ cells.

Figure 4.—

Nup154 and Cup interaction in wild-type egg chambers. (A–C) Superimposed confocal sections of a stage 4 wild-type egg chamber, stained with anti-Nup154 (A) and anti-Cup antibodies (B). The two proteins colocalize at the nurse-cell nuclear rim, as clearly shown in the merged image (C). (D) Western blot analysis of wild-type previtellogenic egg-chamber extracts immunoprecipitated with rat anti-Cup antibodies or control rat IgG. Nup154 copurifies with Cup, but is not detected in control lanes where 50 and 500 μg of control rat IgG were used. The ovarian extract loaded in the first lane was 1/10 of the amount used for immunoprecipitations.

Nup154 genetically interacts with Cup:

The results described above raised the question of whether Nup154 and Cup might also interact functionally. To address this question, we tested the genetic interaction between different mutant alleles with reduced function.

Both Nup154 and Cup are required during oogenesis for a broad range of developmental processes, including germline cyst formation, nurse-cell maturation, and oocyte growth (Keyes and Spradling 1997; Gigliotti et al. 1998). A peculiar phenotype shared by the two genes is the altered chromatin organization of mutant nurse-cell nuclei. During wild-type oogenesis, endoreplicating nurse cells develop tightly condensed polytene chromosomes that eventually dissociate and uniformly distribute their chromatin throughout the nuclei (Dej and Spradling 1999). In nup154 and cup mutants, nurse-cell chromosomes do not properly modify their structure as the egg chamber matures, but retain morphological characteristics of early developmental stages.

We performed a first genetic interaction test in double heterozygous conditions, combining two strong alleles, the nup154 lethal allele l(2)01501 and the cup⁸ allele. Morphological analyses of ovaries from double heterozygous flies did not show evident alterations in egg-chamber development. We therefore decided to look for phenotypic effects of a double homozygous combination and generated a recombinant chromosome bearing a weak cup mutant allele, cup¹³⁵⁵, and the mild nup154 tlp² allele. When tested individually, tlp² and cup¹³⁵⁵ mutant alleles have been shown to be able to sustain egg-chamber development until stage 14 (Keyes and Spradling 1997; Gigliotti et al. 1998). Accordingly, single-mutant ovarioles displayed a normal array of progressively more mature egg chambers (Figure 5, A and B). By combining one mutant copy of either gene with two mutant copies of the other gene, we were unable to detect dominant interaction effects, possibly because of the weak nature of the tested alleles. On the contrary, double-mutant ovarioles appeared as long chains of previtellogenic egg chambers that arrested their development around stage 4–5 and finally degenerated (Figure 5C). This enhancement of the mutant phenotype indicated that, in early egg chambers, Nup154 and Cup play a cooperative role essential for germ-cell differentiation.

Figure 5.—

Nup154 and Cup functionally interact during egg-chamber development. Egg chambers from single (A, D, G and B, E, H) or double (C, F, and I) mutants were stained with DAPI and analyzed by conventional epifluorescence. (A) Ovariole from a tlp² homozygote. (B) Ovariole from a cup¹³⁵⁵ homozygote. Egg-chamber maturation proceeds normally in both mutants. (C) Ovariole from a cup¹³⁵⁵, tlp² double homozygote, showing a “string of pearls” morphology due to the arrest of egg-chamber development around stage 4–5. (D) Stage 7 egg chamber from a tlp² homozygote. Slight alterations of the nurse-cell chromatin structure are present only in the two most posterior nurse cells, which also display undersized nuclei (one of them is clearly visible in this focal plane). (E) Stage 7 egg chamber from a cup¹³⁵⁵ homozygote. Nurse-cell chromosomes undergo normal dispersal in this mutant allele. (F) Stage 7 egg chamber from a cup¹³⁵⁵, tlp² double homozygote. Virtually all nurse-cell nuclei display aberrant chromatin configurations, mainly represented by “blob like” and “spiral” chromosomes. (G) Ovariole from a tlp¹ homozygote. (H) Ovariole from a cup²¹ homozygote. (I) Ovariole from a cup²¹, tlp¹ double homozygote.

The interplay occurring between the two genes also became apparent when we compared single- and double-mutant egg chambers with respect to their nurse-cell chromatin structure. In the wild type, a tight developmental control ensures that, by stage 6, all nurse-cell nuclei have properly dispersed their chromosomes. Chromosome structure was not affected in cup¹³⁵⁵ (Figure 5E), as well as in the vast majority of tlp² nurse-cell nuclei (Figure 5D), but was characterized by severe morphological alterations in cup¹³⁵⁵, tlp². In fact, stage 6–7 egg chambers, only occasionally produced in the double mutant, always displayed abnormal chromatin configurations that typically ranged from tight condensed “spiral” chromosomes to “blob-like” chromosomes (Figure 5F). While “spiral” chromosomes are never observed in wild-type nurse cells, the “blob-like” condition is transiently assumed by individual chromosome arms prior to their normal dispersal. It therefore can be inferred that, in cup¹³⁵⁵, tlp², nurse-cell chromosomes are arrested in intermediate steps of the processes regulating their morphology, as it happens in strong cup and nup154 mutants (Keyes and Spradling 1997; Gigliotti et al. 1998).

The genetic interaction between nup154 and cup was further confirmed by testing the behavior of a different allelic combination (Figure 5, G–I).

DISCUSSION

Nucleoporins are classified into two groups on the basis of the presence or absence in their sequence of FG repeat motifs, which serve as docking sites for transport receptors (reviewed in Suntharalingam and Wente 2003). Nup154 belongs to the group of nucleoporins devoid of FG repeats, considered to represent the predominant structural constituents of the NPC (Gigliotti et al. 1998). One of the two Nup154 yeast homologs, Nup170, is required for the correct positioning of FG-containing nucleoporins and other NPC components directly involved in nuclear transport mechanisms (Kenna et al. 1996; Shulga and Goldfarb 2003). In addition, it has been recently demonstrated that both Caenorhabditis elegans and Xenopus Nup154 counterparts are necessary for the accumulation of several nucleoporins at the nuclear periphery during nuclear envelope formation at the end of mitosis (Franz et al. 2005). On the basis of these data and the observation that Nup154 is an essential gene required in all cell types, it is reasonable to assume that the Drosophila nucleoporin also plays an important structural role in the context of the NPC (Kiger et al. 1999).

In our overexpression assay, we demonstrated that a vast Nup154 cytoplasmic excess is fully tolerated in differentiating germline cells, including the oocyte, but is lethal in fertilized eggs, which stop their development during the first cleavage divisions. This suggests that increased Nup154 levels do not affect cell viability per se, but specifically impair nuclear divisions. It is interesting to note that Nup155 is required for both NPC assembly and nuclear envelope formation in C. elegans embryos (Franz et al. 2005). An altered ratio generated between Nup154 and other NPC components stored in the egg cytoplasm might therefore directly or indirectly interfere with NPC biogenesis and/or nuclear membrane fusion, resulting in embryo death. On the other hand, the absence of any detectable morphological alterations in germline cells that have stopped dividing indicates that unbalanced Nup154 levels do not affect nuclear envelope growth and NPC assembly during interphase. It is of course important to consider that, in early embryogenesis, the occurrence of a very rapid sequence of nuclear divisions requires massive NPC assembly. This might be sufficient to explain the different sensitivity to Nup154 cytoplasmic accumulation displayed by early embryos and terminally differentiating germ cells. In addition, it is not known whether NPC insertion in a nuclear envelope lacking NPCs relies on the same mechanisms as NPC assembly in a nuclear envelope containing preexisting NPCs. During early embryonic nuclear divisions, the nuclear envelope does not break down, but NPCs are disassembled in prophase, dispersed in the cytoplasm, and reassembled in telophase and G₁ (Harel et al. 1989). If mechanistic differences exist between NPC assembly in the semiclosed mitosis of early embryos and NPC formation during interphase, Nup154 overexpression might have a different impact on either of the two processes. In any case, our data are in agreement with the finding that overexpression of Nup170, one of the two yeast homologs of Nup154, does not alter cell growth (Aitchison et al. 1995). Actually, in yeast cells, NPCs do not disassemble at mitosis and their formation takes place throughout the cell cycle.

By demonstrating that Nup154 genetically and physically interacts with Cup, a regulatory protein playing crucial functions during oogenesis, we have uncovered in this study a novel property of this nucleoporin. Our findings indicate that a clear-cut distinction between nucleoporins directly involved in NPC functions and structural nucleoporins might be not appropriate. More importantly, our data delineate an intriguing functional link between an ubiquitous NPC component and a cell-type-specific protein implicated in key developmental processes.

Both Nup154 and Cup have been shown to be required during oogenesis for a broad range of developmental processes, including the dynamic changes responsible for the modulation of the nurse-cell chromosome morphology (Keyes and Spradling 1997; Gigliotti et al. 1998). cup alleles have been classified in three groups on the basis of the developmental stage reached by mutant egg chambers prior to degenerating (Keyes and Spradling 1997). Class I alleles prevent oogenesis from proceeding beyond previtellogenic stages. Class II alleles cause a later developmental arrest, sustaining egg-chamber growth until mid-oogenesis. Class III alleles produce mature eggs displaying a cup shape. Several members of these classes, especially those belonging to class II, are characterized by a variety of aberrant nurse-cell chromatin configurations that may represent distinct intermediates of the chromosome reorganization process. Also, nup154 mutant alleles can be grouped with respect to their relative strength and a first distinction can be made between loss-of-function alleles, which are lethal, and hypomorphic alleles, which affect fertility in both sexes. On the basis of their impact on oogenesis, the latter alleles can be described, in turn, as severe alleles, causing previtellogenic arrest of egg-chamber development, and as weak alleles, which are able to complete oogenesis but lay morphologically altered eggs closely resembling the eggs produced by class III cup mutants (Gigliotti et al. 1998; Kiger et al. 1999). Nurse-cell chromosome dispersal defects are present in both homo- and heteroallelic combinations, albeit with varying degrees of penetrance. When we combined a class III cup mutant allele with a nup154 weak allele, we produced striking oogenesis defects that were much more severe than those observed in either homozygote alone. Egg-chamber development was blocked in the double mutant in previtellogenic stages, generating an overall phenotype closely resembling that of class I cup alleles or of severe homo- and heteroallelic nup154 mutant combinations. This result, obtained in two distinct genetic tests with different mutant alleles, indicated that Nup154 and Cup affect common developmental processes during early oogenesis and might exert their function acting in a shared pathway. The regulation of the nurse-cell chromosome structure is an integral part of the germ-cell developmental program. Accordingly, nurse-cell nuclear morphology was heavily altered in the double mutants, which retained blob-like chromosomes in post-stage 5 egg chambers. Chromosome dispersal has been proposed to take place during a mitosis-like phase that occurs only in the fifth nurse-cell endocycle (Dej and Spradling 1999). This hypothesis has been corroborated by the finding that the activity of the anaphase-promoting complex/cyclosome is required for the polytene chromosome breakdown (Kashevsky et al. 2002). It therefore appears that the transition from polyteny to polyploidy is achieved in the nurse cells through the control of crucial cell cycle regulators. Nup154 and Cup could be directly involved in this process or play an indirect role, affecting nurse-cell differentiation events that are tightly linked to the correct progression of the endomitotic program. The latter hypothesis seems more likely, since a defect in chromosome dispersal might not be sufficient to explain the oogenesis arrest observed in the double mutant. In fact, several examples exist of mutations that impair nurse-cell chromosome structure but are able to sustain egg-chamber maturation (Volpe et al. 2001; Van Buskirk and Schupbach 2002; Goodrich et al. 2004). However, the two aforementioned hypotheses are not mutually exclusive and a combinatory effect could also explain the severity of the double-mutant phenotype.

When overexpressed in the germline, Nup154 was ectopically found in the nurse-cell cytoplasm, where it associated with particulate structures that also contained Cup. While this condition had never been observed in the wild type, it was nonetheless suggestive of the ability of Nup154 to directly or indirectly interact with Cup in vivo. Physical interaction between the two proteins was in fact detected by co-immunoprecipitation in the experimental background where Nup154 was overexpressed and was also found in wild-type conditions. It has been reported that Cup is closely associated with the nurse-cell nuclear membrane during the stages when chromosomes are tightly condensed, but nothing is known about the molecular bases of this association (Keyes and Spradling 1997). Interestingly, the perinuclear Cup localization pattern largely overlaps with that of Nup154 in early oogenesis. We propose that, after having reached the nurse-cell nuclear rim, Cup interacts with Nup154 and that this interaction is essential for Cup function(s) in early germ cells. Several evidences support the hypothesis that Cup employs different mechanisms to affect different developmental processes during oogenesis and early embryogenesis. While Cup's ability to bind eIF4E has been demonstrated to be essential for oskar and nanos mRNA translational regulation, it seems dispensable for the role played in other developmental events (Wilhelm et al. 2003; Nakamura et al. 2004; Nelson et al. 2004). Actually, females expressing a truncated form of Cup that fails to bind eIF4E do not display any defects in egg-chamber maturation nor in nurse-cell morphology (Nakamura et al. 2004). Moreover, the egg-laying defect observed in the cup³/cup¹³⁵⁵ heteroallelic combination can be rescued by a mutant Cup protein that does not interact with eIF4E (Nelson et al. 2004). Our finding that Cup can physically interact with a NPC component suggests that at least some of the defects associated with mutations in the cup gene might result from alterations in a Cup molecular function taking place at the nuclear level. Further analyses will be needed to shed light on the exact nature of the mechanisms underlying the interaction between Nup154 and Cup. One possibility is that Cup interacts with Nup154 while translocating through the nuclear pore. In vitro experiments carried out in transfected Schneider cells in fact have uncovered the ability of Cup to shuttle between the nucleus and the cytoplasm (Zappavigna et al. 2004). In addition, even if Cup has never been found inside the nurse-cell nuclei, its close association with the nurse-cell nuclear envelope during early egg-chamber developmental stages supports the idea that it is engaged in nucleo-cytoplasmic exchange activities at this time (Keyes and Spradling 1997). Since no prominent alterations of the Cup distribution pattern have been observed in nup154 mutant alleles (our unpublished data), Nup154 might not play a major role in the initial docking step bringing Cup in contact with the nuclear pore, but might be involved in the translocation process.

Since a direct demonstration that Cup actually shuttles between the nucleus and the cytoplasm in the nurse cells is still missing, an alternative explanation of the interaction occurring between Cup and Nup154 during oogenesis should be also considered. It has been previously suggested that Cup may affect the structural or functional properties of the nuclear envelope (Keyes and Spradling 1997). Nup154 might directly support this activity of Cup by mediating and/or stabilizing its interaction with the NPC. When localized at the NPC, Cup in fact could influence the nucleo-cytoplasmic transport of effector molecules required for early egg-chamber development. Since NPCs have been demonstrated to play a pivotal role in genome organization (Casolari et al. 2004), Cup could also interact with Nup154 to control specific NPC features responsible for the correct execution of the nurse-cell chromosome dispersal program.