Drosophila Past1 is involved in endocytosis and is required for germline development and survival of the adult fly Free

Endocytosis in eukaryotes mediates a variety of key cellular processes, such as maintenance of homeostasis, development, uptake of nutrients and downregulation of signal transduction (Le Roy and Wrana, 2005; Mukherjee et al., 1997), which is highly regulated by a complex network of interacting proteins. One of the protein classes involved in endocytosis is the family of Eps15 homology (EH)-domain-containing proteins (Confalonieri and Di Fiore, 2002). It includes proteins such as Eps15, intersectin and the evolutionarily conserved EHD subfamily. The EHD proteins contain an N-terminal nucleotide-binding domain, a central coiled-coil module and a single C-terminal EH domain (Lee et al., 2005). There are four members of the EHD family in mammals, EHD1-EHD4 (Mintz et al., 1999; Pohl et al., 2000). EHD1 is localized mainly in the endosomal recycling compartment (ERC) (Mintz et al., 1999; Lin et al., 2001). However, it recycles from the plasma membrane, through coated pits, coated vesicles and early endosomes, to the recycling compartment (Rapaport et al., 2006). EHD2 resides in the plasma membrane (Blume et al., 2007; Daumke et al., 2007; George et al., 2007; Guilherme et al., 2004). EHD3 resides within the tubular structures of the recycling compartment (Galperin et al., 2002) or early endosomes (George et al., 2007; Naslavsky and Caplan, 2005), whereas EHD4 appears to be located in endosomal membranes (Blume et al., 2007; Shao et al., 2002; Sharma et al., 2008).

In mammals, the existence of several members in a protein family frequently constitutes a major hurdle for genetic and phenotypic analysis of their roles, because they often have partially redundant functions. Indeed, the EHD1-knockout mice we have generated, do not display any obvious mutant phenotype (Rapaport et al., 2006). Lower organisms, such as nematodes and fruit flies, constitute an ideal alternative because they often contain single members of the family and are highly amenable to genetic analysis. Caenorhabditis elegans and Drosophila melanogaster each has a single EHD ortholog termed rme-1 and Past1, respectively (Grant et al., 2001; Smith et al., 2004). In the worm, individuals carrying either the G65R substitution (a mutation in the P-loop of the N-terminal domain) or the G429R mutation (in a residue close to the C-terminal EH domain) failed to recycle yolk receptors, which accumulated in endosomal compartments (Grant et al., 2001). A recent report demonstrated that the single Drosophila EHD ortholog Past1 is expressed ubiquitously during early embryogenesis, exhibits both plasma-membrane-associated and punctate cytosolic staining, and is capable of binding, in vitro, to the adaptor protein Numb (Smith et al., 2004). Numb is a conserved membrane-associated protein that antagonizes Notch signaling. It binds to the intracellular domain of Notch and to α-adaptin. Numb-mediated inhibition of Notch appears to require α-adaptin function, suggesting that Numb is directly involved in targeting Notch for endocytosis (Berdnik et al., 2002).

The importance of receptor-mediated endocytosis has been demonstrated in Notch receptor signaling in the fly. In the sending cell, ligand endocytosis is mandatory for its presentation as an active ligand that binds Notch on the plasma membrane of the receiving cell (Le Borgne, 2006; Le Borgne et al., 2005). In the receiving cell, Notch endocytosis leads to the activation of the cleaved Notch receptor and might lead to downregulation of the non-cleaved form of the receptor (Le Borgne, 2006; Le Borgne et al., 2005; Wilkin and Baron, 2005). After binding its ligand (Serrate or Delta in Drosophila), Notch undergoes two successive proteolytic cleavages. The first metalloprotease-dependent S2 cleavage generates an activated membrane-bound form, which is further processed by γ-secretase to release the Notch intracellular domain (Kopan, 2002). The intracellular domain is translocated into the nucleus, where, together with Suppressor of Hairless, it constitutes a transcription factor. Hairless sequesters Suppressor of Hairless and inhibits its DNA binding, thereby downregulating the function of the Notch Suppressor of Hairless complex as a transcriptional activator (Lecourtois and Schweisguth, 1997).

In the present study, we characterized the Drosophila Past1 gene and its mutants. We observed that the two Past1 gene transcripts, predicted by FlyBase (Drysdale and Crosby, 2005) (Release 5.6), are differentially expressed both temporally and spatially during the life cycle of the fly and the corresponding proteins are localized to the plasma membrane of cultured Drosophila cells. The Past1 mutants we generated reached adulthood, but died precociously. They were temperature sensitive and infertile. Garland cells from homozygous mutant larvae exhibited a marked decrease in their ability to endocytose fluorescently labeled avidin. We showed a genetic interaction between Past1 and both Notch and its repressor Hairless, suggesting a role for Past1 in the control of Notch signaling. Taken together, our results indicate that the Drosophila Past1 is involved in endocytosis and is required for germline development and survival of the adult fly.

Four EHD homologs exist in vertebrates, which regulate different stages of endocytosis (Naslavsky and Caplan, 2005; Blume et al., 2007). There is only one Drosophila EHD homolog (CG6148), designated Past1 (putative achaete scute target 1), because it was originally cloned as a putative target of achaete scute (Drysdale and Crosby, 2005).

The Drosophila database predicts two Past1 mRNA transcripts: RNA-A (2610 bp) and RNA-B (2679 bp), which encode two nearly identical proteins of 64 kDa (Past1A has 540 amino acids whereas Past1B has 6 amino acids fewer at the N-terminus of the protein) (see Fig. 1A) (Drysdale and Crosby, 2005). The predicted structure of the Drosophila Past1 protein is similar to that of the other members of the mammalian EHD family (Mintz et al., 1999), with an N-terminal domain, containing several conserved nucleotide-binding motifs (P-loop, NKxD, DxxG), a central coiled-coil region and a C-terminus containing an EH domain. Phylogenetic analysis indicates that of all the mammalian orthologs, Past1 is evolutionary closest to EHD2 (Fig. 1B) with 70% sequence homology at the protein level.

To explore the possibility of differential expression of the two Drosophila EHD (Past1) transcripts, RT-PCR was performed on RNA extracted from successive stages of development, using 5′ primers specific for each transcript and a common 3′ primer (PAST1 for RNA-A, PAST5 for RNA-B and PAST2 for both) (Fig. 1A). The results indicated that although RNA-B is expressed throughout development as well as in adults in both sexes, RNA-A is expressed only from the third larval stage onwards, and only in males (Fig. 2A). In males, the expression of RNA-A is restricted to the testes (Fig. 2B). Since tudor males, which lack germ cells (Boswell and Mahowald, 1985), do not express RNA-A (Fig. 2C), this RNA species is germ-cell specific.

Western blot analysis performed on lysates of larvae, adult males and females as well as cultured Drosophila cells, using polyclonal anti-recombinant-Past1 serum (for details see Materials and Methods), indicated that all contain a Past1 protein (Fig. 2D). Taken together, these results show that Past1 is expressed throughout the life cycle of the fly. However, the two transcripts have a different pattern of expression.

Immunofluorescence studies were used to unravel the intracellular localization of endogenous Past1 using anti-Past1 serum (see Materials and Methods). In Drosophila-derived SR+ Schneider cells there was mainly plasma-membrane staining of Past1, as detected by colocalization with myristylated RFP, a marker that stains primarily the plasma membrane (Fig. 3). There was little colocalization with GFP-rab5, a marker for early endosomes, or GFP-rab11, a marker for the ERC (Mohrmann and van der Sluijs, 1999). In the same cells transfected with UAS-GFP-Past1A or UAS-GFP-Past1B and the transcription activator actin-GAL4 vector, colocalization was detected between both GFP-Past1 isoforms and myristylated RFP (data not shown). Thus, Past1A and Past1B appear to associate mainly with the plasma membrane.

To elucidate the physiological role of Past1, we generated Drosophila mutants by mobilizing a homozygous viable P transposable element insert (EY01852), located in the first intron of RNA-A and within the 5′ UTR of RNA-B (Fig. 4A). Imprecise excision of the P-element produced four independent lines that harbored deletions in the Past1 gene. The excision of the P-element and the extent of each deletion were verified by PCR and sequencing. As seen in Fig. 4B, line Past1^60-4 carries a deletion of 1463 bp 3′ to the P-element insertion site, disrupting the second exon of RNA-A and the first exon of RNA-B. Line Past1^88-1 harbors a deletion of 776 bp 5′ to the insertion site, affecting the first exon of RNA-A and the 5′UTR of RNA-B (in this line 55 nucleotides from the 3′ end of the P-element are retained) (see Fig. 4B). Lines Past1^55-1 and Past1^110-1 have 1150 bp and 1741 bp deletions, respectively, spanning both sides of the insertion site and affecting both transcripts (Fig. 4B).

Mutants homozygous for the deletions in all four lines developed at a rate ∼20% slower than their heterozygous siblings. They reached adulthood but died precociously, three to five days after eclosion, whereas their siblings or wild-type control flies lived for over a month. This phenotype was not complemented by the chromosomal deletion Df(3R)kar-Sz37, which removes polytene bands 87C5-87D14 including band 87C6, where Past1 resides. There was no complementation between the different mutants, which is to be expected if they affect the same gene.

Expression of Past1 RNA in the mutant flies was analyzed by RT-PCR, using primers flanking the genomic deletion in each mutant line and sequencing the resultant PCR products. In each line, we found transcripts corresponding to the mutant gene. In line Past1^60-4, only RNA-A was expressed (Fig. 4C) and only in males (not shown). However, this RNA had a frame shift, which precluded the production of a corresponding protein. Interestingly, in the mutant lines Past1^55-1, Past1^88-1 and Past1^110-1 (but not in line Past1^60-4 or the wild type) the Past1 transcript was fused to RNA expressed from the upstream gene (CG14394), which is an unstudied gene (Fig. 4A,C). Since the open reading frame derived from RNA-B in mutant line Past1^88-1 remained intact, the possibility existed that Past1 protein is expressed in this mutant, provided there is an internal ribosome entry site (IRES) between the CG14394 and Past1 ORFs in the fused RNA. Similarly, the sequence of the ORF derived from the fused transcript in mutant Past1^55-1 suggested that a chimeric protein might be produced that lacks the 56 N-terminal amino acids of Past1. In the Past1^110-1 mutant, the sequence of the fused RNA harbors stop codons in all three reading frames, precluding translation of a protein from this transcript. It is worth noting that the mutant line Past1^60-4 has a normal sequence of CG14394 whereas in Past1^110-1 its ORF is not affected.

To examine whether the mutants indeed produce Past1 protein, western blot analysis was performed on lysates of adult flies from the wild type and the four mutant lines. As evident from Fig. 4D, none of the mutants produced Past1 protein. This suggests that no functional IRES exists in the transcript of mutant Past1^88-1, and that if a chimeric protein is produced in line Past1^55-1 it is highly unstable.

A characteristic of several endocytic mutants in Drosophila (e.g. shibire, stoned, comatose) is their temperature sensitivity (Poodry and Edgar, 1979; Grigliatti et al., 1973; Siddiqi and Benzer, 1976). To examine whether the survival of the Past1 mutants is affected by temperature, the balanced strains of the Past1^60-4 and Past1^110-1 mutants, which contain both homozygotes and heterozygotes, were grown throughout their lifetime at either 25°C or 29°C. The number of homozygous and heterozygous offspring in each strain was counted daily. As evident from the results (Table 1A), the proportion of the homozygotes in the population of each mutant strain tested was lower than the expected ratio of 33%. In both lines, the proportion of the homozygous mutants was significantly lower at 29°C than 25°C, regardless of the sex of the flies. For example, the proportion of Past1^60-4 homozygotes was 2% at 29°C versus 16% at 25°C (Table 1A). Survival appeared to be impaired at the high temperature even before adulthood, because monitoring the development of Past1^60-4 homozygotes from the first larval stage onwards (see Materials and Methods) revealed that only about 50% of the homozygotes that reached the pupal stage succeeded to eclose (Table 1B). The fact that in the Past1^60-4 strain itself, only 2% reached adulthood probably reflects strong competition with their heterozygous siblings in the relatively crowded bottles (see Materials and Methods). Similar results were observed for the transheterozygotes Past1^60-4/Past1^110-1 generated by crossing the two mutant strains to each other (data not shown). These findings led us to conclude that the Past1 mutants are sensitive to high temperature. The fact that all mutant lines presented the same phenotype argues that it was due to the absence of Past1 (and not CG14394).

The fertility of Past1 mutants was compromised as demonstrated by the reduced number of offspring of single pair-matings of females and males from two mutant lines, Past1^60-4 and Past1^110-1, with wild-type mates (Fig. 5A). For example, the mean number of offspring from Past1^60-4 females was about 40 times lower than that obtained from the control females.

To test whether the partial sterility of the Past1 mutants was due, at least in part, to defects in their reproductive system, we examined their testes and ovaries. The testes of the Past1 mutants (e.g. Past1^60-4/Past1^110-1) were much thinner than those of the wild type, and contained significantly fewer bundles of elongating and elongated cysts (Fig. 5B). Nevertheless, we could not detect any gross defects in Past1 mutant pre-meiotic spermatocytes and post-meiotic round spermatids. During spermatid terminal differentiation, a process known as spermatid individualization takes place. It involves extrusion of the bulk cytoplasm of the elongated spermatids into a waste bag in an apoptosis-like process that results in highly compact, viable sperm (Arama et al., 2003). The majority of mutations that affect male fertility cause defects during the spermatid individualization process (Wakimoto et al., 2004). Therefore, we examined the Past1 mutant elongated spermatids during their terminal differentiation process. For this, we stained wild-type and mutant testes for active caspase-3, which is expressed in the cytoplasm of individualizing spermatids and with Phalloidin, which binds to actin cones in the individualization complex (Fig. 5B). Past1 mutants displayed defects during the spermatid individualization process. More specifically, although an individualization complex appeared to assemble normally in the vicinity of the nuclei, Past1 mutant spermatids failed to extrude their entire cytoplasm into a waste bag, leaving trails of cytoplasm in what should have been the post-individualized region of the spermatids (marked by active caspase-3) (Fig. 5B). Consequently, Past1 cystic bulges and waste bags were reduced in size or appeared flat (Fig. 5B). Additionally, although the nuclei of the wild-type spermatids bundled at one edge (Arama et al., 2003), the nuclei of some of the Past1 mutant spermatids appeared scattered (data not shown). According to these results, semi-sterility of the Past1 mutant males is expected.

The ovaries of mature mated homozygous mutants as well as transheterozygous females were highly abnormal compared with ovaries of mated wild-type and heterozygous females of the same age (Fig. 5C; Table 2). They were much smaller, less organized and contained fewer well-differentiated oocytes (stages 12-14) than those in the control. For example Past1^110-1 homozygotes contained on average 5.5 egg chambers at stage 9-11 and only 1.6 fully developed chambers (stage 12-14), whereas the control heterozygous siblings contained 7.7 egg chambers and 9.8 fully developed chambers. Transheterozygotes displayed a similar, although less severe, abnormality. In addition, whereas yolk deposition into the oocyte was evident in all egg chambers beyond stage 8 during the normal oogenesis seen in heterozygotes, in the Past1 mutants, all oocytes from stage 9 onwards were devoid of yolk. This indicates that the process of oogenesis is highly impaired in Past1 mutants. Taken together, our results imply that Past1 is required for the development of germ cells in Drosophila.

To gain insight into the role of Past1 in Drosophila development, we performed crosses between Past1 mutants and mutants of the Notch signaling pathway. The Notch signaling pathway, extensively studied in Drosophila, is pivotal for many developmental processes and is regulated by endocytosis (Le Borgne et al., 2005; Le Borgne, 2006; Wilkin and Baron, 2005). We found a genetic interaction between Past1 and Notch in flies carrying either the Past1^60-4 or Past1^110-1 homozygous mutation in addition to the Notch N^nd-1 allele (hemizygous males and homozygous females). The wing scalloping of N^nd-1 is thought to be due to reduced Notch signaling (Royet et al., 1998). The N^nd-1; Past1^–/– (Past1^60-4 or Past1^110-1 homozygotes) flies reached adulthood at a very low frequency, died after 1 day and displayed a more severe wing notching phenotype than the N^nd-1 mutant alone (Table 3A), although wings of Past1 homozygous mutants were completely normal. Genetic interaction was also found between Past1 and Hairless mutants. Hairless is a repressor of the Notch signaling pathway. Its binding to Suppressor of Hairless inhibits the DNA binding and transcription activation by the Notch Suppressor of Hairless complex. Double heterozygotes of the dominant loss-of-function mutant H² and either Past1^60-4 or Past1^110-1 showed a less defective wing venation phenotype than the H² mutant alone (Table 3B). The occurrence of shortened L2 and L4 longitudinal veins was much lower (6%) in wings of the double heterozygotes than in wings of H² heterozygotes (36%) (Table 3B). Past1 heterozygotes had normal wing vein venation. These results implicate Past1 in the control of developmental processes associated with the Notch signaling pathway.

The Drosophila Past1 has four mammalian homologs, EHD1-EHD4. EHD1, the most studied EHD, is localized mainly to the ERC and regulates recycling. Expression of its dominant-negative mutants in tissue culture cells, its RNAi, or its absence in animal models led to attenuated recycling (George et al., 2007; Naslavsky et al., 2004; Rapaport et al., 2006). EHD2 is a plasma-membrane protein that controls actin cytoskeleton reorganization. Its overexpression results in a marked decrease of internalization (Guilherme et al., 2004) (S.B. and M.H., unpublished). EHD3 regulates trafficking from the early endosomes to the ERC (Naslavsky et al., 2006) whereas EHD4 has been shown recently to control trafficking at the early endosomes and regulates exit of cargo toward the recycling compartment and the late endocytic pathway (Sharma et al., 2008).

To test whether Past1 has a role in endocytosis, we assessed the ability of normal and mutant Garland cells to endocytose fluorescently labeled avidin. Larval Garland cells are similar to nephrocytes, have a rapid rate of fluid phase endocytosis, and are an established system for the study of endocytic dynamics (Chang et al., 2002; Dermaut et al., 2005; Kosaka and Ikeda, 1983). We first tested the ability of normal Garland cells to endocytose Texas-red-labeled avidin. We did so by incubating wild-type Garland cells with Texas-red-avidin, after which the cells were transferred onto ice and incubated for 5 minutes with fluorescein-labeled avidin, to allow its binding to the membrane. As shown in Fig. 6 and Table 4, fluorescein-labeled avidin marked the membrane whereas Texas-red-avidin staining was seen inside the cells. We then tested the ability of Past1 mutant Garland cells to endocytose fluorescently labeled avidin. As shown in Fig. 6 and Table 4, avidin was readily endocytosed by all wild-type Garland cells. However, Garland cells isolated from Past1 mutant larvae displayed attenuated endocytosis. Mutant Garland cells could be divided into three groups: cells that endocytosed avidin similarly to wild-type cells; cells that presented reduced endocytosis of avidin, and cells that did not display any endocytosis. Our results strongly suggest that Past1 is directly involved in endocytosis.

We used Drosophila as a model to study the EHD homolog, Past1. Two transcripts, encoding two nearly identical proteins, are transcribed from the Past1 gene. One transcript is expressed throughout development, as well as in adults, in both sexes; the other transcript is expressed only from the third larval stage onwards, and only in the male testes. The differential expression of these two transcripts suggests that they are under distinct control. Although the regulatory regions of the gene remain to be identified, there is probably a separate promoter for each transcript. Since they do not share an upstream region, alternative splicing cannot account for the two transcripts (our unpublished results). The fact that Past1A is expressed in testes only from the third larval stage onwards, suggests that this protein must be precluded from these cells in early developmental stages.

Localization of endogenous Past1, as well as of the two Past1 proteins expressed from plasmids, to the plasma membrane in Schneider cells, suggests a role in early stages of endocytosis, similar to EHD2 (Guilherme et al., 2004) (S.B. and M.H., unpublished). Moreover, there is not only functional resemblance between Past1 and EHD2, but also structural homology, as suggested by their evolutionary conservation (Fig. 1B).

Results of endocytosis assays, in which we followed the fate of fluorescently labeled avidin in larval garland cells, showed a remarkable attenuation in internalization of the marker in Past1 mutant Garland cells. It is tempting to postulate that the higher the proportion of cells defective in endocytosis, the lower the likelihood of the larvae to survive. Indeed, we observed varying degrees of viability among larvae with the same Past1 mutation. These results strongly suggest that Past1 is involved in endocytosis.

The mutants that we generated did not express Past1 protein and exhibited temperature sensitivity, infertility and death at early adulthood. All Past1 mutants expressed abortive transcripts. Our results indicate that the Past1 RNA expressed from the Past1^55-1, Past1^88-1 and Past1^110-1 mutants, initiated from the promoter of the upstream CG14394 gene. Temperature sensitivity characterizes endocytic mutants of two types: missense and null mutants. Missense mutations cause the mutant protein to lose its correct structure and/or function (for example shibire) at high temperature (Grant et al., 1998). The absence of a protein because of a null mutation such as Dap160 or Past1 causes the complex in which it is normally involved, to become less stable or unstable, especially at high temperature (Koh et al., 2004).

Past1 mutants exhibited impaired gonads and a significant reduction in fertility. Mutant male sperm differentiation is derailed at the individualization step, and mutant females have ovaries that contain yolkless egg chambers. Yolk is produced in the nurse cells and is transferred into the oocyte at stage 8 onwards. The lack of yolk in a large number of the mutant oocytes suggests that Past1 is required for yolk trafficking from the nurse cells into the oocytes. Interestingly, other fly mutants defective in endocytosis also display male or female sterility, and the latter is associated with abnormal oogenesis, e.g. yolkless, quitPK61, stillPS34, Merlin and clathrin-heavy-chain 4 (Bazinet et al., 1993; DiMario and Mahowald, 1987; Dorogova et al., 2008; Gutzeit and Arendt, 1994), and mutants in the C. elegans EHD ortholog rme-1 failed to recycle yolk receptors (Grant et al., 2001). These results highlight the importance of Past1, and endocytosis, in development and maturation of the germlines.

The transition from the third instar larva to pupa is the stage at which the primordial germ cells are selected to become germline stem cells (Gilboa and Lehmann, 2004). This might be the trigger for expression of the testis-specific RNA-A transcript. The fact that both male and female germ cell development are impaired in the absence of Past1 expression strongly implies that both Past1 proteins have a role in early germ cell differentiation.

The possible involvement of Past1 in Notch signaling led us to test genetic interactions between Past1 mutants and mutants in the Notch signaling pathway. Mutated Past1 in flies carrying the Notch N^nd-1 allele caused an enhanced wing notching phenotype compared with that of the N^nd-1 mutants alone. Additionally, reduced expression of Past1 ameliorated the venation defects of the haploinsufficient H² mutant in the Hairless gene.

Notch has an inductive role in the development of the wing margin (Portin, 2002). Notch signaling between the dorsal and ventral compartments of the future wing specifies the wing margin – a line of cells that organizes the outgrowth of the wing. Loss of Notch signaling eliminates the wing margin and wing tissue, producing notched wings (Lai, 2004). Hairless is known to sequester Suppressor of Hairless and thus inhibit its ability for DNA binding and transcription activation together with Notch. Our results show that the Notch N^nd-1 phenotype was augmented by the absence of Past1 and the severity of the Hairless H² phenotype was diminished by the reduction of Past1 expression. These observations strongly suggest that Notch and Past1 function synergistically, whereas Hairless and Past1 are antagonistic, suggesting that Past1 promotes Notch signaling.

It is well documented that endocytosis has a crucial role in Notch signaling, both in the sending cells and in the receiving cells (Emery et al., 2005; Le Borgne, 2006; Polo and Di Fiore, 2006). In the sending cell, endocytosis of the ligand (Delta or Serrate in Drosophila) is required for its activation, whereas in the receiving cell, endocytosis is central to activation of the cleaved Notch receptor or downregulation of the membranous, non-cleaved form (Le Borgne, 2006; Wilkin and Baron, 2005). The fact that Past1 synergizes with Notch in its effect on the wing margin suggests that Past1 is required as an activator of Notch signaling. We suggest that in the absence of Past1 in the sending cell there is less active ligand, leading to less Notch activation. The antagonistic effect of Past1 on Hairless, presumably due to abrogation of ligand-dependent Notch activation in the sending cell, leads to partial restoration of vein formation. It is still possible that Notch undergoes Past1-regulated recycling in the receiving cell.

Notch signaling has an important role during oogenensis as well as during spermatogenesis (Polesello and Tapon, 2007; Shcherbata et al., 2004; Song et al., 2007; Xu et al., 1992). This can explain the sterility of both sexes in the Past1 mutant flies.

In conclusion, lack of Past1 renders flies partially infertile and temperature sensitive, and shortens their life expectancy. Furthermore, Past1 mutants enhance the effect of the Notch mutation on wing scalloping and ameliorate the effect of the Hairless mutation on wing vein formation, implicating Past1 in the control of endocytosis of Notch ligands in a key developmental signaling pathway. Our results provide the first evidence for a physiological role of Past1 proteins, which are members of the EHD protein family, in development.

Drosophila Schneider SR+ cells were cultured in Schneider medium (Beit Haemek, Israel) supplemented with 10% fetal calf serum at 25°C.

For transfection, plasmid DNA was introduced into Drosophila Schneider SR+ cells, grown on coverslips, using the ESCORT IV transfection reagent (Sigma). Actin-GAL4 expression vector was a kind gift from T. Volk at the Weizmann Institute of Science, Rehovot, Israel. UAS-GFP-rab5 and UAS-GFP-rab11 expression vectors were a kind gift from B. Shilo also at the Weizmann Institute of Science. UAS-myr-RFP was a kind gift from Henry Chang, Yale University School of Medicine, New Haven, CT (Chang et al., 2002). GFP-EHD1 was previously described (Mintz et al., 1999).

For immunofluorescence studies, cells were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100 18 hours after transfection. Following incubation with 20% normal goat serum and 1% BSA in PBS, to block nonspecific interactions, the cells were incubated with the appropriate antibodies. Following washing in PBS, the cells were incubated with secondary antibodies and, after several additional PBS washes, the coverslips were mounted on glass slides with galvanol. The images were obtained using either an LSM510 confocal microscope or an LSM510 Meta confocal microscope (Carl Zeiss, Germany).

For RNA extraction, embryos, larvae, pupae, adult flies or testes from larvae and adults, were collected, frozen in liquid nitrogen and then homogenized in TRI reagent solution (MRC, Cincinnati, OH). The extraction was performed according to the manufacturer's recommendations. 1 μg RNA was reverse transcribed, using M-MLV reverse transcriptase kit (Promega, Madison, WI), in the presence of random hexamer primers and RNAsin (Promega). The resulting cDNA was used to amplify Past1 transcripts with Ready Mix Taq polymerase kit (Bio-Lab, Jerusalem, Israel) and the following primers: PAST1, 5′-GCTAGCCGCGTTTTAGTG-3′; PAST2, 5′-CTTCGCATCGAAGTCTGGATCCT-3′; PAST3, 5′-GTGGCAAGCAAGTGTA-3′; PAST4, 5′-GGGCATGCTGGACATCGC-3′; PAST5, 5′-GGCGTGGTCACGTTGATTG-3′; PAST6, 5′-CGCCTCCGTGACATCCG-3′; PAST7, 5′-CCAGGTATACGATCGCATCCAG-3′; PAST8, 5′-CATGGTCGTCGATGACTCC-3′. The positions of the primers used are shown in Fig. 1A and Fig. 4A. Thermal cycling consisted of 94°C for 10 minutes, followed by 30 cycles of denaturation (94°C, 1 minute), annealing (56°C, 1 minute), extension (72°C, 1 minute), and a final extension at 72°C for 10 minutes.

For each preparation, 10 flies (or 5×10⁶ S2 Schneider cells) were homogenized in RIPA lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS) containing protease inhibitors (10 μg/ml leupeptin, 10 μg/ml aprotinin and 0.1 mM phenylmethylsulfonylfluoride, all from Sigma). 100 μg protein was separated by 8% or 10% SDS-PAGE and transferred onto a nitrocellulose membrane (Schleicher and Schuell, Dassel, Germany). After blocking in 5% skimmed milk and 0.1% Tween 20 in Tris-buffered saline (TBS) at 4°C overnight, the membranes were reacted overnight with rabbit anti-Past1 serum diluted 1:1000. Following washes in TBST (TBS containing 0.1% Tween 20), the membranes were reacted with HRP-conjugated goat anti rabbit antibodies for 1 hour (diluted 1:10,000; Jackson ImmunoResearch, West Grove, PA). After additional washes, proteins were detected using Western blotting Luminol Reagent (Santa Cruz Biotechnology, Santa Cruz, CA).

A BamHI-SalI fragment of the ORF of Past1 (lacking sequences encoding 54 amino acids from the N-terminus and 44 amino acids from the C-terminus of the protein), was cloned into the pET29b expression vector (Novagen, San Diego, CA) at its BamHI-XhoI sites. Clones of transformed E. coli (strain DE3), after IPTG induction at 30°C, were lysed and extracts were analyzed on 10% SDS-PAGE. Extracts from a highly expressing clone were purified using the QIAexpressionist Ni-NTA protein purification system according to the manufacturer's recommendations (Qiagen, Hilden, Germany). Polyclonal anti-Past1 antibodies were prepared by immunizing rabbits with five injections of 0.5-1 mg purified protein at 4-week intervals. Animals were bled 10 days after each boost. Serum was prepared from the blood and stored at –80°C.

Canton-S flies (CS) served as a wild-type control. Strains were maintained and crosses were made on cornmeal molasses medium at 25°C, taking special care to avoid overcrowding. Strain EY01852, containing a P-element in the Past1 gene, was obtained from the Bloomington Stock Center (strain #15082). A strain carrying the stable source of P-transposase (Δ2,3) was used to mobilize the P-element. The resulting mutants were balanced over the TM3, Ser act-GFP balancer, to allow identification of homozygous mutants at the larval stage. A strain containing a chromosomal deletion that removes Past1 and adjacent genes [Df(3R)kar-Sz37, Bloomington #6306] was used to confirm the lesions generated in the Past1 gene by the mobilization of the P-element. The Notch allele N^nd-1 (strain #2209) and the Hairless H² allele (strain #517) are from Bloomington.

The P-element in the EY01852 strain carries the dominant w⁺ eye color marker. Reciprocal crosses of this strain and a strain carrying the Δ2,3 stable source of transposase were conducted. Imprecise excisions of the P-element were verified in F2 by PCR amplification and sequencing.

For survival measurements, 25 males and 25 females from balanced Past1 mutant lines were allowed to mate and lay eggs at 25°C or 29°C for 7 days, after which they were discarded. The resulting adult offspring were separated into homozygotes and heterozygotes, according to the Stubble marker on the TM3, Sb balancer, and were counted daily for a week after they had started to eclose. The experiment was repeated at least four times. For assessing survival rate during development, first instar larvae from the Past1 mutant strain were collected and separated into homozygotes and heterozygotes, according to the GFP marker on the TM3, Ser act-GFP balancer. Larval development was monitored daily and the number of survivors in every developmental stage was counted. At least 20 homozygous larvae were collected. The experiment was repeated twice.

To assay female fertility, a virgin Past1 mutant female and a virgin Canton-S male (or a pair of Canton-S male and female as control), collected within 8 hours of eclosion, were placed together in a culture vial for 3 days, at which time the pair was removed from the vial. The vial was cultured for an additional 17 days, and the number of offspring that eclosed from the vial during that time was counted. To assay male fertility, a virgin Canton-S female and a virgin Past1 mutant male were placed together, following the same procedure that was used for female fertility. The number of offspring produced per vial was counted until 20 days after mating for vials in which the parental pair survived the 3-day breeding period. The data was analyzed statistically: Two outliers, were eliminated using the Grubbs test (α=0.01) (Grubbs, 1969), and were not included in the mean. The mean rank of each experimental group was compared with the mean rank of Canton-S using the nonparametric Dunn test (Dunn, 1964).

Testes of young adult males were dissected, fixed and stained as described (Arama et al., 2003) with anti-active mouse caspase-3 (Cell Signaling, Danvers, MA), TRITC-phalloidin (Sigma) and DAPI (Vector Laboratories, Burlingame, CA). Ovaries of 3-day-old mated adult females were dissected, separated from extra-ovarian tissue, spread on a microscope slide and visualized under a light microscope. Wings were dissected from young adults, mounted in Hoyer's medium and visualized under a light microscope.

Garland cells were dissected from third instar larvae and incubated with 10 μg/ml of Texas-red-conjugated avidin (A-820, Molecular Probes) in PBS for 1 minute at 25°C, in six-well plates. The cells were washed with PBS, chased for 20 minutes in PBS at 25°C and fixed with 4% paraformaldehyde. The cells were transferred onto glass slides, which were mounted with gavanol reagent and covered with coverslips. Cells were visualized with a LSM510 Meta confocal microscope. The ImageJ program was used to quantify the fluorescence intensity of individual cells.