Past1 Modulates Drosophila Eye Development

Endocytosis is a multi-step process involving a large number of proteins, both general factors, such as clathrin and adaptor protein complexes, and unique proteins, which modulate specialized endocytic processes, like the EHD proteins. EHDs are a family of Eps15 Homology Domain containing proteins that consists of four mammalian homologs, one C. elegans, one Drosophila melanogaster and two plants orthologs. These membrane-associated proteins are involved in different steps of endocytic trafficking pathways. We have previously shown that the Drosophila EHD ortholog, PAST1, associates predominantly with the plasma membrane. Mutations in Past1 result in defects in endocytosis, male sterility, temperature sensitivity and premature death of the flies. Also, Past1 genetically interacts with Notch. In the present study, we investigated the role of PAST1 in the developing fly eye. In mutant flies lacking PAST1, abnormal differentiation of photoreceptors R1, R6 and R7 was evident, with partial penetrance. Likewise, five cone cells were present instead of four. Expression of transgenic PAST1 resulted in a dominant negative effect, with a phenotype similar to that of the deletion mutant, and appearance of additional inter-ommatidial pigment cells. Our results strongly suggest a role for PAST1 in differentiation of photoreceptors R1/R6/R7 and cone cells of the fly ommatidia.


Introduction
Drosophila PAST1 is an endocytic protein, which belongs to the family of Eps15 homology (EH) domain-containing proteins. The Past1 gene encodes two transcripts, RNA-A and RNA-B. The corresponding proteins, PAST1A and PAST1B, differ in seven amino acids missing from the N-terminus of PAST1B [1]. The two Past1 transcripts are differentially expressed both temporally and spatially during the life cycle of the fly [2]. Past1B is expressed throughout development in both males and females, while Past1A is only expressed from the third larval stage onwards, and is restricted to the male testes [2]. Past1 null mutant flies, generated by imprecise excision of a P-element, reach adulthood, but are temperature sensitive, infertile and die precociously. As expected, these flies present a defect in endocytosis. Past1 shows genetic a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Immunofluorescence Third instar larval eye disc or pupal and adult retinas were fixed in 4% paraformaldehyde for 20 and 60 minutes, respectively, followed by rinsing with PBT (1x PBS supplemented with 0.3% Triton X-100). First antibodies were diluted in BBT (1x PBS supplemented with 0.1% BSA, 0.1% Tween-20 and 250 mM NaCl) and were added for an overnight incubation at 4˚C with shaking. Following several washes with PBT, secondary antibodies were added and incubated with shaking for 2 hours at room temperature (RT). After several washes with PBT the preparations were mounted with Galvanol. Cells were visualized using LSM510 Meta (ZEISS) confocal microscope.
For analysis of adult or pupal eye sections a minimum of 50 individual ommatidia were scored for each genotype. When using the GMRGal4-driver, all adult, pupae and larvae used were males, to avoid background effect in females homozygous for the GMRGal4.

Endocytosis in Garland cells
Garland cells were dissected from third instar larvae and stained as previously described [2] Cells were visualized using LSM510 Meta confocal microscope.
Measurements of fluorescence intensity in individual cells were performed using the ImageJ software (http://rsb.info.nih.gov/ij/index.html).

Electron Microscopy
For scanning electron microscopy (SEM), three-days-old adult flies were decapitated and their heads were fixed in 2.5% glutaraldehyde in PBS. They were then washed, dehydrated in graded ethanol solutions, dried from CO 2 at critical point, coated with gold (Polaron SEM coating unit E5100) and examined with a Jeol JSM 840A SEM. Images were captured using scandium SEM imaging platform (Olympus).
For transmission electron microscopy (TEM), three-days-old adult flies were fixed in 2.5% glutaraldehyde in PBS. They were then washed, fixed in 1% OsO 4 in PBS, and re-washed. After dehydration in graded ethanol solutions, the samples were embedded in Glycid ether (Serva, Heidelberg, Germany). Ultrathin sections were stained with uranyl acetate and lead citrate and examined with a Jeol 1200 EX TEM. Images were captured using SIS Megaview III and digitalized with the iTEM program the TEM imaging platform (Olympus).

SDS/PAGE and western blotting
Flies (usually 10 flies in each preparation) 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-Aldrich, Israel). Proteins were separated by 10% SDS-PAGE and transferred onto a nitrocellulose membrane (Schleicher and Schuell BioScience, Keene, NH, USA). After blocking in 5% skim milk and 0.1% Tween-20 in Tris-buffered saline (TBS) for 1 hour at RT, the membranes were incubated with the primary antibody for 1 hour at RT. The membranes were then washed three times in 0.1% Tween-20 in TBS and incubated with the appropriate secondary antibody for 1 hour at RT. After additional washes, membranes were reacted with ECL detection reagents (Santa Cruz Biotechnology, CA, USA) and analyzed by luminescent image analyzer (Kodak, Medical x-ray Processor 102, Rochester, NY, USA). The intensity of each band was measured by the GelQuant software (BiochemLab-Solutions, San Francisco, CA, USA).

Data analysis and quantification
Non-parametric, Mann-Whitney Test, was used for statistical analyses, unless otherwise indicated.

Results
The effect of null Past1 mutation on fly eye development To examine the possibility that PAST1 plays a role during eye development in Drosophila, we explored the effect of a null Past1 mutation. No external morphological defects were detected by scanning electron microscopy in the eyes of three-days-old adults homozygous for a null Past1 mutation compared to wild type flies ( Fig 1C). Transmission electron microscopy analysis of Past1 mutant eyes revealed some ommatidia with a decreased number of rhabdomeres, compared to wild type control ( Fig 1D). In some of these abnormal ommatidia two potential R7 photoreceptors were observed (Fig 1D). To quantify this abnormality, staining of all the rhabdomers in late pupa eyes of both wild type and Past1 mutants was performed using phalloidin. The results showed that 41% of the ommatidia in the Past1 mutant eyes were abnormal, containing only six or five rhabdomeres (Fig 2A and 2B). We could confirm the abnormality in the number of photoreceptors in the developing Past1 mutant eye at early-mid pupal stage, by staining with anti-chaoptin antibody, which at this stage is localized to the plasma membrane of all photoreceptors [32,33]. The results (Fig 2C and 2D) indicated that 14% of the ommatidia at the early-mid pupal stage presented abnormal number of photoreceptors compared to the wild type.
The first photoreceptor to differentiate in the developing fly eye is R8 [18,34]. To determine whether PAST1 plays a role in R8 differentiation, eye discs of third instar Past1 mutant larvae were immunostained with an antibody against the R8-specific transcription factor senseless (sense) [34,35]. No abnormalities in the number of R8 photoreceptors were detected (Fig 3A), implying that differentiation of R8 is unaffected by the absence of PAST1. Since there were no indications for abnormality in differentiation of R2/R5 and of R3/R4 photoreceptors, we tested whether there are any defects in the next stage of differentiation, namely, development of R1/ R6 photoreceptors. We first stained late pupal eyes with anti-Rhodopsin 1 (Rh1) antibody, which recognizes rhabdomeres R1-R6 only [14]. The results revealed an occasional absence of either one or two photoreceptors. According to their localization within the developing ommatidia, they could have been either R1 or R6 photoreceptors, or both ( Fig 3B). More so, in some of these ommatidia, there was a duplication of what looked like R7 photoreceptor ( Fig  3B-V). To directly test whether the absence of Past1 leads to improper development of photoreceptors R1, R6 and R7, eye discs of third instar larvae of Past1 mutant were stained with anti-BarH1 antibodies. BarH1 is expressed in the nuclei of photoreceptors R1 and R6, especially during the third instar larval stage (Higashijima et al 1992; Charlton-Perkins et al., 2011). The results, presented in Fig 4A and 4B, indicated that in 10% of the ommatidia derived from Past1 mutant eye discs there was only one R1/R6 instead of two photoreceptors.
In light of the aberrant differentiation of R1/R6 photoreceptors, we wondered whether lack of PAST1 affects the differentiation of R7. To that end, early pupal eyes were immunostained with anti-prospero (pros), which specifically marks R7 nuclei as well as cone cells. To confirm visualization of the R7 confocal plane and not that of cone cells, we co-stained the samples with anti-Elav antibody such that co-localization of pros and Elav represents an R7 cell. The results (Fig 4C and 4D) showed that 8.6% of the ommatidia in early pupal Past1 mutant eyes contained two R7 photoreceptors instead of the single R7 photoreceptor normally present in each ommatidium.
We also followed the development of the four non-neuronal cone cells, which are specified after R7 differentiation in early-mid pupae. Staining with anti-cut antibody revealed that 9.1% of Past1 mutant ommatidia had five cone cells instead of the normal four ( Fig 5A and 5B). Interestingly, the improper development of the photoreceptors and cone cells in the Past1 mutant was not associated with an abnormal development of the IPC, as detected by immunostaining of early-mid and late pupa stages with anti-Discs large (Dlg) antibodies, which mark the cellular membrane of these cells (Fig 5C and 5D).
Taken together, our results show that PAST1 is important for normal development of photoreceptors R1/R6/R7, as well as of cone cells.
To demonstrate that the described Past1 null phenotype was indeed a result of PAST1 loss, a rescue experiment was performed using PAST1B transgenes (Fig 6A-6C). Homozygous null Past1 mutant flies expressing a GFP-PAST1B transgene at comparable levels in their eyes (see Fig 6A) displayed only 6% of abnormal ommatidia with less than seven rhabdomeres (Fig 6B  and 6C). This result strongly implies that the absence of PAST1 in the Past1 null mutant flies leads to abrogated development of their eyes.

Transgenic PAST1 has a dominant negative effect on eye development
To further explore the effect of PAST1 on eye development we tested transgenic flies overexpressing GFP-PAST1A or GFP-PAST1B in their eyes under the control of GMR-Gal4 (see Fig  7). While Past1 mutant flies did not display any external abnormalities in the morphology of the adult eye (Fig 1C), the eyes of flies overexpressing either PAST1A or PAST1B exhibited ommatidial fusion in various degrees of severity ( Fig 7A). The severity of eye defects was in correlation with the level of transgenic PAST1 expression (Fig 7B).
Since endogenous PAST1A is not expressed in the fly eye [2] we focused our further analysis on PAST1B. To analyze the development of photoreceptors in the presence of transgenic PAST1B, two lines were chosen, one with a high expression level of the transgene (line 2) and the other (line 1) presents a low expression level (Fig 7). Phalloidin-stained late pupal GFP-PAST1B eyes displayed abnormal number of rhabdomeres (Fig 8A and 8B). Overall, 10% of the ommatidia in the GFP-PAST1B line 1 and 34% of the ommatidia in the GFP-PAST1B line 2 flies were abnormal (Fig 8B).
Staining with anti-pros antibody in early pupal eyes of GFP-PAST1B line 1 under the control of GMR-Gal4 resulted in 9% of the ommatidia containing two R7 photoreceptors (Fig 9A   Fig 3 and 9B). We employed an additional driver, mirr-Gal4, whose expression is limited to the dorsal area of the eye, thus enabling comparison between the dorsal region overexpressing GFP-PAST1B and the ventral region, used as a control. Transgenic expression of GFP-PAST1B line 1 also led to abnormal differentiation of the cone cells. Instead of four cone cells, we observed either three or five cone cells in 10% of the GFP-PAST1B overexpressing ommatidia (Fig 9C and 9D). Abnormalities were observed in the inter-pigment cells as well, in the form of an extra IPC (Fig 9E), a phenomenon not observed in the eyes of the Past1 null mutant flies (Fig 5C and 5D).
In summary, our results indicated that overexpression of PAST1B in the fly eye leads to defects in the development of photoreceptors, cone cells and IPC cells.

PAST1 modulates endocytosis
We have previously shown that the absence of PAST1 leads to decreased endocytosis of fluorescently labeled Texas-Red avidin in Garland cells [2]. In light of this finding, we aimed to assess the effect of PAST1 overexpression on endocytosis using the same assay. As presented in Fig 10A, avidin was readily endocytosed by all wild-type Garland cells. However, Garland cells isolated from GFP-PAST1A (line 1) or GFP-PAST1B (line 1) overexpressing third instar larvae displayed attenuated endocytosis (Fig 10A). The Garland cells in these larvae could be classified into three groups: cells that endocytosed avidin similarly to wild-type cells, cells that presented reduced level of avidin endocytosis and cells that did not display any endocytosis of avidin ( Fig 10A). As summarized in Fig 10B, overexpressed PAST1 had a dominant negative effect with approximately 50%-60% attenuation of endocytosis. To summarize, our results Since Boss, expressed on the surface of R8 photoreceptor, is internalized by the neighboring R7 cell [20][21][22][23][24], we tested whether this endocytic process is also impaired in eyes of Past1 mutant and transgenic flies. To do so, we immunostained eye discs from Past1 null and PAST1 overexpressing flies with anti-Boss antibody. The results, presented in S1 Fig, indicated no change in Boss staining of R7 cells of mutant or transgenic GFP-PAST1B eye discs.

Discussion
In the present work we utilized the Drosophila eye to study the physiological role of the endocytic protein PAST1 during development. Recently it has been shown that EHD1 is necessary for proper differentiation of ocular tissues including the lens, cornea and retina in mice [29].
In the Past1 null mutant flies no external eye abnormalities were observed, but the development of photoreceptors R1, R6 and R7, as well as of cone cells, was abnormal. In flies overexpressing PAST1 in the eye external abnormalities were noted, as well as an abnormal number of photoreceptors, cone cells and IPCs. Defects in ommatidium development accompanied by no external abnormalities in the eye have been previously reported for other mutations. For example, the external eye of mutants in Muscleblind (Mbl), a regulator of alternative splicing, appear to be normal, yet tangential sections revealed that they harbored ommatidia defects [36]. Another example involves Crumbs (Crb), a transmembrane protein which is essential for biogenesis of adherens junctions and for establishing apical-basal polarity in Drosophila epithelia by downregulating endocytosis of Notch and Delta during eye development [4,[37][38][39]. External morphology of mutant Crb (crb S87-2 ) was reported to be normal. However, internal structure of adult Crb S87-2 mutant ommatidia was found to be defective with shortened and bulkier rhabdomeres, often in contact with each other rather than being distinct as in wild type eyes [38].
Interestingly in Past1 null flies, the ommatidia always contained at least one R7 photoreceptor, with some having two R7 photoreceptors with a concomitant loss of either R1 or R6 or both. These results imply that differentiation to R1 or R6 was abrogated and instead, at least one R7 photoreceptor was developed. No change in Boss staining was detected in R7 cells of eye discs from Past1 null mutants or from transgenic GFP-PAST1 flies (S1 Fig), while in Past1 null mutants the ommatidia always contained at least four cone cells and some ommatidia had five cone cells.
We have previously shown that Past1 genetically interacts with Notch in the wing [2]. The results of the present study indicated that Past1 plays a role during eye development, which is known to be regulated by Notch [17,19,27,40]. Interestingly, endocytosis of Notch seemed abnormal in early-mid pupal mutant eyes. Thus, higher level of Notch staining was detected at the vicinity of the plasma membrane of mutant pigment cells, indicating abrogated endocytosis (S2 Fig). This result implies abnormal internalization of Notch in the absence of PAST1 and hints to the possibility that PAST1 modulates internalization of Notch during eye development.
Our results well fit the model suggested by Tomlinson et al., (2011) for the development of photoreceptors R1/R6/R7 and cone cells. According to this model Notch overexpression in either R1 or R6 photoreceptors, leads to development of one of these cells into an extra R7 photoreceptor. Additionally, overexpression of Notch in R1 or R6 combined with an absence of Sev leads to the appearance of an extra cone cell [17]. We propose that lack of PAST1 elevates Notch activation in R1 and R6 photoreceptors, which subsequently develop into R7 or into cone cells (Fig 11). We hypothesize that PAST1 negatively regulates Notch signaling in R1 and R6. Since penetrance of Past1 mutation in the eye is low, PAST1 is, most probably, not a major regulator of Notch signaling, but fine-tunes it. Further experiments are needed to establish the role of PAST1 in endocytosis of Notch. Eyes of flies overexpressing PAST1 were rough with regions of fused ommatidia ( Fig 7A) and exhibited extra IPCs (Fig 9E). Mutations that abrogate the normal cell death process of IPCs have been shown to lead to rough eyes [25,41]. More so, overexpression of Notch (N FL or N ICD ) in the pupal retina led to appearance of extra IPCs [27,[42][43][44]. Therefore, we assume that the rough eyes phenotype in the PAST1 transgenic flies is due to abnormal PCD of IPCs.
In the present study we examined the ability of Garland cells overexpressing PAST1 to internalize fluorescently labeled Texas-Red avidin, as was examined for Past1 mutants [2]. Garland cells overexpressing Past1 displayed attenuated endocytosis. Overexpression of PAST1 had a dominant negative effect, showing approximately 50-60% attenuation in endocytosis (Fig 10). A similar dominant negative effect has also been previously shown in HeLa cells overexpressing the mammalian EHD2, leading to attenuation of plasma membrane internalization [45].  In conclusion, the results presented in this work highlight the importance of PAST1 for development of R1,R6 and R7 photoreceptors of the fly ommatidia.