|
|
Dev Biol. Author manuscript; available in PMC 2008 July 15. Published in final edited form as: Published online 2007 May 3. doi: 10.1016/j.ydbio.2007.04.037. | PMCID: PMC2140239 NIHMSID: NIHMS27385 |
Spitz from the Retina Regulates Genes Transcribed in the Second Mitotic Wave, Peripodial Epithelium, Glia and Plasmatocytes of the Drosophila Eye Imaginal Disc Lucy C. Firth1 and Nicholas E. Baker1$ 1 Department of Molecular Genetics, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461 Proliferation, differentiation, and other processes must be coordinated during the development of multi-cellular animals. A discrete and regulated cell division, the Second Mitotic Wave (SMW), occurs concomitantly with early cell fate decisions in the Drosophila developing retina. Signals from the Epidermal Growth Factor Receptor (EGFR) are required to promote cell cycle arrest of specified cells and antagonize S-phase entry in the SMW. Cells that do not receive any EGFR activity enter S-phase in the SMW in response to the Notch pathway. To identify genes with potential roles in the SMW, we used microarrays and genetic manipulation of the EGFR pathway to seek transcripts regulated during the SMW. RNA in situ hybridization of 126 differentially transcribed genes revealed genes that have novel expression patterns in cells closely associated with the SMW. In addition, other genes’ transcripts were regulated in the differentiating photoreceptor cells, retinal basal glia, the peripodial epithelium, and blood cells (plasmatocytes) associated with the developing retina. These novel targets suggest that during eye development, EGFR activity coordinates transcriptional programs in other tissues with retinal differentiation. The development of multicellular animals requires the co-ordination of multiple signals that control proliferation and differentiation. Often, the extracellular signaling pathways required for differentiation also regulate the cell cycle. How these signals are choreographed together to produce properly sized and patterned tissue is just beginning to be understood ( Baker, 2007). In the Drosophila eye, the cell-cell signals that control differentiation have been well characterized ( Zipursky and Rubin, 1994; Freeman, 1997; Greenwood and Struhl, 1999; Curtiss and Mlodzik, 2000; Simon, 2000) and several of these signals are also required to regulate proliferation in the developing retina ( Penton et al., 1997; Horsfield et al., 1998; Baonza and Freeman, 2005; Firth and Baker, 2005). The retina and other limbs each develop from epithelial sacs or imaginal discs ( Auerbach, 1936; Cohen, 1993). Each imaginal disc is made of 2 opposing epithelial layers. The layer that forms the adult eye or limb is composed of columnar cells, the disc proper (DP) and overlaying the DP is a layer of squamous epithelial cells, the peripodial epithelium (PE) ( Cohen, 1993) (). Although the PE does not form part of the adult eye, signals from the PE to the DP are important for retinal development ( Cho et al., 2000; Gibson and Schubiger, 2000). Differentiation starts at the posterior edge of the presumptive retinal epithelium and progresses anteriorly ( Wolff and Ready, 1993). The front of this wave of differentiation is marked by the apical constriction of cells called the morphogenetic furrow (MF). The MF separates the undifferentiated and differentiating portions of the eye. Anterior to the MF, cells are randomly proliferating. Just ahead of the furrow all cells undergo a prolonged G1 arrest during which time signals to specify early cell fate decisions are received and neurogenesis begins ( Thomas et al., 1994). Some cells then re-enter the cell cycle and undergo a single round of proliferation, the Second Mitotic Wave (SMW) ( Wolff and Ready, 1991). Not all cells enter the SMW however. Groups of cells remain arrested in G1, permanently withdraw from the cell cycle and differentiate ( Wolff and Ready, 1993). During eye development signals from the Epidermal Growth Factor Receptor (EGFR) are required at numerous stages. First, EGFR is activated by the ligand Spitz (Spi) in four cells surrounding the founding photoreceptor cell R8, resulting in the G1 arrest and recruitment of these four cells (R2, R5, R3 and R4) into the photoreceptor precluster ( Freeman, 1994; Dominguez et al., 1998; Kumar et al., 1998; Baker and Yu, 2001). The surrounding cells, without any EGFR activity, reenter the cell cycle and perform DNA synthesis. After S-phase, progression from G2 phase into mitosis requires EGFR activity. EGFR is activated in G2 cells that are in contact with precluster cells by Spi ( Baker and Yu, 2001). Later expression of Spi recruits these post-mitotic cells into the remaining retinal cell fates ( Freeman, 1996). Interestingly, Spi protein travels down the photoreceptor axons into the brain lamina, where it triggers the differentiation of the synaptic cartridge units ( Huang et al., 1998). Retinal basal glia cells are closely associated with the developing photoreceptor axons and are important for their guidance down the optic stalk to the brain lamina ( Choi and Benzer, 1994; Rangarajan et al., 1999). Also, larval hemocytes/plasmatocytes are found on the outer surface of the eye disc. Both of these cell types are of a different linage to the imaginal disc. Only some gene targets of EGFR signaling have been identified. The canonical Ras/MAPK pathway downstream of EGFR transduces the EGFR signal to the ETS transcription factor, Pointed (Pnt) ( Brunner et al., 1994; O Neill et al., 1994; Yang and Baker, 2003). Phyllopod also appears to be a target during photoreceptor differentiation ( Chang et al., 1995; Dickson et al., 1995; Wassarman et al., 1995). EGFR signaling targets the Cdc25 homolog, String ( Stg), for progression from G2 to mitosis in the SMW ( Baonza et al., 2002; Baker and Yu, 2001). The signal to enter S-phase in the SMW is dependant on the Notch (N) pathway; cells defective for N signaling are unable to progress from G1 phase into S-phase ( Baonza and Freeman, 2005; Firth and Baker, 2005). In the absence of both EGFR and N, cells remain in G1 despite lacking EGFR activity. Both N and EGFR signaling are thought to regulate S-phase entry through transcriptional targets, but these targets remain unknown ( Firth and Baker, 2005). To understand the events surrounding the SMW we have taken a genome wide approach to isolate genes expressed within and around the SMW. Other groups have analyzed the gene expression profile during wild type eye development using microarrays or SAGE analysis of FACs sorted cells ( Jasper et al., 2002; Klebes et al., 2002; Michaut et al., 2003) and showed that many hundreds of genes are transcriptionaly upregulated during eye development. It is likely that many are targets of EGFR or N regulation. We designed a microarray gene expression screen to identify genes regulated specifically during the SMW, rather than all the targets of EGFR or N signaling during eye development. The SMW was eliminated genetically through the manipulation of the EGFR pathway and the gene expression profile was compared to similar retinas with a SMW. This approach gave a number of candidate genes. We then determined the mRNA expression patterns of differentially transcribed genes by RNA in situ hybridization. We have identified genes whose expression is regulated by EGFR signaling that are transcribed in cells participating in the SMW or the cell cycle arrested differentiating cells as expected, but also many genes transcribed in the PE, glia and the plasmatocytes. We think that our specific strategy also selected for genes that EGFR regulates in other tissues in response to the eye disc. RNA isolation, probe preparation for array and data analysis Total RNA was extracted from 100 eye/antennal discs using Trizol and analyzed for quality on the Agilent Bioanalyzer ( www.chem.agilent.com). For both GMR>RasV12 and GMR>sSpi, 3 independent samples of total RNA were prepared for the GeneChip® arrays according to the manufacturer’s specifications and hybridized to DrosGenome1 expression arrays ( www.affymetrix.com) (Platform accession no. GPL72). The raw data reported in this paper has been submitted to the NCBI Gene Expression Omnibus, www.ncbi.nlm.nih.gov/geo (accession series no. GSE6300). Data analysis was performed using Microarry Suite 5.0 (MAS 5.0), the Data Mining Tool software (Affymetrix) and Microsoft Excel. Single array analyses for each array were performed in MAS 5.0. All nine comparison replicates were performed in the Data Mining Tool with GMR>sSpi as the baseline and GMR>RasV12 as the experiment. A Signal Log Ratio (SLR) greater than 1 is the same as a Fold Change of 2 ( Wodicka et al., 1997). Gene expression changes that satisfied both the T-Test and the non-parametric Mann-Whitney Test (p <0.05) were used to evaluate gene expression changes between the GMR>RasV12 and GMR>sSpi ( Affymetrix, 2003). A cut off was applied and genes that consistently had a Signal Log Ratio of 1 or greater were deemed significantly upregulated in GMR>RasV12 (or downregulated in GMR>sSpi) and those with a Signal log ratio less than −1 were deemed significantly downregulated in GMR>RasV12 (or upregulated in GMR>sSpi). RNA in situ probe design, preparation and hybridization RNA probes were designed against the contiguous cDNA sequence of differentially expressed genes. A PCR strategy for rapid generation of template DNA for synthesis of labeled RNA probes was used. Probes were designed to be between 400–800 base pairs long and correspond to unique sequences of the cDNA as determined by Blast ( http://flybase.net/blast). Primers were designed using the Primer3 program ( http://frodo.wi.mit.edu/cgi-bin/primer3/primer3). The following linkers were added to the 5′ end of each primer: Forward Primer: 5′ GGCCGCGG 3′; Reverse Primer: 5′ CCCGGGGC 3′. Primers for each sequence amplified can be found in the supplementary table 1. PCR was checked by gel electrophoresis. If a single product was obtained, 5μl of the PCR reaction was treated with ExoSAP-IT (USB Corp. #78202). 1–2μl of this was used for a second PCR reaction with universal primers containing sequences that hybridize to the linker sequence and promoter sequences for the T7 and SP6 polymerases. Universal forward primer sequence (adds T7 promoter and EcoR1 restriction site for subcloning): 5′GAGAATTCTAATACGACTCACTATAGGGCCGCGG 3′. Universal reverse primer sequence (adds SP6 promoter and a BamH1 restriction site for subcloning): 5′AGGGATCCATTTAGGTGACACTATAGAACCCGGGGC 3′. A standard PCR program with an annealing temperature of 45°C was used. The second PCR products, containing the SP6 and T7 promoter sequences, were cleaned up by gel extraction. Sense (T7) and anti-sense (SP6) RNA DIG probes were made directly from this PCR product (Roche DIG RNA labeling Kit #1 277 073). In cases where the first PCR reaction gave multiple bands, the correct band was gel extracted before the second PCR reaction. DIG probe was precipitated with LiCl and re-suspended in 100μl of 50% Formamide in HSW solution (see below). RNA in situ hybridization was performed as described ( Cornell et al., 1999). Immunofluorescence Labeling of eye discs was performed as described ( Firth et al., 2006). Preparations were examined on the BioRad Radiance 2000 Confocal microscope. Images were processed using Adobe Photoshop 4.0 and NIH Image J software. Rabbit anti-phosphoHistone3 was from Cell Signaling Technology (#9701). Rabbit anti-Caspase 3 (CM1) ( Srinivasan et al., 1998). Anti-MDP-1 ( Hortsch et al., 1998). Anti-arm (N2 7A1), anti-Dlg (4F3), anti-Repo and anti-ElaV were from the Developmental Studies Hybridoma Bank, maintained by the University of Iowa, Department of Biological Sciences, Iowa City IA52242, USA under contract N01-HD-7-3263 from the NICHD. Screen design and strategy for identifying genes expressed in the SMW during Drosophila eye development Using the Drosophila developing eye a microarray gene expression screen was designed to identify genes involved in the SMW. Since the cells in the SMW could not reliably be dissected out from the surrounding tissue, an approach where the SMW was ablated genetically was taken. The SMW was blocked in the developing eye by expressing the EGFR activating ligand, Spitz (Spi). Spi was over expressed in all cells in the differentiating part of the retina posterior to the MF using the UAS-Gal4 system; GMR>sSpi ( GMRGal4/+; UAS sSpi/+). GMR Gal4 is active in all cells posterior to the MF ( Hay et al., 1994; Freeman, 1996). Expression of sSpi promotes G1 arrest and as a result there is no SMW ( Yang and Baker, 2003). Due to increased activity of the EGFR/Ras pathway ectopic differentiation also occurs () ( Freeman, 1996). The program of eye development and photoreceptor differentiation leads to the induction of many genes ( Jasper et al., 2002; Klebes et al., 2002; Michaut et al., 2003). To circumvent the expected enrichment of all EGFR targets throughout the retinal eye field in GMR>sSpi, the gene expression profile of GMR>sSpi eye/antennal discs was compared to that of GMR>RasV12 ( GMRGal4/UAS RasV12), rather than to wild type discs. Hyper-activation of the EGFR pathway with oncogenic Ras V12 results in a similar ectopic differentiation phenotype; the effects of GMR>RasV12 are autonomous, however and the EGFR/Ras/MAPK pathway is activated after the SMW begins, so that GMR>RasV12 does not prevent the SMW () ( Baker and Yu, 2001). Due to the absence of cells participating in the SMW, we hypothesized that transcripts of genes expressed in the SMW would be downregulated in GMR>sSpi when compared to GMR>RasV12. There also might be other differences between GMR>sSpi and GMR>RasV12, such as differences in genes specific to the early columns of retinal specification that occur during the SMW. For example, GMR>sSpi contains very few R8 photoreceptors, whereas normal numbers of R8 cells are specified in GMR>RasV12 ( Lesokhin et al., 1999; Baker and Yu, 2001). The mRNA expression profiles of GMR>sSpi and GMR>RasV12 whole eye/antennal imaginal discs were determined using oligonucleotide arrays representing approximately 13,500 known and predicted genes in the Drosophila genome (GeneChip® expression array DrosGenome1, Affymetrix). For both genotypes, 3 independent RNA samples were prepared and hybridized to the arrays. Using Affymetrix Data Mining Tool Software whole genome gene expression changes between GMR>sSpi and GMR>RasV12 were evaluated ( Affymetrix, 2003). 140 genes were differentially transcribed between GMR>sSpi and GMR>RasV12. 72 of these genes were downregulated and 68 were upregulated in GMR>sSpi compared to GMR>RasV12 (). Characterized genes associated with eye development Known transcripts downregulated in GMR>sSpi compared to GMR>RasV12 72 transcripts were downregulated in GMR>sSpi compared to GMR>RasV12. 14 of these genes have been previously characterized during eye development. Because the SMW is absent in GMR>sSpi, we expected genes expressed in the SMW cells to be included in this group. As expected 4 of the 14 genes characterized previously are known to be expressed more highly in the cycling cells of the SMW and the in the unspecified cells i.e. in cells for 2/3 columns at the posterior edge of the MF: lozenge ( lz), BarH1 ( B-H1), Traf1 and Spn43A ( Higashijima et al., 1992; Daga et al., 1996; Flores et al., 1998; Lai et al., 2000; Preiss et al., 2001). Three of the downregulated genes are required for R8 photoreceptor development: atonal ( ato), bearded ( brd) and big brain ( bib) ( Jarman et al., 1994; Singson et al., 1994; Li and Baker, 2001). R8 specific transcripts were expected to be downregulated because there are fewer R8s in GMR>sSpi (Lesokhin et al., 1999). Another 4 genes downregulated in GMR>sSpi are reported to be present in other differentiating photoreceptors: SoxNeuro ( SoxN/ SoxB1), Fasciclin 2 ( Fas2), klingon ( klg) and tartan ( trn) ( Chang et al., 1993; Butler et al., 1997; Pignoni et al., 1997; Cremazy et al., 2001). Finally, transcripts of 4 downregulated genes are expressed in a pattern associated with the MF: E(Spl)m2, brd, arc ( a) and spineless ( ss) ( Singson et al., 1994; Duncan et al., 1998; Lai et al., 2000; Liu and Lengyel, 2000). As anticipated, the array analysis identified genes already known to be expressed in cells within the SMW or required for the development R8 photoreceptor cells. The remaining 58 genes, whose transcription was uncharacterized or genes uncharacterized with respect to eye development are described later in the results. They include both, SMW and R8-associated transcripts, as well as genes expressed in other tissues. Known transcripts upregulated in GMR>sSpi compared to GMR>RasV12 68 transcripts were upregulated in GMR>sSpi compared to GMR>RasV12. Since GMR>sSpi promotes G1 arrest and interferes with the differentiation of R8 photoreceptors, we expect this group to contain genes required for photoreceptor neurons other than R8, possibly required for G1 arrest, or negative regulators of the SMW. Such genes should be expressed in groups of cells corresponding to the precluster cells just posterior to the MF. Six genes known to be necessary for eye development were upregulated in GMR>sSpi; 5 of these are expressed in differentiating cells in the eye and required for neuronal development: kekkon ( kek), seven-up ( svp), βTub60D and neuroglian ( nrg) and pointed ( pnt) ( Mlodzik et al., 1990; Bellen et al., 1992; Desai et al., 1994; O Neill et al., 1994; Musacchio and Perrimon, 1996; Hoyle et al., 2000). Transcripts for one negative cell cycle regulator, scribble ( scrib) ( Bilder and Perrimon, 2000), were up regulated in GMR>sSpi. As expected, the microarray analysis identified transcripts known to be expressed in differentiating photoreceptor cells. The remaining 62 transcripts enriched in GMR>sSpi were uncharacterized with respect to eye development and are described later in the results. They include not only genes transcribed in photoreceptors, but also genes expressed in other tissues (). Identification of genes’ expression patterns during Drosophila eye development Using RNA in situ hybridization we examined the wild type expression patterns of 126 genes whose transcripts were determined to be significantly different between GMR>sSpi and GMR>RasV12 (). We found 27 genes downregulated and 30 genes upregulated in GMR>sSpi that were expressed in the cells of the DP (). A single gene, CG31676, downregulated in GMR>sSpi, was expressed in the retinal basal glia (). Surprisingly, 12 of the genes downregulated and 17 genes upregulated in GMR>sSpi were expressed in the PE of the eye imaginal disc (). In addition, transcripts of 10 downregulated and 6 upregulated genes in GMR>sSpi were expressed in the larval plasmatocytes present on the basal outer surface of the eye imaginal disc () ( Wolff and Ready, 1993). Some of the genes are expressed in overlapping locations. For example, 3 genes, Pde8, CG8502 and Neu3 were expressed in the DP, PE and plasmatocytes. The Venn diagram in summarizes the expression pattern data and tables 1 and 2 list the novel and known expression patterns of the differentially transcribed genes that are downregulated and upregulated in GMR>sSpi compared to GMR>RasV12 respectively. | Table 1RNA in situ hybridization patterns of genes downregulated in GMR>sSpi vs. GMR>RasV12 |
| Table 2RNA in situ hybridization patterns of genes upregulated in GMR>sSpi vs. GMR>RasV12 |
We have established the expression pattern of 35 uncharacterized genes expressed in the DP, 29 genes expressed in the PE, 15 genes expressed in the plasmatocytes and 1 uncharacterized gene expressed in glia, suggesting that EGFR signaling may regulate gene expression in these different cell types. The RNA in situ hybridization patterns of all genes examined can be found in Supplementary Table 2. This includes expression patterns in some other imaginal discs and larval tissues. 45 of the differentially regulated transcripts were not detected during normal eye development; 9 of these genes were expressed in other imaginal discs or tissues however including 2 ( Lim1 and al) in the antennal disc. The genes for which no transcript was detected in the eye disc are either expressed at levels too low to detect by in situ hybridization, not detected by the probe generated or else ectopically induced by either GMR>RasV12 and GMR>sSpi. Seven genes remain untested due to technical difficulties. Transcripts of uncharacterized genes associated with the SMW and MF Six uncharacterized genes are expressed in a patterned band of cells associated with the SMW. Two of these genes, CG15630 and CG8502 appear to be expressed in the dividing cells, and are down regulated in the absence of the SMW in GMR>sSpi. CG15630 is expressed in groups of 2 or 3 cells and CG8502 in single cells, both at the posterior edge of the MF for 2–3 columns (). The other four genes were upregulated when the SMW was blocked by GMR>sSpi, and are expressed in a band cells associated with the SMW: CG11339, CG31176, CG11382 and Lip1 (). CG11339, CG11382 and Lip1 are expressed in the non-dividing cells whereas CG31176 appears to be expressed in all cells near the SMW. These four genes maybe associated with maintaining the G1 arrest of differentiating cells or early photoreceptor differentiation. The gene CAP, which was downregulated in GMR>sSpi, was expressed in the MF prior to the onset of the SMW. E(Spl)m2, brd, arc and ss are also reported to be expressed within the MF ( Singson et al., 1994; Duncan et al., 1998; Lai et al., 2000; Liu and Lengyel, 2000). We also report the expression of 8 uncharacterized genes in the anterior part of the eye disc ahead of the MF. Four of these, SP1029, Pde8, CG13966 and CG14598 were expressed a band of cells anterior to the MF (). The other four genes, GM130, CG5929, Neu3 and Tsp42E1 were expressed throughout the anterior part of the eye disc (). Transcripts of uncharacterized genes expressed in differentiating cells 8 genes downregulated and 17 genes upregulated in GMR>sSpi are expressed in the differentiating photoreceptors ( Tables 1 and 2). The expression of 7 out of the 8 photoreceptor expressed genes downregulated in GMR>sSpi is known ( Table 1). These genes may be regulated specifically in cells that arrest and differentiate in the first 4–5 columns, before the SMW is complete. Transcripts of the remaining gene, Obp44a, were detected in the axons of late differentiating photoreceptors and as they exit the eye disc via the optic stalk. The expression of 11 out of the 17 genes upregulated in GMR>sSpi is novel and 8 of these genes are uncharacterized: CG14275, CG15522, CG30337, CG30188, CG9487, SP2353, CG9336 and CG32030 (); 2 genes, squeeze ( sqz) and couch potato ( cpo), have previously implicated in the development of the nervous system but not shown to be expressed in the photoreceptors () ( Bellen et al., 1992; McGovern et al., 2003) and 1 gene, Neu3 is known to be expressed in plasmatocytes but it’s expression in the photoreceptors has not been reported before ( Asha et al., 2003) ( Table 2). Transcripts of CG9336 and CG14275 were both detected in the axons of differentiating cells (). A noteworthy expression pattern is that of CG15522 which was upregulated in GMR>sSpi and is expressed in the groups of cells along the posterior edge of the presumptive eye field (). These groups of cells may correspond to a distinct subset of peripheral ommatidia; an expression pattern of this kind has not been previously reported. Consequences of increased Spitz and Ras activity for cells outside of the retina Many of the differentially regulated transcripts were predominantly expressed outside of the retinal epithelium: in the PE, retinal basal glia or larval plasmatocytes. We have therefore investigated the role of Spi and Ras signaling with respect to these different eye imaginal disc cell types. Genes expressed in the peripodial epithelium Transcripts of 29 differentially regulated genes were expressed in PE cells; 12 genes downregulated and 17 genes upregulated in GMR>sSpi ( Tables 1 and 2). A change in the relative sizes of the retina and the PE would lead to a consistent increase or decrease in the relative abundance of PE-expressed genes, so these results instead suggest more specific regulation of particular genes. Consistent with this idea, there was no expansion or reduction of the PE in response to Ras V12 expression in the retina (data not shown). The morphology of the PE in GMR>RasV12 and GMR>sSpi eye discs was also investigated. To examine the size and shape of the PE cells eye discs were labeled with Armadillo (Arm), which localizes to the adheren junctions ( Riggleman et al., 1990). Although the PE of both GMR>RasV12 and GMR>sSpi contained more cells than wild type no discernable difference in cell size or shape between the genotypes was observed (). The difference in genes expressed in the PE between GMR>RasV12 and GMR>sSpi is not due to gross differences in cellular morphology, cell size or PE size. The majority of the PE-expressed genes are uncharacterized or little characterized previously ( Tables 1 and 2). Notably, all the transcripts were also expressed in the PE of other imaginal discs examined, the leg and or wing discs ( Supplementary Fig. 1). At third instar the eye/antennal PE is composed of squamous epithelial cells. The leg and wing PE are made of up both squamous cells and at the margins of the PE, cuboidal cells (margin cells) ( Auerbach, 1936; Cohen, 1993). The difference in squamous and margin cells is reflected in the expression of some of the PE specific transcripts. For instance, transcripts of CG8502 are only present in the squamous cells ( Supplementary Fig. 1, A–B). In contrast, CG3893, CG2657, Osi23 and dScam transcripts were more readily detected in the margin cells ( Supplementary Fig. 1, C–G). The other genes were transcribed throughout the PE ( Supplementary Fig. 1, H–I). The PE-expressed genes can be divided into 2 categories: (I). Genes expressed in PE cells only. This includes most of the PE-expressed genes. Nine genes downregulated in GMR>sSpi: Cyp4e2, Osi23, CG11073, CG3893, beat-Ic, wbl, CG4408, CG2657 and CG15370, and 10 genes upregulated in GMR>sSpi: CG32354, CG18854, sda, Rac2, Sulf1, CG9699, Aplip1, tun, Dscam and CG13890. (II). Genes expressed in the PE and elsewhere in the eye disc and/or plasmatocytes. Three genes downregulated in GMR>sSpi: Pde8, CG8502 and CG13203, and 7 genes upregulated in GMR>sSpi: CG13532, CG13041, CG9487, Neu3, CG9336, eiger, and Lip1 were expressed in this manner ( Tables 1 and 2). For example, CG9487, CG9336 are expressed in the PE and DP whereas CG13203 and CG13041 are expressed in the PE and plasmatocytes. As these genes are also expressed outside of the PE, the microarray may have detected the altered expression levels due to either changes in the PE or elsewhere. Because GMR Gal4 drives UAS transgene transcription in DP cells and not PE cells (data not shown), changes in Group I genes that are only expressed in the PE, indicate EGFR-dependent signaling from the DP to the PE. The simplest hypothesis is that direct targets of EGFR are activated in PE cells by Spi from the DP. However, the results would also be consistent with indirect effects on the PE of EGFR signaling in the DP, if the PE is affected by the SMW, early- differentiating photoreceptors, or cell survival differences that distinguish GMR>RasV12 from GMR>sSpi. CG31676 is expressed in the retinal basal glia CG31676 is downregulated in GMR>sSpi and is expressed in the retinal basal glial (). CG31676 message was also detected along the Bolwig nerve ( Supplementary Table 2). In the differentiating eye disc, glia are present along the basal surface of the eye epithelium associated with photoreceptor axons where they are required for guiding axons into the optic stalk () ( Choi and Benzer, 1994; Rangarajan et al., 1999). In situ hybridization of the CG31676 probe to the original microarray genotypes revealed that the levels of CG31676 mRNA were comparable to wild type in GMR>RasV12 but greatly reduced in GMR>sSpi (). Due to non-autonomous Spi secretion it is likely that CG31676 was down regulated in the glia in response to EGFR/Ras signaling. We hypothesize that activation of EGFR in the glia by Spi blocked the expression of CG31676. To investigate the effects of ectopic Spi on glia, we examined the retinal basal glia in GMR>RasV12 and GMR>sSpi with the glial cell marker, Repo. The glia in GMR>RasV12 retinas were comparable to wild type (). The glia in GMR>sSpi displayed defects in both migration and localization (). Normally the anterior border of glial cell migration is posterior to the MF around column 6, the point at which axons begin to turn posteriorly, known as the axonal boundary. Basal glia in GMR>sSpi retinas migrated anterior to the MF past differentiating photoreceptors (). In addition, glia were also found apical to the photoreceptors and along the Bolwig nerve (). We propose that activation of EGFR in glia, by Spi from the photoreceptor axons, promotes the motility of and alters the localization of retinal basal glia, possibly through CG31676. Genes transcribed in the larval plasmatocytes Ten genes downregulated in GMR>sSpi are expressed in the plasmatocytes present on the basal surface of the wild type retinal eye field: Pde8, twe, CG12508, CG8502, CG13203, CG15911, PGRP-SC2, Smg1, robl62A and CG18547 (). Seven genes enriched in GMR>sSpi were also expressed in the plasmatocytes: CG13521, CG32406, CG14598, CG13041, CG30022, CG33275 and Neu3. Other genes expressed in larval plasmatocytes such as, serpent (s rp), croqumort ( crq) or peroxidasin ( pxn) were not differentially regulated in our microarray analysis ( Nelson et al., 1994; Franc et al., 1996; Rehorn et al., 1996). Although some genes, such as CG8502 and Pde8, are expressed in other eye disc cells, 7 genes are expressed only in the plasmatocytes, implying a difference in the plasmatocytes between GMR>RasV12 and GMR>sSpi. To identify the plasmatocytes on the developing retina we labeled GMR>RasV12 and GMR>sSpi discs with Arm. More plasmatocytes were present in both GMR>RasV12 and GMR>sSpi than in wild type; there was little difference in number between GMR>RasV12 and GMR>sSpi, however (). To investigate further we labeled wild type, GMR>RasV12 and GMR>sSpi developing retinas with the mature plasmatocyte marker MDP-1 ( Hortsch et al., 1998). MDP-1 is present in plasmatocytes on the surface of wild type eye imaginal discs () and plasmatocytes in GMR>RasV12 and GMR>sSpi both expressed MDP-1 suggesting that they mature in both genotypes (). During development plasmatocytes are required for the phagocytosis of apoptotic cells ( Rizki, 1978). EGFR signaling serves as a survival signal, and activation of the EGFR/Ras/MAPK pathway promotes cell survival in the developing retina ( Bergmann et al., 1998; Kurada and White, 1998). To examine cell death in GMR>RasV12 and GMR>sSpi we labeled developing retinas with the CM1 antibody that recognizes activated Drice ( Srinivasan et al., 1998; Yu et al., 2002). As previously reported, no cell death occurred in GMR>RasV12 () ( Yang and Baker, 2003). Some cells in GMR>sSpi eye discs labeled positively for CM1, however; the apoptosis was observed in the posterior regions of the retinal eye field (). Apoptosis in GMR>sSpi is one potential explanation for changes in gene expression in plasmatocytes. Retinal EGFR signaling regulates basal lamina composition Transcripts of the proteoglycan papilin ( ppn) were downregulated in GMR>sSpi ( Table 1). Ppn is transcribed by imaginal disc cells and accumulates in the extracellular matrix where it forms part of the basal lamina ( Campbell et al., 1987; Kramerova et al., 2000) ( Supplementary Table 2). Ppn is recognized by the monoclonal antibody MDP-1 that we used to label plasmatocytes ( Hortsch et al., 1998) (J. Fessler, Pers. Comm.). In wild type retinas, Ppn was detected in the basal lamina under epithelial cells posterior to the MF (). Ppn was not detected in the basal lamina of GMR>sSpi eye imaginal discs (). In GMR>RasV12 eye discs, Ppn was only present in the in the basal lamina next to cells at the posterior edge of the MF (). If the ppn mRNA expression level is a reflection of Ppn in the basal lamina in the microarray analysis, a difference in ppn expression in the DP between GMR>RasV12 and GMR>sSpi may have been detected. Since Ppn in the basal lamina is anterior to ectopic Ras V12 activity in the DP, the increased activation of the EGFR pathway in the retinal epithelium possibly negatively regulates ppn transcription in retinal epithelial cells, thereby affecting the composition of the basal lamina. Microarray predicted changes translate to alterations of transcript levels in vivo To verify our array results, the RNA in situ hybridization pattern of six genes were re-examined in the original microarray genotypes, GMR>RasV12 and GMR>sSpi: CG8502, CG15522, SP1029, CG31676, Smg1 and Dscam. CG8502, CG15522 and SP1029 are all expressed in the DP. CG8502 was downregulated in GMR>sSpi with a signal log ratio (SLR) of −1.7. Transcripts of CG8502 are expressed in 3 columns of cells at the posterior edge of the MF in both wild type and GMR>RasV12 eye discs; in agreement with the array data, this expression was not detected in GMR>sSpi (). The microarray analysis determined a SLR of +4 for CG15522. By in situ hybridization, we also observed an up regulation in the expression level and an expansion of cells expressing CG15522 in GMR>sSpi compared to GMR>RasV12 (). The normal expression of SP1029 in a band of cells anterior to the MF was absent in GMR>sSpi () also consistent with the predicted downregulation of the message in GMR>sSpi (SLR = −3.45). In wild type eye discs, CG31676 is expressed in the retinal basal glia. We could not detect any message by in situ for CG31676 in GMR>sSpi eye discs. The in situ for CG31676 in GMR>RasV12 was comparable to wild type. This is consistent with a SLR of −3.2 for CG31676 in the array analysis (). The microarray analysis predicted changes of Smg1, a plasmatocyte expressed gene, and Dscam a PE expressed gene (SLR of −1.44 and +1.36 respectively) were both consistent with changes observed between GMR>sSpi and GMR>RasV12 by in situ hybridization (). Because the gene expression changes observed by microarray analysis, were confirmed in each of these cases, it is likely that the microarray results provide a generally accurate picture of transcriptional differences between GMR>RasV12 and GMR>sSpi in the DP, PE, glia and plasmatocytes. We have used microarray to identify genes transcribed in the SMW during Drosophila eye development. By manipulating the EGFR pathway, we blocked the SMW, and compared the RNA profile of these developing retinas to retinas with a SMW. Both genotypes activated EGFR signaling posterior to the SMW, controlling for many of the other roles of EGFR. We performed RNA in situ on most of the 140 differentially transcribed genes. Together this analysis has identified of uncharacterized genes that are expressed in SMW and differentiating cells. Although our strategy avoided identifying many thousands of EGFR-dependant genes, many of the genes found were not expressed during the SMW. Instead, these genes are expressed in cells of the PE, larval plasmatocytes and glia. Having investigated the significance of ectopic EGFR signaling on these different cell types, we suggest that : (1) There are targets of EGFR signaling in the PE; (2) Ectopic EGFR signaling leads to the differential regulation of several genes expressed in plasmatocytes; (3) Ectopic EGFR signaling in the glia effects the migration and localization of glia in the developing retina and (4) EGFR signaling also regulates the composition of the the basal lamina. These genes are probably identified because they are targets of Spi secreted from the DP, although some could be regulated indirectly, in response to the SMW or other differences between GMR>sSpi and GMR>RasV12 discs. Thus an unexpected bonus of the approach was to uncover genes that are candidates to mediate responses in other tissues to Spi secretion from the DP. In future it will be interesting to explore these predictions by direct investigation of the non-autonomous role of retinal Spi. Identification of genes transcribed in cells associated with the SMW Our primary goal was to identify genes transcribed in the SMW. By comparing eye discs with and without a SMW we have identified 10 genes that are expressed in cells associated with the SMW; 6 of these genes have novel expression patterns in the SMW. The SMW is a specific patterned cell cycle regulated by the developmental signaling pathways EGFR and N. The transcriptional targets of these pathways in the SMW remain unknown. Cell cycle regulators such as Stg, Cyclin B, or Dacapo (Dap) were not differentially regulated between the sSpi and RasV12 expressing eye discs; this was probably due to the many cycling cells in the anterior part of the eye disc that are unaffected by GMR>sSpi and GMR>RasV12. We expected genes expressed in the SMW cells to be transcribed in a patterned band of cells at the posterior edge of the MF for 2–3 columns of retinal development. The 6 uncharacterized genes are expressed in this manner: CG8502 and CG15630 were down regulated in GMR>sSpi and CG11339, CG11382, CG31167 and Lip were upregulated in GMR>sSpi. The 4 characterized genes that were all downregulated in GMR>sSpi are lz, B-H1, Traf1 and Spn43A their potential cell cycle roles have not been previously examined. Since early cell fate decisions occur simultaneously with the SMW, further work will be required to distinguish whether expression of these genes is associated with cell cycle progression or arrest, or the differentiation of R8 or other early-specified cells. Ten genes were expressed ahead of the MF in the region of the eye disc where the cells are unspecified and asynchronously cycling: Spn43A, SP1029, Pde8, Klg, CG13966, GM130, trn and CG5929 were downregulated in GMR>sSpi, and CG14598 and Tsp42E1 were upregulated in GMR>sSpi. One of these genes, SP1029, a metallopeptidase located 3′ of stg that exhibits a similar mRNA expression pattern to stg ( Alphey et al., 1992). We speculate that SP1029 may come under the regulation of the SMW enhancer of stg; this enhancer not yet been mapped and could be 3′ to stg ( Lehman et al., 1999). Five of the genes downregulated in response to sSpi are expressed within the MF: E(Spl)m2, CAP, Traf1, Brd and ss. The EGFR pathway might have transcriptional targets anterior to and within the MF, although this has not been reported previously. Alternatively these targets may be indirect and depend on secreted signals produced during the SMW. EGFR targets in differentiating cells Even though GMR>sSpi and GMR>RasV12 lead to ectopic photoreceptor specification to similar degrees, 25 genes expressed in the differentiating cells were uncovered. SoxN, lz, Fas2, B-H1, thisbe ( ths), klg, trn and Obp44a were downregulated in GMR>sSpi, and CG14275, CG15522, CG30337, kek1, CG30188, CG9487, CG30022, Neu3, svp, CG9336, sqz, SP2353, CG32030, βTub60D, Nrg, cpo and pnt were upregulated in GMR>sSpi. These genes are potential targets of the EGFR during photoreceptor differentiation. More of the genes expressed in differentiating cells are upregulated in GMR>sSpi than downregulated. Although photoreceptor differentiation occurs only slightly earlier in GMR>sSpi, this maybe sufficient to increase the proportion of transcripts from such cells; alternatively there may also be changes in the type of photoreceptor specified. For example, Spi interferes with R8 photoreceptor development ( Lesokhin et al., 1999; Frankfort and Mardon, 2004). The mRNAs of three genes affected by EGFR signaling, Obp44a, CG9336 and CG14275, were detected in the axons of photoreceptors neurons. The transport of mRNAs to the axon of young neurons has an important role regulating nerve cell maturation ( Mohr and Richter, 2000). It will be interesting to determine whether EGFR regulates the transport and/or localization of mRNAs. EGFR activity in the DP also regulates the composition of extracellular structures preventing the addition of glycoprotein, Ppn, to the adjacent basal lamina. Eight genes expressed in the differentiating cells have putative or known roles either cell adhesion or cytoskeletal organization. EGFR signaling alters the adhesive properties of cells in the eye disc so that normal cell shape re-arrangements occur ( Brown et al., 2006). These genes maybe downstream of EGFR in maintaining proper cell adhesion during the G1 arrest of differentiating cells. Possible targets of the EGFR/Ras pathway in the peripodial epithelium In this study we have identified 29 genes that were not previously known to be expressed in the PE. 17 are uncharacterized genes that are expressed exclusively in some or all cells of the PE and not elsewhere in the imaginal disc. Until now Ultrabithorax ( Ubx) was the only gene expressed in all cells of the wing PE but not the DP, although Ubx is expressed in the DP other imaginal discs ( Brower, 1987). Coronin-Gal4 is also detected in the wing PE only but in a subset of cells ( Pallavi and Shashidhara, 2003). Thus, we uncovered an unexpected pool of PE-specific genes, suggesting that gene expression in the PE is regulated by Spi secreted from the DP, either directly or in response to the SMW or R8 differentiation. It is well established that signals from the PE affect the development of the DP ( Cho et al., 2000; Gibson and Schubiger, 2000; Pallavi and Shashidhara, 2003; Pallavi and Shashidhara, 2005). Our data suggests that the reverse may also be true; verticals signal from the DP to the PE have an important role in regulating the development and interactions between two disc epithelia. As PE cells do not change in size and morphology between the two genotypes it will be interesting to discover the nature of the response. Spitz affects glial cell migration and localization CG31676 was the only gene found to be altered in the retinal basal glia. We speculate that activation of EGFR in the glia, by Spi secreted from the axons, negatively regulates the expression of CG31676. Since only one glia gene was affected, this is unlikely to reflect a change in the relative number of glial cells after Spi expression. We found that migration and localization of glia was affected in GMR>sSpi. Glia normally migrate up to the axonal boundary, the point at which the axons of differentiating photoreceptors turn posteriorly to the optic stalk ( Choi and Benzer, 1994). Glia in GMR>sSpi migrated beyond the axonal boundary and photoreceptor differentiation. In addition, glia were present apical to photoreceptor and along the Bolwig nerve. The actual molecular mechanism of glia migration is unknown ( Rangarajan et al., 1999). We speculate that CG31676 maybe directly involved in glia migration and/or localization in response to sSpi from photoreceptor cells, as it was the only gene whose transcription was affected. Spitz affects larval plasmatocytes Ectopic Spi in the retinal epithelium affected gene expression in plasmatocytes that are associated with the eye imaginal disc. 7 of 17 differentially transcribed genes were not expressed elsewhere in the eye imaginal discs. Most of the plasmatocyte expressed genes uncharacterized. However, Neu3 and PGRP- SC2, have previously been demonstrated by microarray to be expressed in plasmatocytes ( De Gregorio et al., 2001; Asha et al., 2003). We propose that Spi from the DP must affect the plasmatocytes directly or indirectly. Interestingly, we detected no difference in number or differentiation (as assessed by the MDP-1 antigen). EGFR activity coordinates development of the eye disc with other associated tissues In summary, combining microarray and RNA in situ hybridization has led to the identification of genes that are transcribed in cells associated with the SMW, but we also uncovered as many targets of EGFR in the other cell types associated with the eye imaginal disc. Because our microarray was designed to identify a narrow subset of targets, and appears to have excluded the majority of genes directly or indirectly regulated by EGFR in the eye disc as a whole, we think that most of these genes are different because they are targets of Spi in cells where GMRGal4 does not express. Thus, we inadvertently selected for genes in other tissues that respond to Spi made in the DP. These findings suggest that, in addition to regulating multiple aspects of retinal differentiation, and regulating brain differentiation in response to retinal innervation ( Huang et al., 1998), changes in EGFR activity during eye disc differentiation could also serve to coordinate the developmental programs of the glia, PE, and plasmatocytes with the eye disc proper together comprising an organ system of cells from multiple origins (). We thank C. Hindnavis for technical assistance; M. Wilkin and D. Dimova for helpful technical advice; M. Hortsch for the MDP-1 antibody and U. Banerjee for helpful advice. The manuscript was improved by suggestions from W. Li and E. Bach. Confocal microscopy was performed at the Analytical Imaging Facility and array processing was performed at the Affymetrix facility at the Albert Einstein College of Medicine. This work was supported by grant GM47892 from the National Institutes of Health. NEB is a Scholar of the Irma T. Hirschl Foundation for Biomedical Sciences.
|
The publisher's final edited version of this article is available at 
