The discovery of direct downstream targets of transcription factors (TFs) is necessary for understanding the genetic mechanisms underlying complex, highly regulated processes such as development. In this report, a combinatorial strategy was used to conduct a genome-wide search for novel direct targets of Eyeless (Ey), a key transcription factor controlling early eye development in Drosophila. To overcome the lack of high-quality consensus binding site sequences, phylogenetic shadowing of known Ey binding sites in sine oculis (so) was used to construct a position weight matrix (PWM) of the Ey protein. This PWM was then used for in silico prediction of potential binding sites in the Drosophila melanogaster genome. To reduce the false positive rate, conservation of these potential binding sites was assessed by comparing the genomic sequences from seven Drosophila species. In parallel, microarray analysis of wild-type versus ectopic ey-expressing tissue, followed by microarray-based epistasis experiments in an atonal (ato) mutant background, identified 188 genes induced by ey. Intersection of in silico predicted conserved Ey binding sites with the candidate gene list produced through expression profiling yielded a list of 20 putative ey-induced, eye-enriched, ato-independent, direct targets of Ey. The accuracy of this list of genes was confirmed using both in vitro and in vivo methods. Initial analysis reveals three genes, eyes absent, shifted, and Optix, as novel direct targets of Ey. These results suggest that the integrated strategy of computational biology, genomics, and genetics is a powerful approach to identify direct downstream targets for any transcription factor genome-wide (Ostrin, 2006).
To further understand the basis of the shf phenotype, the expression and distribution of Hh itself was examined. Hh transcription is not obviously affected by shf, as assessed using the hh-lacZ enhancer trap. However, the accumulation of Hh protein is strongly disrupted. In a wt wing imaginal disc, Hh can be detected at the cell surface of the posterior secreting cells, along the entire apical/basal axis without any preferential localization. In neighboring anterior cells, Hh is found in vesicles where it colocalizes with its receptor, Ptc. Confocal sections at different levels along the apical/basal axis show that the apical accumulation of Hh is not significantly different in shf and wt cells. However, basolateral anti-Hh staining is strongly reduced in the posterior compartment of shf2 and shfx33 wing discs. Z sections of these discs confirm that the basolateral accumulation of Hh is strongly defective in shf mutants. The same results are obtained with all three alleles tested and using two different anti-Hh antisera. Similarly, the levels of Hh in the anterior compartment are reduced. Some punctate anti-Hh staining was still detected in anterior cells in the shf2 and shfx33 mutant backgrounds, and as in wt, much of this staining colocalizes with anti-Ptc staining. However, vesicles with colocalized staining do not extend as far anterior in shf as in wt discs, suggesting that the range of Hh movement into the anterior compartment is reduced in shf mutants (Glise, 2005).
One possibility is that this apparent reduced movement of Hh is the consequence of the reduced accumulation of Hh in the posterior compartment. To test this, a GFP-tagged version of Hh was expressed in the dorsal compartment using ap-Gal4 and the range of Hh-GFP movement into the ventral compartment was found to be still reduced in shfx33 discs, despite the higher than normal levels of Hh present in the dorsal compartment. This reduction is observed even in the posterior of the wing and is thus not a byproduct of altered interactions with Ptc. Faint residual Hh-GFP was occasionally observed in ventral cells, suggesting that some Hh movement occurs in the absence of Shf; this is consistent with what is observed with endogenous Hh and with the mutant phenotype. However, such movement is severely reduced (Glise, 2005).
The phenotype of shf is quite similar to that observed after the loss of HSPGs in the wing; both Shf and the HSPGs are required for the accumulation of Hh in the posterior compartment and its movement into the anterior compartment. Much of the effects of heparan sulfate on Hh distribution and signaling are mediated by the Drosophila glypicans Dally and Dally-like (Dlp). However, the results suggest that the shf phenotype is not caused by changes in the levels of the Dally or Dlp, as assessed by Dlp antibody staining and a dally-lacZ reporter construct (Glise, 2005).
Hh is made as a precursor molecule, consisting of a C-terminal protease domain and an N-terminal signaling domain. The C-terminal domain is removed by autocatalytic cleavage, and during this cleavage process, a cholesterol molecule is covalently attached to the C terminus of the signaling domain. Some evidence suggests that Hh that lacks cholesterol is less sensitive to the loss of HSPGs. Constructs that contain only the N terminus of Hh (HhN) generate a protein that is not processed and thus lacks the cholesterol. The effect of this on the movement of Hh and its vertebrate homologs apparently depends on the developmental context and the presence of wt Hh. In wt wings, HhN signals over a longer range than cholesterol-modified Hh. While the range of Hh signaling is reduced in a shf background, this is not true of HhN. ap-Gal4 was used to drive expression of UAS-hh throughout the dorsal compartment. In both wt and shf2 discs, this results in the ectopic expression of the Hh target col-kn in the entire dorsal anterior quadrant of the wing pouch, consistent with the competence of anterior cells to interpret the Hh signal. In wt discs, this also induces ectopic col-kn expression in several rows of ventral anterior cells adjacent to the dorsoventral compartment boundary. In shf2/Y discs, the ectopic ventral expression of col-kn is drastically reduced. However, when the same experiments are repeated using ap-gal4 and UAS-hhN, the long-range activation of col-kn expression observed in the wt background is not decreased in a shf2 background. Together, these experiments show that shf is required for normal range of movement of cholesterol-modified Hh in the wing disc (Glise, 2005).
The Dally and Dlp glypicans are bound to the cell surface by a GPI linkage, and while this linkage can be cleaved, the effects of loss of HSPGs from clones in the wing are nearly cell autonomous. Hh accumulation is affected throughout posterior HSPG clones, and Hh can signal only over a very short distance into anterior clones. However, the same is not true of Shf. First, shf2 and shf919 mitotic clones were induced in wing discs and the extent of Hh signaling was visualized using col-kn expression. A shf phenotype was never observed, even when the mutant clones encompassed large portions of either the anterior or posterior compartments and abutted the A/P compartment boundary along most of its length. Thus, shf activity in the wing can be provided at a distance by wt cells, consistent with the prediction that Shf is a secreted protein. Supporting this conclusion, the adult wing phenotype of shf2 and shfx33 flies is rescued by UAS-shf, even when expression of the Gal4 driver is limited to posterior (en-Gal4), anterior (dpp-Gal4 or ptc-Gal4), or dorsal (ap-Gal4) cells (Glise, 2005).
To confirm this, a UAS-shf was generated containing a C-terminal GFP tag; this construct can rescue shf using a variety of Gal4 drivers. When UAS-shf-GFP was expressed in dorsal cells using ap-Gal4, GFP was found not only dorsally, but also throughout the ventral wing pouch. Thus, Shf is secreted and can move across the length of the prospective wing blade (Glise, 2005).
The distribution of Shf-GFP also provided some clues about potential interactions between Shf and other extracellular factors. Although ap-Gal4 drives uniform expression throughout the dorsal compartment, the distribution of Shf-GFP driven by ap-Gal4 was not uniform: in both dorsal and ventral cells, it shows a slight posterior emphasis and decreased levels anterior to the A/P boundary. This suggests that Shf is accumulating in response to some factor expressed only in the posterior compartment. Among known extracellular components of Hh signaling, only Hh itself accumulates at higher levels specifically in the posterior compartment (Glise, 2005).
To see whether the endogenous Shf protein shows a similar distribution, a polyclonal antiserum against Shf was generated. Wing discs show a pattern of staining similar to that observed using in situ hybridization, with stronger expression in the far anterior and posterior regions of the wing pouch, as well as regions of the hinge and notum. However, unlike the in situ stains, cell surface anti-Shf staining is higher in the posterior compartment and shows a sharp decrease congruent with the A/P boundary. The anti-Shf staining is lost from shfx33 discs, confirming the specificity of the antiserum and indicating that this allele is likely a null (Glise, 2005).
The wt Shf distribution appears to be caused by a difference in the accumulation of Shf protein, rather than in shf expression. The shf2 and shf919 alleles are caused by point mutations in the third EGF repeat and show normal patterns of transcription. However, anti-Shf staining in these mutants was abnormal; the sharp demarcation between anterior and posterior anti-Shf staining was lost, and posterior cell surface staining was less distinct. The remaining staining pattern in these mutants closely resembles that of message transcription, with stronger anterior and posterior expression distant from the compartment boundary. This result may indicate that a normal third EGF repeat is specifically required for interaction with a posteriorly expressed protein (Glise, 2005).
To test whether Hh itself plays a role in Shf accumulation, anti-Shf staining was examined in discs homozygous for the temperature-sensitive hhts2 allele. This allele is caused by an amino acid change in the N-terminal region of Hh, which likely renders the protein sensitive to misfolding at the restrictive temperature. When upshifted from 18°C to 30°C for 24 hr, hhts2 discs showed an abnormal pattern of anti-Shf staining, similar to that observed in shf919; the stronger posterior staining was lost, and the sharp demarcation between the anterior and posterior compartments was no longer apparent. Thus, the posterior accumulation of anti-Shf staining depends on the presence of normal Hh. Consistent with these results, an independent study on Shf by Gorfinkiel (2005) reports coimmunoprecipitation of Shf and Hh from Shf-misexpressing salivary glands. As yet, this result has not been reproduced using tagged forms of Shf and Hh expressed in the Drosophila S2 cell line. While such negative results should be treated with caution, they suggest either that the association between Shf and Hh is indirect or that S2 cells lack some cofactor or processing necessary for the interaction (Glise, 2005).
Cells in the hhts2 disc still retain anti-Shf staining at the cell surface, and this suggests that Shf might be binding to some additional cell surface component. Therefore whether Shf accumulation responds to changes in HSPGs was tested. Cells lacking the EXT copolymerases Tout velu (Ttv) and Sister of tout velu (Sotv) are deficient in HS synthesis. ttv sotv mutant clones were generated and a marked, cell-autonomous decrease in the levels of anti-Shf staining was seen. A similar but slightly weaker effect was observed after loss of the Drosophila glypicans Dally and Dally-like. The effect was most striking in the posterior compartment and thus might be caused in part by the loss of Hh from such clones. However, weak decreases were also observed in anterior clones distant from the compartment boundary, indicating that the decrease cannot be entirely accounted for by the loss of Hh. This indicates that strong Shf accumulation requires the HSPGs (Glise, 2005).
Hsieh (1999) showed that a construct lacking the EGF repeats of hWIF-1 was nearly as effective as full-length protein in binding Wg and Xenopus Wnt8 and inhibiting their ability to signal. However, the localization of the shf2 and shf919 point mutations to the third EGF repeat of Shf suggests that the EGF domains play a critical role in regulating Hh signaling. Moreover, the second and third EGF repeats of Shf are by BLAST analysis more similar to the EGF repeats of the Hedgehog binding protein HIP than to most other types of EGF domains, suggesting that they help mediate interactions with Hh or some other component of the Hh signaling complex (Glise, 2005).
Therefore constructs lacking either the WIF domain (UAS-shfΔWIF) or the EGF domains (UAS-shfΔEGF) were generated and expressed in flies. Driving expression using ap-GAL4 does not rescue the shf phenotype. This failure is not caused by a decrease in protein stability; both truncated proteins are recognized by polyclonal anti-Shf antiserum, and the level of staining in these discs is well above that of the endogenous protein. Thus, unlike hWIF-1, the EGF domains of Shf are necessary for its function (Glise, 2005).
The shf phenotype could not be rescued by expressing zebrafish WIF-1 in the place of Shf. Misexpression with ap-Gal4, ptc-Gal4, or en-Gal4 also does not induce obvious Wg loss-of-function phenotypes. It should be noted that human WIF-1, while unable to rescue shf, can induce Wg loss-of-function phenotypes (Gorfinkiel, 2005). While these negative results should be treated with caution, they suggest that vertebrate WIF-1 proteins may differ in their functions (Glise, 2005).
Therefore shf is required for the accumulation and movement of cholesterol-modified Hh in the developing wing imaginal disc. shf encodes the only Drosophila member of the WIF-1 family of proteins. The only known role of human WIF-1 is to bind to Wnts and inhibit Wnt signaling. Misexpression of WIF-1 blocks Wnt signaling, and inhibiting WIF-1 function and adding Wnt4 have opposite effects in the mouse retina on rod photoreceptor proliferation (Hsieh, 1999; Hunter, 2004). However, the results indicate that Shf does not have a similar function (Glise, 2005).
Rather, the results strongly suggest that secreted Shf affects Hh signaling by interacting with Hh and the HSPGs. This is supported first by the similarity of their effects of Shf and the HSPGs on Hh accumulation and movement. Moreover, Shf accumulation in the posterior compartment of the wing disc depends in part on the presence of normal Hh; this is consistent with the binding observed by Gorfinkiel (2005). Shf accumulation is also strongly reduced by blocking HSPG synthesis or expression. The simplest model that explains these findings is that Shf binds to both HSPGs and a Hh-containing complex and is required to stabilize that interaction, allowing Hh to move over a long distance (Glise, 2005).
However, whereas HSPGs are required for Hh signaling during embryogenesis and in a number of adult tissues, defects in shf null allele are detected only in the wing blade, notum, and eye. The interaction between Hh and HSPGs cannot be totally dependent on Shf. The sensitivity of specific tissues to the loss of shf may reflect the different requirements for long-range Hh movement in these tissues. In the wing imaginal disc, Hh signals over 12 rows of anterior cells, as shown by the induction of its most sensitive known target, dpp. In contrast, in the embryonic ectoderm Hh normally signals over a maximum of three to four rows of cells (Glise, 2005).
That a WIF-1 family member is required for Hh signaling is both unexpected and exciting, as it suggests that WIF-1s have the potential to act as multimodal modifiers of signaling. Vertebrate WIF-1s are expressed in a variety of developing tissues, including nonsegmented paraxial mesoderm but not somites, in notochord adjacent to mature somites, in developing head structures, and in the developing retina (Hsieh, 1999; Hunter, 2004). Moreover, human WIF-1 expression is downregulated in a high percentage of prostate, breast, and non-small lung cell tumors (Mazieres, 2004; Wissmann, 2003) and upregulated in intestinal adenomas and colon carcinomas (Cebrat, 2004). It will be interesting to see if the functions of vertebrate WIF-1s are limited to Wnt signaling or whether these molecules play a wider role (Glise, 2005).
To confirm that Shf/DmWIF is a secreted protein, the chimeric protein Yellow Fluorescent Protein (YFP)-Shf was generated. YFP was fused in-frame downstream from the signal peptide and upstream from the WIF domain. This chimeric protein was able to rescue the shf phenotype as the wild-type protein. The distribution of YFP-Shf was analyzed when overexpressed using the ap-GAL4 line. The presence of YFP-Shf in the non-expressing cells, shows that Shf/DmWIF is secreted and has a long-range diffusion. In these cells, the YFP-Shf staining pattern seemed to be extra-cellular. By incubation with an anti-GFP antibody before cell fixation (Torroja, 2004), it was possible to confirm the extracellular localization of Shf/DmWIF. Using an anti-Hh antibody, the extracellular YFP-Shf was observed colocalized with Hh. In these discs, it was also noted that the extracellular levels of YFP-Shf were higher in the P compartment than in the A compartment, suggesting that Hh might stabilize Shf. It was also found that Hh interacts with Shf/DmWIF by immunoprecipitation assays in salivary gland cells expressing YFP-Shf and using anti-GFP antibody. Therefore, all these results suggest that Hh and Shf/DmWIF interact probably in the extracellular matrix (Gorfinkiel, 2005).
The decreased Hh signaling and reduced Hh protein levels at the plasma membrane in shf mutants, and the colocalization of YFP-Shf with Hh at the extracellular matrix are reminiscent of HSPG function in Hh signaling. It has been argued that HSPGs are required for the stable retention of Hh on the cell surface and for efficient diffusion through the A and P compartments. Furthermore, when synthesis of HSPG is impeded in ttv and sotv clones in the P compartment, Hh levels are reduced in the P compartment, and in mutant ttv and sotv clones in the A compartment, Hh signals only to the first row of cells touching the A/P compartment border. These two phenomena also occur in shf. The question of whether Shf and HSPG might interact was posed. To test this hypothesis, ttv clones were generated in a shf2 or shf EY03173 background. If the shf and ttv phenotypes were additive, one would expect to observe a considerable reduction in Hh levels in a shf background. Although a fall in Hh levels was seen when ttv− clones were induced in a wild-type background, no further reduction was observed when ttv− clones were induced either in shf2 or shfEY03173 background. This result suggests that shf is required for the stabilization of Hh by the HSPGs (Gorfinkiel, 2005).
A common feature shared by ttv and shf is their requirement for the diffusion of lipid-modified Hh. Thus, ttv is not required for the diffusion of non-cholesterol-modified Hh and shf does not affect the diffusion of either cholesterol-free or palmitic acid-free Hh. The explanation for this specific control has to lie in the mechanism of Hh transport. A close analysis of this process shows that lipid modified Hh is anchored to the membrane, and only in this state is its diffusion restricted. However, unmodified Hh is not anchored to the plasma membrane or to the extracellular matrix. Accordingly, spreading of this Hh is unlimited and a proper gradient is not formed. This might therefore suggest that HSPGs and Shf mediate the anchoring or stabilization of lipid modified Hh to the extracellular matrix and, in this way, control its movement (Gorfinkiel, 2005).
The role of HSPG in morphogen stabilization and transport is not exclusive to Hh. Mutations in ttv, sotv, and botv also affect Dpp and Wg signaling and reduce morphogen levels, indicating that HSPG-dependent diffusion is a common mechanism for forming gradients of the three morphogens. Other enzymes involved in HSPG biosynthesis such as those coded by sugarless and sulfateless also affect both Wg and Hh signaling. In addition, dally, which encodes an HSPG protein core, is required for Wg and Dpp activities. Dally-like (Dlp), the other Drosophila glypican, participates in Wg signaling and is also required for Hh signal transduction both in cultured Drosophila cells and Drosophila embryos. Dlp embryos show an altered Hh distribution, and Dlp-GFP colocalizes with Hh. Both observations are consistent with a role for Dlp in Hh stabilization and movement. Next whether Dlp could in someway interact with shf was explored, and it was found that Dlp colocalizes with YFP-Shf. Based on genetic and cytological data, it is suggested that Shf might interact with glypicans to regulate Hh movement by stabilizing Hh protein (Gorfinkiel, 2005).
It is not known how the specificity of HSPG for different ligands during development is achieved. Recently, it has been reported that Notum/Wingful is a secreted enzyme that acts on Dlp to determine whether Dlp reduces or promotes Wg signaling. In the case of Hh signaling, Shf could render HSPG specific for Hh (Gorfinkiel, 2005).
Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.