Hyperplastic Discs Differentially Regulates the Transcriptional Outputs of Hedgehog Signaling

Abstract

Hedgehog (Hh) acts as a morphogen to activate the transcription of diverse target genes via its downstream effector Cubitus interruptus (Ci). Currently, it is less understood how Ci recruits cofactors to activate transcription. Here we report that hyperplastic discs (hyd), an E3 ubiquitin ligase, can differentially regulate the transcriptional outputs of Hh signaling. We show that loss of Hyd activity caused upregulation of some, but not all of Hh target genes. Importantly, Hyd does not affect the stability of Ci. Our data suggest that Hyd differentially restrains the transcriptional activity of Ci via selective association with respective promoters.

INTRODUCTION

Morphogens are secreted signaling molecules which form a concentration gradient and induce distinct cellular responses in a concentration-dependent manner. The roles of morphogen molecules in coordinating growth and patterning are reflected by the time- and tissue- specific control of the transcription of a variety of target genes. The Hh family of proteins are well-studied morphogen molecules which serve as global organizers to control tissue growth and patterning during metazoan development [1, 2]. The role of Hh as a morphogen in development is particularly well illustrated during wing disc development [3–5]. The larval wing imaginal disc, which finally gives rise to the adult wing blade after metamorphosis, is subdivided into distinctive anterior (A) and posterior (P) compartments. Hh expressed in the P compartment moves anteriorly to transduce its signal in a narrow stripe in the A compartment abutting the A/P boundary [6]. Hh signaling induces the localized expression of a number of target genes including decapentaplegic (dpp), patched (ptc), collier (col, also known as knot) and engrailed (en) in a concentration-dependant manner [5, 7–9].

Hedgehog signaling is subject to regulation at multiple levels [10, 11]. These levels converge to control the stability and activity of the DNA binding protein, Ci [12, 13]. Full-length Ci (Ci-155 or CiA) is the primary transcriptional mediator in response to Hh signaling. In the absence of Hh ligand, Ci is proteolytically processed into a shorter form (Ci-75 or CiR) that acts as a transcriptional repressor of target genes or is degraded by the proteasomes. Exposure of cells to Hh blocks the production of Ci-75 and stimulates the nuclear translocation and activation of Ci-155. Ci processing requires a complex for priming phosphorylation [14, 15], which is composed of the atypical kinesin protein Costal 2 (Cos2), the serine threonine kinase Fused (Fu), PKA, and the PEST domain protein Suppressor of fused (Sufu) [16]. Ci phosphorylation results in the recruitment of the F-box protein Slimb/β-TRCP of the SCF ubiquitin ligase complex, which targets Ci for processing or proteasome-dependent degradation [13, 17, 18].

It is well established that Ci is the sole mediator in activating target transcription, but the detailed mechanisms underlying co-factor recruitment and differential target gene transcription remain elusive. It has been shown that the transcriptional co-activator CBP binds to Ci-155 but not Ci-75 and CBP is required for normal Hh signaling [19]. In mammalian systems, Su(fu) can enter the nucleus accompanied by Ci to help inhibit target gene expression by recruiting SAP18, a component of the SMRT/Sin3 histone deacetylase (HDAC) co-repressor complex [20]. In Drosophila, genetic impairment of hyrax (hyx) decreased the expression of the high-threshold Hh target gene col/knot, but not other targets, suggesting that differential regulation of Ci-dependent transcription is subject to more complicated and elaborate controls [21].

The ubiquitin ligase gene hyperplastic discs (hyd, the Drosophila homolog of UBR5, CG9484) has been reported to negatively regulate Hh signaling in eye and wing development, possibly through Ci ubiquitination and degradation [22]. In this study we further investigated the roles of Hyd in Hh signaling using the Drosophila wing disc as the model system. We showed that loss of hyd function results in upregulated transcription of ptc and dpp without noticeable effect on en or col transcription. We demonstrated that the ubiquitin ligase activity is strictly required for Hyd’s function. Importantly, we found that neither over-expression nor knockout of hyd alters Ci stability. We propose a model in which Hyd, through ubiquitination, participates in Ci-dependent transactivation of Hh signaling at selective promoters.

RESULTS AND DISCUSSION

In a genetic screen for ubiquitin ligases involved in Hh signaling, we uncovered the role of Hyd in regulating the transcription of a subset of Hh target genes (Fig 1). Four Hh targets have been examined, including ptc, dpp, en and col, either by anti-β-Galactosidase immunostaining for the reporters (ptc-lacZ [7] and dpp-lacZ [23] ) or by direct antibody staining against the gene product (En, Col and Ptc). As revealed by Hyd antibody staining, Hyd expression is not detectable in wing disc cells either homozygous for the null allele hyd¹⁵, or expressing hyd RNAi (Fig 1A″,B″,C″, Fig 3A″ and Fig S1B′. Note that flies were from the same crosses in Fig 1D–D″ and Fig S1B–B″). While loss of hyd activities strongly enhanced the transcription of ptc and dpp in a cell-autonomous manner (Fig 1A,B and Fig 3A), the levels of Col and En remained unchanged (Fig 1C,D and Fig S4B). It is worth mentioning that the upregulated dpp and ptc transcription only occurs in the anterior cells in a manner that follows the Hh gradient, indicating that Hyd loss modifies the signaling strength in Hh-responsive cells, rather than inducing ectopic Hh activation in cells normally not responding to Hh signaling. Meanwhile, over-expression of hyd has no effect on the transcription of any of the four target genes (Fig S2). Next we tested whether Hyd is generally involved in signal transduction. As shown in Fig S1 and S5, Dpp signaling does not respond to any manipulation of Hyd activity, indicating that Hyd is not a general transcription factor in the wing disc.

Fig 1. hyd selectively controls the transcription of Hh target genes.

All the wing discs presented in this and other figures were derived from third instar larvae with anterior oriented to the left and dorsal up. (A–A‴, C–C‴) wing discs with clones of cells (CD8-GFP positive) over-expressing hyd RNAi under the control of actin-Gal4. (B–B‴) a wing disc with clones of cells (CD8-GFP negative) carrying homozygous hyd¹⁵ allele. (D–D″) a wing disc over-expressing hyd RNAi under the control of ap-Gal4 in the dorsal compartment with the lines marking the dorsal-ventral boundary. The efficacy of RNAi-mediated knockdown in D–D″ is confirmed in Fig S1. With lost or reduced hyd activity, expression of ptc-lacZ (ptcZ) and dpp-lacZ (dppZ) is upregulated, while no notable change is observed for the expression of Col, En and Hh. The gradient of Hh and En also remains normal upon RNAi-mediated knockdown of hyd. Scale bar, 50μm.

Fig. 3. hyd controls Hh signaling in an E3 activity-dependent manner.

(A–C‴) Genetic rescue experiments using MARCM system. Over-expression of V5hyd (B–B‴) but not the E3-dead form V5hydCS (C–C‴) can block the upregulated expression of Ptc in hyd¹⁵ clones (CD8-GFP positive). (D–E‴) wing discs with clones of cells (CD8-GFP positive) over-expressing actin-Gal4-driven hyd RNAi were dissected and cultured in M3 medium containing DMSO (D–D‴) or MG132 (100 μM) for up to 6 hours before immunostaining. MG132 treatment caused a uniform anterior accumulation of Ci (E″) while it does not block the upregulation of ptc-lacZ in hyd RNAi clones (CD8-GFP positive). Scale bar, 50μm.

To determine the possible role of Hyd in Hh ligand production and distribution, we stained Hh protein in wing discs expressing hyd RNAi in the dorsal compartment. No change was observed on protein levels or gradient formation of the Hh ligand when Hyd activity is knocked down (Fig 1D′). Taken together with the nuclear localization revealed by antibody staining (Fig S1), these data argue that hyd is selectively involved in the regulation of specific Hh target genes by a nuclear process downstream of ligand reception.

Previously, Lee J. et al reported that Ci levels were upregulated in hyd mutant clones, and thus they proposed a model that hyd acts by targeting Ci for degradation [22]. Our studies provided several lines of evidence against that Ci is the direct target of hyd. First, the alteration of Ci levels cannot explain the differential regulation of hh targets by hyd. Second, while we did observe similar changes in Ci distribution upon hyd loss in some cases, these changes only occurred in cells with disrupted morphology, and should therefore be considered as a secondary effect. In cells which are morphologically normal, as shown in Fig 2A–A‴, 2B–B‴ and S4A–A′, Ci levels are not subject to any alteration by Hyd impairment or disruption. Furthermore, our epistasis analysis showed that loss of Ci function eliminated all transcriptional output of Hh signaling even in cells with reduced Hyd activity (Fig 2C–E‴), suggesting that Hyd acts upstream of, or in parallel to, Ci in Hh signaling.

Fig. 2. hyd controls Ci-dependent transcription without affecting Ci stability.

(A–A‴) In hyd¹⁵ clones (GFP negative), Ci levels remain unchanged. (B–B‴) In V5Hyd over-expressing clones (CD8-GFP positive), Ci levels are unaltered. (C–E‴) CD8-GFP marks cells co-expressing ci RNAi and hyd RNAi. When both Ci and Hyd are knocked down, the expression of Col, ptc-lacZ and dpp-lacZ is all eliminated. Scale bar, 50μm.

Hyd and its homologs belong to the HECT family of E3 ubiquitin ligases. Several substrates have been identified for the HECT domain activity of Hyd’s mammalian homologs [24–29]. To characterize the role of the HECT domain of Hyd in Hh signaling, we generated a HECT domain mutated form of Hyd by converting the catalytic cysteine at aa2854 to serine (HydCS) [30]. When expressed in cells deficient in endogenous hyd, wild-type Hyd can completely rescue the upregulation of ptc transcription from loss-function of hyd (Fig 3B–B‴) while C2854S mutation renders it incapable of rescuing elevated ptc levels, arguing that the E3 ligase activity of Hyd is required for its role in Hh signaling. Interestingly, over-expression of hydCS doesn’t affect Ci accumulation or hh signaling activity as shown in Fig S3, further arguing against that Ci is the direct target for the E3 ligase activity of Hyd.

HECT domain-mediated ubiquitination of proteins has been reported to regulate numerous biological processes through affecting the trafficking, function, and stability of respective substrates [31]. We further examined whether ubiquitination-mediated degradation is involved in the regulation of Hh signaling by Hyd. In this assay, cultured wing discs were treated either with DMSO control or with MG132 in DMSO, which can effectively block the proteasome-dependent degradation of ubiquitinated proteins. Indeed, treatment with MG132 resulted in dramatic accumulation of Ci uniformly in the anterior compartment (Fig 3, compare E″ to D″) leading to a generally enhanced and expanded ptc transcription domain (Fig 3, compare E to D). However, loss of hyd function caused a further increase in both the level and the range of ptc transcription in the already elevated background (Fig 3E), again arguing against the previous hypothesis that Hyd acts on Ci degradation. In addition, we failed to detect any change in Ci ubiquitination in cultured cells when Hyd or HydCS is expressed (Fig S6), suggesting that Ci is not the direct target for the E3 ligase activity of Hyd.

To further investigate how Hyd regulates the transcriptional activity of Ci, we examined the physical interaction between these two proteins. Intriguingly, both the full-length Ci and CiM (aa 440-1160) were found to interact with Flag-tagged Hyd in cultured S2 cells while CiN (aa1-450) and CiC (aa1161-1377a) both of which lack the DNA-binding zinc finger domain failed to do so (Fig 4A and B). This result provided the molecular basis for the specific role of Hyd in restraining the signaling strength under Hh activation. Next we asked how Hyd functions in a target gene-restricted manner. One simple explanation would be that Hyd is selectively associated with different target genes. To test the hypothesis, we checked whether hyd is selectively enriched for the seven target genes of Hh and Dpp signaling with ChIP coupled real time PCR, and we also incorporate the house-keeping gene, rp49 as a control. In support of the hypothesis, the ChIP-PCR results with cultured cells indicated that hyd is only enriched around the transcription start site (TSS) of ptc and dpp promoters at a ratio of 0.7% and 0.5% respectively in both S2 and Clone 8 cells, while the enrichment for en and col promoters is less than 0.2% (Fig 4 C and D, Fig S7). Since S2 cells do not express Ci while Clone 8 cells have endogenous Ci expression, these results indicate that Hyd acts as a pre-existing factor at selective promoters whose binding is independent from the presence of Ci. Moreover, in agreement with no participation in Dpp signaling, Hyd was not enriched at any of the dpp-responsive promoters examined here (Fig S7).

Fig. 4. hyd selectively associates with promoters of hh targets.

(A) Co-immunoprecipitation (IP) assays in clone 8 cells. IB: immunoblot. Hyd is found in a complex with Ci but not CiN.

(B) Co-immunoprecipitation (IP) assays in clone 8 cells. IB: immunoblot. Hyd is found in a complex with CiM but not CiN and CiC. * indicated the expressed bands of Myc CiM and CiN.

(C) ChIP-PCR assays in cultured cells. Hyd selectively associates with the transcription start sites (TSSs) of Hh-responsive promoters in S2 cells.

(D) ChIP-PCR assays in cultured cells. Hyd does not associate with transcription start site (TSS) of dpp responsive promoters in Clone 8 cells.

Error bars show the standard deviation (mean plus SD) of the ChIP PCR reactions performed in triplicate. * indicates statistically significant difference (P<0.05) calculated by unpaired two-tailed Student’s t-test.

Based on the above results, we made the following conclusions: first, Hyd selectively restrains the maximal transcriptional activity of Ci for specific Hh target genes including dpp and ptc; second, this regulation absolutely requires the E3 ligase activity of Hyd; third, in regulating Hh signaling, Hyd does not act on the proteasome-mediated degradation of its substrate. Our findings strongly suggest that Ci recruits and assembles differential sets of transcriptional cofactors at respective target genes. Further studies are needed to identify the substrate for Hyd and to characterize the detailed process that Hyd is involved in Hh signaling.

Previously, several Ci co-factors have been reported to affect Hh signaling in a target gene-restricted manner. The mediator complex subunits Skuld (Skd) and Kohtalo (Kto) are involved in controlling the transcription of cell affinity-regulating genes, yet not ptc and dpp [32]. As one of the key components of the PAF1 complex, Hyx is a positive regulator of col transcription, yet does not affect ptc, dpp or en [21]. Different target genes have distinctive requirements for the threshold of Hh signaling, with en and col for the high threshold, ptc for the intermediate, and dpp for the lowest. The role of Hyd to specifically restrict the transcriptional outputs of intermediate- to high-threshold target genes in Hh signaling provides a good example toward the question ‘how organisms achieve functional robustness and plasticity for morphogen signaling’. On one hand, the maintenance of a maximal transcriptional potential helps organisms resist to detrimental signaling over-activation in face of genetic and environmental perturbations, such as ectopic Ci activation in this case; on the other hand, differential control of a subset of target genes fine-tunes the sensitivity of cellular responses, conferring cells the flexibility to adapt to developmental dynamics. Altogether, our studies provide more insights into the mechanisms underlying the precise regulation of morphogen signaling.

METHODS

Drosophila Husbandry

All crosses were reared using standard techniques at 25°C. Flies were raised on standard cornmeal-yeast-agar Drosophila medium at 25°C unless otherwise indicated. The EMS allele of hyd¹⁵ was obtained from Bloomington Drosophila Stock Center (BDSC) and the allele was then recombined with P{FRT(whs)}2A P{neoFRT}82B for mosaic clone analysis. Mosaic clones are produced by cross of the recombined null allele with fly stock yw, hsflp; FRT82B Ubi-GFP/TM6B. MARCM clones are produced by crossing recombined null allele with fly stock yw, hsflp, UAS-CD8GFP; TubGal4 FRT82B TubGal80/TM6B. Flipout clones are produced by crossing RNAi with fly stock yw, hsflp/+; Act> y+> Gal4-UAS-CD8GFP/CyO; MKRS/TM6B. Heat shocks were performed for 60 minutes at 37°C at second instar larval stage. RNAi lines targeting hyd, P(GD14227)v44676, P(GD14227)v44675 were obtained from Vienna Drosophila RNAi Center (VDRC). The ptc-lacZ and dpp-lacZ reporters are enhancer trap lines.

Antibody production and purification

Hyd antibodies were generated by immunizing mouse and rabbit with affinity purified GST fusion antigenic fragments (aa570-670) in the vector pGEX-4T-1 from bacterial culture. To obtain purified antibody for CHIP assay, the same antigenic fragment was also subcloned into pET28 for His-tagged protein production. The His-tagged fragment was covalently crosslinked with the AminoLink™ Coupling Resin and Immobilization Kit (44890, Thermo scientific). The His-tagged antigen-coupled resin was then used to affinity purify the rabbit anti-serum against hyd to produce a highly purified polyclonal antibody. The full length cDNA of Colier was cloned into the pGEX-4T-1 and the protein was affinity purified with GST fusion resin from bacterial culture. The polyclonal mouse or rabbit α-Col antibody were prepared and the antibody was purified with the antigenic fragment according to standard protocols.

Immunostaining

Discs were dissected from wandering third instar larvae and immunofluorescence staining was carried out as previously described [33]. Briefly, larvae are dissected in cold PBS (pH7.4) and fixed in 4% formaldehyde in PBS for 15 minutes. Samples were then permeabilized in PBS + 0.1% Triton X-100 and blocked in 5% normal horse serum. Primary antibodies in 5% normal horse serum were incubated overnight at 4°C or 2 hours at RT. Primary antibodies used for the immunostainings were mouse or rabbit α-V5 (Life Technology), rabbit anti-pMad (1:200) (Cell Signaling), rabbit α-Spalt major (1:100), anti-hedgehog (1:40), and chicken α-lacZ (1:100). Mouse anti-patched (DSHB), mouse anti-engrailed (DSHB), rat anti-Ci 2A1 (DSHB) were obtained from Developmental Studies Hybridoma Bank (DSHB), University of Iowa, Department of Biological Sciences. Fluorophore conjugated secondary antibodies were from Jackson Immunoresearch. Confocal images were collected as single frames on a Zeiss LSM 780 confocal microscope. MG132 (100 μM; Calbiochem) in M3 medium (Sigma) was used to treat wing discs for up to 6 hours before immunostaining.

Plasmids and transgenes

Molecular cloning was performed in accordance with standard protocol. The cDNA of hyd is assembled from various fragments by synthetic methods, and detailed cloning procedures will be available on request. The mutation of cyteine to serine at aa2854 was introduced by Quickchange protocol, then substituted with the corresponding C terminal fragment of full length hyd cDNA and sequencing is used to confirm the replacement. The molecular cloning details are available on request. Hyd cDNA (WT) and the CS mutant were then subcloned into pUAST-attB-V5 for generation of transgenic flies according to standard protocol. The transgenic flies, UAS-V5-hyd and UAS-V5-hydC2854S, are generated according to site-directed integration protocol and used for rescue analysis. Ci and its mutant plasmids were described previously[34].

• Hyd differentially regulate hedgehog responsive genes in wing imaginal disc.

• Hyd do not affect the activity and stability of Ci.

• The E3 ligase activity of hyd is required to regulate Ci-dependent transcription.

• Hyd selectively associate with hedgehog responsive promoters.