Smoothened Oligomerization/Higher Order Clustering in Lipid Rafts Is Essential for High Hedgehog Activity Transduction*

Background: How the activity of Smo on the plasma membrane is regulated is not fully understood. Results: Smo forms oligomers/higher order clusters in lipid rafts of cell plasma membrane to transduce high level Hh signaling. Conclusion: Both lipid rafts and the oligomerization/higher order clustering of Smo are critical for high Hh activity transduction. Signification: Lipid rafts are essential for Smo activation and Hh gradient signal transduction. The Hedgehog (Hh) signaling pathway plays evolutionarily conserved roles in controlling embryonic development and tissue homeostasis, and its dysregulation has been implicated in many human diseases including congenital disorder and cancer. The Hh pathway has a unique signal reception system that includes two membrane proteins, the receptor Patched (Ptc) and the transducer Smoothened (Smo). In the Hh signaling cascade, Smo plays a critical role in controlling transduction of Hh gradient signal from the outside into the inside of cells. Although the Smo downstream signal transduction has been intensively studied, the mechanism by which Smo on the plasma membrane is regulated has not been fully understood. As a specific membrane structure of metazoan cells, lipid rafts act as a platform to regulate signal transduction by forming a nanoscale cluster through protein-protein or protein-lipid interactions. However, it remains largely unknown whether lipid rafts are also involved in the regulation of Hh signal transduction. Here, we show that Smo extracellular domain (N terminus) and transmembrane domains form oligomers/higher order clusters in response to Hh signal. Furthermore, we identify that lipid rafts on the plasma membrane are essential for high level activity of Smo during the Hh signal transduction. Finally, our observation suggests that oligomerization/higher order clustering of Smo C-terminal cytoplasmic tail (C-tail) is essential for the transduction of high level Hh signal. Collectively, our data support that in response to Hh gradient signals, Smo transduces high level Hh signal by forming oligomers/higher order clusters in the lipid rafts of cell plasma membrane.


The Hedgehog (Hh) signaling pathway plays evolutionarily conserved roles in controlling embryonic development and tissue homeostasis, and its dysregulation has been implicated in many human diseases including congenital disorder and cancer. The Hh pathway has a unique signal reception system that includes two membrane proteins, the receptor Patched (Ptc) and the transducer Smoothened (Smo). In the Hh signaling cascade, Smo plays a critical role in controlling transduction of Hh gradient signal from the outside into the inside of cells. Although the Smo downstream signal transduction
has been intensively studied, the mechanism by which Smo on the plasma membrane is regulated has not been fully understood. As a specific membrane structure of metazoan cells, lipid rafts act as a platform to regulate signal transduction by forming a nanoscale cluster through protein-protein or protein-lipid interactions. However, it remains largely unknown whether lipid rafts are also involved in the regulation of Hh signal transduction. Here, we show that Smo extracellular domain (N terminus) and transmembrane domains form oligomers/higher order clusters in response to Hh signal. Furthermore, we identify that lipid rafts on the plasma membrane are essential for high level activity of Smo during the Hh signal transduction. Finally, our observation suggests that oligomerization/higher order clustering of Smo C-terminal cytoplasmic tail (C-tail) is essential for the transduction of high level Hh signal. Collectively, our data support that in response to Hh gradient signals, Smo transduces high level Hh signal by forming oligomers/higher order clusters in the lipid rafts of cell plasma membrane.
Hedgehog (Hh) 6 signaling is essential for organ patterning of both invertebrates and vertebrates (1,2). Malfunction of the Hh signaling pathway leads to numerous human congenital disorders and cancers (3). The core components and regulatory mechanisms of the Hh signaling are conserved among invertebrates and vertebrates with few exceptions (1,4). In Drosophila wing discs, Hh proteins secreted by posterior (P) compartment cells move into the anterior (A) compartment to form a local concentration gradient. Genetic and biochemical studies revealed that two multipass transmembrane proteins, Patched (Ptc, 12-transmembrane domain protein) and Smoothened (Smo, 7-transmembrane domain protein), function as a reception system for signal transduction in Hh-receiving cells (5).
Correlative studies in Drosophila have revealed several important steps in the regulation of Smo activity and the Hh signaling (3,6). In the absence of Hh, Ptc inhibits Smo activity, thereby blocking the Hh signal transduction. Under this condition, Smo plasma membrane accumulation is prevented and its C-terminal cytoplasmic tail (C-tail) assumes a "closed" inactive conformation, whereas its extracellular N terminus forms a constitutive dimer. At the same time, full-length Cubitus interruptus (Ci), the transcription factor of Hh pathway, is sequentially phosphorylated by protein kinase A, glycogen synthase kinase 3, and casein kinase I, and processed to generate Ci75, to block downstream gene expression as a transcriptional repressor (7,8). In the presence of Hh, the Hh ligand physically inter-acts with Ptc and relieves its inhibition on Smo, resulting in Smo accumulation on plasma membrane and its C-tail conformational switch to an "open" active dimer form, which regulates distinct downstream target gene expression in a Hh concentration-dependent manner through controlling Ci nuclei translocation (3, 9 -13). Low levels of Hh are sufficient to induce the expression of decapentaplegic (dpp), whereas middle and high levels of Hh are required to induce patched (ptc) and engrailed (en) expression, respectively (10,14).
In the Hh signaling pathway, the mechanism by which Ci leads to downstream target gene expression is understood in general (2,3). However, the regulation mechanism of Smo on the plasma membrane is poorly defined. Lipid rafts are known as a spatiotemporally regulated dynamic lipid-dependent segregation of specific membrane components. They range from unstable nanoscale clusters of protein to relatively stable microdomains formed from preexisting small clusters through protein-protein or protein-lipid interactions, mediating signal transduction as platforms on the plasma membrane. For example, lipid rafts have been found to aid protein dimerization during certain signal transductions (15)(16)(17)(18). Given that Smo dimerization is essential for the Hh signal transduction across the plasma membrane (6,9), it is interesting to study whether Smo accumulates in lipid rafts to form dimers and even oligomers/higher order clusters in the presence of Hh.
In the present study, we demonstrated that the extracellular and transmembrane domains of Smo form oligomers/higher order clusters in the presence of Hh. Our results also suggested that the closer proximity of Smo intracellular subdomain 3 (third internal loop) is triggered by Hh. Importantly, we identified that lipid rafts on the plasma membrane are essential for Smo activity induced by high level of Hh. Furthermore, we showed that oligomerization of Smo C-tail is essential for high level Hh signaling transduction.  (9). For constructs of Myr-FLAG-CCm/-CCd/-CCt-SmoCT, CCm, CCd, or CCt (19 -21) were inserted downstream of myristoylation signal (Myr-) (amino acid sequence: MGNKCCSKRQ) and FLAG tag, upstream of the Smo C-tail.

EXPERIMENTAL PROCEDURES
Fly Stocks-Drosophila strains used in this study were maintained under standard conditions. The yw strain was used as host for all of the P-element mediated transformations. MS1096, Ci-Gal4, Hh-Gal4, dpp-LacZ, ptc-LacZ, FLAG-SmoWT, FLAG-SmoSD and HA-Ci were described before (9,11). Myr-FLAG-CCm/-CCd/-CCt-SmoCT transgenic flies were generated following standard protocols.
Cell Culture, Transfection, Immunoprecipitation, and Western Blotting-S2 cells were cultured in Schneider's Drosophila Medium (Invitrogen) and transfected with Lipofectamine 2000 (Invitrogen) under standard conditions and protocol (13). To detect the Myc-Smo⌬C-EphB2CT phosphorylation, S2 cells were transfected with the indicated constructs. After 48-h transfection, S2 cells were harvested, and the cell lysate was treated with the phosphatase, LAR (Sigma), a protein-tyrosine phosphatase, for 30 min at 30°C. The sample was immunoprecipitated with anti-Myc antibody and subjected to standard SDS-PAGE and Western blotting with anti-Myc and 4G10 antibodies. To detect the oligomerization of Smo, S2 cells were transfected with the indicated constructs. The cell lysates were treated as described (22), and then Smo signals were detected by Western blotting with anti-Myc antibody. Immunoprecipitation, Western blotting, and the native gel electrophoresis were performed following standard protocols (22,23).
Fluorescence Resonance Energy Transfer (FRET) Analysis-FRET was performed as described previously (9,24). Briefly, S2 cells were transfected with the indicated constructs. Cells were then treated and mounted as described. Fluorescence signals were acquired with Leica TCS SP5 Confocal microscope. The CFP signal was obtained once before (BP) and once after (AP) photobleaching YFP at the top half of each cell, leaving the bottom half of each cell as the internal control. Each data set was based on Ͼ30 individual cells. Three to five interest areas in the bleached or unbleached half of cell were selected for analysis. The intensity change of CFP was analyzed using Leica FRET AB Wizard software. FRET efficiency was calculated using the formula: FRET% ϭ ((CFP AP Ϫ CFP BP )/CFP AP ) ϫ 100. Statistical significance was determined using Student's t test.
Gradient Centrifugation-The centrifugation was conducted as reported previously (17,25) and modified for Drosophila cells. Specifically, S2 cells were transfected with Myc-Smo and HA-dG q constructs and treated with or without Hh. The cell lysates were centrifuged at 10,000 ϫ g for 45min at 4°C. The membrane pellet was resolved in the lysis buffer, loaded onto 1.12 M sucrose cushion in the same lysis buffer, centrifuged at 100,000 ϫ g for 2 h at 4°C. The fraction of 1 ml at the interface, which contained the plasma membrane, was collected and mixed with 1 ml of 1 M Na 2 CO 3 (pH 11.0) and 2 ml of 90% sucrose in 50 mM MES, 0.3 M NaCl (pH 6.5). The mixture was placed at the bottom of the 12.5-ml tube, overlaid with 3 ml of 35%, 4 ml of 21%, 1 ml of 5% sucrose solution in 0.25 M Na 2 CO 3 (final pH 11.0), 50 mM MES, 0.3 M NaCl. Percentage sucrose here is standard for weight/volume. The samples were then centrifuged at 200,000 ϫ g for 20 h at 4°C. Twenty-four fractions were collected for detection.
Luciferase Assay-S2 cells cultured in a 24-well plate were co-transfected with 250 ng of Gal4, 250 ng of Ci, 250 ng of ptc-luciferase reporter, 25 ng of Renilla, 250 ng of the indicated constructs, and 250 ng HhN or pUAST in each well. After transfection for 48 h, dual-luciferase measurements were performed in triplicate using the Dual-Glo TM luciferase assay system (Promega). Statistical significance was determined using Student's t test.
Lipid Raft Disruption-Lipid raft stability was disrupted by perturbation with exogenous polyunsaturated fatty acid (eicosapentaenoic acid (20:5(n-3), PUFA)) (Sigma). Stearic acid (18:0) (SA) (Sigma) was used as the negative control as recommended (26). For S2 cells, after transfection for 12 h, cell culture medium was replaced with serum-free one supplemented with 0.4% BSA and 50 M PUFA or SA. Cells were cultured for another 48 h following detection. For in vivo assay, the dissected discs from third instar larvae were cultured in serum-free M3 Medium with 5% BSA and 50 M PUFA or SA for 6 h at 25°C. After treatment, the immunostaining was followed.
Cell Surface Immunostaining and Z Section Image Processing-Cell surface staining was described before (9). The Z sections were scanned under Leica Confocal microscope, x-y images were scanned in the z axis at 0.25-m intervals for three-dimensional reconstruction. The Z plane sections, Z plane stacks, and videos were analyzed and exported using Leica software.

Smo Extracellular Domains Form Oligomers/Higher Order Clusters in the Presence of Hh-Regardless of Hh, the Smo N
termini form a constitutive dimer, and its C-tail has a conformational switch between closed and open states (6,9). To understand the molecular mechanism by which Smo activity is regulated on the plasma membrane, we sought to investigate whether the extracellular domains of Smo form oligomers/ higher order clusters in the presence of Hh. First, we used antibody to label the endogenous Smo in Drosophila wing imaginal disc. As shown in Fig. 1, A-AЉ, in the A compartment of wing disc where Hh signal was absent, Smo is located in the cytoplasm. However, in the P compartment of wing disc where high level Hh was present, Smo accumulated on the plasma membrane and formed clusters.
Phosphorylation of Smo C-tail is very important for its activity and localization and could be mimicked by SmoSD mutations, in which three protein kinase A sites (Ser 667 , Ser 687 , and Ser 740 ) and adjacent casein kinase I sites are mutated to Asp (9,11). We found that overexpressed SmoSD also aggregated and formed clusters on the cell membrane, as wild-type Smo upon Hh treatment (Fig. 1, B-BЉ).
Smo belongs to G protein-coupled receptor (GPCR) superfamily (27). Certain studies have suggested that GPCRs might form oligomeric structures for signal transduction (28,29 EphB2 is a receptor tyrosine kinase. As part of cell signaling machinery, the ligand binding-induced N-terminal dimerization/higher order clustering of EphB2 has been correlated with its C-terminal autophosphorylation level (32). To further verify our notion, we generated a fusion protein, Smo-EphB2CT, in which the Smo C-tail was replaced with the kinase tail of EphB2. The tyrosine phosphorylation levels of the fusion protein were then detected by 4G10, a specific tyrosine phosphorylation antibody. As shown in Fig. 1D, Myc-Smo⌬C-EphB2CT was  MAY 3, 2013 • VOLUME 288 • NUMBER 18 JOURNAL OF BIOLOGICAL CHEMISTRY 12607 autophosphorylated, which indicated a constitutive dimer of the Smo N terminus, consistent with a previous finding (9). After treated with Hh, the detected tyrosine phosphorylation level was dramatically increased (Fig. 1D). Moreover, this effect could be blocked if the lysate was treated with a specific protein-tyrosine phosphatase, LAR, before immunoprecipitation and the gel electrophoresis. Taken together, these observations suggest that in response to Hh, the Smo N terminus/transmembrane domains (TMs)-mediated constitutive dimer may further aggregate to form oligomers/higher order clusters.

Smo Forms Oligomers/Higher Order Clusters in Lipid Rafts
Hh Consistent with the previous finding that Smo C-tails form a dimer upon Hh treatment (9), we observed FRET between SmoCFP C and SmoYFP C changed from ϳ1.4% to ϳ12.3% in S2 cells upon Hh treatment (Fig. 2, A and B) (Fig. 2, A and B), and FRET between SmoCFP N and SmoYFP N increased from ϳ11.5% to ϳ16.3% (Fig. 2B). FRET between negative control pairs (CFP/YFP) was Ͻ0.5% regardless of Hh (Fig. 2B). These observations strongly indicate that Hh may induce close proximity of Smo N terminus and TMs during Smo oligomerization.
We next performed FRET analysis using phosphorylationmimicking mutants SmoSDCFP L3 and SmoSDYFP L3 . As shown in Fig. 2,A and B 2C). The wild-type and mutant Smo were subject to native gel and immunoprecipitation analysis (Fig. 2, D and E). Without the Smo N terminus and TMs, SmoCT could not form dimer/ oligomers (Fig. 2E). The tetramer of Smo⌬N⌬C or Smo⌬C was obviously observed in the lysates (Fig. 2E). To further confirm this finding, different tagged wild-type or truncated Smo were co-expressed in S2 cells. After being immunoprecipitated with FLAG, the samples were detected by anti-FLAG or anti-Myc antibodies (Fig. 2D). The result suggested the indicated Smo, beside SmoCT, could form dimer/oligomer. Taking these results and those from FRET experiments (Fig. 2B), it suggested that both the N terminus and the TMs play an important role in Smo oligomer/higher order cluster formation.
Lipid Rafts Are Essential for Smo Activation Induced by High Level Hh-As aforementioned, lipid rafts may function as a signaling platform to mediate signal transduction, especially those involving membrane protein dimerization/oligomerization (15,18). To determine whether lipid rafts play a role in the regulation of Hh signal transduction, specifically their effect on Smo activation, we performed gradient centrifugation assay. As shown in Fig. 3A, not like lipid raft marker, dG q , Smo was isolated in low and middle density insoluble membrane fractions only in the presence of Hh. This finding suggested that Smo locates in lipid rafts on the plasma membrane in the presence of Hh.
PUFA can disrupt lipid rafts on the membrane and inhibit related signal transduction, whereas the noninhibitory fatty acid SA (18:0) does not affect the lipid raft structure (26,34). Indeed, we found that PUFA but not SA treatment could dramatically decrease ptc reporter luciferase activity induced by Hh as well as SmoSD (Fig. 3B). However, PUFA treatment could not fully inhibit the activity of SmoSD or Smo stimulated by Hh (see more below). Together with the results shown in Fig. 3A, these observations indicated that lipid rafts are important in the Hh signal transduction from the outside into the inside of cell.
Given that Smo C-tail forms dimer (9), we further investigated whether lipid rafts are important for the oligomerization and activation of Smo. We found that PUFA decreased the FRET induced by Hh between Smo C termini (SmoCFP C / SmoYFP C ), yet SA did not (Fig. 3C and data not shown). Meanwhile, the FRET induced by Hh between SmoCFP L3 and SmoYFP L3 was blocked by PUFA (Fig. 3, C and D, and data not shown). Smo clustering on the plasma membrane induced by Hh was also reduced by PUFA although Smo membrane accumulation was not affected (Fig. 3, E-HЉ   To further investigate whether Smo locates in lipid rafts in vivo, we overexpressed Smo with Ci-Gal4 or Hh-Gal4 in the A or P compartment, respectively. We then examined whether Smo could co-localize with cholera toxin B subunit (CTB), which has been widely used as a marker of membrane lipid rafts (35). As shown in Fig. 4, A-Bٞ, we found that Smo locates in the cytosol and does not co-localize with CTB in the absence of Hh (A compartment). However, Smo co-localizes with CTB in the presence of Hh (P compartment) (Fig. 4, C-Dٞ). This finding suggested that Smo relocates to lipid rafts only in the presence of Hh.
Because PUFA treatment decreased ptc reporter activity induced by Hh (Fig. 3B), we asked whether it could down-regulate Hh pathway activity in vivo. To address this question, we  it could turn on dpp expression. However, the dpp expression level was not affected by PUFA (data not shown). We further investigated whether lipid rafts play an important role in high Hh activity transduction. SmoSD has been used to mimic highly activated Smo in response to high level of Hh signal. When overexpressed in wing disc and treated with SA, SmoSD cannot only turn on ptc expression, which responds to the middle level of Hh, but also en expression, which responds to high Hh level (Fig. 5, A-Aٞ and C-Cٞ). After being treated by PUFA, the ptc expression level had no significant alteration (Fig. 5, A and B). However, the en expression level decreased dramatically by PUFA treatment (Fig. 5, C and D). The magnified image demonstrated that PUFA treatment decreased clustering level of SmoSD in vivo (Fig. 5, E, EЈ, F, and FЈ). However, PUFA did not reduce the en expression level induced by overexpressing downstream component of Hh pathway, HA-tagged wild-type full-length Ci (Fig. 5, G, GЈ, H, and HЈ). In addition, Western blotting results suggested that PUFA treatment reduced the oligomerization level of Smo in the presence of Hh (Fig. 5, I and  J). These results suggested that PUFA treatment mainly affects the high level Hh activity by reducing Smo oligomerization/ higher order clustering on the cell membrane. Combining these result with the observation in Figs. 3, B and C, and 4, D-Dٞ, our results indicated that lipid raft localization and oligomerization/higher order clustering of Smo on the plasma membrane are essential for transduction of high level Hh activity.  Fig. 6A, right). A coiled-coil dimerization motif from yeast GCN4 (21) (referred to as CCd), and its dimerization-deficient version (referred to as CCm) or its tetramerization version (referred to as CCt) (19,20) Fig. 6, D, G, J, and M). These results suggested that, along with Smo C-tail oligomerization ranging from monomer, dimer to high order cluster, Smo activates the Hh pathway in a gradient manner in vivo. It appears that Smo C-tail oligomerization/higher order clustering is essential for transduction of high level Hh signaling.

DISCUSSION
It is generally accepted that Smo transduces the Hh signaling via intracellular complexes consisting of many proteins such as Fu/Cos2 and Ci and that Smo activation is regulated via a cytoplasmic conformational switch as well as its extracellular loops (9, 13, 36 -38 A graded signal mediated by morphogen, including the Hh graded signal, has been proposed to regulate differential gene expressions in a concentration-dependent manner (3,39). However, as a fundamental aspect of the Hh signaling pathway, how Smo activity on the plasma membrane is regulated by Hh gradient has not been fully addressed. Drosophila wing discs provide an excellent model system to study the regulatory mechanism of the Hh signal transduction. In wing discs, Hh signal molecules are secreted from the P compartment cells and form a signal concentration gradient in the A compartment cells adjacent to the A/P boundary, which has a high level of Hh, to induce the expression of downstream target genes, including dpp, ptc, and en. Because Smo has a critical positive role in the Hh signaling pathway, its activity must be tightly controlled by the concentration of Hh to ensure proper strength and duration of the Hh signaling. Previous studies have proposed a model by which the Smo N terminus forms a constitutive dimer and its C-tails have a conformational change to transduce Hh gradient information (6,9). However, it is still unknown whether Smo forms oligomers/higher order clusters in responding to Hh gradient. In this study, we demonstrated that the N terminus and the TMs contribute to the oligomerization/higher order clustering of Smo on the cell plasma membrane. Meanwhile, our data suggested that the N terminus and the TMs of Smo may induce its C-tails to form oligomers/higher order clusters for transducing the information of differential Hh concentration.
To test this hypothesis, we mimicked Smo C-tail monomer, dimer, and tetramer (oligomer) on the plasma membrane by expressing Smo mutants (Myr-FLAG-CCm-SmoCT, Myr-FLAG-CCd-SmoCT, or Myr-FLAG-CCt-SmoCT), respectively. We found that, by mimicking N terminus oligomers/ higher order clusters, oligomerized Smo C-tails exhibit higher level of Hh signaling activity than dimerized or monomer one. We concluded that the oligomerization/higher order clustering of the Smo C-tail is essential for high Hh activity transduction.
Lipid rafts play a critical role in mediating certain signal transductions by acting as signaling and oligomerization platforms (15)(16)(17)(18)35). However, the functional relevance of lipid rafts regarding Hh signaling has not been studied. In this work, we identified an important role of lipid rafts in Hh pathway. We found that Smo locates in lipid rafts on the plasma membrane in the presence of Hh, implying that lipid rafts are involved in  regulating Hh signal transduction. Given that the disruption of lipid rafts down-regulates high level Hh pathway activity and decreases the oligomers/higher order clusters of Smo on the plasma membrane, lipid rafts appear to play a critical role in transducing high level Hh activity by acting as a platform for Smo oligomers/higher order clusters. In fact, lipid rafts function not only in cell signaling and but also in protein trafficking (35). As a future direction, it would be intriguing to investigate their potential function in regulating Smo trafficking in the presence of Hh. In addition, the functional conservation of Smo in development of vertebrate and invertebrate warrants further studies into the importance of lipid rafts in mammalian Hh signaling.