The Drosophila Hedgehog receptor component Interference hedgehog (Ihog) mediates cell–cell interactions through trans-homophilic binding

Hedgehog (Hh) signaling is crucial for establishing complex cellular patterns in embryonic tissues and maintaining homeostasis in adult organs. In Drosophila, Interference hedgehog (Ihog) or its close paralogue Brother of Ihog (Boi) forms a receptor complex with Patched to mediate intracellular Hh signaling. Ihog proteins (Ihog and Boi) also contribute to cell segregation in wing imaginal discs through an unknown mechanism independent of their role in transducing the Hh signal. Here, we report a molecular mechanism by which the Ihog proteins mediate cell–cell interactions. We found that Ihog proteins are enriched at the site of cell–cell contacts and engage in trans-homophilic interactions in a calcium-independent manner. The region that we identified as mediating the trans-Ihog–Ihog interaction overlaps with the Ihog–Hh interface on the first fibronectin repeat of the extracellular domain of Ihog. We further demonstrate that Hh interferes with Ihog-mediated homophilic interactions by competing for Ihog binding. These results, thus, not only reveal a mechanism for Ihog-mediated cell–cell interactions but also suggest a direct Hh-mediated regulation of both intracellular signaling and cell adhesion through Ihog.

Hedgehog (Hh) signaling is crucial for establishing complex cellular patterns in embryonic tissues and maintaining homeostasis in adult organs. In Drosophila, Interference hedgehog (Ihog) or its close paralogue Brother of Ihog (Boi) forms a receptor complex with Patched to mediate intracellular Hh signaling. Ihog proteins (Ihog and Boi) also contribute to cell segregation in wing imaginal discs through an unknown mechanism independent of their role in transducing the Hh signal. Here, we report a molecular mechanism by which the Ihog proteins mediate cell-cell interactions. We found that Ihog proteins are enriched at the site of cell-cell contacts and engage in transhomophilic interactions in a calcium-independent manner. The region that we identified as mediating the trans-Ihog-Ihog interaction overlaps with the Ihog-Hh interface on the first fibronectin repeat of the extracellular domain of Ihog. We further demonstrate that Hh interferes with Ihog-mediated homophilic interactions by competing for Ihog binding. These results, thus, not only reveal a mechanism for Ihog-mediated cell-cell interactions but also suggest a direct Hh-mediated regulation of both intracellular signaling and cell adhesion through Ihog.
Hedgehog (Hh) 3 signaling is essential for establishing the complex cellular patterns in various embryonic tissues and plays key roles in maintaining adult organ homeostasis. Hh pathway dysfunction during development can cause birth defects in humans, such as holoprosencephaly (1), and postem-bryonic malfunction of this pathway is linked to various proliferative disorders, such as the growth of malignant tumors (2).
The mature Hh ligand is derived from the Hh protein precursor by autoprocessing and lipid modification (3). Pathway activity is triggered by binding of the dually lipidated Hh ligand to Patched (Ptc), a transporter-like protein that, in the absence of Hh, suppresses the activity of Smoothened (Smo). The release of Smo inhibition upon Hh binding to Ptc activates an intracellular signal cascade that stimulates transcriptional activation of pathway target genes (4).
The Drosophila Hh receptor is composed of Interference hedgehog (Ihog) proteins, or the related Brother of Ihog (Boi) proteins, and Ptc. Ihog or Boi is required for Hh reception and stimulation of biological responses as well as for sequestration of Hh to limit long-range signaling (5)(6)(7)(8)(9)(10)(11). Drosophila Ihog proteins are type I single-span transmembrane proteins with four or five extracellular immunoglobulin (Ig) domains and two extracellular fibronectin type III (FNIII) domains; consequently, these proteins resemble cell adhesion molecules (12). Previously, our laboratory identified a function of the Ihog proteins that is independent of their role in transducing the Hh signal (13). Specifically, ectopic Ihog expression leads to aggregation of otherwise nonadherent cells, and loss of Ihog activity in the context of the Drosophila wing disc disrupts cell segregation, even in the presence of downstream genetic rescue of the intracellular response to Hh (13). Considering the structural similarity between Ihog proteins and cell adhesion molecules, we proposed that Ihog proteins function similarly to cell adhesion molecules to directly mediate cell-cell interactions. However, mechanistic understanding of this function, how the dual roles of Ihog proteins (transduction of the Hh signal and cell-cell interaction) are coordinated, and their functional interplay is lacking.
To address these questions, we explored the cell adhesion and homophilic interaction properties of Ihog and the effect of Hh on these properties. We used Drosophila S2 cells, which lack a Hh signal response and are intrinsically nonadhesive, to investigate the properties of ectopically expressed Ihog in cellcell interactions. We found that Ihog proteins are enriched at the site of cell-cell contacts and engaged in calcium-independent homophilic trans-interactions. By mapping the Ihog-Ihog trans-homophilic binding site, we determined that it overlaps with the Ihog-Hh interface on the first fibronectin repeat in the extracellular domain of Ihog. We further demonstrated that Hh interferes with Ihog-mediated trans-homophilic interactions by competing for Ihog binding. These results, thus, not only reveal a molecular mechanism for Ihog-mediated cell-cell interactions but also suggest a direct Hh-mediated regulation of both intracellular signaling and cell adhesion through Ihog.

Ihog proteins concentrate at the site of cell-cell contact and engage in trans-homophilic interactions
The Hh co-receptors Ihog and Boi resemble typical cell adhesion molecules with a single transmembrane domain, four Ig domains, and two FNIII domains (12). Our laboratory reported that the Ihog proteins possess cell adhesion functions independent of their role in transducing the Hh signal (13). To investigate the molecular mechanism, we expressed Ihog in Drosophila S2 cells. Hh ligands and the intracellular Hh mediator Cubitus interruptus (Ci) are absent from S2 cells, and these cells are intrinsically nonadhesive (14 -16), thus enabling the assessment of Ihog-mediated cell-cell interactions without the complication of co-occurring Hh signaling.
Consistent with our previous observations (13), Drosophila S2 cells transiently transfected with Ihog tagged with hemagglutinin (HA) at the C-terminal intracellular side formed multicellular aggregates, whereas the untransfected cells remained mostly dispersed (Fig. 1A). When we stained the cells with an antibody recognizing the extracellular domain (anti-IhogECD) and an antibody recognizing the intracellular HA tag (anti-HA), we found that anti-IhogECD detected Ihog along the outer surface of the multicell aggregates and that anti-HA revealed Ihog at both the outer surface of the clusters and along the contacting cell surfaces (Fig. 1, B and C). We predicted that the extensive contacts formed among the Ihog-expressing cells prevented antibodies from diffusing into the multicell aggregates, thus excluding the anti-IhogECD antibodies from cell-cell contacts (Fig. 1B). The anti-HA staining pattern suggested that Ihog proteins were enriched at cell-cell contacts (Fig. 1C), and quantification of fluorescence intensity in the stained cells confirmed this enrichment (Fig. 1D). We also showed that the enrichment of Ihog proteins at cell-cell contacts was severalfold higher than the enrichment of coexpressed membranelocalized mCD8-GFP proteins (Fig. S1), suggesting that the increased signal of Ihog at the cell-cell contacts was not simply due to membranes of adjacent cells being in close proximity.
Drosophila S2 cells, which are derived from phagocytic hematopoietic cells, lack DE-cadherin at the cell surface and do not form Ca 2ϩ -dependent cell aggregates (17). To exclude the possibility that Ihog caused S2 cell aggregation by indirectly activating or inducing the production of other endogenous adhesion molecules, we stained S2 cells for DE-cadherin, DNcadherin, and fasciclin II (Drosophila neural cell adhesion molecule (NCAM)). As expected, the S2 cells had only background staining for these proteins, and their abundances were unaffected by transfection with Ihog-YFP constructs (Fig. 2, A-C). Furthermore, Ihog-dependent cell aggregation occurred when Ca 2ϩ in the medium was eliminated with the addition of EGTA (Fig. 2, D and E). Thus, the data indicated that the Ihog proteins, in addition to functioning as a Hh co-receptor (5), engage in Ca 2ϩ -independent, homophilic trans-interactions to mediate cell-cell interactions.

The first FNIII domain is essential for Ihog-mediated formation of cell-cell contacts
The Ig (Ig1-4) domains and the FNIII (Fn1 and Fn2) domains in the extracellular portion of Ihog proteins are potentially capable of mediating cell adhesion (12). To map the region in the extracellular portion of Ihog required for the formation of cell-cell contacts, we generated a series of truncations of the extracellular domain of Ihog-YFP, expressed them in S2 cells, and monitored their distribution in live cells using YFP fluorescence. We also monitored the distribution of the small, singletransmembrane-domain protein CD8 tagged with GFP (mCD8-GFP (18)) as a control protein that localizes to the membrane. As expected, mCD8-GFP was found in intracellular structures (likely intermediates along the biosynthetic and trafficking pathway) and along the cell surface (Fig. 3A). The distribution of mCD8-GFP was similar in individual cells, two closely positioned cells, and clusters of multiple cells (Fig. 3A). Similar to the full-length Ihog protein, Ihog truncated at the cytoplasmic C-terminal domain (CTD) was enriched along the contacting sides of cells that both expressed the truncated Ihog (Fig. 3, B and C), indicating that the intracellular domain is not necessary for the enrichment and homophilic interaction. We found that all the extracellular domain-truncated Ihog variants with an intact Fn1 domain were enriched at the sites of cell-cell contact in closely positioned two-cell and multiple-cell aggregates (Fig. 3, C-F). In contrast, Ihog lacking the Fn1 domain was

Ihog proteins mediate trans-homophilic interactions
present along the plasma membrane but did not concentrate at the site of cell-cell contact (Fig. 3, G-I). Similar results were also seen for cells cotransfected with the Ihog paralogue Boi.
We tested the ability of the truncated Ihog variants to induce aggregation of transfected S2 cells. We monitored aggregation induced by coexpression of mCherry with Ihog or Ihog lacking the CTD (Ihog⌬CTD), lacking the Fn2 domain (Ihog⌬Fn2), lacking all four Ig domains (Ihog⌬IG), and lacking the Fn1 domain (Ihog⌬Fn1) (Fig. 4A). Control cells expressed mCherry without any coexpressed Ihog. We dissociated the cells and then quantified the percentage of cell clusters 4 h after mixing (Fig. 4B). Cells expressing wildtype (WT) Ihog protein or any of the Ihog truncated proteins containing the Fn1 domain (Ihog⌬CTD, Ihog⌬Fn2, and Ihog⌬IG) exhibited significant aggregation of the transfected S2 cells (Fig. 4, A and B). In contrast, cells expressing Ihog⌬Fn1 did not form any recognizable cell aggregates and were indistinguishable from those express-ing mCherry alone (Fig. 4, A and B). We confirmed by immunoblotting that the transfected cells used for the aggregation assays had similar amounts of Ihog or Ihog mutant (Fig. 4C).
To test for the Ihog-Ihog interaction biochemically, we assessed whether the extracellular domain of Ihog binds to itself using a pulldown assay. We used only the extracellular domain to avoid any aggregation due to the hydrophobic transmembrane domain or unphysiological interactions between the extracellular domain and the intracellular domain. We expressed variants of the extracellular domain of Ihog tagged with FLAG (FLAG-tagged IhogECD) along with YFP-tagged WT IhogECD in S2 cells and collected the media. We captured FLAG-tagged IhogECD variants with FLAG antibodycoated beads and detected the amount of YFP-tagged IhogECD bound by immunoblotting. Consistent with the cellular data, FLAG-tagged IhogECD variants with the Fn1 domain, but not those with Fn1 deletions, pulled down YFP-tagged IhogECD (Fig. 4D). Thus, S2 cell-based results and the biochemical data indicated that Ihog proteins bind to each other through the Fn1 domain, which serves as the basis for Ihog-mediated homophilic cell-cell interactions.

The Ihog trans-homophilic binding site overlaps with the Ihog-Hh interface
Our data showed that the Ihog Fn1 domain is essential for Ihog-Ihog homophilic interactions, and previous biochemical and structural studies showed that the Ihog Fn1 contributes to the binding site for the N-terminal Hh signaling domain (HhN) in a heparin-dependent manner (7,8). We explored the structural properties that enable the same FNIII domain to possess these two distinct functions of ligand binding and homophilic trans-interaction. We used the IhogFn1-2 (PDB code 2IBB) and the IhogFn1-2-HhN (PDB code 2IBG) complex structures (8) to identify surface residues on the Ihog Fn1 domain. We divided those residues that are nonoverlapping with the Hh-Ihog interface into six major regions, which we called M1-6, and subjected these regions to mutagenesis ( Fig. 5A and Table S1). Based on previous reports (5,8), we also tested mutant Ihog proteins with surface residue substitutions that disrupt Ihog-Hh (Ihog xHh ) or Ihog-Ptc (Ihog xPtc ) interactions, respectively ( Fig. 5A and Table S1).
Because the side chains of the selected residues are solventexposed, we expected that the folded structures of these proteins would not be affected. Indeed, the abundances of the mutant proteins were similar to that of WT Ihog expressed in the S2 cells, and the mutants properly localized to the cell surface of S2 cells (Fig. S2). We expressed mCherry alone or mCherry with WT or the mutated Ihog proteins and assessed their ability to induce cell clustering by quantifying cell cluster formation 4 h after cell dissociation (Fig. 5, B and C). The only region that impaired cell aggregation activity when mutated was the region where Ihog interacts with heparin and mediates the Hh interaction (Ihog xHh ). Furthermore, S2 cells expressing Ihog xHh mutants neither formed clusters with S2 cells expressing WT Ihog proteins (Fig. 6, A-C) nor concentrated at the cell-cell contact sites (Fig. 6D). In contrast, Ihog xPtc was enriched at cell-cell contacts (Fig. 6E), and Ihog xPtc -expressing cells formed clusters at the same proportion as WT Ihog-

Ihog proteins mediate trans-homophilic interactions
expressing cells. Together, these data with the S2 cells indicated that the Ihog homophilic binding site involves the heparinbinding site and overlaps with the Ihog-Hh interface in the Ihog Fn1 domain.

Hh binding and trans-homophilic interactions are simultaneously incompatible for a single Ihog molecule
Intriguingly, our data indicated that the two distinct functions of Ihog proteins involve the same extracellular domain on

Ihog proteins mediate trans-homophilic interactions
an overlapping surface of Ihog (Figs. [3][4][5][6], suggesting that the signal transduction and trans-homophilic binding activities are simultaneously incompatible for a single Ihog molecule. Thus, we predicted that the presence of Hh ligands would compete for the Ihog homophilic binding or vice versa, depending on the relative abundance and affinities of the Fn1 domain for Hh and another Ihog Fn1 domain. Indeed, we found that transient coexpression of HhN by S2 cells reduced the interaction of differentially tagged IhogECD in a coimmunoprecipitation assay (Fig. 7A). A competitive binding assay using purified HhN revealed a concentration-dependent interference between the interactions of differentially tagged IhogECD proteins (Figs. 7B and S3). These results are consistent with competition between homophilic Ihog-Ihog interactions and heterophilic Ihog-Hh interactions.
The published HhN-IhogFn1-2 crystal structure reveals a 2:2 complex, in which each HhN molecule contacts a single Ihog molecule, and a pair of 1:1 Hh-Ihog complexes forms a dimeric 2:2 complex that is entirely mediated by cis-interactions between the Ihog proteins (8). In this complex, Hh does not interfere with the cis-Ihog-Ihog interaction. Consistent with a cis-interaction that is not destabilized by Hh, the HhNmediated disruption of Ihog-Ihog binding was incomplete, even when the concentration of HhN was 10 times higher than the reported dissociation constant of IhogFn1-2 for HhN (8) (Fig. 7B, lane 2).
The Ihog-Ihog interaction that persisted in the presence of excess HhN was likely due to Ihog-Ihog homophilic cis-interactions that were not competed by HhN (8). Furthermore, due to the coexistence of Ihog-Ihog cis-homophilic interactions that are not competed by Hh, in vitro competition assays, such as that used here (Fig. 7), cannot determine the binding affinity of Ihog-Hh versus Ihog-Ihog trans-homophilic interactions. Taken together, we showed that the Hedgehog co-receptor Ihog mediates trans-homophilic binding and signal reception via an overlapping surface, and the presence of excess Hh ligands interferes with Ihog-mediated homophilic interactions by competing for Ihog binding.

Discussion
The Ihog family proteins are type I single-span transmembrane proteins with Ig and FNIII domains, resembling typical cell adhesion molecules in the Ig cell adhesion molecule (Ig-CAM) superfamily. We previously found that Ihog proteins not only play an essential role in Hh signal transduction but also contribute to cell segregation in the Drosophila wing imaginal , which is shown in the same orientation as the front surface representation. The back surface representation is rotated 180°about a vertical axis relative to the front representation. Surface-exposed residues selected for mutagenesis were grouped into differentially colored regions: M1-M6, Ptc-interacting residues (xPtc), and Hh-interacting residues (xHh). See Table S1 for details of mutations. B, S2 cells were transfected with empty vector (Ev; negative control) or plasmids expressing HA-tagged WT or Ihog variants and mCherry. 48 h after transfection, cells were dissociated by trypsin treatment and then mixed for 4 h to allow aggregation to occur. The brightfield channel shows both transfected and untransfected S2 cells. Scale bar, 100 m. C, the aggregation effect from experiments like those shown in B was quantified as the ratio of transfected cells within a cluster to total transfected cells. Each bar shows the mean Ϯ S.D. from n ϭ 30 different images. Unpaired two-tailed t test was used for statistical analysis. n.s., not significant (p Ͼ 0.05); ****, p Ͻ 0.0001. Error bars represent S.D.

Ihog proteins mediate trans-homophilic interactions
discs (5,13). In this study, we showed that, when transfected into the relatively nonadhesive Drosophila S2 cells lacking the ability to transduce the Hh signal, Ihog proteins concentrate at cell-cell contacts and mediate cell-cell interactions in a homophilic, calcium-independent manner.
The region that we identified as mediating the trans-Ihog-Ihog interaction overlaps with the region that mediates the interaction with Hh and includes the region where the nega-tively charged glycan heparin binds. Heparin is required for the Ihog-HhN heterophilic interaction by bridging positively charged patches of both proteins at the interaction interface (8). In the in vitro immunoprecipitation assay and competition assay (Figs. 4 and 7), binding between differentially tagged Ihog ECDs and Hh-mediated disruption of Ihog-Ihog interactions required heparin in the immunoprecipitation buffer (see "Experimental procedures"). In the cell-based assays, heparan sulfate proteoglycans are present because these are naturally produced by the S2 cells (19). Thus, the Ihog-Ihog homophilic trans-interactions likely occurred through heparin-dependent bridging of positively charged surfaces on the two opposing Fn1 domains, in a manner similar to heparin-bridged Ihog-Hh interactions (Fig. S4). These results thus provided a mechanistic basis for the role of Ihog proteins in wing disc cell segregation (13).
Hh release occurs at the apical or basal side of the singlelayered disc epithelium (20 -25), providing Hh at sites where Ihog-Ihog trans-interactions are unlikely to happen. In contrast, the Ihog-Ihog trans-interactions would contribute to cell-cell interactions along the lateral sides of epithelia, which are farther from the source of secreted Hh and less likely affected by the apically or basally secreted Hh. We thus propose that the dominant function of Ihog proteins depends on their subcellular localization and the availability of the Hh ligands (Fig. 8).
Of note, this dual function is not unique to Ihog proteins. Like Ihog, other members of the Ig-CAM family, such as the Netrin receptor Deleted in Colorectal Cancer (DCC), the Slit receptor Robo, and NCAM, have dual roles. These proteins act as "glue" that holds cells together and as molecular sensors to mediate cellular responses, such as motility, proliferation, and survival (26,27). Whereas ligand binding and cell adhesion are often structurally separated involving different extracellular domains (28 -30), the Ihog protein couples two distinct functions within the same region. Further studies are needed to puzzle out the physiological relevance of coupling the two distinct functions of the Ihog proteins in the same region of the protein.
Cdo and Boc are vertebrate homologs of the Drosophila Ihog proteins. Both Cdo and Boc contribute to aspects of Hh signaling (7, 31-33) by binding to mammalian Hh proteins through a nonorthologous FNIII repeat (7,33,34). Cdo and Boc may not mediate cell-cell adhesion (35)(36)(37), based on the lack of cell aggregation when Cdo was overexpressed in the fibroblast-derived cell lines Rat 6 and C1-T24 (35). However, recombinant, soluble fusion proteins that contain the entire Cdo extracellular region coupled to either alkaline phosphatase or the Fc region of human IgG not only bind to the surface of various cell lines (37) but also interact with the extracellular domain of Boc (36). Furthermore, a secreted form of the extracellular domain of Cdo functions as a dominant-negative form of Cdo and inhibits myogenic differentiation, whereas an analogous form of Boc or the full-length Boc promotes such differentiation (36). It is possible that Boc and Cdo interact in a heterophilic complex and that their interactions with themselves or each other may be affected by ligands or other proteins with which they interact. Figure 6. Mutations that disrupt Hh binding also abolish Ihog-mediated trans-homophilic interactions. A, schematic diagram of the mutation sites on the Ihog proteins lacking Hh-or Ptc-binding capacity. The Hh-binding site (Ihog xHh ) and Ptc-binding site (Ihog xPtc ) were mutated separately for analysis. B, S2 cells were transfected with plasmids expressing WT or mutant Ihog along with GFP or mCherry as indicated. Cells were dissociated by trypsin treatment and then mixed to allow aggregation to occur. Scale bar, 100 m. C, the aggregation effect from experiments like those in B was quantified as the ratio of transfected cells within a cluster to total transfected cells. Each bar shows the mean Ϯ S.D. from n ϭ 30 different images. Unpaired two-tailed t test was used for statistical analysis. n.s., not significant, p Ͼ 0.05; ****, p Ͻ 0.0001. D and E, S2 cells transfected with plasmids expressing HA-tagged mutant Ihog xHh or Ihog xPtc were fixed and stained with an HA antibody. Scale bar, 5 m. Error bars represent S.D.

Ihog proteins mediate trans-homophilic interactions
Our study identified a ligand-dependent competitive mechanism for coordinating the biophysical (cell-cell contact) and biochemical (subunit of a signaling receptor) functions of a single protein. Additionally, our data indicate that this competitive mechanism involves just the ligand-receptor interaction without intracellular signal transduction, which is a previously unreported function for Hh. As such, our results and model not only have implications for understanding development in the fly but also for understanding how cells can switch from an adhesive phenotype to a motile phenotype. Because disruption of cell contacts and cell adhesion are involved in physiological tissue repair and pathological tumor invasion and metastasis,

Ihog proteins mediate trans-homophilic interactions
the mechanisms by which Hh regulates cell adhesion through both transcriptional and nontranscriptional (as described here) processes may be relevant in developmental, homeostatic, and pathological contexts.

Constructs
Expression constructs of GFP, mCherry, mCD8-GFP, HhN, Hh, and differentially tagged Ihog variants used in Drosophila cell culture were cloned into the pAcSV expression vector. The IhogECD-3XFLAG and IhogECD-HA-YFP variants were prepared by replacing the Fc tag in the pIB/Fc vector (38) with 3XFLAG tag and HA-YFP tag, respectively.

Antibodies
The following antibodies and dilutions were used: anti-␤tubulin E7 (

Cell culture and transfection
Drosophila S2 cells (Drosophila Genomics Resource Center) were cultured in Drosophila Schneider's medium supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin-glutamine (Thermo Fisher) at 25°C in a humidified incubator. Transfection was performed with FuGENE 6 transfection reagent (Promega).

Cell aggregation assay
S2 cells were transfected separately with plasmids expressing desired proteins. Forty-eight hours after transfection, S2 cells were washed with PBS and dissociated by 0.05% trypsin treatment for 5 min at 25°C. The dissociated cells were resuspended in fresh medium with 10% fetal bovine serum or supplemented with EGTA or purified HhN at concentrations indicated in the figure legends. The resuspended cells were then incubated in 1.5 ml ultra-low-adhesion microcentrifuge tubes with gentle rotation at room temperature for the time indicated in the figure legends. Cells were then transferred into glass-bottom dishes (D35-20-1.5-N, In Vitro Scientific) for live imaging by microscopy. In the experiments involving mixing differentially labeled red and green cells, cells coexpressing GFP or mCherry with the plasmid expressing the protein of interest were counted under microscope and mixed with an equal number of transfected cells prior to incubation with rotation.
To assess cell aggregation, low-magnification fields of similar cell density were randomly taken from each cell aggregation experiment, and the cell clusters were scored if they contained three or more cells. The aggregation effect was quantified as the ratio of certain transfected cells within clusters to total trans-fected cells (both clustered and nonclustered). Each bar shows the mean Ϯ S.D. from 20 -30 different images. Unpaired twotailed t test was used for statistical analysis. Statistical analysis was performed using GraphPad Prism software.

Cell immunostaining and imaging
Forty-eight hours after transfection, dissociated S2 cells were allowed to settle and adhere for 60 min on a glass coverslip. Cells were then washed twice with PBS, fixed in 4% formaldehyde (Ted Pella) in PBS, blocked and permeabilized by 1.5% normal goat serum and 0.1% Triton X-100 in PBS, incubated with primary antibody in PBS containing 1.5% normal goat serum and 0.1% Triton X-100 for 1 h at room temperature, washed three times with 0.1% Triton X-100 in PBS, incubated with secondary antibody with DAPI, and washed with 0.1% Triton X-100 in PBS. Cell surface staining was carried out in the absence of Triton X-100.

MBP-HhN purification
The MBP-HhN expression plasmid was a gift from Dr. Daniel Leahy (The University of Texas at Austin). A DNA fragment encoding the Drosophila melanogaster Hh residues 85-248 (HhN) was cloned into the MBP-HTSHP expression vector, which was modified based on the pMal-c2x vector (New England Biolabs) by including a linker region with various tags (His-TEV-Strep-His-PreScission). Similar to the procedure described previously (8), the fusion proteins were expressed in Escherichia coli strain B834 (DE3) by induction with 1 mM isopropyl 1-thio-␤-D-galactopyranoside overnight at 16°C. Cells were harvested, lysed, and centrifuged, and the supernatant was passed over nickel-nitrilotriacetic acid resin (Qiagen). Proteins were eluted with imidazole according to the manufacturer's suggestions. The elution was then placed into 6000 -8000 molecular weight-cutoff 40-mm dialysis tubing and dialyzed against 20 mM Tris (pH 8.0) and 200 mM NaCl.

Immunoprecipitation
S2 cells were cotransfected with HA-and YFP-tagged IhogECD (residues 1-707, representing the entire extracellular domain of Ihog) and FLAG-tagged IhogECD variants in the presence or absence of Hh as indicated in the figure legend. Media were harvested 48 h post-transfection by centrifuging for 5 min at 500 ϫ g (4°C) to remove cells and then for 15 min at 18,000 ϫ g (4°C) followed by filtering (0.2 m; low protein binding) to remove cell debris. The filtered media were further precleaned by incubating with Protein A/G-Sepharose beads (GE Healthcare) for 3 h. The precleaned media were then incubated overnight at 4°C with EZview TM Red Anti-FLAG M2 Affinity Gel (Sigma) to capture FLAG-tagged protein. Beads were washed with PBS containing 0.1% Tween 20 supplemented with 40 M low-molecular-weight heparin (Sigma). Proteins were recovered directly in SDS-PAGE sample buffer. Proteins samples were resolved by SDS-PAGE and transferred to PVDF membranes (Millipore) for Western blot analysis.

Western blot analysis
Forty-eight hours after transfection, S2 cells were lysed in 1% Nonidet P-40 (50 mM Tris-HCl at pH 6.8, 150 mM NaCl, and Ihog proteins mediate trans-homophilic interactions protease inhibitors) for 30 min at room temperature. The lysate was clarified by centrifugation, and proteins were recovered directly in SDS-PAGE sample buffer. Proteins were separated by SDS-PAGE under reducing conditions and then transferred onto PVDF membranes (Millipore). After protein transfer, the membranes were blocked and then immunostained with primary antibodies and HRP-conjugated secondary antibodies.